From singer@apple.com Fri Mar 4 15:56:29 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 510203A6921 for ; Fri, 4 Mar 2011 15:56:29 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -106.599 X-Spam-Level: X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=0.001, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id skrPX-iT3ZnQ for ; Fri, 4 Mar 2011 15:56:28 -0800 (PST) Received: from mail-out3.apple.com (mail-out.apple.com [17.254.13.22]) by core3.amsl.com (Postfix) with ESMTP id 0DC953A68A9 for ; Fri, 4 Mar 2011 15:56:28 -0800 (PST) Received: from relay15.apple.com (relay15.apple.com [17.128.113.54]) by mail-out3.apple.com (Postfix) with ESMTP id 1205AD437C21 for ; Fri, 4 Mar 2011 15:57:38 -0800 (PST) X-AuditID: 11807136-b7c6bae000004a34-f5-4d717c719581 Received: from singda.apple.com (singda.apple.com [17.197.32.11]) by relay15.apple.com (Apple SCV relay) with SMTP id 78.EA.18996.17C717D4; Fri, 4 Mar 2011 15:57:38 -0800 (PST) From: David Singer Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable Date: Fri, 4 Mar 2011 15:57:37 -0800 References: <20110304235407.070953A6A17@core3.amsl.com> To: DISPATCH list Message-Id: Mime-Version: 1.0 (Apple Message framework v1082) X-Mailer: Apple Mail (2.1082) X-Brightmail-Tracker: AAAAAA== Subject: [dispatch] Fwd: New Version Notification for draft-gellens-mime-bucket-bis-03 X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 04 Mar 2011 23:56:29 -0000 Friends this is a minor update to the previous mime-bucket-bis, which in turn is = a minor update to RFC 4281. It adds a profiles parameter in addition to = the codecs parameter, and tries to make it clearer what files these = parameters apply to. We'll be trying to finish this revision sometime soon, so, if you would = like to, please review and comment soon! Thanks Begin forwarded message: > From: IETF I-D Submission Tool > Date: March 4, 2011 15:54:07 PST > To: singer@apple.com > Cc: randy@qualcomm.com, Per.Frojdh@ericsson.com > Subject: New Version Notification for draft-gellens-mime-bucket-bis-03 >=20 >=20 > A new version of I-D, draft-gellens-mime-bucket-bis-03.txt has been = successfully submitted by David Singer and posted to the IETF = repository. >=20 > Filename: draft-gellens-mime-bucket-bis > Revision: 03 > Title: The Codecs and Profiles Parameters for "Bucket" = Media Types > Creation_date: 2011-03-04 > WG ID: Independent Submission > Number_of_pages: 23 >=20 > Abstract: > Several MIME type/subtype combinations exist that can contain > different media formats. A receiving agent thus needs to examine the > details of such media content to determine if the specific elements > can be rendered given an available set of codecs. Especially when > the end system has limited resources, or the connection to the end > system has limited bandwidth, it is helpful to know from the Content- > Type alone if the content can be rendered. >=20 > This document specifies two parameters, "codecs" and "profiles", > which are used with various MIME types or type/subtype combinations > to allow for unambiguous specification of the codecs and/or profiles > employed by the media formats contained within. >=20 > By labeling content with the specific codecs indicated to render the > contained media, receiving systems can determine if the codecs are > supported by the end system, and if not, can take appropriate action > (such as rejecting the content, sending notification of the > situation, transcoding the content to a supported type, fetching and > installing the required codecs, further inspection to determine if it > will be sufficient to support a subset of the indicated codecs, etc.) >=20 > Similarly, the profiles can provide an overall indication, to the > receiver, of the specifications with which the content complies. The > receiver may be able to work out the extent to which it can handle > and render the content by examining to see which of the declared > profiles it supports, and what they mean. >=20 >=20 >=20 > The IETF Secretariat. >=20 >=20 David Singer Multimedia and Software Standards, Apple Inc. From henry.sinnreich@gmail.com Fri Mar 4 16:17:52 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 673D43A6A01 for ; Fri, 4 Mar 2011 16:17:52 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -2.918 X-Spam-Level: X-Spam-Status: No, score=-2.918 tagged_above=-999 required=5 tests=[AWL=0.681, BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id j46bJLTyCwgg for ; Fri, 4 Mar 2011 16:17:51 -0800 (PST) Received: from mail-gx0-f182.google.com (mail-gx0-f182.google.com [209.85.161.182]) by core3.amsl.com (Postfix) with ESMTP id 66E7A3A692D for ; Fri, 4 Mar 2011 16:17:51 -0800 (PST) Received: by gxk7 with SMTP id 7so1340482gxk.27 for ; Fri, 04 Mar 2011 16:19:01 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=kxf4iL7QBGLtmSR7cG+U5OeKx2AX0z/3Xyy1d4mtmL8=; b=XEMZjQH5rbwMHbN6SHKNlLjd2wpxkHDYJzbhZVoM8vgYoidyIUOZ/oou5FXsr1ebXx aiisu6Q2qlkGk++acm7pV1NEk2WLR6NAA5B7C4QvqthGbof7OiYkCTARs3nVbORmT0sQ pbD0dyI85PHWzSkIUPBKR2r5H1Zoin/o12TrM= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=Po76X/GLmkF6tfdAkRhPw/DFXyqjUCWIaR93el1zUANhXAHxk3XnRZiMD0DgYjcAwv RyR/x/jVF+a309t4JhgVXOxAhJDnNIndWg5DKDGT/O/o29rJS5U5xrBujogSWUJuGekJ MyXQGGqEJMnckuHC/8avRZxn7bYTzfIng/eog= Received: by 10.90.244.3 with SMTP id r3mr1868984agh.170.1299284339383; Fri, 04 Mar 2011 16:18:59 -0800 (PST) Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id 35sm3560999ano.11.2011.03.04.16.18.55 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 04 Mar 2011 16:18:57 -0800 (PST) User-Agent: Microsoft-Entourage/12.28.0.101117 Date: Fri, 04 Mar 2011 18:18:53 -0600 From: Henry Sinnreich To: David Singer , DISPATCH list Message-ID: Thread-Topic: [dispatch] Fwd: New Version Notification for draft-gellens-mime-bucket-bis-03 Thread-Index: AcvayuiNzxR9KNAWakyxI9AJp5/qCg== In-Reply-To: Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Subject: Re: [dispatch] Fwd: New Version Notification for draft-gellens-mime-bucket-bis-03 X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 05 Mar 2011 00:17:52 -0000 This I-D offers IMO a most interesting alternative to the SIP/SDP codec negotiation and needless to say is working very well. Just wondering (naively no doubt) what may also be applicable to the media component for the RTC-Web. Disclaimer: Just another rant against using SDP in the signaling component. Thanks, Henry On 3/4/11 5:57 PM, "David Singer" wrote: > Friends > > this is a minor update to the previous mime-bucket-bis, which in turn is a > minor update to RFC 4281. It adds a profiles parameter in addition to the > codecs parameter, and tries to make it clearer what files these parameters > apply to. > > We'll be trying to finish this revision sometime soon, so, if you would like > to, please review and comment soon! Thanks > > > Begin forwarded message: > >> From: IETF I-D Submission Tool >> Date: March 4, 2011 15:54:07 PST >> To: singer@apple.com >> Cc: randy@qualcomm.com, Per.Frojdh@ericsson.com >> Subject: New Version Notification for draft-gellens-mime-bucket-bis-03 >> >> >> A new version of I-D, draft-gellens-mime-bucket-bis-03.txt has been >> successfully submitted by David Singer and posted to the IETF repository. >> >> Filename: draft-gellens-mime-bucket-bis >> Revision: 03 >> Title: The Codecs and Profiles Parameters for "Bucket" Media Types >> Creation_date: 2011-03-04 >> WG ID: Independent Submission >> Number_of_pages: 23 >> >> Abstract: >> Several MIME type/subtype combinations exist that can contain >> different media formats. A receiving agent thus needs to examine the >> details of such media content to determine if the specific elements >> can be rendered given an available set of codecs. Especially when >> the end system has limited resources, or the connection to the end >> system has limited bandwidth, it is helpful to know from the Content- >> Type alone if the content can be rendered. >> >> This document specifies two parameters, "codecs" and "profiles", >> which are used with various MIME types or type/subtype combinations >> to allow for unambiguous specification of the codecs and/or profiles >> employed by the media formats contained within. >> >> By labeling content with the specific codecs indicated to render the >> contained media, receiving systems can determine if the codecs are >> supported by the end system, and if not, can take appropriate action >> (such as rejecting the content, sending notification of the >> situation, transcoding the content to a supported type, fetching and >> installing the required codecs, further inspection to determine if it >> will be sufficient to support a subset of the indicated codecs, etc.) >> >> Similarly, the profiles can provide an overall indication, to the >> receiver, of the specifications with which the content complies. The >> receiver may be able to work out the extent to which it can handle >> and render the content by examining to see which of the declared >> profiles it supports, and what they mean. >> >> >> >> The IETF Secretariat. >> >> > > David Singer > Multimedia and Software Standards, Apple Inc. > > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch From magnus.westerlund@ericsson.com Mon Mar 7 02:20:15 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 512AD3A695C for ; Mon, 7 Mar 2011 02:20:15 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -106.418 X-Spam-Level: X-Spam-Status: No, score=-106.418 tagged_above=-999 required=5 tests=[AWL=0.181, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UfydJKeAtlEt for ; Mon, 7 Mar 2011 02:20:14 -0800 (PST) Received: from mailgw10.se.ericsson.net (mailgw10.se.ericsson.net [193.180.251.61]) by core3.amsl.com (Postfix) with ESMTP id B251C3A6950 for ; Mon, 7 Mar 2011 02:20:13 -0800 (PST) X-AuditID: c1b4fb3d-b7bbbae000005311-27-4d74b1a6400d Received: from esessmw0247.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw10.se.ericsson.net (Symantec Mail Security) with SMTP id D6.AF.21265.6A1B47D4; Mon, 7 Mar 2011 11:21:26 +0100 (CET) Received: from [147.214.183.8] (153.88.115.8) by esessmw0247.eemea.ericsson.se (153.88.115.94) with Microsoft SMTP Server id 8.2.234.1; Mon, 7 Mar 2011 11:21:24 +0100 Message-ID: <4D74B1A4.9050309@ericsson.com> Date: Mon, 7 Mar 2011 11:21:24 +0100 From: Magnus Westerlund User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; sv-SE; rv:1.9.2.14) Gecko/20110221 Thunderbird/3.1.8 MIME-Version: 1.0 To: DISPATCH list X-Enigmail-Version: 1.1.1 Content-Type: multipart/mixed; boundary="------------020101000203080901090702" X-Brightmail-Tracker: AAAAAA== Subject: [dispatch] Fwd: I-D Action:draft-perkins-rtcweb-rtp-usage-00.txt X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 10:20:15 -0000 --------------020101000203080901090702 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 8bit Hi, Colin Perkins, Joerg Ott and I have created this draft about RTC-WEB's usage of RTP. Feedback and discussion are most appreciated. Cheers Magnus Westerlund ---------------------------------------------------------------------- Multimedia Technologies, Ericsson Research EAB/TVM ---------------------------------------------------------------------- Ericsson AB | Phone +46 10 7148287 Färögatan 6 | Mobile +46 73 0949079 SE-164 80 Stockholm, Sweden| mailto: magnus.westerlund@ericsson.com ---------------------------------------------------------------------- --------------020101000203080901090702 Content-Type: message/rfc822; name="I-D Action:draft-perkins-rtcweb-rtp-usage-00_txt.eml" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="I-D Action:draft-perkins-rtcweb-rtp-usage-00_txt.eml" Received: from mailgw7.ericsson.se (153.88.115.8) by esessmw0184.eemea.ericsson.se (153.88.115.83) with Microsoft SMTP Server id 8.2.234.1; Mon, 7 Mar 2011 11:16:27 +0100 Received: from mail.ietf.org (mail.ietf.org [64.170.98.32]) by mailgw7.ericsson.se (Symantec Mail Security) with SMTP id 54.25.25859.A70B47D4; Mon, 7 Mar 2011 11:16:27 +0100 (CET) Received: from [127.0.0.1] (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E1D2B3A6948; Mon, 7 Mar 2011 02:15:03 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5F0A63A6937 for ; Mon, 7 Mar 2011 02:15:02 -0800 (PST) Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7SgcUQfgzcnG for ; Mon, 7 Mar 2011 02:15:01 -0800 (PST) Received: from [127.0.0.1] (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 724543A67DF for ; Mon, 7 Mar 2011 02:15:01 -0800 (PST) From: "Internet-Drafts@ietf.org" To: "i-d-announce@ietf.org" Sender: "i-d-announce-bounces@ietf.org" Date: Mon, 7 Mar 2011 11:15:01 +0100 Subject: I-D Action:draft-perkins-rtcweb-rtp-usage-00.txt Thread-Topic: I-D Action:draft-perkins-rtcweb-rtp-usage-00.txt Thread-Index: AcvcsLiDV0OucwdxS4C4sIsDk69kTw== Message-ID: <20110307101501.31346.62344.idtracker@localhost> List-Help: List-Subscribe: , List-Unsubscribe: , Reply-To: "internet-drafts@ietf.org" Accept-Language: en-US Content-Language: en-US X-MS-Exchange-Organization-AuthAs: Anonymous X-MS-Exchange-Organization-AuthSource: esessmw0184.eemea.ericsson.se X-MS-Has-Attach: yes X-Auto-Response-Suppress: All X-MS-TNEF-Correlator: x-auditid: c1b4fb30-b7ba2ae000006503-65-4d74b07a6de3 x-brightmail-tracker: AAAAAA== errors-to: i-d-announce-bounces@ietf.org x-virus-scanned: amavisd-new at amsl.com x-mailman-version: 2.1.9 list-post: list-archive: list-id: Internet Draft Announcements only x-beenthere: i-d-announce@ietf.org x-spam-level: x-spam-status: No, score=-102.547 tagged_above=-999 required=5 tests=[AWL=0.052, BAYES_00=-2.599, USER_IN_WHITELIST=-100] x-spam-score: -102.547 x-spam-flag: NO delivered-to: i-d-announce@core3.amsl.com x-original-to: i-d-announce@core3.amsl.com Content-Type: multipart/mixed; boundary="_003_201103071015013134662344idtrackerlocalhost_" MIME-Version: 1.0 --_003_201103071015013134662344idtrackerlocalhost_ Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable A New Internet-Draft is available from the on-line Internet-Drafts director= ies. Title : RTP Requirements for RTC-Web Author(s) : C. Perkins, et al. Filename : draft-perkins-rtcweb-rtp-usage-00.txt Pages : 16 Date : 2011-03-07 This document discusses usage of RTP in the context of RTC-WEB work. It discusses important factors of RTP to consider by other parts of the solution, it also discusses which RTP profile to support and which RTP extensions that should be supported. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-perkins-rtcweb-rtp-usage-00.txt Internet-Drafts are also available by anonymous FTP at: ftp://ftp.ietf.org/internet-drafts/ Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --_003_201103071015013134662344idtrackerlocalhost_ Content-Type: text/plain; charset="us-ascii"; name="draft-perkins-rtcweb-rtp-usage-00.txt" Content-Description: draft-perkins-rtcweb-rtp-usage-00.txt Content-Disposition: attachment; filename="draft-perkins-rtcweb-rtp-usage-00.txt"; size=162; creation-date="Mon, 07 Mar 2011 11:16:27 GMT"; modification-date="Mon, 07 Mar 2011 11:16:27 GMT" Content-Transfer-Encoding: base64 RklMRSBRVUFSQU5USU5FRA0KDQpNaWNyb3NvZnQgRm9yZWZyb250IFNlY3VyaXR5IGZvciBFeGNo YW5nZSBTZXJ2ZXIgcmVtb3ZlZCBhIGZpbGUgc2luY2UgaXQgd2FzIGZvdW5kIHRvIG1hdGNoIGEg ZmlsdGVyLg0KRmlsZSBuYW1lOiAiZHJhZnQtcGVya2lucy1ydGN3ZWItcnRwLXVzYWdlLTAwLnVy bCINCkZpbHRlciBuYW1lOiAiRklMRSBGSUxURVI9IHVubmFtZWQ6ICoudXJsIg0K --_003_201103071015013134662344idtrackerlocalhost_ Content-Type: text/plain; name="ATT00001..txt" Content-Description: ATT00001..txt Content-Disposition: attachment; filename="ATT00001..txt"; size=322; creation-date="Mon, 07 Mar 2011 11:16:27 GMT"; modification-date="Mon, 07 Mar 2011 11:16:27 GMT" Content-Transfer-Encoding: base64 X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18NCkktRC1Bbm5v dW5jZSBtYWlsaW5nIGxpc3QNCkktRC1Bbm5vdW5jZUBpZXRmLm9yZw0KaHR0cHM6Ly93d3cuaWV0 Zi5vcmcvbWFpbG1hbi9saXN0aW5mby9pLWQtYW5ub3VuY2UNCkludGVybmV0LURyYWZ0IGRpcmVj dG9yaWVzOiBodHRwOi8vd3d3LmlldGYub3JnL3NoYWRvdy5odG1sDQpvciBmdHA6Ly9mdHAuaWV0 Zi5vcmcvaWV0Zi8xc2hhZG93LXNpdGVzLnR4dA0K --_003_201103071015013134662344idtrackerlocalhost_-- --------------020101000203080901090702-- From ian_elz@yahoo.co.uk Mon Mar 7 03:49:51 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D7EF03A6944 for ; Mon, 7 Mar 2011 03:49:51 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -1.109 X-Spam-Level: X-Spam-Status: No, score=-1.109 tagged_above=-999 required=5 tests=[BAYES_05=-1.11, HTML_MESSAGE=0.001] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fUU4ydiT+doK for ; Mon, 7 Mar 2011 03:49:51 -0800 (PST) Received: from nm13-vm1.bullet.mail.ird.yahoo.com (nm13-vm1.bullet.mail.ird.yahoo.com [77.238.189.209]) by core3.amsl.com (Postfix) with SMTP id AC7473A63C9 for ; Mon, 7 Mar 2011 03:49:50 -0800 (PST) Received: from [77.238.189.234] by nm13.bullet.mail.ird.yahoo.com with NNFMP; 07 Mar 2011 11:51:03 -0000 Received: from [212.82.108.247] by tm15.bullet.mail.ird.yahoo.com with NNFMP; 07 Mar 2011 11:51:03 -0000 Received: from [127.0.0.1] by omp1012.mail.ird.yahoo.com with NNFMP; 07 Mar 2011 11:51:03 -0000 X-Yahoo-Newman-Property: ymail-5 X-Yahoo-Newman-Id: 634187.49718.bm@omp1012.mail.ird.yahoo.com Received: (qmail 20741 invoked by uid 60001); 7 Mar 2011 11:51:03 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.co.uk; s=s1024; t=1299498663; bh=84Atr0DbjSt0cnnjRXvXncLbet0MfCh+rvUiq2tj74E=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=B6mpYzUb+DtOgZQKxzwNF05qjpEopKdNj+IIDKpSClHs/xmzkc5VBhmLPVpw1/CyPjgL/vmfdDVKa4mrbS5JehX+GuG8REwYdYMuE8ojiB8FLB1caNcMhzFdLzClRG0wKj9nw+MmZJodPt331Vp9t5r/Fop31O3ABiYWdJxedns= DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=U/H2WBdtjH8HxgTX0WH9sEnBK4KfFxoXxTX1cRAr2Wv5cZnBFJ+5eVwTI7BjM764wnrkA5neIOTVvPwMCHO+htw7Hign+686S2XzQzcJT8HMYGvQ9mG9pwosfNlTpfaxuvjPAR2u2sM4ifHYhWPUD4/RRIYYmJmIkLWMrtU7ayQ=; Message-ID: <517621.17498.qm@web29103.mail.ird.yahoo.com> X-YMail-OSG: yIaVwLIVM1n5e0TxYP._LvnPO1tZZ5MbHBqEijqWj1e46Xv Djj9vEEzhkDpYvRThhN7nTX7Nsz5GFUeKbGghScsvjvGWqEMHnLSUiaVKz0k 7QvscuEFDXJOMitPne7fUr5BrZK0_LQnbrhpDwyBuA8QoqGZhZ7_vOz3PH0m G73kb3NYN6KzAfQjt5JkNcb1QhxaxraqTTHtis7usPLeXFsrVvUCMyeum270 5O74- Received: from [80.231.29.53] by web29103.mail.ird.yahoo.com via HTTP; Mon, 07 Mar 2011 11:51:03 GMT X-Mailer: YahooMailRC/559 YahooMailWebService/0.8.109.292656 Date: Mon, 7 Mar 2011 11:51:03 +0000 (GMT) From: Ian Elz To: dispatch@ietf.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-644207344-1299498663=:17498" Subject: [dispatch] New Draft - draft-elz-dispatch-req-sip-3rdpartyprivacy-00 X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 11:49:52 -0000 --0-644207344-1299498663=:17498 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable All,=0A=0AI have submitted a new draft relating to third party privacy in S= IP messages.=0A=0AThere are two specific issues included, how to specify th= ird party privacy in a =0ASIP message; the precedence of privacy settings b= etween Privacy header values =0Aand Privacy URI parameters.=0A=0AThe things= to consider are:=0A=0AShould this work be pursued and if so in which forum= .=0A=0AA suitable solution to meet the requirements.=0A=0AHow any changes s= hould be implemented: a new draft which modifies RFC3323 or an =0Aupdate to= RFC3323 which includes the additional work.=0A=0AAll comments welcome.=0A= =0ASorry I forgot to write the Abstract in version -00. I will include this= in the =0Anext version. ( I leave abstracts to last and overlooked this be= fore =0Asubmission.)=0A=0AIan Elz=0A=0A=0A=0A --0-644207344-1299498663=:17498 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable
All,

I have submitted a new draft relating to third party pr= ivacy in SIP messages.

There are two specific issues included, how t= o specify third party privacy in a SIP message; the precedence of privacy s= ettings between Privacy header values and Privacy URI parameters.

Th= e things to consider are:

Should this work be pursued and if so in w= hich forum.

A suitable solution to meet the requirements.

How= any changes should be implemented: a new draft which modifies RFC3323 or a= n update to RFC3323 which includes the additional work.

All comments= welcome.

Sorry I forgot to write the Abstract in version -00. I wil= l include this in the next version. ( I leave abstracts to last and overloo= ked this before submission.)

Ian Elz

=0A=0A=0A= =0A --0-644207344-1299498663=:17498-- From christer.holmberg@ericsson.com Mon Mar 7 10:11:46 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C5A573A680A for ; Mon, 7 Mar 2011 10:11:46 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -6.476 X-Spam-Level: X-Spam-Status: No, score=-6.476 tagged_above=-999 required=5 tests=[AWL=-0.104, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, SARE_SUB_OBFU_Q1=0.227] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2HvpXJf9tZDg for ; Mon, 7 Mar 2011 10:11:39 -0800 (PST) Received: from mailgw9.se.ericsson.net (mailgw9.se.ericsson.net [193.180.251.57]) by core3.amsl.com (Postfix) with ESMTP id 544E03A67DA for ; Mon, 7 Mar 2011 10:11:38 -0800 (PST) X-AuditID: c1b4fb39-b7c6dae0000023f2-39-4d7520234703 Received: from esessmw0237.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw9.se.ericsson.net (Symantec Mail Security) with SMTP id F1.C4.09202.320257D4; Mon, 7 Mar 2011 19:12:51 +0100 (CET) Received: from ESESSCMS0356.eemea.ericsson.se ([169.254.1.30]) by esessmw0237.eemea.ericsson.se ([153.88.115.90]) with mapi; Mon, 7 Mar 2011 19:12:50 +0100 From: Christer Holmberg To: "dispatch@ietf.org" Date: Mon, 7 Mar 2011 19:12:50 +0100 Thread-Topic: Draft new: draft-holmberg-rtcweb-ucreqs-00 (Web Real-Time Communication Use-cases and Requirements) Thread-Index: Acvc8zpusbRq2Zd8SSSGSTsT3UivlQ== Message-ID: <7F2072F1E0DE894DA4B517B93C6A05851948F0B2F9@ESESSCMS0356.eemea.ericsson.se> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-Brightmail-Tracker: AAAAAA== Subject: [dispatch] Draft new: draft-holmberg-rtcweb-ucreqs-00 (Web Real-Time Communication Use-cases and Requirements) X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 18:11:47 -0000 Hi browser lovers,=20 We've submitted a use-case/requirement draft, draft-holmberg-rtcweb-ucreqs-= 00.txt, that we think could be used as a base when those discussions start.= =20 The document describes some use-cases, and based on those proposes browser = requirement, and application-browser API requirements. It focues on media r= elated issue. Ie issues related to privacy, signalling between the browser = and web server etc, are currenly not considered.=20 Best regards,=20 Christer From vkg@bell-labs.com Mon Mar 7 14:23:52 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1ED563A6964 for ; Mon, 7 Mar 2011 14:23:52 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -106.599 X-Spam-Level: X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MceWpCLxRFmd for ; Mon, 7 Mar 2011 14:23:50 -0800 (PST) Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id D8CEA3A694D for ; Mon, 7 Mar 2011 14:23:49 -0800 (PST) Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p27MP2q9016003 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Mon, 7 Mar 2011 16:25:02 -0600 (CST) Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p27MP2Ch017650 for ; Mon, 7 Mar 2011 16:25:02 -0600 (CST) Message-ID: <4D755B9E.6030008@bell-labs.com> Date: Mon, 07 Mar 2011 16:26:38 -0600 From: "Vijay K. Gurbani" Organization: Bell Laboratories, Alcatel-Lucent User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7 MIME-Version: 1.0 To: "dispatch@ietf.org" Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35 Subject: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 22:23:52 -0000 Hello: The work going on in the SOC working group has highlighted the need for some guidelines or strategies for SIP load balancing. RjS had started some preliminary work on this way back in 2003 [1], but that effort seems not to have proceeded further. Furthermore, our collective knowledge on SIP has increased since 2003 such that some of the mechanisms discussed in the early work may no longer be directly applicable (e.g., use of 503 and resulting problems with overload, modifying route-sets may impinge on signing the route set [2], etc.) My co-author (Thierry Bessis) and I have put together a short draft on load balancing that I have just submitted to the IETF archives [3]. It will be good to have some folks take a look at it and suggest improvements or to point out any shortcomings, of which I am sure there will be many. If there is interest, we can discuss this in the dispatch working group in Prague. [1] http://tools.ietf.org/html/draft-sparks-sipping-load-00 [2] http://tools.ietf.org/html/rfc5658 [3] http://tools.ietf.org/html/draft-bessis-dispatch-adaptive-load-balancing-00 Thanks, - vijay -- Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA) Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com Web: http://ect.bell-labs.com/who/vkg/ From pkyzivat@cisco.com Mon Mar 7 15:12:57 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 942C93A6833 for ; Mon, 7 Mar 2011 15:12:57 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -110.529 X-Spam-Level: X-Spam-Status: No, score=-110.529 tagged_above=-999 required=5 tests=[AWL=0.070, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id guide4qei8ze for ; Mon, 7 Mar 2011 15:12:56 -0800 (PST) Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87]) by core3.amsl.com (Postfix) with ESMTP id 814903A6809 for ; Mon, 7 Mar 2011 15:12:56 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=2119; q=dns/txt; s=iport; t=1299539650; x=1300749250; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=u3KfBvqd45BD0GLW20YJGfiqTUlOlrPc9zGxBg7E23A=; b=eGASv1pqcjr6zuuKcZySQY0/W2P3UQHzxTzw4Y3uNC0VTUrPz97ci75x uBlazKWcnmpuPpqG+6GcwlP1AuCSjfWCvhU/AxwNTQpv51NlOXb4ITbDo 7sEZJwVx+XhNgGxzAn51FGzXDL3j7IeqjTLvd0Ap+xv6E7DGoXT9PXgVe E=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AjUIAAn2dE2tJV2a/2dsb2JhbACYT44IdKIcnCCFYgSFHIcUg0M X-IronPort-AV: E=Sophos;i="4.62,279,1297036800"; d="scan'208";a="341947330" Received: from rcdn-core-3.cisco.com ([173.37.93.154]) by sj-iport-5.cisco.com with ESMTP; 07 Mar 2011 23:14:10 +0000 Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rcdn-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p27NE9Hn031139 for ; Mon, 7 Mar 2011 23:14:10 GMT Message-ID: <4D7566C1.40605@cisco.com> Date: Mon, 07 Mar 2011 18:14:09 -0500 From: Paul Kyzivat User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7 MIME-Version: 1.0 To: dispatch@ietf.org References: <4D755B9E.6030008@bell-labs.com> In-Reply-To: <4D755B9E.6030008@bell-labs.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 23:12:57 -0000 Vijay, IIUC, the algorithm you have proposed is similar to the one Paul Jones has proposed - with the main difference being the numeric range of the weight. Do I have that right? It isn't clear to me from your draft if you are specifying the algorithm by which the relative weights are used to determine the probability of a request going to one server or another. ISTM that it *does* matter. And it especially matters if there is an implication that the upstream node's algorithm may depend on some assumption or knowledge about the downstream nodes. And of course this method of reporting weights isn't sufficient to drive a balancing algorithm that is sensitive to some requests requiring more or different resources than others. (That seems to be a real problem, though I don't know if we have a real solution for it.) Thanks, Paul On 3/7/2011 5:26 PM, Vijay K. Gurbani wrote: > Hello: The work going on in the SOC working group has highlighted > the need for some guidelines or strategies for SIP load balancing. > > RjS had started some preliminary work on this way back in 2003 [1], > but that effort seems not to have proceeded further. Furthermore, > our collective knowledge on SIP has increased since 2003 such that > some of the mechanisms discussed in the early work may no longer be > directly applicable (e.g., use of 503 and resulting problems with > overload, modifying route-sets may impinge on signing the route set > [2], etc.) > > My co-author (Thierry Bessis) and I have put together a short draft > on load balancing that I have just submitted to the IETF > archives [3]. It will be good to have some folks take a look at it > and suggest improvements or to point out any shortcomings, of which > I am sure there will be many. > > If there is interest, we can discuss this in the dispatch working > group in Prague. > > [1] http://tools.ietf.org/html/draft-sparks-sipping-load-00 > [2] http://tools.ietf.org/html/rfc5658 > [3] > http://tools.ietf.org/html/draft-bessis-dispatch-adaptive-load-balancing-00 > > Thanks, > > - vijay From vkg@bell-labs.com Mon Mar 7 15:34:55 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 62DF828C122 for ; Mon, 7 Mar 2011 15:34:55 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -106.599 X-Spam-Level: X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qEQel18H6KA7 for ; Mon, 7 Mar 2011 15:34:54 -0800 (PST) Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 1BC5028C137 for ; Mon, 7 Mar 2011 15:34:54 -0800 (PST) Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p27Na7ZF014636 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Mon, 7 Mar 2011 17:36:07 -0600 (CST) Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p27Na78c004683 for ; Mon, 7 Mar 2011 17:36:07 -0600 (CST) Message-ID: <4D756C47.1000509@bell-labs.com> Date: Mon, 07 Mar 2011 17:37:43 -0600 From: "Vijay K. Gurbani" Organization: Bell Laboratories, Alcatel-Lucent User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7 MIME-Version: 1.0 To: dispatch@ietf.org References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> In-Reply-To: <4D7566C1.40605@cisco.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37 Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2011 23:34:55 -0000 Paul: Thank you for your input. More inline. On 03/07/2011 05:14 PM, Paul Kyzivat wrote: > Vijay, > IIUC, the algorithm you have proposed is similar to the one Paul > Jones has proposed - with the main difference being the numeric range > of the weight. Do I have that right? I believe that Paul J.'s algorithm is more suited towards overload control, not load balancing (and I will await Paul J.'s correction to my assertion). The intent in draft-bessis-dispatch-adaptive-load-balancing is to *always* report the current load to give the upstream SIP proxy an indication of the health of the SIP server in the cluster. If one waits until overload is imminent, then it may be too late since a sudden and unanticipated burst of messages will put the SIP server in overload rather quickly. > It isn't clear to me from your draft if you are specifying the > algorithm by which the relative weights are used to determine the > probability of a request going to one server or another. Yes, same way as RFC2782 does it. > ISTM that it *does* matter. And it especially matters if there is an > implication that the upstream node's algorithm may depend on some > assumption or knowledge about the downstream nodes. No knowledge, just a reported number that the upstream SIP proxy can treat as a RFC2782-like weight indicator. Today, rfc3263-resolution starts off with the SRV indicators in RFC2782, without any regard to additional knowledge or assumption about downstream nodes. dispatch-adaptive-load-balancing is trying to use the same model, except render the weight to be sensitive to the current traffic in the cluster in real-time. > And of course this method of reporting weights isn't sufficient to > drive a balancing algorithm that is sensitive to some requests > requiring more or different resources than others. (That seems to be > a real problem, though I don't know if we have a real solution for > it.) Yes, that is a real problem, as is the problem of a SIP server in a cluster that has a low weight and is thus not contacted with enough frequency to merit more participation to the goodput. It is well known in SIP that some methods are more onerous to process than others --- from a proxying point of view, processing an INVITE with forking and response aggregation is far more detrimental to resource usage than is processing a BYE. Some existing literature techniques thus add a coefficient to the weight of sending an INVITE versus a BYE. At this time, the process outlined in dispatch-adaptive-load-balancing treats all requests the same, but we can definitely look at ways to change this behaviour. Thanks, - vijay -- Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA) Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com Web: http://ect.bell-labs.com/who/vkg/ From paulej@packetizer.com Mon Mar 7 21:04:13 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7B25A3A68D4 for ; Mon, 7 Mar 2011 21:04:13 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -2.599 X-Spam-Level: X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ecMV5oI0mH6e for ; Mon, 7 Mar 2011 21:04:11 -0800 (PST) Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id AB2853A6847 for ; Mon, 7 Mar 2011 21:04:10 -0800 (PST) Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2855GqP001481 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Tue, 8 Mar 2011 00:05:22 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299560722; bh=Hgkule4Oj0XoxVZgvOb7ZPqNb+E90wcuInevj2MuJwc=; h=From:To:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type:Content-Transfer-Encoding; b=sR33r5HSQkRyo8ZkzJ8wVRIXOj3TeK3pWIkNG22sEhbijiZSitMItP1JaQXpwbq/f WzPSnzwg7qNhWIdFAIMCceWviRPxhCW7RSYUwQF5C+3x5ldc/C5cnRxDBDtjjQX27K 5aarTZRa88qEzumaSIqXkWZ2UyJAuFB6M5g8DsTA= From: "Paul E. Jones" To: "'Vijay K. Gurbani'" , References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> In-Reply-To: <4D756C47.1000509@bell-labs.com> Date: Tue, 8 Mar 2011 00:05:13 -0500 Message-ID: <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Outlook 14.0 Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3mTxRFo0A== Content-language: en-us Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 05:04:13 -0000 Vijay, The Session Capacity Estimate (SCE) idea is actually quite similar to the "adaptive load balancing" draft, from what I can see. SCE was proposed as a solution to the "overload" problem, the approach being one where we try to avoid overload by selecting a device that reports a higher capacity value. Thus, one could view it as "overload control through load balancing". We didn't explicitly call it load balancing, though, because we were not trying to balance load (i.e., place the same load on all servers). Rather, we just wanted to direct traffic to devices that we knew had available capacity. In practice, the way SCE would be employed is the following: * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from peer devices * The selector would direct traffic toward the peer devices in a round-robin fashion, taking into consideration the SCE value so as to prevent exhausting resources on any given peer device * Once the peer's SCE value dropped below a given "low-water mark", the peer device would be less favored and not given traffic until the SCE value increases above the low-water mark (OPTIONS "ping" messages may be used to determine the SCE value in the event that there is no other traffic) * The current SOC overload control mechanism would be used to to prevent message overloading, as SCE only focuses on resource overloading (e.g., memory, CPU, disk space, DSPs) Note that the "round-robin" approach may or may not be preferred and is not required in our draft. I've found this approach works pretty well, but one may prefer to use a particular weighting that is more complex. I don't think local decisions in the box need to be subject to standardization: we just need to standardize the signaling to enable such decisions. In any case, it certainly looks like we're thinking along the same lines. Paul > -----Original Message----- > From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On > Behalf Of Vijay K. Gurbani > Sent: Monday, March 07, 2011 6:38 PM > To: dispatch@ietf.org > Subject: Re: [dispatch] SIP load balancing > > Paul: Thank you for your input. More inline. > > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: > > Vijay, > > IIUC, the algorithm you have proposed is similar to the one Paul Jones > > has proposed - with the main difference being the numeric range of the > > weight. Do I have that right? > > I believe that Paul J.'s algorithm is more suited towards overload > control, not load balancing (and I will await Paul J.'s correction to my > assertion). > > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy an > indication of the health of the SIP server in the cluster. If one waits > until overload is imminent, then it may be too late since a sudden and > unanticipated burst of messages will put the SIP server in overload > rather quickly. > > > It isn't clear to me from your draft if you are specifying the > > algorithm by which the relative weights are used to determine the > > probability of a request going to one server or another. > > Yes, same way as RFC2782 does it. > > > ISTM that it *does* matter. And it especially matters if there is an > > implication that the upstream node's algorithm may depend on some > > assumption or knowledge about the downstream nodes. > > No knowledge, just a reported number that the upstream SIP proxy can > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution > starts off with the SRV indicators in RFC2782, without any regard to > additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model, except > render the weight to be sensitive to the current traffic in the cluster > in real-time. > > > And of course this method of reporting weights isn't sufficient to > > drive a balancing algorithm that is sensitive to some requests > > requiring more or different resources than others. (That seems to be a > > real problem, though I don't know if we have a real solution for > > it.) > > Yes, that is a real problem, as is the problem of a SIP server in a > cluster that has a low weight and is thus not contacted with enough > frequency to merit more participation to the goodput. > > It is well known in SIP that some methods are more onerous to process > than others --- from a proxying point of view, processing an INVITE with > forking and response aggregation is far more detrimental to resource > usage than is processing a BYE. Some existing literature techniques > thus add a coefficient to the weight of sending an INVITE versus a BYE. > At this time, the process outlined in dispatch-adaptive-load-balancing > treats all requests the same, but we can definitely look at ways to > change this behaviour. > > Thanks, > > - vijay > -- > Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent > 1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA) > Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com > Web: http://ect.bell-labs.com/who/vkg/ > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch From ian_elz@yahoo.co.uk Tue Mar 8 00:14:32 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0246C3A6907 for ; Tue, 8 Mar 2011 00:14:31 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -1.854 X-Spam-Level: X-Spam-Status: No, score=-1.854 tagged_above=-999 required=5 tests=[AWL=0.744, BAYES_00=-2.599, HTML_MESSAGE=0.001] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XSooVxf+QP0M for ; Tue, 8 Mar 2011 00:14:30 -0800 (PST) Received: from nm16-vm1.bullet.mail.ird.yahoo.com (nm16-vm1.bullet.mail.ird.yahoo.com [77.238.189.88]) by core3.amsl.com (Postfix) with SMTP id 23A633A6915 for ; Tue, 8 Mar 2011 00:14:29 -0800 (PST) Received: from [77.238.189.49] by nm16.bullet.mail.ird.yahoo.com with NNFMP; 08 Mar 2011 08:15:41 -0000 Received: from [212.82.108.253] by tm2.bullet.mail.ird.yahoo.com with NNFMP; 08 Mar 2011 08:15:41 -0000 Received: from [127.0.0.1] by omp1018.mail.ird.yahoo.com with NNFMP; 08 Mar 2011 08:15:41 -0000 X-Yahoo-Newman-Property: ymail-5 X-Yahoo-Newman-Id: 382227.80135.bm@omp1018.mail.ird.yahoo.com Received: (qmail 77433 invoked by uid 60001); 8 Mar 2011 08:15:41 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.co.uk; s=s1024; t=1299572141; bh=KlykqCIb34Iudup+djCzxi4jX0YXV7reNlipqkoIfpA=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=oNX3txFqZFB0meEkzJ3LIlHZZNzoM+9UVS+iOb4WnWSmoDKmncNc3XwVRndqnqTAg3sa0g2cCs6O0sspuR4u5tiSZLUBvf9i/N7XRwHdPo9igR2+4XD8f0Ah5ysigEofd6Pujl2tXfSYUQwnFC2/+kk+jTWfBtsmW5kWOHFySDw= DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type; b=ylxvRJgEDRzbzU/3NpWexgi24reHWqAWDptHoE8qxOzfVeUH/bJdnXXZ1eFkihyLvUd0Hd5ttefwtkKlYcY6DW31kshZs5TIz+Ip0GhwBCeUCfqigU6trXJ7dstbj0s/rJrE/XSkDoMxYoydAUCJ4UvlA/FqzcwaahON2swUVWU=; Message-ID: <294181.75599.qm@web29112.mail.ird.yahoo.com> X-YMail-OSG: Lr3DwTcVM1nYrhNEuw4ZQhiY7Rj9iPap5vo6ftO7Z0a3Hlx e9jdvUBH3XXQuAMwnEZ9_MF2ZfaRQG5m2U4Y0o0ngfhCBjaHVcMAnJUtM2Mh bpoArg3hRA9nD0W6D51DFZMdSYzHYl5NSVH2dmwk7MqxrDpe8oKGE2Vy6OmC pej0F.bDvDmn45PiT02ww9RX02NYCp_UK05czWgtrV1aQfGbBM4CEy3Sedxy pVEusTdyczYQ9_P2qJUgDUzkqRVAqwaYqsUVl9oBxj0c- Received: from [80.231.29.53] by web29112.mail.ird.yahoo.com via HTTP; Tue, 08 Mar 2011 08:15:41 GMT X-Mailer: YahooMailRC/559 YahooMailWebService/0.8.109.292656 Date: Tue, 8 Mar 2011 08:15:41 +0000 (GMT) From: Ian Elz To: dispatch@ietf.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-1477614842-1299572141=:75599" Subject: [dispatch] New Draft - draft-elz-dispatch-req-sip-3rdpartyprivacy-00 X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 08:14:32 -0000 --0-1477614842-1299572141=:75599 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable All,=0A=0AI have submitted a new draft relating to third party privacy in S= IP messages.=0A=0Ahttp://www.ietf.org/id/draft-elz-dispatch-req-sip-3rdpart= yprivacy-00.txt (Link =0Athat I forgot in the previous posting.)=0A=0ATher= e are two specific issues included, how to specify third party privacy in a= =0ASIP message; the precedence of privacy settings between Privacy header = values =0Aand Privacy URI parameters.=0A=0AThe things to consider are:=0A= =0AShould this work be pursued and if so in which forum.=0A=0AA suitable so= lution to meet the requirements.=0A=0AHow any changes should be implemented= : a new draft which modifies RFC3323 or an =0Aupdate to RFC3323 which inclu= des the additional work.=0A=0AAll comments welcome.=0A=0ASorry I forgot to = write the Abstract in version -00. I will include this in the =0Anext versi= on. ( I leave abstracts to last and overlooked this before =0Asubmission.)= =0A=0AIan Elz=0A=0A=0A --0-1477614842-1299572141=:75599 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable
All,

I have submitted a new draft relating to third party pr= ivacy in SIP messages.

http://www.ietf.org/id/draft-elz-di= spatch-req-sip-3rdpartyprivacy-00.txt  (Link that I forgot in the = previous posting.)

There=0Aare two specific issues included, how to = specify third party privacy in=0Aa SIP message; the precedence of privacy s= ettings between Privacy=0Aheader values and Privacy URI parameters.

= The things to consider are:

Should this work be pursued and if so in= which forum.

A suitable solution to meet the requirements.

H= ow=0Aany changes should be implemented: a new draft which modifies RFC3323= =0Aor an update to RFC3323 which includes the additional work.

All c= omments welcome.

Sorry=0AI forgot to write the Abstract in version -= 00. I will include this in=0Athe next version. ( I leave abstracts to last = and overlooked this=0Abefore submission.)

Ian Elz
=0A

= =0A=0A=0A=0A --0-1477614842-1299572141=:75599-- From partr@cisco.com Tue Mar 8 01:51:02 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 79D753A67B8 for ; Tue, 8 Mar 2011 01:51:02 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -10.1 X-Spam-Level: X-Spam-Status: No, score=-10.1 tagged_above=-999 required=5 tests=[AWL=0.499, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9CFEHBHf-ZKK for ; Tue, 8 Mar 2011 01:51:01 -0800 (PST) Received: from sj-iport-3.cisco.com (sj-iport-3.cisco.com [171.71.176.72]) by core3.amsl.com (Postfix) with ESMTP id 35A4F3A6810 for ; Tue, 8 Mar 2011 01:51:01 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=7334; q=dns/txt; s=iport; t=1299577936; x=1300787536; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to; bh=z0Czl//ZpCk9WpaQC42Lp7WM0UgbM8O6BbL7YJYSOAE=; b=LfOIotPnj4gAkZWXVTdorL+qs/OiFh3vOHLgFfRF6qW4OTz2/Bc1Hn1E +HxLyrSCV1jISD5Uv1YJzMjjYQQfDaWnNfzCPRMPplFmRM/a0nCdN7Q41 fG8FOcnQzmQOdH74SN+tjEwJbb2hAH+SDyeUtC9HmQu6uR/VdqNPvd+ab 8=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AvMAAA6LdU1AaMHG/2dsb2JhbACYE44qdKF+nEeFYgSFG4pj X-IronPort-AV: E=Sophos;i="4.62,283,1297036800"; d="scan'208";a="274715282" Received: from syd-core-1.cisco.com ([64.104.193.198]) by sj-iport-3.cisco.com with ESMTP; 08 Mar 2011 09:52:14 +0000 Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by syd-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id p289q6V8004156; Tue, 8 Mar 2011 09:52:13 GMT Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675); Tue, 8 Mar 2011 15:22:05 +0530 X-MimeOLE: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Date: Tue, 8 Mar 2011 15:22:09 +0530 Message-ID: In-Reply-To: <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [dispatch] SIP load balancing Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3mTxRFo0IAASaig References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> From: "Parthasarathi R (partr)" To: "Paul E. Jones" , "Vijay K. Gurbani" , X-OriginalArrivalTime: 08 Mar 2011 09:52:05.0611 (UTC) FILETIME=[7B624BB0:01CBDD76] Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 09:51:02 -0000 Vijay, As Paul K & Paul J mentioned, I'm looking at your proposal with lot of similarity to SCE. One more key stuff, I could not understand the need for two separate mechanism for 1) Load balancing ("weight") 2) Overload control ("oc") In case feedback mechanism to the upstream is generic enough, the load balancer will be able to balance based on value (capacity) received from downstream and reduce the load drastically or stop in case the value is too low in terms of capacity. Even though any overload control algorithm is used ("oc" or "weight" or "SCE" or "resource-availability" other), the locally overload control mechanism is required in downstream as per mentioned in draft-ietf-soc-overload-design. I wish that we will look into these aspect because overload control is the extreme case for load balancing the cluster of downstream entity and load balancing mechanism shall help to avoid the overload control based on feedback. I guess that your mechanism assumes about downstream SIP UA as signaling entity only and does not handle media. Please correct me in case I wrongly understand. draft-partha-dispatch-resource-availability-00 was written to handle both resource monitoring and load balancing of SIP based media servers wherein load balancing has to take care of overload as well. I'm also interested in discussing SIP load balancing topic during IETF-80 meeting. Thanks Partha -----Original Message----- From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On Behalf Of Paul E. Jones Sent: Tuesday, March 08, 2011 10:35 AM To: 'Vijay K. Gurbani'; dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing Vijay, The Session Capacity Estimate (SCE) idea is actually quite similar to the "adaptive load balancing" draft, from what I can see. SCE was proposed as a solution to the "overload" problem, the approach being one where we try to avoid overload by selecting a device that reports a higher capacity value. Thus, one could view it as "overload control through load balancing". We didn't explicitly call it load balancing, though, because we were not trying to balance load (i.e., place the same load on all servers). Rather, we just wanted to direct traffic to devices that we knew had available capacity. In practice, the way SCE would be employed is the following: * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from peer devices * The selector would direct traffic toward the peer devices in a round-robin fashion, taking into consideration the SCE value so as to prevent exhausting resources on any given peer device * Once the peer's SCE value dropped below a given "low-water mark", the peer device would be less favored and not given traffic until the SCE value increases above the low-water mark (OPTIONS "ping" messages may be used to determine the SCE value in the event that there is no other traffic) * The current SOC overload control mechanism would be used to to prevent message overloading, as SCE only focuses on resource overloading (e.g., memory, CPU, disk space, DSPs) Note that the "round-robin" approach may or may not be preferred and is not required in our draft. I've found this approach works pretty well, but one may prefer to use a particular weighting that is more complex. I don't think local decisions in the box need to be subject to standardization: we just need to standardize the signaling to enable such decisions. In any case, it certainly looks like we're thinking along the same lines. Paul > -----Original Message----- > From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On=20 > Behalf Of Vijay K. Gurbani > Sent: Monday, March 07, 2011 6:38 PM > To: dispatch@ietf.org > Subject: Re: [dispatch] SIP load balancing >=20 > Paul: Thank you for your input. More inline. >=20 > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: > > Vijay, > > IIUC, the algorithm you have proposed is similar to the one Paul=20 > > Jones has proposed - with the main difference being the numeric=20 > > range of the weight. Do I have that right? >=20 > I believe that Paul J.'s algorithm is more suited towards overload=20 > control, not load balancing (and I will await Paul J.'s correction to=20 > my assertion). >=20 > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy an=20 > indication of the health of the SIP server in the cluster. If one=20 > waits until overload is imminent, then it may be too late since a=20 > sudden and unanticipated burst of messages will put the SIP server in=20 > overload rather quickly. >=20 > > It isn't clear to me from your draft if you are specifying the=20 > > algorithm by which the relative weights are used to determine the=20 > > probability of a request going to one server or another. >=20 > Yes, same way as RFC2782 does it. >=20 > > ISTM that it *does* matter. And it especially matters if there is an > > implication that the upstream node's algorithm may depend on some=20 > > assumption or knowledge about the downstream nodes. >=20 > No knowledge, just a reported number that the upstream SIP proxy can=20 > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution=20 > starts off with the SRV indicators in RFC2782, without any regard to=20 > additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model,=20 > except render the weight to be sensitive to the current traffic in the > cluster in real-time. >=20 > > And of course this method of reporting weights isn't sufficient to=20 > > drive a balancing algorithm that is sensitive to some requests=20 > > requiring more or different resources than others. (That seems to be > > a real problem, though I don't know if we have a real solution for > > it.) >=20 > Yes, that is a real problem, as is the problem of a SIP server in a=20 > cluster that has a low weight and is thus not contacted with enough=20 > frequency to merit more participation to the goodput. >=20 > It is well known in SIP that some methods are more onerous to process=20 > than others --- from a proxying point of view, processing an INVITE=20 > with forking and response aggregation is far more detrimental to=20 > resource usage than is processing a BYE. Some existing literature=20 > techniques thus add a coefficient to the weight of sending an INVITE versus a BYE. > At this time, the process outlined in dispatch-adaptive-load-balancing > treats all requests the same, but we can definitely look at ways to=20 > change this behaviour. >=20 > Thanks, >=20 > - vijay > -- > Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane,=20 > Rm. 9C-533, Naperville, Illinois 60566 (USA) > Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com > Web: http://ect.bell-labs.com/who/vkg/ > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch From pkyzivat@cisco.com Tue Mar 8 05:48:34 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A2F913A686C for ; Tue, 8 Mar 2011 05:48:34 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -110.53 X-Spam-Level: X-Spam-Status: No, score=-110.53 tagged_above=-999 required=5 tests=[AWL=0.069, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id T-Rdv3PhULEo for ; Tue, 8 Mar 2011 05:48:33 -0800 (PST) Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id 621E93A63CB for ; Tue, 8 Mar 2011 05:48:33 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=3668; q=dns/txt; s=iport; t=1299592188; x=1300801788; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=niX/ySzIZ4bIMyTja7bQKhcSnYvotFKS4w9uYAFYXj0=; b=OalGmA5P3Pg+WhtE0uqUjDbu0M52bKUXnXeH8W8EhD+S68j7c6m4Y8cd aowFvomhYwLDQB7BS5hTq2iMu+srqRqmv7LSLRbikcYLYsxQ6z5SQlA7I TYeqwsPuFiCw34FV7jvvj7o8oSJSPwD5Man17bu7t/5BP9HSb8vSzPZKs M=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AvsEAJrCdU1AZnwN/2dsb2JhbACmQHSiWpxJhWMEhR2HFYNH X-IronPort-AV: E=Sophos;i="4.62,284,1297036800"; d="scan'208";a="223222349" Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-1.cisco.com with ESMTP; 08 Mar 2011 13:49:45 +0000 Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id p28Dnjlg027691 for ; Tue, 8 Mar 2011 13:49:45 GMT Message-ID: <4D7633F9.7000608@cisco.com> Date: Tue, 08 Mar 2011 08:49:45 -0500 From: Paul Kyzivat User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7 MIME-Version: 1.0 To: dispatch@ietf.org References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> In-Reply-To: <4D756C47.1000509@bell-labs.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 13:48:34 -0000 Vijay, Thanks for clarifications. Now I think I understand what you are proposing. It is certainly simple, and better than nothing. It *does* require that all the members of the load balancing set assign their own weights in a consistent way. This of course is no different than when the weights are assigned in DNS, except that its likely to be one person who puts all the weights into DNS, while the weights coming from the servers can be influenced by who implemented them. If they aren't all implemented consistently it can be a problem. But this is a second order problem, and is probably unavoidable. (AFAIK every scheme for this space has a similar constraint.) It might need to be the subject of additional standardization/profiling, though it might not be ietf work. Thanks, Paul On 3/7/2011 6:37 PM, Vijay K. Gurbani wrote: > Paul: Thank you for your input. More inline. > > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: >> Vijay, >> IIUC, the algorithm you have proposed is similar to the one Paul >> Jones has proposed - with the main difference being the numeric range >> of the weight. Do I have that right? > > I believe that Paul J.'s algorithm is more suited towards overload > control, not load balancing (and I will await Paul J.'s correction to > my assertion). > > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy > an indication of the health of the SIP server in the cluster. If one > waits until overload is imminent, then it may be too late since a sudden > and unanticipated burst of messages will put the SIP server in > overload rather quickly. > >> It isn't clear to me from your draft if you are specifying the >> algorithm by which the relative weights are used to determine the >> probability of a request going to one server or another. > > Yes, same way as RFC2782 does it. > >> ISTM that it *does* matter. And it especially matters if there is an >> implication that the upstream node's algorithm may depend on some >> assumption or knowledge about the downstream nodes. > > No knowledge, just a reported number that the upstream SIP proxy can > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution > starts off with the SRV indicators in RFC2782, without any regard > to additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model, > except render the weight to be sensitive to the current traffic > in the cluster in real-time. > >> And of course this method of reporting weights isn't sufficient to >> drive a balancing algorithm that is sensitive to some requests >> requiring more or different resources than others. (That seems to be >> a real problem, though I don't know if we have a real solution for >> it.) > > Yes, that is a real problem, as is the problem of a SIP server in a > cluster that has a low weight and is thus not contacted with enough > frequency to merit more participation to the goodput. > > It is well known in SIP that some methods are more onerous to process > than others --- from a proxying point of view, processing an INVITE > with forking and response aggregation is far more detrimental to > resource usage than is processing a BYE. Some existing literature > techniques thus add a coefficient to the weight of sending an INVITE > versus a BYE. At this time, the process outlined in > dispatch-adaptive-load-balancing treats all requests the same, but we > can definitely > look at ways to change this behaviour. > > Thanks, > > - vijay From pkyzivat@cisco.com Tue Mar 8 06:11:34 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1A7FE3A6843 for ; Tue, 8 Mar 2011 06:11:31 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -110.531 X-Spam-Level: X-Spam-Status: No, score=-110.531 tagged_above=-999 required=5 tests=[AWL=0.068, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FgqwWm7sByyw for ; Tue, 8 Mar 2011 06:11:26 -0800 (PST) Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87]) by core3.amsl.com (Postfix) with ESMTP id 7F1C33A67BD for ; Tue, 8 Mar 2011 06:11:25 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=1681; q=dns/txt; s=iport; t=1299593560; x=1300803160; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=+o7nmAN2kcTe2FEi/DxqtTm5igr101Gyx22j2KT43kA=; b=XKBIDPBnxZgrrhnkJrmCLHP01V1UncfKI5L/aheTUZ4JpbodwL/iDd+t v/zWRRmePt8e/Hggxi7VY0ENz20EKx5KRboesxt5FwlhqR7tPQjgCxg6h w7YS6392chwVH7DRbodLNno+SvDw7tMn0Kxtpnqq8gxEbVS/V7KVSiPYW Q=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AuMFAMLHdU2rRN+K/2dsb2JhbACELJQmjW50olmLDJE3gSeDRnYEhR2HFYNH X-IronPort-AV: E=Sophos;i="4.62,284,1297036800"; d="scan'208";a="342292800" Received: from sj-core-4.cisco.com ([171.68.223.138]) by sj-iport-5.cisco.com with ESMTP; 08 Mar 2011 14:12:39 +0000 Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by sj-core-4.cisco.com (8.13.8/8.14.3) with ESMTP id p28ECdSg011138 for ; Tue, 8 Mar 2011 14:12:39 GMT Message-ID: <4D763957.7050300@cisco.com> Date: Tue, 08 Mar 2011 09:12:39 -0500 From: Paul Kyzivat User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7 MIME-Version: 1.0 To: dispatch@ietf.org References: <294181.75599.qm@web29112.mail.ird.yahoo.com> In-Reply-To: <294181.75599.qm@web29112.mail.ird.yahoo.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [dispatch] New Draft - draft-elz-dispatch-req-sip-3rdpartyprivacy-00 X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 14:11:34 -0000 Ian, I'd like to thank you for submitting a clear requirements draft that is unburdened by mechanism proposals. IMO the hard part of this issue is sorting out whether there is a need for a "generalized" mechanism dealing with arbitrary 3rd party identities, or just for a mechanism to deal with a specific enumerated set of 3rd party identities. While its easy to wish for a generalized mechanism, I think it will be very difficult to adequately specify. If constrained to a specific set than it is a solvable problem. Thanks, Paul On 3/8/2011 3:15 AM, Ian Elz wrote: > All, > > I have submitted a new draft relating to third party privacy in SIP > messages. > > http://www.ietf.org/id/draft-elz-dispatch-req-sip-3rdpartyprivacy-00.txt > (Link that I forgot in the previous posting.) > > There are two specific issues included, how to specify third party > privacy in a SIP message; the precedence of privacy settings between > Privacy header values and Privacy URI parameters. > > The things to consider are: > > Should this work be pursued and if so in which forum. > > A suitable solution to meet the requirements. > > How any changes should be implemented: a new draft which modifies > RFC3323 or an update to RFC3323 which includes the additional work. > > All comments welcome. > > Sorry I forgot to write the Abstract in version -00. I will include this > in the next version. ( I leave abstracts to last and overlooked this > before submission.) > > Ian Elz > > > > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch From hmmr@cisco.com Tue Mar 8 06:19:03 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id F1D5D3A67B7 for ; Tue, 8 Mar 2011 06:19:02 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -10.599 X-Spam-Level: X-Spam-Status: No, score=-10.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nObSFO87AW4M for ; Tue, 8 Mar 2011 06:19:01 -0800 (PST) Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id DE7F53A6874 for ; Tue, 8 Mar 2011 06:19:00 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=hmmr@cisco.com; l=8060; q=dns/txt; s=iport; t=1299594016; x=1300803616; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to; bh=mD+1Jj1qM9kV8txqKjghJHUxvq5TDXHVrrHvqsD++PE=; b=elS3W76WyGca0bw0Ds5QS2EW6DLa01oQGj1H6zCi8T2haipIeAKFvVpp H7aa6ZuL5warrLKintztYOuAhxBIobGmnCmOS/AgVc5dkmBl6XAcD6cvX wR3H5iRwioFDyoWbG1MvLWJ98h5eA7+E2OPzgMUnwcQIhSj3+/cpk+K4u 0=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: Aj0BAKPJdU2tJV2Z/2dsb2JhbACYE44tdGihY5NviFSFYwSFHYpi X-IronPort-AV: E=Sophos;i="4.62,284,1297036800"; d="scan'208";a="223231641" Received: from rcdn-core-2.cisco.com ([173.37.93.153]) by rtp-iport-1.cisco.com with ESMTP; 08 Mar 2011 14:20:15 +0000 Received: from xbh-rcd-202.cisco.com (xbh-rcd-202.cisco.com [72.163.62.201]) by rcdn-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p28EKFbU002315; Tue, 8 Mar 2011 14:20:15 GMT Received: from xmb-rcd-111.cisco.com ([72.163.62.153]) by xbh-rcd-202.cisco.com with Microsoft SMTPSVC(6.0.3790.4675); Tue, 8 Mar 2011 08:20:15 -0600 X-MimeOLE: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable Date: Tue, 8 Mar 2011 08:20:13 -0600 Message-ID: In-Reply-To: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [dispatch] SIP load balancing Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3mTxRFo0IAASaiggABUXlA= References: <4D755B9E.6030008@bell-labs.com><4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com><086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> From: "Mike Hammer (hmmr)" To: "Parthasarathi R (partr)" , "Paul E. Jones" , "Vijay K. Gurbani" , X-OriginalArrivalTime: 08 Mar 2011 14:20:15.0336 (UTC) FILETIME=[F19B7E80:01CBDD9B] Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 14:19:03 -0000 Is there a need to inter-twine media with signaling? I think if there is an assumption that the media and signaling take the same path,=20 then you've lost some flexibility, and worse make sub-optimal choices when=20 one or the other is over-capacity and the other is not even in the picture. Mike -----Original Message----- From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On Behalf Of Parthasarathi R (partr) Sent: Tuesday, March 08, 2011 4:52 AM To: Paul E. Jones; Vijay K. Gurbani; dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing Vijay, As Paul K & Paul J mentioned, I'm looking at your proposal with lot of similarity to SCE. One more key stuff, I could not understand the need for two separate mechanism for 1) Load balancing ("weight") 2) Overload control ("oc") In case feedback mechanism to the upstream is generic enough, the load balancer will be able to balance based on value (capacity) received from downstream and reduce the load drastically or stop in case the value is too low in terms of capacity. Even though any overload control algorithm is used ("oc" or "weight" or "SCE" or "resource-availability" other), the locally overload control mechanism is required in downstream as per mentioned in draft-ietf-soc-overload-design. I wish that we will look into these aspect because overload control is the extreme case for load balancing the cluster of downstream entity and load balancing mechanism shall help to avoid the overload control based on feedback. I guess that your mechanism assumes about downstream SIP UA as signaling entity only and does not handle media. Please correct me in case I wrongly understand. draft-partha-dispatch-resource-availability-00 was written to handle both resource monitoring and load balancing of SIP based media servers wherein load balancing has to take care of overload as well. I'm also interested in discussing SIP load balancing topic during IETF-80 meeting. Thanks Partha -----Original Message----- From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On Behalf Of Paul E. Jones Sent: Tuesday, March 08, 2011 10:35 AM To: 'Vijay K. Gurbani'; dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing Vijay, The Session Capacity Estimate (SCE) idea is actually quite similar to the "adaptive load balancing" draft, from what I can see. SCE was proposed as a solution to the "overload" problem, the approach being one where we try to avoid overload by selecting a device that reports a higher capacity value. Thus, one could view it as "overload control through load balancing". We didn't explicitly call it load balancing, though, because we were not trying to balance load (i.e., place the same load on all servers). Rather, we just wanted to direct traffic to devices that we knew had available capacity. In practice, the way SCE would be employed is the following: * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from peer devices * The selector would direct traffic toward the peer devices in a round-robin fashion, taking into consideration the SCE value so as to prevent exhausting resources on any given peer device * Once the peer's SCE value dropped below a given "low-water mark", the peer device would be less favored and not given traffic until the SCE value increases above the low-water mark (OPTIONS "ping" messages may be used to determine the SCE value in the event that there is no other traffic) * The current SOC overload control mechanism would be used to to prevent message overloading, as SCE only focuses on resource overloading (e.g., memory, CPU, disk space, DSPs) Note that the "round-robin" approach may or may not be preferred and is not required in our draft. I've found this approach works pretty well, but one may prefer to use a particular weighting that is more complex. I don't think local decisions in the box need to be subject to standardization: we just need to standardize the signaling to enable such decisions. In any case, it certainly looks like we're thinking along the same lines. Paul > -----Original Message----- > From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On=20 > Behalf Of Vijay K. Gurbani > Sent: Monday, March 07, 2011 6:38 PM > To: dispatch@ietf.org > Subject: Re: [dispatch] SIP load balancing >=20 > Paul: Thank you for your input. More inline. >=20 > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: > > Vijay, > > IIUC, the algorithm you have proposed is similar to the one Paul=20 > > Jones has proposed - with the main difference being the numeric=20 > > range of the weight. Do I have that right? >=20 > I believe that Paul J.'s algorithm is more suited towards overload=20 > control, not load balancing (and I will await Paul J.'s correction to=20 > my assertion). >=20 > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy an=20 > indication of the health of the SIP server in the cluster. If one=20 > waits until overload is imminent, then it may be too late since a=20 > sudden and unanticipated burst of messages will put the SIP server in=20 > overload rather quickly. >=20 > > It isn't clear to me from your draft if you are specifying the=20 > > algorithm by which the relative weights are used to determine the=20 > > probability of a request going to one server or another. >=20 > Yes, same way as RFC2782 does it. >=20 > > ISTM that it *does* matter. And it especially matters if there is an > > implication that the upstream node's algorithm may depend on some=20 > > assumption or knowledge about the downstream nodes. >=20 > No knowledge, just a reported number that the upstream SIP proxy can=20 > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution=20 > starts off with the SRV indicators in RFC2782, without any regard to=20 > additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model,=20 > except render the weight to be sensitive to the current traffic in the > cluster in real-time. >=20 > > And of course this method of reporting weights isn't sufficient to=20 > > drive a balancing algorithm that is sensitive to some requests=20 > > requiring more or different resources than others. (That seems to be > > a real problem, though I don't know if we have a real solution for > > it.) >=20 > Yes, that is a real problem, as is the problem of a SIP server in a=20 > cluster that has a low weight and is thus not contacted with enough=20 > frequency to merit more participation to the goodput. >=20 > It is well known in SIP that some methods are more onerous to process=20 > than others --- from a proxying point of view, processing an INVITE=20 > with forking and response aggregation is far more detrimental to=20 > resource usage than is processing a BYE. Some existing literature=20 > techniques thus add a coefficient to the weight of sending an INVITE versus a BYE. > At this time, the process outlined in dispatch-adaptive-load-balancing > treats all requests the same, but we can definitely look at ways to=20 > change this behaviour. >=20 > Thanks, >=20 > - vijay > -- > Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane,=20 > Rm. 9C-533, Naperville, Illinois 60566 (USA) > Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com > Web: http://ect.bell-labs.com/who/vkg/ > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch From partr@cisco.com Tue Mar 8 06:30:59 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 084FD3A6908 for ; Tue, 8 Mar 2011 06:30:59 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -10.154 X-Spam-Level: X-Spam-Status: No, score=-10.154 tagged_above=-999 required=5 tests=[AWL=0.445, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jtFWmFqOagnt for ; Tue, 8 Mar 2011 06:30:56 -0800 (PST) Received: from sj-iport-2.cisco.com (sj-iport-2.cisco.com [171.71.176.71]) by core3.amsl.com (Postfix) with ESMTP id 50FFB3A68B1 for ; Tue, 8 Mar 2011 06:30:56 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=8601; q=dns/txt; s=iport; t=1299594731; x=1300804331; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to; bh=dvb5ap22XgQjbbpfhLgWmk6AdHWB+rqkdt1N2o+znT4=; b=Hk0nPJclw8DBzZ18V3MC1zexfVKE6H7hGv8hlS8u8h2CXLKwUgvJ5JGe +sHAOiffXmWQ6ZWQGmbF5PJ/NNFBExRV/9GYUceekgq1A5vCyytNYw6/U uSY/9oIWEOit6L6uyYLx2+8X4db47wUePM5yUBp+EqsoZrSXdmvCttw5W A=; X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AvMAAPrLdU1AaMHG/2dsb2JhbACYE44tdKJYnEaFYwSFG4pk X-IronPort-AV: E=Sophos;i="4.62,284,1297036800"; d="scan'208";a="317952133" Received: from syd-core-1.cisco.com ([64.104.193.198]) by sj-iport-2.cisco.com with ESMTP; 08 Mar 2011 14:32:10 +0000 Received: from xbh-bgl-412.cisco.com (xbh-bgl-412.cisco.com [72.163.129.202]) by syd-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id p28EVd59006861; Tue, 8 Mar 2011 14:32:09 GMT Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-412.cisco.com with Microsoft SMTPSVC(6.0.3790.4675); Tue, 8 Mar 2011 20:01:55 +0530 X-MimeOLE: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Date: Tue, 8 Mar 2011 20:01:53 +0530 Message-ID: In-Reply-To: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [dispatch] SIP load balancing Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3mTxRFo0IAASaiggABUXlCAAAJssA== References: <4D755B9E.6030008@bell-labs.com><4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com><086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> From: "Parthasarathi R (partr)" To: "Mike Hammer (hmmr)" , "Paul E. Jones" , "Vijay K. Gurbani" , X-OriginalArrivalTime: 08 Mar 2011 14:31:55.0105 (UTC) FILETIME=[92B3C510:01CBDD9D] Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 14:30:59 -0000 Mike, The mechanism is required for the SIP based media servers to load balancer. Media servers may be based on IETF mediactrl sort of interfaces or SIP devices like SBC, PSTN GW. As you mentioned, the problem in SIP based media server is complex as more resources involved at the given time. =20 Thanks Partha -----Original Message----- From: Mike Hammer (hmmr)=20 Sent: Tuesday, March 08, 2011 7:50 PM To: Parthasarathi R (partr); Paul E. Jones; Vijay K. Gurbani; dispatch@ietf.org Subject: RE: [dispatch] SIP load balancing Is there a need to inter-twine media with signaling? I think if there is an assumption that the media and signaling take the same path, then you've lost some flexibility, and worse make sub-optimal choices when one or the other is over-capacity and the other is not even in the picture. Mike -----Original Message----- From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On Behalf Of Parthasarathi R (partr) Sent: Tuesday, March 08, 2011 4:52 AM To: Paul E. Jones; Vijay K. Gurbani; dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing Vijay, As Paul K & Paul J mentioned, I'm looking at your proposal with lot of similarity to SCE. One more key stuff, I could not understand the need for two separate mechanism for 1) Load balancing ("weight") 2) Overload control ("oc") In case feedback mechanism to the upstream is generic enough, the load balancer will be able to balance based on value (capacity) received from downstream and reduce the load drastically or stop in case the value is too low in terms of capacity. Even though any overload control algorithm is used ("oc" or "weight" or "SCE" or "resource-availability" other), the locally overload control mechanism is required in downstream as per mentioned in draft-ietf-soc-overload-design. I wish that we will look into these aspect because overload control is the extreme case for load balancing the cluster of downstream entity and load balancing mechanism shall help to avoid the overload control based on feedback. I guess that your mechanism assumes about downstream SIP UA as signaling entity only and does not handle media. Please correct me in case I wrongly understand. draft-partha-dispatch-resource-availability-00 was written to handle both resource monitoring and load balancing of SIP based media servers wherein load balancing has to take care of overload as well. I'm also interested in discussing SIP load balancing topic during IETF-80 meeting. Thanks Partha -----Original Message----- From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On Behalf Of Paul E. Jones Sent: Tuesday, March 08, 2011 10:35 AM To: 'Vijay K. Gurbani'; dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing Vijay, The Session Capacity Estimate (SCE) idea is actually quite similar to the "adaptive load balancing" draft, from what I can see. SCE was proposed as a solution to the "overload" problem, the approach being one where we try to avoid overload by selecting a device that reports a higher capacity value. Thus, one could view it as "overload control through load balancing". We didn't explicitly call it load balancing, though, because we were not trying to balance load (i.e., place the same load on all servers). Rather, we just wanted to direct traffic to devices that we knew had available capacity. In practice, the way SCE would be employed is the following: * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from peer devices * The selector would direct traffic toward the peer devices in a round-robin fashion, taking into consideration the SCE value so as to prevent exhausting resources on any given peer device * Once the peer's SCE value dropped below a given "low-water mark", the peer device would be less favored and not given traffic until the SCE value increases above the low-water mark (OPTIONS "ping" messages may be used to determine the SCE value in the event that there is no other traffic) * The current SOC overload control mechanism would be used to to prevent message overloading, as SCE only focuses on resource overloading (e.g., memory, CPU, disk space, DSPs) Note that the "round-robin" approach may or may not be preferred and is not required in our draft. I've found this approach works pretty well, but one may prefer to use a particular weighting that is more complex. I don't think local decisions in the box need to be subject to standardization: we just need to standardize the signaling to enable such decisions. In any case, it certainly looks like we're thinking along the same lines. Paul > -----Original Message----- > From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On=20 > Behalf Of Vijay K. Gurbani > Sent: Monday, March 07, 2011 6:38 PM > To: dispatch@ietf.org > Subject: Re: [dispatch] SIP load balancing >=20 > Paul: Thank you for your input. More inline. >=20 > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: > > Vijay, > > IIUC, the algorithm you have proposed is similar to the one Paul=20 > > Jones has proposed - with the main difference being the numeric=20 > > range of the weight. Do I have that right? >=20 > I believe that Paul J.'s algorithm is more suited towards overload=20 > control, not load balancing (and I will await Paul J.'s correction to=20 > my assertion). >=20 > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy an=20 > indication of the health of the SIP server in the cluster. If one=20 > waits until overload is imminent, then it may be too late since a=20 > sudden and unanticipated burst of messages will put the SIP server in=20 > overload rather quickly. >=20 > > It isn't clear to me from your draft if you are specifying the=20 > > algorithm by which the relative weights are used to determine the=20 > > probability of a request going to one server or another. >=20 > Yes, same way as RFC2782 does it. >=20 > > ISTM that it *does* matter. And it especially matters if there is an > > implication that the upstream node's algorithm may depend on some=20 > > assumption or knowledge about the downstream nodes. >=20 > No knowledge, just a reported number that the upstream SIP proxy can=20 > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution=20 > starts off with the SRV indicators in RFC2782, without any regard to=20 > additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model,=20 > except render the weight to be sensitive to the current traffic in the > cluster in real-time. >=20 > > And of course this method of reporting weights isn't sufficient to=20 > > drive a balancing algorithm that is sensitive to some requests=20 > > requiring more or different resources than others. (That seems to be > > a real problem, though I don't know if we have a real solution for > > it.) >=20 > Yes, that is a real problem, as is the problem of a SIP server in a=20 > cluster that has a low weight and is thus not contacted with enough=20 > frequency to merit more participation to the goodput. >=20 > It is well known in SIP that some methods are more onerous to process=20 > than others --- from a proxying point of view, processing an INVITE=20 > with forking and response aggregation is far more detrimental to=20 > resource usage than is processing a BYE. Some existing literature=20 > techniques thus add a coefficient to the weight of sending an INVITE versus a BYE. > At this time, the process outlined in dispatch-adaptive-load-balancing > treats all requests the same, but we can definitely look at ways to=20 > change this behaviour. >=20 > Thanks, >=20 > - vijay > -- > Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane,=20 > Rm. 9C-533, Naperville, Illinois 60566 (USA) > Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com > Web: http://ect.bell-labs.com/who/vkg/ > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch From spromano@unina.it Tue Mar 8 06:39:27 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EADDD3A689E for ; Tue, 8 Mar 2011 06:39:26 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -100.719 X-Spam-Level: X-Spam-Status: No, score=-100.719 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, USER_IN_WHITELIST=-100] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id u7ojc3qEERNm for ; Tue, 8 Mar 2011 06:39:22 -0800 (PST) Received: from smtp2.unina.it (smtp2.unina.it [192.132.34.62]) by core3.amsl.com (Postfix) with ESMTP id 776103A67BD for ; Tue, 8 Mar 2011 06:39:22 -0800 (PST) Received: from [143.225.229.230] ([143.225.229.230]) (authenticated bits=0) by smtp2.unina.it (8.14.4/8.14.4) with ESMTP id p28EeVYt003984 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 15:40:31 +0100 Message-ID: <4D763FDD.8010702@unina.it> Date: Tue, 08 Mar 2011 15:40:29 +0100 From: Simon Pietro Romano User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; it; rv:1.9.2.14) Gecko/20110221 Lightning/1.0b2 Thunderbird/3.1.8 MIME-Version: 1.0 To: "Vijay K. Gurbani" References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> In-Reply-To: <4D756C47.1000509@bell-labs.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Cc: dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 14:39:27 -0000 Hi Vijay, interesting draft, indeed. It brought to my mind that some of the ideas you are proposing can be found in Mobicent's implementation of Distributed Load Balancing, which makes use of 'Via' and 'Record-Route' headers to keep the Load Balancer in the loop, and exploits (in place of your piggybacking proposal) a proprietary 'heartbit' protocol (implemented with Java RMI) to report health status from SIP Servers to the SIP Proxy. Interestingly enough, inside Mobicents the balancer algorithms are 'pluggable': the original implementation is shipped with three different built-in algorithms, but nothing prevents implementers from proposing new ones. As a general comment, I think that solutions like the one you are proposing have strong liaisons, besides Overload Control, with the broader issue of (transparent) fail-over management. My 2 cents, Simon Il 08/03/2011 00:37, Vijay K. Gurbani ha scritto: > Paul: Thank you for your input. More inline. > > On 03/07/2011 05:14 PM, Paul Kyzivat wrote: >> Vijay, >> IIUC, the algorithm you have proposed is similar to the one Paul >> Jones has proposed - with the main difference being the numeric range >> of the weight. Do I have that right? > > I believe that Paul J.'s algorithm is more suited towards overload > control, not load balancing (and I will await Paul J.'s correction to > my assertion). > > The intent in draft-bessis-dispatch-adaptive-load-balancing is to > *always* report the current load to give the upstream SIP proxy > an indication of the health of the SIP server in the cluster. If one > waits until overload is imminent, then it may be too late since a sudden > and unanticipated burst of messages will put the SIP server in > overload rather quickly. > >> It isn't clear to me from your draft if you are specifying the >> algorithm by which the relative weights are used to determine the >> probability of a request going to one server or another. > > Yes, same way as RFC2782 does it. > >> ISTM that it *does* matter. And it especially matters if there is an >> implication that the upstream node's algorithm may depend on some >> assumption or knowledge about the downstream nodes. > > No knowledge, just a reported number that the upstream SIP proxy can > treat as a RFC2782-like weight indicator. Today, rfc3263-resolution > starts off with the SRV indicators in RFC2782, without any regard > to additional knowledge or assumption about downstream nodes. > dispatch-adaptive-load-balancing is trying to use the same model, > except render the weight to be sensitive to the current traffic > in the cluster in real-time. > >> And of course this method of reporting weights isn't sufficient to >> drive a balancing algorithm that is sensitive to some requests >> requiring more or different resources than others. (That seems to be >> a real problem, though I don't know if we have a real solution for >> it.) > > Yes, that is a real problem, as is the problem of a SIP server in a > cluster that has a low weight and is thus not contacted with enough > frequency to merit more participation to the goodput. > > It is well known in SIP that some methods are more onerous to process > than others --- from a proxying point of view, processing an INVITE > with forking and response aggregation is far more detrimental to > resource usage than is processing a BYE. Some existing literature > techniques thus add a coefficient to the weight of sending an INVITE > versus a BYE. At this time, the process outlined in > dispatch-adaptive-load-balancing treats all requests the same, but we > can definitely > look at ways to change this behaviour. > > Thanks, > > - vijay -- _\\|//_ ( O-O ) ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~ Simon Pietro Romano Universita' di Napoli Federico II Computer Science Department Phone: +39 081 7683823 -- Fax: +39 081 7684219 e-mail: spromano@unina.it http://www.comics.unina.it/simonpietro.romano <>. Magritte. oooO ~~~~~~~~~~~~~~~~~~~~~~( )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~ \ ( ( ) \_) ) / (_/ From Thierry.Bessis@alcatel-lucent.com Tue Mar 8 09:33:41 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C5A6A3A6936 for ; Tue, 8 Mar 2011 09:33:41 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -6.597 X-Spam-Level: X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id VdsQ9MMAWs1E for ; Tue, 8 Mar 2011 09:33:39 -0800 (PST) Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 2B1CF3A6945 for ; Tue, 8 Mar 2011 09:33:39 -0800 (PST) Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p28HYn5l012751 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 11:34:50 -0600 (CST) Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28HYnEr024208; Tue, 8 Mar 2011 11:34:49 -0600 (CST) Message-ID: <4D7668B9.1020105@alcatel-lucent.com> Date: Tue, 08 Mar 2011 11:34:49 -0600 From: Thierry Bessis User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9 MIME-Version: 1.0 To: "Paul E. Jones" References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> In-Reply-To: <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> Content-Type: multipart/alternative; boundary="------------000605040804040307080105" X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35 Cc: dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 17:33:41 -0000 This is a multi-part message in MIME format. --------------000605040804040307080105 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Paul, Please let me address some of your comments. I don't know what the SCE is, so please pardon me if I misunderstand something. On 07-Mar-2011 23:05, Paul E. Jones wrote: > Vijay, > > The Session Capacity Estimate (SCE) idea is actually quite similar to the > "adaptive load balancing" draft, from what I can see. SCE was proposed as a > solution to the "overload" problem, the approach being one where we try to > avoid overload by selecting a device that reports a higher capacity value. > Thus, one could view it as "overload control through load balancing". We > didn't explicitly call it load balancing, though, because we were not trying > to balance load (i.e., place the same load on all servers). Rather, we just > wanted to direct traffic to devices that we knew had available capacity. Agreed, the goal seems to be the same: avoiding or at least delaying overload. And your other comment too: our proposal does try to balance the load on all the servers. The rational is that it is very easy for each server to know where it is in terms of resource usage (for instance: I'm 60 percent resource loaded at this moment), and this is just the information you need to balance the traffic. Now it seems (but it is not clear, see below) that the SCE "available capacity" is actually a real time feedback, so the information would be very similar actually. (What's the difference ? ) > In practice, the way SCE would be employed is the following: > * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from > peer devices > * The selector would direct traffic toward the peer devices in a > round-robin fashion, taking into consideration the SCE value so > as to prevent exhausting resources on any given peer device I see a contradiction here: if it is round-robin, it does not take into account the SCE value. Is that something more like a weighted round-robin algorithm ? > * Once the peer's SCE value dropped below a given "low-water mark", > the peer device would be less favored and not given traffic until > the SCE value increases above the low-water mark (OPTIONS "ping" > messages may be used to determine the SCE value in the event that > there is no other traffic) Why would we wait for a lower mark before to influence the distribution ? Also is that less favored, or given no traffic ? The later has stability issues. And why do we need ping Options when there is no traffic ? If there is no traffic (or very little), should we care about optimal distribution from this source ? > * The current SOC overload control mechanism would be used to > to prevent message overloading, as SCE only focuses on resource > overloading (e.g., memory, CPU, disk space, DSPs) I don't see this difference here (message overload versus resource overload ? ) . To me, SOC focuses on source shedding, while SCE seems to focus on improved load distribution, just like our proposal. > Note that the "round-robin" approach may or may not be preferred and is not > required in our draft. I've found this approach works pretty well, but one > may prefer to use a particular weighting that is more complex. I don't > think local decisions in the box need to be subject to standardization: we > just need to standardize the signaling to enable such decisions. I believe the signaling is not enough. The deep meaning of the feedback parameter needs to be clarified too. For example, in our proposal, the feedback is simply the current utilization rate of the node (very simple). Also the expectation needs to be clarified because it has a very important impact on the recommended behavior in case of overload. The solution that we propose constantly tries to balance load in the cluster of servers. Therefore, in case the SOC shedding kicks in, there is NOT NEED to repeat the message elsewhere in the cluster because it is balanced already (So no other server has more capacity). This is a huge advantage for the network stability as message repeat tends to spread the overload and collapse the cluster. > In any case, it certainly looks like we're thinking along the same lines. Yes I agree. Cordially, Thierry > Paul > >> -----Original Message----- >> From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On >> Behalf Of Vijay K. Gurbani >> Sent: Monday, March 07, 2011 6:38 PM >> To: dispatch@ietf.org >> Subject: Re: [dispatch] SIP load balancing >> >> Paul: Thank you for your input. More inline. >> >> On 03/07/2011 05:14 PM, Paul Kyzivat wrote: >>> Vijay, >>> IIUC, the algorithm you have proposed is similar to the one Paul Jones >>> has proposed - with the main difference being the numeric range of the >>> weight. Do I have that right? >> I believe that Paul J.'s algorithm is more suited towards overload >> control, not load balancing (and I will await Paul J.'s correction to my >> assertion). >> >> The intent in draft-bessis-dispatch-adaptive-load-balancing is to >> *always* report the current load to give the upstream SIP proxy an >> indication of the health of the SIP server in the cluster. If one waits >> until overload is imminent, then it may be too late since a sudden and >> unanticipated burst of messages will put the SIP server in overload >> rather quickly. >> >>> It isn't clear to me from your draft if you are specifying the >>> algorithm by which the relative weights are used to determine the >>> probability of a request going to one server or another. >> Yes, same way as RFC2782 does it. >> >>> ISTM that it *does* matter. And it especially matters if there is an >>> implication that the upstream node's algorithm may depend on some >>> assumption or knowledge about the downstream nodes. >> No knowledge, just a reported number that the upstream SIP proxy can >> treat as a RFC2782-like weight indicator. Today, rfc3263-resolution >> starts off with the SRV indicators in RFC2782, without any regard to >> additional knowledge or assumption about downstream nodes. >> dispatch-adaptive-load-balancing is trying to use the same model, except >> render the weight to be sensitive to the current traffic in the cluster >> in real-time. >> >>> And of course this method of reporting weights isn't sufficient to >>> drive a balancing algorithm that is sensitive to some requests >>> requiring more or different resources than others. (That seems to be a >>> real problem, though I don't know if we have a real solution for >>> it.) >> Yes, that is a real problem, as is the problem of a SIP server in a >> cluster that has a low weight and is thus not contacted with enough >> frequency to merit more participation to the goodput. >> >> It is well known in SIP that some methods are more onerous to process >> than others --- from a proxying point of view, processing an INVITE with >> forking and response aggregation is far more detrimental to resource >> usage than is processing a BYE. Some existing literature techniques >> thus add a coefficient to the weight of sending an INVITE versus a BYE. >> At this time, the process outlined in dispatch-adaptive-load-balancing >> treats all requests the same, but we can definitely look at ways to >> change this behaviour. >> >> Thanks, >> >> - vijay >> -- >> Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent >> 1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA) >> Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com >> Web: http://ect.bell-labs.com/who/vkg/ >> _______________________________________________ >> dispatch mailing list >> dispatch@ietf.org >> https://www.ietf.org/mailman/listinfo/dispatch > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch -- -- Cordially, Thierry Bessis IMS Solution: Architecture & Product Management - ALTA Member Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM Alcatel-Lucent 1960 Lucent Lane - Room 6L-517 Naperville, IL 60563 U S Tel/Fax: +1 630 979 7989 Corporate IM: tbessis - Yahoo IM: thierry_bessis My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline Conference information: 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989 others countries see: http://all.alcatel-lucent.com/audio Upcoming planned Business Trip: 12 April - 14 April 2011 Upcoming planned Vacation: None --------------000605040804040307080105 Content-Type: multipart/related; boundary="------------020402090103080608090109" --------------020402090103080608090109 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Paul,

Please let me address some of your comments.
I don't know what the SCE is, so please pardon me if I misunderstand something.

On 07-Mar-2011 23:05, Paul E. Jones wrote: 
Vijay,

The Session Capacity Estimate (SCE) idea is actually quite similar to the
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a
solution to the "overload" problem, the approach being one where we try to
avoid overload by selecting a device that reports a higher capacity value.
Thus, one could view it as "overload control through load balancing".  We
didn't explicitly call it load balancing, though, because we were not trying
to balance load (i.e., place the same load on all servers).  Rather, we just
wanted to direct traffic to devices that we knew had available capacity.
Agreed, the goal seems to be the same: avoiding or at least delaying overload.
And your other comment too: our proposal does try to balance the load on all the servers. The rational is that it is very easy for each server to know where it is in terms of resource usage (for instance: I'm 60 percent resource loaded at this moment), and this is just the information you need to balance the traffic. Now it seems (but it is not clear, see below)  that the SCE "available capacity" is actually a real time feedback, so the information would be very similar actually. (What's the difference ? ) 
In practice, the way SCE would be employed is the following:
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
   peer devices
 * The selector would direct traffic toward the peer devices in a
   round-robin fashion, taking into consideration the SCE value so
   as to prevent exhausting resources on any given peer device
I see a contradiction here: if it is round-robin, it does not take into account the SCE value.
Is that something more like a weighted round-robin algorithm ? 
 * Once the peer's SCE value dropped below a given "low-water mark",
   the peer device would be less favored and not given traffic until
   the SCE value increases above the low-water mark (OPTIONS "ping"
   messages may be used to determine the SCE value in the event that
   there is no other traffic)
Why would we wait for a lower mark before to influence the distribution ? Also is that less favored, or given no traffic ?
The later has stability issues.  And why do we need ping Options when there is no traffic ?
 If there is no traffic (or very little), should we care about optimal distribution from this source ? 
 * The current SOC overload control mechanism would be used to
   to prevent message overloading, as SCE only focuses on resource
   overloading (e.g., memory, CPU, disk space, DSPs)
I don't see this difference here (message overload versus resource overload ? ) . To me, SOC focuses on source shedding, while SCE seems to focus on improved load distribution, just like our proposal. 
Note that the "round-robin" approach may or may not be preferred and is not
required in our draft.  I've found this approach works pretty well, but one
may prefer to use a particular weighting that is more complex.  I don't
think local decisions in the box need to be subject to standardization: we
just need to standardize the signaling to enable such decisions.
I believe the signaling is not enough. The deep meaning of the feedback parameter needs to be clarified too.
For example, in our proposal, the feedback is simply the current utilization rate of the node (very simple).

Also the expectation needs to be clarified because it has a very important impact on the recommended behavior in case of overload.
The solution that we propose constantly tries to balance load in the cluster of servers. Therefore, in case the SOC shedding kicks in, there is NOT NEED to repeat the message elsewhere in the cluster because it is balanced already (So no other server has more capacity). This is a huge advantage for the network stability as message repeat tends to spread the overload and collapse the cluster. 
In any case, it certainly looks like we're thinking along the same lines.
Yes I agree.

Cordially,

Thierry
Paul


-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Vijay K. Gurbani
Sent: Monday, March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

Paul: Thank you for your input.  More inline.

On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul Jones
has proposed - with the main difference being the numeric range of the
weight. Do I have that right?

I believe that Paul J.'s algorithm is more suited towards overload
control, not load balancing (and I will await Paul J.'s correction to my
assertion).

The intent in draft-bessis-dispatch-adaptive-load-balancing is to
*always* report the current load to give the upstream SIP proxy an
indication of the health of the SIP server in the cluster.  If one waits
until overload is imminent, then it may be too late since a sudden and
unanticipated burst of messages will put the SIP server in overload
rather quickly.


It isn't clear to me from your draft if you are specifying the
algorithm by which the relative weights are used to determine the
probability of a request going to one server or another.

Yes, same way as RFC2782 does it.


ISTM that it *does* matter. And it especially matters if there is an
implication that the upstream node's algorithm may depend on some
assumption or knowledge about the downstream nodes.

No knowledge, just a reported number that the upstream SIP proxy can
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
starts off with the SRV indicators in RFC2782, without any regard to
additional knowledge or assumption about downstream nodes.
dispatch-adaptive-load-balancing is trying to use the same model, except
render the weight to be sensitive to the current traffic in the cluster
in real-time.


And of course this method of reporting weights isn't sufficient to
drive a balancing algorithm that is sensitive to some requests
requiring more or different resources than others. (That seems to be a
real problem, though I don't know if we have a real solution for
it.)

Yes, that is a real problem, as is the problem of a SIP server in a
cluster that has a low weight and is thus not contacted with enough
frequency to merit more participation to the goodput.

It is well known in SIP that some methods are more onerous to process
than others --- from a proxying point of view, processing an INVITE with
forking and response aggregation is far more detrimental to resource
usage than is processing a BYE.  Some existing literature techniques
thus add a coefficient to the weight of sending an INVITE versus a BYE.
At this time, the process outlined in dispatch-adaptive-load-balancing
treats all requests the same, but we can definitely look at ways to
change this behaviour.

Thanks,

- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

--
Thierry's signature --
Cordially,
Thierry Bessis
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None
--------------020402090103080608090109 Content-Type: image/jpeg; name="al_corp_h_3d_16mm.jpg" Content-Transfer-Encoding: base64 Content-ID: Content-Disposition: inline; filename="al_corp_h_3d_16mm.jpg" /9j/4AAQSkZJRgABAgEAYABgAAD/4QreRXhpZgAATU0AKgAAAAgABwESAAMAAAABAAEAAAEa AAUAAAABAAAAYgEbAAUAAAABAAAAagEoAAMAAAABAAIAAAExAAIAAAAdAAAAcgEyAAIAAAAU AAAAj4dpAAQAAAABAAAApAAAANAAAABgAAAAAQAAAGAAAAABQWRvYmUgUGhvdG9zaG9wIENT IE1hY2ludG9zaAAyMDA4OjAyOjEyIDExOjE5OjU0AAAAA6ABAAMAAAAB//8AAKACAAQAAAAB AAABK6ADAAQAAAABAAAAaAAAAAAAAAAGAQMAAwAAAAEABgAAARoABQAAAAEAAAEeARsABQAA AAEAAAEmASgAAwAAAAEAAgAAAgEABAAAAAEAAAEuAgIABAAAAAEAAAmoAAAAAAAAAEgAAAAB AAAASAAAAAH/2P/gABBKRklGAAECAQBIAEgAAP/tAAxBZG9iZV9DTQAB/+4ADkFkb2JlAGSA AAAAAf/bAIQADAgICAkIDAkJDBELCgsRFQ8MDA8VGBMTFRMTGBEMDAwMDAwRDAwMDAwMDAwM DAwMDAwMDAwMDAwMDAwMDAwMDAENCwsNDg0QDg4QFA4ODhQUDg4ODhQRDAwMDAwREQwMDAwM DBEMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwM/8AAEQgAOACgAwEiAAIRAQMRAf/dAAQA Cv/EAT8AAAEFAQEBAQEBAAAAAAAAAAMAAQIEBQYHCAkKCwEAAQUBAQEBAQEAAAAAAAAAAQAC AwQFBgcICQoLEAABBAEDAgQCBQcGCAUDDDMBAAIRAwQhEjEFQVFhEyJxgTIGFJGhsUIjJBVS wWIzNHKC0UMHJZJT8OHxY3M1FqKygyZEk1RkRcKjdDYX0lXiZfKzhMPTdePzRieUpIW0lcTU 5PSltcXV5fVWZnaGlqa2xtbm9jdHV2d3h5ent8fX5/cRAAICAQIEBAMEBQYHBwYFNQEAAhED ITESBEFRYXEiEwUygZEUobFCI8FS0fAzJGLhcoKSQ1MVY3M08SUGFqKygwcmNcLSRJNUoxdk RVU2dGXi8rOEw9N14/NGlKSFtJXE1OT0pbXF1eX1VmZ2hpamtsbW5vYnN0dXZ3eHl6e3x//a AAwDAQACEQMRAD8A9VSSSSUpJJJJSkkkklKSSSSUpJJJJSkkkklKSSSSUpJJCycrGxKHZGVa yiivV9tjgxo/rPfDUlJUlj1fW76vXE+jleqBy9lVrm/9uNqNf/SWljZeLl1C7FtZfUTG+twc JHLZb+ciYyG4I8wpMkkkgp//0PVUkkLKyacTGtyr3bKaGOtteeAxgL3u/stakppYv1h6Vl9Z yuiU27s/CY2y6uDG12z6D/ov2erV6n9dWOn9Sxeo1WW4xcW03WY79wLffS4028/m72rzfoPV MijL6F1W7p+ZiOzMvJOflOqa3Gtb1V3qY+y/e+zZVczB9P8AR1I1xrczCwup2GjoGV1vqjOq P3mppe197sHGvtYW/oLrt/qbvZ7ElPpiq5PUsPGy8TCuftyM9z2YzIJ3Gtjr7f6rWVsXIdVw vq4bum9A6ZhYeS0nJvx/tV7xh1uDmDMYymt1rMzL/S+zE/7S+/8AmGLHqwsDLp+rfTOsGuzp 1Wd1XFcPUeKgyoZDMaqvJss9f0mbK2Y2+/1PT9NJT6iqmN1PEys7MwaS439PdW3IBaQAbWC+ va/6L/0bl54/EyOqZnXbHnDr6ri5b24eflZtmPfiY9Ya/plmNi14t1bML0rPW9Rt3696tvrL osH9onrf1uHTjUM/fieibZ9Pf9kq+lt9yQFmlPQ39SxaOoYvTrC77RmNtfSACQRTs9Xc783+ eYra5TqeW3C+sn1cu6vfRj2jEzRkWF4rq9Utw94qdcW+3fu9NLF6x0ln1wzsp2djjGyOm476 bzaz03tqtzG3vrt3em70Xfznu9iSnq0l5V9XaMfIw/q1036yONfR7sXKvoque6uu/LOU5+PX kat9X9TsbfjV2fzm/wDsLS6jh9NyOu0dGbRh5vSG9PaOk4+VmW00E77qst+JbVTnfaMupnp1 /wA4y/Gq99KSn0NVsvqGLhvx673EPzLRj47QC4usLXWR7R7Wtrqsse9/sXEYnQ7M/qXQunde yh1A19NynXPoue6rIY2/H+y132tFD8qttZpfbvb+mvo/SeoqDcbp9vTsTpXUCHdKxPrLk4Nd dtjg1lDK8puPj+sX+p7d36P9Ikp9LvvpxqLMi9wrppa6yx54a1o3vcf6rQvHOp9W6n9auqst c19gtfs6bgDRrAf5v2/Q+0OZ+kycl/8AN/yMepei/Wqiuj6i59HTobj04JZSGEuHoMYBDX+7 cz7O1cx9UB0/o3R7frTnRZba52LgUsIL54e1v7l1763fS/msWn1P8JYrGCgJSq5XwxHiUhu9 GweldMyMrB2/bczAxnZHU81riGVWR+gwcNk/T+m97/8Atz+c9PH0MZ9ttTc7HpOF1ypgdZSf azKa3+cot2/orf3a7m+/Ht/7bWI/Nud9TczLsYyn9o5TcSmqoQxtLHera2T+kusts+3PvyLv 0t9lqycDPODhWX02Bp+1VtrqaRILa7X5N3p/ms2Px6N/+Fs/4hTe2ZCRJ1B4f++/wU0X1fDy qszFpy6STVextjJ0MOG6HD81zfzkZY31Qu9f6v49wENsfc9n9V11rmrZVKQqRHYkLX//0fVV C6mnIpfRfW22m1pZZW8BzXNOjmPY72ua5TSSUhtxMS6gY11NdlDdsUvaHMGwh9X6Nw2fo3Ma 6v8AcUXdPwH49uK/Gqdj3uc+6ksaWPc8+pa+2vbssdY/3v3qwkkpoP6B0J+IzCf07FdiVOL6 8c0VmtrjO57Kdnpse7cpO6J0Z2E3p7sDGOEx29uKaazUHane2jZ6W73O/NV1JJTRyeh9FzHV vy+n42Q6hoZS62mt5Y0fRZXvY7Yxv8lWa8bHqtturqYy28g3WNaA55aPTYbXj3WbGDYzeipJ Kamd0npXUtn7RwsfM9KfT+0VMt27o37PVa/bu2NQ7OgdBtxqsS3puI/Gxy40UOorNbC87rPS qLNle9309ivpJKa+VgYObj/ZszHqycfQ+jaxr2e36P6OwOZ7UK7o3R8jFqw78HGtxaI9HHfU x1bIG0elU5vp1+0/mK6kkpBVgYNBqdTj1VOorNNJYxrSys7Saato/R1exn6NihZ0rpduPbi2 YdD8fIebb6XVMLLLCfUddbWW7LLXPG/1Hq0kkpCcPEOGcH0WDENfo+gAAz09vp+i1jfa2v0/ ZsXjHWOi5f1X6g7DymuOI536lmH6FrPzWucPa3KrYPTur+n+ez9EvbULJxsbLofj5VTL6LBD 6rWh7HDn3sfua5SYspxmx9QkGnzTF619WLfqvidN6r9pfdhXvvGPjgt9Uudc5gOR/NtqdXk+ /wDTUXf9/wAyinqP1t6uMfApZjV1MZSPSbNOHjNLixjn+z1bvda9rXfpMm//AEWP/Negn/F7 9Tjb6v7MYDM7WvsDP+2G2ej/AOBrbw8LDwaG42FRXjUN+jVU0MaJ59rAFKeYiLMQeI2fVtHi 34UmSsHDowMOjCxm7aMattVbefa0bWyf3kdJJVlr/9L1VJfKqSSn6qSXyqkkp+qkl8qpJKfq pJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn/9n/7S24 UGhvdG9zaG9wIDMuMAA4QklNBCUAAAAAABAAAAAAAAAAAAAAAAAAAAAAOEJJTQPqAAAAAB2t PD94bWwgdmVyc2lvbj0iMS4wIiBlbmNvZGluZz0iVVRGLTgiPz4KPCFET0NUWVBFIHBsaXN0 IFBVQkxJQyAiLS8vQXBwbGUgQ29tcHV0ZXIvL0RURCBQTElTVCAxLjAvL0VOIiAiaHR0cDov L3d3dy5hcHBsZS5jb20vRFREcy9Qcm9wZXJ0eUxpc3QtMS4wLmR0ZCI+CjxwbGlzdCB2ZXJz aW9uPSIxLjAiPgo8ZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1I b3Jpem9udGFsUmVzPC9rZXk+Cgk8ZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNr ZXQuY3JlYXRvcjwva2V5PgoJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwv c3RyaW5nPgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5pdGVtQXJyYXk8L2tleT4K CQk8YXJyYXk+CgkJCTxkaWN0PgoJCQkJPGtleT5jb20uYXBwbGUucHJpbnQuUGFnZUZvcm1h dC5QTUhvcml6b250YWxSZXM8L2tleT4KCQkJCTxyZWFsPjcyPC9yZWFsPgoJCQkJPGtleT5j b20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5PgoJCQkJPHN0cmluZz5jb20uYXBw bGUucHJpbnRpbmdtYW5hZ2VyPC9zdHJpbmc+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50 aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0OFo8L2Rh dGU+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuc3RhdGVGbGFnPC9rZXk+CgkJ CQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQk8L2RpY3Q+CgkJPC9hcnJheT4KCTwvZGljdD4K CTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1PcmllbnRhdGlvbjwva2V5PgoJ PGRpY3Q+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8 c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5h cHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4K CQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1PcmllbnRhdGlvbjwva2V5 PgoJCQkJPGludGVnZXI+MTwvaW50ZWdlcj4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp Y2tldC5jbGllbnQ8L2tleT4KCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdl cjwvc3RyaW5nPgoJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0Lm1vZERhdGU8L2tl eT4KCQkJCTxkYXRlPjIwMDgtMDItMTJUMTY6MTg6NDhaPC9kYXRlPgoJCQkJPGtleT5jb20u YXBwbGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJPGludGVnZXI+MDwvaW50 ZWdlcj4KCQkJPC9kaWN0PgoJCTwvYXJyYXk+Cgk8L2RpY3Q+Cgk8a2V5PmNvbS5hcHBsZS5w cmludC5QYWdlRm9ybWF0LlBNU2NhbGluZzwva2V5PgoJPGRpY3Q+CgkJPGtleT5jb20uYXBw bGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmlu dGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRl bUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4KCQkJCTxrZXk+Y29tLmFwcGxlLnBy aW50LlBhZ2VGb3JtYXQuUE1TY2FsaW5nPC9rZXk+CgkJCQk8cmVhbD4xPC9yZWFsPgoJCQkJ PGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5PgoJCQkJPHN0cmluZz5j b20uYXBwbGUucHJpbnRpbmdtYW5hZ2VyPC9zdHJpbmc+CgkJCQk8a2V5PmNvbS5hcHBsZS5w cmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0 OFo8L2RhdGU+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuc3RhdGVGbGFnPC9r ZXk+CgkJCQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQk8L2RpY3Q+CgkJPC9hcnJheT4KCTwv ZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFJlczwv a2V5PgoJPGRpY3Q+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tl eT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5 PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8 ZGljdD4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFJl czwva2V5PgoJCQkJPHJlYWw+NzI8L3JlYWw+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50 aWNrZXQuY2xpZW50PC9rZXk+CgkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFn ZXI8L3N0cmluZz4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9r ZXk+CgkJCQk8ZGF0ZT4yMDA4LTAyLTEyVDE2OjE4OjQ4WjwvZGF0ZT4KCQkJCTxrZXk+Y29t LmFwcGxlLnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCTxpbnRlZ2VyPjA8L2lu dGVnZXI+CgkJCTwvZGljdD4KCQk8L2FycmF5PgoJPC9kaWN0PgoJPGtleT5jb20uYXBwbGUu cHJpbnQuUGFnZUZvcm1hdC5QTVZlcnRpY2FsU2NhbGluZzwva2V5PgoJPGRpY3Q+CgkJPGtl eT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8c3RyaW5nPmNvbS5h cHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50 aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4KCQkJCTxrZXk+Y29t LmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFNjYWxpbmc8L2tleT4KCQkJCTxy ZWFsPjE8L3JlYWw+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuY2xpZW50PC9r ZXk+CgkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJ CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJCQk8ZGF0ZT4y MDA4LTAyLTEyVDE2OjE4OjQ4WjwvZGF0ZT4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp Y2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCTxpbnRlZ2VyPjA8L2ludGVnZXI+CgkJCTwvZGlj dD4KCQk8L2FycmF5PgoJPC9kaWN0PgoJPGtleT5jb20uYXBwbGUucHJpbnQuc3ViVGlja2V0 LnBhcGVyX2luZm9fdGlja2V0PC9rZXk+Cgk8ZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmlu dC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYWdlUmVjdDwva2V5PgoJCTxkaWN0PgoJCQk8a2V5 PmNvbS5hcHBsZS5wcmludC50aWNrZXQuY3JlYXRvcjwva2V5PgoJCQk8c3RyaW5nPmNvbS5h cHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJPGtleT5jb20uYXBwbGUucHJpbnQu dGlja2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4KCQkJCQk8a2V5 PmNvbS5hcHBsZS5wcmludC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYWdlUmVjdDwva2V5PgoJ CQkJCTxhcnJheT4KCQkJCQkJPHJlYWw+MC4wPC9yZWFsPgoJCQkJCQk8cmVhbD4wLjA8L3Jl YWw+CgkJCQkJCTxyZWFsPjczNDwvcmVhbD4KCQkJCQkJPHJlYWw+NTc2PC9yZWFsPgoJCQkJ CTwvYXJyYXk+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5 PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwvc3RyaW5nPgoJCQkJ CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJCQkJPGRhdGU+ MjAwOC0wMi0xMlQxNjoxODo0OVo8L2RhdGU+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQu dGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJCTxpbnRlZ2VyPjA8L2ludGVnZXI+CgkJCQk8 L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0PgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBh Z2VGb3JtYXQuUE1BZGp1c3RlZFBhcGVyUmVjdDwva2V5PgoJCTxkaWN0PgoJCQk8a2V5PmNv bS5hcHBsZS5wcmludC50aWNrZXQuY3JlYXRvcjwva2V5PgoJCQk8c3RyaW5nPmNvbS5hcHBs ZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlj a2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4KCQkJCQk8a2V5PmNv bS5hcHBsZS5wcmludC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYXBlclJlY3Q8L2tleT4KCQkJ CQk8YXJyYXk+CgkJCQkJCTxyZWFsPi0xODwvcmVhbD4KCQkJCQkJPHJlYWw+LTE4PC9yZWFs PgoJCQkJCQk8cmVhbD43NzQ8L3JlYWw+CgkJCQkJCTxyZWFsPjU5NDwvcmVhbD4KCQkJCQk8 L2FycmF5PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5jbGllbnQ8L2tleT4K CQkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJCQk8 a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJCTxkYXRlPjIw MDgtMDItMTJUMTY6MTg6NDlaPC9kYXRlPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp Y2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQkJPC9k aWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC5QYXBl ckluZm8uUE1QYXBlck5hbWU8L2tleT4KCQk8ZGljdD4KCQkJPGtleT5jb20uYXBwbGUucHJp bnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmluZz5jb20uYXBwbGUucHJpbnQucG0u UG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRl bUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0PgoJCQkJCTxrZXk+Y29tLmFwcGxl LnByaW50LlBhcGVySW5mby5QTVBhcGVyTmFtZTwva2V5PgoJCQkJCTxzdHJpbmc+bmEtbGV0 dGVyPC9zdHJpbmc+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwv a2V5PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50LnBtLlBvc3RTY3JpcHQ8L3N0cmlu Zz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJ CTxkYXRlPjIwMDAtMDctMjhUMjI6NTc6MDRaPC9kYXRlPgoJCQkJCTxrZXk+Y29tLmFwcGxl LnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4xPC9pbnRlZ2Vy PgoJCQkJPC9kaWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5w cmludC5QYXBlckluZm8uUE1VbmFkanVzdGVkUGFnZVJlY3Q8L2tleT4KCQk8ZGljdD4KCQkJ PGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmluZz5j b20uYXBwbGUucHJpbnQucG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5hcHBs ZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0PgoJ CQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYWdlUmVj dDwva2V5PgoJCQkJCTxhcnJheT4KCQkJCQkJPHJlYWw+MC4wPC9yZWFsPgoJCQkJCQk8cmVh bD4wLjA8L3JlYWw+CgkJCQkJCTxyZWFsPjczNDwvcmVhbD4KCQkJCQkJPHJlYWw+NTc2PC9y ZWFsPgoJCQkJCTwvYXJyYXk+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNs aWVudDwva2V5PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwvc3Ry aW5nPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJ CQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0OFo8L2RhdGU+CgkJCQkJPGtleT5jb20uYXBw bGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJCTxpbnRlZ2VyPjA8L2ludGVn ZXI+CgkJCQk8L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0PgoJCTxrZXk+Y29tLmFwcGxl LnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYXBlclJlY3Q8L2tleT4KCQk8ZGljdD4K CQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmlu Zz5jb20uYXBwbGUucHJpbnQucG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5h cHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0 PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYXBl clJlY3Q8L2tleT4KCQkJCQk8YXJyYXk+CgkJCQkJCTxyZWFsPi0xODwvcmVhbD4KCQkJCQkJ PHJlYWw+LTE4PC9yZWFsPgoJCQkJCQk8cmVhbD43NzQ8L3JlYWw+CgkJCQkJCTxyZWFsPjU5 NDwvcmVhbD4KCQkJCQk8L2FycmF5PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tl dC5jbGllbnQ8L2tleT4KCQkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8 L3N0cmluZz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5 PgoJCQkJCTxkYXRlPjIwMDgtMDItMTJUMTY6MTg6NDhaPC9kYXRlPgoJCQkJCTxrZXk+Y29t LmFwcGxlLnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4wPC9p bnRlZ2VyPgoJCQkJPC9kaWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5h cHBsZS5wcmludC5QYXBlckluZm8ucHBkLlBNUGFwZXJOYW1lPC9rZXk+CgkJPGRpY3Q+CgkJ CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5jcmVhdG9yPC9rZXk+CgkJCTxzdHJpbmc+ Y29tLmFwcGxlLnByaW50LnBtLlBvc3RTY3JpcHQ8L3N0cmluZz4KCQkJPGtleT5jb20uYXBw bGUucHJpbnQudGlja2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4K CQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC5QYXBlckluZm8ucHBkLlBNUGFwZXJOYW1lPC9r ZXk+CgkJCQkJPHN0cmluZz5MZXR0ZXI8L3N0cmluZz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5w cmludC50aWNrZXQuY2xpZW50PC9rZXk+CgkJCQkJPHN0cmluZz5jb20uYXBwbGUucHJpbnQu cG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tl dC5tb2REYXRlPC9rZXk+CgkJCQkJPGRhdGU+MjAwMC0wNy0yOFQyMjo1NzowNFo8L2RhdGU+ CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJ CTxpbnRlZ2VyPjE8L2ludGVnZXI+CgkJCQk8L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0 PgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5BUElWZXJzaW9uPC9rZXk+CgkJPHN0 cmluZz4wMC4yMDwvc3RyaW5nPgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5wcml2 YXRlTG9jazwva2V5PgoJCTxmYWxzZS8+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0 LnR5cGU8L2tleT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludC5QYXBlckluZm9UaWNrZXQ8 L3N0cmluZz4KCTwvZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5BUElWZXJz aW9uPC9rZXk+Cgk8c3RyaW5nPjAwLjIwPC9zdHJpbmc+Cgk8a2V5PmNvbS5hcHBsZS5wcmlu dC50aWNrZXQucHJpdmF0ZUxvY2s8L2tleT4KCTxmYWxzZS8+Cgk8a2V5PmNvbS5hcHBsZS5w cmludC50aWNrZXQudHlwZTwva2V5PgoJPHN0cmluZz5jb20uYXBwbGUucHJpbnQuUGFnZUZv cm1hdFRpY2tldDwvc3RyaW5nPgo8L2RpY3Q+CjwvcGxpc3Q+CgA4QklNA+kAAAAAAHgAAwAA AEgASAAAAAAC3gJA/+7/7gMGAlIDZwUoA/wAAgAAAEgASAAAAAAC2AIoAAEAAABkAAAAAQAD AwMAAAABf/8AAQABAAAAAAAAAAAAAAAAaAgAGQGQAAAAAAAgAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAA4QklNA+0AAAAAABAAYAAAAAEAAQBgAAAAAQABOEJJTQQmAAAAAAAOAAAAAAAA AAAAAD+AAAA4QklNBA0AAAAAAAQAAAB4OEJJTQQZAAAAAAAEAAAAHjhCSU0D8wAAAAAACQAA AAAAAAAAAQA4QklNBAoAAAAAAAEAADhCSU0nEAAAAAAACgABAAAAAAAAAAE4QklNA/UAAAAA AEgAL2ZmAAEAbGZmAAYAAAAAAAEAL2ZmAAEAoZmaAAYAAAAAAAEAMgAAAAEAWgAAAAYAAAAA AAEANQAAAAEALQAAAAYAAAAAAAE4QklNA/gAAAAAAHAAAP////////////////////////// //8D6AAAAAD/////////////////////////////A+gAAAAA//////////////////////// /////wPoAAAAAP////////////////////////////8D6AAAOEJJTQQIAAAAAAAQAAAAAQAA AkAAAAJAAAAAADhCSU0EHgAAAAAABAAAAAA4QklNBBoAAAAAA0kAAAAGAAAAAAAAAAAAAABo AAABKwAAAAoAVQBuAHQAaQB0AGwAZQBkAC0AMQAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAQAA AAAAAAAAAAABKwAAAGgAAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAQAAAA AQAAAAAAAG51bGwAAAACAAAABmJvdW5kc09iamMAAAABAAAAAAAAUmN0MQAAAAQAAAAAVG9w IGxvbmcAAAAAAAAAAExlZnRsb25nAAAAAAAAAABCdG9tbG9uZwAAAGgAAAAAUmdodGxvbmcA AAErAAAABnNsaWNlc1ZsTHMAAAABT2JqYwAAAAEAAAAAAAVzbGljZQAAABIAAAAHc2xpY2VJ RGxvbmcAAAAAAAAAB2dyb3VwSURsb25nAAAAAAAAAAZvcmlnaW5lbnVtAAAADEVTbGljZU9y aWdpbgAAAA1hdXRvR2VuZXJhdGVkAAAAAFR5cGVlbnVtAAAACkVTbGljZVR5cGUAAAAASW1n IAAAAAZib3VuZHNPYmpjAAAAAQAAAAAAAFJjdDEAAAAEAAAAAFRvcCBsb25nAAAAAAAAAABM ZWZ0bG9uZwAAAAAAAAAAQnRvbWxvbmcAAABoAAAAAFJnaHRsb25nAAABKwAAAAN1cmxURVhU AAAAAQAAAAAAAG51bGxURVhUAAAAAQAAAAAAAE1zZ2VURVhUAAAAAQAAAAAABmFsdFRhZ1RF WFQAAAABAAAAAAAOY2VsbFRleHRJc0hUTUxib29sAQAAAAhjZWxsVGV4dFRFWFQAAAABAAAA AAAJaG9yekFsaWduZW51bQAAAA9FU2xpY2VIb3J6QWxpZ24AAAAHZGVmYXVsdAAAAAl2ZXJ0 QWxpZ25lbnVtAAAAD0VTbGljZVZlcnRBbGlnbgAAAAdkZWZhdWx0AAAAC2JnQ29sb3JUeXBl ZW51bQAAABFFU2xpY2VCR0NvbG9yVHlwZQAAAABOb25lAAAACXRvcE91dHNldGxvbmcAAAAA AAAACmxlZnRPdXRzZXRsb25nAAAAAAAAAAxib3R0b21PdXRzZXRsb25nAAAAAAAAAAtyaWdo dE91dHNldGxvbmcAAAAAADhCSU0EKAAAAAAADAAAAAE/8AAAAAAAADhCSU0EFAAAAAAABAAA AAM4QklNBAwAAAAACcQAAAABAAAAoAAAADgAAAHgAABpAAAACagAGAAB/9j/4AAQSkZJRgAB AgEASABIAAD/7QAMQWRvYmVfQ00AAf/uAA5BZG9iZQBkgAAAAAH/2wCEAAwICAgJCAwJCQwR CwoLERUPDAwPFRgTExUTExgRDAwMDAwMEQwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwB DQsLDQ4NEA4OEBQODg4UFA4ODg4UEQwMDAwMEREMDAwMDAwRDAwMDAwMDAwMDAwMDAwMDAwM DAwMDAwMDAwMDP/AABEIADgAoAMBIgACEQEDEQH/3QAEAAr/xAE/AAABBQEBAQEBAQAAAAAA AAADAAECBAUGBwgJCgsBAAEFAQEBAQEBAAAAAAAAAAEAAgMEBQYHCAkKCxAAAQQBAwIEAgUH BggFAwwzAQACEQMEIRIxBUFRYRMicYEyBhSRobFCIyQVUsFiMzRygtFDByWSU/Dh8WNzNRai soMmRJNUZEXCo3Q2F9JV4mXys4TD03Xj80YnlKSFtJXE1OT0pbXF1eX1VmZ2hpamtsbW5vY3 R1dnd4eXp7fH1+f3EQACAgECBAQDBAUGBwcGBTUBAAIRAyExEgRBUWFxIhMFMoGRFKGxQiPB UtHwMyRi4XKCkkNTFWNzNPElBhaisoMHJjXC0kSTVKMXZEVVNnRl4vKzhMPTdePzRpSkhbSV xNTk9KW1xdXl9VZmdoaWprbG1ub2JzdHV2d3h5ent8f/2gAMAwEAAhEDEQA/APVUkkklKSSS SUpJJJJSkkkklKSSSSUpJJJJSkkkklKSSQsnKxsSh2RlWsoor1fbY4MaP6z3w1JSVJY9X1u+ r1xPo5XqgcvZVa5v/bjajX/0lpY2Xi5dQuxbWX1ExvrcHCRy2W/nImMhuCPMKTJJJIKf/9D1 VJJCysmnExrcq92ymhjrbXngMYC97v7LWpKaWL9YelZfWcrolNu7PwmNsurgxtds+g/6L9nq 1ep/XVjp/UsXqNVluMXFtN1mO/cC330uNNvP5u9q836D1TIoy+hdVu6fmYjszLyTn5Tqmtxr W9Vd6mPsv3vs2VXMwfT/AEdSNca3MwsLqdho6Bldb6ozqj95qaXtfe7Bxr7WFv6C67f6m72e xJT6YquT1LDxsvEwrn7cjPc9mMyCdxrY6+3+q1lbFyHVcL6uG7pvQOmYWHktJyb8f7Ve8Ydb g5gzGMprdazMy/0vsxP+0vv/AJhix6sLAy6fq30zrBrs6dVndVxXD1HioMqGQzGqrybLPX9J mytmNvv9T0/TSU+oqpjdTxMrOzMGkuN/T3VtyAWkAG1gvr2v+i/9G5eePxMjqmZ12x5w6+q4 uW9uHn5WbZj34mPWGv6ZZjYteLdWzC9Kz1vUbd+verb6y6LB/aJ639bh041DP34nom2fT3/Z KvpbfckBZpT0N/UsWjqGL06wu+0ZjbX0gAkEU7PV3O/N/nmK2uU6nltwvrJ9XLur30Y9oxM0 ZFheK6vVLcPeKnXFvt37vTSxesdJZ9cM7KdnY4xsjpuO+m82s9N7arcxt767d3pu9F3857vY kp6tJeVfV2jHyMP6tdN+sjjX0e7Fyr6KrnurrvyzlOfj15GrfV/U7G341dn85v8A7C0uo4fT cjrtHRm0Yeb0hvT2jpOPlZltNBO+6rLfiW1U532jLqZ6df8AOMvxqvfSkp9DVbL6hi4b8eu9 xD8y0Y+O0AuLrC11ke0e1ra6rLHvf7FxGJ0OzP6l0Lp3XsodQNfTcp1z6LnuqyGNvx/std9r RQ/KrbWaX272/pr6P0nqKg3G6fb07E6V1Ah3SsT6y5ODXXbY4NZQyvKbj4/rF/qe3d+j/SJK fS776caizIvcK6aWusseeGtaN73H+q0LxzqfVup/WrqrLXNfYLX7Om4A0awH+b9v0PtDmfpM nJf/ADf8jHqXov1qoro+oufR06G49OCWUhhLh6DGAQ1/u3M+ztXMfVAdP6N0e36050WW2udi 4FLCC+eHtb+5de+t30v5rFp9T/CWKxgoCUquV8MR4lIbvRsHpXTMjKwdv23MwMZ2R1PNa4hl VkfoMHDZP0/pve//ALc/nPTx9DGfbbU3Ox6ThdcqYHWUn2symt/nKLdv6K392u5vvx7f+21i PzbnfU3My7GMp/aOU3EpqqEMbSx3q2tk/pLrLbPtz78i79LfZasnAzzg4Vl9NgaftVba6mkS C2u1+Td6f5rNj8ejf/hbP+IU3tmQkSdQeH/vv8FNF9Xw8qrMxacukk1XsbYydDDhuhw/Nc38 5GWN9ULvX+r+PcBDbH3PZ/Vdda5q2VSkKkR2JC1//9H1VQuppyKX0X1ttptaWWVvAc1zTo5j 2O9rmuU0klIbcTEuoGNdTXZQ3bFL2hzBsIfV+jcNn6NzGur/AHFF3T8B+PbivxqnY97nPupL Glj3PPqWvtr27LHWP9796sJJKaD+gdCfiMwn9OxXYlTi+vHNFZra4zueynZ6bHu3KTuidGdh N6e7AxjhMdvbimms1B2p3to2elu9zvzVdSSU0cnofRcx1b8vp+NkOoaGUutpreWNH0WV72O2 Mb/JVmvGx6rbbq6mMtvIN1jWgOeWj02G1491mxg2M3oqSSmpndJ6V1LZ+0cLHzPSn0/tFTLd u6N+z1Wv27tjUOzoHQbcarEt6biPxscuNFDqKzWwvO6z0qizZXvd9PYr6SSmvlYGDm4/2bMx 6snH0Po2sa9nt+j+jsDme1Cu6N0fIxasO/BxrcWiPRx31MdWyBtHpVOb6dftP5iupJKQVYGD QanU49VTqKzTSWMa0srO0mmraP0dXsZ+jYoWdK6Xbj24tmHQ/HyHm2+l1TCyywn1HXW1luyy 1zxv9R6tJJKQnDxDhnB9FgxDX6PoAAM9Pb6fotY32tr9P2bF4x1jouX9V+oOw8prjiOd+pZh +haz81rnD2tyq2D07q/p/ns/RL21CycbGy6H4+VUy+iwQ+q1oexw597H7muUmLKcZsfUJBp8 0xetfVi36r4nTeq/aX3YV77xj44LfVLnXOYDkfzbanV5Pv8A01F3/f8AMop6j9berjHwKWY1 dTGUj0mzTh4zS4sY5/s9W73Wva136TJv/wBFj/zXoJ/xe/U42+r+zGAzO1r7Az/thtno/wDg a28PCw8GhuNhUV41Dfo1VNDGiefawBSnmIizEHiNn1bR4t+FJkrBw6MDDowsZu2jGrbVW3n2 tG1sn95HSSVZa//S9VSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn 6qSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp//ZOEJJTQQhAAAAAABTAAAAAQEAAAAPAEEA ZABvAGIAZQAgAFAAaABvAHQAbwBzAGgAbwBwAAAAEgBBAGQAbwBiAGUAIABQAGgAbwB0AG8A cwBoAG8AcAAgAEMAUwAAAAEAOEJJTQQGAAAAAAAHAAgAAAABAQD/4RgFaHR0cDovL25zLmFk b2JlLmNvbS94YXAvMS4wLwA8P3hwYWNrZXQgYmVnaW49J++7vycgaWQ9J1c1TTBNcENlaGlI enJlU3pOVGN6a2M5ZCc/Pgo8eDp4bXBtZXRhIHhtbG5zOng9J2Fkb2JlOm5zOm1ldGEvJyB4 OnhtcHRrPSdYTVAgdG9vbGtpdCAzLjAtMjgsIGZyYW1ld29yayAxLjYnPgo8cmRmOlJERiB4 bWxuczpyZGY9J2h0dHA6Ly93d3cudzMub3JnLzE5OTkvMDIvMjItcmRmLXN5bnRheC1ucyMn IHhtbG5zOmlYPSdodHRwOi8vbnMuYWRvYmUuY29tL2lYLzEuMC8nPgoKIDxyZGY6RGVzY3Jp cHRpb24gcmRmOmFib3V0PSd1dWlkOmE0NDZmNTA1LWRhZWQtMTFkYy05N2EzLWRjOWRjYjU2 NWRiYycKICB4bWxuczpleGlmPSdodHRwOi8vbnMuYWRvYmUuY29tL2V4aWYvMS4wLyc+CiAg PGV4aWY6Q29sb3JTcGFjZT40Mjk0OTY3Mjk1PC9leGlmOkNvbG9yU3BhY2U+CiAgPGV4aWY6 UGl4ZWxYRGltZW5zaW9uPjI5OTwvZXhpZjpQaXhlbFhEaW1lbnNpb24+CiAgPGV4aWY6UGl4 ZWxZRGltZW5zaW9uPjEwNDwvZXhpZjpQaXhlbFlEaW1lbnNpb24+CiA8L3JkZjpEZXNjcmlw dGlvbj4KCiA8cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0ndXVpZDphNDQ2ZjUwNS1kYWVk LTExZGMtOTdhMy1kYzlkY2I1NjVkYmMnCiAgeG1sbnM6cGRmPSdodHRwOi8vbnMuYWRvYmUu Y29tL3BkZi8xLjMvJz4KIDwvcmRmOkRlc2NyaXB0aW9uPgoKIDxyZGY6RGVzY3JpcHRpb24g cmRmOmFib3V0PSd1dWlkOmE0NDZmNTA1LWRhZWQtMTFkYy05N2EzLWRjOWRjYjU2NWRiYycK ICB4bWxuczpwaG90b3Nob3A9J2h0dHA6Ly9ucy5hZG9iZS5jb20vcGhvdG9zaG9wLzEuMC8n PgogIDxwaG90b3Nob3A6SGlzdG9yeT48L3Bob3Rvc2hvcDpIaXN0b3J5PgogPC9yZGY6RGVz Y3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUt ZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnRpZmY9J2h0dHA6Ly9ucy5h ZG9iZS5jb20vdGlmZi8xLjAvJz4KICA8dGlmZjpPcmllbnRhdGlvbj4xPC90aWZmOk9yaWVu dGF0aW9uPgogIDx0aWZmOlhSZXNvbHV0aW9uPjk2LzE8L3RpZmY6WFJlc29sdXRpb24+CiAg PHRpZmY6WVJlc29sdXRpb24+OTYvMTwvdGlmZjpZUmVzb2x1dGlvbj4KICA8dGlmZjpSZXNv bHV0aW9uVW5pdD4yPC90aWZmOlJlc29sdXRpb25Vbml0PgogPC9yZGY6RGVzY3JpcHRpb24+ CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUtZGFlZC0xMWRj LTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnhhcD0naHR0cDovL25zLmFkb2JlLmNvbS94 YXAvMS4wLyc+CiAgPHhhcDpDcmVhdGVEYXRlPjIwMDgtMDItMTJUMTE6MTk6NTQtMDU6MDA8 L3hhcDpDcmVhdGVEYXRlPgogIDx4YXA6TW9kaWZ5RGF0ZT4yMDA4LTAyLTEyVDExOjE5OjU0 LTA1OjAwPC94YXA6TW9kaWZ5RGF0ZT4KICA8eGFwOk1ldGFkYXRhRGF0ZT4yMDA4LTAyLTEy VDExOjE5OjU0LTA1OjAwPC94YXA6TWV0YWRhdGFEYXRlPgogIDx4YXA6Q3JlYXRvclRvb2w+ QWRvYmUgUGhvdG9zaG9wIENTIE1hY2ludG9zaDwveGFwOkNyZWF0b3JUb29sPgogPC9yZGY6 RGVzY3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1 MDUtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnhhcE1NPSdodHRwOi8v bnMuYWRvYmUuY29tL3hhcC8xLjAvbW0vJz4KICA8eGFwTU06RG9jdW1lbnRJRD5hZG9iZTpk b2NpZDpwaG90b3Nob3A6YTQ0NmY1MDQtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjPC94 YXBNTTpEb2N1bWVudElEPgogPC9yZGY6RGVzY3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlv biByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJj JwogIHhtbG5zOmRjPSdodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyc+CiAgPGRj OmZvcm1hdD5pbWFnZS9qcGVnPC9kYzpmb3JtYXQ+CiA8L3JkZjpEZXNjcmlwdGlvbj4KCjwv cmRmOlJERj4KPC94OnhtcG1ldGE+CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAog ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAK ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIAo8P3hwYWNrZXQgZW5kPSd3Jz8+/+ICOElDQ19QUk9GSUxFAAEBAAAC KEFEQkUCEAAAbW50clJHQiBYWVogB88ABgADAAAAAAAAYWNzcEFQUEwAAAAAbm9uZQAAAAAA AAAAAAAAAAAAAAAAAPbWAAEAAAAA0y1BREJFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAKY3BydAAAAPwAAAAyZGVzYwAAATAAAABkd3RwdAAAAZQA AAAUYmtwdAAAAagAAAAUclRSQwAAAbwAAAAOZ1RSQwAAAcwAAAAOYlRSQwAAAdwAAAAOclhZ WgAAAewAAAAUZ1hZWgAAAgAAAAAUYlhZWgAAAhQAAAAUdGV4dAAAAABDb3B5cmlnaHQgMTk5 OSBBZG9iZSBTeXN0ZW1zIEluY29ycG9yYXRlZAAAAGRlc2MAAAAAAAAACkFwcGxlIFJHQgAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAABYWVogAAAAAAAA81EAAQAAAAEWzFhZWiAAAAAAAAAA AAAAAAAAAAAAY3VydgAAAAAAAAABAc0AAGN1cnYAAAAAAAAAAQHNAABjdXJ2AAAAAAAAAAEB zQAAWFlaIAAAAAAAAHm9AABBUgAABLlYWVogAAAAAAAAVvgAAKwvAAAdA1hZWiAAAAAAAAAm IgAAEn8AALFw/+4ADkFkb2JlAGRAAAAAAf/bAIQAAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEB AQEBAQEBAQEBAQEBAQEBAQEBAQICAgICAgICAgICAwMDAwMDAwMDAwEBAQEBAQEBAQEBAgIB AgIDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMD/8AA EQgAaAErAwERAAIRAQMRAf/dAAQAJv/EAaIAAAAGAgMBAAAAAAAAAAAAAAcIBgUECQMKAgEA CwEAAAYDAQEBAAAAAAAAAAAABgUEAwcCCAEJAAoLEAACAQMEAQMDAgMDAwIGCXUBAgMEEQUS BiEHEyIACDEUQTIjFQlRQhZhJDMXUnGBGGKRJUOhsfAmNHIKGcHRNSfhUzaC8ZKiRFRzRUY3 R2MoVVZXGrLC0uLyZIN0k4Rlo7PD0+MpOGbzdSo5OkhJSlhZWmdoaWp2d3h5eoWGh4iJipSV lpeYmZqkpaanqKmqtLW2t7i5usTFxsfIycrU1dbX2Nna5OXm5+jp6vT19vf4+foRAAIBAwIE BAMFBAQEBgYFbQECAxEEIRIFMQYAIhNBUQcyYRRxCEKBI5EVUqFiFjMJsSTB0UNy8BfhgjQl klMYY0TxorImNRlUNkVkJwpzg5NGdMLS4vJVZXVWN4SFo7PD0+PzKRqUpLTE1OT0laW1xdXl 9ShHV2Y4doaWprbG1ub2Z3eHl6e3x9fn90hYaHiImKi4yNjo+DlJWWl5iZmpucnZ6fkqOkpa anqKmqq6ytrq+v/aAAwDAQACEQMRAD8A3+Pfuvde9+691737r3Xvfuvde9+691737r3Xvfuv de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+ 691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917 37r3Xvfuvde9+691737r3Xvfuvde9+691//Q3+Pfuvde9+691737r3Xvfuvde9+691737r3X vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+ 691737r3Xvfuvde9+691737r3Xvfuvde9+691//R3+Pfuvde9+691737r3Xvfuvde9+69173 7r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3X vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691//S3+Pfuvde9+691737r3Xvfuvde9+6 91737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+69173 7r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3XvfuvdNG4NwYHaeCzO 6N05vEba2zt3F12b3BuLcGSo8NgsFhcXTS1uTy+Zy+RmpsfjMXjqOF5p6ieRIoYkZnYKCfd4 45JpEhhjZ5WICqASSTgAAZJJwAOPXgCTQcetWb5p/wDCnzqPr/NZPYPwj66g73zeOqKiiq+3 ewTmdudTR1dO3jvtjbdC2N3rvuiMoZWqJKjAU50h6d6qJ1k9zVyz7N3t+sdzzFdm2hND4cdG lp/SY1RD8qOfUAinS1LNiKyGg9PPqr3bf8+D+Yx2buCN63vzDdaU1TK9sPsjp7p6fFU6zMpV Ipd/7I31mWSnAITXWMxBOosbESintLyVbQf8kdp3A4vNMCf+ccka/wAunPAhXBX+Z6td6K+e f8yLN09HuDb3bvV/e1OtPF93tPsnq/bWAqayMa2mkpMp1PH17LS5KZX0o7JLToVUmBxqVwXv HI/I8JaGbb7izNcPFKzAfaJfEqP2H59NPHAMUI6s7+Pn80/YO89y4rqr5NbFyvxb7fytemJw y7lr/wCM9T7wyE8vix9Jt3sf7PGR4nL5Ij0UWZpaBTK6QU9TVzOFMd757dX1lBJuOxXS7hti irFBSVB5loqmoH8SFsVLKo6aaAgakOpf5/s6ta9xz0x1737r3Xvfuvde9+691737r3X/09/j 37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3WKSoghenjmmiikqpW p6VJJER6mdYJqpoadWYNNKtNTSSFVuRHGzWspI917rL7917r3v3Xuve/de697917r3v3Xuve /de697917r3v3Xuve/de697917r3v3Xuve/de697917r3v3XuscksUKh5pI4kMkUQaR1RTLP KkEEYZiAZJppFRB9WZgByR7917rJ7917r3v3Xuvnkfz4P5ue7vmJ3Ruz4odFbqqcb8UOo9yV G3c/WYGtlhTvnsHbtaafM5/M1dLL4cp11tvN0rwYGkQyUla8H8VkaUy0aUeTPtpyRFtdpDvF /DXdplqtR/ZIeAHo7DLniK6MUbUcWltoUSMP1D/If5+qKNq7fqayqocdj6Oor8hXVFPR0VFR 08tTV1lZVSpBTUlLTQI81RU1EzqkaICzsQACT7myKNY0LuwCAVJPAAcST8vM9LCmMjrcm+Iv 8tv4w/yzfjinzf8A5mGLxG6Oz5aWCfYnReYpsfuDGbczmTpGrNu7PptqVh/h2++4MhHTyPOK ppMRgoxJKwBpJK+OB99505g563scpchyNHY1pJcKSpZQaM5cZjhFRSlHkNB+IISmSRp5DHB8 Pr/q8urEfmF2/wBa7F+FHx3+Wk3Te0Olew+1dxdZVeCwG34MXHlsZh9+bbzu8YNtZ3LYnEbe bNfabRohLUJNShKatBVEQ3JDvJ22bhe84b7ywN1lu9vto5gzMW0lo3WPWqszaauaChyvEnpO AdbpWoHSu2Vjfjt/ME6Vm23ubH4WXc9RiZIoaiSKmM7SPDpJBseWbgE3IPtjcRvvI27C4gZ/ o9WRnrys0bVB6kfAXvjsLpXt6t/l+fIXO1udqaDGZHKfGXsDO1ktXltw7YwFPJV5jqbMV9Tq qMpltp4Onevw9S7ySS4mnqaeRlFFT+ZBzpstluW2pznskISNmAuo1FArsaCVQMAO3a48nKkf EaWlRWXxU/Mf5erofcVdJ+ve/de697917r3v3Xuv/9Tf49+691737r3XvfuvdNuZzGL29h8r n83WwYzC4PG12Yy+Sqm0U2PxeMpZa2vral7HRBSUkDyObcKp9+6918s3Ed59yYz5c7d/4UC1 xy1F1Fn/AOahuDr6tpp6KaqyFFs6bbtBvWq2VBURkvUQH4+5eu2/ABHqY0D6GZ45Au+nfLT8 ut6b/hQBW0eR/kz/ADWyGPq6avoK/rzr2toa6iniqqOto6ruDrSemq6Spgd4aimqIXV45EYq 6kEEg+9dUX4h1D/lP90dXfHb+R98SO7e6t6Ybrzq3rf43UO4t47vz07Q4/E42HPZaniVY4km q6/JZKvqIaShoqaOarr62eKmp45Z5Y429145Y9EL/wCgrz4l5bcGbr9ifEb5r9gdJ7XyFTQb l7o251/tOXHY2OjSCafJtg23g0FJj/tJvuQuRyWOrFpjG8kEbOUT3W9B9erPPnh/OD+O/wAC 63prae6etvkF3Z2t37tebePWPU3Q3XUO8d25Pb0DUsUmRyTZDOYTG0kaz1QjMNLLXVupSfB4 7OfdVCk9Fe+J/wDwoa+OHf3yR2Z8S+5/j98lvhl3h2ZW0mO61w3yE2ZTYLB7uyuWllg25gly Bq6PPYjN7pqIvBjRVYpKCsqyKdKszvDHL7rZXFQeqtv5sn8wfKfDX/hQx8NN8dsVnd3Yfxx6 E+Ped3zTdH9S01DuXJnefbXWHyU6qrN2bZ2Nmty7Q23ktxO+YoPvayprYqiHE0soidgohk31 tRVT69Xxw/zovhtjfgPiP5ifZCdr9MdNbk3JuLZe1tkdn7NxOP7q3TvPb+Yz2Gj2rgdkbZ3V unGVmWzr7bqqmkJyiUyUMZqKqWmiWRk11XSa06Il1t/wp1+KOY331/g++fjT8uvid1l3BkKW j6s767u68xeL6uztHWpRPT5zM5Wizcz4/AWr4mlrsZ/G6OlhlSeolhpy8qe63pPr0V7/AIVG VVNXdhfydq2iqYKyjrPlNuCqpKulmjqKaqpqjN9DywVNNPEzxTwTxOGR1JVlIIJB9+9etpwb rb39+6p1W9/OF/7dY/zAP/FUu5f/AHjsl791scR0VH/hNd/25U+F/wD5cX/8Fj3t791tviPR tfk7/M76E+KHy2+Kfw07E2j29mez/mBlqDD9aZ7ZeA2ZkdhYOpyO6qbaEEm+cpnN/bc3DjIF yVUrucfi8mwgBYKXsh91oAkE9Cv87/m11V/L0+Ne8PlN3Rt/sHc/X2ysttHD5XD9X4rbma3j UVO9NzYzauLkx+P3XuvZWElggyOVjeoMmQiZIVYorsAh914Cpp0kuwP5kPxY6i+EXXfz97f3 dkOsOjO0+rusu0dm0G6qCkk7Fy8Xbm0MZvTZmxaDaW3snnUzPYdTisoBPQ46qraen8FRO1R9 nTy1S+69Q1p1UP0f/wAKevjj3d2rsDaGP+H3zM231d2l2NtPrDZfeOW2TtKq2iu4N67potnY Kp3SuO3XPi8Ni/43koUqWo8nk6iEFtMMjroPut6TTj0fX+Yl/OY+MH8u/eOz+mdx7d7R75+S /YePpMvs348dEbYXdO+ajEZGpr6TG5jOz1NVRY/DUWTqcXUpTU8RrcrUGIvHRNBeZfdeCk/Z 0kPgJ/PC+Mvzl7jy/wAYsp193P8AFj5TYfFyZdOi/kXtem2nuDc1JS4tM3k12dVxVsz5HIYr CyLXS0NfTYvIy0Gqqp6eanhqJYfdeKkZ8uk/8mv+FAnwR+Ivyg+Q3xP7ux/d23+wPjnsPb+9 s7nqHZ+y8ps3f9bu3aXW28NsbA6ukg7HXd+f35mqHtChiSCtw+MoIJKWtlqKyGjpzVN7r2kk A9A18O/+FJvxA+UvyT2v8Vt59Sd+/GHs/sXN4/b3WLd0YHb9LgN053OeMbW23kqnFZyfL7Y3 NvCWoiixcU1DLQVc8qQrWCWSBJfdeKkCvRDf583zq3Z8TP5uf8sqr3dne3NxfGvrHbNB31u7 pLquKjyVdu7eCb07H2qM9SbVrs1tjE7r3NR4mgp4aFcjXJFRRiZqcxNPOZfdbUVU+vWxX/L1 /mCdVfzI+l9z949Qdd92dZ7a2r2fmuqK7A987S25szd9XnMHtTZW76nL43F7Y3pvqgn23PQb 6poYah6uOZqqnqEMKqiPJ7qpFMHoI/50XygynxG/lr/JvtLbFecdvrMbToup+v6uGpNLkKHd fb2ZoOvos7iJgQVy20MPnazNwfX143kEXHsTcnbam7cybXaSrWHxNbDyKoNVD8mICn7enraP xJo18q1/Z18wHamORIYdK86FH0/w95rWUICKadCaNK463Kv5Av8ALY21tnbEv8yT5O4+gxG2 Np0GZ3D0Rj91pHT4rEYzbFPWTbk75zkVV+3DR4aOinTAvMoERgmySL6cfUe4T92ec55Zl5I2 Fi1xIVWcpxYuRpt1p5tUeJTjUIfxr0WbhOdQtIcseNP5D/P+zrh8xeo/mH/Ms+TnQ3Z3aW0s l038T98967S6U+NvWm+q2uwHY259gZWWu3Lv7taDrpIP43ja/JbB2pX53IVmTTGmCgioqOne oWOOaQ05U3PlfkLYN5sNvuVuuZIbN57qWMBokkWixwmWukgSOsaqmurF2YLUgWi+ntYpUVg0 yqSxHAHgBX7TTHz6Hf8An2fI7rTY/c/xt6D3T1fQ9n7L2H1luPsSv6/TeW49iYaDMbwy9Ls7 YtVU12znp8lHJtXCbFyy09IkkIMeVV39ARXQeyew7hd7TzBvdtuTW13NcJEJfDSRtMamSQAS VHe0iVahyhAzWiK1gaSOSQGhrT/L0Rj4jfJ3rjI5PMz9JYDdPSvZGz8DnN7J1fX77ruw9jdi 7P2hja3cW8E2Rm8ti8fuza28Nq7VoKnJNjsjU5amy1DRVBhqaapjipasd8zbHfxxQpvU8V5t 0zrH4wjEUkUkhCJ4iqSjxu5Ca1CFGZaqykslJIiuGyOjnfKnv6DtDrjYHyY67qY07e6J3Ftz sjAVNLP9vUVNbtaup8o1BPIscxehycMD0tVG6SJLTTOjo6MysCdl2H93Xd/sF4v+6y7R4yDk AMCKj5g5HoQDx69EpUlCO0inW01sbeGG7C2Ts/f23ZXm2/vja2394YKaVDHLLhtzYmkzWMlk jPMbyUVahKn6E29423dtLZXVzZzik0UjI3+mUlT/ADHSQggkHiOlT7T9a697917r3v3Xuv/V 3+Pfuvde9+691737r3VJv/Chf5Rf7K1/Km+SGUx9Z9pu3u3G0Hxs2Zap+0klru31rMXu4086 sJ1q6Dqyh3BWQ+IF/LTL+ldTr7qyipHWo9uT5s/yvKr/AITw4P8Al5Y3tnOy/KzD0OI7njx7 dTdjxY5u/ans+Tfeews265dtHbb/AG2zcxWbXirlqFpWhjjkEhX1Nvq1DqrTHVvG4vlAflb/ AMJIt87wyWQSu3h1v0tsjoHfY8z1FXFnelu7Ot9l4epylRKzyT5bcOw6HDZeodiWeTIknk+/ dapR+q3P5oO8N84X/hNb/KG2nhZq2l2JvXcWE/v1LTKyU9dkNubW7IyezsLkJ1/VTTzy1tak DHTJPj0ktqhUjXWx8R63yvjX090x0V0F1T1J0Dg9v4bpraWxcBjtjU+3oKP+GZbAzYyCoh3H PU0kax5rJbqE5yFbXya5sjVVMlRK7ySMx91Q8eqff5hH813ePQ3y+6/+AvwI+KuG+UXz639s nG1ta+4a6i2j131RsGSPIbixdJvHP08uMy2TpKXFUc2VqKJsjhcbj6Sqp5xVy1E4pffutgVF Scda/X81aP8AmfUvzy/kt70/mKf7Jxg8/lPlVSS9NUHxCXs+DN7WGF7n+MVXu6g7DzO/0qJJ /sqvI4mfAnHZTILBJJkGd0Z01b6sKUanVhfyI/7i+fgf/wCKpbw/+B/+ZnvXl1ofAege/wCF M+e7ByP8w3+UR19ieuaTvDbw3lJuXanQm5914jYew+5Oy63tzrvFR7D3FuzdEkO0sfBuiGgx uKnnrmEdJSZGQFo1qWLe68vBs9Cn/MmxX87b+ZH8St6/FjfP8lLYuxkz+W2fuPaPYtD85/i7 ufM9dbl2huTG5eLNYDE1+7MZA0+VwUFbhanTUQMcdlKhVYEj3vrwoDXV0VD+cp132x1F8Zv+ E4nVfesEtH3F1xuPY+yOxsfUZrGbjqsXurbGG+N+GymIrNw4WsyWIztbiJaP7aaspaqrp6mW JpI551YSvr162v4utjH+cdjP5gWV6L2BT/AWo3jBnhv8v2eOs8nSYjsiTbgxcn8BGFraiopK tcEuZ1HJLRSrVMft7hqb7ke5R9qpOSI95vW51WIw+B+j4oLRaq92oUI1afg1Cnxfi09L9rNk Jn+tpTTivCvn+fp0HHywpfkFRfyB/kfSfKmoap+QEHwa7ZTsmWafG1Vd/FBtjPHHR5mqw4/h lVuGDAGkTJSwM6SV6zMHkvrYI82tsb8y7w3La02QzHwuIFKCukNkKW1FAeC0wOHSW6MBupTb f2GrH+r0rw6h/wDCa7/typ8L/wDy4v8A+Cx729hzphviPVbv86NlT+fJ/IzLsqg772ooLEKC z95YVEW5sNTuwAH5Jt7969bX4W6PH/wqDqIIf5O3fEc00UUlVv8A6Fp6VJJER6mde3tp1TQ0 6swaaVaamkkKrciONmtZSR7rS/EOtfr+Ydj6LtbEf8JWvjV3RXVuO+Mm/fjv8PqfftE0tXis blZd3Yf417I3xHkclFUUyQVNBs2SGljqifJh4spPMDaZg2+rD8Z6378Ds3aO1tq4bY22tr7e 2/srbmJx2B2/tHDYfH43bWEwmIhhp8ViMVg6OnhxtBjcbBTxpBBFGscSooUAAe9dN9aHXRnb fzEwX/ChL+aP2d8b/hjtv5x9+bVzO+9o4jCb2706+6OyXVvVWH3VtTYdDura9X2XXUcGV8u2 MVhsJ5KH96DH1hXUIatg++nDTSM9G/7h+M/84r5s/wAz3+XJ8w+xf5am0/iFR/Gft7ryLtzs bbfyt+O/ZmS3J03S9mbez+6KfNQYHelPuaugwmz2z1JTUdJQ109QuYljXQCLe60CACK9MmxO mth9vf8ACwj5N1m/cDi9xw9Q9R9e9wbVx2Zx9PlMfDvfBfFT4ybY25mjSVYemGQ23NvB8jQT lHekyFLBPFomjjkT3Xq0QdLP/hUTtPb2M+RX8n7tLGYumxvYR+R+X2xJu/Hr9lnZ8Hid7dHZ 7C4+oyNN46meLBZirqKih1MTSS1U7RaTNJq169eTg3Si/mnf9xJ/8nH/AMMXB/8AvxO5PfvL rw+A9beHv3VOtbD/AIVUUuQn/lhYmpow/wBrjPlD0/W5llqI4VTHSYPsLHQtLG8sbVSHMV9I ojQOwYh9OlCyyH7Ysq80R6vOF6fbVT/gB6WWJpMf9Kf8nWpt/KJ+D2T+f3ym2N1TPFX03WWB j/v13Tn6FZo2xPXuCqKf7vGU9dHYUeZ3jkJoMTROCXhkqmqArpTyAZJc280x8p8s3G5gqb5/ 04VPnIwwaeaoKs3qBpqKjo7mult4GkH9pwA+f+xx63y/5inz+6F/lrdDYXb8WA2zuDsKu25R YDonoOijpoMauP21DBjsJms5QUwvt3rfZ70EKK6KklTLTrS0dnWSWnxu5G5K3jnzeJZfGkSx Vy1xcGpNWyyqT8Ur1PyAOpsUDFNhYzbhMTqIjB7m/wA3qT/xfVIX8k7f/dvzz/mM9nfLb5D7 trd7ZbpzqbMptpHX7bb2xcz2Xk121gtvbLwSMaHbeBg2fHuCMRwjzTuTLUSTTyTSvLvuzZbR yZyJt/LWx2whju7ldfm0ixDWzSNxZtfh8cDgoAAAON4ihsrGO2gSmps+ppk1Pma06qk/nAdx S9t/zHPk9lxVmbHbM3unU2Kg9Xix8XVOKx+xctSwh7sBNufC19TJyQZqhytlKgSX7X7cu1ch 8vRaaPLD4zfPxmMik/7RlA+QHTlpAEsYKjJWv7c/4D0g/wCXjg63N9+Zffk0lTR7M6F6V717 t7GzcAUjDbX2h1Xumlols5VZ6jcG7svjMVBACJJ565UTk+1XPN6kWyx2S0N1eXdvBEv8TvMh P5Kis5PABa9F94gC0pkkAft6RfT3yWr8ThsntHK1ryUFfQ1FHJHK90eKaFo7EMSPz7d3XbIp 5EnRRrU16ThaN19BX+XvUZKq+B/wznyyRpVv8XeidJjlM/noU6z20mKrJJWklZqiuxawzS3N xLIwIFrDCnnIRrzdzOIz2/vC4/b4rVH2A1A+XRbL/aSf6Y/4ejgew11Tr3v3Xuve/de6/9bf 49+691737r3XvfuvdUE/zqv5WnyO/mj75+FuzNo726d2p8Xum+zKrfHyCxG89z77xfYW76bM 5famKrl2JiNudd7lwNZksD15RZqKgNblsWs1dlikjxRr5h7qykCvr1YTUfyuP5Z1TBNTyfy8 fg6sdRFJC7U/xR6JpZ1SVGRmhqqXYkNTTShW9MkbpIhsVIIB9+61U+vVAnQP8hf5tdHfy+f5 n/wAh7a+OWb2L8p957I3b8XcjNvbtJF2nVYDfONm3NW9vUx6ZvhclufrvZu3YycN/G1TIUDx EtCUnHurahUHqzrDfyhNtdsfye+ov5ZPyuy+Cqty9fddYTGRdkdW1NfmaHZPam2MhmK/be/d i1O6cFtjI5OCgGVemqYKuioWrsdVVdKTEJvKvutau6o6rD6V/l+/8KWviXsCH4s9CfOP4k7j 6FwEX93OsexOx6LL5rfXW2zfJUQ0lPh6LcnTG68pQNi6Nl+1xNTV7ix+NRYqajmSBNK+63VT mnRh/mZ/KF+dJ+ZHTf8AMn/l/fJDqzb/AMuNsdO7W6y7uwvfFHmX2F2rV7a2RSbHyedx1Ti9 q7sZo96YSEQVeOrKWj8E9NBW01fT1HMfutAilDw6LT3V/LG/ni/Pv5N/CDtr5q7t+AG2Nn/D PuPH9mYmPqXM9sY7L5bE5LffVW597Y6LH1OzNzRZfN1dF1ZRJQLPVY2libya5fWLb62CoBAr 0cL+an/Kh+Y/dHzf+Pn8yv8Al090dYdZ/JrpTYp68yWA7fXJptzMYiI72oYsthauHau+cdUV eY2x2TlcPlMdWUVLTSUIjlhqEqdWvXWgRQg8OuPfP8o/5YfzMPg3gdjfzIO5eptsfOnq3tzd HYvx77v+PWGr6vZ/W+28thtpUa7EzuKlwnX+RyFFuPI7feoyMtJK1VS1dNjqqCsmWmeik916 oBwMdBrtHpj/AIVR7H2jTdRU/wAlf5em86bGQRYPFd974p965jsFMZBTpBR5fJE9SRUOay0E Cr5ZsjgchVVFQGaaSoJMre69VfTox383P+Vx8mP5gc/8uqv677G6fgzXxP7OG/O4c32hlN07 Wn3oZn6tmyNds3HbE633VjXymRq9lV0zU06YmkiM8SowQsI/deUgV6v19+6r0Ur559Dbv+Uf wv8AlB8c+v8AJbbw+9u6+k9/9bbVym8azKY/auPzm6sBV4rHVe4K7C4fcGXpcTDUVAaaSmoa uZUBKxOeD7rYwQetYr41/wArX/hTf8QelNl/Hb47fzDfgx150515/eP+5+z/AOBU+7f4R/e3 dme3zuD/AH8G+fgbubdOQ/iG6dzV1V/lVdP4vP449EKRxpvqxKnJHRs/lT/Jn+d/zm+Inxcz fyO+W/WsP8074kdh7+3z198g9i4mXEdTZ+iz+9MZuHau3si2z+q+rsntnIbYh2lgamjzNHta pmoa2inX7ar+7eoTXWgwBNBjotPbP8pf+ez/ADIIesej/wCZN8yPjlgvi3sfdGF3XvKm6SxX m332BlcLFXUcOUfE4rqzYeHyGdfHVtRDC9bX0mJx8tStYuNqp4hH731uqitBnq47+Zn/ACee jP5hvxJ66+OFNkpOo9zfHrG46m+NPY9DQy5+TrqPEbdxm2BtfK4+avoqrN7M3DhcFQQZCFam Kp8tDS1KSGSnCvrrQahr1Xn0H8Xf+FN+xsl1n1F2D81viPmuk9mby2iu5e0quhrN2dubh6s2 1mMVNk9rw19f0nistmcluHb9DJSNVZCWny7NO7PlkfTOPdeqvp0LP8xD+TJ33v35h4f+ZT/L L+ReA+MPzJhxdPiewMTvahqper+14aHD0e3oMhlp8dgN2pSVtftzHw0GToK/B5fGZb7elqLU VVTvUVHuvBhSh4dDX8Qdrfz/AGr+RPVmW+bXZPwYxfxk2rBug9o7X6Pxu55+yOzJ6vr7deI2 glJVZ/ZNXQ4v+F9g1eKytc1JX4QPBSNHGssbvTv7rx00NK16gdVfyxO+9jfz3Pkt/M+y27uo ajoPuXpPA9bbY2jjs/vObt+gzmL6v6A2VUVeewFTsGk2ZS4l8r1VkZEkp9wVUxp5qdjEHeWO H3Xq9oHXv5z38sTvv+Yru/4H5/pPd3UO16P4vd2Zjsnf8Xaef3nhKnL4PIZPrWthpNnptLYO 94shlli2bVBo618fCGkitKQzlPdeUgV6S/8AOV/lS/Ir5n9pfFf5e/CfuTZPUXyy+JGSrKna Z7JhrY9pblx0edx27MC8WWoNt7xWhzO383RVSrR1+KrMXlock8VS9NHGzS+68rUqDw6P7/Lx wf8AMbwvWG9G/mVby6I3b2tkN9PUbFj6Cpaym27g+v4sDiKVKDNms2rtlps7U7ihram6NWKK aWIeRCDEnutGnl0if5x3xWy/zJ/lt/KTpPamLkzPYEmxU7C6zxtKjvkcn2B1Tlsf2Lt3AYlU 4OT3fLtx8NFrshORIYqCWB/yxuQ2nfduvXakQfSx/osNJJ+QrX8unrd/DmRjwrQ/njr5938p L+ax25/Li3tunP7D23tXfeyuycficZ2X15uaCPGVGck2yMs22Mhid5UlDU7g23lNv1WaqwqA VNBPDVTLNSvL4J6fKDeOVdr57261tr25khuYSTHIudOqmoFCQrBgB6MCBRqVBNp4EmAVyRTg epfyf+UvY/zL7y3z8gu08iKjce9MgXocNTVFTLhdm7ZpC0W3tmbahqZJHpMFt+gtHGP1zzGS omL1E8sjypyzy/t3LWy2Wz7alLeJcsQNTufid6cWY5+QooooAAhsFiigSKLgP5n1+3rbK/4T C7GgoPjt8l+zFhAqN2907e2JLUeKINLD13sai3BTwmcSGaQQSdoysEZVVPISpJZguN/3hLwy b7sG317YrRpKfOWQqcfPwR+zoj5ik1T28fohP7TT/J1qpUPXPdvzD+TW9ttdO7A3N2R2V2T2 PvPcZweExlTFNTNnd1VtdkMtn58oKOm2rhaGpyINXWZSWkpqIN+/JHz7yBl3LauWdgtJtyvI 4LGCCNdTEZ0oAAoFS7GmFQEt5A9G8zxW8Cl2ARVA/l/P8ujbfO/pXOfyu6Kj+MO0PkRT7y7C +Q/RO3D8yNibdpo/4LtTJ0G/6PfGysFicytPT5CLH1lDQRIIqhkrqqkSeqnhiostS08QM5Z3 yHntm3652bwrOyu2+ikY9zgxlHZlrSoJOR2g0UEtGzElSQXVZTHRVbtPr5f6v9jqrPpfYW++ /O6usekOsaWXIb67S3phNl7fhjjlligqczWR00+Ur/ArvFiMJRmWsrZraYKSnkkYhVJAq3Xd 7barC83G8eltDGXb5gCtB8ycAeZIHVJKKC7cB19YzYuz8R17snZ2wdvpJHgdj7V29s/CRylD KmI21iKTC41JDGkcZkWiokB0qq3+gA494GXdzJeXVzeTH9aWRnb7WJY/zPRISSSTxPSq9p+t de9+691737r3X//X3+Pfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfu vde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9 +691737r3Xvfuvde9+691737r3XvfuvdfOm/4UIfyft3/Dbundnzd+OW1Mrmvir3HuOv3V2j g9v42srY/j32buHJGqzk2Rgo4pYsX1TvjM1xqMZVEpS4vIVEmNZYIv4cJ579u+dRJHDtd7LS 8jACEn+0UcP9so4jiQK5zQ3tZxKojc/qDh8x/n6oS2b2DDPBETOCpUcFhyDb+p495FbZvKSI oZujCKZomBB620P5Dn82v4p/EHp/tvor5N7wyfW2OznZUnbOzt6Q7P3ZvPDV9VmtqbW2jntu 5Ck2Jg9ybix9fTR7MoqimkajelmSWYNLE6Ks0P8Au9yPvfNW5bbvOwwLO6W/gumtEYBXd1YF 2VSDrYEVqKDBBwxuUEl68c0IBIWhFQPMmuft6HH5Sf8ACgT4kfHvae79hfyu+kNoUW8N6Vdf kc13AeqsR1dsGiy+SlqHn3PSbIXFYXdG/t0o8rvEc3SY+kp3ZGdayMPTew/s3tVv+7T213z3 u0jW8QAWHxTLIQKdperLGnroLE5+E0bpmOwmlKteSkgeVan9vkPs609Oze2d3dk7w3Lvzfm5 szvPfO88zW53cu5c7Wz5LNZ3NZOczVVbW1czPLPUTyvwBwosqgKABOSfS7fbRWdpEsdrGoVV UUVVHAAenS5gqgKooo63Wf8AhO7/ACldydE4+n+d3yU23X4Dt7eu3K7G9G9dZ2lWnynXWw9y 0scWS37uKhqYfvsRvfe2Id6SkpWMUuPwlTOtQpmrWipcb/dHnlN2c8vbXKGsUcGVwcO68EB4 FEOSfNwKYWpJbycOfDQ9o49bW/uGOkPXvfuvde9+691737r3X//Q3+Pfuvde9+691737r3Xv fuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvd e9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3XvfuvdNuaw uH3Hh8tt7cOJxuewGextdhc5g81Q0uUw+Zw+UpZaHJ4nLYyuinosjjcjRTvDPBMjxTROyOpU ke7KzIyujEODUEYII4EH168CQajj1qe/Or/hKZ8fu289kex/g12Ufipuivmq6/JdT7jxeV3x 0jla+od5rbckTKxbv6vhkqJWZ4qc5rGwxhYaSgpYwLSVsPuXue2hYtxjNxEODA6XA+eKN+dD 5knpfFfMBSVdQ9fP/Z6oq3r/AMJt/wCcHsKsrqPb/W3UXb9PRPVpS5TrvvDZuLo8rHTLC0El HB2xL1lkIf4h5SsQqYYCrRt5PGNDPJln7ubEyL480sZ9GRjT/eNXSpb6EGoYj8v81ehC6l/4 TRfzU+w8nSU3YeJ6R6Iw7zIMjlN6dq0G7a+npTFFLNLQ4jquj3zDkKtS5jSKWrpY3lQ3lRCs h3ee8OxRI3gCaZ6YCppH5lyv+A9WfcYaYqT9n+frZs/l2/8ACev4mfCXO4TtfsrI1Pyk7+wd TT5HB7v3xgKXC9fbGylOUmp8lsbrEV+cpI85RVKh4cnlq3J1NPLHHNRrRSAkxVzL7kb3v6SW sP8AitgwoVQksw9GfBofNVCg8DqHRfNeSSjSO1P9Xn1f37jvpJ1737r3Xvfuvde9+691737r 3X//0d/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+ 691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917 37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3 Xvfuvdf/0t/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+ 691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917 37r3Xvfuvdf/09/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3X vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+ 691737r3Xvfuvdf/2Q== --------------020402090103080608090109-- --------------000605040804040307080105-- From Thierry.Bessis@alcatel-lucent.com Tue Mar 8 10:16:04 2011 Return-Path: X-Original-To: dispatch@core3.amsl.com Delivered-To: dispatch@core3.amsl.com Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 91D0B3A6957 for ; Tue, 8 Mar 2011 10:16:04 -0800 (PST) X-Virus-Scanned: amavisd-new at amsl.com X-Spam-Flag: NO X-Spam-Score: -6.597 X-Spam-Level: X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4] Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RVnf3tuJL3r7 for ; Tue, 8 Mar 2011 10:16:02 -0800 (PST) Received: from ihemail4.lucent.com (ihemail4.lucent.com [135.245.0.39]) by core3.amsl.com (Postfix) with ESMTP id 10D743A6936 for ; Tue, 8 Mar 2011 10:16:01 -0800 (PST) Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail4.lucent.com (8.13.8/IER-o) with ESMTP id p28IHBRG003368 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 12:17:12 -0600 (CST) Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28IHAWq000194; Tue, 8 Mar 2011 12:17:10 -0600 (CST) Message-ID: <4D7672A6.6040409@alcatel-lucent.com> Date: Tue, 08 Mar 2011 12:17:10 -0600 From: Thierry Bessis User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9 MIME-Version: 1.0 To: "Parthasarathi R (partr)" References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> In-Reply-To: Content-Type: multipart/alternative; boundary="------------030405020704000403040309" X-Scanned-By: MIMEDefang 2.57 on 135.245.2.39 Cc: dispatch@ietf.org Subject: Re: [dispatch] SIP load balancing X-BeenThere: dispatch@ietf.org X-Mailman-Version: 2.1.9 Precedence: list List-Id: DISPATCH Working Group Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 08 Mar 2011 18:16:04 -0000 This is a multi-part message in MIME format. --------------030405020704000403040309 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Part, On 08-Mar-2011 03:52, Parthasarathi R (partr) wrote: > Vijay, > > As Paul K& Paul J mentioned, I'm looking at your proposal with lot of > similarity to SCE. One more key stuff, I could not understand the need > for two separate mechanism for > > 1) Load balancing ("weight") > 2) Overload control ("oc") See my previous Email: OC is a request to shed: Whatever you did before is not enough: you need to shed now. Load Balancing is a mean to constantly balance the load, which is an optimized way to prevent localized overload (but won't prevent the cluster overload of course !). However, the same transport mechanism can certainly be used and both mechanisms should work very well together, > In case feedback mechanism to the upstream is generic enough, the load > balancer will be able to balance based on value (capacity) received from > downstream and reduce the load drastically or stop in case the value is > too low in terms of capacity. You may decide shed from the load feedback value, but this complicated to setup: we did that before inside our IMS core product, but we decided that having a load indication separated from the shedding request is much easier and more efficient. The main issue is to agree about the threshold: which load threshold do you use to decide to shed ? and how much to shed based on the load ? Some servers will want you to shed before others for the same indicated load and/or more (or less) than others. The rejection rate allows each server to request exactly the source shedding that it thinks is needed: this is actually very powerful. > Even though any overload control algorithm > is used ("oc" or "weight" or "SCE" or "resource-availability" other), > the locally overload control mechanism is required in downstream as per > mentioned in draft-ietf-soc-overload-design. Right. A difficulty is when a downstream server will have both oc compliant clients, and non oc-compliant. Obviously the local OLC needs to take into account that, otherwise "non compliant" will get a huge and unfair advantage (one stage OLC instead of two). In our implementation, until we reach a critical level, we do NOT apply the local overload to olc compliant nodes (we trust them). we only apply it to non compliant. > I wish that we will look > into these aspect because overload control is the extreme case for load > balancing the cluster of downstream entity and load balancing mechanism > shall help to avoid the overload control based on feedback. > > I guess that your mechanism assumes about downstream SIP UA as signaling > entity only and does not handle media. Yes this is correct. Although we plan to use the same thing in our MGCF which has media too (i.e. the resource feedback will ALSO take into account the available media count). Cordially, Thierry > Please correct me in case I > wrongly understand. draft-partha-dispatch-resource-availability-00 was > written to handle both resource monitoring and load balancing of SIP > based media servers wherein load balancing has to take care of overload > as well. > > I'm also interested in discussing SIP load balancing topic during > IETF-80 meeting. > > Thanks > Partha > > -----Original Message----- > From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On > Behalf Of Paul E. Jones > Sent: Tuesday, March 08, 2011 10:35 AM > To: 'Vijay K. Gurbani'; dispatch@ietf.org > Subject: Re: [dispatch] SIP load balancing > > Vijay, > > The Session Capacity Estimate (SCE) idea is actually quite similar to > the "adaptive load balancing" draft, from what I can see. SCE was > proposed as a solution to the "overload" problem, the approach being one > where we try to avoid overload by selecting a device that reports a > higher capacity value. > Thus, one could view it as "overload control through load balancing". > We didn't explicitly call it load balancing, though, because we were not > trying to balance load (i.e., place the same load on all servers). > Rather, we just wanted to direct traffic to devices that we knew had > available capacity. > > In practice, the way SCE would be employed is the following: > * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from > peer devices > * The selector would direct traffic toward the peer devices in a > round-robin fashion, taking into consideration the SCE value so > as to prevent exhausting resources on any given peer device > * Once the peer's SCE value dropped below a given "low-water mark", > the peer device would be less favored and not given traffic until > the SCE value increases above the low-water mark (OPTIONS "ping" > messages may be used to determine the SCE value in the event that > there is no other traffic) > * The current SOC overload control mechanism would be used to > to prevent message overloading, as SCE only focuses on resource > overloading (e.g., memory, CPU, disk space, DSPs) > > Note that the "round-robin" approach may or may not be preferred and is > not required in our draft. I've found this approach works pretty well, > but one may prefer to use a particular weighting that is more complex. > I don't think local decisions in the box need to be subject to > standardization: we just need to standardize the signaling to enable > such decisions. > > In any case, it certainly looks like we're thinking along the same > lines. > > Paul > >> -----Original Message----- >> From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On >> Behalf Of Vijay K. Gurbani >> Sent: Monday, March 07, 2011 6:38 PM >> To: dispatch@ietf.org >> Subject: Re: [dispatch] SIP load balancing >> >> Paul: Thank you for your input. More inline. >> >> On 03/07/2011 05:14 PM, Paul Kyzivat wrote: >>> Vijay, >>> IIUC, the algorithm you have proposed is similar to the one Paul >>> Jones has proposed - with the main difference being the numeric >>> range of the weight. Do I have that right? >> I believe that Paul J.'s algorithm is more suited towards overload >> control, not load balancing (and I will await Paul J.'s correction to >> my assertion). >> >> The intent in draft-bessis-dispatch-adaptive-load-balancing is to >> *always* report the current load to give the upstream SIP proxy an >> indication of the health of the SIP server in the cluster. If one >> waits until overload is imminent, then it may be too late since a >> sudden and unanticipated burst of messages will put the SIP server in >> overload rather quickly. >> >>> It isn't clear to me from your draft if you are specifying the >>> algorithm by which the relative weights are used to determine the >>> probability of a request going to one server or another. >> Yes, same way as RFC2782 does it. >> >>> ISTM that it *does* matter. And it especially matters if there is an >>> implication that the upstream node's algorithm may depend on some >>> assumption or knowledge about the downstream nodes. >> No knowledge, just a reported number that the upstream SIP proxy can >> treat as a RFC2782-like weight indicator. Today, rfc3263-resolution >> starts off with the SRV indicators in RFC2782, without any regard to >> additional knowledge or assumption about downstream nodes. >> dispatch-adaptive-load-balancing is trying to use the same model, >> except render the weight to be sensitive to the current traffic in the >> cluster in real-time. >> >>> And of course this method of reporting weights isn't sufficient to >>> drive a balancing algorithm that is sensitive to some requests >>> requiring more or different resources than others. (That seems to be >>> a real problem, though I don't know if we have a real solution for >>> it.) >> Yes, that is a real problem, as is the problem of a SIP server in a >> cluster that has a low weight and is thus not contacted with enough >> frequency to merit more participation to the goodput. >> >> It is well known in SIP that some methods are more onerous to process >> than others --- from a proxying point of view, processing an INVITE >> with forking and response aggregation is far more detrimental to >> resource usage than is processing a BYE. Some existing literature >> techniques thus add a coefficient to the weight of sending an INVITE > versus a BYE. >> At this time, the process outlined in dispatch-adaptive-load-balancing >> treats all requests the same, but we can definitely look at ways to >> change this behaviour. >> >> Thanks, >> >> - vijay >> -- >> Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane, >> Rm. 9C-533, Naperville, Illinois 60566 (USA) >> Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com >> Web: http://ect.bell-labs.com/who/vkg/ >> _______________________________________________ >> dispatch mailing list >> dispatch@ietf.org >> https://www.ietf.org/mailman/listinfo/dispatch > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch > _______________________________________________ > dispatch mailing list > dispatch@ietf.org > https://www.ietf.org/mailman/listinfo/dispatch -- -- Cordially, Thierry Bessis IMS Solution: Architecture & Product Management - ALTA Member Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM Alcatel-Lucent 1960 Lucent Lane - Room 6L-517 Naperville, IL 60563 U S Tel/Fax: +1 630 979 7989 Corporate IM: tbessis - Yahoo IM: thierry_bessis My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline Conference information: 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989 others countries see: http://all.alcatel-lucent.com/audio Upcoming planned Business Trip: 12 April - 14 April 2011 Upcoming planned Vacation: None --------------030405020704000403040309 Content-Type: multipart/related; boundary="------------000507050006020109080407" --------------000507050006020109080407 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Part,

On 08-Mar-2011 03:52, Parthasarathi R (partr) wrote: 
Vijay,

As Paul K & Paul J mentioned, I'm looking at your proposal with lot of
similarity to SCE. One more key stuff, I could not understand the need
for two separate mechanism for

1) Load balancing ("weight")
2) Overload control ("oc")
See my previous Email:
OC is a request to shed: Whatever you did before is not enough: you need to shed now.
Load Balancing is a mean to constantly balance the load, which is an optimized way to prevent localized overload (but won't prevent the cluster overload of course !).

However, the same transport mechanism can certainly be used and both mechanisms should work very well together,
In case feedback mechanism to the upstream is generic enough, the load
balancer will be able to balance based on value (capacity) received from
downstream and reduce the load drastically or stop in case the value is
too low in terms of capacity.
You may decide shed from the load feedback value, but this complicated to setup: we did that before inside our IMS core product, but we decided that having a load indication separated from the shedding request is much easier and more efficient.
The main issue is to agree about the threshold: which load threshold do you use to decide to shed ? and how much to shed based on the load ? Some servers will want you to shed before others for the same indicated load and/or more (or less) than others.
The rejection rate allows each server to request exactly the source shedding that it thinks is needed: this is actually very powerful. 
Even though any overload control algorithm
is used ("oc" or "weight" or "SCE" or "resource-availability" other),
the locally overload control mechanism is required in downstream as per
mentioned in draft-ietf-soc-overload-design.
Right. A difficulty is when a downstream server will have both oc compliant clients, and non oc-compliant. Obviously the local OLC needs to take into account that, otherwise "non compliant" will get a huge and unfair advantage (one stage OLC instead of two).
In our implementation, until we reach a critical level, we do NOT apply the local overload to olc compliant nodes (we trust them). we only apply it to non compliant. 
I wish that we will look
into these aspect because overload control is the extreme case for load
balancing the cluster of downstream entity and load balancing mechanism
shall help to avoid the overload control based on feedback.

I guess that your mechanism assumes about downstream SIP UA as signaling
entity only and does not handle media.
Yes this is correct. Although we plan to use the same thing in our MGCF which has media too (i.e. the resource feedback will ALSO take into account the available media count).

Cordially,

Thierry
Please correct me in case I
wrongly understand. draft-partha-dispatch-resource-availability-00 was
written to handle both resource monitoring and load balancing of SIP
based media servers wherein load balancing has to take care of overload
as well.

I'm also interested in discussing SIP load balancing topic during
IETF-80 meeting.

Thanks
Partha

-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Paul E. Jones
Sent: Tuesday, March 08, 2011 10:35 AM
To: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

Vijay,

The Session Capacity Estimate (SCE) idea is actually quite similar to
the "adaptive load balancing" draft, from what I can see.  SCE was
proposed as a solution to the "overload" problem, the approach being one
where we try to avoid overload by selecting a device that reports a
higher capacity value.
Thus, one could view it as "overload control through load balancing".
We didn't explicitly call it load balancing, though, because we were not
trying to balance load (i.e., place the same load on all servers).
Rather, we just wanted to direct traffic to devices that we knew had
available capacity.

In practice, the way SCE would be employed is the following:
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
   peer devices
 * The selector would direct traffic toward the peer devices in a
   round-robin fashion, taking into consideration the SCE value so
   as to prevent exhausting resources on any given peer device
 * Once the peer's SCE value dropped below a given "low-water mark",
   the peer device would be less favored and not given traffic until
   the SCE value increases above the low-water mark (OPTIONS "ping"
   messages may be used to determine the SCE value in the event that
   there is no other traffic)
 * The current SOC overload control mechanism would be used to
   to prevent message overloading, as SCE only focuses on resource
   overloading (e.g., memory, CPU, disk space, DSPs)

Note that the "round-robin" approach may or may not be preferred and is
not required in our draft.  I've found this approach works pretty well,
but one may prefer to use a particular weighting that is more complex.
I don't think local decisions in the box need to be subject to
standardization: we just need to standardize the signaling to enable
such decisions.

In any case, it certainly looks like we're thinking along the same
lines.

Paul


-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On 
Behalf Of Vijay K. Gurbani
Sent: Monday, March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

Paul: Thank you for your input.  More inline.

On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul 
Jones has proposed - with the main difference being the numeric 
range of the weight. Do I have that right?

I believe that Paul J.'s algorithm is more suited towards overload 
control, not load balancing (and I will await Paul J.'s correction to 
my assertion).

The intent in draft-bessis-dispatch-adaptive-load-balancing is to
*always* report the current load to give the upstream SIP proxy an 
indication of the health of the SIP server in the cluster.  If one 
waits until overload is imminent, then it may be too late since a 
sudden and unanticipated burst of messages will put the SIP server in 
overload rather quickly.


It isn't clear to me from your draft if you are specifying the 
algorithm by which the relative weights are used to determine the 
probability of a request going to one server or another.

Yes, same way as RFC2782 does it.


ISTM that it *does* matter. And it especially matters if there is an


      

        

          
implication that the upstream node's algorithm may depend on some 
assumption or knowledge about the downstream nodes.


        

        
No knowledge, just a reported number that the upstream SIP proxy can 
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution 
starts off with the SRV indicators in RFC2782, without any regard to 
additional knowledge or assumption about downstream nodes.
dispatch-adaptive-load-balancing is trying to use the same model, 
except render the weight to be sensitive to the current traffic in the


      

      

      

        
cluster in real-time.



        

          
And of course this method of reporting weights isn't sufficient to 
drive a balancing algorithm that is sensitive to some requests 
requiring more or different resources than others. (That seems to be


        

      

      

      

        

          
a real problem, though I don't know if we have a real solution for
it.)


        

        
Yes, that is a real problem, as is the problem of a SIP server in a 
cluster that has a low weight and is thus not contacted with enough 
frequency to merit more participation to the goodput.

It is well known in SIP that some methods are more onerous to process 
than others --- from a proxying point of view, processing an INVITE 
with forking and response aggregation is far more detrimental to 
resource usage than is processing a BYE.  Some existing literature 
techniques thus add a coefficient to the weight of sending an INVITE


      

      
versus a BYE.


      

        
At this time, the process outlined in dispatch-adaptive-load-balancing


      

      

      

        
treats all requests the same, but we can definitely look at ways to 
change this behaviour.

Thanks,

- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane, 
Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


      

      
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------000507050006020109080407
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------000507050006020109080407--

--------------030405020704000403040309--

From Thierry.Bessis@alcatel-lucent.com  Tue Mar  8 10:36:55 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2ADD23A68AF for ; Tue,  8 Mar 2011 10:36:55 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pN7G6hgp5HIo for ; Tue,  8 Mar 2011 10:36:53 -0800 (PST)
Received: from ihemail4.lucent.com (ihemail4.lucent.com [135.245.0.39]) by core3.amsl.com (Postfix) with ESMTP id E018D3A68A7 for ; Tue,  8 Mar 2011 10:36:52 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail4.lucent.com (8.13.8/IER-o) with ESMTP id p28IbN0U021956 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 12:37:23 -0600 (CST)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28IbNxi016679; Tue, 8 Mar 2011 12:37:23 -0600 (CST)
Message-ID: <4D767762.1090508@alcatel-lucent.com>
Date: Tue, 08 Mar 2011 12:37:22 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com>
In-Reply-To: <4D7633F9.7000608@cisco.com>
Content-Type: multipart/alternative; boundary="------------010306000509060100090004"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.39
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 18:36:55 -0000

This is a multi-part message in MIME format.
--------------010306000509060100090004
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Paul,

On 08-Mar-2011 07:49, Paul Kyzivat wrote:
> Vijay,
>
> Thanks for clarifications. Now I think I understand what you are 
> proposing. It is certainly simple, and better than nothing.
I think it is much better than nothing  ! :-)
It is true that some transactions use much more resources than others, 
but this does not damage the load balancing:
If one server is unlucky enough to get much more "loaded transactions" 
than others, it will quickly report a higher resource usage and will 
consequently receive less requests than others: despite the fact that 
one server get more of such transactions, the cluster will still be 
evenly loaded.

>
> It *does* require that all the members of the load balancing set 
> assign their own weights in a consistent way. This of course is no 
> different than when the weights are assigned in DNS, except that its 
> likely to be one person who puts all the weights into DNS, while the 
> weights coming from the servers can be influenced by who implemented 
> them. If they aren't all implemented consistently it can be a problem.
This is true, but the mechanism is still robust. In practice, the most 
important is that all use the same range. But the system is robust to 
things like linearity issues: as long as the feedback reported load 
curve is a monotonic function of the actual load, it works.

Cordially,

Thierry
>
> But this is a second order problem, and is probably unavoidable. 
> (AFAIK every scheme for this space has a similar constraint.) It might 
> need to be the subject of additional standardization/profiling, though 
> it might not be ietf work.
>
>     Thanks,
>     Paul
>
> On 3/7/2011 6:37 PM, Vijay K. Gurbani wrote:
>> Paul: Thank you for your input. More inline.
>>
>> On 03/07/2011 05:14 PM, Paul Kyzivat wrote:
>>> Vijay,
>>> IIUC, the algorithm you have proposed is similar to the one Paul
>>> Jones has proposed - with the main difference being the numeric range
>>> of the weight. Do I have that right?
>>
>> I believe that Paul J.'s algorithm is more suited towards overload
>> control, not load balancing (and I will await Paul J.'s correction to
>> my assertion).
>>
>> The intent in draft-bessis-dispatch-adaptive-load-balancing is to
>> *always* report the current load to give the upstream SIP proxy
>> an indication of the health of the SIP server in the cluster. If one
>> waits until overload is imminent, then it may be too late since a sudden
>> and unanticipated burst of messages will put the SIP server in
>> overload rather quickly.
>>
>>> It isn't clear to me from your draft if you are specifying the
>>> algorithm by which the relative weights are used to determine the
>>> probability of a request going to one server or another.
>>
>> Yes, same way as RFC2782 does it.
>>
>>> ISTM that it *does* matter. And it especially matters if there is an
>>> implication that the upstream node's algorithm may depend on some
>>> assumption or knowledge about the downstream nodes.
>>
>> No knowledge, just a reported number that the upstream SIP proxy can
>> treat as a RFC2782-like weight indicator. Today, rfc3263-resolution
>> starts off with the SRV indicators in RFC2782, without any regard
>> to additional knowledge or assumption about downstream nodes.
>> dispatch-adaptive-load-balancing is trying to use the same model,
>> except render the weight to be sensitive to the current traffic
>> in the cluster in real-time.
>>
>>> And of course this method of reporting weights isn't sufficient to
>>> drive a balancing algorithm that is sensitive to some requests
>>> requiring more or different resources than others. (That seems to be
>>> a real problem, though I don't know if we have a real solution for
>>> it.)
>>
>> Yes, that is a real problem, as is the problem of a SIP server in a
>> cluster that has a low weight and is thus not contacted with enough
>> frequency to merit more participation to the goodput.
>>
>> It is well known in SIP that some methods are more onerous to process
>> than others --- from a proxying point of view, processing an INVITE
>> with forking and response aggregation is far more detrimental to
>> resource usage than is processing a BYE. Some existing literature
>> techniques thus add a coefficient to the weight of sending an INVITE
>> versus a BYE. At this time, the process outlined in
>> dispatch-adaptive-load-balancing treats all requests the same, but we
>> can definitely
>> look at ways to change this behaviour.
>>
>> Thanks,
>>
>> - vijay
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------010306000509060100090004
Content-Type: multipart/related;
 boundary="------------030603080100030709070503"


--------------030603080100030709070503
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Paul, 

    

    On 08-Mar-2011 07:49, Paul Kyzivat wrote:
    
Vijay,
      

      

      Thanks for clarifications. Now I think I understand what you are
      proposing. It is certainly simple, and better than nothing.
      

    

    I think it is much better than nothing  !  :-) 

    It is true that some transactions use much more resources than
    others, but this does not damage the load balancing: 

    If one server is unlucky enough to get much more "loaded
    transactions" than others, it will quickly report a higher resource
    usage and will consequently receive less requests than others:
    despite the fact that one server get more of such transactions, the
    cluster will still be evenly loaded. 

    

    

      

      It *does* require that all the members of the load balancing set
      assign their own weights in a consistent way. This of course is no
      different than when the weights are assigned in DNS, except that
      its likely to be one person who puts all the weights into DNS,
      while the weights coming from the servers can be influenced by who
      implemented them. If they aren't all implemented consistently it
      can be a problem.
      

    

    This is true, but the mechanism is still robust. In practice, the
    most important is that all use the same range. But the system is
    robust to things like linearity issues: as long as the feedback
    reported load curve is a monotonic function of the actual load, it
    works. 

    

    Cordially, 

    

    Thierry

    

      

      But this is a second order problem, and is probably unavoidable.
      (AFAIK every scheme for this space has a similar constraint.) It
      might need to be the subject of additional
      standardization/profiling, though it might not be ietf work.
      

      

          Thanks,
      

          Paul
      

      

      On 3/7/2011 6:37 PM, Vijay K. Gurbani wrote:
      

      
Paul: Thank you for your input. More
        inline.
        

        

        On 03/07/2011 05:14 PM, Paul Kyzivat wrote:
        

        
Vijay,
          

          IIUC, the algorithm you have proposed is similar to the one
          Paul
          

          Jones has proposed - with the main difference being the
          numeric range
          

          of the weight. Do I have that right?
          

        

        

        I believe that Paul J.'s algorithm is more suited towards
        overload
        

        control, not load balancing (and I will await Paul J.'s
        correction to
        

        my assertion).
        

        

        The intent in draft-bessis-dispatch-adaptive-load-balancing is
        to
        

        *always* report the current load to give the upstream SIP proxy
        

        an indication of the health of the SIP server in the cluster. If
        one
        

        waits until overload is imminent, then it may be too late since
        a sudden
        

        and unanticipated burst of messages will put the SIP server in
        

        overload rather quickly.
        

        

        
It isn't clear to me from your draft if
          you are specifying the
          

          algorithm by which the relative weights are used to determine
          the
          

          probability of a request going to one server or another.
          

        

        

        Yes, same way as RFC2782 does it.
        

        

        
ISTM that it *does* matter. And it
          especially matters if there is an
          

          implication that the upstream node's algorithm may depend on
          some
          

          assumption or knowledge about the downstream nodes.
          

        

        

        No knowledge, just a reported number that the upstream SIP proxy
        can
        

        treat as a RFC2782-like weight indicator. Today,
        rfc3263-resolution
        

        starts off with the SRV indicators in RFC2782, without any
        regard
        

        to additional knowledge or assumption about downstream nodes.
        

        dispatch-adaptive-load-balancing is trying to use the same
        model,
        

        except render the weight to be sensitive to the current traffic
        

        in the cluster in real-time.
        

        

        
And of course this method of reporting
          weights isn't sufficient to
          

          drive a balancing algorithm that is sensitive to some requests
          

          requiring more or different resources than others. (That seems
          to be
          

          a real problem, though I don't know if we have a real solution
          for
          

          it.)
          

        

        

        Yes, that is a real problem, as is the problem of a SIP server
        in a
        

        cluster that has a low weight and is thus not contacted with
        enough
        

        frequency to merit more participation to the goodput.
        

        

        It is well known in SIP that some methods are more onerous to
        process
        

        than others --- from a proxying point of view, processing an
        INVITE
        

        with forking and response aggregation is far more detrimental to
        

        resource usage than is processing a BYE. Some existing
        literature
        

        techniques thus add a coefficient to the weight of sending an
        INVITE
        

        versus a BYE. At this time, the process outlined in
        

        dispatch-adaptive-load-balancing treats all requests the same,
        but we
        

        can definitely
        

        look at ways to change this behaviour.
        

        

        Thanks,
        

        

        - vijay
        

      

      _______________________________________________
      

      dispatch mailing list
      

      dispatch@ietf.org
      

      https://www.ietf.org/mailman/listinfo/dispatch
      

    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------030603080100030709070503
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------030603080100030709070503--

--------------010306000509060100090004--

From vkg@bell-labs.com  Tue Mar  8 12:02:14 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 70B973A6924 for ; Tue,  8 Mar 2011 12:02:14 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 1GkK52OSOJ3G for ; Tue,  8 Mar 2011 12:02:13 -0800 (PST)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 771833A6814 for ; Tue,  8 Mar 2011 12:02:13 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p28K3QX7013214 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 14:03:26 -0600 (CST)
Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28K3Q4Q002010; Tue, 8 Mar 2011 14:03:26 -0600 (CST)
Message-ID: <4D768BEE.2060400@bell-labs.com>
Date: Tue, 08 Mar 2011 14:05:02 -0600
From: "Vijay K. Gurbani" 
Organization: Bell Laboratories, Alcatel-Lucent
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com>
In-Reply-To: <4D7633F9.7000608@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 20:02:14 -0000

On 03/08/2011 07:49 AM, Paul Kyzivat wrote:
> Vijay,
>
> Thanks for clarifications. Now I think I understand what you are
> proposing.  It is certainly simple, and better than nothing.

Paul: Thanks for taking the time to tease out the larger picture.

More inline.

> It *does* require that all the members of the load balancing set
> assign their own weights in a consistent way. This of course is no
> different than when the weights are assigned in DNS, except that its
> likely to be one person who puts all the weights into DNS, while the
> weights coming from the servers can be influenced by who implemented
> them. If they aren't all implemented consistently it can be a
> problem.

While I do not claim that the "weight" scheme is perfect, I think the
bounds for ambiguity are much less with this approach.  All the
upstream SIP proxy needs is a "weight" indicator that it will
use in a RFC2782-type algorithm to proportionally send as much traffic
to a downstream SIP server as it has indicated (i.e., higher "weight"
values will get proportionally more traffic).  This is much better
than round-robin, and furthermore, there is no requirement that each
downstream SIP server in the cluster set its "weight" in a manner
consistent with other servers, or even that the "weight" values of two
SIP servers be comparable.

I have also been thinking about the email exchange we had yesterday.
You had mentioned that:

   > And of course this method of reporting weights isn't sufficient to
   > drive a balancing algorithm that is sensitive to some requests
   > requiring more or different resources than others. (That seems to be
   > a real problem, though I don't know if we have a real solution for
   > it.)

I am starting to believe that this property --- being sensitive to
the processing requirements of some requests versus others --- is
more pertinent to overload control than it is to load balancing.
The "weight" approach normalizes the processing load at a SIP server
in the cluster and reports a canonical value regardless of the specific
*type* of request the SIP server can handle.  If the SIP server is
handling a lot of resource-heavy INVITE transactions, then it will
report a lower "weight" value.

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From paulej@packetizer.com  Tue Mar  8 12:46:54 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AAD663A635F for ; Tue,  8 Mar 2011 12:46:54 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ko2Ys9bu4XyC for ; Tue,  8 Mar 2011 12:46:40 -0800 (PST)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id D170A3A6359 for ; Tue,  8 Mar 2011 12:46:39 -0800 (PST)
Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p28KlkMF002255 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Tue, 8 Mar 2011 15:47:52 -0500
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299617273; bh=9TOapL8vaEbSMGLskhJWzjT81p79hVJEzuKm8rIlLyE=; h=From:To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type; b=WgWRPqDVa5f9BUIVn83d3fWGcpnKHOGacrIrxeVsUkcqWctmsmUeiW0VtJMCzHCjq toNtNeGDOl1D5XlffOSlwgLxaO5ycxzgNTL2ZuInmbL0NLXT5adGAK4wxYS0muhEc6 Zy+DpCnV+9fY4ITUAurLBMw464dVYNGLvls6I0tQ=
From: "Paul E. Jones" 
To: "'Thierry Bessis'" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com>
In-Reply-To: <4D7668B9.1020105@alcatel-lucent.com>
Date: Tue, 8 Mar 2011 15:47:42 -0500
Message-ID: <004601cbddd2$15b11ab0$41135010$@packetizer.com>
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----=_NextPart_000_0047_01CBDDA8.2CE1F080"
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3kCCnyKVgIDWSzZk6WhUeA=
Content-Language: en-us
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 20:46:54 -0000

This is a multipart message in MIME format.

------=_NextPart_000_0047_01CBDDA8.2CE1F080
Content-Type: multipart/alternative;
	boundary="----=_NextPart_001_0048_01CBDDA8.2CE1F080"


------=_NextPart_001_0048_01CBDDA8.2CE1F080
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

Thierry,

 

Please see my comments below in green:

 

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Paul, 

Please let me address some of your comments. 
I don't know what the SCE is, so please pardon me if I misunderstand
something. 

On 07-Mar-2011 23:05, Paul E. Jones wrote: 

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to the
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a
solution to the "overload" problem, the approach being one where we try to
avoid overload by selecting a device that reports a higher capacity value.
Thus, one could view it as "overload control through load balancing".  We
didn't explicitly call it load balancing, though, because we were not trying
to balance load (i.e., place the same load on all servers).  Rather, we just
wanted to direct traffic to devices that we knew had available capacity.

Agreed, the goal seems to be the same: avoiding or at least delaying
overload. 
And your other comment too: our proposal does try to balance the load on all
the servers. The rational is that it is very easy for each server to know
where it is in terms of resource usage (for instance: I'm 60 percent
resource loaded at this moment), and this is just the information you need
to balance the traffic. Now it seems (but it is not clear, see below)  that
the SCE "available capacity" is actually a real time feedback, so the
information would be very similar actually. (What's the difference ? )

 

PEJ: I think it is very similar.  SCE provides real-time feedback of the
available resource capacity of a peer device.  The intent is to convey a
single integer value that represents an estimate of the number of sessions
the peer device can handle.  Note, this is not "message per second" or
other, but the number of sessions.  The focus on SCE is on devices like
"call agents", "session border controllers", stateful SIP proxies, etc. 



 
In practice, the way SCE would be employed is the following:
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
   peer devices
 * The selector would direct traffic toward the peer devices in a
   round-robin fashion, taking into consideration the SCE value so
   as to prevent exhausting resources on any given peer device

I see a contradiction here: if it is round-robin, it does not take into
account the SCE value. 
Is that something more like a weighted round-robin algorithm ? 

 

PEJ: As I mention below, the algorithm used to select a peer device is not a
part of the SCE draft.  One can use any mechanism.  The SCE draft merely
advertises a session capacity estimate.  We've found round-robin to be a
simple approach that works, as it distributes load relatively uniformly, but
stops directing traffic toward devices that report an SCE value that is
below a specified threshold. Other input could be considered, but we
deliberately left the algorithm outside the scope of SCE.



 * Once the peer's SCE value dropped below a given "low-water mark",
   the peer device would be less favored and not given traffic until
   the SCE value increases above the low-water mark (OPTIONS "ping"
   messages may be used to determine the SCE value in the event that
   there is no other traffic)

Why would we wait for a lower mark before to influence the distribution ?
Also is that less favored, or given no traffic ? 

 

PEJ: With our use of SCE, we're not trying to "balance" the load, per se.
One certainly could do that, but in order to balance the load on peer
devices when they are out of balance, one must direct multiple new sessions
toward the same device.  That is not always desirable since new sessions
usually consume more resources temporarily than established sessions.  So,
we try to burden a peer device by throwing a number of new sessions toward
the peer by distributing them in a round-robin fashion.  Again, though, the
approach taken is outside the scope of the SCE draft.

 


The later has stability issues.  And why do we need ping Options when there
is no traffic ?
 If there is no traffic (or very little), should we care about optimal
distribution from this source ?

 

PEJ: The assumption is that a peer device might also have several peers.
So, while device A1 may not be directing traffic toward device B, perhaps
A2, A3, and A4 are.  We can use OPTIONS "ping" messages to get current SCE
information.





 * The current SOC overload control mechanism would be used to
   to prevent message overloading, as SCE only focuses on resource
   overloading (e.g., memory, CPU, disk space, DSPs)

I don't see this difference here (message overload versus resource overload
? ) . To me, SOC focuses on source shedding, while SCE seems to focus on
improved load distribution, just like our proposal.

 

PEJ: A peer device might be able to handle 18,000 new sessions, but perhaps
it can only handle them at a rate of 100 per second.  We would use SCE to
determine that a peer can handle new sessions, while we would use the
methods SOC is defining to prevent message overload so we do not send 200
INVITEs per second to the peer.  (Use of the SOC mechanism might prove to
work better than our current round-robin approach, but we've not tested it.)




 
Note that the "round-robin" approach may or may not be preferred and is not
required in our draft.  I've found this approach works pretty well, but one
may prefer to use a particular weighting that is more complex.  I don't
think local decisions in the box need to be subject to standardization: we
just need to standardize the signaling to enable such decisions.

I believe the signaling is not enough. The deep meaning of the feedback
parameter needs to be clarified too. 
For example, in our proposal, the feedback is simply the current utilization
rate of the node (very simple).

 

PEJ: That sounds similar to SCE.  An SCE-capable device advertises to its
peer how many new sessions it can handle given the number of sessions it is
currently handling. 

Also the expectation needs to be clarified because it has a very important
impact on the recommended behavior in case of overload. 
The solution that we propose constantly tries to balance load in the cluster
of servers. Therefore, in case the SOC shedding kicks in, there is NOT NEED
to repeat the message elsewhere in the cluster because it is balanced
already (So no other server has more capacity). This is a huge advantage for
the network stability as message repeat tends to spread the overload and
collapse the cluster. 



In any case, it certainly looks like we're thinking along the same lines.

Yes I agree. 



PEJ: If you haven't already, have a look at our draft:

 
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

 

It is now expired, but your draft suggests there is still life in this
approach.  We should look to harmonize them, perhaps.

 

Paul

 


Cordially, 

Thierry



 
Paul
 

-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Vijay K. Gurbani
Sent: Monday, March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Paul: Thank you for your input.  More inline.
 
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul Jones
has proposed - with the main difference being the numeric range of the
weight. Do I have that right?

 
I believe that Paul J.'s algorithm is more suited towards overload
control, not load balancing (and I will await Paul J.'s correction to my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing is to
*always* report the current load to give the upstream SIP proxy an
indication of the health of the SIP server in the cluster.  If one waits
until overload is imminent, then it may be too late since a sudden and
unanticipated burst of messages will put the SIP server in overload
rather quickly.
 

It isn't clear to me from your draft if you are specifying the
algorithm by which the relative weights are used to determine the
probability of a request going to one server or another.

 
Yes, same way as RFC2782 does it.
 

ISTM that it *does* matter. And it especially matters if there is an
implication that the upstream node's algorithm may depend on some
assumption or knowledge about the downstream nodes.

 
No knowledge, just a reported number that the upstream SIP proxy can
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
starts off with the SRV indicators in RFC2782, without any regard to
additional knowledge or assumption about downstream nodes.
dispatch-adaptive-load-balancing is trying to use the same model, except
render the weight to be sensitive to the current traffic in the cluster
in real-time.
 

And of course this method of reporting weights isn't sufficient to
drive a balancing algorithm that is sensitive to some requests
requiring more or different resources than others. (That seems to be a
real problem, though I don't know if we have a real solution for
it.)

 
Yes, that is a real problem, as is the problem of a SIP server in a
cluster that has a low weight and is thus not contacted with enough
frequency to merit more participation to the goodput.
 
It is well known in SIP that some methods are more onerous to process
than others --- from a proxying point of view, processing an INVITE with
forking and response aggregation is far more detrimental to resource
usage than is processing a BYE.  Some existing literature techniques
thus add a coefficient to the weight of sending an INVITE versus a BYE.
At this time, the process outlined in dispatch-adaptive-load-balancing
treats all requests the same, but we can definitely look at ways to
change this behaviour.
 
Thanks,
 
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} 
/ vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 


------=_NextPart_001_0048_01CBDDA8.2CE1F080
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Thierry's signature
Thierry,
 
Please see my comments below in green:
 
 
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] =

Sent: Tuesday, March 08, 2011 12:35 PM
To: Paul E. =
Jones
Cc: 'Vijay K. Gurbani'; =
dispatch@ietf.org
Subject: Re: [dispatch] SIP load =
balancing
 
Paul, =


Please let me address some of your comments. 
I don't know =
what the SCE is, so please pardon me if I misunderstand something. =


On 07-Mar-2011 23:05, Paul E. Jones wrote: =

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to =
the
"adaptive load balancing" draft, from =
what I can see.  SCE was proposed as =
a
solution to the "overload" problem, the =
approach being one where we try to
avoid overload =
by selecting a device that reports a higher capacity =
value.
Thus, one could view it as "overload =
control through load balancing".  =
We
didn't explicitly call it load balancing, =
though, because we were not trying
to balance load =
(i.e., place the same load on all servers).  Rather, we =
just
wanted to direct traffic to devices that we =
knew had available capacity.
Agreed, the goal seems to be the same: avoiding or at =
least delaying overload. 
And your other comment too: our proposal =
does try to balance the load on all the servers. The rational is that it =
is very easy for each server to know where it is in terms of resource =
usage (for instance: I'm 60 percent resource loaded at this moment), and =
this is just the information you need to balance the traffic. Now it =
seems (but it is not clear, see below)  that the SCE =
"available capacity" is actually a real time feedback, so the =
information would be very similar actually. (What's the difference ? =
)
 
PEJ: I think it is very similar.  SCE provides real-time =
feedback of the available resource capacity of a peer device.  The =
intent is to convey a single integer value that represents an estimate =
of the number of sessions the peer device can handle.  Note, this =
is not “message per second” or other, but the number of =
sessions.  The focus on SCE is on devices like “call =
agents”, “session border controllers”, stateful SIP =
proxies, etc. =


 
In practice, the =
way SCE would be employed is the following:
* SIP =
proxy, B2BUA, etc. (selector) would receive SCE capacity =
from
   peer =
devices
 * The selector would direct traffic toward =
the peer devices in a
   round-robin =
fashion, taking into consideration the SCE value =
so
   as to prevent exhausting resources =
on any given peer device
I see a =
contradiction here: if it is round-robin, it does not take into account =
the SCE value. 
Is that something more like a weighted round-robin =
algorithm ? 
 
PEJ: As I mention below, the algorithm used to select a peer device =
is not a part of the SCE draft.  One can use any mechanism.  =
The SCE draft merely advertises a session capacity estimate.  =
We’ve found round-robin to be a simple approach that works, as it =
distributes load relatively uniformly, but stops directing traffic =
toward devices that report an SCE value that is below a specified =
threshold. Other input could be considered, but we deliberately left the =
algorithm outside the scope of SCE.

 * =
Once the peer's SCE value dropped below a given "low-water =
mark",
   the peer device would be =
less favored and not given traffic =
until
   the SCE value increases above =
the low-water mark (OPTIONS =
"ping"
   messages may be used =
to determine the SCE value in the event =
that
   there is no other =
traffic)
Why would we wait for a =
lower mark before to influence the distribution ? Also is that less =
favored, or given no traffic ? 
 
PEJ: With our use of SCE, we’re not trying to =
“balance” the load, per se.  One certainly could do =
that, but in order to balance the load on peer devices when they are out =
of balance, one must direct multiple new sessions toward the same =
device.  That is not always desirable since new sessions usually =
consume more resources temporarily than established sessions.  So, =
we try to burden a peer device by throwing a number of new sessions =
toward the peer by distributing them in a round-robin fashion.  =
Again, though, the approach taken is outside the scope of the SCE =
draft.
 

The later has =
stability issues.  And why do we need ping Options when there is no =
traffic ?
 If there is no traffic (or very little), should we =
care about optimal distribution from this source ?
 
PEJ: The assumption is that a peer device might also have several =
peers.  So, while device A1 may not be directing traffic toward =
device B, perhaps A2, A3, and A4 are.  We can use OPTIONS =
“ping” messages to get current SCE =
information.
 =


 * The current SOC overload control =
mechanism would be used to
   to prevent =
message overloading, as SCE only focuses on =
resource
   overloading (e.g., memory, =
CPU, disk space, DSPs)
I don't see =
this difference here (message overload versus resource overload ? ) . To =
me, SOC focuses on source shedding, while SCE seems to focus on improved =
load distribution, just like our proposal.
 
PEJ: A peer device might be able to handle 18,000 new sessions, but =
perhaps it can only handle them at a rate of 100 per second.  We =
would use SCE to determine that a peer can handle new sessions, while we =
would use the methods SOC is defining to prevent message overload so we =
do not send 200 INVITEs per second to the peer.  (Use of the SOC =
mechanism might prove to work better than our current round-robin =
approach, but we’ve not tested it.) =


 
Note that =
the "round-robin" approach may or may not be preferred and is =
not
required in our draft.  I've found this =
approach works pretty well, but one
may prefer to =
use a particular weighting that is more complex.  I =
don't
think local decisions in the box need to be =
subject to standardization: we
just need to =
standardize the signaling to enable such decisions.
I believe the signaling is not enough. The deep =
meaning of the feedback parameter needs to be clarified too. 
For =
example, in our proposal, the feedback is simply the current utilization =
rate of the node (very simple).
 
PEJ: That sounds similar to SCE.  An SCE-capable device =
advertises to its peer how many new sessions it can handle given the =
number of sessions it is currently handling. 

Also the =
expectation needs to be clarified because it has a very important impact =
on the recommended behavior in case of overload. 
The solution that =
we propose constantly tries to balance load in the cluster of servers. =
Therefore, in case the SOC shedding kicks in, there is NOT NEED to =
repeat the message elsewhere in the cluster because it is balanced =
already (So no other server has more capacity). This is a huge advantage =
for the network stability as message repeat tends to spread the overload =
and collapse the cluster. 

In any case, it =
certainly looks like we're thinking along the same =
lines.
Yes I agree. 

PEJ: If you haven’t already, have a look at our =
draft:
http://tools.ietf.org/html/draft-jones-sip-overlo=
ad-sce-00
 
It is now expired, but your draft suggests there is still life in =
this approach.  We should look to harmonize them, =
perhaps.
 
Paul
 

Cordially, =


Thierry

 
Pa=
ul
 
-----Original =
Message-----
From: dispatch-bounces@ietf.org =
[mailto:dispatch-bounces@ietf.or=
g] On
Behalf Of Vijay K. =
Gurbani
Sent: Monday, March 07, 2011 6:38 =
PM
To: dispatch@ietf.org
<=
pre>Subject: Re: [dispatch] SIP load =
balancing
 
Paul: Thank =
you for your input.  More =
inline.
 
On 03/07/2011 =
05:14 PM, Paul Kyzivat wrote:
Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul =
Jones
has proposed - with the main difference being =
the numeric range of the
weight. Do I have that =
right?
 
I =
believe that Paul J.'s algorithm is more suited towards =
overload
control, not load balancing (and I will =
await Paul J.'s correction to =
my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing =
is to
*always* report the current load to give the =
upstream SIP proxy an
indication of the health of =
the SIP server in the cluster.  If one =
waits
until overload is imminent, then it may be =
too late since a sudden and
unanticipated burst of =
messages will put the SIP server in overload
rather =
quickly.
 
It isn't clear to me =
from your draft if you are specifying the
algorithm =
by which the relative weights are used to determine =
the
probability of a request going to one server or =
another.
 
Ye=
s, same way as RFC2782 does =
it.
 
ISTM that it *does* =
matter. And it especially matters if there is =
an
implication that the upstream node's algorithm =
may depend on some
assumption or knowledge about =
the downstream =
nodes.
 
No =
knowledge, just a reported number that the upstream SIP proxy =
can
treat as a RFC2782-like weight indicator.  =
Today, rfc3263-resolution
starts off with the SRV =
indicators in RFC2782, without any regard =
to
additional knowledge or assumption about =
downstream nodes.
dispatch-adaptive-load-balancing =
is trying to use the same model, except
render the =
weight to be sensitive to the current traffic in the =
cluster
in =
real-time.
 
And of course this =
method of reporting weights isn't sufficient =
to
drive a balancing algorithm that is sensitive to =
some requests
requiring more or different resources =
than others. (That seems to be a
real problem, =
though I don't know if we have a real solution =
for
it.)
&nbs=
p;
Yes, that is a real problem, as is the problem of a =
SIP server in a
cluster that has a low weight and =
is thus not contacted with enough
frequency to =
merit more participation to the =
goodput.
 
It is well =
known in SIP that some methods are more onerous to =
process
than others --- from a proxying point of =
view, processing an INVITE with
forking and =
response aggregation is far more detrimental to =
resource
usage than is processing a BYE.  Some =
existing literature techniques
thus add a =
coefficient to the weight of sending an INVITE versus a =
BYE.
At this time, the process outlined in =
dispatch-adaptive-load-balancing
treats all =
requests the same, but we can definitely look at ways =
to
change this =
behaviour.
 
Thanks,<=
/o:p>
 
- =
vijay
--
Vijay K. Gurbani, =
Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, =
Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: =
vkg@{bell-labs.com,acm.or=
g} / vijay.gurbani@alcatel-lu=
cent.com
Web:   http://ect.bell-labs.com/who/v=
kg/
____________________________________________=
___
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
=
 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis 
  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming =
planned Business Trip: 12 April - 14 April 2011
Upcoming =
planned Vacation: None =


------=_NextPart_001_0048_01CBDDA8.2CE1F080--

------=_NextPart_000_0047_01CBDDA8.2CE1F080
Content-Type: image/jpeg;
	name="image001.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
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------=_NextPart_000_0047_01CBDDA8.2CE1F080--


From vkg@bell-labs.com  Tue Mar  8 12:58:37 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 05ED93A672F for ; Tue,  8 Mar 2011 12:58:37 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UJQDvTMOVtPU for ; Tue,  8 Mar 2011 12:58:35 -0800 (PST)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 9C5953A659A for ; Tue,  8 Mar 2011 12:58:35 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p28Kxo5n012829 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 14:59:50 -0600 (CST)
Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28KxoDa019207; Tue, 8 Mar 2011 14:59:50 -0600 (CST)
Message-ID: <4D769926.1050103@bell-labs.com>
Date: Tue, 08 Mar 2011 15:01:26 -0600
From: "Vijay K. Gurbani" 
Organization: Bell Laboratories, Alcatel-Lucent
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7
MIME-Version: 1.0
To: paulej@packetizer.com
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>
In-Reply-To: <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 20:58:37 -0000

On 03/07/2011 11:05 PM, Paul E. Jones wrote:
> Vijay,
>
> The Session Capacity Estimate (SCE) idea is actually quite similar
> to the "adaptive load balancing" draft, from what I can see.

Paul: Thank you much for your input.  More inline.

> SCE was proposed as a solution to the "overload" problem, the
> approach being one where we try to avoid overload by selecting a
> device that reports a higher capacity value. Thus, one could view it
> as "overload control through load balancing".

I think overload control and load balancing are distinct, even though
it is natural to relate them.  Load balancing can delay, or prevent,
overload; but once a SIP host is overloaded, load balancing by itself
will not help too much.

I think it is best that we treat them separately, if not for anything
else than the fact that SOC is not chartered to create a solution for
load balancing.  We (i.e., the dispatch WG) can decide to send this
work to SOC, but I think that decision is a bit premature at this
point.  Furthermore, I understand that before the formation of the
SOC WG, there was some debate on whether or not to include load-
balancing in the charter and it was decided that load-balancing will not
be included.

I believe that we should take the view that load balancing and overload-
control are separate, and that when a cluster of SIP servers is put in
service, all of them start off providing load balancing information to
the upstream SIP proxy.  The upstream SIP proxy will do load balancing
to try to spread the load proportionally to the resources available at
the downstream SIP servers (not evenly, proportionally).  As time
progresses, some of them will enter an overload state and at this time,
the overload control kicks in.  When the overload condition is
alleviated, the overloaded peer goes back to reporting its load.

Now ... to the details of "sce" and "weight" as proposed in our draft.

In the large, I agree that we are thinking along the same lines.  This
is good.

In addition to the email exchange between you and Thierry, there
is one more property that is very attractive: statelessness.

I think that it is advantageous that we impose the property of
statelessness on a load balancing mechanism.  With "sce",
the upstream SIP proxy must keep track of how many requests it
has sent to a downstream SIP server since the latter sent it the
last "sce" value.  Furthermore, each time a request is sent to the
downstream SIP server, the SIP proxy must decrement its request count.
Consequently, the state on how many requests have been sent to the
downstream SIP server since the SIP proxy got the last "sce" parameter,
what the low water-mark is, is all lost if not replicated upon a
reboot.

With "weight", you bypass all of this state maintenance business.

All the SIP proxy needs is a "weight" indicator that it will use
in a RFC2782-type algorithm to proportionally send as much traffic
to a downstream SIP server as it has indicated (i.e., higher "weight"
values will get proportionally more traffic).  This is much better
than round-robin, and furthermore, there is no requirement that each
downstream SIP server in the cluster set its "weight" in a consistent
manner, or even that the "weight" values of two SIP servers be
comparable.

If a SIP server thinks that it has infinite resources, it
can set the "weight" to be 65,535.  Consequently, it will get a
proportionally more traffic if the "weight" of other SIP servers
in the cluster is set to, say, 10 or 20.  Similarly, if the SIP server
in a cluster wants to accept less number of messages, it sets a
lower, albeit non-zero, value for its "weight" parameter.   If the
cluster has capacity, other load control weights will be higher, and
will thus get a proportionate share of the traffic.  As all SIP servers
in the cluster reach capacity, the load distribution algorithm arranges
for a uniform load distribution to all of the SIP servers until either
the offered load subsides or the SIP servers enter into overload mode
and start shedding traffic using the mechanism defined in the overload
control draft.

> I don't think local decisions in the box need to be subject to
> standardization: we just need to standardize the signaling to enable
> such decisions.

I could not agree more.  Local decisions on how each SIP server in a
cluster generates a "weight" value are not subject to standardization.
Transporting this value and specifying a default algorithm to allow
a SIP proxy to perform load balancing seem to be a good avenue for
standardization.

Note that I think specifying a default algorithm to do load balancing
is important, and our draft chooses to use the well-understood RFC2782
usage of "weight" to do this.  This does not preclude any implementer
from coming up with their own algorithm and using that instead.

> In any case, it certainly looks like we're thinking along the same
> lines.

Indeed.  If this work progresses, it'll be nice to have a coherent
view going forward.  If you will be in Prague, it will be good to
hook up and chat about this.

Once again, thanks for your input.

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From pkyzivat@cisco.com  Tue Mar  8 12:59:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 873833A672F for ; Tue,  8 Mar 2011 12:59:51 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.532
X-Spam-Level: 
X-Spam-Status: No, score=-110.532 tagged_above=-999 required=5 tests=[AWL=0.067, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cygmmAp6+5fj for ; Tue,  8 Mar 2011 12:59:50 -0800 (PST)
Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id 349113A659A for ; Tue,  8 Mar 2011 12:59:50 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=3514; q=dns/txt; s=iport; t=1299618066; x=1300827666; h=message-id:date:from:mime-version:to:cc:subject: references:in-reply-to:content-transfer-encoding; bh=+sSqcIPZ5HJCmdv+pcYxeYZRiYaU5OgpHb28t2wCZdM=; b=D/JB7IIlX04aSeVYKMv/QSsxGC+y+z+cGAcbMArflnHo7pzuYrJ0EtP4 aGOKcUDc+hYl841WsVBirpudeKvKgFnEt/dQPVPmW8S4qiXByp9rXb60/ YAw/Yj6eVGK3LEFJz7pXaMVk1umPawTq/khxgV4M/OU8ltkaqvs8dTMnv c=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAJ4ndk1AZnwM/2dsb2JhbACmZXSjYZwshWMEhR2HFYNH
X-IronPort-AV: E=Sophos;i="4.62,285,1297036800"; d="scan'208";a="223367437"
Received: from rtp-core-1.cisco.com ([64.102.124.12]) by rtp-iport-1.cisco.com with ESMTP; 08 Mar 2011 21:00:48 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rtp-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id p28L0jm5019131; Tue, 8 Mar 2011 21:00:45 GMT
Message-ID: <4D7698FD.4030501@cisco.com>
Date: Tue, 08 Mar 2011 16:00:45 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: "Vijay K. Gurbani" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>
In-Reply-To: <4D768BEE.2060400@bell-labs.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 20:59:51 -0000

On 3/8/2011 3:05 PM, Vijay K. Gurbani wrote:
> On 03/08/2011 07:49 AM, Paul Kyzivat wrote:
>> Vijay,
>>
>> Thanks for clarifications. Now I think I understand what you are
>> proposing. It is certainly simple, and better than nothing.
>
> Paul: Thanks for taking the time to tease out the larger picture.
>
> More inline.
>
>> It *does* require that all the members of the load balancing set
>> assign their own weights in a consistent way. This of course is no
>> different than when the weights are assigned in DNS, except that its
>> likely to be one person who puts all the weights into DNS, while the
>> weights coming from the servers can be influenced by who implemented
>> them. If they aren't all implemented consistently it can be a
>> problem.
>
> While I do not claim that the "weight" scheme is perfect, I think the
> bounds for ambiguity are much less with this approach. All the
> upstream SIP proxy needs is a "weight" indicator that it will
> use in a RFC2782-type algorithm to proportionally send as much traffic
> to a downstream SIP server as it has indicated (i.e., higher "weight"
> values will get proportionally more traffic). This is much better
> than round-robin, and furthermore, there is no requirement that each
> downstream SIP server in the cluster set its "weight" in a manner
> consistent with other servers, or even that the "weight" values of two
> SIP servers be comparable.

I don't understand that.

Suppose we have two servers that are identical except that they have 
different understandings of the weight assignment algorithm.

One thinks it is a powerful server, and sets its initial (unloaded) 
weight to 32k. The other thinks that it is not especially powerful as 
servers go, and so sets its weight to 4k.

That means that initially the first will get sent 8 times as much 
traffic as the other. So of course it will get more heavily loaded. As 
it does, it will presumably reduce its weight, but for a given load it 
will always have a larger number than the other.

> I have also been thinking about the email exchange we had yesterday.
> You had mentioned that:
>
>  > And of course this method of reporting weights isn't sufficient to
>  > drive a balancing algorithm that is sensitive to some requests
>  > requiring more or different resources than others. (That seems to be
>  > a real problem, though I don't know if we have a real solution for
>  > it.)
>
> I am starting to believe that this property --- being sensitive to
> the processing requirements of some requests versus others --- is
> more pertinent to overload control than it is to load balancing.
> The "weight" approach normalizes the processing load at a SIP server
> in the cluster and reports a canonical value regardless of the specific
> *type* of request the SIP server can handle. If the SIP server is
> handling a lot of resource-heavy INVITE transactions, then it will
> report a lower "weight" value.

The thing is, a server may be able to handle a lot of some type of load 
while being totally out of capacity for some other type of load.

(E.g. a box that is both an SBC to a provider sip trunk and a gw to some 
PSTN lines. It might be totally out of PSTN lines yet be able to handle 
lots of calls that happen to be addressed to that they go to the sip trunk.)

Also, shedding load is harder when the consumed resource is dialog stateful.

I think this is a hard problem.

	Thanks,
	Paul

> - vijay

From vkg@bell-labs.com  Tue Mar  8 13:59:21 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AB51E3A676A for ; Tue,  8 Mar 2011 13:59:21 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id scGTATOsfCjR for ; Tue,  8 Mar 2011 13:59:20 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 77A123A6767 for ; Tue,  8 Mar 2011 13:59:20 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p28M0YTs028981 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 16:00:34 -0600 (CST)
Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28M0Xcm008301; Tue, 8 Mar 2011 16:00:33 -0600 (CST)
Message-ID: <4D76A762.6090708@bell-labs.com>
Date: Tue, 08 Mar 2011 16:02:10 -0600
From: "Vijay K. Gurbani" 
Organization: Bell Laboratories, Alcatel-Lucent
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com>
In-Reply-To: <4D7698FD.4030501@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 21:59:21 -0000

On 03/08/2011 03:00 PM, Paul Kyzivat wrote:
> I don't understand that.
>
> Suppose we have two servers that are identical except that they have
> different understandings of the weight assignment algorithm.
>
> One thinks it is a powerful server, and sets its initial (unloaded)
> weight to 32k. The other thinks that it is not especially powerful as
> servers go, and so sets its weight to 4k.
>
> That means that initially the first will get sent 8 times as much
> traffic as the other. So of course it will get more heavily loaded.
> As it does, it will presumably reduce its weight, but for a given
> load it will always have a larger number than the other.

Paul: Good point.  However, note that if the server used AIMD strategy,
it will converge to an equilibrium point where it is reporting < 4K
"weight".  At that time, the other server becomes the preferred server.

I suspect that a good algorithm for reducing the "weight" value would
use some sort of AIMD strategy.

> The thing is, a server may be able to handle a lot of some type of
> load while being totally out of capacity for some other type of
> load.
>
> (E.g. a box that is both an SBC to a provider sip trunk and a gw to
> some PSTN lines. It might be totally out of PSTN lines yet be able to
> handle lots of calls that happen to be addressed to that they go to
> the sip trunk.)

True, but at least it has enough processing capacity to reject
the calls to PSTN lines and accept calls to the SIP trunk.  Rejecting
the PSTN calls at the sink is better than forcing an upstream server
to shed them.

> Also, shedding load is harder when the consumed resource is dialog
> stateful.
>
> I think this is a hard problem.

No doubt, but I will sweep the which-message-to-shed problem
under the overload rug, and not worry about it when dusting the
load balancing rug ;-)

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From vkg@bell-labs.com  Tue Mar  8 14:25:47 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8C6033A672F for ; Tue,  8 Mar 2011 14:25:47 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id YgNJPc5+EiyU for ; Tue,  8 Mar 2011 14:25:46 -0800 (PST)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 69C1C3A659B for ; Tue,  8 Mar 2011 14:25:46 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p28MR0IZ020729 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 8 Mar 2011 16:27:00 -0600 (CST)
Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p28MR0Jr026832; Tue, 8 Mar 2011 16:27:00 -0600 (CST)
Message-ID: <4D76AD94.4000108@bell-labs.com>
Date: Tue, 08 Mar 2011 16:28:36 -0600
From: "Vijay K. Gurbani" 
Organization: Bell Laboratories, Alcatel-Lucent
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7
MIME-Version: 1.0
To: "Parthasarathi R (partr)" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> 
In-Reply-To: 
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 08 Mar 2011 22:25:47 -0000

On 03/08/2011 03:52 AM, Parthasarathi R (partr) wrote:
>[...] Even though any overload control algorithm
> is used ("oc" or "weight" or "SCE" or "resource-availability" other),
> the locally overload control mechanism is required in downstream as per
> mentioned in draft-ietf-soc-overload-design. I wish that we will look
> into these aspect because overload control is the extreme case for load
> balancing the cluster of downstream entity and load balancing mechanism
> shall help to avoid the overload control based on feedback.

Partha: Thank you for your input.  I do agree that overload control
is the extreme case for load balancing.  However, as my email to
Paul J. indicated, treating them differently may help, or we could
make load balancing an extension under overload control.  I don't
know what the answer is ... we'd have to hash it out in dispatch
on where to send load balancing to.  My understanding is that the
charter for SOC does not account for load balancing.

> I guess that your mechanism assumes about downstream SIP UA as signaling
> entity only and does not handle media. Please correct me in case I
> wrongly understand.

Our mechanism does not distinguish between media or signaling in the
sense that the "weight" that is reported is a normalized value that
is calculated by the SIP server in a cluster.  The SIP server could be
a SIP bridge, for instance, in which case it may be able to handle a
lot of point-to-point calls, but can only handle few calls
requiring full mesh conferencing capabilities.  As long as the SIP
server is handling the low-resource point-to-point calls, it reports
a higher "weight" value.  But as soon as a conference call with a large
number of participants commences, using other resources like mixers
and transcoders, etc., the SIP server starts to report lower "weight"
values.

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From ian_elz@yahoo.co.uk  Tue Mar  8 16:20:05 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 88B973A67D6 for ; Tue,  8 Mar 2011 16:20:05 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5efQm0+a9Mrc for ; Tue,  8 Mar 2011 16:20:04 -0800 (PST)
Received: from nm8-vm1.bullet.mail.ird.yahoo.com (nm8-vm1.bullet.mail.ird.yahoo.com [77.238.189.198]) by core3.amsl.com (Postfix) with SMTP id CDB6F3A67D2 for ; Tue,  8 Mar 2011 16:20:03 -0800 (PST)
Received: from [77.238.189.48] by nm8.bullet.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 00:21:16 -0000
Received: from [212.82.108.114] by tm1.bullet.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 00:21:16 -0000
Received: from [127.0.0.1] by omp1023.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 00:21:16 -0000
X-Yahoo-Newman-Property: ymail-5
X-Yahoo-Newman-Id: 234733.51170.bm@omp1023.mail.ird.yahoo.com
Received: (qmail 268 invoked by uid 60001); 9 Mar 2011 00:21:16 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.co.uk; s=s1024; t=1299630076; bh=AIGXZbZkOUi5sukdhKVT2dfaPyFOc6Dn7R6FBZCyiYY=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Reply-To:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=XRoS0DTsMhafjzA4tn7B50gZ35ihLw1f+4KhKnNKhWqV+IZEI5Q2N4u03oCOj4NsQTMG4P6F+NIeQE46OWtTMWXFyX+Ea/7IhBMROoe6hv1lcs8ZHwVq4wbfxBS3oFTqI1fXf4fdisUrj+VwzUJmVKadFxr+2ARbGKz7YBw2Tzw=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Reply-To:Subject:To:Cc:MIME-Version:Content-Type:Content-Transfer-Encoding; b=oiVf0OAEfXru/0MJSH3BOulQLbwITrK2AmgbdA5V2dKbmJMl8962re0YjpEf8VHLGaENz+RsNFU2zEqM9SjExxPlG/O8t9ndsOujm6rmzyCaW5gfI43xY0bk23MRl485PZNMqiaJ+UUNd14eiEN6WBtKsEF3+Ku4+doLC4KVkDg=;
Message-ID: <120652.93381.qm@web29106.mail.ird.yahoo.com>
X-YMail-OSG: _3jvT.YVM1kpYQwQTzQsLDYFnfYnV3CNrGfcbI_nCVsEcTy hB.SESih6mmSlR9F2ZWE0hATN_tSByoUgkn1W_jY1XHTfYVRYj6u.9HsuRKr Fge18q3kWd_CfvLPLiDG5jOHaF6KmxEo.eCaF4OAP3z4qSRQg00.Z0m98DnU Kujj4H0EW1Ei30NCkmtjA9GNikSL6o1I3LY0.WNZ_4LOShqGwF8YXxmOJsBP Wxwl4WXQpcYocPAP_0CLVQx5VGIJZXo.TspZyg25Y10g-
Received: from [86.20.66.247] by web29106.mail.ird.yahoo.com via HTTP; Wed, 09 Mar 2011 00:21:16 GMT
X-Mailer: YahooMailWebService/0.8.109.295617
Date: Wed, 9 Mar 2011 00:21:16 +0000 (GMT)
From: Ian Elz 
To: Paul Kyzivat 
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Cc: dispatch@ietf.org
Subject: Re: [dispatch] New Draft -	draft-elz-dispatch-req-sip-3rdpartyprivacy-00
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: Ian Elz 
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 00:20:05 -0000

Paul,=0A=0AI don't believe that a generalized mechanism is difficult. It ju=
st needs a little thought.=0A=0AHistory-info has defined a mechanism for ad=
ding privacy to an individual uri. It just need extending so that any uri c=
an use a similar mechanism.=0A=0AThe second issue regarding the precedence =
is the one which people may have opinions about.=0A=0AIan=0A=0A----- Origin=
al Message -----=0AFrom: "Paul Kyzivat" =0ATo: dispatch=
@ietf.org=0ASent: Tuesday, 8 March, 2011 2:12:39 PM=0ASubject: Re: [dispatc=
h] New Draft -=09draft-elz-dispatch-req-sip-3rdpartyprivacy-00=0A=0AIan,=0A=
=0AI'd like to thank you for submitting a clear requirements draft that is =
=0Aunburdened by mechanism proposals.=0A=0AIMO the hard part of this issue =
is sorting out whether there is a need =0Afor a "generalized" mechanism dea=
ling with arbitrary 3rd party =0Aidentities, or just for a mechanism to dea=
l with a specific enumerated =0Aset of 3rd party identities.=0A=0AWhile its=
 easy to wish for a generalized mechanism, I think it will be =0Avery diffi=
cult to adequately specify. If constrained to a specific set =0Athan it is =
a solvable problem.=0A=0A=09Thanks,=0A=09Paul=0A=0AOn 3/8/2011 3:15 AM, Ian=
 Elz wrote:=0A> All,=0A>=0A> I have submitted a new draft relating to third=
 party privacy in SIP=0A> messages.=0A>=0A> http://www.ietf.org/id/draft-el=
z-dispatch-req-sip-3rdpartyprivacy-00.txt=0A> (Link that I forgot in the pr=
evious posting.)=0A>=0A> There are two specific issues included, how to spe=
cify third party=0A> privacy in a SIP message; the precedence of privacy se=
ttings between=0A> Privacy header values and Privacy URI parameters.=0A>=0A=
> The things to consider are:=0A>=0A> Should this work be pursued and if so=
 in which forum.=0A>=0A> A suitable solution to meet the requirements.=0A>=
=0A> How any changes should be implemented: a new draft which modifies=0A> =
RFC3323 or an update to RFC3323 which includes the additional work.=0A>=0A>=
 All comments welcome.=0A>=0A> Sorry I forgot to write the Abstract in vers=
ion -00. I will include this=0A> in the next version. ( I leave abstracts t=
o last and overlooked this=0A> before submission.)=0A>=0A> Ian Elz=0A>=0A>=
=0A>=0A> _______________________________________________=0A> dispatch maili=
ng list=0A> dispatch@ietf.org=0A> https://www.ietf.org/mailman/listinfo/dis=
patch=0A=0A=0A=0A=0A      

From pkyzivat@cisco.com  Tue Mar  8 16:24:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0F2243A67D6 for ; Tue,  8 Mar 2011 16:24:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.533
X-Spam-Level: 
X-Spam-Status: No, score=-110.533 tagged_above=-999 required=5 tests=[AWL=0.066, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7kGGVfT658Fp for ; Tue,  8 Mar 2011 16:24:58 -0800 (PST)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 8FAE73A67D4 for ; Tue,  8 Mar 2011 16:24:57 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=3038; q=dns/txt; s=iport; t=1299630373; x=1300839973; h=message-id:date:from:mime-version:to:cc:subject: references:in-reply-to:content-transfer-encoding; bh=IUiKGUYYhgpSdi/BhIb/bWyh9B6AkNMiTiUVnk+EvDY=; b=lYUs1rErEYOcfE96xo/c/VKiVPJffp4B8Kl/AbFndRFhP1tXafloRVPm DArgP/057zT7jFlparfdnh4aSWII+uxNyTnC01m4iuZMS6B15DxfXMhsE oFdZu/No92yXNfgD4eqH9ZEHt9QzSa6gw8FMFcwSwIs1JazQ4u8GeJC+x I=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAG9Xdk1AZnwN/2dsb2JhbACmZ3SlHZw0hWMEhR2HFYNH
X-IronPort-AV: E=Sophos;i="4.62,286,1297036800"; d="scan'208";a="223788904"
Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-2.cisco.com with ESMTP; 09 Mar 2011 00:26:12 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id p290QCQk026772; Wed, 9 Mar 2011 00:26:12 GMT
Message-ID: <4D76C924.7050900@cisco.com>
Date: Tue, 08 Mar 2011 19:26:12 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: "Vijay K. Gurbani" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>
In-Reply-To: <4D76A762.6090708@bell-labs.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 00:24:59 -0000

On 3/8/2011 5:02 PM, Vijay K. Gurbani wrote:
> On 03/08/2011 03:00 PM, Paul Kyzivat wrote:
>> I don't understand that.
>>
>> Suppose we have two servers that are identical except that they have
>> different understandings of the weight assignment algorithm.
>>
>> One thinks it is a powerful server, and sets its initial (unloaded)
>> weight to 32k. The other thinks that it is not especially powerful as
>> servers go, and so sets its weight to 4k.
>>
>> That means that initially the first will get sent 8 times as much
>> traffic as the other. So of course it will get more heavily loaded.
>> As it does, it will presumably reduce its weight, but for a given
>> load it will always have a larger number than the other.
>
> Paul: Good point. However, note that if the server used AIMD strategy,
> it will converge to an equilibrium point where it is reporting < 4K
> "weight". At that time, the other server becomes the preferred server.

I guess I need more explanation of how that algorithm applies to this 
problem. Are you saying that the "weight" is the "window size", and that 
the server increases/decreases it whenever the incoming load is not 
keeping it at the optimal loading? Sounds interesting. But I know next 
to nothing about that algorithm or the statistical behavior of such 
algorithms.

> I suspect that a good algorithm for reducing the "weight" value would
> use some sort of AIMD strategy.
>
>> The thing is, a server may be able to handle a lot of some type of
>> load while being totally out of capacity for some other type of
>> load.
>>
>> (E.g. a box that is both an SBC to a provider sip trunk and a gw to
>> some PSTN lines. It might be totally out of PSTN lines yet be able to
>> handle lots of calls that happen to be addressed to that they go to
>> the sip trunk.)
>
> True, but at least it has enough processing capacity to reject
> the calls to PSTN lines and accept calls to the SIP trunk. Rejecting
> the PSTN calls at the sink is better than forcing an upstream server
> to shed them.

Yes, up to a point. If we are balancing across a lot of servers, and 
most of them are out of pstn lines (because we are close to max 
capacity) then rejecting at the server may result in the same request 
having to be retried many times before finding a server that can handle 
the request. Once that number gets above a few the call wait time gets 
too high to be acceptable. You need a better way to pick the box that 
has resources to handle the particular request.

But I don't see how one characterizes the resource availability/limits 
in a way that the upstream node can effectively use to make a good decision.

	Thanks,
	Paul

>> Also, shedding load is harder when the consumed resource is dialog
>> stateful.
>>
>> I think this is a hard problem.
>
> No doubt, but I will sweep the which-message-to-shed problem
> under the overload rug, and not worry about it when dusting the
> load balancing rug ;-)
>
> Thanks,
>
> - vijay

From pkyzivat@cisco.com  Tue Mar  8 16:58:25 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A9D093A67D6 for ; Tue,  8 Mar 2011 16:58:25 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.535
X-Spam-Level: 
X-Spam-Status: No, score=-110.535 tagged_above=-999 required=5 tests=[AWL=0.064, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mlE5I+qVtc07 for ; Tue,  8 Mar 2011 16:58:24 -0800 (PST)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 78DB53A67E3 for ; Tue,  8 Mar 2011 16:58:24 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=2768; q=dns/txt; s=iport; t=1299632380; x=1300841980; h=message-id:date:from:mime-version:to:cc:subject: references:in-reply-to:content-transfer-encoding; bh=W56L1eINujKsYRtDEwXSCYbILlzHfd2IZ9af5PRN3Uw=; b=gHLNplNWbGNlR4ByujbREemA/Y+ASW6lIPIqWULKO+lGgQa3amcibsEV nnv/TObzellsn9Pm4OtzcuQKKRob6AkU9QvMEcgT7sN6EVaEVGrqfgvJ3 rvW6dP/MrfRsijLauJSUEhlF70EH4cE8BRfzdH60RfXlkRmjNLWtwUWnI I=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Av0EAKNfdk1AZnwM/2dsb2JhbACELKI7dKV1iwyRLYEng0Z2BIUdhxWDRw
X-IronPort-AV: E=Sophos;i="4.62,286,1297036800"; d="scan'208";a="223794507"
Received: from rtp-core-1.cisco.com ([64.102.124.12]) by rtp-iport-2.cisco.com with ESMTP; 09 Mar 2011 00:59:39 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rtp-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id p290xdma026226; Wed, 9 Mar 2011 00:59:39 GMT
Message-ID: <4D76D0FB.80005@cisco.com>
Date: Tue, 08 Mar 2011 19:59:39 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: Ian Elz 
References: <120652.93381.qm@web29106.mail.ird.yahoo.com>
In-Reply-To: <120652.93381.qm@web29106.mail.ird.yahoo.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: dispatch@ietf.org
Subject: Re: [dispatch] New Draft -	draft-elz-dispatch-req-sip-3rdpartyprivacy-00
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 00:58:25 -0000

On 3/8/2011 7:21 PM, Ian Elz wrote:
> Paul,
>
> I don't believe that a generalized mechanism is difficult. It just needs a little thought.
>
> History-info has defined a mechanism for adding privacy to an individual uri. It just need extending so that any uri can use a similar mechanism.

The problem with generalizing is that boxes implementing privacy need to 
be updated every time there is a new place header, etc. defined that can 
contain something requiring privacy. At the moment that happens the 
world is filled with boxes that don't support privacy for that thing. 
How do you determine if you have the privacy you need?

	Thanks,
	Paul

> The second issue regarding the precedence is the one which people may have opinions about.
>
> Ian
>
> ----- Original Message -----
> From: "Paul Kyzivat"
> To: dispatch@ietf.org
> Sent: Tuesday, 8 March, 2011 2:12:39 PM
> Subject: Re: [dispatch] New Draft -	draft-elz-dispatch-req-sip-3rdpartyprivacy-00
>
> Ian,
>
> I'd like to thank you for submitting a clear requirements draft that is
> unburdened by mechanism proposals.
>
> IMO the hard part of this issue is sorting out whether there is a need
> for a "generalized" mechanism dealing with arbitrary 3rd party
> identities, or just for a mechanism to deal with a specific enumerated
> set of 3rd party identities.
>
> While its easy to wish for a generalized mechanism, I think it will be
> very difficult to adequately specify. If constrained to a specific set
> than it is a solvable problem.
>
> 	Thanks,
> 	Paul
>
> On 3/8/2011 3:15 AM, Ian Elz wrote:
>> All,
>>
>> I have submitted a new draft relating to third party privacy in SIP
>> messages.
>>
>> http://www.ietf.org/id/draft-elz-dispatch-req-sip-3rdpartyprivacy-00.txt
>> (Link that I forgot in the previous posting.)
>>
>> There are two specific issues included, how to specify third party
>> privacy in a SIP message; the precedence of privacy settings between
>> Privacy header values and Privacy URI parameters.
>>
>> The things to consider are:
>>
>> Should this work be pursued and if so in which forum.
>>
>> A suitable solution to meet the requirements.
>>
>> How any changes should be implemented: a new draft which modifies
>> RFC3323 or an update to RFC3323 which includes the additional work.
>>
>> All comments welcome.
>>
>> Sorry I forgot to write the Abstract in version -00. I will include this
>> in the next version. ( I leave abstracts to last and overlooked this
>> before submission.)
>>
>> Ian Elz
>>
>>
>>
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>
>
>
>
>
>

From ian_elz@yahoo.co.uk  Wed Mar  9 00:18:40 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 06C313A688C for ; Wed,  9 Mar 2011 00:18:40 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.226
X-Spam-Level: 
X-Spam-Status: No, score=-2.226 tagged_above=-999 required=5 tests=[AWL=0.372,  BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AdtSaO2ie89z for ; Wed,  9 Mar 2011 00:18:38 -0800 (PST)
Received: from nm9-vm1.bullet.mail.ird.yahoo.com (nm9-vm1.bullet.mail.ird.yahoo.com [77.238.189.94]) by core3.amsl.com (Postfix) with SMTP id 3DB0C3A6841 for ; Wed,  9 Mar 2011 00:18:38 -0800 (PST)
Received: from [77.238.189.234] by nm9.bullet.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 08:19:49 -0000
Received: from [212.82.108.116] by tm15.bullet.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 08:19:48 -0000
Received: from [127.0.0.1] by omp1025.mail.ird.yahoo.com with NNFMP; 09 Mar 2011 08:19:48 -0000
X-Yahoo-Newman-Property: ymail-5
X-Yahoo-Newman-Id: 911090.97573.bm@omp1025.mail.ird.yahoo.com
Received: (qmail 20583 invoked by uid 60001); 9 Mar 2011 08:19:48 -0000
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.co.uk; s=s1024; t=1299658788; bh=5rqm+blSstQw/OYXKvrMrY0+RVSW2Ka+CihASeGtxjY=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:References:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type; b=2e8KraONsitm6xsNwdwWfRyUy6G9XrqBjK1ls62dogdB8pCQ+F6K35xbolwoPnWafOoqwvOBD8+pFJXCMUiqdaQ9CsM7+lEMD0rz0NgJz18SyIZV1iYXiwI1qIerRIKFsGk1Egltr4dzzcQtwbyf4L0eqPyNqp0Onlbnbd9SuBA=
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk; h=Message-ID:X-YMail-OSG:Received:X-Mailer:References:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type; b=jFsveFRY2X+w+41lbcYdiK/tuAqlJslvUeqmR6JeVGh8v1xHOXjdsfxJesfOt95iVEXl720ilh91qRlYwAc2/q+MHGejAEBHZWbsQzA7+mRZgen4fGqoeFy9QgoyGJOh9UNRQw97ZpOzDFyv/ajzw8yG8T87oT6mOmjVxGnctu4=;
Message-ID: <777439.29308.qm@web29101.mail.ird.yahoo.com>
X-YMail-OSG: xjFbeE8VM1kE.BSciItFjMLz_ehr84U0nMlJiNspvvH036m 0PXY1qJHXxiaZk.E__XaoYKX9webxyJZTxkc__loR2djK98rbP5SBYQ74YIW 28S9Z5moQQvoMmiHwT9zCUU02RF_0A6_khW6JcqMn38YbLFBCNLDNlNsJw6H ktn8LDDzBFEKFk8KkDKzZ3QdL_JxFItq7aP9lSO76UcJiidShlHodtrMkskO pVr3eGBLwa2AIL709RaSCWoZA1Pik.HOZwZ_DprRakITsk2w-
Received: from [80.231.29.53] by web29101.mail.ird.yahoo.com via HTTP; Wed, 09 Mar 2011 08:19:48 GMT
X-Mailer: YahooMailRC/559 YahooMailWebService/0.8.109.295617
References: <120652.93381.qm@web29106.mail.ird.yahoo.com> <4D76D0FB.80005@cisco.com>
Date: Wed, 9 Mar 2011 08:19:48 +0000 (GMT)
From: Ian Elz 
To: Paul Kyzivat 
In-Reply-To: <4D76D0FB.80005@cisco.com>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="0-1236837020-1299658788=:29308"
Cc: dispatch@ietf.org
Subject: Re: [dispatch] New Draft - draft-elz-dispatch-req-sip-3rdpartyprivacy-00
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 08:18:40 -0000

--0-1236837020-1299658788=:29308
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Paul,=0A=0AI would have thought that using the present system that any new =
header added =0Athat requires privacy would require that the box providing =
the privacy service =0Awould need to be updated. E.g. If a new header "Xxxx=
" is defined which requires =0Aprivacy and can be classified as "user" priv=
acy in the existing mechanism then =0Athe privacy box would need to be upda=
ted to ensure that Header "Xxxx" was =0Aremoved when "user" privacy was ind=
icated but not if "header" privacy is =0Aindicated.=0A=0AIf we can defined =
a generic mechanism that allows the privacy box to identify =0Awhether priv=
acy is required irrespective of the header then there may be less =0Achange=
s required at the privacy box. E.g. If it was possible to mark a header =0A=
containing a uri with a privacy header indicating that privacy is required =
then =0Athe privacy box would know to remove the header even if it did not =
understand =0Athe specific header. This could be as follows: =0A=0A=0AXxxx:=
 =0ATo: Ian Elz =0ACc=
: dispatch@ietf.org=0ASent: Wed, 9 March, 2011 0:59:39=0ASubject: Re: [disp=
atch] New Draft - =0Adraft-elz-dispatch-req-sip-3rdpartyprivacy-00=0A=0A=0A=
=0AOn 3/8/2011 7:21 PM, Ian Elz wrote:=0A> Paul,=0A>=0A> I don't believe th=
at a generalized mechanism is difficult. It just needs a =0A>little thought=
.=0A>=0A> History-info has defined a mechanism for adding privacy to an ind=
ividual uri. =0A>It just need extending so that any uri can use a similar m=
echanism.=0A=0AThe problem with generalizing is that boxes implementing pri=
vacy need to =0Abe updated every time there is a new place header, etc. def=
ined that can =0Acontain something requiring privacy. At the moment that ha=
ppens the =0Aworld is filled with boxes that don't support privacy for that=
 thing. =0AHow do you determine if you have the privacy you need?=0A=0A    =
Thanks,=0A    Paul=0A=0A> The second issue regarding the precedence is the =
one which people may have =0A>opinions about.=0A>=0A> Ian=0A>=0A> ----- Ori=
ginal Message -----=0A> From: "Paul Kyzivat"=0A> To: di=
spatch@ietf.org=0A> Sent: Tuesday, 8 March, 2011 2:12:39 PM=0A> Subject: Re=
: [dispatch] New Draft -    =0A>draft-elz-dispatch-req-sip-3rdpartyprivacy-=
00=0A>=0A> Ian,=0A>=0A> I'd like to thank you for submitting a clear requir=
ements draft that is=0A> unburdened by mechanism proposals.=0A>=0A> IMO the=
 hard part of this issue is sorting out whether there is a need=0A> for a "=
generalized" mechanism dealing with arbitrary 3rd party=0A> identities, or =
just for a mechanism to deal with a specific enumerated=0A> set of 3rd part=
y identities.=0A>=0A> While its easy to wish for a generalized mechanism, I=
 think it will be=0A> very difficult to adequately specify. If constrained =
to a specific set=0A> than it is a solvable problem.=0A>=0A>     Thanks,=0A=
>     Paul=0A>=0A> On 3/8/2011 3:15 AM, Ian Elz wrote:=0A>> All,=0A>>=0A>> =
I have submitted a new draft relating to third party privacy in SIP=0A>> me=
ssages.=0A>>=0A>> http://www.ietf.org/id/draft-elz-dispatch-req-sip-3rdpart=
yprivacy-00.txt=0A>> (Link that I forgot in the previous posting.)=0A>>=0A>=
> There are two specific issues included, how to specify third party=0A>> p=
rivacy in a SIP message; the precedence of privacy settings between=0A>> Pr=
ivacy header values and Privacy URI parameters.=0A>>=0A>> The things to con=
sider are:=0A>>=0A>> Should this work be pursued and if so in which forum.=
=0A>>=0A>> A suitable solution to meet the requirements.=0A>>=0A>> How any =
changes should be implemented: a new draft which modifies=0A>> RFC3323 or a=
n update to RFC3323 which includes the additional work.=0A>>=0A>> All comme=
nts welcome.=0A>>=0A>> Sorry I forgot to write the Abstract in version -00.=
 I will include this=0A>> in the next version. ( I leave abstracts to last =
and overlooked this=0A>> before submission.)=0A>>=0A>> Ian Elz=0A>>=0A>>=0A=
>>=0A>> _______________________________________________=0A>> dispatch maili=
ng list=0A>> dispatch@ietf.org=0A>> https://www.ietf.org/mailman/listinfo/d=
ispatch=0A>=0A>=0A>=0A>=0A>=0A>=0A=0A=0A=0A      
--0-1236837020-1299658788=:29308
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable


Paul,

I would have thought that using the present system that any=
 new header added that requires privacy would require that the box providin=
g the privacy service would need to be updated. E.g. If a new header "Xxxx"=
 is defined which requires privacy and can be classified as "user" privacy =
in the existing mechanism then the privacy box would need to be updated to =
ensure that Header "Xxxx" was removed when "user" privacy was indicated but=
 not if "header" privacy is indicated.

If we can defined a generic m=
echanism that allows the privacy box to identify whether privacy is require=
d irrespective of the header then there may be less changes required at the=
 privacy box. E.g. If it was possible to mark a header containing a uri wit=
h a privacy header indicating that privacy is required then the privacy
 box would know to remove the header even if it did not understand the spec=
ific header. This could be as follows: 

Xxxx: <alice@example.com;=
'private'    (not suggesting that this is the mechanism for =
marking privacy)

The privacy box could identify the privacy setting =
and either remove the header if it is optional or anonymise the header if i=
t is required in SIP.

At present I am more interested in whether the=
re is agreement with pursuing the work than in agreeing a solution.

=
Thanks

Ian
From: Paul Kyzivat <pk=
yzivat@cisco.com>
To: Ian Elz <ian_elz@yahoo.co.uk>
Cc: dispatch@ietf.org
Sent: Wed, 9 March, 2011 0:59:39
Subject: Re: [dispatch] New Draft - draft-elz-d=
ispatch-req-sip-3rdpartyprivacy-00



On 3/8/2011 7:21 P=
M, Ian Elz wrote:
> Paul,
>
> I don't believe that a gene=
ralized mechanism is difficult. It just needs a little thought.
>
=
> History-info has defined a mechanism for adding privacy to an individu=
al uri. It just need extending so that any uri can use a similar mechanism.=


The problem with generalizing is that boxes implementing privacy ne=
ed to 
be updated every time there is a new place header, etc. defined t=
hat can 
contain something requiring privacy. At the moment that happens=
 the 
world is filled with boxes that don't support privacy for that thi=
ng. 
How do you determine if you have the privacy you
 need?

    Thanks,
    Paul

=
> The second issue regarding the precedence is the one which people may =
have opinions about.
>
> Ian
>
> ----- Original Mes=
sage -----
> From: "Paul Kyzivat"<pkyzivat@cisco.com>
&=
gt; To: dispatch@ietf.org
> Sent: Tuesday, 8 March, 2011 2:12:39 P=
M
> Subject: Re: [dispatch] New Draft -    draft-elz-d=
ispatch-req-sip-3rdpartyprivacy-00
>
> Ian,
>
> I'd=
 like to thank you for submitting a clear requirements draft that is
>=
; unburdened by mechanism proposals.
>
> IMO the hard part of t=
his issue is sorting out whether there is a need
> for a "generalized=
" mechanism dealing with arbitrary 3rd party
> identities, or just fo=
r
 a mechanism to deal with a specific enumerated
> set of 3rd party id=
entities.
>
> While its easy to wish for a generalized mechanis=
m, I think it will be
> very difficult to adequately specify. If cons=
trained to a specific set
> than it is a solvable problem.
>>     Thanks,
>     Paul
>> On 3/8/2011 3:15 AM, Ian Elz wrote:
>> All,
>>
&=
gt;> I have submitted a new draft relating to third party privacy in SIP=

>> messages.
>>
>> =
http://www.ietf.org/id/draft-elz-dispatch-req-sip-3rdpartyprivacy-00.txt
>> (Link that I forgot in the previous posting.)
>>
=
>> There are two specific issues included, how to specify third party=

>> privacy in a SIP message; the precedence of privacy
 settings between
>> Privacy header values and Privacy URI paramet=
ers.
>>
>> The things to consider are:
>>
>=
;> Should this work be pursued and if so in which forum.
>>
=
>> A suitable solution to meet the requirements.
>>
>&=
gt; How any changes should be implemented: a new draft which modifies
&g=
t;> RFC3323 or an update to RFC3323 which includes the additional work.<=
br>>>
>> All comments welcome.
>>
>> Sorry=
 I forgot to write the Abstract in version -00. I will include this
>=
> in the next version. ( I leave abstracts to last and overlooked this>> before submission.)
>>
>> Ian Elz
>><=
br>>>
>>
>> _______________________________________=
________
>> dispatch mailing list
>> dispatch@ietf.org
>> http=
s://www.ietf.org/mailman/listinfo/dispatch
>
>
>
&=
gt;
>
>
 

=0A=0A=0A=0A      
--0-1236837020-1299658788=:29308--

From pkyzivat@cisco.com  Wed Mar  9 06:01:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 29C4E3A69E0 for ; Wed,  9 Mar 2011 06:01:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.536
X-Spam-Level: 
X-Spam-Status: No, score=-110.536 tagged_above=-999 required=5 tests=[AWL=0.063, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pG35eYQ7jUFe for ; Wed,  9 Mar 2011 06:01:31 -0800 (PST)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 5BB303A69BE for ; Wed,  9 Mar 2011 06:01:31 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=300; q=dns/txt; s=iport; t=1299679368; x=1300888968; h=message-id:date:from:mime-version:to:cc:subject: references:in-reply-to:content-transfer-encoding; bh=LC65t1CA4CFHr7zjnSuizKwKvINDuXqNWn27ou0NoDQ=; b=Uv4kf1oKcQCPmO7xhP+FnSPKpM95bFi4H+aAEoMVcNYRk2nIqgmEE6Ij z1TkSSApeSdpjr4dA3QrsrMm3MstrI8XYknQq8NxxOVBcVVEPBBBNAhQR ewkmXbFYi17giG8gJBCzuSP5litx1gWdbCF+o/KNqJTUXMADvV2WIyIiI o=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Av0EAHIWd01AZnwN/2dsb2JhbACELaJAdKcCiwyRPYEng0h2BIUihxiDSA
X-IronPort-AV: E=Sophos;i="4.62,290,1297036800"; d="scan'208";a="223941236"
Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-2.cisco.com with ESMTP; 09 Mar 2011 14:02:47 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id p29E2lHf012725; Wed, 9 Mar 2011 14:02:47 GMT
Message-ID: <4D778887.90503@cisco.com>
Date: Wed, 09 Mar 2011 09:02:47 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: Ian Elz 
References: <120652.93381.qm@web29106.mail.ird.yahoo.com> <4D76D0FB.80005@cisco.com> <777439.29308.qm@web29101.mail.ird.yahoo.com>
In-Reply-To: <777439.29308.qm@web29101.mail.ird.yahoo.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: dispatch@ietf.org
Subject: Re: [dispatch] New Draft - draft-elz-dispatch-req-sip-3rdpartyprivacy-00
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 14:01:59 -0000

On 3/9/2011 3:19 AM, Ian Elz wrote:
> Paul,
...
> At present I am more interested in whether there is agreement with
> pursuing the work than in agreeing a solution.

I think that is the right focus. So lets keep it at that for now, and 
worry about the details later.

	Thanks,
	Paul

From tanakai@nttdocomo.co.jp  Wed Mar  9 06:53:49 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0B7973A69F2 for ; Wed,  9 Mar 2011 06:53:49 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.834
X-Spam-Level: *
X-Spam-Status: No, score=1.834 tagged_above=-999 required=5 tests=[AWL=2.925,  BAYES_00=-2.599, HELO_EQ_JP=1.244, HOST_EQ_JP=1.265, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id YOxrSJRkzRc9 for ; Wed,  9 Mar 2011 06:53:47 -0800 (PST)
Received: from zcsg-mailro12.is.nttdocomo.co.jp (dish-sg.nttdocomo.co.jp [202.19.227.74]) by core3.amsl.com (Postfix) with ESMTP id D3D833A6874 for ; Wed,  9 Mar 2011 06:53:46 -0800 (PST)
Received: from zcsg-mailmt12.is.nttdocomo.co.jp (zcsg-mailmt10.is.nttdocomo.co.jp [10.160.86.41]) by zcsg-mailro12.is.nttdocomo.co.jp (Postfix) with ESMTP id 1318D34005 for ; Wed,  9 Mar 2011 23:55:02 +0900 (JST)
Received: from zcsg-mailmi13.is.nttdocomo.co.jp (zcsg-mailmi10.is.nttdocomo.co.jp [10.160.86.49]) by zcsg-mailmt12.is.nttdocomo.co.jp (NTT DoCoMo Mail System) with SMTP id <0LHS00M5KPFQXR40@NTTDoCoMo.co.jp> for dispatch@ietf.org; Wed, 09 Mar 2011 23:55:02 +0900 (JST)
Received: from unknown (HELO zcsg-mailvs12.is.nttdocomo.co.jp) (10.160.86.48) by 0 with SMTP; Wed, 09 Mar 2011 23:55:01 +0900
Received: from zcsg-mailvs12.is.nttdocomo.co.jp (localhost [127.0.0.1]) by localhost.is.nttdocomo.co.jp (Postfix) with ESMTP id DA9B38003; Wed, 09 Mar 2011 23:55:01 +0900 (JST)
Received: from zcsg-mailsa12.is.nttdocomo.co.jp (zcsg-mailsa10.is.nttdocomo.co.jp [10.160.86.46]) by zcsg-mailvs12.is.nttdocomo.co.jp (Postfix) with ESMTP id CF7FD8002; Wed, 09 Mar 2011 23:55:01 +0900 (JST)
Received: from AKTANAKAI223 (atnakamurahi222.docomo.docomogr.net [10.18.212.17]) by zcsg-mailsa12.is.nttdocomo.co.jp (NTT DoCoMo Mail System) with ESMTPA id <0LHS002N8PFN1W30@NTTDoCoMo.co.jp>; Wed, 09 Mar 2011 23:55:01 +0900 (JST)
Date: Wed, 09 Mar 2011 23:54:59 +0900
From: Itsuma TANAKA 
To: dispatch@ietf.org
Message-id: 
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
X-Mailer: Microsoft Office Outlook 11
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7bit
Thread-index: AcveafX/UNE//jAPRxC3GPIUezdMVA==
X-DoCoMo: ZCSG
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 14:53:49 -0000

Dear All,

I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP meetings (SA2,
CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming Expert
Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  I serve
as a Vice Chair of GSMA IREG RILTE and Packet subgroups.

I would like to express my support of draft-montemurro-gsma-imei-urn and the
associated draft-allen-dispatch-imei-urn-as-instanceid.

The reasons are exactly same as what Ricky Kaura (Samsung) expressed.  A
various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and TS24.292), and
GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are using
these drafts.
 
These internet drafts are very important to support various key features in
3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would like ask
that this work is progressed as quickly as possible in IETF. 


Best regards,

Itsuma Tanaka
NTT DOCOMO



From fluffy@cisco.com  Wed Mar  9 08:06:21 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 479C03A6859 for ; Wed,  9 Mar 2011 08:06:21 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.559
X-Spam-Level: 
X-Spam-Status: No, score=-110.559 tagged_above=-999 required=5 tests=[AWL=0.040, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Aot598qVtibD for ; Wed,  9 Mar 2011 08:06:20 -0800 (PST)
Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87]) by core3.amsl.com (Postfix) with ESMTP id 320003A67D8 for ; Wed,  9 Mar 2011 08:06:20 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=2075; q=dns/txt; s=iport; t=1299686856; x=1300896456; h=from:content-transfer-encoding:subject:date:references: to:message-id:mime-version; bh=hYppJy9aAcY77KKr7m5aL/3RZhaPE+W87k6uWIfwkFE=; b=B8gv4V04ReaR7nTM9IAjgyv32Z5fMU9fwQLfbQT/t8jpWhKp/QmmuzMq DTVplI/PgD3JmVL9IqnT3oQX+P1Nlgv95XrFICCfwjCP6ThhZ0cglLCr0 N9xFpjzZBYVoZENpy+2sprLZk1v5Xg28zFvogZCx96gWlwQy4ikF5OIQW c=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAHI0d02rRN+K/2dsb2JhbACmbnSnCJxPgxiCTQSFIoV/gRmDSA
X-IronPort-AV: E=Sophos;i="4.62,291,1297036800"; d="scan'208";a="343051050"
Received: from sj-core-4.cisco.com ([171.68.223.138]) by sj-iport-5.cisco.com with ESMTP; 09 Mar 2011 16:07:36 +0000
Received: from [192.168.4.100] (rcdn-fluffy-8712.cisco.com [10.99.9.19]) by sj-core-4.cisco.com (8.13.8/8.14.3) with ESMTP id p29G7ZSI025212; Wed, 9 Mar 2011 16:07:36 GMT
From: Cullen Jennings 
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Date: Wed, 9 Mar 2011 09:10:10 -0700
References: <6EA67834-47AB-48CD-B6A7-178B4B45FBB8@w3.org>
To: rtc-web@alvestrand.no, DISPATCH list 
Message-Id: 
Mime-Version: 1.0 (Apple Message framework v1082)
X-Mailer: Apple Mail (2.1082)
Subject: [dispatch] Fwd: [new-work] Proposed W3C Charter: Web Real-Time Communications Working Group (until 2011-03-18)
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 16:06:21 -0000

FYI

Begin forwarded message:

> From: Ian Jacobs 
> Date: February 17, 2011 11:12:50 AM MST
> To: new-work@ietf.org
> Subject: [new-work] Proposed W3C Charter: Web Real-Time Communications =
Working Group (until 2011-03-18)
>=20
> Hello,
>=20
> Today W3C Advisory Committee Representatives received a Proposal to =
revise the Ubiquitous Web Applications Activity [0] (see the W3C Process =
Document description of Activity Proposals [1]). This proposal includes =
a draft charter for the Web Real-Time Communications Working Group:
>  http://www.w3.org/2010/12/webrtc-charter
>=20
> As part of ensuring that the community is aware of proposed work at =
W3C, this draft charter is public during the Advisory Committee review =
period.
>=20
> W3C invites public comments through 2011-03-18 on the proposed =
charter. Please send comments to public-new-work@w3.org, which has a =
public archive:
>  http://lists.w3.org/Archives/Public/public-new-work/
>=20
> Other than comments sent in formal responses by W3C Advisory Committee =
Representatives, W3C cannot guarantee a response to comments. If you =
work for a W3C Member [2], please coordinate your comments with your =
Advisory Committee Representative. For example, you may wish to make =
public comments via this list and have your Advisory Committee =
Representative refer to it from his or her formal review comments.
>=20
> If you should have any questions or need further information, please =
contact Fran=E7ois Daoust, Team Contact .
>=20
> Thank you,
>=20
> Ian Jacobs, Head of W3C Communications
>=20
> [0] http://www.w3.org/2007/uwa/
> [1]
> http://www.w3.org/2005/10/Process-20051014/activities#ActivityCreation
> [2] http://www.w3.org/Consortium/Member/List
> --
> Ian Jacobs (ij@w3.org)    http://www.w3.org/People/Jacobs/
> Tel:                                      +1 718 260 9447
>=20
> _______________________________________________
> new-work mailing list
> new-work@ietf.org
> https://www.ietf.org/mailman/listinfo/new-work


From vkg@bell-labs.com  Wed Mar  9 09:06:39 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B874F3A68AD for ; Wed,  9 Mar 2011 09:06:39 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lgwZUZzTrFO7 for ; Wed,  9 Mar 2011 09:06:38 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 90C6D3A688F for ; Wed,  9 Mar 2011 09:06:38 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p29H72d5016027 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 9 Mar 2011 11:07:02 -0600 (CST)
Received: from shoonya.ih.lucent.com (shoonya.ih.lucent.com [135.185.238.233]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p29H72EV020698; Wed, 9 Mar 2011 11:07:02 -0600 (CST)
Message-ID: <4D77B416.5070601@bell-labs.com>
Date: Wed, 09 Mar 2011 11:08:38 -0600
From: "Vijay K. Gurbani" 
Organization: Bell Laboratories, Alcatel-Lucent
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.13) Gecko/20101209 Fedora/3.1.7-0.35.b3pre.fc14 Thunderbird/3.1.7
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com> <4D76C924.7050900@cisco.com>
In-Reply-To: <4D76C924.7050900@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:06:39 -0000

On 03/08/2011 06:26 PM, Paul Kyzivat wrote:
> I guess I need more explanation of how that algorithm applies to
> this problem.

Paul: I am referring to the stock algorithm in RFC2782 for
ordering RRs using the "weight" component.  Better algorithms
can be substituted, of course, but as a default, the one in
RFC2782 suffices.

> Are you saying that the "weight" is the "window size", and that the
> server increases/decreases it whenever the incoming load is not
> keeping it at the optimal loading? Sounds interesting.

Yes, sort of.  But because traffic is flowing one way --- from
the SIP proxy (the load balancer) to the SIP server in the
cluster --- there is only one window size (that of the latter).

> But I know next to nothing about that algorithm or the statistical
> behavior of such algorithms.

Taking your example, let's say that there are two servers and
srv1 advertises a "weight" of 32,000 and srv2 advertises a
"weight" of 4,000.  As load increases, srv1 decreases its
"weight" by 1/2 each time period (i.e., at time T0, "weight" was
32,000; at T1 it advertises a "weight" of 16,000; at T2 8,000;
at T3 4,000; at T4 2,000).  Let's also assume that srv2 keeps
its "weight" constant at 4,000.

At T4, the "weight" of srv1 is less than that of srv2.  Up until T4,
srv1 will get about 80% of the traffic (assuming only these two
servers and a constant "weight" of 4,000 for srv2).  At T4 and beyond,
srv2 picks up more than 90% of the traffic, relieving srv1.

Convergence is predicated on the delta between T_n and T_(n+1).
If the delta is larger, it will take a bit more time to converge.

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From dworley@avaya.com  Wed Mar  9 09:09:05 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0E5713A68B1 for ; Wed,  9 Mar 2011 09:09:05 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.539
X-Spam-Level: 
X-Spam-Status: No, score=-102.539 tagged_above=-999 required=5 tests=[AWL=0.060, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zsjCEnlgD0ky for ; Wed,  9 Mar 2011 09:09:04 -0800 (PST)
Received: from co300216-co-outbound.net.avaya.com (co300216-co-outbound.net.avaya.com [198.152.13.100]) by core3.amsl.com (Postfix) with ESMTP id 14CDD3A68AD for ; Wed,  9 Mar 2011 09:09:04 -0800 (PST)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU3GmAcF/2dsb2JhbACmZXSkeQKZFoVhBIUcinA
X-IronPort-AV: E=Sophos;i="4.62,291,1297054800"; d="scan'208";a="268592474"
Received: from unknown (HELO co300216-co-erhwest.avaya.com) ([198.152.7.5]) by co300216-co-outbound.net.avaya.com with ESMTP; 09 Mar 2011 12:10:20 -0500
X-IronPort-AV: E=Sophos;i="4.62,291,1297054800"; d="scan'208";a="592882404"
Received: from unknown (HELO DC-US1HCEX3.global.avaya.com) ([135.11.52.22]) by co300216-co-erhwest-out.avaya.com with ESMTP; 09 Mar 2011 12:10:20 -0500
Received: from DC-US1MBEX4.global.avaya.com ([169.254.1.187]) by DC-US1HCEX3.global.avaya.com ([135.11.52.22]) with mapi; Wed, 9 Mar 2011 12:10:19 -0500
From: "Worley, Dale R (Dale)" 
To: Paul Kyzivat , "Vijay K. Gurbani" 
Date: Wed, 9 Mar 2011 12:10:18 -0500
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd8Jwuu595L4ZlROesYLB3rZg1awAiT73N
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,<4D76C924.7050900@cisco.com>
In-Reply-To: <4D76C924.7050900@cisco.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:09:05 -0000

This is a fascinating subject.  On the way home from work, I started to fra=
me a set of differential equations to model it.

The key, of course, is to have all of the servers reach full capacity *at t=
he same time* if the presented load increases steadily.  That is how we max=
imize goodput under such circumstances.  (If there is spare capacity within=
 the cluster of servers, it doesn't matter if the load on the servers is ba=
lanced or not, as goodput equals presented load.)

The crucial questions are "What is the algorithm for generating the number =
that the server presents to the load balancer?" and "What algorithm does th=
e load balancer use for assigning work, given the numbers presented by the =
servers?"

The draft (draft-bessis-dispatch-adaptive-load-balancing-00) seems to assum=
e that the assignment algorithm is much the same as RFC 3263, to statistica=
lly send Wi/(sum(Wj)) of the load to server i (with Wi being the number rep=
orted by server i).  The draft gives little or no hint of what the generati=
on algorithm will be.

In order for the mechanism to be standardized in a meaningful way, it must =
specify the generation algorithm sufficiently that a combination of servers=
 that use generation algorithms in conformance with the standard will work =
together correctly -- as I said above, as the entire cluster nears full cap=
acity, all servers much reach full capacity simultaneously.

t seems that we need to do a careful analysis of the behavior of the system=
 near the point where "unused capacity of server i" is near 0 for all serve=
rs i.  The first question is to categorize generation functions by their be=
havior near the point "unused capacity =3D 0", that is, does the reported "=
weight" approach a non-zero number, does it approach zero linearly, does it=
 approach zero quadratically, etc.?  I suspect that the system behaves well=
 only when all servers use functions of the same form, and with the same co=
nstant on the leading term.  Intuitively, the correct answer seems to be to=
 have the weight be "the unused capacity, measured in number of calls that =
can be handled".

This gets messier when we consider that the load is comes in quantized unit=
s (single calls) and has significant statistical fluctuations.  In particul=
ar, some functions that appear to work correctly in the "analytical" analys=
is outlined above my have bad behavior when presented with statistical fluc=
tuations.

Additionally, the feedback from the servers to the load balancer will not b=
e instantaneous, and the delays will introduce "control loop delays", which=
 if not handled correctly might cause the system to oscillate.

A different problem results from the fact that different calls may present =
decidedly different loads to a server, and that the server may have several=
 different types of resource.  How does the server sensibly report the natu=
re of its resources and the loading on each?  How does the load balancer es=
timate the resources required by an individual call?  (Note that this probl=
em is largely nonexistent in circuit-switched systems, as all calls present=
 exactly the same load.)  (This is the problem that has bedevilled attempts=
 to solve the problem in draft-partha-sip-overload-resource-availability.)

But I don't see any of the needed analysis in the draft, and I don't see an=
y indication in the draft of a realization that this analysis needs to be d=
one.  Until that is fixed, I predict that a solution will not be found, or =
worse, a solution that appears to work on paper but does not work in practi=
ce.

Dale

From Thierry.Bessis@alcatel-lucent.com  Wed Mar  9 09:25:15 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BA7AF3A68AD for ; Wed,  9 Mar 2011 09:25:14 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FfJNdQAmiqzR for ; Wed,  9 Mar 2011 09:24:59 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 22F693A65A6 for ; Wed,  9 Mar 2011 09:24:58 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p29HQESo019503 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 9 Mar 2011 11:26:14 -0600 (CST)
Received: from [135.244.3.108] (tbessis.lra.lucent.com [135.244.3.108]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p29HQAqk007246; Wed, 9 Mar 2011 11:26:10 -0600 (CST)
Message-ID: <4D77B831.7000404@alcatel-lucent.com>
Date: Wed, 09 Mar 2011 11:26:09 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Paul E. Jones" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com>
In-Reply-To: <004601cbddd2$15b11ab0$41135010$@packetizer.com>
Content-Type: multipart/alternative; boundary="------------020004030201040501060401"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:25:15 -0000

This is a multi-part message in MIME format.
--------------020004030201040501060401
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Thanks Paul for you detailed answer !

I've looked at your SCE proposal and your comments and my impression is 
that our two proposals are actually very different both from their 
intent perspective, and from their realization perspective.

The intent of our load balancing proposal is clearly continuous load 
balancing. Our suggested weight does try to figure any kind of remaining 
capacity in terms of sessions: it only indicates the instant (slightly 
smoothed) available resource, which integrates CPU, Memory and any other 
constrained resource. For the kind of usage we have in mind, (and for 
example), it does not matter if we have lot of sessions "left" when we 
have very little CPU to register those sessions: it is optimal to 
temporary offload this busy node rather than risking a CPU induced 
rejection or failure. When the node will have more CPU available, it 
will take more sessions, and may be even more than others, because it 
may have now more memory than others. If this node never have more CPU 
available, then its "session" capacity is actually unusable and the 
system correctly reflects that.

Also, the SOC was not intended for "dynamic" load distribution: it only 
provides the feedback when some shedding is expected (So already one or 
multiple nodes are about to be overloaded). Also, integrating the 
information from both SOC and SCE to take the distribution decision 
seems rather complicated to me. In our proposal, the distribution 
algorithm is simple, never changes, and is independent from SOC. We 
selected a node with weights always the same way, and we apply the SOC 
rejection when applicable to this node.

Cordially,

Thierry

On 08-Mar-2011 14:47, Paul E. Jones wrote:
>
> Thierry,
>
> Please see my comments below in green:
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Tuesday, March 08, 2011 12:35 PM
> *To:* Paul E. Jones
> *Cc:* 'Vijay K. Gurbani'; dispatch@ietf.org
> *Subject:* Re: [dispatch] SIP load balancing
>
> Paul,
>
> Please let me address some of your comments.
> I don't know what the SCE is, so please pardon me if I misunderstand 
> something.
>
> On 07-Mar-2011 23:05, Paul E. Jones wrote:
>
> Vijay,
>   
> The Session Capacity Estimate (SCE) idea is actually quite similar to the
> "adaptive load balancing" draft, from what I can see.  SCE was proposed as a
> solution to the "overload" problem, the approach being one where we try to
> avoid overload by selecting a device that reports a higher capacity value.
> Thus, one could view it as "overload control through load balancing".  We
> didn't explicitly call it load balancing, though, because we were not trying
> to balance load (i.e., place the same load on all servers).  Rather, we just
> wanted to direct traffic to devices that we knew had available capacity.
>
> Agreed, the goal seems to be the same: avoiding or at least delaying 
> overload.
> And your other comment too: our proposal does try to balance the load 
> on all the servers. The rational is that it is very easy for each 
> server to know where it is in terms of resource usage (for instance: 
> I'm 60 percent resource loaded at this moment), and this is just the 
> information you need to balance the traffic. Now it seems (but it is 
> not clear, see below)  that the SCE "available capacity" is actually a 
> real time feedback, so the information would be very similar actually. 
> (What's the difference ? )
>
> PEJ: I think it is very similar.  SCE provides real-time feedback of 
> the available resource capacity of a peer device.  The intent is to 
> convey a single integer value that represents an estimate of the 
> number of sessions the peer device can handle.  Note, this is not 
> "message per second" or other, but the number of sessions.  The focus 
> on SCE is on devices like "call agents", "session border controllers", 
> stateful SIP proxies, etc.
>
>   
> In practice, the way SCE would be employed is the following:
> * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
>     peer devices
>   * The selector would direct traffic toward the peer devices in a
>     round-robin fashion, taking into consideration the SCE value so
>     as to prevent exhausting resources on any given peer device
>
> I see a contradiction here: if it is round-robin, it does not take 
> into account the SCE value.
> Is that something more like a weighted round-robin algorithm ?
>
> PEJ: As I mention below, the algorithm used to select a peer device is 
> not a part of the SCE draft.  One can use any mechanism.  The SCE 
> draft merely advertises a session capacity estimate.  We've found 
> round-robin to be a simple approach that works, as it distributes load 
> relatively uniformly, but stops directing traffic toward devices that 
> report an SCE value that is below a specified threshold. Other input 
> could be considered, but we deliberately left the algorithm outside 
> the scope of SCE.
>
>   * Once the peer's SCE value dropped below a given "low-water mark",
>     the peer device would be less favored and not given traffic until
>     the SCE value increases above the low-water mark (OPTIONS "ping"
>     messages may be used to determine the SCE value in the event that
>     there is no other traffic)
>
> Why would we wait for a lower mark before to influence the 
> distribution ? Also is that less favored, or given no traffic ?
>
> PEJ: With our use of SCE, we're not trying to "balance" the load, per 
> se.  One certainly could do that, but in order to balance the load on 
> peer devices when they are out of balance, one must direct multiple 
> new sessions toward the same device.  That is not always desirable 
> since new sessions usually consume more resources temporarily than 
> established sessions.  So, we try to burden a peer device by throwing 
> a number of new sessions toward the peer by distributing them in a 
> round-robin fashion.  Again, though, the approach taken is outside the 
> scope of the SCE draft.
>
>
> The later has stability issues.  And why do we need ping Options when 
> there is no traffic ?
>  If there is no traffic (or very little), should we care about optimal 
> distribution from this source ?
>
> PEJ: The assumption is that a peer device might also have several 
> peers.  So, while device A1 may not be directing traffic toward device 
> B, perhaps A2, A3, and A4 are.  We can use OPTIONS "ping" messages to 
> get current SCE information.
>
>
>
>   * The current SOC overload control mechanism would be used to
>     to prevent message overloading, as SCE only focuses on resource
>     overloading (e.g., memory, CPU, disk space, DSPs)
>
> I don't see this difference here (message overload versus resource 
> overload ? ) . To me, SOC focuses on source shedding, while SCE seems 
> to focus on improved load distribution, just like our proposal.
>
> PEJ: A peer device might be able to handle 18,000 new sessions, but 
> perhaps it can only handle them at a rate of 100 per second.  We would 
> use SCE to determine that a peer can handle new sessions, while we 
> would use the methods SOC is defining to prevent message overload so 
> we do not send 200 INVITEs per second to the peer.  (Use of the SOC 
> mechanism might prove to work better than our current round-robin 
> approach, but we've not tested it.)
>
>   
> Note that the "round-robin" approach may or may not be preferred and is not
> required in our draft.  I've found this approach works pretty well, but one
> may prefer to use a particular weighting that is more complex.  I don't
> think local decisions in the box need to be subject to standardization: we
> just need to standardize the signaling to enable such decisions.
>
> I believe the signaling is not enough. The deep meaning of the 
> feedback parameter needs to be clarified too.
> For example, in our proposal, the feedback is simply the current 
> utilization rate of the node (very simple).
>
> PEJ: That sounds similar to SCE.  An SCE-capable device advertises to 
> its peer how many new sessions it can handle given the number of 
> sessions it is currently handling.
>
> Also the expectation needs to be clarified because it has a very 
> important impact on the recommended behavior in case of overload.
> The solution that we propose constantly tries to balance load in the 
> cluster of servers. Therefore, in case the SOC shedding kicks in, 
> there is NOT NEED to repeat the message elsewhere in the cluster 
> because it is balanced already (So no other server has more capacity). 
> This is a huge advantage for the network stability as message repeat 
> tends to spread the overload and collapse the cluster.
>
> In any case, it certainly looks like we're thinking along the same lines.
>
> Yes I agree.
>
> PEJ: If you haven't already, have a look at our draft:
>
> http://tools.ietf.org/html/draft-jones-sip-overload-sce-00
>
> It is now expired, but your draft suggests there is still life in this 
> approach.  We should look to harmonize them, perhaps.
>
> Paul
>
>
> Cordially,
>
> Thierry
>
>   
> Paul
>   
>
>     -----Original Message-----
>
>     From:dispatch-bounces@ietf.org    [mailto:dispatch-bounces@ietf.org] On
>
>     Behalf Of Vijay K. Gurbani
>
>     Sent: Monday, March 07, 2011 6:38 PM
>
>     To:dispatch@ietf.org  
>
>     Subject: Re: [dispatch] SIP load balancing
>
>       
>
>     Paul: Thank you for your input.  More inline.
>
>       
>
>     On 03/07/2011 05:14 PM, Paul Kyzivat wrote:
>
>         Vijay,
>
>         IIUC, the algorithm you have proposed is similar to the one Paul Jones
>
>         has proposed - with the main difference being the numeric range of the
>
>         weight. Do I have that right?
>
>       
>
>     I believe that Paul J.'s algorithm is more suited towards overload
>
>     control, not load balancing (and I will await Paul J.'s correction to my
>
>     assertion).
>
>       
>
>     The intent in draft-bessis-dispatch-adaptive-load-balancing is to
>
>     *always* report the current load to give the upstream SIP proxy an
>
>     indication of the health of the SIP server in the cluster.  If one waits
>
>     until overload is imminent, then it may be too late since a sudden and
>
>     unanticipated burst of messages will put the SIP server in overload
>
>     rather quickly.
>
>       
>
>         It isn't clear to me from your draft if you are specifying the
>
>         algorithm by which the relative weights are used to determine the
>
>         probability of a request going to one server or another.
>
>       
>
>     Yes, same way as RFC2782 does it.
>
>       
>
>         ISTM that it *does* matter. And it especially matters if there is an
>
>         implication that the upstream node's algorithm may depend on some
>
>         assumption or knowledge about the downstream nodes.
>
>       
>
>     No knowledge, just a reported number that the upstream SIP proxy can
>
>     treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
>
>     starts off with the SRV indicators in RFC2782, without any regard to
>
>     additional knowledge or assumption about downstream nodes.
>
>     dispatch-adaptive-load-balancing is trying to use the same model, except
>
>     render the weight to be sensitive to the current traffic in the cluster
>
>     in real-time.
>
>       
>
>         And of course this method of reporting weights isn't sufficient to
>
>         drive a balancing algorithm that is sensitive to some requests
>
>         requiring more or different resources than others. (That seems to be a
>
>         real problem, though I don't know if we have a real solution for
>
>         it.)
>
>       
>
>     Yes, that is a real problem, as is the problem of a SIP server in a
>
>     cluster that has a low weight and is thus not contacted with enough
>
>     frequency to merit more participation to the goodput.
>
>       
>
>     It is well known in SIP that some methods are more onerous to process
>
>     than others --- from a proxying point of view, processing an INVITE with
>
>     forking and response aggregation is far more detrimental to resource
>
>     usage than is processing a BYE.  Some existing literature techniques
>
>     thus add a coefficient to the weight of sending an INVITE versus a BYE.
>
>     At this time, the process outlined in dispatch-adaptive-load-balancing
>
>     treats all requests the same, but we can definitely look at ways to
>
>     change this behaviour.
>
>       
>
>     Thanks,
>
>       
>
>     - vijay
>
>     --
>
>     Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
>
>     1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
>
>     Email:vkg@{bell-labs.com,acm.org}    /vijay.gurbani@alcatel-lucent.com  
>
>     Web:http://ect.bell-labs.com/who/vkg/
>
>     _______________________________________________
>
>     dispatch mailing list
>
>     dispatch@ietf.org  
>
>     https://www.ietf.org/mailman/listinfo/dispatch
>
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------020004030201040501060401
Content-Type: multipart/related;
 boundary="------------050008040701040305050904"


--------------050008040701040305050904
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Thanks Paul for you detailed answer ! 

    

    I've looked at your SCE proposal and your comments and my impression
    is that our two proposals are actually very different both from
    their intent perspective, and from their realization perspective. 

    

    The intent of our load balancing proposal is clearly continuous load
    balancing. Our suggested weight does try to figure any kind of
    remaining capacity in terms of sessions: it only indicates the
    instant (slightly smoothed) available resource, which integrates
    CPU, Memory and any other constrained resource. For the kind of
    usage we have in mind, (and for example), it does not matter if we
    have lot of sessions "left" when we have very little CPU to register
    those sessions: it is optimal to temporary offload this busy node
    rather than risking a CPU induced rejection or failure. When the
    node will have more CPU available, it will take more sessions, and
    may be even more than others, because it may have now more memory
    than others. If this node never have more CPU available, then its
    "session" capacity is actually unusable and the system correctly
    reflects that. 

    

    Also, the SOC was not intended for "dynamic" load distribution: it
    only provides the feedback when some shedding is expected (So
    already one or multiple nodes are about to be overloaded). Also,
    integrating the information from both SOC and SCE to take the
    distribution decision seems rather complicated to me. In our
    proposal, the distribution algorithm is simple, never changes, and
    is independent from SOC. We selected a node with weights always the
    same way, and we apply the SOC rejection when applicable to this
    node.

    

    Cordially, 

    

    Thierry

    

    On 08-Mar-2011 14:47, Paul E. Jones wrote:
    

      
      
      
      Thierry's signature
      
      

        
Thierry,

        
 

        
Please see my comments below in green:

        
 

        
 

        

          

            

              
From: Thierry Bessis
                  [mailto:Thierry.Bessis@alcatel-lucent.com] 

                  Sent: Tuesday, March 08, 2011 12:35 PM

                  To: Paul E. Jones

                  Cc: 'Vijay K. Gurbani'; dispatch@ietf.org

                  Subject: Re: [dispatch] SIP load balancing

            

          

          
 

          
Paul, 

            

            Please let me address some of your comments. 

            I don't know what the SCE is, so please pardon me if I
            misunderstand something. 

            

            On 07-Mar-2011 23:05, Paul E. Jones wrote: 

          
Vijay,

          
 

          
The Session Capacity Estimate (SCE) idea is actually quite similar to the

          
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a

          
solution to the "overload" problem, the approach being one where we try to

          
avoid overload by selecting a device that reports a higher capacity value.

          
Thus, one could view it as "overload control through load balancing".  We

          
didn't explicitly call it load balancing, though, because we were not trying

          
to balance load (i.e., place the same load on all servers).  Rather, we just

          
wanted to direct traffic to devices that we knew had available capacity.

          
Agreed, the goal seems to be the same:
            avoiding or at least delaying overload. 

            And your other comment too: our proposal does try to balance
            the load on all the servers. The rational is that it is very
            easy for each server to know where it is in terms of
            resource usage (for instance: I'm 60 percent resource loaded
            at this moment), and this is just the information you need
            to balance the traffic. Now it seems (but it is not clear,
            see below)  that the SCE "available capacity" is actually a
            real time feedback, so the information would be very similar
            actually. (What's the difference ? )

          
 

          
PEJ: I think it is very similar. 
              SCE provides real-time feedback of the available resource
              capacity of a peer device.  The intent is to convey a
              single integer value that represents an estimate of the
              number of sessions the peer device can handle.  Note, this
              is not “message per second” or other, but the number of
              sessions.  The focus on SCE is on devices like “call
              agents”, “session border controllers”, stateful SIP
              proxies, etc. 

            

            

          
 

          
In practice, the way SCE would be employed is the following:

          
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from

          
   peer devices

          
 * The selector would direct traffic toward the peer devices in a

          
   round-robin fashion, taking into consideration the SCE value so

          
   as to prevent exhausting resources on any given peer device

          
I see a contradiction here: if it is
            round-robin, it does not take into account the SCE value. 

            Is that something more like a weighted round-robin algorithm
            ? 

          
 

          
PEJ: As I mention below, the
              algorithm used to select a peer device is not a part of
              the SCE draft.  One can use any mechanism.  The SCE draft
              merely advertises a session capacity estimate.  We’ve
              found round-robin to be a simple approach that works, as
              it distributes load relatively uniformly, but stops
              directing traffic toward devices that report an SCE value
              that is below a specified threshold. Other input could be
              considered, but we deliberately left the algorithm outside
              the scope of SCE.

            

            

          
 * Once the peer's SCE value dropped below a given "low-water mark",

          
   the peer device would be less favored and not given traffic until

          
   the SCE value increases above the low-water mark (OPTIONS "ping"

          
   messages may be used to determine the SCE value in the event that

          
   there is no other traffic)

          
Why would we wait for a lower mark before
            to influence the distribution ? Also is that less favored,
            or given no traffic ? 

          
 

          
PEJ: With our use of SCE, we’re
              not trying to “balance” the load, per se.  One certainly
              could do that, but in order to balance the load on peer
              devices when they are out of balance, one must direct
              multiple new sessions toward the same device.  That is not
              always desirable since new sessions usually consume more
              resources temporarily than established sessions.  So, we
              try to burden a peer device by throwing a number of new
              sessions toward the peer by distributing them in a
              round-robin fashion.  Again, though, the approach taken is
              outside the scope of the SCE draft.

          
 

          


            The later has stability issues.  And why do we need ping
            Options when there is no traffic ?

             If there is no traffic (or very little), should we care
            about optimal distribution from this source ?

          
 

          
PEJ: The assumption is that a
              peer device might also have several peers.  So, while
              device A1 may not be directing traffic toward device B,
              perhaps A2, A3, and A4 are.  We can use OPTIONS “ping”
              messages to get current SCE information.

          
 

            

            

          
 * The current SOC overload control mechanism would be used to

          
   to prevent message overloading, as SCE only focuses on resource

          
   overloading (e.g., memory, CPU, disk space, DSPs)

          
I don't see this difference here (message
            overload versus resource overload ? ) . To me, SOC focuses
            on source shedding, while SCE seems to focus on improved
            load distribution, just like our proposal.

          
 

          
PEJ: A peer device might be able
              to handle 18,000 new sessions, but perhaps it can only
              handle them at a rate of 100 per second.  We would use SCE
              to determine that a peer can handle new sessions, while we
              would use the methods SOC is defining to prevent message
              overload so we do not send 200 INVITEs per second to the
              peer.  (Use of the SOC mechanism might prove to work
              better than our current round-robin approach, but we’ve
              not tested it.) 

            

            

          
 

          
Note that the "round-robin" approach may or may not be preferred and is not

          
required in our draft.  I've found this approach works pretty well, but one

          
may prefer to use a particular weighting that is more complex.  I don't

          
think local decisions in the box need to be subject to standardization: we

          
just need to standardize the signaling to enable such decisions.

          
I believe the signaling is not enough.
            The deep meaning of the feedback parameter needs to be
            clarified too. 

            For example, in our proposal, the feedback is simply the
            current utilization rate of the node (very simple).

          
 

          
PEJ: That sounds similar to SCE. 
              An SCE-capable device advertises to its peer how many new
              sessions it can handle given the number of sessions it is
              currently handling. 

            

            Also the expectation needs to be clarified because it has a
            very important impact on the recommended behavior in case of
            overload. 

            The solution that we propose constantly tries to balance
            load in the cluster of servers. Therefore, in case the SOC
            shedding kicks in, there is NOT NEED to repeat the message
            elsewhere in the cluster because it is balanced already (So
            no other server has more capacity). This is a huge advantage
            for the network stability as message repeat tends to spread
            the overload and collapse the cluster. 

            

            

          
In any case, it certainly looks like we're thinking along the same lines.

          
Yes I agree. 

            

            

          
PEJ: If you haven’t already, have
              a look at our draft:

          
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

          
 

          
It is now expired, but your draft
              suggests there is still life in this approach.  We should
              look to harmonize them, perhaps.

          
 

          
Paul

          
 

          


            Cordially, 

            

            Thierry

            

            

          
 

          
Paul

          
 

          

            
-----Original Message-----

            
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On

            
Behalf Of Vijay K. Gurbani

            
Sent: Monday, March 07, 2011 6:38 PM

            
To: dispatch@ietf.org

            
Subject: Re: [dispatch] SIP load balancing

            
 

            
Paul: Thank you for your input.  More inline.

            
 

            
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

            

              
Vijay,

              
IIUC, the algorithm you have proposed is similar to the one Paul Jones

              
has proposed - with the main difference being the numeric range of the

              
weight. Do I have that right?

            

            
 

            
I believe that Paul J.'s algorithm is more suited towards overload

            
control, not load balancing (and I will await Paul J.'s correction to my

            
assertion).

            
 

            
The intent in draft-bessis-dispatch-adaptive-load-balancing is to

            
*always* report the current load to give the upstream SIP proxy an

            
indication of the health of the SIP server in the cluster.  If one waits

            
until overload is imminent, then it may be too late since a sudden and

            
unanticipated burst of messages will put the SIP server in overload

            
rather quickly.

            
 

            

              
It isn't clear to me from your draft if you are specifying the

              
algorithm by which the relative weights are used to determine the

              
probability of a request going to one server or another.

            

            
 

            
Yes, same way as RFC2782 does it.

            
 

            

              
ISTM that it *does* matter. And it especially matters if there is an

              
implication that the upstream node's algorithm may depend on some

              
assumption or knowledge about the downstream nodes.

            

            
 

            
No knowledge, just a reported number that the upstream SIP proxy can

            
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution

            
starts off with the SRV indicators in RFC2782, without any regard to

            
additional knowledge or assumption about downstream nodes.

            
dispatch-adaptive-load-balancing is trying to use the same model, except

            
render the weight to be sensitive to the current traffic in the cluster

            
in real-time.

            
 

            

              
And of course this method of reporting weights isn't sufficient to

              
drive a balancing algorithm that is sensitive to some requests

              
requiring more or different resources than others. (That seems to be a

              
real problem, though I don't know if we have a real solution for

              
it.)

            

            
 

            
Yes, that is a real problem, as is the problem of a SIP server in a

            
cluster that has a low weight and is thus not contacted with enough

            
frequency to merit more participation to the goodput.

            
 

            
It is well known in SIP that some methods are more onerous to process

            
than others --- from a proxying point of view, processing an INVITE with

            
forking and response aggregation is far more detrimental to resource

            
usage than is processing a BYE.  Some existing literature techniques

            
thus add a coefficient to the weight of sending an INVITE versus a BYE.

            
At this time, the process outlined in dispatch-adaptive-load-balancing

            
treats all requests the same, but we can definitely look at ways to

            
change this behaviour.

            
 

            
Thanks,

            
 

            
- vijay

            
--

            
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent

            
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)

            
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com

            
Web:   http://ect.bell-labs.com/who/vkg/

            
_______________________________________________

            
dispatch mailing list

            
dispatch@ietf.org

            
https://www.ietf.org/mailman/listinfo/dispatch

          

          
 

          
_______________________________________________

          
dispatch mailing list

          
dispatch@ietf.org

          
https://www.ietf.org/mailman/listinfo/dispatch

          
 

          

            
-- 

              -- 

                  Cordially, 

                  Thierry Bessis 

            
  IMS Solution:
                Architecture & Product Management - ALTA Member

                  Organization: ALU > Carrier group > Wireline BU
                >  IMS Sol Arch, SE, PdM

              Alcatel-Lucent

                1960 Lucent
                Lane - Room 6L-517

                  Naperville, IL  60563  U S

                  Tel/Fax: +1 630 979 7989

                  Corporate IM: tbessis    - Yahoo IM:    
                thierry_bessis

              

              My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

              

              Conference information:

                2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 -
                Access Code: 9797989

                others countries see: http://all.alcatel-lucent.com/audio

              

              Upcoming planned Business Trip: 12
                  April - 14 April 2011

              Upcoming planned Vacation: None
              

          

        

      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------050008040701040305050904
Content-Type: image/jpeg
Content-Transfer-Encoding: base64
Content-ID: 
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--------------050008040701040305050904
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------050008040701040305050904--

--------------020004030201040501060401--

From hmmr@cisco.com  Wed Mar  9 09:35:31 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1C2C03A68F1 for ; Wed,  9 Mar 2011 09:35:31 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.599
X-Spam-Level: 
X-Spam-Status: No, score=-10.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uww2lhNK-VrX for ; Wed,  9 Mar 2011 09:35:29 -0800 (PST)
Received: from sj-iport-5.cisco.com (sj-iport-5.cisco.com [171.68.10.87]) by core3.amsl.com (Postfix) with ESMTP id BB4E13A68EE for ; Wed,  9 Mar 2011 09:35:29 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=hmmr@cisco.com; l=4143; q=dns/txt; s=iport; t=1299692206; x=1300901806; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to:cc; bh=NiCbxgo0Phtd169ocj0zHaqiIXIrZC93f9TFqlzWGzU=; b=EstLwNeLEtUhrfBYTQ65D0CS8WLfBQi9NV6wx/PVH+RXzDSDy3cAH0CJ RNOANZuYfAWAYczs1x9S6iee4mnTGBdddCfjTk+6Hb0RPXm6anKilyUzL XsFgxxZ92Sy+q2v/6Hxcb80BYKoHnGB3m8zF874Uvzrwja39Dwfi0MTEK I=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AkgBAIpJd02tJXHA/2dsb2JhbACYQo4sdGimHZN9iFSFZQSFIopm
X-IronPort-AV: E=Sophos;i="4.62,291,1297036800"; d="scan'208";a="343109437"
Received: from rcdn-core2-5.cisco.com ([173.37.113.192]) by sj-iport-5.cisco.com with ESMTP; 09 Mar 2011 17:36:44 +0000
Received: from xbh-rcd-301.cisco.com (xbh-rcd-301.cisco.com [72.163.63.8]) by rcdn-core2-5.cisco.com (8.14.3/8.14.3) with ESMTP id p29Hai47007260;  Wed, 9 Mar 2011 17:36:44 GMT
Received: from xmb-rcd-111.cisco.com ([72.163.62.153]) by xbh-rcd-301.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Wed, 9 Mar 2011 11:36:44 -0600
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Date: Wed, 9 Mar 2011 11:36:43 -0600
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd8Jwuu595L4ZlROesYLB3rZg1awAiT73NAAGnwWA=
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com><4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com><4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com> 
From: "Mike Hammer (hmmr)" 
To: "Worley, Dale R (Dale)" , "Paul Kyzivat (pkyzivat)" , "Vijay K. Gurbani" 
X-OriginalArrivalTime: 09 Mar 2011 17:36:44.0536 (UTC) FILETIME=[8EEE6780:01CBDE80]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:35:31 -0000

Agree with this.  If loads are not uniform, but bursty, then the
algorithm for the former may fail the latter.

Mike


-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Worley, Dale R (Dale)
Sent: Wednesday, March 09, 2011 12:10 PM
To: Paul Kyzivat (pkyzivat); Vijay K. Gurbani
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

This is a fascinating subject.  On the way home from work, I started to
frame a set of differential equations to model it.

The key, of course, is to have all of the servers reach full capacity
*at the same time* if the presented load increases steadily.  That is
how we maximize goodput under such circumstances.  (If there is spare
capacity within the cluster of servers, it doesn't matter if the load on
the servers is balanced or not, as goodput equals presented load.)

The crucial questions are "What is the algorithm for generating the
number that the server presents to the load balancer?" and "What
algorithm does the load balancer use for assigning work, given the
numbers presented by the servers?"

The draft (draft-bessis-dispatch-adaptive-load-balancing-00) seems to
assume that the assignment algorithm is much the same as RFC 3263, to
statistically send Wi/(sum(Wj)) of the load to server i (with Wi being
the number reported by server i).  The draft gives little or no hint of
what the generation algorithm will be.

In order for the mechanism to be standardized in a meaningful way, it
must specify the generation algorithm sufficiently that a combination of
servers that use generation algorithms in conformance with the standard
will work together correctly -- as I said above, as the entire cluster
nears full capacity, all servers much reach full capacity
simultaneously.

t seems that we need to do a careful analysis of the behavior of the
system near the point where "unused capacity of server i" is near 0 for
all servers i.  The first question is to categorize generation functions
by their behavior near the point "unused capacity =3D 0", that is, does
the reported "weight" approach a non-zero number, does it approach zero
linearly, does it approach zero quadratically, etc.?  I suspect that the
system behaves well only when all servers use functions of the same
form, and with the same constant on the leading term.  Intuitively, the
correct answer seems to be to have the weight be "the unused capacity,
measured in number of calls that can be handled".

This gets messier when we consider that the load is comes in quantized
units (single calls) and has significant statistical fluctuations.  In
particular, some functions that appear to work correctly in the
"analytical" analysis outlined above my have bad behavior when presented
with statistical fluctuations.

Additionally, the feedback from the servers to the load balancer will
not be instantaneous, and the delays will introduce "control loop
delays", which if not handled correctly might cause the system to
oscillate.

A different problem results from the fact that different calls may
present decidedly different loads to a server, and that the server may
have several different types of resource.  How does the server sensibly
report the nature of its resources and the loading on each?  How does
the load balancer estimate the resources required by an individual call?
(Note that this problem is largely nonexistent in circuit-switched
systems, as all calls present exactly the same load.)  (This is the
problem that has bedevilled attempts to solve the problem in
draft-partha-sip-overload-resource-availability.)

But I don't see any of the needed analysis in the draft, and I don't see
any indication in the draft of a realization that this analysis needs to
be done.  Until that is fixed, I predict that a solution will not be
found, or worse, a solution that appears to work on paper but does not
work in practice.

Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

From pkyzivat@cisco.com  Wed Mar  9 09:42:13 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D91FC3A68F7 for ; Wed,  9 Mar 2011 09:42:12 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.537
X-Spam-Level: 
X-Spam-Status: No, score=-110.537 tagged_above=-999 required=5 tests=[AWL=0.062, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id niITLlwqytZB for ; Wed,  9 Mar 2011 09:42:10 -0800 (PST)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id C49AE3A6765 for ; Wed,  9 Mar 2011 09:42:09 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=1196; q=dns/txt; s=iport; t=1299692606; x=1300902206; h=message-id:date:from:mime-version:to:cc:subject: references:in-reply-to:content-transfer-encoding; bh=J9dQln0PaEV9KFNQr6ZtS1rjNodyXb3Hgo23Cp9ubQU=; b=OhcY2UYduul30+3+U/jw26Q3pN/x5jn5+odLlQEGcy0K280i8WF5XG3w /PtPMQrDaoO+RM6Skc0yfV9zwiFU4kcOkpS3KeaYThntReB9U5LHmpBZV eVc3ME60a1AhCqP1YZ/tKAEBpNQg60W7dUi12UFYgWAnOznaGUdffMCnM 8=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAC9Ld02tJV2a/2dsb2JhbACmbnSnCpxUhWUEhSKHGINI
X-IronPort-AV: E=Sophos;i="4.62,291,1297036800"; d="scan'208";a="224003647"
Received: from rcdn-core-3.cisco.com ([173.37.93.154]) by rtp-iport-2.cisco.com with ESMTP; 09 Mar 2011 17:43:23 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rcdn-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p29HhNwR003181;  Wed, 9 Mar 2011 17:43:23 GMT
Message-ID: <4D77BC3A.9090103@cisco.com>
Date: Wed, 09 Mar 2011 12:43:22 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: "Vijay K. Gurbani" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com> <4D76C924.7050900@cisco.com> <4D77B416.5070601@bell-labs.com>
In-Reply-To: <4D77B416.5070601@bell-labs.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:42:13 -0000

On 3/9/2011 12:08 PM, Vijay K. Gurbani wrote:

> Taking your example, let's say that there are two servers and
> srv1 advertises a "weight" of 32,000 and srv2 advertises a
> "weight" of 4,000. As load increases, srv1 decreases its
> "weight" by 1/2 each time period (i.e., at time T0, "weight" was
> 32,000; at T1 it advertises a "weight" of 16,000; at T2 8,000;
> at T3 4,000; at T4 2,000). Let's also assume that srv2 keeps
> its "weight" constant at 4,000.
>
> At T4, the "weight" of srv1 is less than that of srv2. Up until T4,
> srv1 will get about 80% of the traffic (assuming only these two
> servers and a constant "weight" of 4,000 for srv2). At T4 and beyond,
> srv2 picks up more than 90% of the traffic, relieving srv1.
>
> Convergence is predicated on the delta between T_n and T_(n+1).
> If the delta is larger, it will take a bit more time to converge.

Hmm. Clearly I'm missing something, since I see no mention of capacity 
in the above. Certainly remaining capacity of the server needs to enter 
into the computation of the weight.

I suspect maybe this needs to be thoroughly discussed over beer.
And with Dale in the discussion.

	Thanks,
	Paul

From Thierry.Bessis@alcatel-lucent.com  Wed Mar  9 09:48:13 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 79DE33A680F for ; Wed,  9 Mar 2011 09:48:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mq0+ZQYs593O for ; Wed,  9 Mar 2011 09:48:11 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 8C7633A67EE for ; Wed,  9 Mar 2011 09:48:10 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p29HnOwp013934 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 9 Mar 2011 11:49:24 -0600 (CST)
Received: from [135.244.3.108] (tbessis.lra.lucent.com [135.244.3.108]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p29HnLnq028459; Wed, 9 Mar 2011 11:49:22 -0600 (CST)
Message-ID: <4D77BDA1.7020508@alcatel-lucent.com>
Date: Wed, 09 Mar 2011 11:49:21 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com>
In-Reply-To: <4D7698FD.4030501@cisco.com>
Content-Type: multipart/alternative; boundary="------------050308000000020106060506"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 17:48:13 -0000

This is a multi-part message in MIME format.
--------------050308000000020106060506
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi Paul,

Please see my comment.
>>     While I do not claim that the "weight" scheme is perfect, I think 
>> the
>> bounds for ambiguity are much less with this approach. All the
>> upstream SIP proxy needs is a "weight" indicator that it will
>> use in a RFC2782-type algorithm to proportionally send as much traffic
>> to a downstream SIP server as it has indicated (i.e., higher "weight"
>> values will get proportionally more traffic). This is much better
>> than round-robin, and furthermore, there is no requirement that each
>> downstream SIP server in the cluster set its "weight" in a manner
>> consistent with other servers, or even that the "weight" values of two
>> SIP servers be comparable.
>
> I don't understand that.
> Suppose we have two servers that are identical except that they have 
> different understandings of the weight assignment algorithm.
> One thinks it is a powerful server, and sets its initial (unloaded) 
> weight to 32k. The other thinks that it is not especially powerful as 
> servers go, and so sets its weight to 4k.
That is not how we implemented this weight: The weight is NOT an 
estimation of the capacity, it is a (smoothed) _measurement_ of the 
resource available.  Weight of 1 means that you have NO resource 
available (no CPU, or no Memory, or you are missing anything else 
required to process new sessions). Weight of  means that you 
have ALL your resources available (you are IDLE basically). This is a 
relative data: the same weight used by different servers does NOT 
represent the same capacity ! It represents the same relative load. This 
is consistent with the fact that we are trying to load all the server at 
the same relative load. In your example, the server does not need to 
know anything about its power or capacity, and even less about the 
capacity of the neighbors. It does not need to count anything, or 
understand why and how its resource vary with the number of instances: 
it just measures the resources.
>>
>> I am starting to believe that this property --- being sensitive to
>> the processing requirements of some requests versus others --- is
>> more pertinent to overload control than it is to load balancing.
>> The "weight" approach normalizes the processing load at a SIP server
>> in the cluster and reports a canonical value regardless of the specific
>> *type* of request the SIP server can handle. If the SIP server is
>> handling a lot of resource-heavy INVITE transactions, then it will
>> report a lower "weight" value.
>
> The thing is, a server may be able to handle a lot of some type of 
> load while being totally out of capacity for some other type of load.
>
> (E.g. a box that is both an SBC to a provider sip trunk and a gw to 
> some PSTN lines. It might be totally out of PSTN lines yet be able to 
> handle lots of calls that happen to be addressed to that they go to 
> the sip trunk.)
In this case, the mechanism should be applied in two places: one at the 
SBC ingress point, and another one after the PSTN breaking decision has 
been made: I don't see any other way ! :-)
>
> Also, shedding load is harder when the consumed resource is dialog 
> stateful.
I believe we should only shed at the source non Dialog related 
transactions, or new Dialogs.

Cordially,

Thierry
>
> I think this is a hard problem.
>
>     Thanks,
>     Paul
>
>> - vijay
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------050308000000020106060506
Content-Type: multipart/related;
 boundary="------------000403030202040903000009"


--------------000403030202040903000009
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Hi Paul, 

    

    Please see my comment. 

    

      
    While I do not claim that the "weight"
        scheme is perfect, I think the
        

        bounds for ambiguity are much less with this approach. All the
        

        upstream SIP proxy needs is a "weight" indicator that it will
        

        use in a RFC2782-type algorithm to proportionally send as much
        traffic
        

        to a downstream SIP server as it has indicated (i.e., higher
        "weight"
        

        values will get proportionally more traffic). This is much
        better
        

        than round-robin, and furthermore, there is no requirement that
        each
        

        downstream SIP server in the cluster set its "weight" in a
        manner
        

        consistent with other servers, or even that the "weight" values
        of two
        

        SIP servers be comparable.
        

      

      

      I don't understand that.
      

      Suppose we have two servers that are identical except that they
      have different understandings of the weight assignment algorithm.
      

      One thinks it is a powerful server, and sets its initial
      (unloaded) weight to 32k. The other thinks that it is not
      especially powerful as servers go, and so sets its weight to 4k.
      

    

    That is not how we implemented this weight: The weight is NOT an
    estimation of the capacity, it is a (smoothed) measurement
    of the resource available.  Weight of 1 means that you have NO
    resource available (no CPU, or no Memory, or you are missing
    anything else required to process new sessions). Weight of
    <maximum> means that you have ALL your resources available
    (you are IDLE basically). This is a relative data: the same weight
    used by different servers does NOT represent the same capacity ! It
    represents the same relative load. This is consistent with the fact
    that we are trying to load all the server at the same relative load.
    In your example, the server does not need to know anything about its
    power or capacity, and even less about the capacity of the
    neighbors. It does not need to count anything, or understand why and
    how its resource vary with the number of instances: it just measures
    the resources. 

    

      

        

        I am starting to believe that this property --- being sensitive
        to
        

        the processing requirements of some requests versus others ---
        is
        

        more pertinent to overload control than it is to load balancing.
        

        The "weight" approach normalizes the processing load at a SIP
        server
        

        in the cluster and reports a canonical value regardless of the
        specific
        

        *type* of request the SIP server can handle. If the SIP server
        is
        

        handling a lot of resource-heavy INVITE transactions, then it
        will
        

        report a lower "weight" value.
        

      

      

      The thing is, a server may be able to handle a lot of some type of
      load while being totally out of capacity for some other type of
      load.
      

      

      (E.g. a box that is both an SBC to a provider sip trunk and a gw
      to some PSTN lines. It might be totally out of PSTN lines yet be
      able to handle lots of calls that happen to be addressed to that
      they go to the sip trunk.)
      

    

    In this case, the mechanism should be applied in two places: one at
    the SBC ingress point, and another one after the PSTN breaking
    decision has been made: I don't see any other way !  :-) 

    

      

      Also, shedding load is harder when the consumed resource is dialog
      stateful.
      

    

    I believe we should only shed at the source non Dialog related
    transactions, or new Dialogs. 

    

    Cordially,

    

    Thierry

    

      

      I think this is a hard problem.
      

      

          Thanks,
      

          Paul
      

      

      
- vijay
        

      

      _______________________________________________
      

      dispatch mailing list
      

      dispatch@ietf.org
      

      https://www.ietf.org/mailman/listinfo/dispatch
      

    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------000403030202040903000009
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------000403030202040903000009--

--------------050308000000020106060506--

From dworley@avaya.com  Wed Mar  9 10:38:20 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B5C5D3A6933 for ; Wed,  9 Mar 2011 10:38:20 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.543
X-Spam-Level: 
X-Spam-Status: No, score=-102.543 tagged_above=-999 required=5 tests=[AWL=0.056, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QitOem76mp0H for ; Wed,  9 Mar 2011 10:38:19 -0800 (PST)
Received: from de307622-de-outbound.net.avaya.com (de307622-de-outbound.net.avaya.com [198.152.71.100]) by core3.amsl.com (Postfix) with ESMTP id A2A473A692A for ; Wed,  9 Mar 2011 10:38:19 -0800 (PST)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU3GmAcF/2dsb2JhbACmZXSkeQKZFoVhBIUcinA
X-IronPort-AV: E=Sophos;i="4.62,291,1297054800"; d="scan'208";a="235959921"
Received: from unknown (HELO co300216-co-erhwest.avaya.com) ([198.152.7.5]) by de307622-de-outbound.net.avaya.com with ESMTP; 09 Mar 2011 13:39:34 -0500
X-IronPort-AV: E=Sophos;i="4.62,291,1297054800"; d="scan'208";a="592922238"
Received: from dc-us1hcex2.us1.avaya.com (HELO DC-US1HCEX2.global.avaya.com) ([135.11.52.21]) by co300216-co-erhwest-out.avaya.com with ESMTP; 09 Mar 2011 13:39:30 -0500
Received: from DC-US1MBEX4.global.avaya.com ([169.254.1.187]) by DC-US1HCEX2.global.avaya.com ([::1]) with mapi; Wed, 9 Mar 2011 13:39:29 -0500
From: "Worley, Dale R (Dale)" 
To: "Mike Hammer (hmmr)" , "Paul Kyzivat (pkyzivat)" , "Vijay K. Gurbani" 
Date: Wed, 9 Mar 2011 13:39:29 -0500
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd8Jwuu595L4ZlROesYLB3rZg1awAiT73NAAGnwWAAAfzUCA==
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com> <4D7633F9.7000608@cisco.com><4D768BEE.2060400@bell-labs.com> <4D7698FD.4030501@cisco.com><4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com> , 
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 18:38:20 -0000

________________________________________
From: Mike Hammer (hmmr) [hmmr@cisco.com]

Agree with this.  If loads are not uniform, but bursty, then the
algorithm for the former may fail the latter.
_______________________________________________

(1) We *know* the loads are bursty.

(2) If the system has a uniform load that exceeds capacity, the administrat=
ors will install more capacity.

The only interesting case, in practice, is non-uniform loads near capacity.

(We can make a reasonable estimate of what is "near capacity".  If calls ar=
e assigned randomly in proportion to each server's total capacity (via RFC =
3263), each server will on average get its fair share of the traffic, with =
a random variation around that average of approximately the square-root of =
the average traffic it receives.  So if a server can handle N calls, as lon=
g as the average load it is given is less than N minus square-root(N), then=
 RFC 3263 will suffice.  The interesting problem is when the load presented=
 to the system requires loading the servers in excess of that level.)

We still haven't demonstrated that the "obvious" strategy does not work in =
practice:  If a server is overloaded, it rejects the INVITE and the load ba=
lancer forks the INVITE to another server.

Dale

From hmmr@cisco.com  Wed Mar  9 11:01:03 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 591563A68AF for ; Wed,  9 Mar 2011 11:01:03 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.599
X-Spam-Level: 
X-Spam-Status: No, score=-10.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BPmXSnRpbZMm for ; Wed,  9 Mar 2011 11:01:02 -0800 (PST)
Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id 14DAD3A6892 for ; Wed,  9 Mar 2011 11:01:02 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=hmmr@cisco.com; l=2121; q=dns/txt; s=iport; t=1299697339; x=1300906939; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to:cc; bh=QH8rh0rP579sIuxN34cX+5LVyFXNppSxhuSohtwLzY4=; b=HcoMkQA2hedBFOmstQCdneYN81QXsBcJx9xnghBSS/4wQDItDr1y/oY2 0hZeYQbayhS416vRnbcEoPwm7vwK8dA9EollQZaikMHFGWaCe+apbbfBp WHIXaPp3rln5wORhJAVaEtJa1SOjFWNqdk+WiMi6fFZCOxwfDz4g/Zi02 Q=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AkgBAMNcd02tJXHB/2dsb2JhbACYRI4sdGimS5N9iFSFZQSFIopm
X-IronPort-AV: E=Sophos;i="4.62,291,1297036800"; d="scan'208";a="223646397"
Received: from rcdn-core2-6.cisco.com ([173.37.113.193]) by rtp-iport-1.cisco.com with ESMTP; 09 Mar 2011 19:02:18 +0000
Received: from xbh-rcd-201.cisco.com (xbh-rcd-201.cisco.com [72.163.62.200]) by rcdn-core2-6.cisco.com (8.14.3/8.14.3) with ESMTP id p29J2Iav006085;  Wed, 9 Mar 2011 19:02:18 GMT
Received: from xmb-rcd-111.cisco.com ([72.163.62.153]) by xbh-rcd-201.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Wed, 9 Mar 2011 13:02:18 -0600
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Date: Wed, 9 Mar 2011 13:02:17 -0600
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd8Jwuu595L4ZlROesYLB3rZg1awAiT73NAAGnwWAAAfzUCAAAyIjQ
References: <4D755B9E.6030008@bell-labs.com><4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com><4D7633F9.7000608@cisco.com><4D768BEE.2060400@bell-labs.com><4D7698FD.4030501@cisco.com><4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com> ,  
From: "Mike Hammer (hmmr)" 
To: "Worley, Dale R (Dale)" , "Paul Kyzivat (pkyzivat)" , "Vijay K. Gurbani" 
X-OriginalArrivalTime: 09 Mar 2011 19:02:18.0579 (UTC) FILETIME=[830F3230:01CBDE8C]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 19:01:03 -0000

Dale,

I am thinking of loads in two dimensions, if that is on the table:
 - signaling from the arrival rate of SIP INVITEs as an example, can be
bursty while the resource to handle SIP might be not so bursty,
 - media in terms of whether this is only uniform demand of PSTN ports
or voice calls, versus narrowband voice request versus HD video
Telepresence request, where the resources to meet that mix of demand
might be deemed bursty.

This thread has touched on several of these areas, much the way 4 guys
are describing the elephant. :)

Mike


-----Original Message-----
From: Worley, Dale R (Dale) [mailto:dworley@avaya.com]=20
Sent: Wednesday, March 09, 2011 1:39 PM
To: Mike Hammer (hmmr); Paul Kyzivat (pkyzivat); Vijay K. Gurbani
Cc: dispatch@ietf.org
Subject: RE: [dispatch] SIP load balancing

________________________________________
From: Mike Hammer (hmmr) [hmmr@cisco.com]

Agree with this.  If loads are not uniform, but bursty, then the
algorithm for the former may fail the latter.
_______________________________________________

(1) We *know* the loads are bursty.

(2) If the system has a uniform load that exceeds capacity, the
administrators will install more capacity.

The only interesting case, in practice, is non-uniform loads near
capacity.

(We can make a reasonable estimate of what is "near capacity".  If calls
are assigned randomly in proportion to each server's total capacity (via
RFC 3263), each server will on average get its fair share of the
traffic, with a random variation around that average of approximately
the square-root of the average traffic it receives.  So if a server can
handle N calls, as long as the average load it is given is less than N
minus square-root(N), then RFC 3263 will suffice.  The interesting
problem is when the load presented to the system requires loading the
servers in excess of that level.)

We still haven't demonstrated that the "obvious" strategy does not work
in practice:  If a server is overloaded, it rejects the INVITE and the
load balancer forks the INVITE to another server.

Dale

From Thierry.Bessis@alcatel-lucent.com  Wed Mar  9 12:20:18 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AEFBA3A6AAA for ; Wed,  9 Mar 2011 12:20:18 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.097
X-Spam-Level: 
X-Spam-Status: No, score=-4.097 tagged_above=-999 required=5 tests=[AWL=-2.500, BAYES_00=-2.599, GB_SUMOF=5, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CSkytLvSgcPC for ; Wed,  9 Mar 2011 12:20:16 -0800 (PST)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 990883A6AA9 for ; Wed,  9 Mar 2011 12:20:16 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p29KLTXV011807 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 9 Mar 2011 14:21:29 -0600 (CST)
Received: from [135.244.32.47] (tbessis.lra.lucent.com [135.244.32.47]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p29KKubJ009574; Wed, 9 Mar 2011 14:20:56 -0600 (CST)
Message-ID: <4D77E128.8060302@alcatel-lucent.com>
Date: Wed, 09 Mar 2011 14:20:56 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Worley, Dale R (Dale)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com>	<4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com>	<4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090108080503020408080501"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 20:20:18 -0000

This is a multi-part message in MIME format.
--------------090108080503020408080501
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dale,

Please see my comments.

On 09-Mar-2011 11:10, Worley, Dale R (Dale) wrote:
> This is a fascinating subject.  On the way home from work, I started to frame a set of differential equations to model it.
>
> The key, of course, is to have all of the servers reach full capacity *at the same time* if the presented load increases steadily.  That is how we maximize goodput under such circumstances.  (If there is spare capacity within the cluster of servers, it doesn't matter if the load on the servers is balanced or not, as goodput equals presented load.)
>
> The crucial questions are "What is the algorithm for generating the number that the server presents to the load balancer?"
If you have read my previous comments, we suggest there is no algorithm 
per se: this is a measurement of the resource: If you measure the CPU 
and the Memory (for example), you'll take the smallest. It is better to 
smooth (with a 10s sliding window average for instance) the number.

> and "What algorithm does the load balancer use for assigning work, given the numbers presented by the servers?"
Isn't the one described in DNS explicit enough ? You just need to 
requesting a random number from 1 to the sum of the weights and choose 
the server accordingly.
> In order for the mechanism to be standardized in a meaningful way, it must specify the generation algorithm sufficiently that a combination of servers that use generation algorithms in conformance with the standard will work together correctly -- as I said above, as the entire cluster nears full capacity, all servers much reach full capacity simultaneously.
Exactly !
> t seems that we need to do a careful analysis of the behavior of the system near the point where "unused capacity of server i" is near 0 for all servers i.  The first question is to categorize generation functions by their behavior near the point "unused capacity = 0", that is, does the reported "weight" approach a non-zero number, does it approach zero linearly, does it approach zero quadratically, etc.?  I suspect that the system behaves well only when all servers use functions of the same form, and with the same constant on the leading term.  Intuitively, the correct answer seems to be to have the weight be "the unused capacity, measured in number of calls that can be handled".
This is not what we have in mind: the weight should be your relative 
resource usage. (0 - 100%). Using the proposed algorithm, this insure 
that servers reach full capacity simultaneously (even when they have 
vastly different capacities).
> This gets messier when we consider that the load is comes in quantized units (single calls) and has significant statistical fluctuations.  In particular, some functions that appear to work correctly in the "analytical" analysis outlined above my have bad behavior when presented with statistical fluctuations.
Agreed. In a signaling network such as IMS, CPU usage typically needs to 
be averaged over 10s to smooth the curve. Memory usage is already an 
integration, so it can typically be used as is.
> Additionally, the feedback from the servers to the load balancer will not be instantaneous, and the delays will introduce "control loop delays", which if not handled correctly might cause the system to oscillate.
Yes, this is true. We are successfully using the simple algorithm 
mentioned with DNS (where we update the weight every minute). Otherwise 
they are algorithms that can be used to improve the oscillation in the 
feedback chain command. We think that since we are proposing a default 
one that works, this should be OK.
> A different problem results from the fact that different calls may present decidedly different loads to a server, and that the server may have several different types of resource.  How does the server sensibly report the nature of its resources and the loading on each?  How does the load balancer estimate the resources required by an individual call?  (Note that this problem is largely nonexistent in circuit-switched systems, as all calls present exactly the same load.)  (This is the problem that has bedevilled attempts to solve the problem in draft-partha-sip-overload-resource-availability.)
This is why I suggest that the weight should just be a measurement of 
the relative resource usage: then none of those issues seems to exist.
Another advantage is that it is far more robust: the capacity can evolve 
based on whatever is happening on the server, and everything is not call 
related (Consider the case where the server is doing a re-synch with its 
mate: during sometime, its ability to process new calls will be 
considerably less and it would be preferable that less calls get in 
during that time.
> But I don't see any of the needed analysis in the draft, and I don't see any indication in the draft of a realization that this analysis needs to be done.  Until that is fixed, I predict that a solution will not be found, or worse, a solution that appears to work on paper but does not work in practice.
This solution works in practice: we are using it in ALU between 
components and nodes that are in the same NE (we currently use a 
proprietary transport mechanism to return the weights).

Cordially,

Thierry
> Dale
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------090108080503020408080501
Content-Type: multipart/related;
 boundary="------------050602040601090204060709"


--------------050602040601090204060709
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
    
  
  
    Dale, 

    

    Please see my comments.

    

    On 09-Mar-2011 11:10, Worley, Dale R (Dale) wrote:
    

      
This is a fascinating subject.  On the way home from work, I started to frame a set of differential equations to model it.

The key, of course, is to have all of the servers reach full capacity *at the same time* if the presented load increases steadily.  That is how we maximize goodput under such circumstances.  (If there is spare capacity within the cluster of servers, it doesn't matter if the load on the servers is balanced or not, as goodput equals presented load.)

The crucial questions are "What is the algorithm for generating the number that the server presents to the load balancer?" 

    

    If you have read my previous comments, we suggest there is no
    algorithm per se: this is a measurement of the resource: If you
    measure the CPU and the Memory (for example), you'll take the
    smallest. It is better to smooth (with a 10s sliding window average
    for instance) the number. 

    

    

      
and "What algorithm does the load balancer use for assigning work, given the numbers presented by the servers?"


    

    Isn't the one described in DNS explicit enough ? You just need to
    requesting a random number from 1 to the sum of the weights and
    choose the server accordingly. 

    

      
In order for the mechanism to be standardized in a meaningful way, it must specify the generation algorithm sufficiently that a combination of servers that use generation algorithms in conformance with the standard will work together correctly -- as I said above, as the entire cluster nears full capacity, all servers much reach full capacity simultaneously.


    

    Exactly ! 

    

      
t seems that we need to do a careful analysis of the behavior of the system near the point where "unused capacity of server i" is near 0 for all servers i.  The first question is to categorize generation functions by their behavior near the point "unused capacity = 0", that is, does the reported "weight" approach a non-zero number, does it approach zero linearly, does it approach zero quadratically, etc.?  I suspect that the system behaves well only when all servers use functions of the same form, and with the same constant on the leading term.  Intuitively, the correct answer seems to be to have the weight be "the unused capacity, measured in number of calls that can be handled".


    

    This is not what we have in mind: the weight should be your relative
    resource usage. (0 - 100%). Using the proposed algorithm, this
    insure that servers reach full capacity simultaneously (even when
    they have vastly different capacities). 

    

      
This gets messier when we consider that the load is comes in quantized units (single calls) and has significant statistical fluctuations.  In particular, some functions that appear to work correctly in the "analytical" analysis outlined above my have bad behavior when presented with statistical fluctuations.


    

    Agreed. In a signaling network such as IMS, CPU usage typically
    needs to be averaged over 10s to smooth the curve. Memory usage is
    already an integration, so it can typically be used as is. 

    

      
Additionally, the feedback from the servers to the load balancer will not be instantaneous, and the delays will introduce "control loop delays", which if not handled correctly might cause the system to oscillate.


    

    Yes, this is true. We are successfully using the simple algorithm
    mentioned with DNS (where we update the weight every minute).
    Otherwise they are algorithms that can be used to improve the
    oscillation in the feedback chain command. We think that since we
    are proposing a default one that works, this should be OK. 

    

      
A different problem results from the fact that different calls may present decidedly different loads to a server, and that the server may have several different types of resource.  How does the server sensibly report the nature of its resources and the loading on each?  How does the load balancer estimate the resources required by an individual call?  (Note that this problem is largely nonexistent in circuit-switched systems, as all calls present exactly the same load.)  (This is the problem that has bedevilled attempts to solve the problem in draft-partha-sip-overload-resource-availability.)


    

    This is why I suggest that the weight should just be a measurement
    of the relative resource usage: then none of those issues seems to
    exist. 

    Another advantage is that it is far more robust: the capacity can
    evolve based on whatever is happening on the server, and everything
    is not call related (Consider the case where the server is doing a
    re-synch with its mate: during sometime, its ability to process new
    calls will be considerably less and it would be preferable that less
    calls get in during that time. 

    

      
But I don't see any of the needed analysis in the draft, and I don't see any indication in the draft of a realization that this analysis needs to be done.  Until that is fixed, I predict that a solution will not be found, or worse, a solution that appears to work on paper but does not work in practice.


    

    This solution works in practice: we are using it in ALU between
    components and nodes that are in the same NE (we currently use a
    proprietary transport mechanism to return the weights). 

    

    Cordially, 

    

    Thierry

    

      
Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------050602040601090204060709
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------050602040601090204060709--

--------------090108080503020408080501--

From Thierry.Bessis@alcatel-lucent.com  Wed Mar  9 12:34:21 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 58E303A69FD for ; Wed,  9 Mar 2011 12:34:21 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.764
X-Spam-Level: 
X-Spam-Status: No, score=-5.764 tagged_above=-999 required=5 tests=[AWL=0.833,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 38sDBmZXFEWM for ; Wed,  9 Mar 2011 12:34:18 -0800 (PST)
Received: from ihemail4.lucent.com (ihemail4.lucent.com [135.245.0.39]) by core3.amsl.com (Postfix) with ESMTP id 529B23A6969 for ; Wed,  9 Mar 2011 12:34:15 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail4.lucent.com (8.13.8/IER-o) with ESMTP id p29KYgfY018012 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 9 Mar 2011 14:34:42 -0600 (CST)
Received: from [135.244.32.47] (tbessis.lra.lucent.com [135.244.32.47]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p29KYcuR022307; Wed, 9 Mar 2011 14:34:39 -0600 (CST)
Message-ID: <4D77E45E.8090005@alcatel-lucent.com>
Date: Wed, 09 Mar 2011 14:34:38 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Worley, Dale R (Dale)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com><4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com><4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com>	,  
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090809090905070705000909"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.39
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 20:34:21 -0000

This is a multi-part message in MIME format.
--------------090809090905070705000909
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dale,

> We still haven't demonstrated that the "obvious" strategy does not work in practice:  If a server is overloaded, it rejects the INVITE and the load balancer forks the INVITE to another server.
My opinion is the proposed SIP balancing mechanism should be employed 
together with the OLC to get the best results. Both systems are 
complementary and work well together.

The strategy to forks invite to other servers in a cluster in case of 
overload does not work in general. This has been demonstrated by 
simulation made by Volker Hilt, and we have experienced this issue as 
well in our Labs : When a cluster is near full capacity, and if nodes 
fork after overload rejection, the system collapses very fast. If the 
cluster is unevenly loaded, you don't even need to be close to full 
(cluster) capacity to collapse ! (Another good reason to be evenly loaded).

Cordially,

Thierry

> Dale
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------090809090905070705000909
Content-Type: multipart/related;
 boundary="------------080006050606010807020708"


--------------080006050606010807020708
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Dale, 

    

    

      
We still haven't demonstrated that the "obvious" strategy does not work in practice:  If a server is overloaded, it rejects the INVITE and the load balancer forks the INVITE to another server.


    

    My opinion is the proposed SIP balancing mechanism should be
    employed together with the OLC to get the best results. Both systems
    are complementary and work well together. 

    

    The strategy to forks invite to other servers in a cluster in case
    of overload does not work in general. This has been demonstrated by
    simulation made by Volker Hilt, and we have experienced this issue
    as well in our Labs : When a cluster is near full capacity, and if
    nodes fork after overload rejection, the system collapses very fast.
    If the cluster is unevenly loaded, you don't even need to be close
    to full (cluster) capacity to collapse ! (Another good reason to be
    evenly loaded). 

    

    Cordially, 

    

    Thierry

    

    

      
Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------080006050606010807020708
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------080006050606010807020708--

--------------090809090905070705000909--

From henry.sinnreich@gmail.com  Wed Mar  9 13:08:57 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1A0E63A6ABF for ; Wed,  9 Mar 2011 13:08:57 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.992
X-Spam-Level: 
X-Spam-Status: No, score=-2.992 tagged_above=-999 required=5 tests=[AWL=0.607,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uCCJU4f9F4ps for ; Wed,  9 Mar 2011 13:08:55 -0800 (PST)
Received: from mail-yi0-f44.google.com (mail-yi0-f44.google.com [209.85.218.44]) by core3.amsl.com (Postfix) with ESMTP id 538243A6ADC for ; Wed,  9 Mar 2011 13:08:55 -0800 (PST)
Received: by yic13 with SMTP id 13so494490yic.31 for ; Wed, 09 Mar 2011 13:10:12 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=2kPfVrXAqkYbzmCEKiuy2U4EgYWV+2KnW4zjBLq4sfs=; b=E0KypT9gkForr9sI/wB4xKoRVfeMzIBMRPot+IjjcXmqP7ZseLTT6PaazFVak87GRj 3r/Dk99DSRCwEZsanbEAXLxb0AgRB0a1dqU8yAaVdneGimfHagD1ulAupLo2K8beg4ZL 9JVw50pXhunYfrWM4wPL4zHlnynM5FWfhW5nk=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=okMGO20SubbRyOrdBEzG5EcapeU75aRx+L9x1HxTj9ZXIE+lXiiTg4szuwg+uCTSZR 9J26+n7FKBSHR3mEiJ4kU3wEtFwAmULaEUktWw6vJGxVLKCiExTfxCPy1qn3dyfrDTxy rNIjZ1m7E6stuSVWx7CHNdsql73iutmyr8A7g=
Received: by 10.236.190.225 with SMTP id e61mr346579yhn.208.1299705011926; Wed, 09 Mar 2011 13:10:11 -0800 (PST)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id z5sm1567520yhc.35.2011.03.09.13.10.07 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 09 Mar 2011 13:10:10 -0800 (PST)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Wed, 09 Mar 2011 15:10:04 -0600
From: Henry Sinnreich 
To: Itsuma TANAKA , 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcveafX/UNE//jAPRxC3GPIUezdMVAANGYCF
In-Reply-To: 
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 21:08:57 -0000

Such work in the IETF would require IMO an explanation about its
compatibility with the open, e2e architectural and other technical
principles of the Internet. It actually seems to be the opposite.

For this reason, such work may be better accomplished in other
organizations, so as to avoid architectural principle and other technical
conflicts. Simplicity and scalability comes to mind among other, though this
is better discussed elsewhere.

Thanks,

Henry Sinnreich


On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:

> Dear All,
> 
> I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP meetings (SA2,
> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming Expert
> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  I serve
> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
> 
> I would like to express my support of draft-montemurro-gsma-imei-urn and the
> associated draft-allen-dispatch-imei-urn-as-instanceid.
> 
> The reasons are exactly same as what Ricky Kaura (Samsung) expressed.  A
> various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and TS24.292), and
> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are using
> these drafts.
>  
> These internet drafts are very important to support various key features in
> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would like ask
> that this work is progressed as quickly as possible in IETF.
> 
> 
> Best regards,
> 
> Itsuma Tanaka
> NTT DOCOMO
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch



From aallen@rim.com  Wed Mar  9 15:43:58 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6741B3A6AF9 for ; Wed,  9 Mar 2011 15:43:58 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.203
X-Spam-Level: 
X-Spam-Status: No, score=-1.203 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, MIME_QP_LONG_LINE=1.396]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3+Qf0TTLMfq9 for ; Wed,  9 Mar 2011 15:43:57 -0800 (PST)
Received: from mhs060cnc.rim.net (mhs060cnc.rim.net [208.65.73.34]) by core3.amsl.com (Postfix) with ESMTP id 5E17D3A6AF6 for ; Wed,  9 Mar 2011 15:43:56 -0800 (PST)
X-AuditID: 0a41282f-b7b4eae000005e51-ee-4d781104863a
Received: from XCH139CNC.rim.net (xch139cnc.rim.net [10.65.10.235]) by mhs060cnc.rim.net (SBG) with SMTP id 53.0F.24145.401187D4; Wed,  9 Mar 2011 23:45:08 +0000 (GMT)
Received: from XCH02DFW.rim.net ([10.150.100.31]) by XCH139CNC.rim.net with Microsoft SMTPSVC(6.0.3790.3959); Wed, 9 Mar 2011 18:45:13 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
content-transfer-encoding: quoted-printable
Date: Wed, 9 Mar 2011 17:45:11 -0600
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcveafX/UNE//jAPRxC3GPIUezdMVAANGYCFAAVqsB4=
From: "Andrew Allen" 
To: , , 
X-OriginalArrivalTime: 09 Mar 2011 23:45:13.0080 (UTC) FILETIME=[08A67F80:01CBDEB4]
X-Brightmail-Tracker: AAAAAQAAAZE=
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 09 Mar 2011 23:43:58 -0000

Henry

Would you care to elaborate and explain your specific concerns?  How is what=
 is defined in these drafts potentially incompatible with the open, e2e arch=
itectural and in any other way technical conflict with the Internet?

The two drafts do the following things:

One draft simply defines and registers with IANA a URN namespace according t=
o the procedures defined for doing that.

The other draft simply defines how this URN can be used as an instance ID as=
 required by RFC 5626. 

SIP and the mechanisms in RFC 5626 are already being used in networks define=
d by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA have=
 already defined this behavior using their existing device Identifiers for i=
nstance ID generation in their specifications. These drafts complete the pro=
cess doing things the right way by reserving the namespace and defining the=
 usage of it using the IETF defined mechanisms for doing that.

Andrew

----- Original Message -----
From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
Sent: Wednesday, March 09, 2011 03:10 PM
To: Itsuma TANAKA ; dispatch@ietf.org 
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instancei=
d

Such work in the IETF would require IMO an explanation about its
compatibility with the open, e2e architectural and other technical
principles of the Internet. It actually seems to be the opposite.

For this reason, such work may be better accomplished in other
organizations, so as to avoid architectural principle and other technical
conflicts. Simplicity and scalability comes to mind among other, though this
is better discussed elsewhere.

Thanks,

Henry Sinnreich


On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:

> Dear All,
> 
> I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP meetings (SA2=
,
> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming Expert
> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  I serv=
e
> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
> 
> I would like to express my support of draft-montemurro-gsma-imei-urn and t=
he
> associated draft-allen-dispatch-imei-urn-as-instanceid.
> 
> The reasons are exactly same as what Ricky Kaura (Samsung) expressed.  A
> various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and TS24.292), a=
nd
> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are using
> these drafts.
>  
> These internet drafts are very important to support various key features i=
n
> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would like as=
k
> that this work is progressed as quickly as possible in IETF.
> 
> 
> Best regards,
> 
> Itsuma Tanaka
> NTT DOCOMO
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

---------------------------------------------------------------------
This transmission (including any attachments) may contain confidential infor=
mation, privileged material (including material protected by the solicitor-c=
lient or other applicable privileges), or constitute non-public information.=
 Any use of this information by anyone other than the intended recipient is=
 prohibited. If you have received this transmission in error, please immedia=
tely reply to the sender and delete this information from your system. Use,=
 dissemination, distribution, or reproduction of this transmission by uninte=
nded recipients is not authorized and may be unlawful.

From paulej@packetizer.com  Wed Mar  9 16:46:15 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 917613A698A for ; Wed,  9 Mar 2011 16:46:15 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5rRl04gKV+1d for ; Wed,  9 Mar 2011 16:46:14 -0800 (PST)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id 03C433A67D9 for ; Wed,  9 Mar 2011 16:46:13 -0800 (PST)
Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2A0lNdp022023 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Wed, 9 Mar 2011 19:47:29 -0500
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299718049; bh=VffPojotgOWA3NUg4NZwfJoHdDnX4bSm22DsA/SMJ8M=; h=From:To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type:Content-Transfer-Encoding; b=Ux3DWPnWX//kmF+kQc0UgUCgfDKYl/U33OxA4PMQEVTqS+v4WexA3H8ItdZBdLDZ1 pFYf+EkGTmGuEjEyGW8CQ8+drjS2t2qyy2sIZ69hpnmkRMZeC4DkBtegL3AM7Y5Owv 7T+Y129oSfnehgMpaQb/vtR2rHg1TMglkx3OU8wI=
From: "Paul E. Jones" 
To: "'Vijay K. Gurbani'" 
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D769926.1050103@bell-labs.com>
In-Reply-To: <4D769926.1050103@bell-labs.com>
Date: Wed, 9 Mar 2011 19:47:17 -0500
Message-ID: <023e01cbdebc$b8045260$280cf720$@packetizer.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3kCCnyKVgHxyhjfk6gBmTA=
Content-Language: en-us
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 00:46:15 -0000

Vijay,

> I believe that we should take the view that load balancing and overload-
> control are separate

They are different, but one can use overload prevention to balance load.
That was my point.  The focus of our draft was to avoid overload.  A
byproduct is that I can use it to perfectly balance load on a peer server.
 
> Now ... to the details of "sce" and "weight" as proposed in our draft.
> 
> In the large, I agree that we are thinking along the same lines.  This
> is good.
> 
> In addition to the email exchange between you and Thierry, there is one
> more property that is very attractive: statelessness.
> 
> I think that it is advantageous that we impose the property of
> statelessness on a load balancing mechanism.  With "sce", the upstream
> SIP proxy must keep track of how many requests it has sent to a
> downstream SIP server since the latter sent it the last "sce" value.

SCE does not propose keeping track of messages.  As a server replies
forwards a message to a peer, it inserts an SCE value: a value that
indicates how much additional capacity it can handle given its current load.
Imagine an SBC at the edge of my enterprise peering with AT&T.  Perhaps I
have 5 SBCs to which I can direct traffic.  As I forward traffic to the SBC,
I see in response messages SCE values like 1000, 1200, 2300, 1300, 850.
This tells my enterprise SBC roughly how many more sessions I can direct
toward each of those SBCs. The state maintained by the enterprise SBC is
perhaps an ordered list (or using whatever algorithm one chooses) to select
which SBC to direct the next session toward.

> Furthermore, each time a request is sent to the downstream SIP server,
> the SIP proxy must decrement its request count.

The enterprise SBC does not decrement counters, but it will likely be
provided with a new SCE value that might be influenced by other factors
other than the enterprise SBC.

> Consequently, the state on how many requests have been sent to the
> downstream SIP server since the SIP proxy got the last "sce" parameter,
> what the low water-mark is, is all lost if not replicated upon a reboot.

None of this information is needed.  When the enterprise SBC is rebooted, it
will discover the SCE values from its peers the next time it directs traffic
(either a call or an OPTIONS "ping").

> With "weight", you bypass all of this state maintenance business.

The SCE draft does not introduce any of the state you describe.  I agree
entirely that if I had to maintain all the state you described, it would be
complex.  But, SCE does not do it.

> All the SIP proxy needs is a "weight" indicator that it will use in a
> RFC2782-type algorithm to proportionally send as much traffic to a
> downstream SIP server as it has indicated (i.e., higher "weight"
> values will get proportionally more traffic).  This is much better than
> round-robin, and furthermore, there is no requirement that each
> downstream SIP server in the cluster set its "weight" in a consistent
> manner, or even that the "weight" values of two SIP servers be
> comparable.

As RFC 2782 says, "Weight, the server selection field, is not quite
satisfactory, but the actual load on typical servers changes much too
quickly to be kept around in DNS caches."

That is precisely the case with high-transaction systems like SBCs.  It is
even more pronounced when those SBCs are handling phone calls of varying
duration, or processing media flows that vary significantly in terms of CPU
or bandwidth consumption, or when performing transcoding or transrating
functions.

I fully agree that a simple round-robin approach does not work.  We've
proven that.  What does work, though, is a combination of round-robin over a
set of peer devices that we know have available capacity.  That's where SCE
comes into play.  For over 10 years, we have performed peer selection using
an approach very much like SCE + round robin for 10 years now.  It works.
 
> If a SIP server thinks that it has infinite resources, it can set the
> "weight" to be 65,535.  Consequently, it will get a proportionally more
> traffic if the "weight" of other SIP servers in the cluster is set to,
> say, 10 or 20.  Similarly, if the SIP server in a cluster wants to
> accept less number of messages, it sets a
> lower, albeit non-zero, value for its "weight" parameter.   If the
> cluster has capacity, other load control weights will be higher, and
> will thus get a proportionate share of the traffic.  As all SIP servers
> in the cluster reach capacity, the load distribution algorithm arranges
> for a uniform load distribution to all of the SIP servers until either
> the offered load subsides or the SIP servers enter into overload mode
> and start shedding traffic using the mechanism defined in the overload
> control draft.

This is where the weighting mechanism starts to sound lot like SCE. Rather
than providing a "weight", we call it "sce".  I had proposed before that SCE
would be calculated as the minimum "capacity" value from a set of capacity
measurements. The device would estimate its capacity given current memory
utilization with respect to the current number of sessions.  The device
would estimate its capacity given current CPU utilization with respect to
the current number of sessions.  What the device would consider (CPU,
memory, DSPs, whatever) is outside the scope of the draft, but in the end,
the device determines these values and returns the minimum as the SCE value.
(After all, if memory is exhausted, then it does not matter how many other
resources are available: one is out of memory.)  This is the approach we
take, measuring resources on a set interval or certain events.  Again, all
of that is really a matter of implementation.

At the end, our comes an SCE value and it sounds a lot like the weight.
It's a number with which a peer can compare against similar numbers reported
from other peers.

> Indeed.  If this work progresses, it'll be nice to have a coherent view
> going forward.  If you will be in Prague, it will be good to hook up and
> chat about this.

I'd love to, but I will not be there, unfortunately. (I think 1999 was the
last time I had an opportunity to make it to an IETF meeting.)

Paul


From gao.yang2@zte.com.cn  Wed Mar  9 19:25:21 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 060D23A68AC; Wed,  9 Mar 2011 19:25:21 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -94.821
X-Spam-Level: 
X-Spam-Status: No, score=-94.821 tagged_above=-999 required=5 tests=[AWL=-3.014, BAYES_00=-2.599, CHARSET_FARAWAY_HEADER=3.2, CN_SUBJECT_136=0.982, HTML_MESSAGE=0.001, MIME_8BIT_HEADER=0.3, MIME_BASE64_TEXT=1.753, MIME_CHARSET_FARAWAY=2.45, RCVD_DOUBLE_IP_LOOSE=0.76, SARE_SUB_ENC_GB2312=1.345, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id iT-d1vrBADx4; Wed,  9 Mar 2011 19:25:18 -0800 (PST)
Received: from mx5.zte.com.cn (mx5.zte.com.cn [63.217.80.70]) by core3.amsl.com (Postfix) with ESMTP id 777473A6870; Wed,  9 Mar 2011 19:25:17 -0800 (PST)
Received: from [10.30.17.99] by mx5.zte.com.cn with surfront esmtp id 205952302718646; Thu, 10 Mar 2011 11:21:12 +0800 (CST)
Received: from [10.30.3.20] by [192.168.168.15] with StormMail ESMTP id 84746.4283288557; Thu, 10 Mar 2011 11:26:28 +0800 (CST)
Received: from notes_smtp.zte.com.cn ([10.30.1.239]) by mse01.zte.com.cn with ESMTP id p2A3QKGO099281; Thu, 10 Mar 2011 11:26:20 +0800 (GMT-8) (envelope-from gao.yang2@zte.com.cn)
In-Reply-To: <4D769926.1050103@bell-labs.com>
To: "Vijay K. Gurbani" 
MIME-Version: 1.0
X-KeepSent: 58CEE046:1EEC4D1F-4825784F:0012849F; type=4; name=$KeepSent
X-Mailer: Lotus Notes Release 6.5.6 March 06, 2007
Message-ID: 
From: gao.yang2@zte.com.cn
Date: Thu, 10 Mar 2011 11:22:56 +0800
X-MIMETrack: Serialize by Router on notes_smtp/zte_ltd(Release 8.5.1FP4|July 25, 2010) at 2011-03-10 11:26:20, Serialize complete at 2011-03-10 11:26:20
Content-Type: multipart/alternative; boundary="=_alternative 0012F0CF4825784F_="
X-MAIL: mse01.zte.com.cn p2A3QKGO099281
Cc: dispatch-bounces@ietf.org, "dispatch@ietf.org" 
Subject: [dispatch] =?gb2312?b?Rje1xFJGQ7eisrzQws7FuOWjrLmpss6/vCAgIC8v?= =?gb2312?b?tPC4tDogUmU6ICBTSVAgbG9hZCBiYWxhbmNpbmc=?=
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 03:25:21 -0000

This is a multipart message in MIME format.
--=_alternative 0012F0CF4825784F_=
Content-Type: text/plain; charset="GB2312"
Content-Transfer-Encoding: base64

aHR0cDovL3RlY2guMTYzLmNvbS8wOS8wMTIxLzE0LzUwNklNUkZMMDAwOTE1QkUuaHRtbA0KDQpo
dHRwOi8vdGVjaC5xcS5jb20vYS8yMDA5MDEyMS8wMDAyMDQuaHRtDQoNCmh0dHA6Ly9pdC5zb2h1
LmNvbS8yMDA5MDEyMS9uMjYxODgzODE3LnNodG1sDQoNCj09PT09PT09PT09PT09PT09PT09PT09
PT09PT09PT09PT09DQogWmlwICAgIDogMjEwMDEyDQogVGVsICAgIDogODcyMTENCiBUZWwyICAg
OigrODYpLTAyNS01Mjg3NzIxMQ0KIGVfbWFpbCA6IGdhby55YW5nMkB6dGUuY29tLmNuDQo9PT09
PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQ0KDQoNCg0KIlZpamF5IEsuIEd1cmJhbmki
IDx2a2dAYmVsbC1sYWJzLmNvbT4gDQq3orz+yMs6ICBkaXNwYXRjaC1ib3VuY2VzQGlldGYub3Jn
DQoyMDExLTAzLTA5IDA1OjAxDQoNCsrVvP7Iyw0KcGF1bGVqQHBhY2tldGl6ZXIuY29tDQqzrcvN
DQoiZGlzcGF0Y2hAaWV0Zi5vcmciIDxkaXNwYXRjaEBpZXRmLm9yZz4NCtb3zOINClJlOiBbZGlz
cGF0Y2hdIFNJUCBsb2FkIGJhbGFuY2luZw0KDQoNCg0KDQoNCg0KT24gMDMvMDcvMjAxMSAxMTow
NSBQTSwgUGF1bCBFLiBKb25lcyB3cm90ZToNCj4gVmlqYXksDQo+DQo+IFRoZSBTZXNzaW9uIENh
cGFjaXR5IEVzdGltYXRlIChTQ0UpIGlkZWEgaXMgYWN0dWFsbHkgcXVpdGUgc2ltaWxhcg0KPiB0
byB0aGUgImFkYXB0aXZlIGxvYWQgYmFsYW5jaW5nIiBkcmFmdCwgZnJvbSB3aGF0IEkgY2FuIHNl
ZS4NCg0KUGF1bDogVGhhbmsgeW91IG11Y2ggZm9yIHlvdXIgaW5wdXQuICBNb3JlIGlubGluZS4N
Cg0KPiBTQ0Ugd2FzIHByb3Bvc2VkIGFzIGEgc29sdXRpb24gdG8gdGhlICJvdmVybG9hZCIgcHJv
YmxlbSwgdGhlDQo+IGFwcHJvYWNoIGJlaW5nIG9uZSB3aGVyZSB3ZSB0cnkgdG8gYXZvaWQgb3Zl
cmxvYWQgYnkgc2VsZWN0aW5nIGENCj4gZGV2aWNlIHRoYXQgcmVwb3J0cyBhIGhpZ2hlciBjYXBh
Y2l0eSB2YWx1ZS4gVGh1cywgb25lIGNvdWxkIHZpZXcgaXQNCj4gYXMgIm92ZXJsb2FkIGNvbnRy
b2wgdGhyb3VnaCBsb2FkIGJhbGFuY2luZyIuDQoNCkkgdGhpbmsgb3ZlcmxvYWQgY29udHJvbCBh
bmQgbG9hZCBiYWxhbmNpbmcgYXJlIGRpc3RpbmN0LCBldmVuIHRob3VnaA0KaXQgaXMgbmF0dXJh
bCB0byByZWxhdGUgdGhlbS4gIExvYWQgYmFsYW5jaW5nIGNhbiBkZWxheSwgb3IgcHJldmVudCwN
Cm92ZXJsb2FkOyBidXQgb25jZSBhIFNJUCBob3N0IGlzIG92ZXJsb2FkZWQsIGxvYWQgYmFsYW5j
aW5nIGJ5IGl0c2VsZg0Kd2lsbCBub3QgaGVscCB0b28gbXVjaC4NCg0KSSB0aGluayBpdCBpcyBi
ZXN0IHRoYXQgd2UgdHJlYXQgdGhlbSBzZXBhcmF0ZWx5LCBpZiBub3QgZm9yIGFueXRoaW5nDQpl
bHNlIHRoYW4gdGhlIGZhY3QgdGhhdCBTT0MgaXMgbm90IGNoYXJ0ZXJlZCB0byBjcmVhdGUgYSBz
b2x1dGlvbiBmb3INCmxvYWQgYmFsYW5jaW5nLiAgV2UgKGkuZS4sIHRoZSBkaXNwYXRjaCBXRykg
Y2FuIGRlY2lkZSB0byBzZW5kIHRoaXMNCndvcmsgdG8gU09DLCBidXQgSSB0aGluayB0aGF0IGRl
Y2lzaW9uIGlzIGEgYml0IHByZW1hdHVyZSBhdCB0aGlzDQpwb2ludC4gIEZ1cnRoZXJtb3JlLCBJ
IHVuZGVyc3RhbmQgdGhhdCBiZWZvcmUgdGhlIGZvcm1hdGlvbiBvZiB0aGUNClNPQyBXRywgdGhl
cmUgd2FzIHNvbWUgZGViYXRlIG9uIHdoZXRoZXIgb3Igbm90IHRvIGluY2x1ZGUgbG9hZC0NCmJh
bGFuY2luZyBpbiB0aGUgY2hhcnRlciBhbmQgaXQgd2FzIGRlY2lkZWQgdGhhdCBsb2FkLWJhbGFu
Y2luZyB3aWxsIG5vdA0KYmUgaW5jbHVkZWQuDQoNCkkgYmVsaWV2ZSB0aGF0IHdlIHNob3VsZCB0
YWtlIHRoZSB2aWV3IHRoYXQgbG9hZCBiYWxhbmNpbmcgYW5kIG92ZXJsb2FkLQ0KY29udHJvbCBh
cmUgc2VwYXJhdGUsIGFuZCB0aGF0IHdoZW4gYSBjbHVzdGVyIG9mIFNJUCBzZXJ2ZXJzIGlzIHB1
dCBpbg0Kc2VydmljZSwgYWxsIG9mIHRoZW0gc3RhcnQgb2ZmIHByb3ZpZGluZyBsb2FkIGJhbGFu
Y2luZyBpbmZvcm1hdGlvbiB0bw0KdGhlIHVwc3RyZWFtIFNJUCBwcm94eS4gIFRoZSB1cHN0cmVh
bSBTSVAgcHJveHkgd2lsbCBkbyBsb2FkIGJhbGFuY2luZw0KdG8gdHJ5IHRvIHNwcmVhZCB0aGUg
bG9hZCBwcm9wb3J0aW9uYWxseSB0byB0aGUgcmVzb3VyY2VzIGF2YWlsYWJsZSBhdA0KdGhlIGRv
d25zdHJlYW0gU0lQIHNlcnZlcnMgKG5vdCBldmVubHksIHByb3BvcnRpb25hbGx5KS4gIEFzIHRp
bWUNCnByb2dyZXNzZXMsIHNvbWUgb2YgdGhlbSB3aWxsIGVudGVyIGFuIG92ZXJsb2FkIHN0YXRl
IGFuZCBhdCB0aGlzIHRpbWUsDQp0aGUgb3ZlcmxvYWQgY29udHJvbCBraWNrcyBpbi4gIFdoZW4g
dGhlIG92ZXJsb2FkIGNvbmRpdGlvbiBpcw0KYWxsZXZpYXRlZCwgdGhlIG92ZXJsb2FkZWQgcGVl
ciBnb2VzIGJhY2sgdG8gcmVwb3J0aW5nIGl0cyBsb2FkLg0KDQpOb3cgLi4uIHRvIHRoZSBkZXRh
aWxzIG9mICJzY2UiIGFuZCAid2VpZ2h0IiBhcyBwcm9wb3NlZCBpbiBvdXIgZHJhZnQuDQoNCklu
IHRoZSBsYXJnZSwgSSBhZ3JlZSB0aGF0IHdlIGFyZSB0aGlua2luZyBhbG9uZyB0aGUgc2FtZSBs
aW5lcy4gIFRoaXMNCmlzIGdvb2QuDQoNCkluIGFkZGl0aW9uIHRvIHRoZSBlbWFpbCBleGNoYW5n
ZSBiZXR3ZWVuIHlvdSBhbmQgVGhpZXJyeSwgdGhlcmUNCmlzIG9uZSBtb3JlIHByb3BlcnR5IHRo
YXQgaXMgdmVyeSBhdHRyYWN0aXZlOiBzdGF0ZWxlc3NuZXNzLg0KDQpJIHRoaW5rIHRoYXQgaXQg
aXMgYWR2YW50YWdlb3VzIHRoYXQgd2UgaW1wb3NlIHRoZSBwcm9wZXJ0eSBvZg0Kc3RhdGVsZXNz
bmVzcyBvbiBhIGxvYWQgYmFsYW5jaW5nIG1lY2hhbmlzbS4gIFdpdGggInNjZSIsDQp0aGUgdXBz
dHJlYW0gU0lQIHByb3h5IG11c3Qga2VlcCB0cmFjayBvZiBob3cgbWFueSByZXF1ZXN0cyBpdA0K
aGFzIHNlbnQgdG8gYSBkb3duc3RyZWFtIFNJUCBzZXJ2ZXIgc2luY2UgdGhlIGxhdHRlciBzZW50
IGl0IHRoZQ0KbGFzdCAic2NlIiB2YWx1ZS4gIEZ1cnRoZXJtb3JlLCBlYWNoIHRpbWUgYSByZXF1
ZXN0IGlzIHNlbnQgdG8gdGhlDQpkb3duc3RyZWFtIFNJUCBzZXJ2ZXIsIHRoZSBTSVAgcHJveHkg
bXVzdCBkZWNyZW1lbnQgaXRzIHJlcXVlc3QgY291bnQuDQpDb25zZXF1ZW50bHksIHRoZSBzdGF0
ZSBvbiBob3cgbWFueSByZXF1ZXN0cyBoYXZlIGJlZW4gc2VudCB0byB0aGUNCmRvd25zdHJlYW0g
U0lQIHNlcnZlciBzaW5jZSB0aGUgU0lQIHByb3h5IGdvdCB0aGUgbGFzdCAic2NlIiBwYXJhbWV0
ZXIsDQp3aGF0IHRoZSBsb3cgd2F0ZXItbWFyayBpcywgaXMgYWxsIGxvc3QgaWYgbm90IHJlcGxp
Y2F0ZWQgdXBvbiBhDQpyZWJvb3QuDQoNCldpdGggIndlaWdodCIsIHlvdSBieXBhc3MgYWxsIG9m
IHRoaXMgc3RhdGUgbWFpbnRlbmFuY2UgYnVzaW5lc3MuDQoNCkFsbCB0aGUgU0lQIHByb3h5IG5l
ZWRzIGlzIGEgIndlaWdodCIgaW5kaWNhdG9yIHRoYXQgaXQgd2lsbCB1c2UNCmluIGEgUkZDMjc4
Mi10eXBlIGFsZ29yaXRobSB0byBwcm9wb3J0aW9uYWxseSBzZW5kIGFzIG11Y2ggdHJhZmZpYw0K
dG8gYSBkb3duc3RyZWFtIFNJUCBzZXJ2ZXIgYXMgaXQgaGFzIGluZGljYXRlZCAoaS5lLiwgaGln
aGVyICJ3ZWlnaHQiDQp2YWx1ZXMgd2lsbCBnZXQgcHJvcG9ydGlvbmFsbHkgbW9yZSB0cmFmZmlj
KS4gIFRoaXMgaXMgbXVjaCBiZXR0ZXINCnRoYW4gcm91bmQtcm9iaW4sIGFuZCBmdXJ0aGVybW9y
ZSwgdGhlcmUgaXMgbm8gcmVxdWlyZW1lbnQgdGhhdCBlYWNoDQpkb3duc3RyZWFtIFNJUCBzZXJ2
ZXIgaW4gdGhlIGNsdXN0ZXIgc2V0IGl0cyAid2VpZ2h0IiBpbiBhIGNvbnNpc3RlbnQNCm1hbm5l
ciwgb3IgZXZlbiB0aGF0IHRoZSAid2VpZ2h0IiB2YWx1ZXMgb2YgdHdvIFNJUCBzZXJ2ZXJzIGJl
DQpjb21wYXJhYmxlLg0KDQpJZiBhIFNJUCBzZXJ2ZXIgdGhpbmtzIHRoYXQgaXQgaGFzIGluZmlu
aXRlIHJlc291cmNlcywgaXQNCmNhbiBzZXQgdGhlICJ3ZWlnaHQiIHRvIGJlIDY1LDUzNS4gIENv
bnNlcXVlbnRseSwgaXQgd2lsbCBnZXQgYQ0KcHJvcG9ydGlvbmFsbHkgbW9yZSB0cmFmZmljIGlm
IHRoZSAid2VpZ2h0IiBvZiBvdGhlciBTSVAgc2VydmVycw0KaW4gdGhlIGNsdXN0ZXIgaXMgc2V0
IHRvLCBzYXksIDEwIG9yIDIwLiAgU2ltaWxhcmx5LCBpZiB0aGUgU0lQIHNlcnZlcg0KaW4gYSBj
bHVzdGVyIHdhbnRzIHRvIGFjY2VwdCBsZXNzIG51bWJlciBvZiBtZXNzYWdlcywgaXQgc2V0cyBh
DQpsb3dlciwgYWxiZWl0IG5vbi16ZXJvLCB2YWx1ZSBmb3IgaXRzICJ3ZWlnaHQiIHBhcmFtZXRl
ci4gICBJZiB0aGUNCmNsdXN0ZXIgaGFzIGNhcGFjaXR5LCBvdGhlciBsb2FkIGNvbnRyb2wgd2Vp
Z2h0cyB3aWxsIGJlIGhpZ2hlciwgYW5kDQp3aWxsIHRodXMgZ2V0IGEgcHJvcG9ydGlvbmF0ZSBz
aGFyZSBvZiB0aGUgdHJhZmZpYy4gIEFzIGFsbCBTSVAgc2VydmVycw0KaW4gdGhlIGNsdXN0ZXIg
cmVhY2ggY2FwYWNpdHksIHRoZSBsb2FkIGRpc3RyaWJ1dGlvbiBhbGdvcml0aG0gYXJyYW5nZXMN
CmZvciBhIHVuaWZvcm0gbG9hZCBkaXN0cmlidXRpb24gdG8gYWxsIG9mIHRoZSBTSVAgc2VydmVy
cyB1bnRpbCBlaXRoZXINCnRoZSBvZmZlcmVkIGxvYWQgc3Vic2lkZXMgb3IgdGhlIFNJUCBzZXJ2
ZXJzIGVudGVyIGludG8gb3ZlcmxvYWQgbW9kZQ0KYW5kIHN0YXJ0IHNoZWRkaW5nIHRyYWZmaWMg
dXNpbmcgdGhlIG1lY2hhbmlzbSBkZWZpbmVkIGluIHRoZSBvdmVybG9hZA0KY29udHJvbCBkcmFm
dC4NCg0KPiBJIGRvbid0IHRoaW5rIGxvY2FsIGRlY2lzaW9ucyBpbiB0aGUgYm94IG5lZWQgdG8g
YmUgc3ViamVjdCB0bw0KPiBzdGFuZGFyZGl6YXRpb246IHdlIGp1c3QgbmVlZCB0byBzdGFuZGFy
ZGl6ZSB0aGUgc2lnbmFsaW5nIHRvIGVuYWJsZQ0KPiBzdWNoIGRlY2lzaW9ucy4NCg0KSSBjb3Vs
ZCBub3QgYWdyZWUgbW9yZS4gIExvY2FsIGRlY2lzaW9ucyBvbiBob3cgZWFjaCBTSVAgc2VydmVy
IGluIGENCmNsdXN0ZXIgZ2VuZXJhdGVzIGEgIndlaWdodCIgdmFsdWUgYXJlIG5vdCBzdWJqZWN0
IHRvIHN0YW5kYXJkaXphdGlvbi4NClRyYW5zcG9ydGluZyB0aGlzIHZhbHVlIGFuZCBzcGVjaWZ5
aW5nIGEgZGVmYXVsdCBhbGdvcml0aG0gdG8gYWxsb3cNCmEgU0lQIHByb3h5IHRvIHBlcmZvcm0g
bG9hZCBiYWxhbmNpbmcgc2VlbSB0byBiZSBhIGdvb2QgYXZlbnVlIGZvcg0Kc3RhbmRhcmRpemF0
aW9uLg0KDQpOb3RlIHRoYXQgSSB0aGluayBzcGVjaWZ5aW5nIGEgZGVmYXVsdCBhbGdvcml0aG0g
dG8gZG8gbG9hZCBiYWxhbmNpbmcNCmlzIGltcG9ydGFudCwgYW5kIG91ciBkcmFmdCBjaG9vc2Vz
IHRvIHVzZSB0aGUgd2VsbC11bmRlcnN0b29kIFJGQzI3ODINCnVzYWdlIG9mICJ3ZWlnaHQiIHRv
IGRvIHRoaXMuICBUaGlzIGRvZXMgbm90IHByZWNsdWRlIGFueSBpbXBsZW1lbnRlcg0KZnJvbSBj
b21pbmcgdXAgd2l0aCB0aGVpciBvd24gYWxnb3JpdGhtIGFuZCB1c2luZyB0aGF0IGluc3RlYWQu
DQoNCj4gSW4gYW55IGNhc2UsIGl0IGNlcnRhaW5seSBsb29rcyBsaWtlIHdlJ3JlIHRoaW5raW5n
IGFsb25nIHRoZSBzYW1lDQo+IGxpbmVzLg0KDQpJbmRlZWQuICBJZiB0aGlzIHdvcmsgcHJvZ3Jl
c3NlcywgaXQnbGwgYmUgbmljZSB0byBoYXZlIGEgY29oZXJlbnQNCnZpZXcgZ29pbmcgZm9yd2Fy
ZC4gIElmIHlvdSB3aWxsIGJlIGluIFByYWd1ZSwgaXQgd2lsbCBiZSBnb29kIHRvDQpob29rIHVw
IGFuZCBjaGF0IGFib3V0IHRoaXMuDQoNCk9uY2UgYWdhaW4sIHRoYW5rcyBmb3IgeW91ciBpbnB1
dC4NCg0KLSB2aWpheQ0KLS0gDQpWaWpheSBLLiBHdXJiYW5pLCBCZWxsIExhYm9yYXRvcmllcywg
QWxjYXRlbC1MdWNlbnQNCjE5NjAgTHVjZW50IExhbmUsIFJtLiA5Qy01MzMsIE5hcGVydmlsbGUs
IElsbGlub2lzIDYwNTY2IChVU0EpDQpFbWFpbDogdmtnQHtiZWxsLWxhYnMuY29tLGFjbS5vcmd9
IC8gdmlqYXkuZ3VyYmFuaUBhbGNhdGVsLWx1Y2VudC5jb20NCldlYjogICBodHRwOi8vZWN0LmJl
bGwtbGFicy5jb20vd2hvL3ZrZy8NCl9fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19f
X19fX19fX19fX19fDQpkaXNwYXRjaCBtYWlsaW5nIGxpc3QNCmRpc3BhdGNoQGlldGYub3JnDQpo
dHRwczovL3d3dy5pZXRmLm9yZy9tYWlsbWFuL2xpc3RpbmZvL2Rpc3BhdGNoDQoNCg0KDQoNCg0K
LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0N
ClpURSBJbmZvcm1hdGlvbiBTZWN1cml0eSBOb3RpY2U6IFRoZSBpbmZvcm1hdGlvbiBjb250YWlu
ZWQgaW4gdGhpcyBtYWlsIGlzIHNvbGVseSBwcm9wZXJ0eSBvZiB0aGUgc2VuZGVyJ3Mgb3JnYW5p
emF0aW9uLiBUaGlzIG1haWwgY29tbXVuaWNhdGlvbiBpcyBjb25maWRlbnRpYWwuIFJlY2lwaWVu
dHMgbmFtZWQgYWJvdmUgYXJlIG9ibGlnYXRlZCB0byBtYWludGFpbiBzZWNyZWN5IGFuZCBhcmUg
bm90IHBlcm1pdHRlZCB0byBkaXNjbG9zZSB0aGUgY29udGVudHMgb2YgdGhpcyBjb21tdW5pY2F0
aW9uIHRvIG90aGVycy4NClRoaXMgZW1haWwgYW5kIGFueSBmaWxlcyB0cmFuc21pdHRlZCB3aXRo
IGl0IGFyZSBjb25maWRlbnRpYWwgYW5kIGludGVuZGVkIHNvbGVseSBmb3IgdGhlIHVzZSBvZiB0
aGUgaW5kaXZpZHVhbCBvciBlbnRpdHkgdG8gd2hvbSB0aGV5IGFyZSBhZGRyZXNzZWQuIElmIHlv
dSBoYXZlIHJlY2VpdmVkIHRoaXMgZW1haWwgaW4gZXJyb3IgcGxlYXNlIG5vdGlmeSB0aGUgb3Jp
Z2luYXRvciBvZiB0aGUgbWVzc2FnZS4gQW55IHZpZXdzIGV4cHJlc3NlZCBpbiB0aGlzIG1lc3Nh
Z2UgYXJlIHRob3NlIG9mIHRoZSBpbmRpdmlkdWFsIHNlbmRlci4NClRoaXMgbWVzc2FnZSBoYXMg
YmVlbiBzY2FubmVkIGZvciB2aXJ1c2VzIGFuZCBTcGFtIGJ5IFpURSBBbnRpLVNwYW0gc3lzdGVt
Lg0K
--=_alternative 0012F0CF4825784F_=
Content-Type: text/html; charset="GB2312"
Content-Transfer-Encoding: base64
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--=_alternative 0012F0CF4825784F_=--


From gao.yang2@zte.com.cn  Wed Mar  9 19:31:17 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 628803A67F7 for ; Wed,  9 Mar 2011 19:31:17 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -93.592
X-Spam-Level: 
X-Spam-Status: No, score=-93.592 tagged_above=-999 required=5 tests=[AWL=-3.643, BAYES_20=-0.74, CHARSET_FARAWAY_HEADER=3.2, CN_SUBJECT_136=0.982, HTML_MESSAGE=0.001, MIME_8BIT_HEADER=0.3, MIME_BASE64_TEXT=1.753, MIME_CHARSET_FARAWAY=2.45, RCVD_DOUBLE_IP_LOOSE=0.76, SARE_SUB_ENC_GB2312=1.345, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cI5yXQkIGA2U for ; Wed,  9 Mar 2011 19:31:14 -0800 (PST)
Received: from mx5.zte.com.cn (mx6.zte.com.cn [63.218.89.70]) by core3.amsl.com (Postfix) with ESMTP id 89F973A67D6 for ; Wed,  9 Mar 2011 19:31:12 -0800 (PST)
Received: from [10.34.0.130] by mx5.zte.com.cn with surfront esmtp id 3510992332426; Thu, 10 Mar 2011 11:30:08 +0800 (CST)
Received: from [10.30.3.21] by [192.168.168.15] with StormMail ESMTP id 84746.1401481301; Thu, 10 Mar 2011 11:31:32 +0800 (CST)
Received: from notes_smtp.zte.com.cn ([10.30.1.239]) by mse02.zte.com.cn with ESMTP id p2A3VFSP006996; Thu, 10 Mar 2011 11:31:15 +0800 (GMT-8) (envelope-from gao.yang2@zte.com.cn)
In-Reply-To: 
To: gao.yang2@zte.com.cn
MIME-Version: 1.0
X-KeepSent: 220346F6:1EF353B2-4825784F:0012D42F; type=4; name=$KeepSent
X-Mailer: Lotus Notes Release 6.5.6 March 06, 2007
Message-ID: 
From: gao.yang2@zte.com.cn
Date: Thu, 10 Mar 2011 11:27:59 +0800
X-MIMETrack: Serialize by Router on notes_smtp/zte_ltd(Release 8.5.1FP4|July 25, 2010) at 2011-03-10 11:31:16, Serialize complete at 2011-03-10 11:31:16
Content-Type: multipart/alternative; boundary="=_alternative 0013673B4825784F_="
X-MAIL: mse02.zte.com.cn p2A3VFSP006996
Cc: "dispatch@ietf.org" 
Subject: [dispatch] =?gb2312?b?U29ycnkgZm9yIHJlc3BvbnNlIHRoZSB3cm9uZyBt?= =?gb2312?b?YWlsIC8vIEY3tcRSRkO3orK80MLOxbjlo6y5qbLOv7wgICAvL7TwuLQ6IFJl?= =?gb2312?b?OiAgU0lQIGxvYWQgYmFsYW5jaW5n?=
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 03:31:17 -0000

This is a multipart message in MIME format.
--=_alternative 0013673B4825784F_=
Content-Type: text/plain; charset="GB2312"
Content-Transfer-Encoding: base64

PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0NCiBaaXAgICAgOiAyMTAwMTINCiBU
ZWwgICAgOiA4NzIxMQ0KIFRlbDIgICA6KCs4NiktMDI1LTUyODc3MjExDQogZV9tYWlsIDogZ2Fv
LnlhbmcyQHp0ZS5jb20uY24NCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09DQoN
Cg0KDQpnYW8ueWFuZzJAenRlLmNvbS5jbiANCreivP7IyzogIGRpc3BhdGNoLWJvdW5jZXNAaWV0
Zi5vcmcNCjIwMTEtMDMtMTAgMTE6MjINCg0KytW8/sjLDQoiVmlqYXkgSy4gR3VyYmFuaSIgPHZr
Z0BiZWxsLWxhYnMuY29tPg0Ks63LzQ0KZGlzcGF0Y2gtYm91bmNlc0BpZXRmLm9yZywgImRpc3Bh
dGNoQGlldGYub3JnIiA8ZGlzcGF0Y2hAaWV0Zi5vcmc+DQrW98ziDQpbZGlzcGF0Y2hdIEY3tcRS
RkO3orK80MLOxbjlo6y5qbLOv7wgICAvL7TwuLQ6IFJlOiAgU0lQIGxvYWQgYmFsYW5jaW5nDQoN
Cg0KDQoNCg0KDQoNCmh0dHA6Ly90ZWNoLjE2My5jb20vMDkvMDEyMS8xNC81MDZJTVJGTDAwMDkx
NUJFLmh0bWwgDQoNCmh0dHA6Ly90ZWNoLnFxLmNvbS9hLzIwMDkwMTIxLzAwMDIwNC5odG0gDQoN
Cmh0dHA6Ly9pdC5zb2h1LmNvbS8yMDA5MDEyMS9uMjYxODgzODE3LnNodG1sIA0KDQo9PT09PT09
PT09PT09PT09PT09PT09PT09PT09PT09PT09PQ0KWmlwICAgIDogMjEwMDEyDQpUZWwgICAgOiA4
NzIxMQ0KVGVsMiAgIDooKzg2KS0wMjUtNTI4NzcyMTENCmVfbWFpbCA6IGdhby55YW5nMkB6dGUu
Y29tLmNuDQo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PSANCg0KDQoiVmlqYXkg
Sy4gR3VyYmFuaSIgPHZrZ0BiZWxsLWxhYnMuY29tPiANCreivP7IyzogIGRpc3BhdGNoLWJvdW5j
ZXNAaWV0Zi5vcmcgDQoyMDExLTAzLTA5IDA1OjAxIA0KDQoNCsrVvP7Iyw0KcGF1bGVqQHBhY2tl
dGl6ZXIuY29tIA0Ks63LzQ0KImRpc3BhdGNoQGlldGYub3JnIiA8ZGlzcGF0Y2hAaWV0Zi5vcmc+
IA0K1vfM4g0KUmU6IFtkaXNwYXRjaF0gU0lQIGxvYWQgYmFsYW5jaW5nDQoNCg0KDQoNCg0KDQoN
Cg0KT24gMDMvMDcvMjAxMSAxMTowNSBQTSwgUGF1bCBFLiBKb25lcyB3cm90ZToNCj4gVmlqYXks
DQo+DQo+IFRoZSBTZXNzaW9uIENhcGFjaXR5IEVzdGltYXRlIChTQ0UpIGlkZWEgaXMgYWN0dWFs
bHkgcXVpdGUgc2ltaWxhcg0KPiB0byB0aGUgImFkYXB0aXZlIGxvYWQgYmFsYW5jaW5nIiBkcmFm
dCwgZnJvbSB3aGF0IEkgY2FuIHNlZS4NCg0KUGF1bDogVGhhbmsgeW91IG11Y2ggZm9yIHlvdXIg
aW5wdXQuICBNb3JlIGlubGluZS4NCg0KPiBTQ0Ugd2FzIHByb3Bvc2VkIGFzIGEgc29sdXRpb24g
dG8gdGhlICJvdmVybG9hZCIgcHJvYmxlbSwgdGhlDQo+IGFwcHJvYWNoIGJlaW5nIG9uZSB3aGVy
ZSB3ZSB0cnkgdG8gYXZvaWQgb3ZlcmxvYWQgYnkgc2VsZWN0aW5nIGENCj4gZGV2aWNlIHRoYXQg
cmVwb3J0cyBhIGhpZ2hlciBjYXBhY2l0eSB2YWx1ZS4gVGh1cywgb25lIGNvdWxkIHZpZXcgaXQN
Cj4gYXMgIm92ZXJsb2FkIGNvbnRyb2wgdGhyb3VnaCBsb2FkIGJhbGFuY2luZyIuDQoNCkkgdGhp
bmsgb3ZlcmxvYWQgY29udHJvbCBhbmQgbG9hZCBiYWxhbmNpbmcgYXJlIGRpc3RpbmN0LCBldmVu
IHRob3VnaA0KaXQgaXMgbmF0dXJhbCB0byByZWxhdGUgdGhlbS4gIExvYWQgYmFsYW5jaW5nIGNh
biBkZWxheSwgb3IgcHJldmVudCwNCm92ZXJsb2FkOyBidXQgb25jZSBhIFNJUCBob3N0IGlzIG92
ZXJsb2FkZWQsIGxvYWQgYmFsYW5jaW5nIGJ5IGl0c2VsZg0Kd2lsbCBub3QgaGVscCB0b28gbXVj
aC4NCg0KSSB0aGluayBpdCBpcyBiZXN0IHRoYXQgd2UgdHJlYXQgdGhlbSBzZXBhcmF0ZWx5LCBp
ZiBub3QgZm9yIGFueXRoaW5nDQplbHNlIHRoYW4gdGhlIGZhY3QgdGhhdCBTT0MgaXMgbm90IGNo
YXJ0ZXJlZCB0byBjcmVhdGUgYSBzb2x1dGlvbiBmb3INCmxvYWQgYmFsYW5jaW5nLiAgV2UgKGku
ZS4sIHRoZSBkaXNwYXRjaCBXRykgY2FuIGRlY2lkZSB0byBzZW5kIHRoaXMNCndvcmsgdG8gU09D
LCBidXQgSSB0aGluayB0aGF0IGRlY2lzaW9uIGlzIGEgYml0IHByZW1hdHVyZSBhdCB0aGlzDQpw
b2ludC4gIEZ1cnRoZXJtb3JlLCBJIHVuZGVyc3RhbmQgdGhhdCBiZWZvcmUgdGhlIGZvcm1hdGlv
biBvZiB0aGUNClNPQyBXRywgdGhlcmUgd2FzIHNvbWUgZGViYXRlIG9uIHdoZXRoZXIgb3Igbm90
IHRvIGluY2x1ZGUgbG9hZC0NCmJhbGFuY2luZyBpbiB0aGUgY2hhcnRlciBhbmQgaXQgd2FzIGRl
Y2lkZWQgdGhhdCBsb2FkLWJhbGFuY2luZyB3aWxsIG5vdA0KYmUgaW5jbHVkZWQuDQoNCkkgYmVs
aWV2ZSB0aGF0IHdlIHNob3VsZCB0YWtlIHRoZSB2aWV3IHRoYXQgbG9hZCBiYWxhbmNpbmcgYW5k
IG92ZXJsb2FkLQ0KY29udHJvbCBhcmUgc2VwYXJhdGUsIGFuZCB0aGF0IHdoZW4gYSBjbHVzdGVy
IG9mIFNJUCBzZXJ2ZXJzIGlzIHB1dCBpbg0Kc2VydmljZSwgYWxsIG9mIHRoZW0gc3RhcnQgb2Zm
IHByb3ZpZGluZyBsb2FkIGJhbGFuY2luZyBpbmZvcm1hdGlvbiB0bw0KdGhlIHVwc3RyZWFtIFNJ
UCBwcm94eS4gIFRoZSB1cHN0cmVhbSBTSVAgcHJveHkgd2lsbCBkbyBsb2FkIGJhbGFuY2luZw0K
dG8gdHJ5IHRvIHNwcmVhZCB0aGUgbG9hZCBwcm9wb3J0aW9uYWxseSB0byB0aGUgcmVzb3VyY2Vz
IGF2YWlsYWJsZSBhdA0KdGhlIGRvd25zdHJlYW0gU0lQIHNlcnZlcnMgKG5vdCBldmVubHksIHBy
b3BvcnRpb25hbGx5KS4gIEFzIHRpbWUNCnByb2dyZXNzZXMsIHNvbWUgb2YgdGhlbSB3aWxsIGVu
dGVyIGFuIG92ZXJsb2FkIHN0YXRlIGFuZCBhdCB0aGlzIHRpbWUsDQp0aGUgb3ZlcmxvYWQgY29u
dHJvbCBraWNrcyBpbi4gIFdoZW4gdGhlIG92ZXJsb2FkIGNvbmRpdGlvbiBpcw0KYWxsZXZpYXRl
ZCwgdGhlIG92ZXJsb2FkZWQgcGVlciBnb2VzIGJhY2sgdG8gcmVwb3J0aW5nIGl0cyBsb2FkLg0K
DQpOb3cgLi4uIHRvIHRoZSBkZXRhaWxzIG9mICJzY2UiIGFuZCAid2VpZ2h0IiBhcyBwcm9wb3Nl
ZCBpbiBvdXIgZHJhZnQuDQoNCkluIHRoZSBsYXJnZSwgSSBhZ3JlZSB0aGF0IHdlIGFyZSB0aGlu
a2luZyBhbG9uZyB0aGUgc2FtZSBsaW5lcy4gIFRoaXMNCmlzIGdvb2QuDQoNCkluIGFkZGl0aW9u
IHRvIHRoZSBlbWFpbCBleGNoYW5nZSBiZXR3ZWVuIHlvdSBhbmQgVGhpZXJyeSwgdGhlcmUNCmlz
IG9uZSBtb3JlIHByb3BlcnR5IHRoYXQgaXMgdmVyeSBhdHRyYWN0aXZlOiBzdGF0ZWxlc3NuZXNz
Lg0KDQpJIHRoaW5rIHRoYXQgaXQgaXMgYWR2YW50YWdlb3VzIHRoYXQgd2UgaW1wb3NlIHRoZSBw
cm9wZXJ0eSBvZg0Kc3RhdGVsZXNzbmVzcyBvbiBhIGxvYWQgYmFsYW5jaW5nIG1lY2hhbmlzbS4g
IFdpdGggInNjZSIsDQp0aGUgdXBzdHJlYW0gU0lQIHByb3h5IG11c3Qga2VlcCB0cmFjayBvZiBo
b3cgbWFueSByZXF1ZXN0cyBpdA0KaGFzIHNlbnQgdG8gYSBkb3duc3RyZWFtIFNJUCBzZXJ2ZXIg
c2luY2UgdGhlIGxhdHRlciBzZW50IGl0IHRoZQ0KbGFzdCAic2NlIiB2YWx1ZS4gIEZ1cnRoZXJt
b3JlLCBlYWNoIHRpbWUgYSByZXF1ZXN0IGlzIHNlbnQgdG8gdGhlDQpkb3duc3RyZWFtIFNJUCBz
ZXJ2ZXIsIHRoZSBTSVAgcHJveHkgbXVzdCBkZWNyZW1lbnQgaXRzIHJlcXVlc3QgY291bnQuDQpD
b25zZXF1ZW50bHksIHRoZSBzdGF0ZSBvbiBob3cgbWFueSByZXF1ZXN0cyBoYXZlIGJlZW4gc2Vu
dCB0byB0aGUNCmRvd25zdHJlYW0gU0lQIHNlcnZlciBzaW5jZSB0aGUgU0lQIHByb3h5IGdvdCB0
aGUgbGFzdCAic2NlIiBwYXJhbWV0ZXIsDQp3aGF0IHRoZSBsb3cgd2F0ZXItbWFyayBpcywgaXMg
YWxsIGxvc3QgaWYgbm90IHJlcGxpY2F0ZWQgdXBvbiBhDQpyZWJvb3QuDQoNCldpdGggIndlaWdo
dCIsIHlvdSBieXBhc3MgYWxsIG9mIHRoaXMgc3RhdGUgbWFpbnRlbmFuY2UgYnVzaW5lc3MuDQoN
CkFsbCB0aGUgU0lQIHByb3h5IG5lZWRzIGlzIGEgIndlaWdodCIgaW5kaWNhdG9yIHRoYXQgaXQg
d2lsbCB1c2UNCmluIGEgUkZDMjc4Mi10eXBlIGFsZ29yaXRobSB0byBwcm9wb3J0aW9uYWxseSBz
ZW5kIGFzIG11Y2ggdHJhZmZpYw0KdG8gYSBkb3duc3RyZWFtIFNJUCBzZXJ2ZXIgYXMgaXQgaGFz
IGluZGljYXRlZCAoaS5lLiwgaGlnaGVyICJ3ZWlnaHQiDQp2YWx1ZXMgd2lsbCBnZXQgcHJvcG9y
dGlvbmFsbHkgbW9yZSB0cmFmZmljKS4gIFRoaXMgaXMgbXVjaCBiZXR0ZXINCnRoYW4gcm91bmQt
cm9iaW4sIGFuZCBmdXJ0aGVybW9yZSwgdGhlcmUgaXMgbm8gcmVxdWlyZW1lbnQgdGhhdCBlYWNo
DQpkb3duc3RyZWFtIFNJUCBzZXJ2ZXIgaW4gdGhlIGNsdXN0ZXIgc2V0IGl0cyAid2VpZ2h0IiBp
biBhIGNvbnNpc3RlbnQNCm1hbm5lciwgb3IgZXZlbiB0aGF0IHRoZSAid2VpZ2h0IiB2YWx1ZXMg
b2YgdHdvIFNJUCBzZXJ2ZXJzIGJlDQpjb21wYXJhYmxlLg0KDQpJZiBhIFNJUCBzZXJ2ZXIgdGhp
bmtzIHRoYXQgaXQgaGFzIGluZmluaXRlIHJlc291cmNlcywgaXQNCmNhbiBzZXQgdGhlICJ3ZWln
aHQiIHRvIGJlIDY1LDUzNS4gIENvbnNlcXVlbnRseSwgaXQgd2lsbCBnZXQgYQ0KcHJvcG9ydGlv
bmFsbHkgbW9yZSB0cmFmZmljIGlmIHRoZSAid2VpZ2h0IiBvZiBvdGhlciBTSVAgc2VydmVycw0K
aW4gdGhlIGNsdXN0ZXIgaXMgc2V0IHRvLCBzYXksIDEwIG9yIDIwLiAgU2ltaWxhcmx5LCBpZiB0
aGUgU0lQIHNlcnZlcg0KaW4gYSBjbHVzdGVyIHdhbnRzIHRvIGFjY2VwdCBsZXNzIG51bWJlciBv
ZiBtZXNzYWdlcywgaXQgc2V0cyBhDQpsb3dlciwgYWxiZWl0IG5vbi16ZXJvLCB2YWx1ZSBmb3Ig
aXRzICJ3ZWlnaHQiIHBhcmFtZXRlci4gICBJZiB0aGUNCmNsdXN0ZXIgaGFzIGNhcGFjaXR5LCBv
dGhlciBsb2FkIGNvbnRyb2wgd2VpZ2h0cyB3aWxsIGJlIGhpZ2hlciwgYW5kDQp3aWxsIHRodXMg
Z2V0IGEgcHJvcG9ydGlvbmF0ZSBzaGFyZSBvZiB0aGUgdHJhZmZpYy4gIEFzIGFsbCBTSVAgc2Vy
dmVycw0KaW4gdGhlIGNsdXN0ZXIgcmVhY2ggY2FwYWNpdHksIHRoZSBsb2FkIGRpc3RyaWJ1dGlv
biBhbGdvcml0aG0gYXJyYW5nZXMNCmZvciBhIHVuaWZvcm0gbG9hZCBkaXN0cmlidXRpb24gdG8g
YWxsIG9mIHRoZSBTSVAgc2VydmVycyB1bnRpbCBlaXRoZXINCnRoZSBvZmZlcmVkIGxvYWQgc3Vi
c2lkZXMgb3IgdGhlIFNJUCBzZXJ2ZXJzIGVudGVyIGludG8gb3ZlcmxvYWQgbW9kZQ0KYW5kIHN0
YXJ0IHNoZWRkaW5nIHRyYWZmaWMgdXNpbmcgdGhlIG1lY2hhbmlzbSBkZWZpbmVkIGluIHRoZSBv
dmVybG9hZA0KY29udHJvbCBkcmFmdC4NCg0KPiBJIGRvbid0IHRoaW5rIGxvY2FsIGRlY2lzaW9u
cyBpbiB0aGUgYm94IG5lZWQgdG8gYmUgc3ViamVjdCB0bw0KPiBzdGFuZGFyZGl6YXRpb246IHdl
IGp1c3QgbmVlZCB0byBzdGFuZGFyZGl6ZSB0aGUgc2lnbmFsaW5nIHRvIGVuYWJsZQ0KPiBzdWNo
IGRlY2lzaW9ucy4NCg0KSSBjb3VsZCBub3QgYWdyZWUgbW9yZS4gIExvY2FsIGRlY2lzaW9ucyBv
biBob3cgZWFjaCBTSVAgc2VydmVyIGluIGENCmNsdXN0ZXIgZ2VuZXJhdGVzIGEgIndlaWdodCIg
dmFsdWUgYXJlIG5vdCBzdWJqZWN0IHRvIHN0YW5kYXJkaXphdGlvbi4NClRyYW5zcG9ydGluZyB0
aGlzIHZhbHVlIGFuZCBzcGVjaWZ5aW5nIGEgZGVmYXVsdCBhbGdvcml0aG0gdG8gYWxsb3cNCmEg
U0lQIHByb3h5IHRvIHBlcmZvcm0gbG9hZCBiYWxhbmNpbmcgc2VlbSB0byBiZSBhIGdvb2QgYXZl
bnVlIGZvcg0Kc3RhbmRhcmRpemF0aW9uLg0KDQpOb3RlIHRoYXQgSSB0aGluayBzcGVjaWZ5aW5n
IGEgZGVmYXVsdCBhbGdvcml0aG0gdG8gZG8gbG9hZCBiYWxhbmNpbmcNCmlzIGltcG9ydGFudCwg
YW5kIG91ciBkcmFmdCBjaG9vc2VzIHRvIHVzZSB0aGUgd2VsbC11bmRlcnN0b29kIFJGQzI3ODIN
CnVzYWdlIG9mICJ3ZWlnaHQiIHRvIGRvIHRoaXMuICBUaGlzIGRvZXMgbm90IHByZWNsdWRlIGFu
eSBpbXBsZW1lbnRlcg0KZnJvbSBjb21pbmcgdXAgd2l0aCB0aGVpciBvd24gYWxnb3JpdGhtIGFu
ZCB1c2luZyB0aGF0IGluc3RlYWQuDQoNCj4gSW4gYW55IGNhc2UsIGl0IGNlcnRhaW5seSBsb29r
cyBsaWtlIHdlJ3JlIHRoaW5raW5nIGFsb25nIHRoZSBzYW1lDQo+IGxpbmVzLg0KDQpJbmRlZWQu
ICBJZiB0aGlzIHdvcmsgcHJvZ3Jlc3NlcywgaXQnbGwgYmUgbmljZSB0byBoYXZlIGEgY29oZXJl
bnQNCnZpZXcgZ29pbmcgZm9yd2FyZC4gIElmIHlvdSB3aWxsIGJlIGluIFByYWd1ZSwgaXQgd2ls
bCBiZSBnb29kIHRvDQpob29rIHVwIGFuZCBjaGF0IGFib3V0IHRoaXMuDQoNCk9uY2UgYWdhaW4s
IHRoYW5rcyBmb3IgeW91ciBpbnB1dC4NCg0KLSB2aWpheQ0KLS0gDQpWaWpheSBLLiBHdXJiYW5p
LCBCZWxsIExhYm9yYXRvcmllcywgQWxjYXRlbC1MdWNlbnQNCjE5NjAgTHVjZW50IExhbmUsIFJt
LiA5Qy01MzMsIE5hcGVydmlsbGUsIElsbGlub2lzIDYwNTY2IChVU0EpDQpFbWFpbDogdmtnQHti
ZWxsLWxhYnMuY29tLGFjbS5vcmd9IC8gdmlqYXkuZ3VyYmFuaUBhbGNhdGVsLWx1Y2VudC5jb20N
CldlYjogICBodHRwOi8vZWN0LmJlbGwtbGFicy5jb20vd2hvL3ZrZy8NCl9fX19fX19fX19fX19f
X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fDQpkaXNwYXRjaCBtYWlsaW5nIGxpc3QN
CmRpc3BhdGNoQGlldGYub3JnDQpodHRwczovL3d3dy5pZXRmLm9yZy9tYWlsbWFuL2xpc3RpbmZv
L2Rpc3BhdGNoDQoNCg0KDQoNCi0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tLS0tDQpaVEUgSW5mb3JtYXRpb24gU2VjdXJpdHkgTm90aWNlOiBUaGUg
aW5mb3JtYXRpb24gY29udGFpbmVkIGluIHRoaXMgbWFpbCBpcyANCnNvbGVseSBwcm9wZXJ0eSBv
ZiB0aGUgc2VuZGVyJ3Mgb3JnYW5pemF0aW9uLiBUaGlzIG1haWwgY29tbXVuaWNhdGlvbiBpcyAN
CmNvbmZpZGVudGlhbC4gUmVjaXBpZW50cyBuYW1lZCBhYm92ZSBhcmUgb2JsaWdhdGVkIHRvIG1h
aW50YWluIHNlY3JlY3kgYW5kIA0KYXJlIG5vdCBwZXJtaXR0ZWQgdG8gZGlzY2xvc2UgdGhlIGNv
bnRlbnRzIG9mIHRoaXMgY29tbXVuaWNhdGlvbiB0byANCm90aGVycy4NClRoaXMgZW1haWwgYW5k
IGFueSBmaWxlcyB0cmFuc21pdHRlZCB3aXRoIGl0IGFyZSBjb25maWRlbnRpYWwgYW5kIGludGVu
ZGVkIA0Kc29sZWx5IGZvciB0aGUgdXNlIG9mIHRoZSBpbmRpdmlkdWFsIG9yIGVudGl0eSB0byB3
aG9tIHRoZXkgYXJlIGFkZHJlc3NlZC4gDQpJZiB5b3UgaGF2ZSByZWNlaXZlZCB0aGlzIGVtYWls
IGluIGVycm9yIHBsZWFzZSBub3RpZnkgdGhlIG9yaWdpbmF0b3Igb2YgDQp0aGUgbWVzc2FnZS4g
QW55IHZpZXdzIGV4cHJlc3NlZCBpbiB0aGlzIG1lc3NhZ2UgYXJlIHRob3NlIG9mIHRoZSANCmlu
ZGl2aWR1YWwgc2VuZGVyLg0KVGhpcyBtZXNzYWdlIGhhcyBiZWVuIHNjYW5uZWQgZm9yIHZpcnVz
ZXMgYW5kIFNwYW0gYnkgWlRFIEFudGktU3BhbSANCnN5c3RlbS4NCl9fX19fX19fX19fX19fX19f
X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fDQpkaXNwYXRjaCBtYWlsaW5nIGxpc3QNCmRp
c3BhdGNoQGlldGYub3JnDQpodHRwczovL3d3dy5pZXRmLm9yZy9tYWlsbWFuL2xpc3RpbmZvL2Rp
c3BhdGNoDQoNCg0KDQoNCi0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tDQpaVEUgSW5mb3JtYXRpb24gU2VjdXJpdHkgTm90aWNlOiBUaGUgaW5m
b3JtYXRpb24gY29udGFpbmVkIGluIHRoaXMgbWFpbCBpcyBzb2xlbHkgcHJvcGVydHkgb2YgdGhl
IHNlbmRlcidzIG9yZ2FuaXphdGlvbi4gVGhpcyBtYWlsIGNvbW11bmljYXRpb24gaXMgY29uZmlk
ZW50aWFsLiBSZWNpcGllbnRzIG5hbWVkIGFib3ZlIGFyZSBvYmxpZ2F0ZWQgdG8gbWFpbnRhaW4g
c2VjcmVjeSBhbmQgYXJlIG5vdCBwZXJtaXR0ZWQgdG8gZGlzY2xvc2UgdGhlIGNvbnRlbnRzIG9m
IHRoaXMgY29tbXVuaWNhdGlvbiB0byBvdGhlcnMuDQpUaGlzIGVtYWlsIGFuZCBhbnkgZmlsZXMg
dHJhbnNtaXR0ZWQgd2l0aCBpdCBhcmUgY29uZmlkZW50aWFsIGFuZCBpbnRlbmRlZCBzb2xlbHkg
Zm9yIHRoZSB1c2Ugb2YgdGhlIGluZGl2aWR1YWwgb3IgZW50aXR5IHRvIHdob20gdGhleSBhcmUg
YWRkcmVzc2VkLiBJZiB5b3UgaGF2ZSByZWNlaXZlZCB0aGlzIGVtYWlsIGluIGVycm9yIHBsZWFz
ZSBub3RpZnkgdGhlIG9yaWdpbmF0b3Igb2YgdGhlIG1lc3NhZ2UuIEFueSB2aWV3cyBleHByZXNz
ZWQgaW4gdGhpcyBtZXNzYWdlIGFyZSB0aG9zZSBvZiB0aGUgaW5kaXZpZHVhbCBzZW5kZXIuDQpU
aGlzIG1lc3NhZ2UgaGFzIGJlZW4gc2Nhbm5lZCBmb3IgdmlydXNlcyBhbmQgU3BhbSBieSBaVEUg
QW50aS1TcGFtIHN5c3RlbS4NCg==
--=_alternative 0013673B4825784F_=
Content-Type: text/html; charset="GB2312"
Content-Transfer-Encoding: base64
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--=_alternative 0013673B4825784F_=--


From paulej@packetizer.com  Wed Mar  9 20:14:19 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BCF073A67F6 for ; Wed,  9 Mar 2011 20:14:19 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PySaVtJGFJET for ; Wed,  9 Mar 2011 20:14:02 -0800 (PST)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id 97E423A67AA for ; Wed,  9 Mar 2011 20:14:01 -0800 (PST)
Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2A4F9hN025678 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Wed, 9 Mar 2011 23:15:15 -0500
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299730515; bh=LBZTH+/MQwbCUto6m+NPbPSuT6VqMyrwFrFup5JxD7g=; h=From:To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type; b=iY789TcSxbOLSBA61Clrnob3XsM0Dubb4KMr9TC8RFGOTNsWpZGwe0aT67ej6vQZk rinkQk531qY3CuXUm9R0eSPfQYpu1lxeLPJuoAfGy6r8YuCX8HW7PlLiWJsTViPQgO bxH+tzHDMTtAwOLJkboBmx6ah5lSXdSUUa/UFZvA=
From: "Paul E. Jones" 
To: "'Thierry Bessis'" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com>
In-Reply-To: <4D77B831.7000404@alcatel-lucent.com>
Date: Wed, 9 Mar 2011 23:15:03 -0500
Message-ID: <026f01cbded9$be15f290$3a41d7b0$@packetizer.com>
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----=_NextPart_000_0270_01CBDEAF.D54541C0"
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3kCCnyKVgIDWSzZAfKA2HwChasBsJOD+w9g
Content-Language: en-us
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 04:14:19 -0000

This is a multipart message in MIME format.

------=_NextPart_000_0270_01CBDEAF.D54541C0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_001_0271_01CBDEAF.D54541C0"


------=_NextPart_001_0271_01CBDEAF.D54541C0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

Thierry,

 

I think there is more in common than not.

 

You compute a "weight" and we compute an "sce" that is, in my opinion, more
or less the same: it's a value that indicates how much traffic can be
directed toward the box relative to other peer devices.

 

You want to do "load balancing" whereas we want to do "overload prevention".
Preventing overload means distributing load across boxes.  We are not
proposing to perfectly balance the load, but the load is distributed
nonetheless.

 

Honestly, I just don't see a lot of difference, other than terminology.  The
algorithm employed by the device directing traffic, but that algorithm is
independent of the advertisement of a capacity/weight value.

 

Paul

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Wednesday, March 09, 2011 12:26 PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Thanks Paul for you detailed answer ! 

I've looked at your SCE proposal and your comments and my impression is that
our two proposals are actually very different both from their intent
perspective, and from their realization perspective. 

The intent of our load balancing proposal is clearly continuous load
balancing. Our suggested weight does try to figure any kind of remaining
capacity in terms of sessions: it only indicates the instant (slightly
smoothed) available resource, which integrates CPU, Memory and any other
constrained resource. For the kind of usage we have in mind, (and for
example), it does not matter if we have lot of sessions "left" when we have
very little CPU to register those sessions: it is optimal to temporary
offload this busy node rather than risking a CPU induced rejection or
failure. When the node will have more CPU available, it will take more
sessions, and may be even more than others, because it may have now more
memory than others. If this node never have more CPU available, then its
"session" capacity is actually unusable and the system correctly reflects
that. 

Also, the SOC was not intended for "dynamic" load distribution: it only
provides the feedback when some shedding is expected (So already one or
multiple nodes are about to be overloaded). Also, integrating the
information from both SOC and SCE to take the distribution decision seems
rather complicated to me. In our proposal, the distribution algorithm is
simple, never changes, and is independent from SOC. We selected a node with
weights always the same way, and we apply the SOC rejection when applicable
to this node.

Cordially, 

Thierry

On 08-Mar-2011 14:47, Paul E. Jones wrote: 

Thierry,

 

Please see my comments below in green:

 

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Paul, 

Please let me address some of your comments. 
I don't know what the SCE is, so please pardon me if I misunderstand
something. 

On 07-Mar-2011 23:05, Paul E. Jones wrote: 

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to the
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a
solution to the "overload" problem, the approach being one where we try to
avoid overload by selecting a device that reports a higher capacity value.
Thus, one could view it as "overload control through load balancing".  We
didn't explicitly call it load balancing, though, because we were not trying
to balance load (i.e., place the same load on all servers).  Rather, we just
wanted to direct traffic to devices that we knew had available capacity.

Agreed, the goal seems to be the same: avoiding or at least delaying
overload. 
And your other comment too: our proposal does try to balance the load on all
the servers. The rational is that it is very easy for each server to know
where it is in terms of resource usage (for instance: I'm 60 percent
resource loaded at this moment), and this is just the information you need
to balance the traffic. Now it seems (but it is not clear, see below)  that
the SCE "available capacity" is actually a real time feedback, so the
information would be very similar actually. (What's the difference ? )

 

PEJ: I think it is very similar.  SCE provides real-time feedback of the
available resource capacity of a peer device.  The intent is to convey a
single integer value that represents an estimate of the number of sessions
the peer device can handle.  Note, this is not "message per second" or
other, but the number of sessions.  The focus on SCE is on devices like
"call agents", "session border controllers", stateful SIP proxies, etc. 




 
In practice, the way SCE would be employed is the following:
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
   peer devices
 * The selector would direct traffic toward the peer devices in a
   round-robin fashion, taking into consideration the SCE value so
   as to prevent exhausting resources on any given peer device

I see a contradiction here: if it is round-robin, it does not take into
account the SCE value. 
Is that something more like a weighted round-robin algorithm ? 

 

PEJ: As I mention below, the algorithm used to select a peer device is not a
part of the SCE draft.  One can use any mechanism.  The SCE draft merely
advertises a session capacity estimate.  We've found round-robin to be a
simple approach that works, as it distributes load relatively uniformly, but
stops directing traffic toward devices that report an SCE value that is
below a specified threshold. Other input could be considered, but we
deliberately left the algorithm outside the scope of SCE.




 * Once the peer's SCE value dropped below a given "low-water mark",
   the peer device would be less favored and not given traffic until
   the SCE value increases above the low-water mark (OPTIONS "ping"
   messages may be used to determine the SCE value in the event that
   there is no other traffic)

Why would we wait for a lower mark before to influence the distribution ?
Also is that less favored, or given no traffic ? 

 

PEJ: With our use of SCE, we're not trying to "balance" the load, per se.
One certainly could do that, but in order to balance the load on peer
devices when they are out of balance, one must direct multiple new sessions
toward the same device.  That is not always desirable since new sessions
usually consume more resources temporarily than established sessions.  So,
we try to burden a peer device by throwing a number of new sessions toward
the peer by distributing them in a round-robin fashion.  Again, though, the
approach taken is outside the scope of the SCE draft.

 


The later has stability issues.  And why do we need ping Options when there
is no traffic ?
 If there is no traffic (or very little), should we care about optimal
distribution from this source ?

 

PEJ: The assumption is that a peer device might also have several peers.
So, while device A1 may not be directing traffic toward device B, perhaps
A2, A3, and A4 are.  We can use OPTIONS "ping" messages to get current SCE
information.






 * The current SOC overload control mechanism would be used to
   to prevent message overloading, as SCE only focuses on resource
   overloading (e.g., memory, CPU, disk space, DSPs)

I don't see this difference here (message overload versus resource overload
? ) . To me, SOC focuses on source shedding, while SCE seems to focus on
improved load distribution, just like our proposal.

 

PEJ: A peer device might be able to handle 18,000 new sessions, but perhaps
it can only handle them at a rate of 100 per second.  We would use SCE to
determine that a peer can handle new sessions, while we would use the
methods SOC is defining to prevent message overload so we do not send 200
INVITEs per second to the peer.  (Use of the SOC mechanism might prove to
work better than our current round-robin approach, but we've not tested it.)





 
Note that the "round-robin" approach may or may not be preferred and is not
required in our draft.  I've found this approach works pretty well, but one
may prefer to use a particular weighting that is more complex.  I don't
think local decisions in the box need to be subject to standardization: we
just need to standardize the signaling to enable such decisions.

I believe the signaling is not enough. The deep meaning of the feedback
parameter needs to be clarified too. 
For example, in our proposal, the feedback is simply the current utilization
rate of the node (very simple).

 

PEJ: That sounds similar to SCE.  An SCE-capable device advertises to its
peer how many new sessions it can handle given the number of sessions it is
currently handling. 

Also the expectation needs to be clarified because it has a very important
impact on the recommended behavior in case of overload. 
The solution that we propose constantly tries to balance load in the cluster
of servers. Therefore, in case the SOC shedding kicks in, there is NOT NEED
to repeat the message elsewhere in the cluster because it is balanced
already (So no other server has more capacity). This is a huge advantage for
the network stability as message repeat tends to spread the overload and
collapse the cluster. 




In any case, it certainly looks like we're thinking along the same lines.

Yes I agree. 




PEJ: If you haven't already, have a look at our draft:

 
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

 

It is now expired, but your draft suggests there is still life in this
approach.  We should look to harmonize them, perhaps.

 

Paul

 


Cordially, 

Thierry




 
Paul
 

-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Vijay K. Gurbani
Sent: Monday, March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Paul: Thank you for your input.  More inline.
 
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul Jones
has proposed - with the main difference being the numeric range of the
weight. Do I have that right?

 
I believe that Paul J.'s algorithm is more suited towards overload
control, not load balancing (and I will await Paul J.'s correction to my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing is to
*always* report the current load to give the upstream SIP proxy an
indication of the health of the SIP server in the cluster.  If one waits
until overload is imminent, then it may be too late since a sudden and
unanticipated burst of messages will put the SIP server in overload
rather quickly.
 

It isn't clear to me from your draft if you are specifying the
algorithm by which the relative weights are used to determine the
probability of a request going to one server or another.

 
Yes, same way as RFC2782 does it.
 

ISTM that it *does* matter. And it especially matters if there is an
implication that the upstream node's algorithm may depend on some
assumption or knowledge about the downstream nodes.

 
No knowledge, just a reported number that the upstream SIP proxy can
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
starts off with the SRV indicators in RFC2782, without any regard to
additional knowledge or assumption about downstream nodes.
dispatch-adaptive-load-balancing is trying to use the same model, except
render the weight to be sensitive to the current traffic in the cluster
in real-time.
 

And of course this method of reporting weights isn't sufficient to
drive a balancing algorithm that is sensitive to some requests
requiring more or different resources than others. (That seems to be a
real problem, though I don't know if we have a real solution for
it.)

 
Yes, that is a real problem, as is the problem of a SIP server in a
cluster that has a low weight and is thus not contacted with enough
frequency to merit more participation to the goodput.
 
It is well known in SIP that some methods are more onerous to process
than others --- from a proxying point of view, processing an INVITE with
forking and response aggregation is far more detrimental to resource
usage than is processing a BYE.  Some existing literature techniques
thus add a coefficient to the weight of sending an INVITE versus a BYE.
At this time, the process outlined in dispatch-adaptive-load-balancing
treats all requests the same, but we can definitely look at ways to
change this behaviour.
 
Thanks,
 
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} 
/ vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
 
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 

 
 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 


------=_NextPart_001_0271_01CBDEAF.D54541C0
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Thierry's signature
Thierry,
 
I think there is more in common than not.
 
You compute a “weight” and we compute an =
“sce” that is, in my opinion, more or less the same: =
it’s a value that indicates how much traffic can be directed =
toward the box relative to other peer devices.
 
You want to do “load balancing” whereas we want to do =
“overload prevention”.  Preventing overload means =
distributing load across boxes.  We are not proposing to perfectly =
balance the load, but the load is distributed =
nonetheless.
 
Honestly, I just don’t see a lot of difference, other than =
terminology.  The algorithm employed by the device directing =
traffic, but that algorithm is independent of the advertisement of a =
capacity/weight value.
 
Paul
 
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] =

Sent: Wednesday, March 09, 2011 12:26 PM
To: Paul =
E. Jones
Cc: dispatch@ietf.org
Subject: Re: =
[dispatch] SIP load balancing
 
Thanks Paul =
for you detailed answer ! 

I've looked at your SCE proposal and =
your comments and my impression is that our two proposals are actually =
very different both from their intent perspective, and from their =
realization perspective. 

The intent of our load balancing =
proposal is clearly continuous load balancing. Our suggested weight does =
try to figure any kind of remaining capacity in terms of sessions: it =
only indicates the instant (slightly smoothed) available resource, which =
integrates CPU, Memory and any other constrained resource. For the kind =
of usage we have in mind, (and for example), it does not matter if we =
have lot of sessions "left" when we have very little CPU to =
register those sessions: it is optimal to temporary offload this busy =
node rather than risking a CPU induced rejection or failure. When the =
node will have more CPU available, it will take more sessions, and may =
be even more than others, because it may have now more memory than =
others. If this node never have more CPU available, then its =
"session" capacity is actually unusable and the system =
correctly reflects that. 

Also, the SOC was not intended for =
"dynamic" load distribution: it only provides the feedback =
when some shedding is expected (So already one or multiple nodes are =
about to be overloaded). Also, integrating the information from both SOC =
and SCE to take the distribution decision seems rather complicated to =
me. In our proposal, the distribution algorithm is simple, never =
changes, and is independent from SOC. We selected a node with weights =
always the same way, and we apply the SOC rejection when applicable to =
this node.

Cordially, 

Thierry

On 08-Mar-2011 =
14:47, Paul E. Jones wrote: 
Thierry,
 
Please see =
my comments below in green:
 
 
From: Thierry Bessis [mailto:Thierry.Bessis@a=
lcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 =
PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Paul, =


Please let me address some of your comments. 
I don't know =
what the SCE is, so please pardon me if I misunderstand something. =


On 07-Mar-2011 23:05, Paul E. Jones wrote: =

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to =
the
"adaptive load balancing" draft, from =
what I can see.  SCE was proposed as =
a
solution to the "overload" problem, the =
approach being one where we try to
avoid overload =
by selecting a device that reports a higher capacity =
value.
Thus, one could view it as "overload =
control through load balancing".  =
We
didn't explicitly call it load balancing, =
though, because we were not trying
to balance load =
(i.e., place the same load on all servers).  Rather, we =
just
wanted to direct traffic to devices that we =
knew had available capacity.
Agreed, the goal seems to be the same: avoiding or at =
least delaying overload. 
And your other comment too: our proposal =
does try to balance the load on all the servers. The rational is that it =
is very easy for each server to know where it is in terms of resource =
usage (for instance: I'm 60 percent resource loaded at this moment), and =
this is just the information you need to balance the traffic. Now it =
seems (but it is not clear, see below)  that the SCE =
"available capacity" is actually a real time feedback, so the =
information would be very similar actually. (What's the difference ? =
)
 
PEJ: I think it is very similar.  SCE provides real-time =
feedback of the available resource capacity of a peer device.  The =
intent is to convey a single integer value that represents an estimate =
of the number of sessions the peer device can handle.  Note, this =
is not “message per second” or other, but the number of =
sessions.  The focus on SCE is on devices like “call =
agents”, “session border controllers”, stateful SIP =
proxies, etc. =



 
In practice, =
the way SCE would be employed is the following:
* =
SIP proxy, B2BUA, etc. (selector) would receive SCE capacity =
from
   peer =
devices
 * The selector would direct traffic toward =
the peer devices in a
   round-robin =
fashion, taking into consideration the SCE value =
so
   as to prevent exhausting resources =
on any given peer device
I see a =
contradiction here: if it is round-robin, it does not take into account =
the SCE value. 
Is that something more like a weighted round-robin =
algorithm ? 
 
PEJ: As I mention below, the algorithm used to select a peer device =
is not a part of the SCE draft.  One can use any mechanism.  =
The SCE draft merely advertises a session capacity estimate.  =
We’ve found round-robin to be a simple approach that works, as it =
distributes load relatively uniformly, but stops directing traffic =
toward devices that report an SCE value that is below a specified =
threshold. Other input could be considered, but we deliberately left the =
algorithm outside the scope of =
SCE.


 * Once the peer's SCE value =
dropped below a given "low-water =
mark",
   the peer device would be =
less favored and not given traffic =
until
   the SCE value increases above =
the low-water mark (OPTIONS =
"ping"
   messages may be used =
to determine the SCE value in the event =
that
   there is no other =
traffic)
Why would we wait for a =
lower mark before to influence the distribution ? Also is that less =
favored, or given no traffic ? 
 
PEJ: With our use of SCE, we’re not trying to =
“balance” the load, per se.  One certainly could do =
that, but in order to balance the load on peer devices when they are out =
of balance, one must direct multiple new sessions toward the same =
device.  That is not always desirable since new sessions usually =
consume more resources temporarily than established sessions.  So, =
we try to burden a peer device by throwing a number of new sessions =
toward the peer by distributing them in a round-robin fashion.  =
Again, though, the approach taken is outside the scope of the SCE =
draft.
 

The later has =
stability issues.  And why do we need ping Options when there is no =
traffic ?
 If there is no traffic (or very little), should we =
care about optimal distribution from this source ?
 
PEJ: The assumption is that a peer device might also have several =
peers.  So, while device A1 may not be directing traffic toward =
device B, perhaps A2, A3, and A4 are.  We can use OPTIONS =
“ping” messages to get current SCE =
information.



 * The current =
SOC overload control mechanism would be used =
to
   to prevent message overloading, as =
SCE only focuses on resource
   =
overloading (e.g., memory, CPU, disk space, DSPs)
I don't see this difference here (message overload =
versus resource overload ? ) . To me, SOC focuses on source shedding, =
while SCE seems to focus on improved load distribution, just like our =
proposal.
 
PEJ: A peer device might be able to handle 18,000 new sessions, but =
perhaps it can only handle them at a rate of 100 per second.  We =
would use SCE to determine that a peer can handle new sessions, while we =
would use the methods SOC is defining to prevent message overload so we =
do not send 200 INVITEs per second to the peer.  (Use of the SOC =
mechanism might prove to work better than our current round-robin =
approach, but we’ve not tested it.) =



 
Note that =
the "round-robin" approach may or may not be preferred and is =
not
required in our draft.  I've found this =
approach works pretty well, but one
may prefer to =
use a particular weighting that is more complex.  I =
don't
think local decisions in the box need to be =
subject to standardization: we
just need to =
standardize the signaling to enable such decisions.
I believe the signaling is not enough. The deep =
meaning of the feedback parameter needs to be clarified too. 
For =
example, in our proposal, the feedback is simply the current utilization =
rate of the node (very simple).
 
PEJ: That sounds similar to SCE.  An SCE-capable device =
advertises to its peer how many new sessions it can handle given the =
number of sessions it is currently handling. 

Also the =
expectation needs to be clarified because it has a very important impact =
on the recommended behavior in case of overload. 
The solution that =
we propose constantly tries to balance load in the cluster of servers. =
Therefore, in case the SOC shedding kicks in, there is NOT NEED to =
repeat the message elsewhere in the cluster because it is balanced =
already (So no other server has more capacity). This is a huge advantage =
for the network stability as message repeat tends to spread the overload =
and collapse the cluster. 


In any case, =
it certainly looks like we're thinking along the same =
lines.
Yes I agree. =



PEJ: If you haven’t already, have a look at our =
draft:
http://tools.ietf.org/html/draft-jones-sip-overlo=
ad-sce-00
 
It is now expired, but your draft suggests there is still life in =
this approach.  We should look to harmonize them, =
perhaps.
 
Paul
 

Cordially, =


Thierry


 
Paul
 
-----Original =
Message-----
From: dispatch-bounces@ietf.org =
[mailto:dispatch-bounces@ietf.or=
g] On
Behalf Of Vijay K. =
Gurbani
Sent: Monday, March 07, 2011 6:38 =
PM
To: dispatch@ietf.org
<=
pre>Subject: Re: [dispatch] SIP load =
balancing
 
Paul: Thank =
you for your input.  More =
inline.
 
On 03/07/2011 =
05:14 PM, Paul Kyzivat wrote:
Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul =
Jones
has proposed - with the main difference being =
the numeric range of the
weight. Do I have that =
right?
 
I =
believe that Paul J.'s algorithm is more suited towards =
overload
control, not load balancing (and I will =
await Paul J.'s correction to =
my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing =
is to
*always* report the current load to give the =
upstream SIP proxy an
indication of the health of =
the SIP server in the cluster.  If one =
waits
until overload is imminent, then it may be =
too late since a sudden and
unanticipated burst of =
messages will put the SIP server in overload
rather =
quickly.
 
It isn't clear to me =
from your draft if you are specifying the
algorithm =
by which the relative weights are used to determine =
the
probability of a request going to one server or =
another.
 
Ye=
s, same way as RFC2782 does =
it.
 
ISTM that it *does* =
matter. And it especially matters if there is =
an
implication that the upstream node's algorithm =
may depend on some
assumption or knowledge about =
the downstream =
nodes.
 
No =
knowledge, just a reported number that the upstream SIP proxy =
can
treat as a RFC2782-like weight indicator.  =
Today, rfc3263-resolution
starts off with the SRV =
indicators in RFC2782, without any regard =
to
additional knowledge or assumption about =
downstream nodes.
dispatch-adaptive-load-balancing =
is trying to use the same model, except
render the =
weight to be sensitive to the current traffic in the =
cluster
in =
real-time.
 
And of course this =
method of reporting weights isn't sufficient =
to
drive a balancing algorithm that is sensitive to =
some requests
requiring more or different resources =
than others. (That seems to be a
real problem, =
though I don't know if we have a real solution =
for
it.)
 
Yes, that is a real problem, as is the problem of a =
SIP server in a
cluster that has a low weight and =
is thus not contacted with enough
frequency to =
merit more participation to the =
goodput.
 
It is well =
known in SIP that some methods are more onerous to =
process
than others --- from a proxying point of =
view, processing an INVITE with
forking and =
response aggregation is far more detrimental to =
resource
usage than is processing a BYE.  Some =
existing literature techniques
thus add a =
coefficient to the weight of sending an INVITE versus a =
BYE.
At this time, the process outlined in =
dispatch-adaptive-load-balancing
treats all =
requests the same, but we can definitely look at ways =
to
change this =
behaviour.
 
Thanks,<=
/o:p>
 
- =
vijay
--
Vijay K. Gurbani, =
Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, =
Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: =
vkg@{bell-labs.com,acm.or=
g} / vijay.gurbani@alcatel-lu=
cent.com
Web:   http://ect.bell-labs.com/who/v=
kg/
____________________________________________=
___
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 =
;
_______________________________________________
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis =

  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucen=
t.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: 12 April - 14 April =
2011
Upcoming planned Vacation: None =

 
 
_______________________________________________=

dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis 
  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming =
planned Business Trip: 12 April - 14 April 2011
Upcoming =
planned Vacation: None =


------=_NextPart_001_0271_01CBDEAF.D54541C0--

------=_NextPart_000_0270_01CBDEAF.D54541C0
Content-Type: image/jpeg;
	name="image001.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
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------=_NextPart_000_0270_01CBDEAF.D54541C0--


From partr@cisco.com  Thu Mar 10 01:39:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D27DE3A68B7 for ; Thu, 10 Mar 2011 01:39:51 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.185
X-Spam-Level: 
X-Spam-Status: No, score=-10.185 tagged_above=-999 required=5 tests=[AWL=0.414, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id e-kbXu+CAIQx for ; Thu, 10 Mar 2011 01:39:50 -0800 (PST)
Received: from ams-iport-2.cisco.com (ams-iport-2.cisco.com [144.254.224.141]) by core3.amsl.com (Postfix) with ESMTP id 903113A67EF for ; Thu, 10 Mar 2011 01:39:50 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=3252; q=dns/txt; s=iport; t=1299750068; x=1300959668; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to:cc; bh=7lmjdmtorJhUc1s+1Cyk/kV1Te6wJ5dJLULiPBzQDTM=; b=RpXV5n/ahy+YMUS7puBTlkhB3cm1SBZTDK4W/+HGRw/aiDeVZyysD6VC TofCSZ2dyymnp0bFzmvZacJLzOar3MRYPP36QpyGFKmH0dqzf3GJsHPEf maTQ6w5Do/3teRvpSuFtM/nnyOtw+hF9gSl5qY16tXdo5hgtpvrxMNO3C o=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AtQAAHsreE2Q/khNgWdsb2JhbACYVI4sFAEBFiYlpTGcKYMOglQEhSKKcQ
X-IronPort-AV: E=Sophos;i="4.62,295,1297036800"; d="scan'208";a="21146142"
Received: from ams-core-4.cisco.com ([144.254.72.77]) by ams-iport-2.cisco.com with ESMTP; 10 Mar 2011 09:41:06 +0000
Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by ams-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2A9f4SH025823; Thu, 10 Mar 2011 09:41:07 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Thu, 10 Mar 2011 15:11:04 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Date: Thu, 10 Mar 2011 15:11:03 +0530
Message-ID: 
In-Reply-To: <4D76AD94.4000108@bell-labs.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd3/evYX895RcpSc+wMpjgqd0hoQBInvTQ
References: <4D755B9E.6030008@bell-labs.com> <4D7566C1.40605@cisco.com><4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>  <4D76AD94.4000108@bell-labs.com>
From: "Parthasarathi R (partr)" 
To: "Vijay K. Gurbani" 
X-OriginalArrivalTime: 10 Mar 2011 09:41:04.0112 (UTC) FILETIME=[45ED0F00:01CBDF07]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 09:39:51 -0000

Vijay,

Please read inline

Thanks
Partha=20

-----Original Message-----
From: Vijay K. Gurbani [mailto:vkg@bell-labs.com]=20
Sent: Wednesday, March 09, 2011 3:59 AM
To: Parthasarathi R (partr)
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

On 03/08/2011 03:52 AM, Parthasarathi R (partr) wrote:
>[...] Even though any overload control algorithm  is used ("oc" or=20
>"weight" or "SCE" or "resource-availability" other),  the locally=20
>overload control mechanism is required in downstream as per  mentioned=20
>in draft-ietf-soc-overload-design. I wish that we will look  into these

>aspect because overload control is the extreme case for load  balancing

>the cluster of downstream entity and load balancing mechanism  shall=20
>help to avoid the overload control based on feedback.

Partha: Thank you for your input.  I do agree that overload control is
the extreme case for load balancing.  However, as my email to Paul J.
indicated, treating them differently may help, or we could make load
balancing an extension under overload control.  I don't know what the
answer is ... we'd have to hash it out in dispatch on where to send load
balancing to.

 IMO, it is possible to have the single solution for overload
control which shall be used for load balancing as well in case the units
of the feedback are well defined and uniquely  identified across the
different devices. As Paul mentioned in the mail thread, it is worth to
have BAR BoF for this activity during IETF-80 in case lot of folks are
interested. 

  My understanding is that the charter for SOC does not account for load
balancing.

 It took long time for me to understand IETF SOC does not
considered load balancing  :-(...I agree with you now 

> I guess that your mechanism assumes about downstream SIP UA as=20
> signaling entity only and does not handle media. Please correct me in=20
> case I wrongly understand.

Our mechanism does not distinguish between media or signaling in the
sense that the "weight" that is reported is a normalized value that is
calculated by the SIP server in a cluster.=20

 It will be really tough to generalize all devices under
"Weight" as a unit. Because "weight" understanding will vary heavy based
on the implementation of a specific devices. Value 1000 in one device
may not be same as 1000 in another device which may confuse load
balancer 

The SIP server could be a SIP bridge, for instance, in which case it may
be able to handle a lot of point-to-point calls, but can only handle few
calls requiring full mesh conferencing capabilities.  As long as the SIP
server is handling the low-resource point-to-point calls, it reports a
higher "weight" value.  But as soon as a conference call with a large
number of participants commences, using other resources like mixers and
transcoders, etc., the SIP server starts to report lower "weight"
values.

Thanks,

- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent 1960 Lucent Lane,
Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/

From henry.sinnreich@gmail.com  Thu Mar 10 09:01:49 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C9F093A6B2F for ; Thu, 10 Mar 2011 09:01:49 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.052
X-Spam-Level: 
X-Spam-Status: No, score=-3.052 tagged_above=-999 required=5 tests=[AWL=0.547,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WwOEtmq1GpSJ for ; Thu, 10 Mar 2011 09:01:48 -0800 (PST)
Received: from mail-yw0-f44.google.com (mail-yw0-f44.google.com [209.85.213.44]) by core3.amsl.com (Postfix) with ESMTP id 4262F3A6B39 for ; Thu, 10 Mar 2011 09:01:48 -0800 (PST)
Received: by ywi6 with SMTP id 6so987020ywi.31 for ; Thu, 10 Mar 2011 09:03:06 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:mime-version:content-type :content-transfer-encoding; bh=/6MatvV/4G5N4oMWffdBC44o1kYupwk9B+e3lfWs0qw=; b=Sx+I8xepSIo4nPpBwxg7OBP1UHZXzovlSES2HhIWgvcvppR4HfgLEUXedEhGq9QzpK KulItPGi2irbnRb3iA5tXdYypoieVdxJVFTVETng9aJsGYEBmbnsd7p5azoaGkNmaYon B10iPIOpi5RW71KMP5r4K1tz5FP9/l36Q3yzE=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:mime-version:content-type:content-transfer-encoding; b=wmomhCdxcFy0+TqlhZA1Pz+UCZFrbjbZly8Lo+bLGsrr5QT/++dD1XqU8JdpSZNWmF vfROXzUsRsRkBa4c4cuuedYcuSJ6rR6c6UH5/Kl9AUktJBbwx8vPV+JFQRdTEpTVk4NG SVdKzx7bUpvwvLJL2jWEMnmDnnhIWRGEJ8zLs=
Received: by 10.151.25.11 with SMTP id c11mr1215054ybj.445.1299776585148; Thu, 10 Mar 2011 09:03:05 -0800 (PST)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id u37sm2014540yba.19.2011.03.10.09.03.03 (version=TLSv1/SSLv3 cipher=OTHER); Thu, 10 Mar 2011 09:03:04 -0800 (PST)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Thu, 10 Mar 2011 11:03:01 -0600
From: Henry Sinnreich 
To: Andrew Allen , , 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvfRQM5LNYNOSZVz0atfnKVY5yxgA==
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 17:01:49 -0000

Hi Andrew,

I have generic objections to see the publication of yet another
informational RFC about practices in consortiums of closed networks in the
IETF, for the following reasons:

* Using the RFC process as a publishing venue for organizations that don't
explicitly do that in support of the open e2e Internet.

* The resources of the IETF are not unlimited, meeting time and email time
to discuss is scarce and the IETF registries are just not scalable enough to
support registrations for all the extensions that benefit countless external
organizations and companies that deploy Internet technology.
New business plans and the required extensions pop up in an endless fashion.
Big organizations such as you mention have their own resources to publish
their internal technical documents and white papers if they desire to do so.

Members of outside organizations are certainly doing everybody a good
service by informing via email of new standards/extensions in other
organization, such as you have done below, but such information need not
overload the Internet RFC process nor overload the Internet registries.

* Hard for SIP developers and right out impossible for application
developers.
Though SIP works fine indeed across the world; SIPit and the SIP Forum
take care of unavoidable problems still encountered by operators, arguably
increasing SIP complexity for non-Internet purposes is not a good idea, even
if new SIP extensions are labeled as just "informational". It makes it even
more difficult for new developers entering the field to ever understand, let
alone write code for some or all that has been published already about SIP
for example. Checking right now there are 197 RFCs matching SIP and 78 I-Ds
found by the RFC search engine.

It is time to leave SIP alone and move on to other critical areas for the
Internet, such as for example the future of applications that will be
discussed in the next IETF technical plenary on Monday, March 28.

With this in mind, I would rather not go into discussing your I-D in detail,
though at first glance it seems technically correct and if it were discussed
on the list, may indeed technically qualify as an informational RFC - just
what I am arguing against.

These are all personal opinions. The chairs and ADs may or may not discuss
or agree/disagree with. Your I-D is just a good opportunity to raise these
issues. 

Note that I am not suggesting any voting on your I-D, if it will be
discussed, since for various reasons, the majority of folks attending the
DISPATCH WG are from various outside organizations with business plans that
may or may not fully embrace the open e2e Internet; a perfect legitimate
choice for them, but not  in the IETF (I just can't remember when ever there
was a vote against any new I-D for SIP, since no cost was perceived by those
voting).

Thanks,
Henry 


On 3/9/11 5:45 PM, "Andrew Allen"  wrote:

> 
> Henry
> 
> Would you care to elaborate and explain your specific concerns?  How is what
> is defined in these drafts potentially incompatible with the open, e2e
> architectural and in any other way technical conflict with the Internet?
> 
> The two drafts do the following things:
> 
> One draft simply defines and registers with IANA a URN namespace according to
> the procedures defined for doing that.
> 
> The other draft simply defines how this URN can be used as an instance ID as
> required by RFC 5626.
> 
> SIP and the mechanisms in RFC 5626 are already being used in networks defined
> by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA have
> already defined this behavior using their existing device Identifiers for
> instance ID generation in their specifications. These drafts complete the
> process doing things the right way by reserving the namespace and defining the
> usage of it using the IETF defined mechanisms for doing that.
> 
> Andrew
> 
> ----- Original Message -----
> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> Sent: Wednesday, March 09, 2011 03:10 PM
> To: Itsuma TANAKA ; dispatch@ietf.org
> 
> Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> 
> Such work in the IETF would require IMO an explanation about its
> compatibility with the open, e2e architectural and other technical
> principles of the Internet. It actually seems to be the opposite.
> 
> For this reason, such work may be better accomplished in other
> organizations, so as to avoid architectural principle and other technical
> conflicts. Simplicity and scalability comes to mind among other, though this
> is better discussed elsewhere.
> 
> Thanks,
> 
> Henry Sinnreich
> 
> 
> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
> 
>> Dear All,
>> 
>> I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP meetings (SA2,
>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming Expert
>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  I serve
>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>> 
>> I would like to express my support of draft-montemurro-gsma-imei-urn and the
>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>> 
>> The reasons are exactly same as what Ricky Kaura (Samsung) expressed.  A
>> various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and TS24.292), and
>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are using
>> these drafts.
>> 
>> These internet drafts are very important to support various key features in
>> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would like ask
>> that this work is progressed as quickly as possible in IETF.
>> 
>> 
>> Best regards,
>> 
>> Itsuma Tanaka
>> NTT DOCOMO
>> 
>> 
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
> 
> ---------------------------------------------------------------------
> This transmission (including any attachments) may contain confidential
> information, privileged material (including material protected by the
> solicitor-client or other applicable privileges), or constitute non-public
> information. Any use of this information by anyone other than the intended
> recipient is prohibited. If you have received this transmission in error,
> please immediately reply to the sender and delete this information from your
> system. Use, dissemination, distribution, or reproduction of this transmission
> by unintended recipients is not authorized and may be unlawful.






From lconroy@insensate.co.uk  Thu Mar 10 09:39:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4513D3A6810 for ; Thu, 10 Mar 2011 09:39:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.546
X-Spam-Level: 
X-Spam-Status: No, score=-2.546 tagged_above=-999 required=5 tests=[AWL=0.052,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id C8yLJnpY6C0F for ; Thu, 10 Mar 2011 09:39:58 -0800 (PST)
Received: from insensate.co.uk (ghost.insensate.co.uk [213.152.49.121]) by core3.amsl.com (Postfix) with ESMTP id 1EF393A65A6 for ; Thu, 10 Mar 2011 09:39:56 -0800 (PST)
Received: from localhost (localhost [127.0.0.1]) by insensate.co.uk (Postfix) with ESMTP id B1B4A27AA2E; Thu, 10 Mar 2011 17:41:13 +0000 (GMT)
X-Virus-Scanned: amavisd-new at insensate.co.uk
Received: from insensate.co.uk ([127.0.0.1]) by localhost (psyche.insensate.co.uk [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oFHaXBweItkd; Thu, 10 Mar 2011 17:41:13 +0000 (GMT)
Received: from [127.0.0.1] (localhost [127.0.0.1]) by insensate.co.uk (Postfix) with ESMTP id F08EB27AA23; Thu, 10 Mar 2011 17:41:12 +0000 (GMT)
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
From: Lawrence Conroy 
In-Reply-To: 
Date: Thu, 10 Mar 2011 17:41:12 +0000
Content-Transfer-Encoding: quoted-printable
Message-Id: <12CB5769-0242-4C3F-A5F0-AD65CA778556@insensate.co.uk>
References: 
To: Henry Sinnreich 
X-Mailer: Apple Mail (2.1082)
Cc: dispatch@ietf.org
Subject: [dispatch] IETF as an SDO (was Re: Revisionofdraft-allen-dispatch-imei-urn-as-instanceid)
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 17:39:59 -0000

Hi Henry,
 um ...
i. Votes in the IETF; did I miss this, or is this another change
   like actively barring entry to meetings for deadbeats?

ii. Information is good.
   I for one have no axe to grind on this, but believe that this is a =
service to the Internet community.
   Reporting how things work in the real world has been a long-term =
feature of the informational track (as you know).

iii. The fact that imei format is used and can be tied to URNs and IIDs =
seems to me to be both
   a valid and reasonable idea.
   *Requiring* that to be done is a different page of the book, but this =
is NOT a standards track doc.
   How GSMA spectrum members want to run their networks is not an IETF =
topic (viz CONDOR, ...).
   That war was decided waaaaay back -- you were there for that as well =
:).

   Your comments on lack of e2e operation (i.e. if this is not in SIP =
messages they will be
   treated "differently"/with extreme prejudice) is at the very least =
reasonable, but that
   addresses the general topic of the IETF as an SDO, rather than being =
tied to this draft.

iv. you say that SIP time should not be taken up with registrations -- =
that's the iid registration process.
  There's a registration doc that has (if one is fortunate) a "how this =
is used by xxx".
  So ... this should "go through on the nod". You have said that it =
appears to be technically valid.

  Q: Are you arguing that the iid registration process should be =
changed?

all the best,
 Lawrence
--
What comes up must come down.
"Not my department", says Werner Von Braun


On 10 Mar 2011, at 17:03, Henry Sinnreich wrote:
> Hi Andrew,
>=20
> I have generic objections to see the publication of yet another
> informational RFC about practices in consortiums of closed networks in =
the
> IETF, for the following reasons:
>=20
> * Using the RFC process as a publishing venue for organizations that =
don't
> explicitly do that in support of the open e2e Internet.
>=20
> * The resources of the IETF are not unlimited, meeting time and email =
time
> to discuss is scarce and the IETF registries are just not scalable =
enough to
> support registrations for all the extensions that benefit countless =
external
> organizations and companies that deploy Internet technology.
> New business plans and the required extensions pop up in an endless =
fashion.
> Big organizations such as you mention have their own resources to =
publish
> their internal technical documents and white papers if they desire to =
do so.
>=20
> Members of outside organizations are certainly doing everybody a good
> service by informing via email of new standards/extensions in other
> organization, such as you have done below, but such information need =
not
> overload the Internet RFC process nor overload the Internet =
registries.
>=20
> * Hard for SIP developers and right out impossible for application
> developers.
> Though SIP works fine indeed across the world; SIPit and the SIP Forum
> take care of unavoidable problems still encountered by operators, =
arguably
> increasing SIP complexity for non-Internet purposes is not a good =
idea, even
> if new SIP extensions are labeled as just "informational". It makes it =
even
> more difficult for new developers entering the field to ever =
understand, let
> alone write code for some or all that has been published already about =
SIP
> for example. Checking right now there are 197 RFCs matching SIP and 78 =
I-Ds
> found by the RFC search engine.
>=20
> It is time to leave SIP alone and move on to other critical areas for =
the
> Internet, such as for example the future of applications that will be
> discussed in the next IETF technical plenary on Monday, March 28.
>=20
> With this in mind, I would rather not go into discussing your I-D in =
detail,
> though at first glance it seems technically correct and if it were =
discussed
> on the list, may indeed technically qualify as an informational RFC - =
just
> what I am arguing against.
>=20
> These are all personal opinions. The chairs and ADs may or may not =
discuss
> or agree/disagree with. Your I-D is just a good opportunity to raise =
these
> issues.=20
>=20
> Note that I am not suggesting any voting on your I-D, if it will be
> discussed, since for various reasons, the majority of folks attending =
the
> DISPATCH WG are from various outside organizations with business plans =
that
> may or may not fully embrace the open e2e Internet; a perfect =
legitimate
> choice for them, but not  in the IETF (I just can't remember when ever =
there
> was a vote against any new I-D for SIP, since no cost was perceived by =
those
> voting).
>=20
> Thanks,
> Henry=20
>=20
>=20
> On 3/9/11 5:45 PM, "Andrew Allen"  wrote:
>=20
>>=20
>> Henry
>>=20
>> Would you care to elaborate and explain your specific concerns?  How =
is what
>> is defined in these drafts potentially incompatible with the open, =
e2e
>> architectural and in any other way technical conflict with the =
Internet?
>>=20
>> The two drafts do the following things:
>>=20
>> One draft simply defines and registers with IANA a URN namespace =
according to
>> the procedures defined for doing that.
>>=20
>> The other draft simply defines how this URN can be used as an =
instance ID as
>> required by RFC 5626.
>>=20
>> SIP and the mechanisms in RFC 5626 are already being used in networks =
defined
>> by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA =
have
>> already defined this behavior using their existing device Identifiers =
for
>> instance ID generation in their specifications. These drafts complete =
the
>> process doing things the right way by reserving the namespace and =
defining the
>> usage of it using the IETF defined mechanisms for doing that.
>>=20
>> Andrew
>>=20
>> ----- Original Message -----
>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
>> Sent: Wednesday, March 09, 2011 03:10 PM
>> To: Itsuma TANAKA ; dispatch@ietf.org
>> 
>> Subject: Re: [dispatch] =
Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>>=20
>> Such work in the IETF would require IMO an explanation about its
>> compatibility with the open, e2e architectural and other technical
>> principles of the Internet. It actually seems to be the opposite.
>>=20
>> For this reason, such work may be better accomplished in other
>> organizations, so as to avoid architectural principle and other =
technical
>> conflicts. Simplicity and scalability comes to mind among other, =
though this
>> is better discussed elsewhere.
>>=20
>> Thanks,
>>=20
>> Henry Sinnreich
>>=20
>>=20
>> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
>>=20
>>> Dear All,
>>>=20
>>> I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP =
meetings (SA2,
>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming =
Expert
>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  =
I serve
>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>>>=20
>>> I would like to express my support of draft-montemurro-gsma-imei-urn =
and the
>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>>>=20
>>> The reasons are exactly same as what Ricky Kaura (Samsung) =
expressed.  A
>>> various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and =
TS24.292), and
>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are =
using
>>> these drafts.
>>>=20
>>> These internet drafts are very important to support various key =
features in
>>> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would =
like ask
>>> that this work is progressed as quickly as possible in IETF.
>>>=20
>>>=20
>>> Best regards,
>>>=20
>>> Itsuma Tanaka
>>> NTT DOCOMO
>>>=20
>>>=20
>>> _______________________________________________
>>> dispatch mailing list
>>> dispatch@ietf.org
>>> https://www.ietf.org/mailman/listinfo/dispatch
>>=20
>>=20
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>>=20
>> ---------------------------------------------------------------------
>> This transmission (including any attachments) may contain =
confidential
>> information, privileged material (including material protected by the
>> solicitor-client or other applicable privileges), or constitute =
non-public
>> information. Any use of this information by anyone other than the =
intended
>> recipient is prohibited. If you have received this transmission in =
error,
>> please immediately reply to the sender and delete this information =
from your
>> system. Use, dissemination, distribution, or reproduction of this =
transmission
>> by unintended recipients is not authorized and may be unlawful.
>=20
>=20
>=20
>=20
>=20
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


From Thierry.Bessis@alcatel-lucent.com  Thu Mar 10 10:15:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1528D3A68CB for ; Thu, 10 Mar 2011 10:15:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ln0QHNtEmk-M for ; Thu, 10 Mar 2011 10:15:42 -0800 (PST)
Received: from ihemail4.lucent.com (ihemail4.lucent.com [135.245.0.39]) by core3.amsl.com (Postfix) with ESMTP id 4400B3A6906 for ; Thu, 10 Mar 2011 10:15:40 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail4.lucent.com (8.13.8/IER-o) with ESMTP id p2AIGswN028269 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 10 Mar 2011 12:16:54 -0600 (CST)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2AIGrYt008209; Thu, 10 Mar 2011 12:16:54 -0600 (CST)
Message-ID: <4D791595.2010703@alcatel-lucent.com>
Date: Thu, 10 Mar 2011 12:16:53 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Paul E. Jones" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com>
In-Reply-To: <026f01cbded9$be15f290$3a41d7b0$@packetizer.com>
Content-Type: multipart/alternative; boundary="------------020009070706080402000908"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.39
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 18:15:59 -0000

This is a multi-part message in MIME format.
--------------020009070706080402000908
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Paul,

I agree they are many similitudes ! :-)
But the question I think is: Can we make the two approaches compatible ?

For instance, with our approach, it is fundamental that the weight 
represents a "Relative" occupation, and not an absolute capacity.
The first reason is that we found that loading continuously all the 
servers at the same utilization rate (same percentage) is the optimal 
way of loading a cluster. (We can discuss that if not agreed). The 
second reason is that a relative occupation is very easy to derive from 
information already available: any system knows where it is in terms of 
CPU or Memory usage. In contrast, deriving a left capacity in terms of 
remaining sessions can be difficult: Host can be used by different 
applications and perform different duties, which can change over the 
day, all of them impacting their true capacity at a certain moment. 
Beside, as already discussed, different sessions require different 
resources, so there is a whole complication of dealing with a multitude 
of call profiles, that (again) are not even constant over the day.

Also, we do not want to derive some shedding decision from the weight. 
We want to rely on the OLC mechanism for that. As I mentioned before, 
deriving shedding from the weight is very sensitive to the way weights 
are calculated: we would need to specify exactly how the weight is 
calculated: this is impractical ! Also everybody would need to agree 
about specific threshold: What is the weight value threshold for 
shedding ?: Again this is more difficult, and different servers will 
certainly want to use different thresholds.   Also there is a 
contradiction: whether you are using relative or absolute weights, your 
weight should  zero (100% busy, or 0 sessions available) at the moment 
you are getting overloaded. So if you are 100% busy (or in average no 
room for new sessions), how to you tell how many sessions need to be 
shed ? You could be 100% busy with an offered load of 300% nominal (then 
you need to shed 66%), or be 100% busy at just the nominal load, (then 
you don't need to shed). This again shows this threshold issue I 
mentioned. In our implementation, we used to use to a common 
"utilization" information to do both optimal distribution and shedding, 
but we moved away from that because of those reasons: we found that each 
node indicating its requested rejection rate is much more robust and 
requires very little agreements.

Cordially,

Thierry


On 09-Mar-2011 22:15, Paul E. Jones wrote:
>
> Thierry,
>
> I think there is more in common than not.
>
> You compute a "weight" and we compute an "sce" that is, in my opinion, 
> more or less the same: it's a value that indicates how much traffic 
> can be directed toward the box relative to other peer devices.
>
> You want to do "load balancing" whereas we want to do "overload 
> prevention".  Preventing overload means distributing load across 
> boxes.  We are not proposing to perfectly balance the load, but the 
> load is distributed nonetheless.
>
> Honestly, I just don't see a lot of difference, other than 
> terminology.  The algorithm employed by the device directing traffic, 
> but that algorithm is independent of the advertisement of a 
> capacity/weight value.
>
> Paul
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Wednesday, March 09, 2011 12:26 PM
> *To:* Paul E. Jones
> *Cc:* dispatch@ietf.org
> *Subject:* Re: [dispatch] SIP load balancing
>
> Thanks Paul for you detailed answer !
>
> I've looked at your SCE proposal and your comments and my impression 
> is that our two proposals are actually very different both from their 
> intent perspective, and from their realization perspective.
>
> The intent of our load balancing proposal is clearly continuous load 
> balancing. Our suggested weight does try to figure any kind of 
> remaining capacity in terms of sessions: it only indicates the instant 
> (slightly smoothed) available resource, which integrates CPU, Memory 
> and any other constrained resource. For the kind of usage we have in 
> mind, (and for example), it does not matter if we have lot of sessions 
> "left" when we have very little CPU to register those sessions: it is 
> optimal to temporary offload this busy node rather than risking a CPU 
> induced rejection or failure. When the node will have more CPU 
> available, it will take more sessions, and may be even more than 
> others, because it may have now more memory than others. If this node 
> never have more CPU available, then its "session" capacity is actually 
> unusable and the system correctly reflects that.
>
> Also, the SOC was not intended for "dynamic" load distribution: it 
> only provides the feedback when some shedding is expected (So already 
> one or multiple nodes are about to be overloaded). Also, integrating 
> the information from both SOC and SCE to take the distribution 
> decision seems rather complicated to me. In our proposal, the 
> distribution algorithm is simple, never changes, and is independent 
> from SOC. We selected a node with weights always the same way, and we 
> apply the SOC rejection when applicable to this node.
>
> Cordially,
>
> Thierry
>
> On 08-Mar-2011 14:47, Paul E. Jones wrote:
>
> Thierry,
>
> Please see my comments below in green:
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Tuesday, March 08, 2011 12:35 PM
> *To:* Paul E. Jones
> *Cc:* 'Vijay K. Gurbani'; dispatch@ietf.org 
> *Subject:* Re: [dispatch] SIP load balancing
>
> Paul,
>
> Please let me address some of your comments.
> I don't know what the SCE is, so please pardon me if I misunderstand 
> something.
>
> On 07-Mar-2011 23:05, Paul E. Jones wrote:
>
> Vijay,
>   
> The Session Capacity Estimate (SCE) idea is actually quite similar to the
> "adaptive load balancing" draft, from what I can see.  SCE was proposed as a
> solution to the "overload" problem, the approach being one where we try to
> avoid overload by selecting a device that reports a higher capacity value.
> Thus, one could view it as "overload control through load balancing".  We
> didn't explicitly call it load balancing, though, because we were not trying
> to balance load (i.e., place the same load on all servers).  Rather, we just
> wanted to direct traffic to devices that we knew had available capacity.
>
> Agreed, the goal seems to be the same: avoiding or at least delaying 
> overload.
> And your other comment too: our proposal does try to balance the load 
> on all the servers. The rational is that it is very easy for each 
> server to know where it is in terms of resource usage (for instance: 
> I'm 60 percent resource loaded at this moment), and this is just the 
> information you need to balance the traffic. Now it seems (but it is 
> not clear, see below)  that the SCE "available capacity" is actually a 
> real time feedback, so the information would be very similar actually. 
> (What's the difference ? )
>
> PEJ: I think it is very similar.  SCE provides real-time feedback of 
> the available resource capacity of a peer device.  The intent is to 
> convey a single integer value that represents an estimate of the 
> number of sessions the peer device can handle.  Note, this is not 
> "message per second" or other, but the number of sessions.  The focus 
> on SCE is on devices like "call agents", "session border controllers", 
> stateful SIP proxies, etc.
>
>
>   
> In practice, the way SCE would be employed is the following:
> * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
>     peer devices
>   * The selector would direct traffic toward the peer devices in a
>     round-robin fashion, taking into consideration the SCE value so
>     as to prevent exhausting resources on any given peer device
>
> I see a contradiction here: if it is round-robin, it does not take 
> into account the SCE value.
> Is that something more like a weighted round-robin algorithm ?
>
> PEJ: As I mention below, the algorithm used to select a peer device is 
> not a part of the SCE draft.  One can use any mechanism.  The SCE 
> draft merely advertises a session capacity estimate.  We've found 
> round-robin to be a simple approach that works, as it distributes load 
> relatively uniformly, but stops directing traffic toward devices that 
> report an SCE value that is below a specified threshold. Other input 
> could be considered, but we deliberately left the algorithm outside 
> the scope of SCE.
>
>
>   * Once the peer's SCE value dropped below a given "low-water mark",
>     the peer device would be less favored and not given traffic until
>     the SCE value increases above the low-water mark (OPTIONS "ping"
>     messages may be used to determine the SCE value in the event that
>     there is no other traffic)
>
> Why would we wait for a lower mark before to influence the 
> distribution ? Also is that less favored, or given no traffic ?
>
> PEJ: With our use of SCE, we're not trying to "balance" the load, per 
> se.  One certainly could do that, but in order to balance the load on 
> peer devices when they are out of balance, one must direct multiple 
> new sessions toward the same device.  That is not always desirable 
> since new sessions usually consume more resources temporarily than 
> established sessions.  So, we try to burden a peer device by throwing 
> a number of new sessions toward the peer by distributing them in a 
> round-robin fashion.  Again, though, the approach taken is outside the 
> scope of the SCE draft.
>
>
> The later has stability issues.  And why do we need ping Options when 
> there is no traffic ?
>  If there is no traffic (or very little), should we care about optimal 
> distribution from this source ?
>
> PEJ: The assumption is that a peer device might also have several 
> peers.  So, while device A1 may not be directing traffic toward device 
> B, perhaps A2, A3, and A4 are.  We can use OPTIONS "ping" messages to 
> get current SCE information.
>
>
>
>
>   * The current SOC overload control mechanism would be used to
>     to prevent message overloading, as SCE only focuses on resource
>     overloading (e.g., memory, CPU, disk space, DSPs)
>
> I don't see this difference here (message overload versus resource 
> overload ? ) . To me, SOC focuses on source shedding, while SCE seems 
> to focus on improved load distribution, just like our proposal.
>
> PEJ: A peer device might be able to handle 18,000 new sessions, but 
> perhaps it can only handle them at a rate of 100 per second.  We would 
> use SCE to determine that a peer can handle new sessions, while we 
> would use the methods SOC is defining to prevent message overload so 
> we do not send 200 INVITEs per second to the peer.  (Use of the SOC 
> mechanism might prove to work better than our current round-robin 
> approach, but we've not tested it.)
>
>
>   
> Note that the "round-robin" approach may or may not be preferred and is not
> required in our draft.  I've found this approach works pretty well, but one
> may prefer to use a particular weighting that is more complex.  I don't
> think local decisions in the box need to be subject to standardization: we
> just need to standardize the signaling to enable such decisions.
>
> I believe the signaling is not enough. The deep meaning of the 
> feedback parameter needs to be clarified too.
> For example, in our proposal, the feedback is simply the current 
> utilization rate of the node (very simple).
>
> PEJ: That sounds similar to SCE.  An SCE-capable device advertises to 
> its peer how many new sessions it can handle given the number of 
> sessions it is currently handling.
>
> Also the expectation needs to be clarified because it has a very 
> important impact on the recommended behavior in case of overload.
> The solution that we propose constantly tries to balance load in the 
> cluster of servers. Therefore, in case the SOC shedding kicks in, 
> there is NOT NEED to repeat the message elsewhere in the cluster 
> because it is balanced already (So no other server has more capacity). 
> This is a huge advantage for the network stability as message repeat 
> tends to spread the overload and collapse the cluster.
>
>
> In any case, it certainly looks like we're thinking along the same lines.
>
> Yes I agree.
>
>
> PEJ: If you haven't already, have a look at our draft:
>
> http://tools.ietf.org/html/draft-jones-sip-overload-sce-00
>
> It is now expired, but your draft suggests there is still life in this 
> approach.  We should look to harmonize them, perhaps.
>
> Paul
>
>
> Cordially,
>
> Thierry
>
>
>   
> Paul
>   
>
>     -----Original Message-----
>
>     From:dispatch-bounces@ietf.org    [mailto:dispatch-bounces@ietf.org] On
>
>     Behalf Of Vijay K. Gurbani
>
>     Sent: Monday, March 07, 2011 6:38 PM
>
>     To:dispatch@ietf.org  
>
>     Subject: Re: [dispatch] SIP load balancing
>
>       
>
>     Paul: Thank you for your input.  More inline.
>
>       
>
>     On 03/07/2011 05:14 PM, Paul Kyzivat wrote:
>
>         Vijay,
>
>         IIUC, the algorithm you have proposed is similar to the one Paul Jones
>
>         has proposed - with the main difference being the numeric range of the
>
>         weight. Do I have that right?
>
>       
>
>     I believe that Paul J.'s algorithm is more suited towards overload
>
>     control, not load balancing (and I will await Paul J.'s correction to my
>
>     assertion).
>
>       
>
>     The intent in draft-bessis-dispatch-adaptive-load-balancing is to
>
>     *always* report the current load to give the upstream SIP proxy an
>
>     indication of the health of the SIP server in the cluster.  If one waits
>
>     until overload is imminent, then it may be too late since a sudden and
>
>     unanticipated burst of messages will put the SIP server in overload
>
>     rather quickly.
>
>       
>
>         It isn't clear to me from your draft if you are specifying the
>
>         algorithm by which the relative weights are used to determine the
>
>         probability of a request going to one server or another.
>
>       
>
>     Yes, same way as RFC2782 does it.
>
>       
>
>         ISTM that it *does* matter. And it especially matters if there is an
>
>         implication that the upstream node's algorithm may depend on some
>
>         assumption or knowledge about the downstream nodes.
>
>       
>
>     No knowledge, just a reported number that the upstream SIP proxy can
>
>     treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
>
>     starts off with the SRV indicators in RFC2782, without any regard to
>
>     additional knowledge or assumption about downstream nodes.
>
>     dispatch-adaptive-load-balancing is trying to use the same model, except
>
>     render the weight to be sensitive to the current traffic in the cluster
>
>     in real-time.
>
>       
>
>         And of course this method of reporting weights isn't sufficient to
>
>         drive a balancing algorithm that is sensitive to some requests
>
>         requiring more or different resources than others. (That seems to be a
>
>         real problem, though I don't know if we have a real solution for
>
>         it.)
>
>       
>
>     Yes, that is a real problem, as is the problem of a SIP server in a
>
>     cluster that has a low weight and is thus not contacted with enough
>
>     frequency to merit more participation to the goodput.
>
>       
>
>     It is well known in SIP that some methods are more onerous to process
>
>     than others --- from a proxying point of view, processing an INVITE with
>
>     forking and response aggregation is far more detrimental to resource
>
>     usage than is processing a BYE.  Some existing literature techniques
>
>     thus add a coefficient to the weight of sending an INVITE versus a BYE.
>
>     At this time, the process outlined in dispatch-adaptive-load-balancing
>
>     treats all requests the same, but we can definitely look at ways to
>
>     change this behaviour.
>
>       
>
>     Thanks,
>
>       
>
>     - vijay
>
>     --
>
>     Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
>
>     1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
>
>     Email:vkg@{bell-labs.com,acm.org}    /vijay.gurbani@alcatel-lucent.com  
>
>     Web:http://ect.bell-labs.com/who/vkg/
>
>     _______________________________________________
>
>     dispatch mailing list
>
>     dispatch@ietf.org  
>
>     https://www.ietf.org/mailman/listinfo/dispatch
>
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>   
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------020009070706080402000908
Content-Type: multipart/related;
 boundary="------------000400070704020705060401"


--------------000400070704020705060401
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Paul, 

    

    I agree they are many similitudes ! 
        :-) 

    But the question I think is: Can we make the two approaches
    compatible ? 

    

    For instance, with our approach, it is fundamental that the weight
    represents a "Relative" occupation, and not an absolute capacity.

    The first reason is that we found that loading continuously all the
    servers at the same utilization rate (same percentage) is the
    optimal way of loading a cluster. (We can discuss that if not
    agreed). The second reason is that a relative occupation is very
    easy to derive from information already available: any system knows
    where it is in terms of CPU or Memory usage. In contrast, deriving a
    left capacity in terms of remaining sessions can be difficult: Host
    can be used by different applications and perform different duties,
    which can change over the day, all of them impacting their true
    capacity at a certain moment. Beside, as already discussed,
    different sessions require different resources, so there is a whole
    complication of dealing with a multitude of call profiles, that
    (again) are not even constant over the day.

    

    Also, we do not want to derive some shedding decision from the
    weight. We want to rely on the OLC mechanism for that. As I
    mentioned before, deriving shedding from the weight is very
    sensitive to the way weights are calculated: we would need to
    specify exactly how the weight is calculated: this is impractical !
    Also everybody would need to agree about specific threshold: What is
    the weight value threshold for shedding ?: Again this is more
    difficult, and different servers will certainly want to use
    different thresholds.   Also there is a contradiction: whether you
    are using relative or absolute weights, your weight should  zero
    (100% busy, or 0 sessions available) at the moment you are getting
    overloaded. So if you are 100% busy (or in average no room for new
    sessions), how to you tell how many sessions need to be shed ? You
    could be 100% busy with an offered load of 300% nominal (then you
    need to shed 66%), or be 100% busy at just the nominal load, (then
    you don't need to shed). This again shows this threshold issue I
    mentioned. In our implementation, we used to use to a common
    "utilization" information to do both optimal distribution and
    shedding, but we moved away from that because of those reasons: we
    found that each node indicating its requested rejection rate is much
    more robust and requires very little agreements. 

    

    Cordially, 

    

    Thierry

    

    

    On 09-Mar-2011 22:15, Paul E. Jones wrote:
    

      
      
      
      Thierry's signature
      
      

        
Thierry,

        
 

        
I think there is more in common than not.

        
 

        
You compute a “weight” and we compute an “sce”
            that is, in my opinion, more or less the same: it’s a value
            that indicates how much traffic can be directed toward the
            box relative to other peer devices.

        
 

        
You want to do “load balancing” whereas we want
            to do “overload prevention”.  Preventing overload means
            distributing load across boxes.  We are not proposing to
            perfectly balance the load, but the load is distributed
            nonetheless.

        
 

        
Honestly, I just don’t see a lot of difference,
            other than terminology.  The algorithm employed by the
            device directing traffic, but that algorithm is independent
            of the advertisement of a capacity/weight value.

        
 

        
Paul

        
 

        

          

            

              
From: Thierry Bessis
                  [mailto:Thierry.Bessis@alcatel-lucent.com] 

                  Sent: Wednesday, March 09, 2011 12:26 PM

                  To: Paul E. Jones

                  Cc: dispatch@ietf.org

                  Subject: Re: [dispatch] SIP load balancing

            

          

          
 

          
Thanks Paul for you detailed answer ! 

            

            I've looked at your SCE proposal and your comments and my
            impression is that our two proposals are actually very
            different both from their intent perspective, and from their
            realization perspective. 

            

            The intent of our load balancing proposal is clearly
            continuous load balancing. Our suggested weight does try to
            figure any kind of remaining capacity in terms of sessions:
            it only indicates the instant (slightly smoothed) available
            resource, which integrates CPU, Memory and any other
            constrained resource. For the kind of usage we have in mind,
            (and for example), it does not matter if we have lot of
            sessions "left" when we have very little CPU to register
            those sessions: it is optimal to temporary offload this busy
            node rather than risking a CPU induced rejection or failure.
            When the node will have more CPU available, it will take
            more sessions, and may be even more than others, because it
            may have now more memory than others. If this node never
            have more CPU available, then its "session" capacity is
            actually unusable and the system correctly reflects that. 

            

            Also, the SOC was not intended for "dynamic" load
            distribution: it only provides the feedback when some
            shedding is expected (So already one or multiple nodes are
            about to be overloaded). Also, integrating the information
            from both SOC and SCE to take the distribution decision
            seems rather complicated to me. In our proposal, the
            distribution algorithm is simple, never changes, and is
            independent from SOC. We selected a node with weights always
            the same way, and we apply the SOC rejection when applicable
            to this node.

            

            Cordially, 

            

            Thierry

            

            On 08-Mar-2011 14:47, Paul E. Jones wrote: 

          
Thierry,

          
 

          
Please
              see my comments below in green:

          
 

          
 

          

            

              

                
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
                    

                    Sent: Tuesday, March 08, 2011 12:35 PM

                    To: Paul E. Jones

                    Cc: 'Vijay K. Gurbani'; dispatch@ietf.org

                    Subject: Re: [dispatch] SIP load balancing

              

            

            
 

            
Paul, 

              

              Please let me address some of your comments. 

              I don't know what the SCE is, so please pardon me if I
              misunderstand something. 

              

              On 07-Mar-2011 23:05, Paul E. Jones wrote: 

            
Vijay,

            
 

            
The Session Capacity Estimate (SCE) idea is actually quite similar to the

            
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a

            
solution to the "overload" problem, the approach being one where we try to

            
avoid overload by selecting a device that reports a higher capacity value.

            
Thus, one could view it as "overload control through load balancing".  We

            
didn't explicitly call it load balancing, though, because we were not trying

            
to balance load (i.e., place the same load on all servers).  Rather, we just

            
wanted to direct traffic to devices that we knew had available capacity.

            
Agreed, the goal seems to be the same:
              avoiding or at least delaying overload. 

              And your other comment too: our proposal does try to
              balance the load on all the servers. The rational is that
              it is very easy for each server to know where it is in
              terms of resource usage (for instance: I'm 60 percent
              resource loaded at this moment), and this is just the
              information you need to balance the traffic. Now it seems
              (but it is not clear, see below)  that the SCE "available
              capacity" is actually a real time feedback, so the
              information would be very similar actually. (What's the
              difference ? )

            
 

            
PEJ: I think it is very
                similar.  SCE provides real-time feedback of the
                available resource capacity of a peer device.  The
                intent is to convey a single integer value that
                represents an estimate of the number of sessions the
                peer device can handle.  Note, this is not “message per
                second” or other, but the number of sessions.  The focus
                on SCE is on devices like “call agents”, “session border
                controllers”, stateful SIP proxies, etc. 

              

              

              

            
 

            
In practice, the way SCE would be employed is the following:

            
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from

            
   peer devices

            
 * The selector would direct traffic toward the peer devices in a

            
   round-robin fashion, taking into consideration the SCE value so

            
   as to prevent exhausting resources on any given peer device

            
I see a contradiction here: if it is
              round-robin, it does not take into account the SCE value.
              

              Is that something more like a weighted round-robin
              algorithm ? 

            
 

            
PEJ: As I mention below, the
                algorithm used to select a peer device is not a part of
                the SCE draft.  One can use any mechanism.  The SCE
                draft merely advertises a session capacity estimate. 
                We’ve found round-robin to be a simple approach that
                works, as it distributes load relatively uniformly, but
                stops directing traffic toward devices that report an
                SCE value that is below a specified threshold. Other
                input could be considered, but we deliberately left the
                algorithm outside the scope of SCE.

              

              

              

            
 * Once the peer's SCE value dropped below a given "low-water mark",

            
   the peer device would be less favored and not given traffic until

            
   the SCE value increases above the low-water mark (OPTIONS "ping"

            
   messages may be used to determine the SCE value in the event that

            
   there is no other traffic)

            
Why would we wait for a lower mark
              before to influence the distribution ? Also is that less
              favored, or given no traffic ? 

            
 

            
PEJ: With our use of SCE, we’re
                not trying to “balance” the load, per se.  One certainly
                could do that, but in order to balance the load on peer
                devices when they are out of balance, one must direct
                multiple new sessions toward the same device.  That is
                not always desirable since new sessions usually consume
                more resources temporarily than established sessions. 
                So, we try to burden a peer device by throwing a number
                of new sessions toward the peer by distributing them in
                a round-robin fashion.  Again, though, the approach
                taken is outside the scope of the SCE draft.

            
 

            


              The later has stability issues.  And why do we need ping
              Options when there is no traffic ?

               If there is no traffic (or very little), should we care
              about optimal distribution from this source ?

            
 

            
PEJ: The assumption is that a
                peer device might also have several peers.  So, while
                device A1 may not be directing traffic toward device B,
                perhaps A2, A3, and A4 are.  We can use OPTIONS “ping”
                messages to get current SCE information.

            


              

              

              

            
 * The current SOC overload control mechanism would be used to

            
   to prevent message overloading, as SCE only focuses on resource

            
   overloading (e.g., memory, CPU, disk space, DSPs)

            
I don't see this difference here
              (message overload versus resource overload ? ) . To me,
              SOC focuses on source shedding, while SCE seems to focus
              on improved load distribution, just like our proposal.

            
 

            
PEJ: A peer device might be
                able to handle 18,000 new sessions, but perhaps it can
                only handle them at a rate of 100 per second.  We would
                use SCE to determine that a peer can handle new
                sessions, while we would use the methods SOC is defining
                to prevent message overload so we do not send 200
                INVITEs per second to the peer.  (Use of the SOC
                mechanism might prove to work better than our current
                round-robin approach, but we’ve not tested it.) 

              

              

              

            
 

            
Note that the "round-robin" approach may or may not be preferred and is not

            
required in our draft.  I've found this approach works pretty well, but one

            
may prefer to use a particular weighting that is more complex.  I don't

            
think local decisions in the box need to be subject to standardization: we

            
just need to standardize the signaling to enable such decisions.

            
I believe the signaling is not enough.
              The deep meaning of the feedback parameter needs to be
              clarified too. 

              For example, in our proposal, the feedback is simply the
              current utilization rate of the node (very simple).

            
 

            
PEJ: That sounds similar to
                SCE.  An SCE-capable device advertises to its peer how
                many new sessions it can handle given the number of
                sessions it is currently handling. 

              

              Also the expectation needs to be clarified because it has
              a very important impact on the recommended behavior in
              case of overload. 

              The solution that we propose constantly tries to balance
              load in the cluster of servers. Therefore, in case the SOC
              shedding kicks in, there is NOT NEED to repeat the message
              elsewhere in the cluster because it is balanced already
              (So no other server has more capacity). This is a huge
              advantage for the network stability as message repeat
              tends to spread the overload and collapse the cluster. 

              

              

              

            
In any case, it certainly looks like we're thinking along the same lines.

            
Yes I agree. 

              

              

              

            
PEJ: If you haven’t already,
                have a look at our draft:

            
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

            
 

            
It is now expired, but your
                draft suggests there is still life in this approach.  We
                should look to harmonize them, perhaps.

            
 

            
Paul

            
 

            


              Cordially, 

              

              Thierry

              

              

              

            
 

            
Paul

            
 

            

              
-----Original Message-----

              
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On

              
Behalf Of Vijay K. Gurbani

              
Sent: Monday, March 07, 2011 6:38 PM

              
To: dispatch@ietf.org

              
Subject: Re: [dispatch] SIP load balancing

              
 

              
Paul: Thank you for your input.  More inline.

              
 

              
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

              

                
Vijay,

                
IIUC, the algorithm you have proposed is similar to the one Paul Jones

                
has proposed - with the main difference being the numeric range of the

                
weight. Do I have that right?

              

              
 

              
I believe that Paul J.'s algorithm is more suited towards overload

              
control, not load balancing (and I will await Paul J.'s correction to my

              
assertion).

              
 

              
The intent in draft-bessis-dispatch-adaptive-load-balancing is to

              
*always* report the current load to give the upstream SIP proxy an

              
indication of the health of the SIP server in the cluster.  If one waits

              
until overload is imminent, then it may be too late since a sudden and

              
unanticipated burst of messages will put the SIP server in overload

              
rather quickly.

              
 

              

                
It isn't clear to me from your draft if you are specifying the

                
algorithm by which the relative weights are used to determine the

                
probability of a request going to one server or another.

              

              
 

              
Yes, same way as RFC2782 does it.

              
 

              

                
ISTM that it *does* matter. And it especially matters if there is an

                
implication that the upstream node's algorithm may depend on some

                
assumption or knowledge about the downstream nodes.

              

              
 

              
No knowledge, just a reported number that the upstream SIP proxy can

              
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution

              
starts off with the SRV indicators in RFC2782, without any regard to

              
additional knowledge or assumption about downstream nodes.

              
dispatch-adaptive-load-balancing is trying to use the same model, except

              
render the weight to be sensitive to the current traffic in the cluster

              
in real-time.

              
 

              

                
And of course this method of reporting weights isn't sufficient to

                
drive a balancing algorithm that is sensitive to some requests

                
requiring more or different resources than others. (That seems to be a

                
real problem, though I don't know if we have a real solution for

                
it.)

              

              
 

              
Yes, that is a real problem, as is the problem of a SIP server in a

              
cluster that has a low weight and is thus not contacted with enough

              
frequency to merit more participation to the goodput.

              
 

              
It is well known in SIP that some methods are more onerous to process

              
than others --- from a proxying point of view, processing an INVITE with

              
forking and response aggregation is far more detrimental to resource

              
usage than is processing a BYE.  Some existing literature techniques

              
thus add a coefficient to the weight of sending an INVITE versus a BYE.

              
At this time, the process outlined in dispatch-adaptive-load-balancing

              
treats all requests the same, but we can definitely look at ways to

              
change this behaviour.

              
 

              
Thanks,

              
 

              
- vijay

              
--

              
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent

              
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)

              
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com

              
Web:   http://ect.bell-labs.com/who/vkg/

              
_______________________________________________

              
dispatch mailing list

              
dispatch@ietf.org

              
https://www.ietf.org/mailman/listinfo/dispatch

            

            
 

            
_______________________________________________

            
dispatch mailing list

            
dispatch@ietf.org

            
https://www.ietf.org/mailman/listinfo/dispatch

            
 

            

              
-- 

                --
                    

                    Cordially, 

                    Thierry Bessis 

              
  IMS
                  Solution: Architecture & Product Management - ALTA
                  Member

                    Organization: ALU > Carrier group > Wireline
                  BU >  IMS Sol Arch, SE, PdM

                Alcatel-Lucent

                  1960 Lucent
                  Lane - Room 6L-517

                    Naperville, IL  60563  U S

                    Tel/Fax: +1 630 979 7989

                    Corporate IM: tbessis    - Yahoo IM:    
                  thierry_bessis

                

                My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                  My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                  ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

                

                Conference information:

                  2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 -
                  Access Code: 9797989

                  others countries see: http://all.alcatel-lucent.com/audio

                

                Upcoming
                    planned Business Trip: 12 April - 14 April 2011

                Upcoming
                    planned Vacation: None 

            

          

          
 

          
 

          
_______________________________________________

          
dispatch mailing list

          
dispatch@ietf.org

          
https://www.ietf.org/mailman/listinfo/dispatch

          
 

          

            
-- 

              -- 

                  Cordially, 

                  Thierry Bessis 

            
  IMS Solution:
                Architecture & Product Management - ALTA Member

                  Organization: ALU > Carrier group > Wireline BU
                >  IMS Sol Arch, SE, PdM

              Alcatel-Lucent

                1960 Lucent
                Lane - Room 6L-517

                  Naperville, IL  60563  U S

                  Tel/Fax: +1 630 979 7989

                  Corporate IM: tbessis    - Yahoo IM:    
                thierry_bessis

              

              My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

              

              Conference information:

                2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 -
                Access Code: 9797989

                others countries see: http://all.alcatel-lucent.com/audio

              

              Upcoming planned Business Trip: 12
                  April - 14 April 2011

              Upcoming planned Vacation: None
              

          

        

      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------000400070704020705060401
Content-Type: image/jpeg
Content-Transfer-Encoding: base64
Content-ID: 
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--------------000400070704020705060401
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------000400070704020705060401--

--------------020009070706080402000908--

From partr@cisco.com  Thu Mar 10 11:08:20 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2CE613A6826 for ; Thu, 10 Mar 2011 11:08:20 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.495
X-Spam-Level: 
X-Spam-Status: No, score=-9.495 tagged_above=-999 required=5 tests=[AWL=-0.293, BAYES_00=-2.599, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RJG8l90CVqUX for ; Thu, 10 Mar 2011 11:08:18 -0800 (PST)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id F26FC3A6806 for ; Thu, 10 Mar 2011 11:08:17 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=12924; q=dns/txt; s=iport; t=1299784176; x=1300993776; h=mime-version:subject:date:message-id:references:from:to: cc; bh=+eJ3cu5dSja0WNzXp0MR9jhGT6iE8lv7Ng5CNlfHOiY=; b=gXU8/DMju0hv8OX6Koc3Pq4CIrWvO7bwVbxqZ+WPOpo8QOEHKT61iMac QifuRrfbvczr86lYw/VAKFZqjWdhWfJtlUMWJBSo0jN5A6zfv0/AWq1XC e8kRAamet8TFIFN5hY/vpNdHoH/KgIV/JrgYcIM7aBatagAg/sL4ctjxm I=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AqkEAJuweE2Q/khNgWdsb2JhbAClWFgUAQELCyYlpTecNYViBIUiinE
X-IronPort-AV: E=Sophos;i="4.62,297,1297036800"; d="scan'208,217";a="78649472"
Received: from ams-core-4.cisco.com ([144.254.72.77]) by ams-iport-1.cisco.com with ESMTP; 10 Mar 2011 19:09:35 +0000
Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by ams-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2AJ9Xak011632; Thu, 10 Mar 2011 19:09:34 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 11 Mar 2011 00:39:33 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01CBDF56.B08B4278"
Date: Fri, 11 Mar 2011 00:39:33 +0530
Message-ID: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvd8Jwuu595L4ZlROesYLB3rZg1awAiT73NADT0rIQ=
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com> 
From: "Parthasarathi R (partr)" 
To: "Worley, Dale R (Dale)" , "Paul Kyzivat (pkyzivat)" , "Vijay K. Gurbani" 
X-OriginalArrivalTime: 10 Mar 2011 19:09:33.0613 (UTC) FILETIME=[B0C5DDD0:01CBDF56]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 19:08:20 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBDF56.B08B4278
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Dale,
=20
I like the way you described the problem statement.=20

________________________________

From: dispatch-bounces@ietf.org on behalf of Worley, Dale R (Dale)
Sent: Wed 3/9/2011 10:40 PM
To: Paul Kyzivat (pkyzivat); Vijay K. Gurbani
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing



This is a fascinating subject.  On the way home from work, I started to =
frame a set of differential equations to model it.

The key, of course, is to have all of the servers reach full capacity =
*at the same time* if the presented load increases steadily.  That is =
how we maximize goodput under such circumstances.  (If there is spare =
capacity within the cluster of servers, it doesn't matter if the load on =
the servers is balanced or not, as goodput equals presented load.)

The crucial questions are "What is the algorithm for generating the =
number that the server presents to the load balancer?" and "What =
algorithm does the load balancer use for assigning work, given the =
numbers presented by the servers?"

 In case units (number) are well defined in the standard manner, =
it is possible to generate in the downstream entity and upstream (load =
balancer) will be able to understand and shall distribute the load based =
on the unit. IMO, The standard unit may be number of dialog or bandwidth =
available or any other . The selection of unit needs serious discussion =
to conclude somewhere. =20

The draft (draft-bessis-dispatch-adaptive-load-balancing-00) seems to =
assume that the assignment algorithm is much the same as RFC 3263, to =
statistically send Wi/(sum(Wj)) of the load to server i (with Wi being =
the number reported by server i).  The draft gives little or no hint of =
what the generation algorithm will be.

In order for the mechanism to be standardized in a meaningful way, it =
must specify the generation algorithm sufficiently that a combination of =
servers that use generation algorithms in conformance with the standard =
will work together correctly -- as I said above, as the entire cluster =
nears full capacity, all servers much reach full capacity =
simultaneously.

t seems that we need to do a careful analysis of the behavior of the =
system near the point where "unused capacity of server i" is near 0 for =
all servers i.  The first question is to categorize generation functions =
by their behavior near the point "unused capacity =3D 0", that is, does =
the reported "weight" approach a non-zero number, does it approach zero =
linearly, does it approach zero quadratically, etc.?  I suspect that the =
system behaves well only when all servers use functions of the same =
form, and with the same constant on the leading term.  Intuitively, the =
correct answer seems to be to have the weight be "the unused capacity, =
measured in number of calls that can be handled".

 Considering calls as a unit works very well in deployed H.323 =
network wherein most of the calls consume resources lineraly (G711 or =
G729) calls =20

This gets messier when we consider that the load is comes in quantized =
units (single calls) and has significant statistical fluctuations.  In =
particular, some functions that appear to work correctly in the =
"analytical" analysis outlined above my have bad behavior when presented =
with statistical fluctuations.

Additionally, the feedback from the servers to the load balancer will =
not be instantaneous, and the delays will introduce "control loop =
delays", which if not handled correctly might cause the system to =
oscillate.


 This problem shall be solved by creating the lower watermark =
and upper watermark for the action to be taken. Say above 80 means stop =
forwarding and but when it comes back to 75 only start forwarding, this =
mechanism helps in avoiding the oscillation due to delay 


A different problem results from the fact that different calls may =
present decidedly different loads to a server, and that the server may =
have several different types of resource.  How does the server sensibly =
report the nature of its resources and the loading on each?  How does =
the load balancer estimate the resources required by an individual call? =


 Two aspects are required to solve this issue:

1) Report the current state of the resource (mediactrl has some =
mechanism for reporting resource, =
draft-partha-sip-overload-resource-availability mention one another =
mechanism)

2) Maximum Consumption vector for a dialog which is a prior knowledge =
exchanged between servers through other mechanism.=20

In case the amount of available bandwidth required in the system as a =
unit, then the downstream shall report the available bandwidth and the =
amount of maximum consumption of the dialog shall be identified by =
maximum capacity mentioned in the dialog SDP. 

(Note that this problem is largely nonexistent in circuit-switched =
systems, as all calls present exactly the same load.)  (This is the =
problem that has bedevilled attempts to solve the problem in =
draft-partha-sip-overload-resource-availability.)

But I don't see any of the needed analysis in the draft, and I don't see =
any indication in the draft of a realization that this analysis needs to =
be done.  Until that is fixed, I predict that a solution will not be =
found, or worse, a solution that appears to work on paper but does not =
work in practice.

Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch



------_=_NextPart_001_01CBDF56.B08B4278
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Re: [dispatch] SIP load balancing=0A=
=0A=
=0A=
=0A=

=0A=

Dale,
=0A=

 
=0A=

I like the way you =
described the problem statement. 
=0A=


=0A=

=0A=
From: dispatch-bounces@ietf.org on =
behalf of Worley, Dale R (Dale)
Sent: Wed 3/9/2011 10:40 =
PM
To: Paul Kyzivat (pkyzivat); Vijay K. Gurbani
Cc: =
dispatch@ietf.org
Subject: Re: [dispatch] SIP load =
balancing

=0A=

=0A=

This is a fascinating subject.  On the way home =
from work, I started to frame a set of differential equations to model =
it.

The key, of course, is to have all of the servers reach full =
capacity *at the same time* if the presented load increases =
steadily.  That is how we maximize goodput under such =
circumstances.  (If there is spare capacity within the cluster of =
servers, it doesn't matter if the load on the servers is balanced or =
not, as goodput equals presented load.)

The crucial questions are =
"What is the algorithm for generating the number that the server =
presents to the load balancer?" and "What algorithm does the load =
balancer use for assigning work, given the numbers presented by the =
servers?"
=0A=

<Partha> In case units (number) are well =
defined in the standard manner, it is possible to generate in the =
downstream entity and upstream (load balancer) will be able to =
understand and shall distribute the load based on the unit. IMO, The =
standard unit may be number of dialog or bandwidth available or any =
other . The selection of unit needs serious discussion to =
conclude somewhere. </Partha> 

The draft =
(draft-bessis-dispatch-adaptive-load-balancing-00) seems to assume that =
the assignment algorithm is much the same as RFC 3263, to statistically =
send Wi/(sum(Wj)) of the load to server i (with Wi being the number =
reported by server i).  The draft gives little or no hint of what =
the generation algorithm will be.

In order for the mechanism to =
be standardized in a meaningful way, it must specify the generation =
algorithm sufficiently that a combination of servers that use generation =
algorithms in conformance with the standard will work together correctly =
-- as I said above, as the entire cluster nears full capacity, all =
servers much reach full capacity simultaneously.

t seems that we =
need to do a careful analysis of the behavior of the system near the =
point where "unused capacity of server i" is near 0 for all servers =
i.  The first question is to categorize generation functions by =
their behavior near the point "unused capacity =3D 0", that is, does the =
reported "weight" approach a non-zero number, does it approach zero =
linearly, does it approach zero quadratically, etc.?  I suspect =
that the system behaves well only when all servers use functions of the =
same form, and with the same constant on the leading term.  =
Intuitively, the correct answer seems to be to have the weight be "the =
unused capacity, measured in number of calls that can be =
handled".
=0A=

<Partha> Considering calls as a unit works very =
well in deployed H.323 network wherein most of the calls =
consume resources lineraly (G711 or G729) calls =
</Partha> 

This gets messier when we consider that the =
load is comes in quantized units (single calls) and has significant =
statistical fluctuations.  In particular, some functions that =
appear to work correctly in the "analytical" analysis outlined above my =
have bad behavior when presented with statistical =
fluctuations.

Additionally, the feedback from the servers to the =
load balancer will not be instantaneous, and the delays will introduce =
"control loop delays", which if not handled correctly might cause the =
system to oscillate.
=0A=

<Partha> This problem shall be solved by =
creating the lower watermark and upper watermark for the action to be =
taken. Say above 80 means stop forwarding and but when it comes back to =
75 only start forwarding, this mechanism helps in avoiding the =
oscillation due to delay </partha>
=0A=


A different problem results from the fact that different calls =
may present decidedly different loads to a server, and that the server =
may have several different types of resource.  How does the server =
sensibly report the nature of its resources and the loading on =
each?  How does the load balancer estimate the resources required =
by an individual call? 
=0A=

<partha> Two aspects are required to solve this issue:
=0A=

1) Report the current state of the resource (mediactrl has some =
mechanism for reporting resource, =
draft-partha-sip-overload-resource-availability mention one another =
mechanism)
=0A=

2) Maximum Consumption vector for a dialog which is a prior =
knowledge exchanged between servers through other mechanism. 
=0A=

In case the amount of available bandwidth required in the system =
as a unit, then the downstream shall report the available bandwidth and =
the amount of maximum consumption of the dialog shall be identified by =
maximum capacity mentioned in the dialog SDP. </Partha>
=0A=

(Note that this problem is largely nonexistent in circuit-switched =
systems, as all calls present exactly the same load.)  (This is the =
problem that has bedevilled attempts to solve the problem in =
draft-partha-sip-overload-resource-availability.)

But I don't see =
any of the needed analysis in the draft, and I don't see any indication =
in the draft of a realization that this analysis needs to be done.  =
Until that is fixed, I predict that a solution will not be found, or =
worse, a solution that appears to work on paper but does not work in =
practice.

Dale
_______________________________________________<=
BR>dispatch mailing list
dispatch@ietf.org
https://www.ietf.=
org/mailman/listinfo/dispatch

------_=_NextPart_001_01CBDF56.B08B4278--

From fluffy@cisco.com  Thu Mar 10 12:16:45 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CE0D73A6B69 for ; Thu, 10 Mar 2011 12:16:45 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.4
X-Spam-Level: 
X-Spam-Status: No, score=-110.4 tagged_above=-999 required=5 tests=[AWL=0.199,  BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oxvExD5pnnQt for ; Thu, 10 Mar 2011 12:16:45 -0800 (PST)
Received: from sj-iport-2.cisco.com (sj-iport-2.cisco.com [171.71.176.71]) by core3.amsl.com (Postfix) with ESMTP id 8068A3A6962 for ; Thu, 10 Mar 2011 12:16:44 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=607; q=dns/txt; s=iport; t=1299788282; x=1300997882; h=from:content-transfer-encoding:subject:date:message-id: to:mime-version; bh=PjpgrfHyGS8WQicTgQtFDaODMVJKGIaWrZbtIDjUX1I=; b=L8tIE0tENbdaMs6MiE4ldPOBeZBOSvx+8VrC5J5moW130Sg+jQO9cErM K53gNZ9x4SrUeT8qokoHp0KGTwaew+kTc807lHWJf7AL2K/gYVcetYs+D YZ4AJfl2rhhLh4MuGepNkH3lzYZLmgqPNVVkspztvgSRVphT+y6pXh+99 4=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAE/AeE2rR7Hu/2dsb2JhbACmMXelGJwzhWIEhSSHH4NK
X-IronPort-AV: E=Sophos;i="4.62,297,1297036800"; d="scan'208";a="319299502"
Received: from sj-core-5.cisco.com ([171.71.177.238]) by sj-iport-2.cisco.com with ESMTP; 10 Mar 2011 20:18:02 +0000
Received: from [192.168.4.100] (rcdn-fluffy-8712.cisco.com [10.99.9.19]) by sj-core-5.cisco.com (8.13.8/8.14.3) with ESMTP id p2AKI2JV007430 for ; Thu, 10 Mar 2011 20:18:02 GMT
From: Cullen Jennings 
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Date: Thu, 10 Mar 2011 13:20:44 -0700
Message-Id: 
To: DISPATCH list 
Mime-Version: 1.0 (Apple Message framework v1082)
X-Mailer: Apple Mail (2.1082)
Subject: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 20:16:45 -0000

I wrote up the start of a draft on requirements.=20

http://tools.ietf.org/html/draft-jennings-rtcweb-api-00

I think the key thing for the key thing for IETF working is starting to =
think about the requirements and what on the on the wire protocols are =
that we want to select. I included some API definition because I found =
it helpful to be thinking about that in this draft but that would not be =
part of the IETF work. That would be work I think W3C should do. The =
draft also touches on the of the security issues for media and signaling =
originating from browsers.=20

Cullen=20





From dworley@avaya.com  Thu Mar 10 13:31:08 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A27AD3A6ADE for ; Thu, 10 Mar 2011 13:31:08 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.539
X-Spam-Level: 
X-Spam-Status: No, score=-102.539 tagged_above=-999 required=5 tests=[AWL=0.060, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KCnLFCxf1kh0 for ; Thu, 10 Mar 2011 13:31:07 -0800 (PST)
Received: from co300216-co-outbound.net.avaya.com (co300216-co-outbound.net.avaya.com [198.152.13.100]) by core3.amsl.com (Postfix) with ESMTP id 601813A6AD6 for ; Thu, 10 Mar 2011 13:31:07 -0800 (PST)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU3GmAcF/2dsb2JhbACmZXSkeQKZFoVhBIUcinA
X-IronPort-AV: E=Sophos;i="4.62,298,1297054800"; d="scan'208";a="268841600"
Received: from unknown (HELO co300216-co-erhwest.avaya.com) ([198.152.7.5]) by co300216-co-outbound.net.avaya.com with ESMTP; 10 Mar 2011 16:32:25 -0500
X-IronPort-AV: E=Sophos;i="4.62,298,1297054800"; d="scan'208";a="593565215"
Received: from unknown (HELO DC-US1HCEX3.global.avaya.com) ([135.11.52.22]) by co300216-co-erhwest-out.avaya.com with ESMTP; 10 Mar 2011 16:32:25 -0500
Received: from DC-US1MBEX4.global.avaya.com ([169.254.1.187]) by DC-US1HCEX3.global.avaya.com ([135.11.52.22]) with mapi; Thu, 10 Mar 2011 16:32:24 -0500
From: "Worley, Dale R (Dale)" 
To: Henry Sinnreich , Andrew Allen , "tanakai@nttdocomo.co.jp" ,  "dispatch@ietf.org" 
Date: Thu, 10 Mar 2011 16:32:23 -0500
Thread-Topic: [dispatch] Revision of draft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AQHL32qk/anAaDWa+EqtEkwl8nOyUg==
Message-ID: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [dispatch] Revision of draft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 21:31:08 -0000

Yes, there is a concern about the IETF's efforts being diverted into activi=
ties
that are unfriendly to the end-to-end architecture.  On the other hand, it =
is likely
that the walled gardens will not be so completely walled in the future, and=
 it would
be best to maintain as much technical alignment as is reasonable between th=
e
public Internet and various non-public networks

In this case, the tradeoff seems to me to favor advancing this I-D into an =
RFC, as
the IETF bandwidth that will be consumed is fairly small, and the future pr=
oblem to
be avoided is large:  It would be inconvenient to discover that a mobile se=
rvice
provider that wanted to make its SIP network more open was using URNs for
instance-ids that have been defined to have another purpose by the IETF.

Dale

From aallen@rim.com  Thu Mar 10 14:02:57 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5339C3A6AAB for ; Thu, 10 Mar 2011 14:02:57 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.203
X-Spam-Level: 
X-Spam-Status: No, score=-1.203 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, MIME_QP_LONG_LINE=1.396]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yQKzs3olIsCs for ; Thu, 10 Mar 2011 14:02:56 -0800 (PST)
Received: from mhs060cnc.rim.net (mhs060cnc.rim.net [208.65.73.34]) by core3.amsl.com (Postfix) with ESMTP id C2C4A3A6A37 for ; Thu, 10 Mar 2011 14:02:55 -0800 (PST)
X-AuditID: 0a41282f-b7b4eae000005e51-46-4d794ad77676
Received: from XCH138CNC.rim.net (xch138cnc.rim.net [10.65.20.127]) by mhs060cnc.rim.net (SBG) with SMTP id 76.14.24145.7DA497D4; Thu, 10 Mar 2011 22:04:07 +0000 (GMT)
Received: from XCH02DFW.rim.net ([10.150.100.31]) by XCH138CNC.rim.net with Microsoft SMTPSVC(6.0.3790.3959); Thu, 10 Mar 2011 17:04:13 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
content-transfer-encoding: quoted-printable
Date: Thu, 10 Mar 2011 16:04:09 -0600
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvfRQM5LNYNOSZVz0atfnKVY5yxgAAIeg3w
References: 
From: "Andrew Allen" 
To: "Henry Sinnreich" , , 
X-OriginalArrivalTime: 10 Mar 2011 22:04:13.0902 (UTC) FILETIME=[17830EE0:01CBDF6F]
X-Brightmail-Tracker: AAAAAQAAAZE=
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 10 Mar 2011 22:02:57 -0000

Hi Henry,

I think then that it's a fair summary of your email to say that your
objections are of a political and philosophical nature related to your
view of what constitutes the internet and whether the IETF should be
involved in enhancing IETF protocols for stakeholders that utilize those
IETF protocols for deployments that some in the community might not
consider "the Internet" or "the pure Internet" or the "open e2e
Internet" rather than technical issues with the contents of the drafts.

While I don't want to start a general debate on what is "the Internet"
here I will state that in my view it's very difficult to exclude any
particular stakeholders and their deployments of IETF protocols from
being part of the Internet when those stakeholders are in many cases the
very same established telecommunications companies or ISPs that the vast
majority of the population in most countries use to obtain access to the
Internet. 

My understanding also is that IETF nearly a decade ago agreed to
cooperate with 3GPP (and other standards bodies) in developing and
enhancing internet protocols and that IETF specifically requested (more
like insisted) that any extensions or enhancements to IETF protocols
required by 3GPP be done within IETF according to the procedures defined
by IETF for enhancing IETF protocols and for registering namespaces etc.
These drafts simply comply with that agreed way of working between IETF
and 3GPP. To my knowledge this working agreement is still in effect. I
don't think it is the role of the dispatch WG to determine which
applications or deployments of SIP to support or cooperate with counter
to agreements previously made at IESG level.

Refusing cooperation based on what is considered not to be part of the
"open e2e Internet" is only likely to set up two (or more) competing
versions of the Internet resulting in the 21st century equivalent of the
standard rail gauge vs broad rail gauge communications interoperability
issue from the 19th century - effectively killing the goal of an open
e2e internet! As Dale indicated what is considered a walled garden today
may not be walled in the future (if I remember some very significant
walls came tumbling down during just the last decade of the 20th century
alone).

Different business models will be resolved through the market not by
creating competing incompatible technical solutions. 

In terms of the time used on discussion within dispatch I don't think
there is a need for much more discussion in dispatch. These drafts could
go ahead as AD sponsored without further WG involvement. As I understand
it the IMEI URN definition draft did not as part of the URN registration
process need to even be discussed in dispatch at all (although Dale
Worley's review was invaluable in improving the draft technically) as
the NIDs list was the fora for this discussion and where previous
versions of that draft were discussed.

It should be noted all we are talking about with regard to these drafts
is the registration of a URN namespace and definition of its usage as an
Instance ID as required by RFC 5626. If we can't agree that then what
future extensions can we agree?

Andrew

-----Original Message-----
From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com] 
Sent: Thursday, March 10, 2011 11:03 AM
To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org
Subject: Re: [dispatch]
Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

Hi Andrew,

I have generic objections to see the publication of yet another
informational RFC about practices in consortiums of closed networks in
the
IETF, for the following reasons:

* Using the RFC process as a publishing venue for organizations that
don't
explicitly do that in support of the open e2e Internet.

* The resources of the IETF are not unlimited, meeting time and email
time
to discuss is scarce and the IETF registries are just not scalable
enough to
support registrations for all the extensions that benefit countless
external
organizations and companies that deploy Internet technology.
New business plans and the required extensions pop up in an endless
fashion.
Big organizations such as you mention have their own resources to
publish
their internal technical documents and white papers if they desire to do
so.

Members of outside organizations are certainly doing everybody a good
service by informing via email of new standards/extensions in other
organization, such as you have done below, but such information need not
overload the Internet RFC process nor overload the Internet registries.

* Hard for SIP developers and right out impossible for application
developers.
Though SIP works fine indeed across the world; SIPit and the SIP Forum
take care of unavoidable problems still encountered by operators,
arguably
increasing SIP complexity for non-Internet purposes is not a good idea,
even
if new SIP extensions are labeled as just "informational". It makes it
even
more difficult for new developers entering the field to ever understand,
let
alone write code for some or all that has been published already about
SIP
for example. Checking right now there are 197 RFCs matching SIP and 78
I-Ds
found by the RFC search engine.

It is time to leave SIP alone and move on to other critical areas for
the
Internet, such as for example the future of applications that will be
discussed in the next IETF technical plenary on Monday, March 28.

With this in mind, I would rather not go into discussing your I-D in
detail,
though at first glance it seems technically correct and if it were
discussed
on the list, may indeed technically qualify as an informational RFC -
just
what I am arguing against.

These are all personal opinions. The chairs and ADs may or may not
discuss
or agree/disagree with. Your I-D is just a good opportunity to raise
these
issues. 

Note that I am not suggesting any voting on your I-D, if it will be
discussed, since for various reasons, the majority of folks attending
the
DISPATCH WG are from various outside organizations with business plans
that
may or may not fully embrace the open e2e Internet; a perfect legitimate
choice for them, but not  in the IETF (I just can't remember when ever
there
was a vote against any new I-D for SIP, since no cost was perceived by
those
voting).

Thanks,
Henry 


On 3/9/11 5:45 PM, "Andrew Allen"  wrote:

> 
> Henry
> 
> Would you care to elaborate and explain your specific concerns?  How
is what
> is defined in these drafts potentially incompatible with the open, e2e
> architectural and in any other way technical conflict with the
Internet?
> 
> The two drafts do the following things:
> 
> One draft simply defines and registers with IANA a URN namespace
according to
> the procedures defined for doing that.
> 
> The other draft simply defines how this URN can be used as an instance
ID as
> required by RFC 5626.
> 
> SIP and the mechanisms in RFC 5626 are already being used in networks
defined
> by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA
have
> already defined this behavior using their existing device Identifiers
for
> instance ID generation in their specifications. These drafts complete
the
> process doing things the right way by reserving the namespace and
defining the
> usage of it using the IETF defined mechanisms for doing that.
> 
> Andrew
> 
> ----- Original Message -----
> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> Sent: Wednesday, March 09, 2011 03:10 PM
> To: Itsuma TANAKA ; dispatch@ietf.org
> 
> Subject: Re: [dispatch]
Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> 
> Such work in the IETF would require IMO an explanation about its
> compatibility with the open, e2e architectural and other technical
> principles of the Internet. It actually seems to be the opposite.
> 
> For this reason, such work may be better accomplished in other
> organizations, so as to avoid architectural principle and other
technical
> conflicts. Simplicity and scalability comes to mind among other,
though this
> is better discussed elsewhere.
> 
> Thanks,
> 
> Henry Sinnreich
> 
> 
> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
> 
>> Dear All,
>> 
>> I work for NTT DOCOMO, Inc in Japan. and attend various 3GPP meetings
(SA2,
>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming
Expert
>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup meetings.  I
serve
>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>> 
>> I would like to express my support of draft-montemurro-gsma-imei-urn
and the
>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>> 
>> The reasons are exactly same as what Ricky Kaura (Samsung) expressed.
A
>> various 3GPP specifications (IMS Registration in 3GPP TS24.229, IMS
>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and
TS24.292), and
>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are
using
>> these drafts.
>> 
>> These internet drafts are very important to support various key
features in
>> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would
like ask
>> that this work is progressed as quickly as possible in IETF.
>> 
>> 
>> Best regards,
>> 
>> Itsuma Tanaka
>> NTT DOCOMO
>> 
>> 
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
> 
> ---------------------------------------------------------------------
> This transmission (including any attachments) may contain confidential
> information, privileged material (including material protected by the
> solicitor-client or other applicable privileges), or constitute
non-public
> information. Any use of this information by anyone other than the
intended
> recipient is prohibited. If you have received this transmission in
error,
> please immediately reply to the sender and delete this information
from your
> system. Use, dissemination, distribution, or reproduction of this
transmission
> by unintended recipients is not authorized and may be unlawful.






---------------------------------------------------------------------
This transmission (including any attachments) may contain confidential infor=
mation, privileged material (including material protected by the solicitor-c=
lient or other applicable privileges), or constitute non-public information.=
 Any use of this information by anyone other than the intended recipient is=
 prohibited. If you have received this transmission in error, please immedia=
tely reply to the sender and delete this information from your system. Use,=
 dissemination, distribution, or reproduction of this transmission by uninte=
nded recipients is not authorized and may be unlawful.

From partr@cisco.com  Thu Mar 10 18:08:21 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 05FD63A6AC2 for ; Thu, 10 Mar 2011 18:08:21 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -8.977
X-Spam-Level: 
X-Spam-Status: No, score=-8.977 tagged_above=-999 required=5 tests=[AWL=-0.776, BAYES_00=-2.599, EXTRA_MPART_TYPE=1, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GI3lvL62Qq3I for ; Thu, 10 Mar 2011 18:07:57 -0800 (PST)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id 5BCA13A67FF for ; Thu, 10 Mar 2011 18:07:53 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=149082; q=dns/txt; s=iport; t=1299809352; x=1301018952; h=mime-version:subject:date:message-id:references:from:to: cc; bh=Uzze3no6WLZJlhtdbWmH1J4cpXavLqDUNeBBYSCkD1M=; b=gNKptI7exAWp2rlIweXm/jZuYMdndLJKvc0FMQWM358/JEVI3K9i2m7n 2KVZXDIa1q6P1Y2DSHWstVN9xJcvIB5MhpPYW3wKn7xSR8oq4pxYhTqOt NgrTLIw4jAhHwVGWJdcG+Azzi9i+iYNrR2RzMZXuWute73PH95aNYRHAe I=;
X-Files: ATT16304074.jpg, al_corp_h_3d_16mm.jpg : 32899, 32899
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AicCAA8TeU2Q/khMgWdsb2JhbAAwgVtQlgiGSAGGMVgUAQEWJiWlLJw1gw4Hgk0EgXGDMYpx
X-IronPort-AV: E=Sophos;i="4.62,300,1297036800";  d="jpg'145?scan'145,208,217,145";a="78680828"
Received: from ams-core-3.cisco.com ([144.254.72.76]) by ams-iport-1.cisco.com with ESMTP; 11 Mar 2011 02:09:09 +0000
Received: from xbh-bgl-412.cisco.com (xbh-bgl-412.cisco.com [72.163.129.202]) by ams-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p2B2980H012841; Fri, 11 Mar 2011 02:09:08 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-412.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 11 Mar 2011 07:39:08 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----_=_NextPart_001_01CBDF91.4D819962"; type="multipart/alternative"
Date: Fri, 11 Mar 2011 07:39:07 +0530
Message-ID: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvfT4Zg1XuiWYneTJm0hai9UlazlwAPWYDN
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>
From: "Parthasarathi R (partr)" 
To: "Thierry Bessis" , "Paul E. Jones" 
X-OriginalArrivalTime: 11 Mar 2011 02:09:08.0206 (UTC) FILETIME=[4DFFE8E0:01CBDF91]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 02:08:21 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBDF91.4D819962
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_002_01CBDF91.4D819962"


------_=_NextPart_002_01CBDF91.4D819962
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thierry,
=20
Please read inline
=20
Thanks
Partha
________________________________

From: dispatch-bounces@ietf.org on behalf of Thierry Bessis
Sent: Thu 3/10/2011 11:46 PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing


Paul,=20

I agree they are many similitudes ! :-)=20
But the question I think is: Can we make the two approaches compatible ? =


For instance, with our approach, it is fundamental that the weight =
represents a "Relative" occupation, and not an absolute capacity.=20
 The question is how to perform the "relative" operation without =
until understanding the capacity of the peer devices. Say Load balancer =
has 3 devices: 2 devices with 1x capacity and 1 device with 2x capacity. =
Here all devices will be able to provide the "relative" value without =
understanding the capacity of the other device. 

The first reason is that we found that loading continuously all the =
servers at the same utilization rate (same percentage) is the optimal =
way of loading a cluster. (We can discuss that if not agreed).=20
 I agree that same utilization rate is the optimal way but the =
question is how to achieve it.  Round robin algorithm works in case all =
devices have same capacity and same amount of load per call. In reality, =
there are multiple varying factor like call duration, different vareity =
of platform within the cluster (say Generation-1 10 devices and =
Generation-2 5 devies), different resource usage for the call, rate at =
which calls generation. 
=20
The second reason is that a relative occupation is very easy to derive =
from information already available: any system knows where it is in =
terms of CPU or Memory usage.
=20
 Converting CPU or Memory to weight is not explained your draft. =
It is the real algorithm which needs lot of discussion. Because =
preparing the mapping between CPU & Memory to weight will be =
mind-blogging activity 
=20
 In contrast, deriving a left capacity in terms of remaining sessions =
can be difficult: Host can be used by different applications and perform =
different duties, which can change over the day, all of them impacting =
their true capacity at a certain moment. Beside, as already discussed, =
different sessions require different resources, so there is a whole =
complication of dealing with a multitude of call profiles, that (again) =
are not even constant over the day.

Also, we do not want to derive some shedding decision from the weight.
=20
  In case you specify, IETF SoC does not allow and so we want =
another value, I think that we are in same page. If not, I could not =
understand your argument. As you mentioned, all system will know CPU or =
memory usage and it is possible to come up the static or refreshable =
value saying that load shed has to happen when 85% of CPU utilization or =
90% of memory utilization. The different between our proposal is that =
there is no need of two value per message instead 1 value in each =
message & 1 value is static or in periodic manner  
=20
 We want to rely on the OLC mechanism for that. As I mentioned before, =
deriving shedding from the weight is very sensitive to the way weights =
are calculated: we would need to specify exactly how the weight is =
calculated: this is impractical ! Also everybody would need to agree =
about specific threshold: What is the weight value threshold for =
shedding ?: Again this is more difficult, and different servers will =
certainly want to use different thresholds.   Also there is a =
contradiction: whether you are using relative or absolute weights, your =
weight should  zero (100% busy, or 0 sessions available) at the moment =
you are getting overloaded. So if you are 100% busy (or in average no =
room for new sessions), how to you tell how many sessions need to be =
shed ? You could be 100% busy with an offered load of 300% nominal (then =
you need to shed 66%), or be 100% busy at just the nominal load, (then =
you don't need to shed). This again shows this threshold issue I =
mentioned. In our implementation, we used to use to a common =
"utilization" information to do both optimal distribution and shedding, =
but we moved away from that because of those reasons: we found that each =
node indicating its requested rejection rate is much more robust and =
requires very little agreements.=20

Cordially,=20

Thierry


On 09-Mar-2011 22:15, Paul E. Jones wrote:=20

	Thierry,

	=20

	I think there is more in common than not.

	=20

	You compute a "weight" and we compute an "sce" that is, in my opinion, =
more or less the same: it's a value that indicates how much traffic can =
be directed toward the box relative to other peer devices.

	=20

	You want to do "load balancing" whereas we want to do "overload =
prevention".  Preventing overload means distributing load across boxes.  =
We are not proposing to perfectly balance the load, but the load is =
distributed nonetheless.

	=20

	Honestly, I just don't see a lot of difference, other than terminology. =
 The algorithm employed by the device directing traffic, but that =
algorithm is independent of the advertisement of a capacity/weight =
value.

	=20

	Paul

	=20

	From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]=20
	Sent: Wednesday, March 09, 2011 12:26 PM
	To: Paul E. Jones
	Cc: dispatch@ietf.org
	Subject: Re: [dispatch] SIP load balancing

	=20

	Thanks Paul for you detailed answer !=20
=09
	I've looked at your SCE proposal and your comments and my impression is =
that our two proposals are actually very different both from their =
intent perspective, and from their realization perspective.=20
=09
	The intent of our load balancing proposal is clearly continuous load =
balancing. Our suggested weight does try to figure any kind of remaining =
capacity in terms of sessions: it only indicates the instant (slightly =
smoothed) available resource, which integrates CPU, Memory and any other =
constrained resource. For the kind of usage we have in mind, (and for =
example), it does not matter if we have lot of sessions "left" when we =
have very little CPU to register those sessions: it is optimal to =
temporary offload this busy node rather than risking a CPU induced =
rejection or failure. When the node will have more CPU available, it =
will take more sessions, and may be even more than others, because it =
may have now more memory than others. If this node never have more CPU =
available, then its "session" capacity is actually unusable and the =
system correctly reflects that.=20
=09
	Also, the SOC was not intended for "dynamic" load distribution: it only =
provides the feedback when some shedding is expected (So already one or =
multiple nodes are about to be overloaded). Also, integrating the =
information from both SOC and SCE to take the distribution decision =
seems rather complicated to me. In our proposal, the distribution =
algorithm is simple, never changes, and is independent from SOC. We =
selected a node with weights always the same way, and we apply the SOC =
rejection when applicable to this node.
=09
	Cordially,=20
=09
	Thierry
=09
	On 08-Mar-2011 14:47, Paul E. Jones wrote:=20

	Thierry,

	=20

	Please see my comments below in green:

	=20

	=20

	From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]=20
	Sent: Tuesday, March 08, 2011 12:35 PM
	To: Paul E. Jones
	Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
	Subject: Re: [dispatch] SIP load balancing

	=20

	Paul,=20
=09
	Please let me address some of your comments.=20
	I don't know what the SCE is, so please pardon me if I misunderstand =
something.=20
=09
	On 07-Mar-2011 23:05, Paul E. Jones wrote:=20

	Vijay,
	=20
	The Session Capacity Estimate (SCE) idea is actually quite similar to =
the
	"adaptive load balancing" draft, from what I can see.  SCE was proposed =
as a
	solution to the "overload" problem, the approach being one where we try =
to
	avoid overload by selecting a device that reports a higher capacity =
value.
	Thus, one could view it as "overload control through load balancing".  =
We
	didn't explicitly call it load balancing, though, because we were not =
trying
	to balance load (i.e., place the same load on all servers).  Rather, we =
just
	wanted to direct traffic to devices that we knew had available =
capacity.

	Agreed, the goal seems to be the same: avoiding or at least delaying =
overload.=20
	And your other comment too: our proposal does try to balance the load =
on all the servers. The rational is that it is very easy for each server =
to know where it is in terms of resource usage (for instance: I'm 60 =
percent resource loaded at this moment), and this is just the =
information you need to balance the traffic. Now it seems (but it is not =
clear, see below)  that the SCE "available capacity" is actually a real =
time feedback, so the information would be very similar actually. =
(What's the difference ? )

	=20

	PEJ: I think it is very similar.  SCE provides real-time feedback of =
the available resource capacity of a peer device.  The intent is to =
convey a single integer value that represents an estimate of the number =
of sessions the peer device can handle.  Note, this is not "message per =
second" or other, but the number of sessions.  The focus on SCE is on =
devices like "call agents", "session border controllers", stateful SIP =
proxies, etc.=20
=09
=09
=09

	=20
	In practice, the way SCE would be employed is the following:
	* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
	   peer devices
	 * The selector would direct traffic toward the peer devices in a
	   round-robin fashion, taking into consideration the SCE value so
	   as to prevent exhausting resources on any given peer device

	I see a contradiction here: if it is round-robin, it does not take into =
account the SCE value.=20
	Is that something more like a weighted round-robin algorithm ?=20

	=20

	PEJ: As I mention below, the algorithm used to select a peer device is =
not a part of the SCE draft.  One can use any mechanism.  The SCE draft =
merely advertises a session capacity estimate.  We've found round-robin =
to be a simple approach that works, as it distributes load relatively =
uniformly, but stops directing traffic toward devices that report an SCE =
value that is below a specified threshold. Other input could be =
considered, but we deliberately left the algorithm outside the scope of =
SCE.
=09
=09
=09

	 * Once the peer's SCE value dropped below a given "low-water mark",
	   the peer device would be less favored and not given traffic until
	   the SCE value increases above the low-water mark (OPTIONS "ping"
	   messages may be used to determine the SCE value in the event that
	   there is no other traffic)

	Why would we wait for a lower mark before to influence the distribution =
? Also is that less favored, or given no traffic ?=20

	=20

	PEJ: With our use of SCE, we're not trying to "balance" the load, per =
se.  One certainly could do that, but in order to balance the load on =
peer devices when they are out of balance, one must direct multiple new =
sessions toward the same device.  That is not always desirable since new =
sessions usually consume more resources temporarily than established =
sessions.  So, we try to burden a peer device by throwing a number of =
new sessions toward the peer by distributing them in a round-robin =
fashion.  Again, though, the approach taken is outside the scope of the =
SCE draft.

	=20

=09
	The later has stability issues.  And why do we need ping Options when =
there is no traffic ?
	 If there is no traffic (or very little), should we care about optimal =
distribution from this source ?

	=20

	PEJ: The assumption is that a peer device might also have several =
peers.  So, while device A1 may not be directing traffic toward device =
B, perhaps A2, A3, and A4 are.  We can use OPTIONS "ping" messages to =
get current SCE information.

=09
=09
=09
=09

	 * The current SOC overload control mechanism would be used to
	   to prevent message overloading, as SCE only focuses on resource
	   overloading (e.g., memory, CPU, disk space, DSPs)

	I don't see this difference here (message overload versus resource =
overload ? ) . To me, SOC focuses on source shedding, while SCE seems to =
focus on improved load distribution, just like our proposal.

	=20

	PEJ: A peer device might be able to handle 18,000 new sessions, but =
perhaps it can only handle them at a rate of 100 per second.  We would =
use SCE to determine that a peer can handle new sessions, while we would =
use the methods SOC is defining to prevent message overload so we do not =
send 200 INVITEs per second to the peer.  (Use of the SOC mechanism =
might prove to work better than our current round-robin approach, but =
we've not tested it.)=20
=09
=09
=09

	=20
	Note that the "round-robin" approach may or may not be preferred and is =
not
	required in our draft.  I've found this approach works pretty well, but =
one
	may prefer to use a particular weighting that is more complex.  I don't
	think local decisions in the box need to be subject to standardization: =
we
	just need to standardize the signaling to enable such decisions.

	I believe the signaling is not enough. The deep meaning of the feedback =
parameter needs to be clarified too.=20
	For example, in our proposal, the feedback is simply the current =
utilization rate of the node (very simple).

	=20

	PEJ: That sounds similar to SCE.  An SCE-capable device advertises to =
its peer how many new sessions it can handle given the number of =
sessions it is currently handling.=20
=09
	Also the expectation needs to be clarified because it has a very =
important impact on the recommended behavior in case of overload.=20
	The solution that we propose constantly tries to balance load in the =
cluster of servers. Therefore, in case the SOC shedding kicks in, there =
is NOT NEED to repeat the message elsewhere in the cluster because it is =
balanced already (So no other server has more capacity). This is a huge =
advantage for the network stability as message repeat tends to spread =
the overload and collapse the cluster.=20
=09
=09
=09

	In any case, it certainly looks like we're thinking along the same =
lines.

	Yes I agree.=20
=09
=09
=09

	PEJ: If you haven't already, have a look at our draft:

	http://tools.ietf.org/html/draft-jones-sip-overload-sce-00 =
=20

	=20

	It is now expired, but your draft suggests there is still life in this =
approach.  We should look to harmonize them, perhaps.

	=20

	Paul

	=20

=09
	Cordially,=20
=09
	Thierry
=09
=09
=09

	=20
	Paul
	=20

		-----Original Message-----
		From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
		Behalf Of Vijay K. Gurbani
		Sent: Monday, March 07, 2011 6:38 PM
		To: dispatch@ietf.org
		Subject: Re: [dispatch] SIP load balancing
		=20
		Paul: Thank you for your input.  More inline.
		=20
		On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

			Vijay,
			IIUC, the algorithm you have proposed is similar to the one Paul =
Jones
			has proposed - with the main difference being the numeric range of =
the
			weight. Do I have that right?

		=20
		I believe that Paul J.'s algorithm is more suited towards overload
		control, not load balancing (and I will await Paul J.'s correction to =
my
		assertion).
		=20
		The intent in draft-bessis-dispatch-adaptive-load-balancing is to
		*always* report the current load to give the upstream SIP proxy an
		indication of the health of the SIP server in the cluster.  If one =
waits
		until overload is imminent, then it may be too late since a sudden and
		unanticipated burst of messages will put the SIP server in overload
		rather quickly.
		=20

			It isn't clear to me from your draft if you are specifying the
			algorithm by which the relative weights are used to determine the
			probability of a request going to one server or another.

		=20
		Yes, same way as RFC2782 does it.
		=20

			ISTM that it *does* matter. And it especially matters if there is an
			implication that the upstream node's algorithm may depend on some
			assumption or knowledge about the downstream nodes.

		=20
		No knowledge, just a reported number that the upstream SIP proxy can
		treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
		starts off with the SRV indicators in RFC2782, without any regard to
		additional knowledge or assumption about downstream nodes.
		dispatch-adaptive-load-balancing is trying to use the same model, =
except
		render the weight to be sensitive to the current traffic in the =
cluster
		in real-time.
		=20

			And of course this method of reporting weights isn't sufficient to
			drive a balancing algorithm that is sensitive to some requests
			requiring more or different resources than others. (That seems to be =
a
			real problem, though I don't know if we have a real solution for
			it.)

		=20
		Yes, that is a real problem, as is the problem of a SIP server in a
		cluster that has a low weight and is thus not contacted with enough
		frequency to merit more participation to the goodput.
		=20
		It is well known in SIP that some methods are more onerous to process
		than others --- from a proxying point of view, processing an INVITE =
with
		forking and response aggregation is far more detrimental to resource
		usage than is processing a BYE.  Some existing literature techniques
		thus add a coefficient to the weight of sending an INVITE versus a =
BYE.
		At this time, the process outlined in dispatch-adaptive-load-balancing
		treats all requests the same, but we can definitely look at ways to
		change this behaviour.
		=20
		Thanks,
		=20
		- vijay
		--
		Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
		1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
		Email: vkg@{bell-labs.com,acm.org} =
  / =
vijay.gurbani@alcatel-lucent.com
		Web:   http://ect.bell-labs.com/who/vkg/
		_______________________________________________
		dispatch mailing list
		dispatch@ietf.org
		https://www.ietf.org/mailman/listinfo/dispatch

	=20
	_______________________________________________
	dispatch mailing list
	dispatch@ietf.org
	https://www.ietf.org/mailman/listinfo/dispatch

	=20

	--=20
	--=20
	Cordially,=20
	Thierry Bessis=20

	  IMS Solution: Architecture & Product Management - ALTA Member
	  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
	=20
	  1960 Lucent Lane - Room 6L-517
	  Naperville, IL  60563  U S
	  Tel/Fax: +1 630 979 7989
	  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
=09
	My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
	My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
	ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
=09
	Conference information:
	2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
	others countries see: http://all.alcatel-lucent.com/audio
=09
	Upcoming planned Business Trip: 12 April - 14 April 2011
	Upcoming planned Vacation: None=20

	=20
	=20
	_______________________________________________
	dispatch mailing list
	dispatch@ietf.org
	https://www.ietf.org/mailman/listinfo/dispatch

	=20

	--=20
	--=20
	Cordially,=20
	Thierry Bessis=20

	  IMS Solution: Architecture & Product Management - ALTA Member
	  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
=09
	  1960 Lucent Lane - Room 6L-517
	  Naperville, IL  60563  U S
	  Tel/Fax: +1 630 979 7989
	  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
=09
	My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
	My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
	ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
=09
	Conference information:
	2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
	others countries see: http://all.alcatel-lucent.com/audio
=09
	Upcoming planned Business Trip: 12 April - 14 April 2011
	Upcoming planned Vacation: None=20

=09
	_______________________________________________
	dispatch mailing list
	dispatch@ietf.org
	https://www.ietf.org/mailman/listinfo/dispatch


--=20
--=20
Cordially,=20
Thierry Bessis=20
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
=20
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None=20

------_=_NextPart_002_01CBDF91.4D819962
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thierry's signature=0A=
=0A=
=0A=
=0A=

=0A=

Thierry,
=0A=

 
=0A=

Please read =
inline
=0A=

 
=0A=

Thanks
=0A=

Partha
=0A=

=0A=

=0A=
From: dispatch-bounces@ietf.org on =
behalf of Thierry Bessis
Sent: Thu 3/10/2011 11:46 =
PM
To: Paul E. Jones
Cc: =
dispatch@ietf.org
Subject: Re: [dispatch] SIP load =
balancing

=0A=

Paul, 

I agree they are many similitudes ! :-) 
But the question I =
think is: Can we make the two approaches compatible ? 

For =
instance, with our approach, it is fundamental that the weight =
represents a "Relative" occupation, and not an absolute capacity. 
=0A=

<Partha> The question is how to perform the "relative" =
operation without until understanding the capacity of the peer =
devices. Say Load balancer has 3 devices: 2 devices with 1x =
capacity and 1 device with 2x capacity. Here all devices will be able to =
provide the "relative" value without understanding the capacity of the =
other device. </Partha>
=0A=


The first reason is that we found that loading continuously all =
the servers at the same utilization rate (same percentage) is the =
optimal way of loading a cluster. (We can discuss that if not agreed). =

=0A=

<Partha> I agree that same utilization rate is the =
optimal way but the question is how to achieve it.  Round =
robin algorithm works in case all devices have same capacity and same =
amount of load per call. In reality, there are multiple varying factor =
like call duration, different vareity of platform within the =
cluster (say Generation-1 10 devices and Generation-2 5 devies), =
different resource usage for the call, rate at which calls generation. =
</Partha>
=0A=

 
=0A=

The second reason is that a relative occupation is very easy to =
derive from information already available: any system knows where it is =
in terms of CPU or Memory usage.
=0A=

 
=0A=

<Partha> Converting CPU or Memory to weight is not explained =
your draft. It is the real algorithm which needs lot of discussion. =
Because preparing the mapping between CPU & Memory to weight =
will be mind-blogging activity </Partha>
=0A=

 
=0A=

 In contrast, deriving a left capacity in terms of remaining =
sessions can be difficult: Host can be used by different applications =
and perform different duties, which can change over the day, all of them =
impacting their true capacity at a certain moment. Beside, as already =
discussed, different sessions require different resources, so there is a =
whole complication of dealing with a multitude of call profiles, that =
(again) are not even constant over the day.

Also, we do not want =
to derive some shedding decision from the weight.
=0A=

 
=0A=

<Partha>  In case you specify, IETF SoC does not allow =
and so we want another value, I think that we are in same page. If not, =
I could not understand your argument. As you mentioned, all =
system will know CPU or memory usage and it is possible to come up the =
static or refreshable value saying that load shed has to happen when 85% =
of CPU utilization or 90% of memory utilization. The different =
between our proposal is that there is no need of =
two value per message instead 1 value in each message & =
1 value is static or in periodic =
manner  </Partha>
=0A=

 
=0A=

 We want to rely on the OLC mechanism for that. As I mentioned =
before, deriving shedding from the weight is very sensitive to the way =
weights are calculated: we would need to specify exactly how the weight =
is calculated: this is impractical ! Also everybody would need to agree =
about specific threshold: What is the weight value threshold for =
shedding ?: Again this is more difficult, and different servers will =
certainly want to use different thresholds.   Also there is a =
contradiction: whether you are using relative or absolute weights, your =
weight should  zero (100% busy, or 0 sessions available) at the =
moment you are getting overloaded. So if you are 100% busy (or in =
average no room for new sessions), how to you tell how many sessions =
need to be shed ? You could be 100% busy with an offered load of 300% =
nominal (then you need to shed 66%), or be 100% busy at just the nominal =
load, (then you don't need to shed). This again shows this threshold =
issue I mentioned. In our implementation, we used to use to a common =
"utilization" information to do both optimal distribution and shedding, =
but we moved away from that because of those reasons: we found that each =
node indicating its requested rejection rate is much more robust and =
requires very little agreements. 

Cordially, =


Thierry


On 09-Mar-2011 22:15, Paul E. Jones wrote: =

=0A=

=0A=
=0A=
=0A=

=0A=

Thierry,
=0A=

 
=0A=

I think there is more in common than not.
=0A=

 
=0A=

You compute a “weight” and we compute an =
“sce” that is, in my opinion, more or less the same: =
it’s a value that indicates how much traffic can be directed =
toward the box relative to other peer devices.
=0A=

 
=0A=

You want to do “load balancing” whereas we want to do =
“overload prevention”.  Preventing overload means =
distributing load across boxes.  We are not proposing to perfectly =
balance the load, but the load is distributed nonetheless.
=0A=

 
=0A=

Honestly, I just don’t see a lot of difference, other than =
terminology.  The algorithm employed by the device directing =
traffic, but that algorithm is independent of the advertisement of a =
capacity/weight value.
=0A=

 
=0A=

Paul
=0A=

 
=0A=

=0A=

=0A=

=0A=

From: Thierry Bessis [mailto:Thierry.Bessis@a=
lcatel-lucent.com] 
Sent: Wednesday, March 09, 2011 12:26 =
PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
=0A=

 
=0A=

Thanks Paul for you detailed answer ! 

I've =
looked at your SCE proposal and your comments and my impression is that =
our two proposals are actually very different both from their intent =
perspective, and from their realization perspective. 

The intent =
of our load balancing proposal is clearly continuous load balancing. Our =
suggested weight does try to figure any kind of remaining capacity in =
terms of sessions: it only indicates the instant (slightly smoothed) =
available resource, which integrates CPU, Memory and any other =
constrained resource. For the kind of usage we have in mind, (and for =
example), it does not matter if we have lot of sessions "left" when we =
have very little CPU to register those sessions: it is optimal to =
temporary offload this busy node rather than risking a CPU induced =
rejection or failure. When the node will have more CPU available, it =
will take more sessions, and may be even more than others, because it =
may have now more memory than others. If this node never have more CPU =
available, then its "session" capacity is actually unusable and the =
system correctly reflects that. 

Also, the SOC was not intended =
for "dynamic" load distribution: it only provides the feedback when some =
shedding is expected (So already one or multiple nodes are about to be =
overloaded). Also, integrating the information from both SOC and SCE to =
take the distribution decision seems rather complicated to me. In our =
proposal, the distribution algorithm is simple, never changes, and is =
independent from SOC. We selected a node with weights always the same =
way, and we apply the SOC rejection when applicable to this =
node.

Cordially, 

Thierry

On 08-Mar-2011 14:47, =
Paul E. Jones wrote: 
=0A=

Thierry,
=0A=

 
=0A=

Please see my =
comments below in green:
=0A=

 
=0A=

 
=0A=

=0A=

=0A=

=0A=

From: Thierry Bessis [mailto:Thierry.Bessis@a=
lcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 =
PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
=0A=

 
=0A=

Paul, 

Please let me address some of your =
comments. 
I don't know what the SCE is, so please pardon me if I =
misunderstand something. 

On 07-Mar-2011 23:05, Paul E. Jones =
wrote: 
Vijay,
 
The Session Capacity =
Estimate (SCE) idea is actually quite similar to the
"adaptive =
load balancing" draft, from what I can see.  SCE was proposed as =
a
solution to the "overload" problem, the approach being one =
where we try to
avoid overload by selecting a device that =
reports a higher capacity value.
Thus, one could view it as =
"overload control through load balancing".  We
didn't =
explicitly call it load balancing, though, because we were not =
trying
to balance load (i.e., place the same load on all =
servers).  Rather, we just
wanted to direct traffic to =
devices that we knew had available capacity.
=0A=

Agreed, the goal seems to be the same: avoiding or =
at least delaying overload. 
And your other comment too: our proposal =
does try to balance the load on all the servers. The rational is that it =
is very easy for each server to know where it is in terms of resource =
usage (for instance: I'm 60 percent resource loaded at this moment), and =
this is just the information you need to balance the traffic. Now it =
seems (but it is not clear, see below)  that the SCE "available =
capacity" is actually a real time feedback, so the information would be =
very similar actually. (What's the difference ? )
=0A=

 
=0A=

PEJ: I think it is very similar.  SCE provides real-time =
feedback of the available resource capacity of a peer device.  The =
intent is to convey a single integer value that represents an estimate =
of the number of sessions the peer device can handle.  Note, this =
is not “message per second” or other, but the number of =
sessions.  The focus on SCE is on devices like “call =
agents”, “session border controllers”, stateful SIP =
proxies, etc. 


 
In practice, =
the way SCE would be employed is the following:
* SIP proxy, =
B2BUA, etc. (selector) would receive SCE capacity =
from
   peer devices
 * The selector would =
direct traffic toward the peer devices in a
   =
round-robin fashion, taking into consideration the SCE value =
so
   as to prevent exhausting resources on any =
given peer device
=0A=

I see a contradiction here: if it is round-robin, =
it does not take into account the SCE value. 
Is that something more =
like a weighted round-robin algorithm ? 
=0A=

 
=0A=

PEJ: As I mention below, the algorithm used to select a peer =
device is not a part of the SCE draft.  One can use any =
mechanism.  The SCE draft merely advertises a session capacity =
estimate.  We’ve found round-robin to be a simple approach =
that works, as it distributes load relatively uniformly, but stops =
directing traffic toward devices that report an SCE value that is below =
a specified threshold. Other input could be considered, but we =
deliberately left the algorithm outside the scope of =
SCE.


 * Once the peer's SCE value dropped =
below a given "low-water mark",
   the peer device =
would be less favored and not given traffic until
   =
the SCE value increases above the low-water mark (OPTIONS =
"ping"
   messages may be used to determine the SCE =
value in the event that
   there is no other =
traffic)
=0A=

Why would we wait for a lower mark before to =
influence the distribution ? Also is that less favored, or given no =
traffic ? 
=0A=

 
=0A=

PEJ: With our use of SCE, we’re not trying to =
“balance” the load, per se.  One certainly could do =
that, but in order to balance the load on peer devices when they are out =
of balance, one must direct multiple new sessions toward the same =
device.  That is not always desirable since new sessions usually =
consume more resources temporarily than established sessions.  So, =
we try to burden a peer device by throwing a number of new sessions =
toward the peer by distributing them in a round-robin fashion.  =
Again, though, the approach taken is outside the scope of the SCE =
draft.
=0A=

 
=0A=


The later has stability issues.  And why =
do we need ping Options when there is no traffic ?
 If there is =
no traffic (or very little), should we care about optimal distribution =
from this source ?
=0A=

 
=0A=

PEJ: The assumption is that a peer device might also have several =
peers.  So, while device A1 may not be directing traffic toward =
device B, perhaps A2, A3, and A4 are.  We can use OPTIONS =
“ping” messages to get current SCE information.
=0A=




 * The current SOC =
overload control mechanism would be used to
   to =
prevent message overloading, as SCE only focuses on =
resource
   overloading (e.g., memory, CPU, disk =
space, DSPs)
=0A=

I don't see this difference here (message overload =
versus resource overload ? ) . To me, SOC focuses on source shedding, =
while SCE seems to focus on improved load distribution, just like our =
proposal.
=0A=

 
=0A=

PEJ: A peer device might be able to handle 18,000 new sessions, =
but perhaps it can only handle them at a rate of 100 per second.  =
We would use SCE to determine that a peer can handle new sessions, while =
we would use the methods SOC is defining to prevent message overload so =
we do not send 200 INVITEs per second to the peer.  (Use of the SOC =
mechanism might prove to work better than our current round-robin =
approach, but we’ve not tested it.) =



 
Note that the "round-robin" =
approach may or may not be preferred and is not
required in =
our draft.  I've found this approach works pretty well, but =
one
may prefer to use a particular weighting that is more =
complex.  I don't
think local decisions in the box need =
to be subject to standardization: we
just need to standardize =
the signaling to enable such decisions.
=0A=

I believe the signaling is not enough. The deep =
meaning of the feedback parameter needs to be clarified too. 
For =
example, in our proposal, the feedback is simply the current utilization =
rate of the node (very simple).
=0A=

 
=0A=

PEJ: That sounds similar to SCE.  An SCE-capable device =
advertises to its peer how many new sessions it can handle given the =
number of sessions it is currently handling. 

Also the =
expectation needs to be clarified because it has a very important impact =
on the recommended behavior in case of overload. 
The solution that =
we propose constantly tries to balance load in the cluster of servers. =
Therefore, in case the SOC shedding kicks in, there is NOT NEED to =
repeat the message elsewhere in the cluster because it is balanced =
already (So no other server has more capacity). This is a huge advantage =
for the network stability as message repeat tends to spread the overload =
and collapse the cluster. 


In any case, it certainly =
looks like we're thinking along the same lines.
=0A=

Yes I agree. 


=0A=

PEJ: If you haven’t already, have a look at our =
draft:
=0A=

http://tools.ietf.org/html/draft-jones-sip-overload-sce-00=

=0A=

 
=0A=

It is now expired, but your draft suggests there is still life in =
this approach.  We should look to harmonize them, =
perhaps.
=0A=

 
=0A=

Paul
=0A=

 
=0A=


Cordially, =


Thierry


 
Paul
 =
;
=0A=

-----Original Message-----
From: dispatch-bounces@ietf.org =
[mailto:dispatch-bounces@ietf.or=
g] On
Behalf Of Vijay K. Gurbani
Sent: Monday, =
March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject=
: Re: [dispatch] SIP load balancing
 
Paul: =
Thank you for your input.  More =
inline.
 
On 03/07/2011 05:14 PM, Paul Kyzivat =
wrote:
=0A=

Vijay,
IIUC, the algorithm you have proposed is =
similar to the one Paul Jones
has proposed - with the main =
difference being the numeric range of the
weight. Do I have =
that right?
 
I believe that Paul =
J.'s algorithm is more suited towards overload
control, not =
load balancing (and I will await Paul J.'s correction to =
my
assertion).
 
The intent in =
draft-bessis-dispatch-adaptive-load-balancing is to
*always* =
report the current load to give the upstream SIP proxy =
an
indication of the health of the SIP server in the =
cluster.  If one waits
until overload is imminent, then =
it may be too late since a sudden and
unanticipated burst of =
messages will put the SIP server in overload
rather =
quickly.
 
=0A=

It isn't =
clear to me from your draft if you are specifying =
the
algorithm by which the relative weights are used to =
determine the
probability of a request going to one server or =
another.
 
Yes, same way as =
RFC2782 does it.
 
=0A=

ISTM that =
it *does* matter. And it especially matters if there is =
an
implication that the upstream node's algorithm may depend =
on some
assumption or knowledge about the downstream =
nodes.
 
No knowledge, just a =
reported number that the upstream SIP proxy can
treat as a =
RFC2782-like weight indicator.  Today, =
rfc3263-resolution
starts off with the SRV indicators in =
RFC2782, without any regard to
additional knowledge or =
assumption about downstream =
nodes.
dispatch-adaptive-load-balancing is trying to use the =
same model, except
render the weight to be sensitive to the =
current traffic in the cluster
in =
real-time.
 
=0A=

And of =
course this method of reporting weights isn't sufficient =
to
drive a balancing algorithm that is sensitive to some =
requests
requiring more or different resources than others. =
(That seems to be a
real problem, though I don't know if we =
have a real solution =
for
it.)
 
Yes, that is =
a real problem, as is the problem of a SIP server in a
cluster =
that has a low weight and is thus not contacted with =
enough
frequency to merit more participation to the =
goodput.
 
It is well known in SIP that some =
methods are more onerous to process
than others --- from a =
proxying point of view, processing an INVITE with
forking and =
response aggregation is far more detrimental to resource
usage =
than is processing a BYE.  Some existing literature =
techniques
thus add a coefficient to the weight of sending an =
INVITE versus a BYE.
At this time, the process outlined in =
dispatch-adaptive-load-balancing
treats all requests the same, =
but we can definitely look at ways to
change this =
behaviour.
 
Thanks,
 
=
- vijay
--
Vijay K. Gurbani, Bell Laboratories, =
Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, =
Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.or=
g} / vijay.gurbani@alcatel-lu=
cent.com
Web:   http://ect.bell-labs.com/who/v=
kg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.=
org/mailman/listinfo/dispatch
 
_______________________________________________
dispatch =
mailing list
dispatch@ietf.org
https://www.ietf.=
org/mailman/listinfo/dispatch
=0A=

 
=0A=

=0A=

-- 
-- =

Cordially, 
Thierry Bessis 
=0A=

  IMS Solution: Architecture & =
Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 =
Lucent Lane - Room 6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis    - Yahoo IM:     =
thierry_bessis

My WEB =
page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - =
(F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming =
planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None =

 
 
__________________=
_____________________________
dispatch mailing =
list
dispatch@ietf.org
https://www.ietf.=
org/mailman/listinfo/dispatch
=0A=

 
=0A=

=0A=

-- 
-- =

Cordially, 
Thierry Bessis 
=0A=

  IMS Solution: Architecture & =
Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 =
Lucent Lane - Room 6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis    - Yahoo IM:     =
thierry_bessis

My WEB =
page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - =
(F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming =
planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None =

=0A=
_______________________________________________=0A=
dispatch mailing list=0A=
dispatch@ietf.org=0A=
https://www.ietf.=
org/mailman/listinfo/dispatch=0A=

=0A=


=0A=

-- 
-- =

Cordially, 
Thierry =
Bessis =0A=

=0A=

  IMS Solution: =
Architecture & Product Management - ALTA Member
  =
Organization: ALU > Carrier group > Wireline BU =
>  IMS Sol Arch, SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis    - Yahoo IM:     =
thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference =
information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - =
Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: =
12 April - 14 April 2011
Upcoming planned Vacation: None =


------_=_NextPart_002_01CBDF91.4D819962--

------_=_NextPart_001_01CBDF91.4D819962
Content-Type: image/jpeg;
	name="ATT16304074.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Description: ATT16304074.jpg
Content-Location: ATT16304074.jpg
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------_=_NextPart_001_01CBDF91.4D819962
Content-Type: image/jpeg;
	name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Description: al_corp_h_3d_16mm.jpg
Content-Disposition: inline;
	filename="al_corp_h_3d_16mm.jpg"
Content-Location: al_corp_h_3d_16mm.jpg
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------_=_NextPart_001_01CBDF91.4D819962--

From christer.holmberg@ericsson.com  Thu Mar 10 22:55:52 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 898E43A68EB for ; Thu, 10 Mar 2011 22:55:52 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.558
X-Spam-Level: 
X-Spam-Status: No, score=-6.558 tagged_above=-999 required=5 tests=[AWL=0.041,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id v1FPMrAgCsB3 for ; Thu, 10 Mar 2011 22:55:48 -0800 (PST)
Received: from mailgw10.se.ericsson.net (mailgw10.se.ericsson.net [193.180.251.61]) by core3.amsl.com (Postfix) with ESMTP id 4EFC23A687C for ; Thu, 10 Mar 2011 22:55:48 -0800 (PST)
X-AuditID: c1b4fb3d-b7bbbae000005311-0e-4d79c7c2d87e
Received: from esessmw0191.eemea.ericsson.se (Unknown_Domain [153.88.253.124]) by mailgw10.se.ericsson.net (Symantec Mail Security) with SMTP id 48.DF.21265.2C7C97D4; Fri, 11 Mar 2011 07:57:06 +0100 (CET)
Received: from ESESSCMS0356.eemea.ericsson.se ([169.254.1.30]) by esessmw0191.eemea.ericsson.se ([153.88.115.84]) with mapi; Fri, 11 Mar 2011 07:57:06 +0100
From: Christer Holmberg 
To: Andrew Allen , Henry Sinnreich , "tanakai@nttdocomo.co.jp" , "dispatch@ietf.org" 
Date: Fri, 11 Mar 2011 07:57:05 +0100
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvfRQM5LNYNOSZVz0atfnKVY5yxgAAIeg3wABRgSYA=
Message-ID: <7F2072F1E0DE894DA4B517B93C6A05851948FF2D4A@ESESSCMS0356.eemea.ericsson.se>
References:  
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAA==
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 06:55:52 -0000

Hi,

I agree with Andrew.

People have invested, and are investing, lots of resources into SIP based s=
ystems. They are not suddently going to throw everything away just because =
some people in IETF suddently thinks there are more "fun and sexy" stuff to=
 work on than SIP...

Also, regarding the agreement between IETF and 3GPP, if people have issues =
with that agreement I don't think that people who are working based on that=
 agreement should have to suffer.

Instead, people having issues with the agreement should bring those issues =
to a more appropriate instance.

Regards,

Christer


> -----Original Message-----
> From: dispatch-bounces@ietf.org=20
> [mailto:dispatch-bounces@ietf.org] On Behalf Of Andrew Allen
> Sent: 11. maaliskuuta 2011 0:04
> To: Henry Sinnreich; tanakai@nttdocomo.co.jp; dispatch@ietf.org
> Subject: Re: [dispatch]=20
> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>=20
> Hi Henry,
>=20
> I think then that it's a fair summary of your email to say=20
> that your objections are of a political and philosophical=20
> nature related to your view of what constitutes the internet=20
> and whether the IETF should be involved in enhancing IETF=20
> protocols for stakeholders that utilize those IETF protocols=20
> for deployments that some in the community might not consider=20
> "the Internet" or "the pure Internet" or the "open e2e=20
> Internet" rather than technical issues with the contents of=20
> the drafts.
>=20
> While I don't want to start a general debate on what is "the Internet"
> here I will state that in my view it's very difficult to=20
> exclude any particular stakeholders and their deployments of=20
> IETF protocols from being part of the Internet when those=20
> stakeholders are in many cases the very same established=20
> telecommunications companies or ISPs that the vast majority=20
> of the population in most countries use to obtain access to=20
> the Internet.=20
>=20
> My understanding also is that IETF nearly a decade ago agreed=20
> to cooperate with 3GPP (and other standards bodies) in=20
> developing and enhancing internet protocols and that IETF=20
> specifically requested (more like insisted) that any=20
> extensions or enhancements to IETF protocols required by 3GPP=20
> be done within IETF according to the procedures defined by=20
> IETF for enhancing IETF protocols and for registering namespaces etc.
> These drafts simply comply with that agreed way of working=20
> between IETF and 3GPP. To my knowledge this working agreement=20
> is still in effect. I don't think it is the role of the=20
> dispatch WG to determine which applications or deployments of=20
> SIP to support or cooperate with counter to agreements=20
> previously made at IESG level.
>=20
> Refusing cooperation based on what is considered not to be=20
> part of the "open e2e Internet" is only likely to set up two=20
> (or more) competing versions of the Internet resulting in the=20
> 21st century equivalent of the standard rail gauge vs broad=20
> rail gauge communications interoperability issue from the=20
> 19th century - effectively killing the goal of an open e2e=20
> internet! As Dale indicated what is considered a walled=20
> garden today may not be walled in the future (if I remember=20
> some very significant walls came tumbling down during just=20
> the last decade of the 20th century alone).
>=20
> Different business models will be resolved through the market=20
> not by creating competing incompatible technical solutions.=20
>=20
> In terms of the time used on discussion within dispatch I=20
> don't think there is a need for much more discussion in=20
> dispatch. These drafts could go ahead as AD sponsored without=20
> further WG involvement. As I understand it the IMEI URN=20
> definition draft did not as part of the URN registration=20
> process need to even be discussed in dispatch at all=20
> (although Dale Worley's review was invaluable in improving=20
> the draft technically) as the NIDs list was the fora for this=20
> discussion and where previous versions of that draft were discussed.
>=20
> It should be noted all we are talking about with regard to=20
> these drafts is the registration of a URN namespace and=20
> definition of its usage as an Instance ID as required by RFC=20
> 5626. If we can't agree that then what future extensions can we agree?
>=20
> Andrew
>=20
> -----Original Message-----
> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> Sent: Thursday, March 10, 2011 11:03 AM
> To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org
> Subject: Re: [dispatch]
> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>=20
> Hi Andrew,
>=20
> I have generic objections to see the publication of yet=20
> another informational RFC about practices in consortiums of=20
> closed networks in the IETF, for the following reasons:
>=20
> * Using the RFC process as a publishing venue for=20
> organizations that don't explicitly do that in support of the=20
> open e2e Internet.
>=20
> * The resources of the IETF are not unlimited, meeting time=20
> and email time to discuss is scarce and the IETF registries=20
> are just not scalable enough to support registrations for all=20
> the extensions that benefit countless external organizations=20
> and companies that deploy Internet technology.
> New business plans and the required extensions pop up in an=20
> endless fashion.
> Big organizations such as you mention have their own=20
> resources to publish their internal technical documents and=20
> white papers if they desire to do so.
>=20
> Members of outside organizations are certainly doing=20
> everybody a good service by informing via email of new=20
> standards/extensions in other organization, such as you have=20
> done below, but such information need not overload the=20
> Internet RFC process nor overload the Internet registries.
>=20
> * Hard for SIP developers and right out impossible for=20
> application developers.
> Though SIP works fine indeed across the world; SIPit and the=20
> SIP Forum take care of unavoidable problems still encountered=20
> by operators, arguably increasing SIP complexity for=20
> non-Internet purposes is not a good idea, even if new SIP=20
> extensions are labeled as just "informational". It makes it=20
> even more difficult for new developers entering the field to=20
> ever understand, let alone write code for some or all that=20
> has been published already about SIP for example. Checking=20
> right now there are 197 RFCs matching SIP and 78 I-Ds found=20
> by the RFC search engine.
>=20
> It is time to leave SIP alone and move on to other critical=20
> areas for the Internet, such as for example the future of=20
> applications that will be discussed in the next IETF=20
> technical plenary on Monday, March 28.
>=20
> With this in mind, I would rather not go into discussing your=20
> I-D in detail, though at first glance it seems technically=20
> correct and if it were discussed on the list, may indeed=20
> technically qualify as an informational RFC - just what I am=20
> arguing against.
>=20
> These are all personal opinions. The chairs and ADs may or=20
> may not discuss or agree/disagree with. Your I-D is just a=20
> good opportunity to raise these issues.=20
>=20
> Note that I am not suggesting any voting on your I-D, if it=20
> will be discussed, since for various reasons, the majority of=20
> folks attending the DISPATCH WG are from various outside=20
> organizations with business plans that may or may not fully=20
> embrace the open e2e Internet; a perfect legitimate choice=20
> for them, but not  in the IETF (I just can't remember when=20
> ever there was a vote against any new I-D for SIP, since no=20
> cost was perceived by those voting).
>=20
> Thanks,
> Henry=20
>=20
>=20
> On 3/9/11 5:45 PM, "Andrew Allen"  wrote:
>=20
> >=20
> > Henry
> >=20
> > Would you care to elaborate and explain your specific concerns?  How
> is what
> > is defined in these drafts potentially incompatible with=20
> the open, e2e=20
> > architectural and in any other way technical conflict with the
> Internet?
> >=20
> > The two drafts do the following things:
> >=20
> > One draft simply defines and registers with IANA a URN namespace
> according to
> > the procedures defined for doing that.
> >=20
> > The other draft simply defines how this URN can be used as=20
> an instance
> ID as
> > required by RFC 5626.
> >=20
> > SIP and the mechanisms in RFC 5626 are already being used=20
> in networks
> defined
> > by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA
> have
> > already defined this behavior using their existing device=20
> Identifiers
> for
> > instance ID generation in their specifications. These=20
> drafts complete
> the
> > process doing things the right way by reserving the namespace and
> defining the
> > usage of it using the IETF defined mechanisms for doing that.
> >=20
> > Andrew
> >=20
> > ----- Original Message -----
> > From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> > Sent: Wednesday, March 09, 2011 03:10 PM
> > To: Itsuma TANAKA ; dispatch@ietf.org=20
> > 
> > Subject: Re: [dispatch]
> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> >=20
> > Such work in the IETF would require IMO an explanation about its=20
> > compatibility with the open, e2e architectural and other technical=20
> > principles of the Internet. It actually seems to be the opposite.
> >=20
> > For this reason, such work may be better accomplished in other=20
> > organizations, so as to avoid architectural principle and other
> technical
> > conflicts. Simplicity and scalability comes to mind among other,
> though this
> > is better discussed elsewhere.
> >=20
> > Thanks,
> >=20
> > Henry Sinnreich
> >=20
> >=20
> > On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
> >=20
> >> Dear All,
> >>=20
> >> I work for NTT DOCOMO, Inc in Japan. and attend various=20
> 3GPP meetings
> (SA2,
> >> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming
> Expert
> >> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup=20
> meetings.  I
> serve
> >> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
> >>=20
> >> I would like to express my support of=20
> draft-montemurro-gsma-imei-urn
> and the
> >> associated draft-allen-dispatch-imei-urn-as-instanceid.
> >>=20
> >> The reasons are exactly same as what Ricky Kaura (Samsung)=20
> expressed.
> A
> >> various 3GPP specifications (IMS Registration in 3GPP=20
> TS24.229, IMS=20
> >> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and
> TS24.292), and
> >> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are
> using
> >> these drafts.
> >>=20
> >> These internet drafts are very important to support various key
> features in
> >> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would
> like ask
> >> that this work is progressed as quickly as possible in IETF.
> >>=20
> >>=20
> >> Best regards,
> >>=20
> >> Itsuma Tanaka
> >> NTT DOCOMO
> >>=20
> >>=20
> >> _______________________________________________
> >> dispatch mailing list
> >> dispatch@ietf.org
> >> https://www.ietf.org/mailman/listinfo/dispatch
> >=20
> >=20
> > _______________________________________________
> > dispatch mailing list
> > dispatch@ietf.org
> > https://www.ietf.org/mailman/listinfo/dispatch
> >=20
> >=20
> ---------------------------------------------------------------------
> > This transmission (including any attachments) may contain=20
> confidential=20
> > information, privileged material (including material=20
> protected by the=20
> > solicitor-client or other applicable privileges), or constitute
> non-public
> > information. Any use of this information by anyone other than the
> intended
> > recipient is prohibited. If you have received this transmission in
> error,
> > please immediately reply to the sender and delete this information
> from your
> > system. Use, dissemination, distribution, or reproduction of this
> transmission
> > by unintended recipients is not authorized and may be unlawful.
>=20
>=20
>=20
>=20
>=20
>=20
> ---------------------------------------------------------------------
> This transmission (including any attachments) may contain=20
> confidential information, privileged material (including=20
> material protected by the solicitor-client or other=20
> applicable privileges), or constitute non-public information.=20
> Any use of this information by anyone other than the intended=20
> recipient is prohibited. If you have received this=20
> transmission in error, please immediately reply to the sender=20
> and delete this information from your system. Use,=20
> dissemination, distribution, or reproduction of this=20
> transmission by unintended recipients is not authorized and=20
> may be unlawful.
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
> =

From paulej@packetizer.com  Fri Mar 11 00:29:38 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A9B9F3A6BEF for ; Fri, 11 Mar 2011 00:29:38 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gonrlOrdWUtU for ; Fri, 11 Mar 2011 00:29:16 -0800 (PST)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id B69443A6BEE for ; Fri, 11 Mar 2011 00:29:15 -0800 (PST)
Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2B8UO7t029133 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Fri, 11 Mar 2011 03:30:30 -0500
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299832230; bh=n4uKvFPy26gryGIm6cvxyEtfgZfPkvO+KL5UCX8oyXY=; h=From:To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type; b=GBFuFdR1mJcd3emyqVsY5/sNuFPU+AZq93hLuPXeD1bYqpjs69XPkLqShQlHj75+f vyidrYEMa6EwSeNeBymQN7a9kR8fywZBORBFouVmRYpElZQgZz1MkRdAwwFjC9+4xw dWIplxL4GF1+XwCVmG3qe2ry6Y18SokDGSOeda18=
From: "Paul E. Jones" 
To: "'Thierry Bessis'" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>
In-Reply-To: <4D791595.2010703@alcatel-lucent.com>
Date: Fri, 11 Mar 2011 03:30:15 -0500
Message-ID: <060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----=_NextPart_000_0608_01CBDF9C.A6D65E30"
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3kCCnyKVgIDWSzZAfKA2HwChasBsAE4aE9uAgiAjeSTa8y5QA==
Content-Language: en-us
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 08:29:38 -0000

This is a multipart message in MIME format.

------=_NextPart_000_0608_01CBDF9C.A6D65E30
Content-Type: multipart/alternative;
	boundary="----=_NextPart_001_0609_01CBDF9C.A6D65E30"


------=_NextPart_001_0609_01CBDF9C.A6D65E30
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

Thierry,

 

The SCE is not a fixed value that remains constant or decrements by 1,
necessarily, with each new session.  It is a value that is calculated
constantly (or at least with some frequency) and provided to peer devices.
This value can be compared with SCE values from other peer devices to
compare the relative capacities of peer devices at any given moment.

 

The specific ways through which the SCE value is computed may vary by
vendor, though the approach we have used is one where we calculate an
estimate of the number of sessions a device can handle given the current
number of sessions given current memory utilization.  We then do the same
for CPU and other system resources.  We then take the minimum value of each
and report that as the SCE value.  As a matter of course, our devices are
constantly measuring these resources, so computing the SCE if trivial.

 

Paul

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Thursday, March 10, 2011 1:17 PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Paul, 

I agree they are many similitudes ! :-) 
But the question I think is: Can we make the two approaches compatible ? 

For instance, with our approach, it is fundamental that the weight
represents a "Relative" occupation, and not an absolute capacity.
The first reason is that we found that loading continuously all the servers
at the same utilization rate (same percentage) is the optimal way of loading
a cluster. (We can discuss that if not agreed). The second reason is that a
relative occupation is very easy to derive from information already
available: any system knows where it is in terms of CPU or Memory usage. In
contrast, deriving a left capacity in terms of remaining sessions can be
difficult: Host can be used by different applications and perform different
duties, which can change over the day, all of them impacting their true
capacity at a certain moment. Beside, as already discussed, different
sessions require different resources, so there is a whole complication of
dealing with a multitude of call profiles, that (again) are not even
constant over the day.

Also, we do not want to derive some shedding decision from the weight. We
want to rely on the OLC mechanism for that. As I mentioned before, deriving
shedding from the weight is very sensitive to the way weights are
calculated: we would need to specify exactly how the weight is calculated:
this is impractical ! Also everybody would need to agree about specific
threshold: What is the weight value threshold for shedding ?: Again this is
more difficult, and different servers will certainly want to use different
thresholds.   Also there is a contradiction: whether you are using relative
or absolute weights, your weight should  zero (100% busy, or 0 sessions
available) at the moment you are getting overloaded. So if you are 100% busy
(or in average no room for new sessions), how to you tell how many sessions
need to be shed ? You could be 100% busy with an offered load of 300%
nominal (then you need to shed 66%), or be 100% busy at just the nominal
load, (then you don't need to shed). This again shows this threshold issue I
mentioned. In our implementation, we used to use to a common "utilization"
information to do both optimal distribution and shedding, but we moved away
from that because of those reasons: we found that each node indicating its
requested rejection rate is much more robust and requires very little
agreements. 

Cordially, 

Thierry


On 09-Mar-2011 22:15, Paul E. Jones wrote: 

Thierry,

 

I think there is more in common than not.

 

You compute a "weight" and we compute an "sce" that is, in my opinion, more
or less the same: it's a value that indicates how much traffic can be
directed toward the box relative to other peer devices.

 

You want to do "load balancing" whereas we want to do "overload prevention".
Preventing overload means distributing load across boxes.  We are not
proposing to perfectly balance the load, but the load is distributed
nonetheless.

 

Honestly, I just don't see a lot of difference, other than terminology.  The
algorithm employed by the device directing traffic, but that algorithm is
independent of the advertisement of a capacity/weight value.

 

Paul

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Wednesday, March 09, 2011 12:26 PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Thanks Paul for you detailed answer ! 

I've looked at your SCE proposal and your comments and my impression is that
our two proposals are actually very different both from their intent
perspective, and from their realization perspective. 

The intent of our load balancing proposal is clearly continuous load
balancing. Our suggested weight does try to figure any kind of remaining
capacity in terms of sessions: it only indicates the instant (slightly
smoothed) available resource, which integrates CPU, Memory and any other
constrained resource. For the kind of usage we have in mind, (and for
example), it does not matter if we have lot of sessions "left" when we have
very little CPU to register those sessions: it is optimal to temporary
offload this busy node rather than risking a CPU induced rejection or
failure. When the node will have more CPU available, it will take more
sessions, and may be even more than others, because it may have now more
memory than others. If this node never have more CPU available, then its
"session" capacity is actually unusable and the system correctly reflects
that. 

Also, the SOC was not intended for "dynamic" load distribution: it only
provides the feedback when some shedding is expected (So already one or
multiple nodes are about to be overloaded). Also, integrating the
information from both SOC and SCE to take the distribution decision seems
rather complicated to me. In our proposal, the distribution algorithm is
simple, never changes, and is independent from SOC. We selected a node with
weights always the same way, and we apply the SOC rejection when applicable
to this node.

Cordially, 

Thierry

On 08-Mar-2011 14:47, Paul E. Jones wrote: 

Thierry,

 

Please see my comments below in green:

 

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Paul, 

Please let me address some of your comments. 
I don't know what the SCE is, so please pardon me if I misunderstand
something. 

On 07-Mar-2011 23:05, Paul E. Jones wrote: 

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to the
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a
solution to the "overload" problem, the approach being one where we try to
avoid overload by selecting a device that reports a higher capacity value.
Thus, one could view it as "overload control through load balancing".  We
didn't explicitly call it load balancing, though, because we were not trying
to balance load (i.e., place the same load on all servers).  Rather, we just
wanted to direct traffic to devices that we knew had available capacity.

Agreed, the goal seems to be the same: avoiding or at least delaying
overload. 
And your other comment too: our proposal does try to balance the load on all
the servers. The rational is that it is very easy for each server to know
where it is in terms of resource usage (for instance: I'm 60 percent
resource loaded at this moment), and this is just the information you need
to balance the traffic. Now it seems (but it is not clear, see below)  that
the SCE "available capacity" is actually a real time feedback, so the
information would be very similar actually. (What's the difference ? )

 

PEJ: I think it is very similar.  SCE provides real-time feedback of the
available resource capacity of a peer device.  The intent is to convey a
single integer value that represents an estimate of the number of sessions
the peer device can handle.  Note, this is not "message per second" or
other, but the number of sessions.  The focus on SCE is on devices like
"call agents", "session border controllers", stateful SIP proxies, etc. 





 
In practice, the way SCE would be employed is the following:
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
   peer devices
 * The selector would direct traffic toward the peer devices in a
   round-robin fashion, taking into consideration the SCE value so
   as to prevent exhausting resources on any given peer device

I see a contradiction here: if it is round-robin, it does not take into
account the SCE value. 
Is that something more like a weighted round-robin algorithm ? 

 

PEJ: As I mention below, the algorithm used to select a peer device is not a
part of the SCE draft.  One can use any mechanism.  The SCE draft merely
advertises a session capacity estimate.  We've found round-robin to be a
simple approach that works, as it distributes load relatively uniformly, but
stops directing traffic toward devices that report an SCE value that is
below a specified threshold. Other input could be considered, but we
deliberately left the algorithm outside the scope of SCE.





 * Once the peer's SCE value dropped below a given "low-water mark",
   the peer device would be less favored and not given traffic until
   the SCE value increases above the low-water mark (OPTIONS "ping"
   messages may be used to determine the SCE value in the event that
   there is no other traffic)

Why would we wait for a lower mark before to influence the distribution ?
Also is that less favored, or given no traffic ? 

 

PEJ: With our use of SCE, we're not trying to "balance" the load, per se.
One certainly could do that, but in order to balance the load on peer
devices when they are out of balance, one must direct multiple new sessions
toward the same device.  That is not always desirable since new sessions
usually consume more resources temporarily than established sessions.  So,
we try to burden a peer device by throwing a number of new sessions toward
the peer by distributing them in a round-robin fashion.  Again, though, the
approach taken is outside the scope of the SCE draft.

 


The later has stability issues.  And why do we need ping Options when there
is no traffic ?
 If there is no traffic (or very little), should we care about optimal
distribution from this source ?

 

PEJ: The assumption is that a peer device might also have several peers.
So, while device A1 may not be directing traffic toward device B, perhaps
A2, A3, and A4 are.  We can use OPTIONS "ping" messages to get current SCE
information.







 * The current SOC overload control mechanism would be used to
   to prevent message overloading, as SCE only focuses on resource
   overloading (e.g., memory, CPU, disk space, DSPs)

I don't see this difference here (message overload versus resource overload
? ) . To me, SOC focuses on source shedding, while SCE seems to focus on
improved load distribution, just like our proposal.

 

PEJ: A peer device might be able to handle 18,000 new sessions, but perhaps
it can only handle them at a rate of 100 per second.  We would use SCE to
determine that a peer can handle new sessions, while we would use the
methods SOC is defining to prevent message overload so we do not send 200
INVITEs per second to the peer.  (Use of the SOC mechanism might prove to
work better than our current round-robin approach, but we've not tested it.)






 
Note that the "round-robin" approach may or may not be preferred and is not
required in our draft.  I've found this approach works pretty well, but one
may prefer to use a particular weighting that is more complex.  I don't
think local decisions in the box need to be subject to standardization: we
just need to standardize the signaling to enable such decisions.

I believe the signaling is not enough. The deep meaning of the feedback
parameter needs to be clarified too. 
For example, in our proposal, the feedback is simply the current utilization
rate of the node (very simple).

 

PEJ: That sounds similar to SCE.  An SCE-capable device advertises to its
peer how many new sessions it can handle given the number of sessions it is
currently handling. 

Also the expectation needs to be clarified because it has a very important
impact on the recommended behavior in case of overload. 
The solution that we propose constantly tries to balance load in the cluster
of servers. Therefore, in case the SOC shedding kicks in, there is NOT NEED
to repeat the message elsewhere in the cluster because it is balanced
already (So no other server has more capacity). This is a huge advantage for
the network stability as message repeat tends to spread the overload and
collapse the cluster. 





In any case, it certainly looks like we're thinking along the same lines.

Yes I agree. 





PEJ: If you haven't already, have a look at our draft:

 
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

 

It is now expired, but your draft suggests there is still life in this
approach.  We should look to harmonize them, perhaps.

 

Paul

 


Cordially, 

Thierry





 
Paul
 

-----Original Message-----
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On
Behalf Of Vijay K. Gurbani
Sent: Monday, March 07, 2011 6:38 PM
To: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Paul: Thank you for your input.  More inline.
 
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul Jones
has proposed - with the main difference being the numeric range of the
weight. Do I have that right?

 
I believe that Paul J.'s algorithm is more suited towards overload
control, not load balancing (and I will await Paul J.'s correction to my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing is to
*always* report the current load to give the upstream SIP proxy an
indication of the health of the SIP server in the cluster.  If one waits
until overload is imminent, then it may be too late since a sudden and
unanticipated burst of messages will put the SIP server in overload
rather quickly.
 

It isn't clear to me from your draft if you are specifying the
algorithm by which the relative weights are used to determine the
probability of a request going to one server or another.

 
Yes, same way as RFC2782 does it.
 

ISTM that it *does* matter. And it especially matters if there is an
implication that the upstream node's algorithm may depend on some
assumption or knowledge about the downstream nodes.

 
No knowledge, just a reported number that the upstream SIP proxy can
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
starts off with the SRV indicators in RFC2782, without any regard to
additional knowledge or assumption about downstream nodes.
dispatch-adaptive-load-balancing is trying to use the same model, except
render the weight to be sensitive to the current traffic in the cluster
in real-time.
 

And of course this method of reporting weights isn't sufficient to
drive a balancing algorithm that is sensitive to some requests
requiring more or different resources than others. (That seems to be a
real problem, though I don't know if we have a real solution for
it.)

 
Yes, that is a real problem, as is the problem of a SIP server in a
cluster that has a low weight and is thus not contacted with enough
frequency to merit more participation to the goodput.
 
It is well known in SIP that some methods are more onerous to process
than others --- from a proxying point of view, processing an INVITE with
forking and response aggregation is far more detrimental to resource
usage than is processing a BYE.  Some existing literature techniques
thus add a coefficient to the weight of sending an INVITE versus a BYE.
At this time, the process outlined in dispatch-adaptive-load-balancing
treats all requests the same, but we can definitely look at ways to
change this behaviour.
 
Thanks,
 
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: vkg@{bell-labs.com,acm.org} 
/ vijay.gurbani@alcatel-lucent.com
Web:   http://ect.bell-labs.com/who/vkg/
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
 
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 

 
 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
 
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 

 
 
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

 

-- 
-- 
Cordially, 
Thierry Bessis 

  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
Alcatel-Lucent
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None 


------=_NextPart_001_0609_01CBDF9C.A6D65E30
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Thierry's signature
Thierry,
 
The SCE is not a fixed value that remains constant or decrements by =
1, necessarily, with each new session.  It is a value that is =
calculated constantly (or at least with some frequency) and provided to =
peer devices.  This value can be compared with SCE values from =
other peer devices to compare the relative capacities of peer devices at =
any given moment.
 
The specific ways through which the SCE value is computed may vary by =
vendor, though the approach we have used is one where we calculate an =
estimate of the number of sessions a device can handle given the current =
number of sessions given current memory utilization.  We then do =
the same for CPU and other system resources.  We then take the =
minimum value of each and report that as the SCE value.  As a =
matter of course, our devices are constantly measuring these resources, =
so computing the SCE if trivial.
 
Paul
 
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] =

Sent: Thursday, March 10, 2011 1:17 PM
To: Paul E. =
Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] =
SIP load balancing
 
Paul, =


I agree they are many similitudes ! :-) 
But the question I think is: Can we =
make the two approaches compatible ? 

For instance, with our =
approach, it is fundamental that the weight represents a =
"Relative" occupation, and not an absolute capacity.
The =
first reason is that we found that loading continuously all the servers =
at the same utilization rate (same percentage) is the optimal way of =
loading a cluster. (We can discuss that if not agreed). The second =
reason is that a relative occupation is very easy to derive from =
information already available: any system knows where it is in terms of =
CPU or Memory usage. In contrast, deriving a left capacity in terms of =
remaining sessions can be difficult: Host can be used by different =
applications and perform different duties, which can change over the =
day, all of them impacting their true capacity at a certain moment. =
Beside, as already discussed, different sessions require different =
resources, so there is a whole complication of dealing with a multitude =
of call profiles, that (again) are not even constant over the =
day.

Also, we do not want to derive some shedding decision from =
the weight. We want to rely on the OLC mechanism for that. As I =
mentioned before, deriving shedding from the weight is very sensitive to =
the way weights are calculated: we would need to specify exactly how the =
weight is calculated: this is impractical ! Also everybody would need to =
agree about specific threshold: What is the weight value threshold for =
shedding ?: Again this is more difficult, and different servers will =
certainly want to use different thresholds.   Also there is a =
contradiction: whether you are using relative or absolute weights, your =
weight should  zero (100% busy, or 0 sessions available) at the =
moment you are getting overloaded. So if you are 100% busy (or in =
average no room for new sessions), how to you tell how many sessions =
need to be shed ? You could be 100% busy with an offered load of 300% =
nominal (then you need to shed 66%), or be 100% busy at just the nominal =
load, (then you don't need to shed). This again shows this threshold =
issue I mentioned. In our implementation, we used to use to a common =
"utilization" information to do both optimal distribution and =
shedding, but we moved away from that because of those reasons: we found =
that each node indicating its requested rejection rate is much more =
robust and requires very little agreements. 

Cordially, =


Thierry


On 09-Mar-2011 22:15, Paul E. Jones wrote: =

Thierry,
 
I think =
there is more in common than not.
 
You =
compute a “weight” and we compute an “sce” that =
is, in my opinion, more or less the same: it’s a value that =
indicates how much traffic can be directed toward the box relative to =
other peer devices.
 
You want =
to do “load balancing” whereas we want to do “overload =
prevention”.  Preventing overload means distributing load =
across boxes.  We are not proposing to perfectly balance the load, =
but the load is distributed nonetheless.
 
Honestly, =
I just don’t see a lot of difference, other than =
terminology.  The algorithm employed by the device directing =
traffic, but that algorithm is independent of the advertisement of a =
capacity/weight value.
 
Paul=

 
From: Thierry Bessis [mailto:Thierry.Bessis@a=
lcatel-lucent.com] 
Sent: Wednesday, March 09, 2011 12:26 =
PM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Thanks Paul =
for you detailed answer ! 

I've looked at your SCE proposal and =
your comments and my impression is that our two proposals are actually =
very different both from their intent perspective, and from their =
realization perspective. 

The intent of our load balancing =
proposal is clearly continuous load balancing. Our suggested weight does =
try to figure any kind of remaining capacity in terms of sessions: it =
only indicates the instant (slightly smoothed) available resource, which =
integrates CPU, Memory and any other constrained resource. For the kind =
of usage we have in mind, (and for example), it does not matter if we =
have lot of sessions "left" when we have very little CPU to =
register those sessions: it is optimal to temporary offload this busy =
node rather than risking a CPU induced rejection or failure. When the =
node will have more CPU available, it will take more sessions, and may =
be even more than others, because it may have now more memory than =
others. If this node never have more CPU available, then its =
"session" capacity is actually unusable and the system =
correctly reflects that. 

Also, the SOC was not intended for =
"dynamic" load distribution: it only provides the feedback =
when some shedding is expected (So already one or multiple nodes are =
about to be overloaded). Also, integrating the information from both SOC =
and SCE to take the distribution decision seems rather complicated to =
me. In our proposal, the distribution algorithm is simple, never =
changes, and is independent from SOC. We selected a node with weights =
always the same way, and we apply the SOC rejection when applicable to =
this node.

Cordially, 

Thierry

On 08-Mar-2011 =
14:47, Paul E. Jones wrote: 
Thierry,
 
Please see =
my comments below in green:
 
 
From: Thierry Bessis [mailto:Thierry.Bessis@a=
lcatel-lucent.com] 
Sent: Tuesday, March 08, 2011 12:35 =
PM
To: Paul E. Jones
Cc: 'Vijay K. Gurbani'; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
 
Paul, =


Please let me address some of your comments. 
I don't know =
what the SCE is, so please pardon me if I misunderstand something. =


On 07-Mar-2011 23:05, Paul E. Jones wrote: =

Vijay,
 
The Session Capacity Estimate (SCE) idea is actually quite similar to =
the
"adaptive load balancing" draft, from =
what I can see.  SCE was proposed as =
a
solution to the "overload" problem, the =
approach being one where we try to
avoid overload =
by selecting a device that reports a higher capacity =
value.
Thus, one could view it as "overload =
control through load balancing".  =
We
didn't explicitly call it load balancing, =
though, because we were not trying
to balance load =
(i.e., place the same load on all servers).  Rather, we =
just
wanted to direct traffic to devices that we =
knew had available capacity.
Agreed, the goal seems to be the same: avoiding or at =
least delaying overload. 
And your other comment too: our proposal =
does try to balance the load on all the servers. The rational is that it =
is very easy for each server to know where it is in terms of resource =
usage (for instance: I'm 60 percent resource loaded at this moment), and =
this is just the information you need to balance the traffic. Now it =
seems (but it is not clear, see below)  that the SCE =
"available capacity" is actually a real time feedback, so the =
information would be very similar actually. (What's the difference ? =
)
 
PEJ: I think it is very similar.  SCE provides real-time =
feedback of the available resource capacity of a peer device.  The =
intent is to convey a single integer value that represents an estimate =
of the number of sessions the peer device can handle.  Note, this =
is not “message per second” or other, but the number of =
sessions.  The focus on SCE is on devices like “call =
agents”, “session border controllers”, stateful SIP =
proxies, etc. =




 
In =
practice, the way SCE would be employed is the =
following:
* SIP proxy, B2BUA, etc. (selector) =
would receive SCE capacity from
   peer =
devices
 * The selector would direct traffic toward =
the peer devices in a
   round-robin =
fashion, taking into consideration the SCE value =
so
   as to prevent exhausting resources =
on any given peer device
I see a =
contradiction here: if it is round-robin, it does not take into account =
the SCE value. 
Is that something more like a weighted round-robin =
algorithm ? 
 
PEJ: As I mention below, the algorithm used to select a peer device =
is not a part of the SCE draft.  One can use any mechanism.  =
The SCE draft merely advertises a session capacity estimate.  =
We’ve found round-robin to be a simple approach that works, as it =
distributes load relatively uniformly, but stops directing traffic =
toward devices that report an SCE value that is below a specified =
threshold. Other input could be considered, but we deliberately left the =
algorithm outside the scope of =
SCE.



 * Once the peer's SCE =
value dropped below a given "low-water =
mark",
   the peer device would be =
less favored and not given traffic =
until
   the SCE value increases above =
the low-water mark (OPTIONS =
"ping"
   messages may be used =
to determine the SCE value in the event =
that
   there is no other =
traffic)
Why would we wait for a =
lower mark before to influence the distribution ? Also is that less =
favored, or given no traffic ? 
 
PEJ: With our use of SCE, we’re not trying to =
“balance” the load, per se.  One certainly could do =
that, but in order to balance the load on peer devices when they are out =
of balance, one must direct multiple new sessions toward the same =
device.  That is not always desirable since new sessions usually =
consume more resources temporarily than established sessions.  So, =
we try to burden a peer device by throwing a number of new sessions =
toward the peer by distributing them in a round-robin fashion.  =
Again, though, the approach taken is outside the scope of the SCE =
draft.
 

The later has =
stability issues.  And why do we need ping Options when there is no =
traffic ?
 If there is no traffic (or very little), should we =
care about optimal distribution from this source ?
 
PEJ: The assumption is that a peer device might also have several =
peers.  So, while device A1 may not be directing traffic toward =
device B, perhaps A2, A3, and A4 are.  We can use OPTIONS =
“ping” messages to get current SCE =
information.




 * The =
current SOC overload control mechanism would be used =
to
   to prevent message overloading, as =
SCE only focuses on resource
   =
overloading (e.g., memory, CPU, disk space, DSPs)
I don't see this difference here (message overload =
versus resource overload ? ) . To me, SOC focuses on source shedding, =
while SCE seems to focus on improved load distribution, just like our =
proposal.
 
PEJ: A peer device might be able to handle 18,000 new sessions, but =
perhaps it can only handle them at a rate of 100 per second.  We =
would use SCE to determine that a peer can handle new sessions, while we =
would use the methods SOC is defining to prevent message overload so we =
do not send 200 INVITEs per second to the peer.  (Use of the SOC =
mechanism might prove to work better than our current round-robin =
approach, but we’ve not tested it.) =




 
Note =
that the "round-robin" approach may or may not be preferred =
and is not
required in our draft.  I've found =
this approach works pretty well, but one
may prefer =
to use a particular weighting that is more complex.  I =
don't
think local decisions in the box need to be =
subject to standardization: we
just need to =
standardize the signaling to enable such decisions.
I believe the signaling is not enough. The deep =
meaning of the feedback parameter needs to be clarified too. 
For =
example, in our proposal, the feedback is simply the current utilization =
rate of the node (very simple).
 
PEJ: That sounds similar to SCE.  An SCE-capable device =
advertises to its peer how many new sessions it can handle given the =
number of sessions it is currently handling. 

Also the =
expectation needs to be clarified because it has a very important impact =
on the recommended behavior in case of overload. 
The solution that =
we propose constantly tries to balance load in the cluster of servers. =
Therefore, in case the SOC shedding kicks in, there is NOT NEED to =
repeat the message elsewhere in the cluster because it is balanced =
already (So no other server has more capacity). This is a huge advantage =
for the network stability as message repeat tends to spread the overload =
and collapse the cluster. 



In any =
case, it certainly looks like we're thinking along the same =
lines.
Yes I agree. =




PEJ: If you haven’t already, have a look at our =
draft:
http://tools.ietf.org/html/draft-jones-sip-overlo=
ad-sce-00
 
It is now expired, but your draft suggests there is still life in =
this approach.  We should look to harmonize them, =
perhaps.
 
Paul
 

Cordially, =


Thierry



 
Paul
 
-----Original =
Message-----
From: dispatch-bounces@ietf.org =
[mailto:dispatch-bounces@ietf.or=
g] On
Behalf Of Vijay K. =
Gurbani
Sent: Monday, March 07, 2011 6:38 =
PM
To: dispatch@ietf.org
<=
pre>Subject: Re: [dispatch] SIP load =
balancing
 
Paul: Thank =
you for your input.  More =
inline.
 
On 03/07/2011 =
05:14 PM, Paul Kyzivat wrote:
Vijay,
IIUC, the algorithm you have proposed is similar to the one Paul =
Jones
has proposed - with the main difference being =
the numeric range of the
weight. Do I have that =
right?
 
I =
believe that Paul J.'s algorithm is more suited towards =
overload
control, not load balancing (and I will =
await Paul J.'s correction to =
my
assertion).
 
The intent in draft-bessis-dispatch-adaptive-load-balancing =
is to
*always* report the current load to give the =
upstream SIP proxy an
indication of the health of =
the SIP server in the cluster.  If one =
waits
until overload is imminent, then it may be =
too late since a sudden and
unanticipated burst of =
messages will put the SIP server in overload
rather =
quickly.
 
It isn't clear to me =
from your draft if you are specifying the
algorithm =
by which the relative weights are used to determine =
the
probability of a request going to one server or =
another.
 
Ye=
s, same way as RFC2782 does =
it.
 
ISTM that it *does* =
matter. And it especially matters if there is =
an
implication that the upstream node's algorithm =
may depend on some
assumption or knowledge about =
the downstream =
nodes.
 
No =
knowledge, just a reported number that the upstream SIP proxy =
can
treat as a RFC2782-like weight indicator.  =
Today, rfc3263-resolution
starts off with the SRV =
indicators in RFC2782, without any regard =
to
additional knowledge or assumption about =
downstream nodes.
dispatch-adaptive-load-balancing =
is trying to use the same model, except
render the =
weight to be sensitive to the current traffic in the =
cluster
in =
real-time.
 
And of course this =
method of reporting weights isn't sufficient =
to
drive a balancing algorithm that is sensitive to =
some requests
requiring more or different resources =
than others. (That seems to be a
real problem, =
though I don't know if we have a real solution =
for
it.)
 
Yes, that is a real problem, as is the problem of a =
SIP server in a
cluster that has a low weight and =
is thus not contacted with enough
frequency to =
merit more participation to the =
goodput.
 
It is well =
known in SIP that some methods are more onerous to =
process
than others --- from a proxying point of =
view, processing an INVITE with
forking and =
response aggregation is far more detrimental to =
resource
usage than is processing a BYE.  Some =
existing literature techniques
thus add a =
coefficient to the weight of sending an INVITE versus a =
BYE.
At this time, the process outlined in =
dispatch-adaptive-load-balancing
treats all =
requests the same, but we can definitely look at ways =
to
change this =
behaviour.
 
Thanks,<=
/o:p>
 
- =
vijay
--
Vijay K. Gurbani, =
Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane, =
Rm. 9C-533, Naperville, Illinois 60566 (USA)
Email: =
vkg@{bell-labs.com,acm.or=
g} / vijay.gurbani@alcatel-lu=
cent.com
Web:   http://ect.bell-labs.com/who/v=
kg/
____________________________________________=
___
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 =
;
_______________________________________________
dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis =

  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucen=
t.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: 12 April - 14 April =
2011
Upcoming planned Vacation: None =

 
 
_______________________________________________=

dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis =

  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucen=
t.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: 12 April - 14 April =
2011
Upcoming planned Vacation: None =

 
 
_______________________________________________=

dispatch mailing list
dispatch@ietf.org
<=
pre>https://www.ietf.=
org/mailman/listinfo/dispatch
 
-- =

-- 
Cordially, 
Thierry Bessis 
  IMS Solution: Architecture =
& Product Management - ALTA Member
  Organization: ALU > =
Carrier group > Wireline BU >  IMS Sol Arch, =
SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U S
  Tel/Fax: =
+1 630 979 7989
  Corporate IM: tbessis    - =
Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference information:
2801 2801 (US):+1 =
800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others =
countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming =
planned Business Trip: 12 April - 14 April 2011
Upcoming =
planned Vacation: None =


------=_NextPart_001_0609_01CBDF9C.A6D65E30--

------=_NextPart_000_0608_01CBDF9C.A6D65E30
Content-Type: image/jpeg;
	name="image001.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
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------=_NextPart_000_0608_01CBDF9C.A6D65E30--


From Thierry.Bessis@alcatel-lucent.com  Fri Mar 11 08:04:36 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D39983A69D6 for ; Fri, 11 Mar 2011 08:04:36 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RxOusBYRtn49 for ; Fri, 11 Mar 2011 08:04:34 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 9A7F13A6863 for ; Fri, 11 Mar 2011 08:04:34 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p2BG5pBU001251 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 11 Mar 2011 10:05:51 -0600 (CST)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2BG5pI9020018; Fri, 11 Mar 2011 10:05:51 -0600 (CST)
Message-ID: <4D7A485E.2080602@alcatel-lucent.com>
Date: Fri, 11 Mar 2011 10:05:50 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Parthasarathi R (partr)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090602050609070605040309"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 16:04:36 -0000

This is a multi-part message in MIME format.
--------------090602050609070605040309
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Paul,
>
> For instance, with our approach, it is fundamental that the weight 
> represents a "Relative" occupation, and not an absolute capacity.
>  The question is how to perform the "relative" operation 
> without until understanding the capacity of the peer devices. Say Load 
> balancer has 3 devices: 2 devices with 1x capacity and 1 device with 
> 2x capacity. Here all devices will be able to provide the "relative" 
> value without understanding the capacity of the other device. 
>
> The first reason is that we found that loading continuously all the 
> servers at the same utilization rate (same percentage) is the optimal 
> way of loading a cluster. (We can discuss that if not agreed).
>  I agree that same utilization rate is the optimal way but the 
> question is how to achieve it.
Thierry>> we are on line on this one ! :-)
> Round robin algorithm works in case all devices have same capacity and 
> same amount of load per call.
Thierry >> Yes I agree it would work in this case. The reason we are not 
using it is that in the IMS network (where we are already using those 
principles), this is not the case: Devices do not have the same capacity 
because they are on different hardware, or (more often) because they are 
sharing resources with other devices that are doing different things at 
different moments. (which provide a huge resource saving !). The result 
is that it's impossible for any device to calculate its own capacity "in 
isolation": It would need to know exactly which other devices are 
sharing the same resources, their resource usage, and what they are 
doing at any time: an operational nightmare which would break all the 
software isolation rules ! :-) . In addition, we sometime shutdown some 
devices and re-enable them later, (or we grow more devices in the 
cluster): Round robin does not work at all in this case.

> In reality, there are multiple varying factor like call duration, 
> different vareity of platform within the cluster (say Generation-1 10 
> devices and Generation-2 5 devies), different resource usage for the 
> call, rate at which calls generation. 
Thierry>>  yes !
> The second reason is that a relative occupation is very easy to derive 
> from information already available: any system knows where it is in 
> terms of CPU or Memory usage.
>  Converting CPU or Memory to weight is not explained your 
> draft. It is the real algorithm which needs lot of discussion. Because 
> preparing the mapping between CPU & Memory to weight will be 
> mind-blogging activity 
Thierry>> We could be more explicit, but should we ? (see next 
paragraph). Memory and CPU are just the two most obvious examples or 
resources that you need to process sessions. For instance, some of our 
servers have a fixed amount of sessions registers (for instance 50 000). 
In this case, the weight also reflects that. The algorithm we use for 
weights is very simple: this is the smallest remaining (in percentage) 
of each resource that is (1) Required to process a call, (2) Scarce  
(not infinite). Let's consider the following system:
The server has 70% of its memory available for calls, an average of 30% 
CPU usage (average over the last 10s for instance), and 40% of its 
registers available (remember that those resources may be shared by 
other components or devices for any other usage: it does not matter).
Then the weight at this moment is MIN (70, 30, 40) = 30%. You then need 
to normalize this weight to match the weight range (an INT ? )

Now we are not saying that everybody should calculate their weight like 
that. What is important is that the weight reflects your current 
_relative_ capacity. In our case, this "algorithm" balances the load of 
our servers very well. Now if you use a server that has a fixed amount 
of resources and that is dimensioned to process let's say 50 000 
sessions, at a pace of 100 session/s, then you could probably derive a 
weight that would depend on both your count of sessions left, and your 
current rate of session/s.
>  In contrast, deriving a left capacity in terms of remaining sessions 
> can be difficult: Host can be used by different applications and 
> perform different duties, which can change over the day, all of them 
> impacting their true capacity at a certain moment. Beside, as already 
> discussed, different sessions require different resources, so there is 
> a whole complication of dealing with a multitude of call profiles, 
> that (again) are not even constant over the day.
>
> Also, we do not want to derive some shedding decision from the weight.
>   In case you specify, IETF SoC does not allow and so we want 
> another value, I think that we are in same page.
Thierry>> we are on the same page. I thought you suggested to use the 
same value.
> If not, I could not understand your argument. As you mentioned, all 
> system will know CPU or memory usage and it is possible to come up the 
> static or refreshable value saying that load shed has to happen when 
> 85% of CPU utilization or 90% of memory utilization. The different 
> between our proposal is that there is no need of two value per 
> message instead 1 value in each message & 1 value is static or in 
> periodic manner 
Thierry>> We use a single value: see the example above.

--
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------090602050609070605040309
Content-Type: multipart/related;
 boundary="------------090207090601030609020606"


--------------090207090601030609020606
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
    
  
  
    Paul, 

    

      


        For instance, with our approach, it is fundamental that the
        weight represents a "Relative" occupation, and not an absolute
        capacity. 

      
<Partha> The question is how to perform the "relative"
        operation without until understanding the capacity of the peer
        devices. Say Load balancer has 3 devices: 2 devices with 1x
        capacity and 1 device with 2x capacity. Here all devices will be
        able to provide the "relative" value without understanding the
        capacity of the other device. </Partha>

      


        The first reason is that we found that loading continuously all
        the servers at the same utilization rate (same percentage) is
        the optimal way of loading a cluster. (We can discuss that if
        not agreed). 

      
<Partha> I agree that same utilization rate is the
        optimal way but the question is how to achieve it.  

      

    

    Thierry>> we are on line on this one !  :-) 

    

      
Round robin algorithm works in case all devices have same
        capacity and same amount of load per call.

    

    Thierry >> Yes I agree it would work in this case. The reason
    we are not using it is that in the IMS network (where we are already
    using those principles), this is not the case: Devices do not have
    the same capacity because they are on different hardware, or (more
    often) because they are sharing resources with other devices that
    are doing different things at different moments. (which provide a
    huge resource saving !). The result is that it's impossible for any
    device to calculate its own capacity "in isolation": It would need
    to know exactly which other devices are sharing the same resources,
    their resource usage, and what they are doing at any time: an
    operational nightmare which would break all the software isolation
    rules !  :-) . In
    addition, we sometime shutdown some devices and re-enable them
    later, (or we grow more devices in the cluster): Round robin does
    not work at all in this case. 

    

    

      
 In reality, there are multiple varying factor like call
        duration, different vareity of platform within the cluster (say
        Generation-1 10 devices and Generation-2 5 devies), different
        resource usage for the call, rate at which calls generation.
        </Partha>

    

    Thierry>>  yes ! 

    

      
 

      
The second reason is that a relative occupation is very easy
        to derive from information already available: any system knows
        where it is in terms of CPU or Memory usage.

      
 

      
<Partha> Converting CPU or Memory to weight is not
        explained your draft. It is the real algorithm which needs lot
        of discussion. Because preparing the mapping between CPU &
        Memory to weight will be mind-blogging activity </Partha>

    

    Thierry>> We could be more explicit, but should we ? (see next
    paragraph). Memory and CPU are just the two most obvious examples or
    resources that you need to process sessions. For instance, some of
    our servers have a fixed amount of sessions registers (for instance
    50 000). In this case, the weight also reflects that. The algorithm
    we use for weights is very simple: this is the smallest remaining
    (in percentage) of each resource that is (1) Required to process a
    call, (2) Scarce  (not infinite). Let's consider the following
    system: 

    The server has 70% of its memory available for calls, an average of
    30% CPU usage (average over the last 10s for instance), and 40% of
    its registers available (remember that those resources may be shared
    by other components or devices for any other usage: it does not
    matter). 

    Then the weight at this moment is MIN (70, 30, 40) = 30%. You then
    need to normalize this weight to match the weight range (an INT ? )
    

    

    Now we are not saying that everybody should calculate their weight
    like that. What is important is that the weight reflects your
    current relative capacity. In our case, this "algorithm"
    balances the load of our servers very well. Now if you use a server
    that has a fixed amount of resources and that is dimensioned to
    process let's say 50 000 sessions, at a pace of 100 session/s, then
    you could probably derive a weight that would depend on both your
    count of sessions left, and your current rate of session/s. 

    

      
 

      
 In contrast, deriving a left capacity in terms of remaining
        sessions can be difficult: Host can be used by different
        applications and perform different duties, which can change over
        the day, all of them impacting their true capacity at a certain
        moment. Beside, as already discussed, different sessions require
        different resources, so there is a whole complication of dealing
        with a multitude of call profiles, that (again) are not even
        constant over the day.

        

        Also, we do not want to derive some shedding decision from the
        weight.

      
 

      
<Partha>  In case you specify, IETF SoC does not allow
        and so we want another value, I think that we are in same page.

    

    Thierry>> we are on the same page. I thought you suggested to
    use the same value. 

    

      
 If not, I could not understand your argument. As you
        mentioned, all system will know CPU or memory usage and it is
        possible to come up the static or refreshable value saying that
        load shed has to happen when 85% of CPU utilization or 90% of
        memory utilization. The different between our proposal is that
        there is no need of two value per message instead 1 value in
        each message & 1 value is static or in periodic
        manner  </Partha>

    

    Thierry>> We use a single value: see the example above. 

    

    
--

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------090207090601030609020606
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------090207090601030609020606--

--------------090602050609070605040309--

From Thierry.Bessis@alcatel-lucent.com  Fri Mar 11 08:40:13 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 68CC53A69DF for ; Fri, 11 Mar 2011 08:40:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gukz68tSEj9I for ; Fri, 11 Mar 2011 08:39:56 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 4E10B3A6A38 for ; Fri, 11 Mar 2011 08:39:56 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p2BGfFMK014619 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 11 Mar 2011 10:41:15 -0600 (CST)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2BGfEJI024046; Fri, 11 Mar 2011 10:41:14 -0600 (CST)
Message-ID: <4D7A50AA.3060805@alcatel-lucent.com>
Date: Fri, 11 Mar 2011 10:41:14 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Paul E. Jones" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com> <060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>
In-Reply-To: <060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>
Content-Type: multipart/alternative; boundary="------------050204040908060906060900"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 16:40:13 -0000

This is a multi-part message in MIME format.
--------------050204040908060906060900
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Paul,
>
> The specific ways through which the SCE value is computed may vary by 
> vendor, though the approach we have used is one where we calculate an 
> estimate of the number of sessions a device can handle given the 
> current number of sessions given current memory utilization.  We then 
> do the same for CPU and other system resources.  We then take the 
> minimum value of each and report that as the SCE value.  As a matter 
> of course, our devices are constantly measuring these resources, so 
> computing the SCE if trivial.
>
So this mode of operation is very similar to what we also propose. Is 
that number a ratio (percentage of sessions available over the total 
number), or an absolute number ? If this is a ratio, then SCE and our 
load balancing are compatible: the same information and transport 
mechanism could be used for both: A single IETF proposal could be done.

Cordially,

Thierry
>
> Paul
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Thursday, March 10, 2011 1:17 PM
> *To:* Paul E. Jones
> *Cc:* dispatch@ietf.org
> *Subject:* Re: [dispatch] SIP load balancing
>
> Paul,
>
> I agree they are many similitudes ! :-)
> But the question I think is: Can we make the two approaches compatible ?
>
> For instance, with our approach, it is fundamental that the weight 
> represents a "Relative" occupation, and not an absolute capacity.
> The first reason is that we found that loading continuously all the 
> servers at the same utilization rate (same percentage) is the optimal 
> way of loading a cluster. (We can discuss that if not agreed). The 
> second reason is that a relative occupation is very easy to derive 
> from information already available: any system knows where it is in 
> terms of CPU or Memory usage. In contrast, deriving a left capacity in 
> terms of remaining sessions can be difficult: Host can be used by 
> different applications and perform different duties, which can change 
> over the day, all of them impacting their true capacity at a certain 
> moment. Beside, as already discussed, different sessions require 
> different resources, so there is a whole complication of dealing with 
> a multitude of call profiles, that (again) are not even constant over 
> the day.
>
> Also, we do not want to derive some shedding decision from the weight. 
> We want to rely on the OLC mechanism for that. As I mentioned before, 
> deriving shedding from the weight is very sensitive to the way weights 
> are calculated: we would need to specify exactly how the weight is 
> calculated: this is impractical ! Also everybody would need to agree 
> about specific threshold: What is the weight value threshold for 
> shedding ?: Again this is more difficult, and different servers will 
> certainly want to use different thresholds.   Also there is a 
> contradiction: whether you are using relative or absolute weights, 
> your weight should  zero (100% busy, or 0 sessions available) at the 
> moment you are getting overloaded. So if you are 100% busy (or in 
> average no room for new sessions), how to you tell how many sessions 
> need to be shed ? You could be 100% busy with an offered load of 300% 
> nominal (then you need to shed 66%), or be 100% busy at just the 
> nominal load, (then you don't need to shed). This again shows this 
> threshold issue I mentioned. In our implementation, we used to use to 
> a common "utilization" information to do both optimal distribution and 
> shedding, but we moved away from that because of those reasons: we 
> found that each node indicating its requested rejection rate is much 
> more robust and requires very little agreements.
>
> Cordially,
>
> Thierry
>
>
> On 09-Mar-2011 22:15, Paul E. Jones wrote:
>
> Thierry,
>
> I think there is more in common than not.
>
> You compute a "weight" and we compute an "sce" that is, in my opinion, 
> more or less the same: it's a value that indicates how much traffic 
> can be directed toward the box relative to other peer devices.
>
> You want to do "load balancing" whereas we want to do "overload 
> prevention".  Preventing overload means distributing load across 
> boxes.  We are not proposing to perfectly balance the load, but the 
> load is distributed nonetheless.
>
> Honestly, I just don't see a lot of difference, other than 
> terminology.  The algorithm employed by the device directing traffic, 
> but that algorithm is independent of the advertisement of a 
> capacity/weight value.
>
> Paul
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Wednesday, March 09, 2011 12:26 PM
> *To:* Paul E. Jones
> *Cc:* dispatch@ietf.org 
> *Subject:* Re: [dispatch] SIP load balancing
>
> Thanks Paul for you detailed answer !
>
> I've looked at your SCE proposal and your comments and my impression 
> is that our two proposals are actually very different both from their 
> intent perspective, and from their realization perspective.
>
> The intent of our load balancing proposal is clearly continuous load 
> balancing. Our suggested weight does try to figure any kind of 
> remaining capacity in terms of sessions: it only indicates the instant 
> (slightly smoothed) available resource, which integrates CPU, Memory 
> and any other constrained resource. For the kind of usage we have in 
> mind, (and for example), it does not matter if we have lot of sessions 
> "left" when we have very little CPU to register those sessions: it is 
> optimal to temporary offload this busy node rather than risking a CPU 
> induced rejection or failure. When the node will have more CPU 
> available, it will take more sessions, and may be even more than 
> others, because it may have now more memory than others. If this node 
> never have more CPU available, then its "session" capacity is actually 
> unusable and the system correctly reflects that.
>
> Also, the SOC was not intended for "dynamic" load distribution: it 
> only provides the feedback when some shedding is expected (So already 
> one or multiple nodes are about to be overloaded). Also, integrating 
> the information from both SOC and SCE to take the distribution 
> decision seems rather complicated to me. In our proposal, the 
> distribution algorithm is simple, never changes, and is independent 
> from SOC. We selected a node with weights always the same way, and we 
> apply the SOC rejection when applicable to this node.
>
> Cordially,
>
> Thierry
>
> On 08-Mar-2011 14:47, Paul E. Jones wrote:
>
> Thierry,
>
> Please see my comments below in green:
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Tuesday, March 08, 2011 12:35 PM
> *To:* Paul E. Jones
> *Cc:* 'Vijay K. Gurbani'; dispatch@ietf.org 
> *Subject:* Re: [dispatch] SIP load balancing
>
> Paul,
>
> Please let me address some of your comments.
> I don't know what the SCE is, so please pardon me if I misunderstand 
> something.
>
> On 07-Mar-2011 23:05, Paul E. Jones wrote:
>
> Vijay,
>   
> The Session Capacity Estimate (SCE) idea is actually quite similar to the
> "adaptive load balancing" draft, from what I can see.  SCE was proposed as a
> solution to the "overload" problem, the approach being one where we try to
> avoid overload by selecting a device that reports a higher capacity value.
> Thus, one could view it as "overload control through load balancing".  We
> didn't explicitly call it load balancing, though, because we were not trying
> to balance load (i.e., place the same load on all servers).  Rather, we just
> wanted to direct traffic to devices that we knew had available capacity.
>
> Agreed, the goal seems to be the same: avoiding or at least delaying 
> overload.
> And your other comment too: our proposal does try to balance the load 
> on all the servers. The rational is that it is very easy for each 
> server to know where it is in terms of resource usage (for instance: 
> I'm 60 percent resource loaded at this moment), and this is just the 
> information you need to balance the traffic. Now it seems (but it is 
> not clear, see below)  that the SCE "available capacity" is actually a 
> real time feedback, so the information would be very similar actually. 
> (What's the difference ? )
>
> PEJ: I think it is very similar.  SCE provides real-time feedback of 
> the available resource capacity of a peer device.  The intent is to 
> convey a single integer value that represents an estimate of the 
> number of sessions the peer device can handle.  Note, this is not 
> "message per second" or other, but the number of sessions.  The focus 
> on SCE is on devices like "call agents", "session border controllers", 
> stateful SIP proxies, etc.
>
>
>
>   
> In practice, the way SCE would be employed is the following:
> * SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from
>     peer devices
>   * The selector would direct traffic toward the peer devices in a
>     round-robin fashion, taking into consideration the SCE value so
>     as to prevent exhausting resources on any given peer device
>
> I see a contradiction here: if it is round-robin, it does not take 
> into account the SCE value.
> Is that something more like a weighted round-robin algorithm ?
>
> PEJ: As I mention below, the algorithm used to select a peer device is 
> not a part of the SCE draft.  One can use any mechanism.  The SCE 
> draft merely advertises a session capacity estimate.  We've found 
> round-robin to be a simple approach that works, as it distributes load 
> relatively uniformly, but stops directing traffic toward devices that 
> report an SCE value that is below a specified threshold. Other input 
> could be considered, but we deliberately left the algorithm outside 
> the scope of SCE.
>
>
>
>   * Once the peer's SCE value dropped below a given "low-water mark",
>     the peer device would be less favored and not given traffic until
>     the SCE value increases above the low-water mark (OPTIONS "ping"
>     messages may be used to determine the SCE value in the event that
>     there is no other traffic)
>
> Why would we wait for a lower mark before to influence the 
> distribution ? Also is that less favored, or given no traffic ?
>
> PEJ: With our use of SCE, we're not trying to "balance" the load, per 
> se.  One certainly could do that, but in order to balance the load on 
> peer devices when they are out of balance, one must direct multiple 
> new sessions toward the same device.  That is not always desirable 
> since new sessions usually consume more resources temporarily than 
> established sessions.  So, we try to burden a peer device by throwing 
> a number of new sessions toward the peer by distributing them in a 
> round-robin fashion.  Again, though, the approach taken is outside the 
> scope of the SCE draft.
>
>
> The later has stability issues.  And why do we need ping Options when 
> there is no traffic ?
>  If there is no traffic (or very little), should we care about optimal 
> distribution from this source ?
>
> PEJ: The assumption is that a peer device might also have several 
> peers.  So, while device A1 may not be directing traffic toward device 
> B, perhaps A2, A3, and A4 are.  We can use OPTIONS "ping" messages to 
> get current SCE information.
>
>
>
>
>
>   * The current SOC overload control mechanism would be used to
>     to prevent message overloading, as SCE only focuses on resource
>     overloading (e.g., memory, CPU, disk space, DSPs)
>
> I don't see this difference here (message overload versus resource 
> overload ? ) . To me, SOC focuses on source shedding, while SCE seems 
> to focus on improved load distribution, just like our proposal.
>
> PEJ: A peer device might be able to handle 18,000 new sessions, but 
> perhaps it can only handle them at a rate of 100 per second.  We would 
> use SCE to determine that a peer can handle new sessions, while we 
> would use the methods SOC is defining to prevent message overload so 
> we do not send 200 INVITEs per second to the peer.  (Use of the SOC 
> mechanism might prove to work better than our current round-robin 
> approach, but we've not tested it.)
>
>
>
>   
> Note that the "round-robin" approach may or may not be preferred and is not
> required in our draft.  I've found this approach works pretty well, but one
> may prefer to use a particular weighting that is more complex.  I don't
> think local decisions in the box need to be subject to standardization: we
> just need to standardize the signaling to enable such decisions.
>
> I believe the signaling is not enough. The deep meaning of the 
> feedback parameter needs to be clarified too.
> For example, in our proposal, the feedback is simply the current 
> utilization rate of the node (very simple).
>
> PEJ: That sounds similar to SCE.  An SCE-capable device advertises to 
> its peer how many new sessions it can handle given the number of 
> sessions it is currently handling.
>
> Also the expectation needs to be clarified because it has a very 
> important impact on the recommended behavior in case of overload.
> The solution that we propose constantly tries to balance load in the 
> cluster of servers. Therefore, in case the SOC shedding kicks in, 
> there is NOT NEED to repeat the message elsewhere in the cluster 
> because it is balanced already (So no other server has more capacity). 
> This is a huge advantage for the network stability as message repeat 
> tends to spread the overload and collapse the cluster.
>
>
>
> In any case, it certainly looks like we're thinking along the same lines.
>
> Yes I agree.
>
>
>
> PEJ: If you haven't already, have a look at our draft:
>
> http://tools.ietf.org/html/draft-jones-sip-overload-sce-00
>
> It is now expired, but your draft suggests there is still life in this 
> approach.  We should look to harmonize them, perhaps.
>
> Paul
>
>
> Cordially,
>
> Thierry
>
>
>
>   
> Paul
>   
>
>     -----Original Message-----
>
>     From:dispatch-bounces@ietf.org    [mailto:dispatch-bounces@ietf.org] On
>
>     Behalf Of Vijay K. Gurbani
>
>     Sent: Monday, March 07, 2011 6:38 PM
>
>     To:dispatch@ietf.org  
>
>     Subject: Re: [dispatch] SIP load balancing
>
>       
>
>     Paul: Thank you for your input.  More inline.
>
>       
>
>     On 03/07/2011 05:14 PM, Paul Kyzivat wrote:
>
>         Vijay,
>
>         IIUC, the algorithm you have proposed is similar to the one Paul Jones
>
>         has proposed - with the main difference being the numeric range of the
>
>         weight. Do I have that right?
>
>       
>
>     I believe that Paul J.'s algorithm is more suited towards overload
>
>     control, not load balancing (and I will await Paul J.'s correction to my
>
>     assertion).
>
>       
>
>     The intent in draft-bessis-dispatch-adaptive-load-balancing is to
>
>     *always* report the current load to give the upstream SIP proxy an
>
>     indication of the health of the SIP server in the cluster.  If one waits
>
>     until overload is imminent, then it may be too late since a sudden and
>
>     unanticipated burst of messages will put the SIP server in overload
>
>     rather quickly.
>
>       
>
>         It isn't clear to me from your draft if you are specifying the
>
>         algorithm by which the relative weights are used to determine the
>
>         probability of a request going to one server or another.
>
>       
>
>     Yes, same way as RFC2782 does it.
>
>       
>
>         ISTM that it *does* matter. And it especially matters if there is an
>
>         implication that the upstream node's algorithm may depend on some
>
>         assumption or knowledge about the downstream nodes.
>
>       
>
>     No knowledge, just a reported number that the upstream SIP proxy can
>
>     treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution
>
>     starts off with the SRV indicators in RFC2782, without any regard to
>
>     additional knowledge or assumption about downstream nodes.
>
>     dispatch-adaptive-load-balancing is trying to use the same model, except
>
>     render the weight to be sensitive to the current traffic in the cluster
>
>     in real-time.
>
>       
>
>         And of course this method of reporting weights isn't sufficient to
>
>         drive a balancing algorithm that is sensitive to some requests
>
>         requiring more or different resources than others. (That seems to be a
>
>         real problem, though I don't know if we have a real solution for
>
>         it.)
>
>       
>
>     Yes, that is a real problem, as is the problem of a SIP server in a
>
>     cluster that has a low weight and is thus not contacted with enough
>
>     frequency to merit more participation to the goodput.
>
>       
>
>     It is well known in SIP that some methods are more onerous to process
>
>     than others --- from a proxying point of view, processing an INVITE with
>
>     forking and response aggregation is far more detrimental to resource
>
>     usage than is processing a BYE.  Some existing literature techniques
>
>     thus add a coefficient to the weight of sending an INVITE versus a BYE.
>
>     At this time, the process outlined in dispatch-adaptive-load-balancing
>
>     treats all requests the same, but we can definitely look at ways to
>
>     change this behaviour.
>
>       
>
>     Thanks,
>
>       
>
>     - vijay
>
>     --
>
>     Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
>
>     1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)
>
>     Email:vkg@{bell-labs.com,acm.org}    /vijay.gurbani@alcatel-lucent.com  
>
>     Web:http://ect.bell-labs.com/who/vkg/
>
>     _______________________________________________
>
>     dispatch mailing list
>
>     dispatch@ietf.org  
>
>     https://www.ietf.org/mailman/listinfo/dispatch
>
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>   
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>   
>   
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org  
> https://www.ietf.org/mailman/listinfo/dispatch
>
> -- 
> *--
> Cordially,
> Thierry Bessis*
>
> IMS Solution: Architecture & Product Management - ALTA Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/ 
> 
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr 
> 
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> *Upcoming planned Business Trip: 12 April - 14 April 2011*
> *Upcoming planned Vacation: None*
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------050204040908060906060900
Content-Type: multipart/related;
 boundary="------------070001000806000304020300"


--------------070001000806000304020300
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Paul, 

    
    

      

        
The specific ways through which the SCE value is
            computed may vary by vendor, though the approach we have
            used is one where we calculate an estimate of the number of
            sessions a device can handle given the current number of
            sessions given current memory utilization.  We then do the
            same for CPU and other system resources.  We then take the
            minimum value of each and report that as the SCE value.  As
            a matter of course, our devices are constantly measuring
            these resources, so computing the SCE if trivial.

      

    

    So this mode of operation is very similar to what we also propose.
    Is that number a ratio (percentage of sessions available over the
    total number), or an absolute number ? If this is a ratio, then SCE
    and our load balancing are compatible: the same information and
    transport mechanism could be used for both: A single IETF proposal
    could be done. 

    

    Cordially, 

    

    Thierry

    

      

        

        
 

        
Paul

        
 

        

          

            

              
From: Thierry Bessis
                  [mailto:Thierry.Bessis@alcatel-lucent.com] 

                  Sent: Thursday, March 10, 2011 1:17 PM

                  To: Paul E. Jones

                  Cc: dispatch@ietf.org

                  Subject: Re: [dispatch] SIP load balancing

            

          

          
 

          
Paul, 

            

            I agree they are many similitudes ! :-) 

            But the question I think is: Can we make the two approaches
            compatible ? 

            

            For instance, with our approach, it is fundamental that the
            weight represents a "Relative" occupation, and not an
            absolute capacity.

            The first reason is that we found that loading continuously
            all the servers at the same utilization rate (same
            percentage) is the optimal way of loading a cluster. (We can
            discuss that if not agreed). The second reason is that a
            relative occupation is very easy to derive from information
            already available: any system knows where it is in terms of
            CPU or Memory usage. In contrast, deriving a left capacity
            in terms of remaining sessions can be difficult: Host can be
            used by different applications and perform different duties,
            which can change over the day, all of them impacting their
            true capacity at a certain moment. Beside, as already
            discussed, different sessions require different resources,
            so there is a whole complication of dealing with a multitude
            of call profiles, that (again) are not even constant over
            the day.

            

            Also, we do not want to derive some shedding decision from
            the weight. We want to rely on the OLC mechanism for that.
            As I mentioned before, deriving shedding from the weight is
            very sensitive to the way weights are calculated: we would
            need to specify exactly how the weight is calculated: this
            is impractical ! Also everybody would need to agree about
            specific threshold: What is the weight value threshold for
            shedding ?: Again this is more difficult, and different
            servers will certainly want to use different thresholds.  
            Also there is a contradiction: whether you are using
            relative or absolute weights, your weight should  zero (100%
            busy, or 0 sessions available) at the moment you are getting
            overloaded. So if you are 100% busy (or in average no room
            for new sessions), how to you tell how many sessions need to
            be shed ? You could be 100% busy with an offered load of
            300% nominal (then you need to shed 66%), or be 100% busy at
            just the nominal load, (then you don't need to shed). This
            again shows this threshold issue I mentioned. In our
            implementation, we used to use to a common "utilization"
            information to do both optimal distribution and shedding,
            but we moved away from that because of those reasons: we
            found that each node indicating its requested rejection rate
            is much more robust and requires very little agreements. 

            

            Cordially, 

            

            Thierry

            

            

            On 09-Mar-2011 22:15, Paul E. Jones wrote: 

          
Thierry,

          
 

          
I
              think there is more in common than not.

          
 

          
You
              compute a “weight” and we compute an “sce” that is, in my
              opinion, more or less the same: it’s a value that
              indicates how much traffic can be directed toward the box
              relative to other peer devices.

          
 

          
You
              want to do “load balancing” whereas we want to do
              “overload prevention”.  Preventing overload means
              distributing load across boxes.  We are not proposing to
              perfectly balance the load, but the load is distributed
              nonetheless.

          
 

          
Honestly,
              I just don’t see a lot of difference, other than
              terminology.  The algorithm employed by the device
              directing traffic, but that algorithm is independent of
              the advertisement of a capacity/weight value.

          
 

          
Paul

          
 

          

            

              

                
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
                    

                    Sent: Wednesday, March 09, 2011 12:26 PM

                    To: Paul E. Jones

                    Cc: dispatch@ietf.org

                    Subject: Re: [dispatch] SIP load balancing

              

            

            
 

            
Thanks Paul for you detailed answer ! 

              

              I've looked at your SCE proposal and your comments and my
              impression is that our two proposals are actually very
              different both from their intent perspective, and from
              their realization perspective. 

              

              The intent of our load balancing proposal is clearly
              continuous load balancing. Our suggested weight does try
              to figure any kind of remaining capacity in terms of
              sessions: it only indicates the instant (slightly
              smoothed) available resource, which integrates CPU, Memory
              and any other constrained resource. For the kind of usage
              we have in mind, (and for example), it does not matter if
              we have lot of sessions "left" when we have very little
              CPU to register those sessions: it is optimal to temporary
              offload this busy node rather than risking a CPU induced
              rejection or failure. When the node will have more CPU
              available, it will take more sessions, and may be even
              more than others, because it may have now more memory than
              others. If this node never have more CPU available, then
              its "session" capacity is actually unusable and the system
              correctly reflects that. 

              

              Also, the SOC was not intended for "dynamic" load
              distribution: it only provides the feedback when some
              shedding is expected (So already one or multiple nodes are
              about to be overloaded). Also, integrating the information
              from both SOC and SCE to take the distribution decision
              seems rather complicated to me. In our proposal, the
              distribution algorithm is simple, never changes, and is
              independent from SOC. We selected a node with weights
              always the same way, and we apply the SOC rejection when
              applicable to this node.

              

              Cordially, 

              

              Thierry

              

              On 08-Mar-2011 14:47, Paul E. Jones wrote: 

            
Thierry,

            
 

            
Please see
                my comments below in green:

            
 

            
 

            

              

                

                  
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
                      

                      Sent: Tuesday, March 08, 2011 12:35 PM

                      To: Paul E. Jones

                      Cc: 'Vijay K. Gurbani'; dispatch@ietf.org

                      Subject: Re: [dispatch] SIP load balancing

                

              

              
 

              
Paul, 

                

                Please let me address some of your comments. 

                I don't know what the SCE is, so please pardon me if I
                misunderstand something. 

                

                On 07-Mar-2011 23:05, Paul E. Jones wrote: 

              
Vijay,

              
 

              
The Session Capacity Estimate (SCE) idea is actually quite similar to the

              
"adaptive load balancing" draft, from what I can see.  SCE was proposed as a

              
solution to the "overload" problem, the approach being one where we try to

              
avoid overload by selecting a device that reports a higher capacity value.

              
Thus, one could view it as "overload control through load balancing".  We

              
didn't explicitly call it load balancing, though, because we were not trying

              
to balance load (i.e., place the same load on all servers).  Rather, we just

              
wanted to direct traffic to devices that we knew had available capacity.

              
Agreed, the goal seems to be the
                same: avoiding or at least delaying overload. 

                And your other comment too: our proposal does try to
                balance the load on all the servers. The rational is
                that it is very easy for each server to know where it is
                in terms of resource usage (for instance: I'm 60 percent
                resource loaded at this moment), and this is just the
                information you need to balance the traffic. Now it
                seems (but it is not clear, see below)  that the SCE
                "available capacity" is actually a real time feedback,
                so the information would be very similar actually.
                (What's the difference ? )

              
 

              
PEJ: I think it is very similar. 
                  SCE provides real-time feedback of the available
                  resource capacity of a peer device.  The intent is to
                  convey a single integer value that represents an
                  estimate of the number of sessions the peer device can
                  handle.  Note, this is not “message per second” or
                  other, but the number of sessions.  The focus on SCE
                  is on devices like “call agents”, “session border
                  controllers”, stateful SIP proxies, etc. 

                

                

                

                

              
 

              
In practice, the way SCE would be employed is the following:

              
* SIP proxy, B2BUA, etc. (selector) would receive SCE capacity from

              
   peer devices

              
 * The selector would direct traffic toward the peer devices in a

              
   round-robin fashion, taking into consideration the SCE value so

              
   as to prevent exhausting resources on any given peer device

              
I see a contradiction here: if it is
                round-robin, it does not take into account the SCE
                value. 

                Is that something more like a weighted round-robin
                algorithm ? 

              
 

              
PEJ: As I mention below, the
                  algorithm used to select a peer device is not a part
                  of the SCE draft.  One can use any mechanism.  The SCE
                  draft merely advertises a session capacity estimate. 
                  We’ve found round-robin to be a simple approach that
                  works, as it distributes load relatively uniformly,
                  but stops directing traffic toward devices that report
                  an SCE value that is below a specified threshold.
                  Other input could be considered, but we deliberately
                  left the algorithm outside the scope of SCE.

                

                

                

                

              
 * Once the peer's SCE value dropped below a given "low-water mark",

              
   the peer device would be less favored and not given traffic until

              
   the SCE value increases above the low-water mark (OPTIONS "ping"

              
   messages may be used to determine the SCE value in the event that

              
   there is no other traffic)

              
Why would we wait for a lower mark
                before to influence the distribution ? Also is that less
                favored, or given no traffic ? 

              
 

              
PEJ: With our use of SCE, we’re not
                  trying to “balance” the load, per se.  One certainly
                  could do that, but in order to balance the load on
                  peer devices when they are out of balance, one must
                  direct multiple new sessions toward the same device. 
                  That is not always desirable since new sessions
                  usually consume more resources temporarily than
                  established sessions.  So, we try to burden a peer
                  device by throwing a number of new sessions toward the
                  peer by distributing them in a round-robin fashion. 
                  Again, though, the approach taken is outside the scope
                  of the SCE draft.

              
 

              


                The later has stability issues.  And why do we need ping
                Options when there is no traffic ?

                 If there is no traffic (or very little), should we care
                about optimal distribution from this source ?

              
 

              
PEJ: The assumption is that a peer
                  device might also have several peers.  So, while
                  device A1 may not be directing traffic toward device
                  B, perhaps A2, A3, and A4 are.  We can use OPTIONS
                  “ping” messages to get current SCE information.

              


                

                

                

                

              
 * The current SOC overload control mechanism would be used to

              
   to prevent message overloading, as SCE only focuses on resource

              
   overloading (e.g., memory, CPU, disk space, DSPs)

              
I don't see this difference here
                (message overload versus resource overload ? ) . To me,
                SOC focuses on source shedding, while SCE seems to focus
                on improved load distribution, just like our proposal.

              
 

              
PEJ: A peer device might be able to
                  handle 18,000 new sessions, but perhaps it can only
                  handle them at a rate of 100 per second.  We would use
                  SCE to determine that a peer can handle new sessions,
                  while we would use the methods SOC is defining to
                  prevent message overload so we do not send 200 INVITEs
                  per second to the peer.  (Use of the SOC mechanism
                  might prove to work better than our current
                  round-robin approach, but we’ve not tested it.)
                

                

                

                

                

              
 

              
Note that the "round-robin" approach may or may not be preferred and is not

              
required in our draft.  I've found this approach works pretty well, but one

              
may prefer to use a particular weighting that is more complex.  I don't

              
think local decisions in the box need to be subject to standardization: we

              
just need to standardize the signaling to enable such decisions.

              
I believe the signaling is not
                enough. The deep meaning of the feedback parameter needs
                to be clarified too. 

                For example, in our proposal, the feedback is simply the
                current utilization rate of the node (very simple).

              
 

              
PEJ: That sounds similar to SCE.  An
                  SCE-capable device advertises to its peer how many new
                  sessions it can handle given the number of sessions it
                  is currently handling. 

                

                Also the expectation needs to be clarified because it
                has a very important impact on the recommended behavior
                in case of overload. 

                The solution that we propose constantly tries to balance
                load in the cluster of servers. Therefore, in case the
                SOC shedding kicks in, there is NOT NEED to repeat the
                message elsewhere in the cluster because it is balanced
                already (So no other server has more capacity). This is
                a huge advantage for the network stability as message
                repeat tends to spread the overload and collapse the
                cluster. 

                

                

                

                

              
In any case, it certainly looks like we're thinking along the same lines.

              
Yes I agree. 

                

                

                

                

              
PEJ: If you haven’t already, have a
                  look at our draft:

              
http://tools.ietf.org/html/draft-jones-sip-overload-sce-00

              
 

              
It is now expired, but your draft
                  suggests there is still life in this approach.  We
                  should look to harmonize them, perhaps.

              
 

              
Paul

              
 

              


                Cordially, 

                

                Thierry

                

                

                

                

              
 

              
Paul

              
 

              

                
-----Original Message-----

                
From: dispatch-bounces@ietf.org [mailto:dispatch-bounces@ietf.org] On

                
Behalf Of Vijay K. Gurbani

                
Sent: Monday, March 07, 2011 6:38 PM

                
To: dispatch@ietf.org

                
Subject: Re: [dispatch] SIP load balancing

                
 

                
Paul: Thank you for your input.  More inline.

                
 

                
On 03/07/2011 05:14 PM, Paul Kyzivat wrote:

                

                  
Vijay,

                  
IIUC, the algorithm you have proposed is similar to the one Paul Jones

                  
has proposed - with the main difference being the numeric range of the

                  
weight. Do I have that right?

                

                
 

                
I believe that Paul J.'s algorithm is more suited towards overload

                
control, not load balancing (and I will await Paul J.'s correction to my

                
assertion).

                
 

                
The intent in draft-bessis-dispatch-adaptive-load-balancing is to

                
*always* report the current load to give the upstream SIP proxy an

                
indication of the health of the SIP server in the cluster.  If one waits

                
until overload is imminent, then it may be too late since a sudden and

                
unanticipated burst of messages will put the SIP server in overload

                
rather quickly.

                
 

                

                  
It isn't clear to me from your draft if you are specifying the

                  
algorithm by which the relative weights are used to determine the

                  
probability of a request going to one server or another.

                

                
 

                
Yes, same way as RFC2782 does it.

                
 

                

                  
ISTM that it *does* matter. And it especially matters if there is an

                  
implication that the upstream node's algorithm may depend on some

                  
assumption or knowledge about the downstream nodes.

                

                
 

                
No knowledge, just a reported number that the upstream SIP proxy can

                
treat as a RFC2782-like weight indicator.  Today, rfc3263-resolution

                
starts off with the SRV indicators in RFC2782, without any regard to

                
additional knowledge or assumption about downstream nodes.

                
dispatch-adaptive-load-balancing is trying to use the same model, except

                
render the weight to be sensitive to the current traffic in the cluster

                
in real-time.

                
 

                

                  
And of course this method of reporting weights isn't sufficient to

                  
drive a balancing algorithm that is sensitive to some requests

                  
requiring more or different resources than others. (That seems to be a

                  
real problem, though I don't know if we have a real solution for

                  
it.)

                

                
 

                
Yes, that is a real problem, as is the problem of a SIP server in a

                
cluster that has a low weight and is thus not contacted with enough

                
frequency to merit more participation to the goodput.

                
 

                
It is well known in SIP that some methods are more onerous to process

                
than others --- from a proxying point of view, processing an INVITE with

                
forking and response aggregation is far more detrimental to resource

                
usage than is processing a BYE.  Some existing literature techniques

                
thus add a coefficient to the weight of sending an INVITE versus a BYE.

                
At this time, the process outlined in dispatch-adaptive-load-balancing

                
treats all requests the same, but we can definitely look at ways to

                
change this behaviour.

                
 

                
Thanks,

                
 

                
- vijay

                
--

                
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent

                
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60566 (USA)

                
Email: vkg@{bell-labs.com,acm.org} / vijay.gurbani@alcatel-lucent.com

                
Web:   http://ect.bell-labs.com/who/vkg/

                
_______________________________________________

                
dispatch mailing list

                
dispatch@ietf.org

                
https://www.ietf.org/mailman/listinfo/dispatch

              

              
 

              
_______________________________________________

              
dispatch mailing list

              
dispatch@ietf.org

              
https://www.ietf.org/mailman/listinfo/dispatch

              
 

              

                
-- 

                  --
                      

                      Cordially, 

                      Thierry Bessis 

                
  IMS
                    Solution: Architecture & Product Management -
                    ALTA Member

                      Organization: ALU > Carrier group > Wireline
                    BU >  IMS Sol Arch, SE, PdM

                  Alcatel-Lucent

                    1960
                    Lucent Lane - Room 6L-517

                      Naperville, IL  60563  U S

                      Tel/Fax: +1 630 979 7989

                      Corporate IM: tbessis    - Yahoo IM:    
                    thierry_bessis

                  

                  My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                    My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                    ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

                  

                  Conference information:

                    2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674
                    - Access Code: 9797989

                    others countries see: http://all.alcatel-lucent.com/audio

                  

                  Upcoming
                      planned Business Trip: 12 April - 14 April 2011

                  Upcoming
                      planned Vacation: None 

              

            

            
 

            
 

            
_______________________________________________

            
dispatch mailing list

            
dispatch@ietf.org

            
https://www.ietf.org/mailman/listinfo/dispatch

            
 

            

              
-- 

                --
                    

                    Cordially, 

                    Thierry Bessis 

              
  IMS
                  Solution: Architecture & Product Management - ALTA
                  Member

                    Organization: ALU > Carrier group > Wireline
                  BU >  IMS Sol Arch, SE, PdM

                Alcatel-Lucent

                  1960 Lucent
                  Lane - Room 6L-517

                    Naperville, IL  60563  U S

                    Tel/Fax: +1 630 979 7989

                    Corporate IM: tbessis    - Yahoo IM:    
                  thierry_bessis

                

                My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                  My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                  ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

                

                Conference information:

                  2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 -
                  Access Code: 9797989

                  others countries see: http://all.alcatel-lucent.com/audio

                

                Upcoming
                    planned Business Trip: 12 April - 14 April 2011

                Upcoming
                    planned Vacation: None 

            

          

          
 

          
 

          
_______________________________________________

          
dispatch mailing list

          
dispatch@ietf.org

          
https://www.ietf.org/mailman/listinfo/dispatch

          
 

          

            
-- 

              -- 

                  Cordially, 

                  Thierry Bessis 

            
  IMS Solution:
                Architecture & Product Management - ALTA Member

                  Organization: ALU > Carrier group > Wireline BU
                >  IMS Sol Arch, SE, PdM

              Alcatel-Lucent

                1960 Lucent
                Lane - Room 6L-517

                  Naperville, IL  60563  U S

                  Tel/Fax: +1 630 979 7989

                  Corporate IM: tbessis    - Yahoo IM:    
                thierry_bessis

              

              My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

                My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

                ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

              

              Conference information:

                2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 -
                Access Code: 9797989

                others countries see: http://all.alcatel-lucent.com/audio

              

              Upcoming planned Business Trip: 12
                  April - 14 April 2011

              Upcoming planned Vacation: None
              

          

        

      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------070001000806000304020300
Content-Type: image/jpeg
Content-Transfer-Encoding: base64
Content-ID: 
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--------------070001000806000304020300
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------070001000806000304020300--

--------------050204040908060906060900--

From henry.sinnreich@gmail.com  Fri Mar 11 09:27:39 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9CA1D3A6911 for ; Fri, 11 Mar 2011 09:27:39 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.085
X-Spam-Level: 
X-Spam-Status: No, score=-3.085 tagged_above=-999 required=5 tests=[AWL=0.514,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id djmqRSKZJBtQ for ; Fri, 11 Mar 2011 09:27:38 -0800 (PST)
Received: from mail-yw0-f44.google.com (mail-yw0-f44.google.com [209.85.213.44]) by core3.amsl.com (Postfix) with ESMTP id A47983A68BD for ; Fri, 11 Mar 2011 09:27:37 -0800 (PST)
Received: by ywi6 with SMTP id 6so1638991ywi.31 for ; Fri, 11 Mar 2011 09:28:57 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=dEKLuZt+T0uPlc9yNByh6TiJjJghSkiLYgItFhqVjGw=; b=o/3TJsA+MovnRi5mhU+IeVX1nfwa/iji5XCENePGmuXdxp93RsSuyibYRDWTNdNgsV iMU+plh1DqfwzPeTgv3F0yl7Wzg/MuLZLsYMo2qvcHQ7VjFvnbTa9X0wVBQd54Tiyxzl CVlG8n3FW/6H7yT7Tw3rTHAvwCQcaCkcxu6Vc=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=G6UuPXMjeThM0vEhIG5ptSpPutI4E4z/Y5W546bf8P/Ppwk9gEDxB34jL3VVsn5nUk 0M7wcp1VaSQkwPxZDRHwEYgZ1iZc4yHcmSvi5Wccnk1GFN/UDLE/gdz6XMmqJctEK43H eLYgDQqlvCIh2/z6sTMtzCBDZr6mIwE/NKLJw=
Received: by 10.236.95.176 with SMTP id p36mr2800190yhf.112.1299864536561; Fri, 11 Mar 2011 09:28:56 -0800 (PST)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id 68sm3058797yhl.19.2011.03.11.09.28.54 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 11 Mar 2011 09:28:55 -0800 (PST)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Fri, 11 Mar 2011 11:28:53 -0600
From: Henry Sinnreich 
To: Christer Holmberg , Andrew Allen , "tanakai@nttdocomo.co.jp" , "dispatch@ietf.org" 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvfRQM5LNYNOSZVz0atfnKVY5yxgAAIeg3wABRgSYAAFleInQ==
In-Reply-To: <7F2072F1E0DE894DA4B517B93C6A05851948FF2D4A@ESESSCMS0356.eemea.ericsson.se>
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 17:27:39 -0000

Hi Andrew, Christer,

>> My understanding also is that IETF nearly a decade ago agreed to
>> cooperate with 3GPP (and other standards bodies) in developing and
>> enhancing internet protocols

3GPP etc. documents are best served by having them online, up-to-date, on
their respective web sites, where they can be searched and studied.
Using IETF RFCs for this seems rather odd, and certainly less effective.

>some people in IETF suddently thinks there are more "fun and sexy" stuff to
>work on than SIP...

As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost all
SIP apps are telephony features, while 100,000s of apps are all using HTML,
some of them to communicate in new ways - quite different from the SIP
telephone call signaling models. RTC-Web may or may not be based on SIP
signaling _alone_ , if at all. Communication models and user habits have
changed very much (and are still changing) since SIP has been defined.

Don't you think we have a problem here?
If yes, wouldn't listening to the IETF technical plenary presentation on the
future of applications be a good start to address the challenge for SIP
applications?

What do you think?

Henry


On 3/11/11 12:57 AM, "Christer Holmberg" 
wrote:

> 
> Hi,
> 
> I agree with Andrew.
> 
> People have invested, and are investing, lots of resources into SIP based
> systems. They are not suddently going to throw everything away just because
> some people in IETF suddently thinks there are more "fun and sexy" stuff to
> work on than SIP...
> 
> Also, regarding the agreement between IETF and 3GPP, if people have issues
> with that agreement I don't think that people who are working based on that
> agreement should have to suffer.
> 
> Instead, people having issues with the agreement should bring those issues to
> a more appropriate instance.
> 
> Regards,
> 
> Christer
> 
> 
>> -----Original Message-----
>> From: dispatch-bounces@ietf.org
>> [mailto:dispatch-bounces@ietf.org] On Behalf Of Andrew Allen
>> Sent: 11. maaliskuuta 2011 0:04
>> To: Henry Sinnreich; tanakai@nttdocomo.co.jp; dispatch@ietf.org
>> Subject: Re: [dispatch]
>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>> 
>> Hi Henry,
>> 
>> I think then that it's a fair summary of your email to say
>> that your objections are of a political and philosophical
>> nature related to your view of what constitutes the internet
>> and whether the IETF should be involved in enhancing IETF
>> protocols for stakeholders that utilize those IETF protocols
>> for deployments that some in the community might not consider
>> "the Internet" or "the pure Internet" or the "open e2e
>> Internet" rather than technical issues with the contents of
>> the drafts.
>> 
>> While I don't want to start a general debate on what is "the Internet"
>> here I will state that in my view it's very difficult to
>> exclude any particular stakeholders and their deployments of
>> IETF protocols from being part of the Internet when those
>> stakeholders are in many cases the very same established
>> telecommunications companies or ISPs that the vast majority
>> of the population in most countries use to obtain access to
>> the Internet. 
>> 
>> My understanding also is that IETF nearly a decade ago agreed
>> to cooperate with 3GPP (and other standards bodies) in
>> developing and enhancing internet protocols and that IETF
>> specifically requested (more like insisted) that any
>> extensions or enhancements to IETF protocols required by 3GPP
>> be done within IETF according to the procedures defined by
>> IETF for enhancing IETF protocols and for registering namespaces etc.
>> These drafts simply comply with that agreed way of working
>> between IETF and 3GPP. To my knowledge this working agreement
>> is still in effect. I don't think it is the role of the
>> dispatch WG to determine which applications or deployments of
>> SIP to support or cooperate with counter to agreements
>> previously made at IESG level.
>> 
>> Refusing cooperation based on what is considered not to be
>> part of the "open e2e Internet" is only likely to set up two
>> (or more) competing versions of the Internet resulting in the
>> 21st century equivalent of the standard rail gauge vs broad
>> rail gauge communications interoperability issue from the
>> 19th century - effectively killing the goal of an open e2e
>> internet! As Dale indicated what is considered a walled
>> garden today may not be walled in the future (if I remember
>> some very significant walls came tumbling down during just
>> the last decade of the 20th century alone).
>> 
>> Different business models will be resolved through the market
>> not by creating competing incompatible technical solutions.
>> 
>> In terms of the time used on discussion within dispatch I
>> don't think there is a need for much more discussion in
>> dispatch. These drafts could go ahead as AD sponsored without
>> further WG involvement. As I understand it the IMEI URN
>> definition draft did not as part of the URN registration
>> process need to even be discussed in dispatch at all
>> (although Dale Worley's review was invaluable in improving
>> the draft technically) as the NIDs list was the fora for this
>> discussion and where previous versions of that draft were discussed.
>> 
>> It should be noted all we are talking about with regard to
>> these drafts is the registration of a URN namespace and
>> definition of its usage as an Instance ID as required by RFC
>> 5626. If we can't agree that then what future extensions can we agree?
>> 
>> Andrew
>> 
>> -----Original Message-----
>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
>> Sent: Thursday, March 10, 2011 11:03 AM
>> To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org
>> Subject: Re: [dispatch]
>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>> 
>> Hi Andrew,
>> 
>> I have generic objections to see the publication of yet
>> another informational RFC about practices in consortiums of
>> closed networks in the IETF, for the following reasons:
>> 
>> * Using the RFC process as a publishing venue for
>> organizations that don't explicitly do that in support of the
>> open e2e Internet.
>> 
>> * The resources of the IETF are not unlimited, meeting time
>> and email time to discuss is scarce and the IETF registries
>> are just not scalable enough to support registrations for all
>> the extensions that benefit countless external organizations
>> and companies that deploy Internet technology.
>> New business plans and the required extensions pop up in an
>> endless fashion.
>> Big organizations such as you mention have their own
>> resources to publish their internal technical documents and
>> white papers if they desire to do so.
>> 
>> Members of outside organizations are certainly doing
>> everybody a good service by informing via email of new
>> standards/extensions in other organization, such as you have
>> done below, but such information need not overload the
>> Internet RFC process nor overload the Internet registries.
>> 
>> * Hard for SIP developers and right out impossible for
>> application developers.
>> Though SIP works fine indeed across the world; SIPit and the
>> SIP Forum take care of unavoidable problems still encountered
>> by operators, arguably increasing SIP complexity for
>> non-Internet purposes is not a good idea, even if new SIP
>> extensions are labeled as just "informational". It makes it
>> even more difficult for new developers entering the field to
>> ever understand, let alone write code for some or all that
>> has been published already about SIP for example. Checking
>> right now there are 197 RFCs matching SIP and 78 I-Ds found
>> by the RFC search engine.
>> 
>> It is time to leave SIP alone and move on to other critical
>> areas for the Internet, such as for example the future of
>> applications that will be discussed in the next IETF
>> technical plenary on Monday, March 28.
>> 
>> With this in mind, I would rather not go into discussing your
>> I-D in detail, though at first glance it seems technically
>> correct and if it were discussed on the list, may indeed
>> technically qualify as an informational RFC - just what I am
>> arguing against.
>> 
>> These are all personal opinions. The chairs and ADs may or
>> may not discuss or agree/disagree with. Your I-D is just a
>> good opportunity to raise these issues.
>> 
>> Note that I am not suggesting any voting on your I-D, if it
>> will be discussed, since for various reasons, the majority of
>> folks attending the DISPATCH WG are from various outside
>> organizations with business plans that may or may not fully
>> embrace the open e2e Internet; a perfect legitimate choice
>> for them, but not  in the IETF (I just can't remember when
>> ever there was a vote against any new I-D for SIP, since no
>> cost was perceived by those voting).
>> 
>> Thanks,
>> Henry 
>> 
>> 
>> On 3/9/11 5:45 PM, "Andrew Allen"  wrote:
>> 
>>> 
>>> Henry
>>> 
>>> Would you care to elaborate and explain your specific concerns?  How
>> is what
>>> is defined in these drafts potentially incompatible with
>> the open, e2e 
>>> architectural and in any other way technical conflict with the
>> Internet?
>>> 
>>> The two drafts do the following things:
>>> 
>>> One draft simply defines and registers with IANA a URN namespace
>> according to
>>> the procedures defined for doing that.
>>> 
>>> The other draft simply defines how this URN can be used as
>> an instance
>> ID as
>>> required by RFC 5626.
>>> 
>>> SIP and the mechanisms in RFC 5626 are already being used
>> in networks
>> defined
>>> by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA
>> have
>>> already defined this behavior using their existing device
>> Identifiers
>> for
>>> instance ID generation in their specifications. These
>> drafts complete
>> the
>>> process doing things the right way by reserving the namespace and
>> defining the
>>> usage of it using the IETF defined mechanisms for doing that.
>>> 
>>> Andrew
>>> 
>>> ----- Original Message -----
>>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
>>> Sent: Wednesday, March 09, 2011 03:10 PM
>>> To: Itsuma TANAKA ; dispatch@ietf.org
>>> 
>>> Subject: Re: [dispatch]
>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>>> 
>>> Such work in the IETF would require IMO an explanation about its
>>> compatibility with the open, e2e architectural and other technical
>>> principles of the Internet. It actually seems to be the opposite.
>>> 
>>> For this reason, such work may be better accomplished in other
>>> organizations, so as to avoid architectural principle and other
>> technical
>>> conflicts. Simplicity and scalability comes to mind among other,
>> though this
>>> is better discussed elsewhere.
>>> 
>>> Thanks,
>>> 
>>> Henry Sinnreich
>>> 
>>> 
>>> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
>>> 
>>>> Dear All,
>>>> 
>>>> I work for NTT DOCOMO, Inc in Japan. and attend various
>> 3GPP meetings
>> (SA2,
>>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming
>> Expert
>>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup
>> meetings.  I
>> serve
>>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>>>> 
>>>> I would like to express my support of
>> draft-montemurro-gsma-imei-urn
>> and the
>>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>>>> 
>>>> The reasons are exactly same as what Ricky Kaura (Samsung)
>> expressed.
>> A
>>>> various 3GPP specifications (IMS Registration in 3GPP
>> TS24.229, IMS 
>>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and
>> TS24.292), and
>>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are
>> using
>>>> these drafts.
>>>> 
>>>> These internet drafts are very important to support various key
>> features in
>>>> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would
>> like ask
>>>> that this work is progressed as quickly as possible in IETF.
>>>> 
>>>> 
>>>> Best regards,
>>>> 
>>>> Itsuma Tanaka
>>>> NTT DOCOMO
>>>> 
>>>> 
>>>> _______________________________________________
>>>> dispatch mailing list
>>>> dispatch@ietf.org
>>>> https://www.ietf.org/mailman/listinfo/dispatch
>>> 
>>> 
>>> _______________________________________________
>>> dispatch mailing list
>>> dispatch@ietf.org
>>> https://www.ietf.org/mailman/listinfo/dispatch
>>> 
>>> 
>> ---------------------------------------------------------------------
>>> This transmission (including any attachments) may contain
>> confidential 
>>> information, privileged material (including material
>> protected by the
>>> solicitor-client or other applicable privileges), or constitute
>> non-public
>>> information. Any use of this information by anyone other than the
>> intended
>>> recipient is prohibited. If you have received this transmission in
>> error,
>>> please immediately reply to the sender and delete this information
>> from your
>>> system. Use, dissemination, distribution, or reproduction of this
>> transmission
>>> by unintended recipients is not authorized and may be unlawful.
>> 
>> 
>> 
>> 
>> 
>> 
>> ---------------------------------------------------------------------
>> This transmission (including any attachments) may contain
>> confidential information, privileged material (including
>> material protected by the solicitor-client or other
>> applicable privileges), or constitute non-public information.
>> Any use of this information by anyone other than the intended
>> recipient is prohibited. If you have received this
>> transmission in error, please immediately reply to the sender
>> and delete this information from your system. Use,
>> dissemination, distribution, or reproduction of this
>> transmission by unintended recipients is not authorized and
>> may be unlawful.
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>> 



From henry.sinnreich@gmail.com  Fri Mar 11 10:19:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EE02E3A6930 for ; Fri, 11 Mar 2011 10:19:50 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.125
X-Spam-Level: 
X-Spam-Status: No, score=-3.125 tagged_above=-999 required=5 tests=[AWL=0.474,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WXfuK5RozxNM for ; Fri, 11 Mar 2011 10:19:50 -0800 (PST)
Received: from mail-yw0-f44.google.com (mail-yw0-f44.google.com [209.85.213.44]) by core3.amsl.com (Postfix) with ESMTP id 053443A68EE for ; Fri, 11 Mar 2011 10:19:49 -0800 (PST)
Received: by ywi6 with SMTP id 6so1663224ywi.31 for ; Fri, 11 Mar 2011 10:21:09 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:mime-version:content-type :content-transfer-encoding; bh=Rylj/tAuXvGtjhXlBpRtzSzv2Vxd1tl8016TWCu4o4E=; b=L+6HNpWs802aEg4XBqOiWaIhOgxyr81TzoWRBTV/cVGnfkeV0NrwuZH84uOHEDX+B9 7JO3OnK9Bk1F5tX8ycOBK9g5quM0h2AX/ytbG+RTMq0g9pBm2AV0SWqJGQIUmJFL3MxB LTlhQSDuZu9dEnSujaBStNFvgKYEDQ/2GZ/p8=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:mime-version:content-type:content-transfer-encoding; b=Xzc+NwxiBBRZzgOY00OqylScaKAv0mauX+LwnqqLuEZ2uGnC4/l69XKMz0KzHCdM17 tsNHjUQmvX6EnRrOnL7dzMi8fy5Z7fEhMTuIOuXeZ1DZxYvzDRaTL/tr96c6pMib3LJj dL+UfybBO9kETJzVAChhTSCbD3f6l/QcvnzyU=
Received: by 10.90.112.10 with SMTP id k10mr206301agc.55.1299867669297; Fri, 11 Mar 2011 10:21:09 -0800 (PST)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id r8sm3707660ane.39.2011.03.11.10.21.07 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 11 Mar 2011 10:21:08 -0800 (PST)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Fri, 11 Mar 2011 11:26:46 -0600
From: Henry Sinnreich 
To: Andrew Allen , , , Christer Holmberg 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvgEX8A2H+ZlOr37Ema4ZFaWh80Rw==
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 18:19:51 -0000

Hi Andrew, Christer,

> My understanding also is that IETF nearly a decade ago agreed to
> cooperate with 3GPP (and other standards bodies) in developing and
> enhancing internet protocols

3GPP etc. documents are best served by having them online, up-to-date, on
their respective web sites, where they can be easily searched and studied.
Using IETF RFCs for this seem rather odd, and certainly less effective.

>some people in IETF suddently thinks there are more "fun and sexy" stuff to
>work on than SIP...

As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost all
SIP apps are telephony features, while 100,000s of apps are all using HTML,
some of them to communicate in new ways - quite different from the SIP
telephone call signaling models. RTC-Web may or may not be based on SIP
signaling _alone_ , if at all. Communication models and user habits have
changed very much (and are still changing) since SIP has been defined.

Don't you think we have a problem here?
If yes, wouldn't listening to the IETF technical plenary presentation on the
future of applications be a good start to address the challenge for SIP
applications?

What do you think?

Henry









From pkyzivat@cisco.com  Fri Mar 11 10:27:30 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 30AB83A6A58 for ; Fri, 11 Mar 2011 10:27:30 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.541
X-Spam-Level: 
X-Spam-Status: No, score=-110.541 tagged_above=-999 required=5 tests=[AWL=0.058, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qwQ03xpyRHjl for ; Fri, 11 Mar 2011 10:27:29 -0800 (PST)
Received: from sj-iport-4.cisco.com (sj-iport-4.cisco.com [171.68.10.86]) by core3.amsl.com (Postfix) with ESMTP id 0ABCC3A6992 for ; Fri, 11 Mar 2011 10:27:29 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=2499; q=dns/txt; s=iport; t=1299868128; x=1301077728; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=o/kv4e/adSyPmGFq1ieD4XcJFU2xH95CRhtuuR8DTwM=; b=HkQLYVWGGvlhPIptddu9zVoHNZF9VvCwtGkTxhfFlOEhLvT9GvUoXE/h FQ3sX4B+1Wyd00sFTtesKCLcqvfw6PX7reDVZpv5fu/mra9lfSoeY8S5d BdB/VzQWGjmzWQy8RhvnlLsK/ia4u4xKwfQyP7WXAdBVkwRAs6h/cMctN c=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AtwGAEj4eU2tJXG9/2dsb2JhbACZJI0Od6YbnCCFYgSFKYcmg0o
X-IronPort-AV: E=Sophos;i="4.62,304,1297036800"; d="scan'208";a="274873001"
Received: from rcdn-core2-2.cisco.com ([173.37.113.189]) by sj-iport-4.cisco.com with ESMTP; 11 Mar 2011 18:28:48 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by rcdn-core2-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2BISlug008097 for ; Fri, 11 Mar 2011 18:28:47 GMT
Message-ID: <4D7A69DF.3060405@cisco.com>
Date: Fri, 11 Mar 2011 13:28:47 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.13) Gecko/20101207 Thunderbird/3.1.7
MIME-Version: 1.0
To: dispatch@ietf.org
References: 
In-Reply-To: 
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 18:27:30 -0000

Henry,

The following is from fdraft-allen-dispatch-imei-urn-as-instanceid:

    RFC 5626 [1] defines that a UA SHOULD create a Universally Unique
    Identifier (UUID) URN as defined in RFC 4122 [6] as its instance-id
    but allows for the possibility of other URN schemes to be used.  If a
    URN scheme other than UUID is used, the UA MUST only use URNs for
    which an RFC (from the IETF stream) defines how the specific URN
    needs to be constructed and used in the "+sip.instance" Contact
    header field parameter for outbound behavior.  This specification
    meets this requirement by specifying how the GSMA IEMEI URN is used
    in the "+sip.instance" Contact header field parameter for outbound
    behavior and draft-montemurro-gsma-imei-urn-06 [3] defines how the
    GSMA IMEI URN is constructed

They are doing what we required them to do. If we had not required an 
RFC then they wouldn't be here with this draft. How can we fault them 
for following the rules?

	Thanks,
	Paul

On 3/11/2011 12:26 PM, Henry Sinnreich wrote:
> Hi Andrew, Christer,
>
>> My understanding also is that IETF nearly a decade ago agreed to
>> cooperate with 3GPP (and other standards bodies) in developing and
>> enhancing internet protocols
>
> 3GPP etc. documents are best served by having them online, up-to-date, on
> their respective web sites, where they can be easily searched and studied.
> Using IETF RFCs for this seem rather odd, and certainly less effective.
>
>> some people in IETF suddently thinks there are more "fun and sexy" stuff to
>> work on than SIP...
>
> As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost all
> SIP apps are telephony features, while 100,000s of apps are all using HTML,
> some of them to communicate in new ways - quite different from the SIP
> telephone call signaling models. RTC-Web may or may not be based on SIP
> signaling _alone_ , if at all. Communication models and user habits have
> changed very much (and are still changing) since SIP has been defined.
>
> Don't you think we have a problem here?
> If yes, wouldn't listening to the IETF technical plenary presentation on the
> future of applications be a good start to address the challenge for SIP
> applications?
>
> What do you think?
>
> Henry
>
>
>
>
>
>
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>

From paulej@packetizer.com  Fri Mar 11 10:32:03 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D9FCD3A6C67 for ; Fri, 11 Mar 2011 10:32:03 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.598
X-Spam-Level: 
X-Spam-Status: No, score=-2.598 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ESWqfcXk-GGA for ; Fri, 11 Mar 2011 10:31:57 -0800 (PST)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id 5CB0D3A6A5C for ; Fri, 11 Mar 2011 10:31:57 -0800 (PST)
Received: from sydney (rrcs-98-101-155-83.midsouth.biz.rr.com [98.101.155.83]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2BIX8UT029753 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Fri, 11 Mar 2011 13:33:14 -0500
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1299868394; bh=wAw9F9sODJjExFkMr9iDheo6kn2ab8/Rdh149sffzL4=; h=From:To:Cc:References:In-Reply-To:Subject:Date:Message-ID: MIME-Version:Content-Type; b=olDUIFxFD48P4XsCSpTwSSLSr12VOxXhan2cd7wqkE7WSo0QISy7nHi7KXUFFOate 8i4OoKQ/nx1uZOLFdXeyByKxvIbOu+8ghRopKY4f5yGqia4ZzS+fJS4DivkAG6/bA0 vpcB/cf8Ln27439+sUAk1uJYZi1/cX5+5d92Xlw4=
From: "Paul E. Jones" 
To: "'Thierry Bessis'" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com> <060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com> <4D7A50AA.3060805@alcatel-lucent.com>
In-Reply-To: <4D7A50AA.3060805@alcatel-lucent.com>
Date: Fri, 11 Mar 2011 13:32:59 -0500
Message-ID: <074301cbe01a$c2f92330$48eb6990$@packetizer.com>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----=_NextPart_000_0744_01CBDFF0.DA2453B0"
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQHq4aNGJDi8SbEZihBBjeuIo6isswGjVapMAlUxg3kCCnyKVgIDWSzZAfKA2HwChasBsAE4aE9uAgiAjeQBjV7waQIN8UT1k099VKA=
Content-Language: en-us
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 18:32:04 -0000

This is a multipart message in MIME format.

------=_NextPart_000_0744_01CBDFF0.DA2453B0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit

Thierry,

 

The SCE value is en estimate of the number of sessions the device can handle
given the current number of sessions giving respect to the current resource
utilization.  If there are 500 sessions, for example, and the CPU is at 50%
over the past n seconds, then one might assume the device can handle another
500 sessions.  Thus, the SCE might be set to 500.  However, this value is
subject to change every time resources are measured.  There are also "edge
cases" where the current number of sessions and the CPU is at 10%.  In such
cases where arithmetic would not yield a reasonable number, then the device
would need to assume a reasonable value (perhaps based on lab measurements
during development).  I would personally measure each important resource
type and calculate a possible SCE and then take the minimum of those SCE
values.

 

So, the value is not a ratio, but an estimate of the available capacity.

 

Now, looking at the peers to an SBC, what we see are multiple SCE values
advertised, such as (a) 400, (b) 600, (c) 800, and (d) 1000.  This tells us
that device D has the highest available capacity.  We could look at those as
"weights" and send a higher portion of traffic toward (d), but we do not.
Our approach is to then round-robin across those devices so as to avoid
overloading a device with too many SIP messages.  After all, just because a
device can handle more simultaneous calls does not mean that it has a higher
message-handling capacity: this is why we would want to complement SCE with
the current SOC work that is focused on message throttling.  Having said
that, nothing would require one to perform round-robin in SCE.  The
algorithm to select the peer given a set of SCE values was not described in
our document (see Section 6 - we leave that open to implementation).

 

Paul

 

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Friday, March 11, 2011 11:41 AM
To: Paul E. Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing

 

Paul, 



The specific ways through which the SCE value is computed may vary by
vendor, though the approach we have used is one where we calculate an
estimate of the number of sessions a device can handle given the current
number of sessions given current memory utilization.  We then do the same
for CPU and other system resources.  We then take the minimum value of each
and report that as the SCE value.  As a matter of course, our devices are
constantly measuring these resources, so computing the SCE if trivial.

So this mode of operation is very similar to what we also propose. Is that
number a ratio (percentage of sessions available over the total number), or
an absolute number ? If this is a ratio, then SCE and our load balancing are
compatible: the same information and transport mechanism could be used for
both: A single IETF proposal could be done. 

Cordially, 

Thierry 


------=_NextPart_000_0744_01CBDFF0.DA2453B0
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Thierry's signature
Thierry,
 
The SCE value is en estimate of the number of sessions the device can =
handle given the current number of sessions giving respect to the =
current resource utilization.  If there are 500 sessions, for =
example, and the CPU is at 50% over the past n seconds, then one might =
assume the device can handle another 500 sessions.  Thus, the SCE =
might be set to 500.  However, this value is subject to change =
every time resources are measured.  There are also “edge =
cases” where the current number of sessions and the CPU is at =
10%.  In such cases where arithmetic would not yield a reasonable =
number, then the device would need to assume a reasonable value (perhaps =
based on lab measurements during development).  I would personally =
measure each important resource type and calculate a possible SCE and =
then take the minimum of those SCE values.
 
So, the value is not a ratio, but an estimate of the available =
capacity.
 
Now, looking at the peers to an SBC, what we see are multiple SCE =
values advertised, such as (a) 400, (b) 600, (c) 800, and (d) =
1000.  This tells us that device D has the highest available =
capacity.  We could look at those as “weights” and send =
a higher portion of traffic toward (d), but we do not.  Our =
approach is to then round-robin across those devices so as to avoid =
overloading a device with too many SIP messages.  After all, just =
because a device can handle more simultaneous calls does not mean that =
it has a higher message-handling capacity: this is why we would want to =
complement SCE with the current SOC work that is focused on message =
throttling.  Having said that, nothing would require one to perform =
round-robin in SCE.  The algorithm to select the peer given a set =
of SCE values was not described in our document (see Section 6 – =
we leave that open to implementation).
 
Paul
 
From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com] =

Sent: Friday, March 11, 2011 11:41 AM
To: Paul E. =
Jones
Cc: dispatch@ietf.org
Subject: Re: [dispatch] =
SIP load balancing
 
Paul, =


The =
specific ways through which the SCE value is computed may vary by =
vendor, though the approach we have used is one where we calculate an =
estimate of the number of sessions a device can handle given the current =
number of sessions given current memory utilization.  We then do =
the same for CPU and other system resources.  We then take the =
minimum value of each and report that as the SCE value.  As a =
matter of course, our devices are constantly measuring these resources, =
so computing the SCE if trivial.
So this mode of operation is very similar to what we =
also propose. Is that number a ratio (percentage of sessions available =
over the total number), or an absolute number ? If this is a ratio, then =
SCE and our load balancing are compatible: the same information and =
transport mechanism could be used for both: A single IETF proposal could =
be done. 

Cordially, 

Thierry 

------=_NextPart_000_0744_01CBDFF0.DA2453B0--


From mary.ietf.barnes@gmail.com  Fri Mar 11 11:26:13 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AD2293A6BC8 for ; Fri, 11 Mar 2011 11:26:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.955
X-Spam-Level: 
X-Spam-Status: No, score=-102.955 tagged_above=-999 required=5 tests=[AWL=-0.357, BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id frt7ZdJEdy+t for ; Fri, 11 Mar 2011 11:26:10 -0800 (PST)
Received: from mail-vx0-f172.google.com (mail-vx0-f172.google.com [209.85.220.172]) by core3.amsl.com (Postfix) with ESMTP id E1A5F3A68DC for ; Fri, 11 Mar 2011 11:26:09 -0800 (PST)
Received: by vxg33 with SMTP id 33so3451193vxg.31 for ; Fri, 11 Mar 2011 11:27:28 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc:content-type; bh=sJ/6DMz91jtyHn/EDkZHAGsZdYOPvfXxW6D6pgmfVU0=; b=sthTm7KVdvt0nGOhCH5U0c7fGKTLipKSJ5u+FsviPHoeg20CGJhiHW7K1Hinzp3oh2 OpuxyCQfAFiKZOWjXVni2hZUNWTIuXhCk5/vxmSdD0y8Jp9ET0iINde08Azrt6K0Lzeg K0ra/9ntc+QSsh1We48TTCTc/KFvFJFMyPxgo=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=Z0c5TzuJgp/nkqSsTllbfpZ7xOtkX/dZWubOFZTH4GNW5OP5Ur8NvlWS0jSGLbazE4 iBnUofPrO5fnAKV8OoP9cuV41ETeaexP1fQpL0G2fKiR0JqIISTkEM3ANn278oX5dWPL I2fD17Pbk2cjSvE4E4L7/KV8rcX9d6+tHp7wY=
MIME-Version: 1.0
Received: by 10.52.69.46 with SMTP id b14mr8806900vdu.103.1299871648752; Fri, 11 Mar 2011 11:27:28 -0800 (PST)
Received: by 10.52.162.202 with HTTP; Fri, 11 Mar 2011 11:27:28 -0800 (PST)
In-Reply-To: 
References: <7F2072F1E0DE894DA4B517B93C6A05851948FF2D4A@ESESSCMS0356.eemea.ericsson.se> 
Date: Fri, 11 Mar 2011 13:27:28 -0600
Message-ID: 
From: Mary Barnes 
To: Henry Sinnreich 
Content-Type: multipart/alternative; boundary=20cf307d03bae0a559049e39f385
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 19:26:13 -0000

--20cf307d03bae0a559049e39f385
Content-Type: text/plain; charset=ISO-8859-1

Hi Henry,

I'll make just a few comments on this discussion thread, as I don't think
it's a productive discussion on this mailing list - likely it would be
better on the ietf-discussion list or you might want to approach someone on
the IAB about this topic, since the IAB is the responsible party in
establishing and appointing liaisons.

However, as far as the 3gpp specs that are all online, fairly up-to-date as
best as I can tell:
http://www.3gpp.org/-specifications-

As you puruse the specs, you will find that 3gpp uses a number of the specs
in the exact same manner as do "e2e" implementations.  So, I don't think the
situation is as dire as it might appear.  We can debate this topic ad
naseum, but IMS is what it is and is being deployed and nothing any of us
say or think here will change that.  The reality is that a large number of
contributors to the e2e specs are here because the folks that pay them to
contribute to the IETF make money selling those systems.  I will also note
that folks that aren't interested in IMS can use your document that was
published as a baseline "e2e" SIP implementation:
http://tools.ietf.org/html/rfc5638

There is also a spreadsheet of IETF dependencies that is regularly updated
and it includes references to the specs that depend upon core IETF RFCs:
http://www.3gpp.org/-IETF-Dependencies-and-Priorities-

I will note that summaries of issues and actions related to the RAI area
documents has been posted on the RAI area mailing list in the past:
http://www.ietf.org/mail-archive/web/rai/current/msg00955.html
So, I think there has been a reasonable amount of transparency with regards
to this working relationship. Since this relationship between 3GPP and IETF
has been in place, there have been a number of documents that have been
approved and in general it has been a fruitful relationship on both parts.
 In some cases, core changes were made to SIP that benefited both the e2e
model and the walled garden.

Regards,
Mary.

On Fri, Mar 11, 2011 at 11:28 AM, Henry Sinnreich  wrote:

> Hi Andrew, Christer,
>
> >> My understanding also is that IETF nearly a decade ago agreed to
> >> cooperate with 3GPP (and other standards bodies) in developing and
> >> enhancing internet protocols
>
> 3GPP etc. documents are best served by having them online, up-to-date, on
> their respective web sites, where they can be searched and studied.
> Using IETF RFCs for this seems rather odd, and certainly less effective.
>
> >some people in IETF suddently thinks there are more "fun and sexy" stuff
> to
> >work on than SIP...
>
> As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost all
> SIP apps are telephony features, while 100,000s of apps are all using HTML,
> some of them to communicate in new ways - quite different from the SIP
> telephone call signaling models. RTC-Web may or may not be based on SIP
> signaling _alone_ , if at all. Communication models and user habits have
> changed very much (and are still changing) since SIP has been defined.
>
> Don't you think we have a problem here?
> If yes, wouldn't listening to the IETF technical plenary presentation on
> the
> future of applications be a good start to address the challenge for SIP
> applications?
>
> What do you think?
>
> Henry
>
>
> On 3/11/11 12:57 AM, "Christer Holmberg" 
> wrote:
>
> >
> > Hi,
> >
> > I agree with Andrew.
> >
> > People have invested, and are investing, lots of resources into SIP based
> > systems. They are not suddently going to throw everything away just
> because
> > some people in IETF suddently thinks there are more "fun and sexy" stuff
> to
> > work on than SIP...
> >
> > Also, regarding the agreement between IETF and 3GPP, if people have
> issues
> > with that agreement I don't think that people who are working based on
> that
> > agreement should have to suffer.
> >
> > Instead, people having issues with the agreement should bring those
> issues to
> > a more appropriate instance.
> >
> > Regards,
> >
> > Christer
> >
> >
> >> -----Original Message-----
> >> From: dispatch-bounces@ietf.org
> >> [mailto:dispatch-bounces@ietf.org] On Behalf Of Andrew Allen
> >> Sent: 11. maaliskuuta 2011 0:04
> >> To: Henry Sinnreich; tanakai@nttdocomo.co.jp; dispatch@ietf.org
> >> Subject: Re: [dispatch]
> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> >>
> >> Hi Henry,
> >>
> >> I think then that it's a fair summary of your email to say
> >> that your objections are of a political and philosophical
> >> nature related to your view of what constitutes the internet
> >> and whether the IETF should be involved in enhancing IETF
> >> protocols for stakeholders that utilize those IETF protocols
> >> for deployments that some in the community might not consider
> >> "the Internet" or "the pure Internet" or the "open e2e
> >> Internet" rather than technical issues with the contents of
> >> the drafts.
> >>
> >> While I don't want to start a general debate on what is "the Internet"
> >> here I will state that in my view it's very difficult to
> >> exclude any particular stakeholders and their deployments of
> >> IETF protocols from being part of the Internet when those
> >> stakeholders are in many cases the very same established
> >> telecommunications companies or ISPs that the vast majority
> >> of the population in most countries use to obtain access to
> >> the Internet.
> >>
> >> My understanding also is that IETF nearly a decade ago agreed
> >> to cooperate with 3GPP (and other standards bodies) in
> >> developing and enhancing internet protocols and that IETF
> >> specifically requested (more like insisted) that any
> >> extensions or enhancements to IETF protocols required by 3GPP
> >> be done within IETF according to the procedures defined by
> >> IETF for enhancing IETF protocols and for registering namespaces etc.
> >> These drafts simply comply with that agreed way of working
> >> between IETF and 3GPP. To my knowledge this working agreement
> >> is still in effect. I don't think it is the role of the
> >> dispatch WG to determine which applications or deployments of
> >> SIP to support or cooperate with counter to agreements
> >> previously made at IESG level.
> >>
> >> Refusing cooperation based on what is considered not to be
> >> part of the "open e2e Internet" is only likely to set up two
> >> (or more) competing versions of the Internet resulting in the
> >> 21st century equivalent of the standard rail gauge vs broad
> >> rail gauge communications interoperability issue from the
> >> 19th century - effectively killing the goal of an open e2e
> >> internet! As Dale indicated what is considered a walled
> >> garden today may not be walled in the future (if I remember
> >> some very significant walls came tumbling down during just
> >> the last decade of the 20th century alone).
> >>
> >> Different business models will be resolved through the market
> >> not by creating competing incompatible technical solutions.
> >>
> >> In terms of the time used on discussion within dispatch I
> >> don't think there is a need for much more discussion in
> >> dispatch. These drafts could go ahead as AD sponsored without
> >> further WG involvement. As I understand it the IMEI URN
> >> definition draft did not as part of the URN registration
> >> process need to even be discussed in dispatch at all
> >> (although Dale Worley's review was invaluable in improving
> >> the draft technically) as the NIDs list was the fora for this
> >> discussion and where previous versions of that draft were discussed.
> >>
> >> It should be noted all we are talking about with regard to
> >> these drafts is the registration of a URN namespace and
> >> definition of its usage as an Instance ID as required by RFC
> >> 5626. If we can't agree that then what future extensions can we agree?
> >>
> >> Andrew
> >>
> >> -----Original Message-----
> >> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> >> Sent: Thursday, March 10, 2011 11:03 AM
> >> To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org
> >> Subject: Re: [dispatch]
> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> >>
> >> Hi Andrew,
> >>
> >> I have generic objections to see the publication of yet
> >> another informational RFC about practices in consortiums of
> >> closed networks in the IETF, for the following reasons:
> >>
> >> * Using the RFC process as a publishing venue for
> >> organizations that don't explicitly do that in support of the
> >> open e2e Internet.
> >>
> >> * The resources of the IETF are not unlimited, meeting time
> >> and email time to discuss is scarce and the IETF registries
> >> are just not scalable enough to support registrations for all
> >> the extensions that benefit countless external organizations
> >> and companies that deploy Internet technology.
> >> New business plans and the required extensions pop up in an
> >> endless fashion.
> >> Big organizations such as you mention have their own
> >> resources to publish their internal technical documents and
> >> white papers if they desire to do so.
> >>
> >> Members of outside organizations are certainly doing
> >> everybody a good service by informing via email of new
> >> standards/extensions in other organization, such as you have
> >> done below, but such information need not overload the
> >> Internet RFC process nor overload the Internet registries.
> >>
> >> * Hard for SIP developers and right out impossible for
> >> application developers.
> >> Though SIP works fine indeed across the world; SIPit and the
> >> SIP Forum take care of unavoidable problems still encountered
> >> by operators, arguably increasing SIP complexity for
> >> non-Internet purposes is not a good idea, even if new SIP
> >> extensions are labeled as just "informational". It makes it
> >> even more difficult for new developers entering the field to
> >> ever understand, let alone write code for some or all that
> >> has been published already about SIP for example. Checking
> >> right now there are 197 RFCs matching SIP and 78 I-Ds found
> >> by the RFC search engine.
> >>
> >> It is time to leave SIP alone and move on to other critical
> >> areas for the Internet, such as for example the future of
> >> applications that will be discussed in the next IETF
> >> technical plenary on Monday, March 28.
> >>
> >> With this in mind, I would rather not go into discussing your
> >> I-D in detail, though at first glance it seems technically
> >> correct and if it were discussed on the list, may indeed
> >> technically qualify as an informational RFC - just what I am
> >> arguing against.
> >>
> >> These are all personal opinions. The chairs and ADs may or
> >> may not discuss or agree/disagree with. Your I-D is just a
> >> good opportunity to raise these issues.
> >>
> >> Note that I am not suggesting any voting on your I-D, if it
> >> will be discussed, since for various reasons, the majority of
> >> folks attending the DISPATCH WG are from various outside
> >> organizations with business plans that may or may not fully
> >> embrace the open e2e Internet; a perfect legitimate choice
> >> for them, but not  in the IETF (I just can't remember when
> >> ever there was a vote against any new I-D for SIP, since no
> >> cost was perceived by those voting).
> >>
> >> Thanks,
> >> Henry
> >>
> >>
> >> On 3/9/11 5:45 PM, "Andrew Allen"  wrote:
> >>
> >>>
> >>> Henry
> >>>
> >>> Would you care to elaborate and explain your specific concerns?  How
> >> is what
> >>> is defined in these drafts potentially incompatible with
> >> the open, e2e
> >>> architectural and in any other way technical conflict with the
> >> Internet?
> >>>
> >>> The two drafts do the following things:
> >>>
> >>> One draft simply defines and registers with IANA a URN namespace
> >> according to
> >>> the procedures defined for doing that.
> >>>
> >>> The other draft simply defines how this URN can be used as
> >> an instance
> >> ID as
> >>> required by RFC 5626.
> >>>
> >>> SIP and the mechanisms in RFC 5626 are already being used
> >> in networks
> >> defined
> >>> by 3GPP and are being deployed by members of the GSMA. 3GPP and GSMA
> >> have
> >>> already defined this behavior using their existing device
> >> Identifiers
> >> for
> >>> instance ID generation in their specifications. These
> >> drafts complete
> >> the
> >>> process doing things the right way by reserving the namespace and
> >> defining the
> >>> usage of it using the IETF defined mechanisms for doing that.
> >>>
> >>> Andrew
> >>>
> >>> ----- Original Message -----
> >>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
> >>> Sent: Wednesday, March 09, 2011 03:10 PM
> >>> To: Itsuma TANAKA ; dispatch@ietf.org
> >>> 
> >>> Subject: Re: [dispatch]
> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
> >>>
> >>> Such work in the IETF would require IMO an explanation about its
> >>> compatibility with the open, e2e architectural and other technical
> >>> principles of the Internet. It actually seems to be the opposite.
> >>>
> >>> For this reason, such work may be better accomplished in other
> >>> organizations, so as to avoid architectural principle and other
> >> technical
> >>> conflicts. Simplicity and scalability comes to mind among other,
> >> though this
> >>> is better discussed elsewhere.
> >>>
> >>> Thanks,
> >>>
> >>> Henry Sinnreich
> >>>
> >>>
> >>> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrote:
> >>>
> >>>> Dear All,
> >>>>
> >>>> I work for NTT DOCOMO, Inc in Japan. and attend various
> >> 3GPP meetings
> >> (SA2,
> >>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roaming
> >> Expert
> >>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup
> >> meetings.  I
> >> serve
> >>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
> >>>>
> >>>> I would like to express my support of
> >> draft-montemurro-gsma-imei-urn
> >> and the
> >>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
> >>>>
> >>>> The reasons are exactly same as what Ricky Kaura (Samsung)
> >> expressed.
> >> A
> >>>> various 3GPP specifications (IMS Registration in 3GPP
> >> TS24.229, IMS
> >>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and
> >> TS24.292), and
> >>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS") are
> >> using
> >>>> these drafts.
> >>>>
> >>>> These internet drafts are very important to support various key
> >> features in
> >>>> 3GPP and to enable GSMA Voice over LTE.  Therefore, I kindly would
> >> like ask
> >>>> that this work is progressed as quickly as possible in IETF.
> >>>>
> >>>>
> >>>> Best regards,
> >>>>
> >>>> Itsuma Tanaka
> >>>> NTT DOCOMO
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> dispatch mailing list
> >>>> dispatch@ietf.org
> >>>> https://www.ietf.org/mailman/listinfo/dispatch
> >>>
> >>>
> >>> _______________________________________________
> >>> dispatch mailing list
> >>> dispatch@ietf.org
> >>> https://www.ietf.org/mailman/listinfo/dispatch
> >>>
> >>>
> >> ---------------------------------------------------------------------
> >>> This transmission (including any attachments) may contain
> >> confidential
> >>> information, privileged material (including material
> >> protected by the
> >>> solicitor-client or other applicable privileges), or constitute
> >> non-public
> >>> information. Any use of this information by anyone other than the
> >> intended
> >>> recipient is prohibited. If you have received this transmission in
> >> error,
> >>> please immediately reply to the sender and delete this information
> >> from your
> >>> system. Use, dissemination, distribution, or reproduction of this
> >> transmission
> >>> by unintended recipients is not authorized and may be unlawful.
> >>
> >>
> >>
> >>
> >>
> >>
> >> ---------------------------------------------------------------------
> >> This transmission (including any attachments) may contain
> >> confidential information, privileged material (including
> >> material protected by the solicitor-client or other
> >> applicable privileges), or constitute non-public information.
> >> Any use of this information by anyone other than the intended
> >> recipient is prohibited. If you have received this
> >> transmission in error, please immediately reply to the sender
> >> and delete this information from your system. Use,
> >> dissemination, distribution, or reproduction of this
> >> transmission by unintended recipients is not authorized and
> >> may be unlawful.
> >> _______________________________________________
> >> dispatch mailing list
> >> dispatch@ietf.org
> >> https://www.ietf.org/mailman/listinfo/dispatch
> >>
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>

--20cf307d03bae0a559049e39f385
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hi Henry,=A0

I'll make just a few comments on this d=
iscussion thread, as I don't think it's a productive discussion on =
this mailing list - likely it would be better on the ietf-discussion list o=
r you might want to approach someone on the IAB about this topic, since the=
 IAB is the responsible party in establishing and appointing liaisons.=A0



However, as far as the 3gpp specs that are all online, =
fairly up-to-date as best as I can tell:




As you puruse the specs, you will find that 3gpp uses a numb=
er of the specs in the exact same manner as do "e2e" implementati=
ons. =A0So, I don't think the situation is as dire as it might appear. =
=A0We can debate this topic ad naseum, but IMS is what it is and is being d=
eployed and nothing any of us say or think here will change that. =A0The re=
ality is that a large number of contributors to the e2e specs are here beca=
use the folks that pay them to contribute to the IETF make money selling th=
ose systems. =A0I will also note that folks that aren't interested in I=
MS can use your document that was published as a baseline "e2e" S=
IP implementation:=A0http://=
tools.ietf.org/html/rfc5638




There is also a spreadsheet of IETF dependencies that i=
s regularly updated and it includes references to the specs that depend upo=
n core IETF RFCs:




I will note that summaries of issues and actions relate=
d to the RAI area documents has been posted on the RAI area mailing list in=
 the past:


So, I think there has been a reasonable amount of transparency with re=
gards to this working relationship. Since this relationship between 3GPP an=
d IETF has been in place, there have been a number of documents that have b=
een approved and in general it has been a fruitful relationship on both par=
ts. =A0In some cases, core changes were made to SIP that benefited both the=
 e2e model and the walled garden. =A0



Regards,
Mary.=A0


On Fri, Mar 11, 2011 at 11:28 AM, Henry Sinnreich <henry.sinnreich=
@gmail.com> wrote:


Hi Andrew, Christer,




>> My understanding also is that IETF nearly a decade ago agreed to
>> cooperate with 3GPP (and other standards bodies) in developing and=


>> enhancing internet protocols




3GPP etc. documents are best served by having them online, up-to-date=
, on

their respective web sites, where they can be searched and studied.

Using IETF RFCs for this seems rather odd, and certainly less effective.



>some people in IETF suddently thinks there are more "fun and sexy&=
quot; stuff to

>work on than SIP...




As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, al=
most all

SIP apps are telephony features, while 100,000s of apps are all using HTML,=


some of them to communicate in new ways - quite different from the SIP

telephone call signaling models. RTC-Web may or may not be based on SIP

signaling _alone_ , if at all. Communication models and user habits have
changed very much (and are still changing) since SIP has been defined.



Don't you think we have a problem here?

If yes, wouldn't listening to the IETF technical plenary presentation o=
n the

future of applications be a good start to address the challenge for SIP

applications?



What do you think?



Henry





On 3/11/11 12:57 AM, "Christer Holmberg" <christer.holmberg@ericsson.=
com>

wrote:




>

> Hi,

>

> I agree with Andrew.

>

> People have invested, and are investing, lots of resources into SIP ba=
sed

> systems. They are not suddently going to throw everything away just be=
cause

> some people in IETF suddently thinks there are more "fun and sexy=
" stuff to

> work on than SIP...

>

> Also, regarding the agreement between IETF and 3GPP, if people have is=
sues

> with that agreement I don't think that people who are working base=
d on that

> agreement should have to suffer.

>

> Instead, people having issues with the agreement should bring those is=
sues to

> a more appropriate instance.

>

> Regards,

>

> Christer

>

>

>> -----Original Message-----

>> From: dispatch-bounces@ietf.org

>> [mailto:dispatch-bounces@ietf.org] On Behalf Of Andrew Allen

>> Sent: 11. maaliskuuta 2011 0:04

>> To: Henry Sinnreich; tanakai@nttdocomo.co.jp; dispatch@ietf.org

>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>

>> Hi Henry,

>>

>> I think then that it's a fair summary of your email to say

>> that your objections are of a political and philosophical

>> nature related to your view of what constitutes the internet

>> and whether the IETF should be involved in enhancing IETF

>> protocols for stakeholders that utilize those IETF protocols

>> for deployments that some in the community might not consider

>> "the Internet" or "the pure Internet" or the &=
quot;open e2e

>> Internet" rather than technical issues with the contents of
>> the drafts.

>>

>> While I don't want to start a general debate on what is "=
the Internet"

>> here I will state that in my view it's very difficult to

>> exclude any particular stakeholders and their deployments of

>> IETF protocols from being part of the Internet when those

>> stakeholders are in many cases the very same established

>> telecommunications companies or ISPs that the vast majority

>> of the population in most countries use to obtain access to

>> the Internet.

>>

>> My understanding also is that IETF nearly a decade ago agreed

>> to cooperate with 3GPP (and other standards bodies) in

>> developing and enhancing internet protocols and that IETF

>> specifically requested (more like insisted) that any

>> extensions or enhancements to IETF protocols required by 3GPP

>> be done within IETF according to the procedures defined by

>> IETF for enhancing IETF protocols and for registering namespaces e=
tc.

>> These drafts simply comply with that agreed way of working

>> between IETF and 3GPP. To my knowledge this working agreement

>> is still in effect. I don't think it is the role of the

>> dispatch WG to determine which applications or deployments of

>> SIP to support or cooperate with counter to agreements

>> previously made at IESG level.

>>

>> Refusing cooperation based on what is considered not to be

>> part of the "open e2e Internet" is only likely to set up=
 two

>> (or more) competing versions of the Internet resulting in the

>> 21st century equivalent of the standard rail gauge vs broad

>> rail gauge communications interoperability issue from the

>> 19th century - effectively killing the goal of an open e2e

>> internet! As Dale indicated what is considered a walled

>> garden today may not be walled in the future (if I remember

>> some very significant walls came tumbling down during just

>> the last decade of the 20th century alone).

>>

>> Different business models will be resolved through the market

>> not by creating competing incompatible technical solutions.

>>

>> In terms of the time used on discussion within dispatch I

>> don't think there is a need for much more discussion in

>> dispatch. These drafts could go ahead as AD sponsored without

>> further WG involvement. As I understand it the IMEI URN

>> definition draft did not as part of the URN registration

>> process need to even be discussed in dispatch at all

>> (although Dale Worley's review was invaluable in improving

>> the draft technically) as the NIDs list was the fora for this

>> discussion and where previous versions of that draft were discusse=
d.

>>

>> It should be noted all we are talking about with regard to

>> these drafts is the registration of a URN namespace and

>> definition of its usage as an Instance ID as required by RFC

>> 5626. If we can't agree that then what future extensions can w=
e agree?

>>

>> Andrew

>>

>> -----Original Message-----

>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]

>> Sent: Thursday, March 10, 2011 11:03 AM

>> To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org

>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>

>> Hi Andrew,

>>

>> I have generic objections to see the publication of yet

>> another informational RFC about practices in consortiums of

>> closed networks in the IETF, for the following reasons:

>>

>> * Using the RFC process as a publishing venue for

>> organizations that don't explicitly do that in support of the<=
br>
>> open e2e Internet.

>>

>> * The resources of the IETF are not unlimited, meeting time

>> and email time to discuss is scarce and the IETF registries

>> are just not scalable enough to support registrations for all

>> the extensions that benefit countless external organizations

>> and companies that deploy Internet technology.

>> New business plans and the required extensions pop up in an

>> endless fashion.

>> Big organizations such as you mention have their own

>> resources to publish their internal technical documents and

>> white papers if they desire to do so.

>>

>> Members of outside organizations are certainly doing

>> everybody a good service by informing via email of new

>> standards/extensions in other organization, such as you have

>> done below, but such information need not overload the

>> Internet RFC process nor overload the Internet registries.

>>

>> * Hard for SIP developers and right out impossible for

>> application developers.

>> Though SIP works fine indeed across the world; SIPit and the

>> SIP Forum take care of unavoidable problems still encountered

>> by operators, arguably increasing SIP complexity for

>> non-Internet purposes is not a good idea, even if new SIP

>> extensions are labeled as just "informational". It makes=
 it

>> even more difficult for new developers entering the field to

>> ever understand, let alone write code for some or all that

>> has been published already about SIP for example. Checking

>> right now there are 197 RFCs matching SIP and 78 I-Ds found

>> by the RFC search engine.

>>

>> It is time to leave SIP alone and move on to other critical

>> areas for the Internet, such as for example the future of

>> applications that will be discussed in the next IETF

>> technical plenary on Monday, March 28.

>>

>> With this in mind, I would rather not go into discussing your

>> I-D in detail, though at first glance it seems technically

>> correct and if it were discussed on the list, may indeed

>> technically qualify as an informational RFC - just what I am

>> arguing against.

>>

>> These are all personal opinions. The chairs and ADs may or

>> may not discuss or agree/disagree with. Your I-D is just a

>> good opportunity to raise these issues.

>>

>> Note that I am not suggesting any voting on your I-D, if it

>> will be discussed, since for various reasons, the majority of

>> folks attending the DISPATCH WG are from various outside

>> organizations with business plans that may or may not fully

>> embrace the open e2e Internet; a perfect legitimate choice

>> for them, but not =A0in the IETF (I just can't remember when
>> ever there was a vote against any new I-D for SIP, since no

>> cost was perceived by those voting).

>>

>> Thanks,

>> Henry

>>

>>

>> On 3/9/11 5:45 PM, "Andrew Allen" <aallen@rim.com> wrote:

>>

>>>

>>> Henry

>>>

>>> Would you care to elaborate and explain your specific concerns=
? =A0How

>> is what

>>> is defined in these drafts potentially incompatible with

>> the open, e2e

>>> architectural and in any other way technical conflict with the=


>> Internet?

>>>

>>> The two drafts do the following things:

>>>

>>> One draft simply defines and registers with IANA a URN namespa=
ce

>> according to

>>> the procedures defined for doing that.

>>>

>>> The other draft simply defines how this URN can be used as

>> an instance

>> ID as

>>> required by RFC 5626.

>>>

>>> SIP and the mechanisms in RFC 5626 are already being used

>> in networks

>> defined

>>> by 3GPP and are being deployed by members of the GSMA. 3GPP an=
d GSMA

>> have

>>> already defined this behavior using their existing device

>> Identifiers

>> for

>>> instance ID generation in their specifications. These

>> drafts complete

>> the

>>> process doing things the right way by reserving the namespace =
and

>> defining the

>>> usage of it using the IETF defined mechanisms for doing that.<=
br>
>>>

>>> Andrew

>>>

>>> ----- Original Message -----

>>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]

>>> Sent: Wednesday, March 09, 2011 03:10 PM

>>> To: Itsuma TANAKA <tanakai@nttdocomo.co.jp>; dispatch@ietf.org

>>> <dis=
patch@ietf.org>

>>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>>

>>> Such work in the IETF would require IMO an explanation about i=
ts

>>> compatibility with the open, e2e architectural and other techn=
ical

>>> principles of the Internet. It actually seems to be the opposi=
te.

>>>

>>> For this reason, such work may be better accomplished in other=


>>> organizations, so as to avoid architectural principle and othe=
r

>> technical

>>> conflicts. Simplicity and scalability comes to mind among othe=
r,

>> though this

>>> is better discussed elsewhere.

>>>

>>> Thanks,

>>>

>>> Henry Sinnreich

>>>

>>>

>>> On 3/9/11 8:54 AM, "Itsuma TANAKA" <tanakai@nttdocomo.co.jp=
> wrote:

>>>

>>>> Dear All,

>>>>

>>>> I work for NTT DOCOMO, Inc in Japan. and attend various
>> 3GPP meetings

>> (SA2,

>>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking=
, Roaming

>> Expert

>>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup
>> meetings. =A0I

>> serve

>>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>>>>

>>>> I would like to express my support of

>> draft-montemurro-gsma-imei-urn

>> and the

>>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>>>>

>>>> The reasons are exactly same as what Ricky Kaura (Samsung)=


>> expressed.

>> A

>>>> various 3GPP specifications (IMS Registration in 3GPP

>> TS24.229, IMS

>>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 =
and

>> TS24.292), and

>>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice=
 and SMS") are

>> using

>>>> these drafts.

>>>>

>>>> These internet drafts are very important to support variou=
s key

>> features in

>>>> 3GPP and to enable GSMA Voice over LTE. =A0Therefore, I ki=
ndly would

>> like ask

>>>> that this work is progressed as quickly as possible in IET=
F.

>>>>

>>>>

>>>> Best regards,

>>>>

>>>> Itsuma Tanaka

>>>> NTT DOCOMO

>>>>

>>>>

>>>> _______________________________________________

>>>> dispatch mailing list

>>>> dis=
patch@ietf.org

>>>> https://www.ietf.org/mailman/listinfo/dispatch

>>>

>>>

>>> _______________________________________________

>>> dispatch mailing list

>>> dispatc=
h@ietf.org

>>> https://www.ietf.org/mailman/listinfo/dispatch

>>>

>>>

>> ------------------------------------------------------------------=
---

>>> This transmission (including any attachments) may contain

>> confidential

>>> information, privileged material (including material

>> protected by the

>>> solicitor-client or other applicable privileges), or constitut=
e

>> non-public

>>> information. Any use of this information by anyone other than =
the

>> intended

>>> recipient is prohibited. If you have received this transmissio=
n in

>> error,

>>> please immediately reply to the sender and delete this informa=
tion

>> from your

>>> system. Use, dissemination, distribution, or reproduction of t=
his

>> transmission

>>> by unintended recipients is not authorized and may be unlawful=
.

>>

>>

>>

>>

>>

>>

>> ------------------------------------------------------------------=
---

>> This transmission (including any attachments) may contain

>> confidential information, privileged material (including

>> material protected by the solicitor-client or other

>> applicable privileges), or constitute non-public information.

>> Any use of this information by anyone other than the intended

>> recipient is prohibited. If you have received this

>> transmission in error, please immediately reply to the sender

>> and delete this information from your system. Use,

>> dissemination, distribution, or reproduction of this

>> transmission by unintended recipients is not authorized and

>> may be unlawful.

>> _______________________________________________

>> dispatch mailing list

>> dispatch@ie=
tf.org

>> https://www.ietf.org/mailman/listinfo/dispatch

>>





_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.org/mailman/listinfo/dispatch





--20cf307d03bae0a559049e39f385--

From henry.sinnreich@gmail.com  Fri Mar 11 12:44:04 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DA9AE3A6941 for ; Fri, 11 Mar 2011 12:44:04 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.46
X-Spam-Level: 
X-Spam-Status: No, score=-2.46 tagged_above=-999 required=5 tests=[AWL=-0.258,  BAYES_00=-2.599, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xEyXubOl1IMA for ; Fri, 11 Mar 2011 12:44:04 -0800 (PST)
Received: from mail-gy0-f172.google.com (mail-gy0-f172.google.com [209.85.160.172]) by core3.amsl.com (Postfix) with ESMTP id 18B923A691A for ; Fri, 11 Mar 2011 12:44:03 -0800 (PST)
Received: by gyf3 with SMTP id 3so200885gyf.31 for ; Fri, 11 Mar 2011 12:45:23 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:cc:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type; bh=+QLSLVHoPRLfyDpJx5C57OoQdgVXTE8fDh+C2vmRl6k=; b=SJrzENYJk2YQSsRh1BSndily8y4Xy6O0k7Vb9buBd0JbxMcy/4k+F11bRZp6G+ortl QC7TBfh97NgIBJorQCZrd8C1azvQMGJy9lsvcZ6tparM+sz13yuEeT/St72CE0AtSk8+ IgI0FSdL4iJkKqg+c4sdgmG5sgA9kgGPtNxLI=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type; b=VJqaeGtGnLY2CdIPwYLcWDe1PFYltYLjLuDSFZjEtaOWV+uN8IYxj0ZfSO3osU9WHz pwbhH7+yV4mo2nluknzAVbBfyjgOP7v98Icg3/rCSCKciFD9TLFEAKsPhuj+qq7yDw7H q6ScSMac6IT0ZOn8XS0Kfmz7VGuVea1T0irzw=
Received: by 10.236.200.137 with SMTP id z9mr3464737yhn.322.1299876323264; Fri, 11 Mar 2011 12:45:23 -0800 (PST)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id j8sm3190638yha.1.2011.03.11.12.45.20 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 11 Mar 2011 12:45:21 -0800 (PST)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Fri, 11 Mar 2011 14:45:17 -0600
From: Henry Sinnreich 
To: Mary Barnes 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvgLTqCSuk+FfxJ3UiwZArXsPwfYA==
In-Reply-To: 
Mime-version: 1.0
Content-type: multipart/alternative; boundary="B_3382699520_1100847"
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Fri, 11 Mar 2011 20:44:05 -0000

> This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

--B_3382699520_1100847
Content-type: text/plain;
	charset="ISO-8859-1"
Content-transfer-encoding: quoted-printable

Thanks Mary!

The same actually came to my mind as well.

Henry


On 3/11/11 1:27 PM, "Mary Barnes"  wrote:

> Hi Henry,=A0
>=20
> I'll make just a few comments on this discussion thread, as I don't think=
 it's
> a productive discussion on this mailing list - likely it would be better =
on
> the ietf-discussion list or you might want to approach someone on the IAB
> about this topic, since the IAB is the responsible party in establis


--B_3382699520_1100847
Content-type: text/html;
	charset="ISO-8859-1"
Content-transfer-encoding: quoted-printable



Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid=



Thanks Mary!



The same actually came to my mind as well.



Henry





On 3/11/11 1:27 PM, "Mary Barnes" <mary.ietf.barnes@gmail.com> wrote:




<=
SPAN STYLE=3D'font-size:13pt'>Hi Henry,=A0



I'll make just a few comments on this discussion thread, as I don't think i=
t's a productive discussion on this mailing list - likely it would be better=
 on the ietf-discussion list or you might want to approach someone on the IA=
B about this topic, since the IAB is the responsible party in establis







--B_3382699520_1100847--



From Thierry.Bessis@alcatel-lucent.com  Fri Mar 11 16:20:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EE4413A6A14 for ; Fri, 11 Mar 2011 16:20:58 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.972
X-Spam-Level: 
X-Spam-Status: No, score=-5.972 tagged_above=-999 required=5 tests=[AWL=0.625,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bMn1LQMZulip for ; Fri, 11 Mar 2011 16:20:48 -0800 (PST)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 31C763A672F for ; Fri, 11 Mar 2011 16:20:47 -0800 (PST)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p2C0M6br025469 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 11 Mar 2011 18:22:06 -0600 (CST)
Received: from [135.244.21.53] (tbessis.lra.lucent.com [135.244.21.53]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2C0M4E9025327; Fri, 11 Mar 2011 18:22:05 -0600 (CST)
Message-ID: <4D7ABCAC.2070802@alcatel-lucent.com>
Date: Fri, 11 Mar 2011 18:22:04 -0600
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Paul E. Jones" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>	<060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>	<4D7A50AA.3060805@alcatel-lucent.com> <074301cbe01a$c2f92330$48eb6990$@packetizer.com>
In-Reply-To: <074301cbe01a$c2f92330$48eb6990$@packetizer.com>
Content-Type: multipart/alternative; boundary="------------070600070400040400070609"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sat, 12 Mar 2011 00:20:59 -0000

This is a multi-part message in MIME format.
--------------070600070400040400070609
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Paul,

I understand. It seems that the SCE feedback would be as dynamic as the 
it is in the load balancing.
I also agree that SCE could be used to do some load balancing based on 
the available sessions left on each node.

However SCE is using the weight as a representation of the number of 
available sessions, while load balancing is using the weight as an 
utilization ratio. The requirement we used for our load balancing 
requirement is based on the need to balance the utilization ratio on 
each node, (we think the system resists better and longer to 
near-overload situations this way) so the utilization rate must be 
available. In addition to provide what we need, such weight is very 
straightforward to calculate (it's actually more like a measurement): We 
do not have to come up with any assumption, rules or simulations to 
figure out how resources translate to sessions. Another thing is that 
the selection algorithm can be the same as with SRV: Very easy and cheap 
to implement, and still providing continuous load sharing. Also the load 
balancing works great with OLC, but independently: Both algorithm 
operate independently in a serial arrangement (First selection, then 
shedding), and there is no need for a re-selection when the request is 
shed, since the cluster is supposed to be balanced. Here again, the 
implementation can remain modular and very simple.

I agree that SCE can provide similar results, but you need to come up 
with remaining sessions. I agree we could do load distribution, but only 
based on remaining sessions (this probably would also work great, but we 
think load distribution based on utilization rate is better in our case, 
and we know it works). Finally, I think with SCE the algorithm to select 
the final destination when OLC is also used would be more complex.

So I think that the weight should represent the relative occupation of 
the node for those reasons. However we could add an option to provide 
also (or instead) an estimation of the number of sessions available, so 
that those who prefer to distribute based on that instead could do it.
We also did provide for more than one "variant" with OLC ...

Cordially,

Thierry





On 11-Mar-2011 12:32, Paul E. Jones wrote:
>
> Thierry,
>
> The SCE value is en estimate of the number of sessions the device can 
> handle given the current number of sessions giving respect to the 
> current resource utilization.  If there are 500 sessions, for example, 
> and the CPU is at 50% over the past n seconds, then one might assume 
> the device can handle another 500 sessions.  Thus, the SCE might be 
> set to 500.  However, this value is subject to change every time 
> resources are measured.  There are also "edge cases" where the current 
> number of sessions and the CPU is at 10%.  In such cases where 
> arithmetic would not yield a reasonable number, then the device would 
> need to assume a reasonable value (perhaps based on lab measurements 
> during development).  I would personally measure each important 
> resource type and calculate a possible SCE and then take the minimum 
> of those SCE values.
>
> So, the value is not a ratio, but an estimate of the available capacity.
>
> Now, looking at the peers to an SBC, what we see are multiple SCE 
> values advertised, such as (a) 400, (b) 600, (c) 800, and (d) 1000.  
> This tells us that device D has the highest available capacity.  We 
> could look at those as "weights" and send a higher portion of traffic 
> toward (d), but we do not.  Our approach is to then round-robin across 
> those devices so as to avoid overloading a device with too many SIP 
> messages.  After all, just because a device can handle more 
> simultaneous calls does not mean that it has a higher message-handling 
> capacity: this is why we would want to complement SCE with the current 
> SOC work that is focused on message throttling.  Having said that, 
> nothing would require one to perform round-robin in SCE.  The 
> algorithm to select the peer given a set of SCE values was not 
> described in our document (see Section 6 -- we leave that open to 
> implementation).
>
> Paul
>
> *From:*Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
> *Sent:* Friday, March 11, 2011 11:41 AM
> *To:* Paul E. Jones
> *Cc:* dispatch@ietf.org
> *Subject:* Re: [dispatch] SIP load balancing
>
> Paul,
>
> The specific ways through which the SCE value is computed may vary by 
> vendor, though the approach we have used is one where we calculate an 
> estimate of the number of sessions a device can handle given the 
> current number of sessions given current memory utilization.  We then 
> do the same for CPU and other system resources.  We then take the 
> minimum value of each and report that as the SCE value.  As a matter 
> of course, our devices are constantly measuring these resources, so 
> computing the SCE if trivial.
>
> So this mode of operation is very similar to what we also propose. Is 
> that number a ratio (percentage of sessions available over the total 
> number), or an absolute number ? If this is a ratio, then SCE and our 
> load balancing are compatible: the same information and transport 
> mechanism could be used for both: A single IETF proposal could be done.
>
> Cordially,
>
> Thierry
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------070600070400040400070609
Content-Type: multipart/related;
 boundary="------------040403020908080104070205"


--------------040403020908080104070205
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Paul, 

    

    I understand. It seems that the SCE feedback would be as dynamic as
    the it is in the load balancing. 

    I also agree that SCE could be used to do some load balancing based
    on the available sessions left on each node. 

    

    However SCE is using the weight as a representation of the number of
    available sessions, while load balancing is using the weight as an
    utilization ratio. The requirement we used for our load balancing
    requirement is based on the need to balance the utilization ratio on
    each node, (we think the system resists better and longer to
    near-overload situations this way) so the utilization rate must be
    available. In addition to provide what we need, such weight is very
    straightforward to calculate (it's actually more like a
    measurement): We do not have to come up with any assumption, rules
    or simulations to figure out how resources translate to sessions.
    Another thing is that the selection algorithm can be the same as
    with SRV: Very easy and cheap to implement, and still providing
    continuous load sharing. Also the load balancing works great with
    OLC, but independently: Both algorithm operate independently in a
    serial arrangement (First selection, then shedding), and there is no
    need for a re-selection when the request is shed, since the cluster
    is supposed to be balanced. Here again, the implementation can
    remain modular and very simple. 

    

    I agree that SCE can provide similar results, but you need to come
    up with remaining sessions. I agree we could do load distribution,
    but only based on remaining sessions (this probably would also work
    great, but we think load distribution based on utilization rate is
    better in our case, and we know it works). Finally, I think with SCE
    the algorithm to select the final destination when OLC is also used
    would be more complex. 

    

    So I think that the weight should represent the relative occupation
    of the node for those reasons. However we could add an option to
    provide also (or instead) an estimation of the number of sessions
    available, so that those who prefer to distribute based on that
    instead could do it. 

    We also did provide for more than one "variant" with OLC ... 

    

    Cordially, 

    

    Thierry

    

    

    

    

    

    On 11-Mar-2011 12:32, Paul E. Jones wrote:
    

      
      
      Thierry's signature
      
      

        
Thierry,

        
 

        
The SCE value is en estimate of the number of
            sessions the device can handle given the current number of
            sessions giving respect to the current resource
            utilization.  If there are 500 sessions, for example, and
            the CPU is at 50% over the past n seconds, then one might
            assume the device can handle another 500 sessions.  Thus,
            the SCE might be set to 500.  However, this value is subject
            to change every time resources are measured.  There are also
            “edge cases” where the current number of sessions and the
            CPU is at 10%.  In such cases where arithmetic would not
            yield a reasonable number, then the device would need to
            assume a reasonable value (perhaps based on lab measurements
            during development).  I would personally measure each
            important resource type and calculate a possible SCE and
            then take the minimum of those SCE values.

        
 

        
So, the value is not a ratio, but an estimate of
            the available capacity.

        
 

        
Now, looking at the peers to an SBC, what we see
            are multiple SCE values advertised, such as (a) 400, (b)
            600, (c) 800, and (d) 1000.  This tells us that device D has
            the highest available capacity.  We could look at those as
            “weights” and send a higher portion of traffic toward (d),
            but we do not.  Our approach is to then round-robin across
            those devices so as to avoid overloading a device with too
            many SIP messages.  After all, just because a device can
            handle more simultaneous calls does not mean that it has a
            higher message-handling capacity: this is why we would want
            to complement SCE with the current SOC work that is focused
            on message throttling.  Having said that, nothing would
            require one to perform round-robin in SCE.  The algorithm to
            select the peer given a set of SCE values was not described
            in our document (see Section 6 – we leave that open to
            implementation).

        
 

        
Paul

        
 

        

          

            

              
From: Thierry Bessis
                  [mailto:Thierry.Bessis@alcatel-lucent.com] 

                  Sent: Friday, March 11, 2011 11:41 AM

                  To: Paul E. Jones

                  Cc: dispatch@ietf.org

                  Subject: Re: [dispatch] SIP load balancing

            

          

          
 

          
Paul, 

            

            

          

            
The
                specific ways through which the SCE value is computed
                may vary by vendor, though the approach we have used is
                one where we calculate an estimate of the number of
                sessions a device can handle given the current number of
                sessions given current memory utilization.  We then do
                the same for CPU and other system resources.  We then
                take the minimum value of each and report that as the
                SCE value.  As a matter of course, our devices are
                constantly measuring these resources, so computing the
                SCE if trivial.

          

          
So this mode of operation is very similar
            to what we also propose. Is that number a ratio (percentage
            of sessions available over the total number), or an absolute
            number ? If this is a ratio, then SCE and our load balancing
            are compatible: the same information and transport mechanism
            could be used for both: A single IETF proposal could be
            done. 

            

            Cordially, 

            

            Thierry 

        

      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------040403020908080104070205
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------040403020908080104070205--

--------------070600070400040400070609--

From pkyzivat@cisco.com  Fri Mar 11 16:36:54 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6C8403A6ADB for ; Fri, 11 Mar 2011 16:36:54 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.541
X-Spam-Level: 
X-Spam-Status: No, score=-110.541 tagged_above=-999 required=5 tests=[AWL=0.058, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id chMrIxGnNCrZ for ; Fri, 11 Mar 2011 16:36:53 -0800 (PST)
Received: from ams-iport-2.cisco.com (ams-iport-2.cisco.com [144.254.224.141]) by core3.amsl.com (Postfix) with ESMTP id 426353A67B2 for ; Fri, 11 Mar 2011 16:36:53 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=302; q=dns/txt; s=iport; t=1299890293; x=1301099893; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=DEaaDGmQDDip/j2+OD94w9703Fro/OhfzIFKCcO2f9E=; b=aWxB32BL7I1TH7uSf1GVbDK0lgOLInoCOEfxjdrSKMXoGVMGFsxw9s5y n/1YtFZbRZVUg+A1FcnuWImFH7XC1v4Dc5X74xs4IRCSYwHnNy4ypXSUl exS7gJVyTiud3WM3iUwhnvd7a/kGsUqDaHpRqAJ1OmFEgBal/O/CQZwnx o=;
X-IronPort-AV: E=Sophos;i="4.62,306,1297036800"; d="scan'208";a="21405908"
Received: from ams-core-2.cisco.com ([144.254.72.75]) by ams-iport-2.cisco.com with ESMTP; 12 Mar 2011 00:38:12 +0000
Received: from [161.44.174.114] (dhcp-161-44-174-114.cisco.com [161.44.174.114]) by ams-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2C0cCMs004856; Sat, 12 Mar 2011 00:38:12 GMT
Message-ID: <4D7AC073.8010806@cisco.com>
Date: Fri, 11 Mar 2011 19:38:11 -0500
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: dispatch@ietf.org
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>	<060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>	<4D7A50AA.3060805@alcatel-lucent.com>	<074301cbe01a$c2f92330$48eb6990$@packetizer.com> <4D7ABCAC.2070802@alcatel-lucent.com>
In-Reply-To: <4D7ABCAC.2070802@alcatel-lucent.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sat, 12 Mar 2011 00:36:54 -0000

On 3/11/2011 7:22 PM, Thierry Bessis wrote:

...

> The requirement we used for our load balancing
> requirement is based on the need to balance the utilization ratio on
> each node

Ah!!! Finally I understand what you are up to!
It never made sense to me up till now.

	Thanks,
	Paul

From aallen@rim.com  Sat Mar 12 14:59:25 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C39D53A6A34 for ; Sat, 12 Mar 2011 14:59:25 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.651
X-Spam-Level: 
X-Spam-Status: No, score=-1.651 tagged_above=-999 required=5 tests=[AWL=0.447,  BAYES_00=-2.599, HTML_MESSAGE=0.001, MIME_BAD_LINEBREAK=0.5]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id R9AxNVIIM7Ej for ; Sat, 12 Mar 2011 14:59:20 -0800 (PST)
Received: from mhs061cnc.rim.net (mhs061cnc.rim.net [208.65.73.35]) by core3.amsl.com (Postfix) with ESMTP id DE0DC3A6A15 for ; Sat, 12 Mar 2011 14:59:18 -0800 (PST)
X-AuditID: 0a412830-b7ca2ae0000003d5-61-4d7bfb0ba075
Received: from XCH138CNC.rim.net (xch138cnc.rim.net [10.65.20.127]) by mhs061cnc.rim.net (SBG) with SMTP id F2.46.00981.B0BFB7D4; Sat, 12 Mar 2011 23:00:27 +0000 (GMT)
Received: from XCH02DFW.rim.net ([10.150.100.31]) by XCH138CNC.rim.net with Microsoft SMTPSVC(6.0.3790.3959); Sat, 12 Mar 2011 18:00:37 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01CBE108.4BD61574"
Date: Sat, 12 Mar 2011 16:53:25 -0600
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvgIl42kPbjR/gjTkGSZWWm9wulvwA5ezXb
From: "Andrew Allen" 
To: , 
X-OriginalArrivalTime: 12 Mar 2011 23:00:37.0860 (UTC) FILETIME=[4D559240:01CBE109]
X-Brightmail-Tracker: AAAAAQAAAZE=
Cc: dispatch@ietf.org
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sat, 12 Mar 2011 22:59:26 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBE108.4BD61574
Content-Type: text/plain;
	charset="UTF-8"
content-transfer-encoding: base64

DQpIZW5yeQ0KDQpJIHRoaW5rIE1hcnkgaGFzIGNvdmVyZWQgdGhpcyBhbGwgdmVyeSB3ZWxs
Lg0KDQpJIGFtIGhhcHB5IHRvIGVuZ2FnZSBpbiBmdXJ0aGVyIHBoaWxvc29waGljYWwgZGlz
Y3Vzc2lvbnMgb24gdGhlIGZ1dHVyZSBvZiB0aGUgSW50ZXJuZXQgaW4gUHJhZ3VlLg0KDQpB
bmRyZXcNCiANCg0KRnJvbTogTWFyeSBCYXJuZXMgW21haWx0bzptYXJ5LmlldGYuYmFybmVz
QGdtYWlsLmNvbV0gDQpTZW50OiBGcmlkYXksIE1hcmNoIDExLCAyMDExIDAxOjI3IFBNDQpU
bzogSGVucnkgU2lubnJlaWNoIDxoZW5yeS5zaW5ucmVpY2hAZ21haWwuY29tPiANCkNjOiBD
aHJpc3RlciBIb2xtYmVyZyA8Y2hyaXN0ZXIuaG9sbWJlcmdAZXJpY3Nzb24uY29tPjsgQW5k
cmV3IEFsbGVuOyB0YW5ha2FpQG50dGRvY29tby5jby5qcCA8dGFuYWthaUBudHRkb2NvbW8u
Y28uanA+OyBkaXNwYXRjaEBpZXRmLm9yZyA8ZGlzcGF0Y2hAaWV0Zi5vcmc+IA0KU3ViamVj
dDogUmU6IFtkaXNwYXRjaF0gUmV2aXNpb25vZmRyYWZ0LWFsbGVuLWRpc3BhdGNoLWltZWkt
dXJuLWFzLWluc3RhbmNlaWQgDQogDQoNCkhpIEhlbnJ5LCANCg0KSSdsbCBtYWtlIGp1c3Qg
YSBmZXcgY29tbWVudHMgb24gdGhpcyBkaXNjdXNzaW9uIHRocmVhZCwgYXMgSSBkb24ndCB0
aGluayBpdCdzIGEgcHJvZHVjdGl2ZSBkaXNjdXNzaW9uIG9uIHRoaXMgbWFpbGluZyBsaXN0
IC0gbGlrZWx5IGl0IHdvdWxkIGJlIGJldHRlciBvbiB0aGUgaWV0Zi1kaXNjdXNzaW9uIGxp
c3Qgb3IgeW91IG1pZ2h0IHdhbnQgdG8gYXBwcm9hY2ggc29tZW9uZSBvbiB0aGUgSUFCIGFi
b3V0IHRoaXMgdG9waWMsIHNpbmNlIHRoZSBJQUIgaXMgdGhlIHJlc3BvbnNpYmxlIHBhcnR5
IGluIGVzdGFibGlzaGluZyBhbmQgYXBwb2ludGluZyBsaWFpc29ucy4gDQoNCkhvd2V2ZXIs
IGFzIGZhciBhcyB0aGUgM2dwcCBzcGVjcyB0aGF0IGFyZSBhbGwgb25saW5lLCBmYWlybHkg
dXAtdG8tZGF0ZSBhcyBiZXN0IGFzIEkgY2FuIHRlbGw6DQpodHRwOi8vd3d3LjNncHAub3Jn
Ly1zcGVjaWZpY2F0aW9ucy0NCg0KQXMgeW91IHB1cnVzZSB0aGUgc3BlY3MsIHlvdSB3aWxs
IGZpbmQgdGhhdCAzZ3BwIHVzZXMgYSBudW1iZXIgb2YgdGhlIHNwZWNzIGluIHRoZSBleGFj
dCBzYW1lIG1hbm5lciBhcyBkbyAiZTJlIiBpbXBsZW1lbnRhdGlvbnMuICBTbywgSSBkb24n
dCB0aGluayB0aGUgc2l0dWF0aW9uIGlzIGFzIGRpcmUgYXMgaXQgbWlnaHQgYXBwZWFyLiAg
V2UgY2FuIGRlYmF0ZSB0aGlzIHRvcGljIGFkIG5hc2V1bSwgYnV0IElNUyBpcyB3aGF0IGl0
IGlzIGFuZCBpcyBiZWluZyBkZXBsb3llZCBhbmQgbm90aGluZyBhbnkgb2YgdXMgc2F5IG9y
IHRoaW5rIGhlcmUgd2lsbCBjaGFuZ2UgdGhhdC4gIFRoZSByZWFsaXR5IGlzIHRoYXQgYSBs
YXJnZSBudW1iZXIgb2YgY29udHJpYnV0b3JzIHRvIHRoZSBlMmUgc3BlY3MgYXJlIGhlcmUg
YmVjYXVzZSB0aGUgZm9sa3MgdGhhdCBwYXkgdGhlbSB0byBjb250cmlidXRlIHRvIHRoZSBJ
RVRGIG1ha2UgbW9uZXkgc2VsbGluZyB0aG9zZSBzeXN0ZW1zLiAgSSB3aWxsIGFsc28gbm90
ZSB0aGF0IGZvbGtzIHRoYXQgYXJlbid0IGludGVyZXN0ZWQgaW4gSU1TIGNhbiB1c2UgeW91
ciBkb2N1bWVudCB0aGF0IHdhcyBwdWJsaXNoZWQgYXMgYSBiYXNlbGluZSAiZTJlIiBTSVAg
aW1wbGVtZW50YXRpb246IGh0dHA6Ly90b29scy5pZXRmLm9yZy9odG1sL3JmYzU2MzgNCg0K
VGhlcmUgaXMgYWxzbyBhIHNwcmVhZHNoZWV0IG9mIElFVEYgZGVwZW5kZW5jaWVzIHRoYXQg
aXMgcmVndWxhcmx5IHVwZGF0ZWQgYW5kIGl0IGluY2x1ZGVzIHJlZmVyZW5jZXMgdG8gdGhl
IHNwZWNzIHRoYXQgZGVwZW5kIHVwb24gY29yZSBJRVRGIFJGQ3M6DQpodHRwOi8vd3d3LjNn
cHAub3JnLy1JRVRGLURlcGVuZGVuY2llcy1hbmQtUHJpb3JpdGllcy0NCg0KSSB3aWxsIG5v
dGUgdGhhdCBzdW1tYXJpZXMgb2YgaXNzdWVzIGFuZCBhY3Rpb25zIHJlbGF0ZWQgdG8gdGhl
IFJBSSBhcmVhIGRvY3VtZW50cyBoYXMgYmVlbiBwb3N0ZWQgb24gdGhlIFJBSSBhcmVhIG1h
aWxpbmcgbGlzdCBpbiB0aGUgcGFzdDoNCmh0dHA6Ly93d3cuaWV0Zi5vcmcvbWFpbC1hcmNo
aXZlL3dlYi9yYWkvY3VycmVudC9tc2cwMDk1NS5odG1sDQpTbywgSSB0aGluayB0aGVyZSBo
YXMgYmVlbiBhIHJlYXNvbmFibGUgYW1vdW50IG9mIHRyYW5zcGFyZW5jeSB3aXRoIHJlZ2Fy
ZHMgdG8gdGhpcyB3b3JraW5nIHJlbGF0aW9uc2hpcC4gU2luY2UgdGhpcyByZWxhdGlvbnNo
aXAgYmV0d2VlbiAzR1BQIGFuZCBJRVRGIGhhcyBiZWVuIGluIHBsYWNlLCB0aGVyZSBoYXZl
IGJlZW4gYSBudW1iZXIgb2YgZG9jdW1lbnRzIHRoYXQgaGF2ZSBiZWVuIGFwcHJvdmVkIGFu
ZCBpbiBnZW5lcmFsIGl0IGhhcyBiZWVuIGEgZnJ1aXRmdWwgcmVsYXRpb25zaGlwIG9uIGJv
dGggcGFydHMuICBJbiBzb21lIGNhc2VzLCBjb3JlIGNoYW5nZXMgd2VyZSBtYWRlIHRvIFNJ
UCB0aGF0IGJlbmVmaXRlZCBib3RoIHRoZSBlMmUgbW9kZWwgYW5kIHRoZSB3YWxsZWQgZ2Fy
ZGVuLiAgDQoNClJlZ2FyZHMsDQpNYXJ5LiANCg0KDQpPbiBGcmksIE1hciAxMSwgMjAxMSBh
dCAxMToyOCBBTSwgSGVucnkgU2lubnJlaWNoIDxoZW5yeS5zaW5ucmVpY2hAZ21haWwuY29t
PiB3cm90ZToNCg0KDQoJSGkgQW5kcmV3LCBDaHJpc3RlciwNCgkNCg0KCT4+IE15IHVuZGVy
c3RhbmRpbmcgYWxzbyBpcyB0aGF0IElFVEYgbmVhcmx5IGEgZGVjYWRlIGFnbyBhZ3JlZWQg
dG8NCgk+PiBjb29wZXJhdGUgd2l0aCAzR1BQIChhbmQgb3RoZXIgc3RhbmRhcmRzIGJvZGll
cykgaW4gZGV2ZWxvcGluZyBhbmQNCgk+PiBlbmhhbmNpbmcgaW50ZXJuZXQgcHJvdG9jb2xz
DQoJDQoJDQoJM0dQUCBldGMuIGRvY3VtZW50cyBhcmUgYmVzdCBzZXJ2ZWQgYnkgaGF2aW5n
IHRoZW0gb25saW5lLCB1cC10by1kYXRlLCBvbg0KCXRoZWlyIHJlc3BlY3RpdmUgd2ViIHNp
dGVzLCB3aGVyZSB0aGV5IGNhbiBiZSBzZWFyY2hlZCBhbmQgc3R1ZGllZC4NCglVc2luZyBJ
RVRGIFJGQ3MgZm9yIHRoaXMgc2VlbXMgcmF0aGVyIG9kZCwgYW5kIGNlcnRhaW5seSBsZXNz
IGVmZmVjdGl2ZS4NCgkNCg0KCT5zb21lIHBlb3BsZSBpbiBJRVRGIHN1ZGRlbnRseSB0aGlu
a3MgdGhlcmUgYXJlIG1vcmUgImZ1biBhbmQgc2V4eSIgc3R1ZmYgdG8NCgk+d29yayBvbiB0
aGFuIFNJUC4uLg0KCQ0KCQ0KCUFzIG1lbnRpb25lZCwgYWZ0ZXIgMTAwJ3Mgb2YgU0lQIFJG
Q3MgYW5kIEktRHMgb3ZlciAxMisgeWVhcnMsIGFsbW9zdCBhbGwNCglTSVAgYXBwcyBhcmUg
dGVsZXBob255IGZlYXR1cmVzLCB3aGlsZSAxMDAsMDAwcyBvZiBhcHBzIGFyZSBhbGwgdXNp
bmcgSFRNTCwNCglzb21lIG9mIHRoZW0gdG8gY29tbXVuaWNhdGUgaW4gbmV3IHdheXMgLSBx
dWl0ZSBkaWZmZXJlbnQgZnJvbSB0aGUgU0lQDQoJdGVsZXBob25lIGNhbGwgc2lnbmFsaW5n
IG1vZGVscy4gUlRDLVdlYiBtYXkgb3IgbWF5IG5vdCBiZSBiYXNlZCBvbiBTSVANCglzaWdu
YWxpbmcgX2Fsb25lXyAsIGlmIGF0IGFsbC4gQ29tbXVuaWNhdGlvbiBtb2RlbHMgYW5kIHVz
ZXIgaGFiaXRzIGhhdmUNCgljaGFuZ2VkIHZlcnkgbXVjaCAoYW5kIGFyZSBzdGlsbCBjaGFu
Z2luZykgc2luY2UgU0lQIGhhcyBiZWVuIGRlZmluZWQuDQoJDQoJRG9uJ3QgeW91IHRoaW5r
IHdlIGhhdmUgYSBwcm9ibGVtIGhlcmU/DQoJSWYgeWVzLCB3b3VsZG4ndCBsaXN0ZW5pbmcg
dG8gdGhlIElFVEYgdGVjaG5pY2FsIHBsZW5hcnkgcHJlc2VudGF0aW9uIG9uIHRoZQ0KCWZ1
dHVyZSBvZiBhcHBsaWNhdGlvbnMgYmUgYSBnb29kIHN0YXJ0IHRvIGFkZHJlc3MgdGhlIGNo
YWxsZW5nZSBmb3IgU0lQDQoJYXBwbGljYXRpb25zPw0KCQ0KCVdoYXQgZG8geW91IHRoaW5r
Pw0KCQ0KCUhlbnJ5DQoJDQoJDQoJT24gMy8xMS8xMSAxMjo1NyBBTSwgIkNocmlzdGVyIEhv
bG1iZXJnIiA8Y2hyaXN0ZXIuaG9sbWJlcmdAZXJpY3Nzb24uY29tPg0KCXdyb3RlOg0KCQ0K
DQoJPg0KCT4gSGksDQoJPg0KCT4gSSBhZ3JlZSB3aXRoIEFuZHJldy4NCgk+DQoJPiBQZW9w
bGUgaGF2ZSBpbnZlc3RlZCwgYW5kIGFyZSBpbnZlc3RpbmcsIGxvdHMgb2YgcmVzb3VyY2Vz
IGludG8gU0lQIGJhc2VkDQoJPiBzeXN0ZW1zLiBUaGV5IGFyZSBub3Qgc3VkZGVudGx5IGdv
aW5nIHRvIHRocm93IGV2ZXJ5dGhpbmcgYXdheSBqdXN0IGJlY2F1c2UNCgk+IHNvbWUgcGVv
cGxlIGluIElFVEYgc3VkZGVudGx5IHRoaW5rcyB0aGVyZSBhcmUgbW9yZSAiZnVuIGFuZCBz
ZXh5IiBzdHVmZiB0bw0KCT4gd29yayBvbiB0aGFuIFNJUC4uLg0KCT4NCgk+IEFsc28sIHJl
Z2FyZGluZyB0aGUgYWdyZWVtZW50IGJldHdlZW4gSUVURiBhbmQgM0dQUCwgaWYgcGVvcGxl
IGhhdmUgaXNzdWVzDQoJPiB3aXRoIHRoYXQgYWdyZWVtZW50IEkgZG9uJ3QgdGhpbmsgdGhh
dCBwZW9wbGUgd2hvIGFyZSB3b3JraW5nIGJhc2VkIG9uIHRoYXQNCgk+IGFncmVlbWVudCBz
aG91bGQgaGF2ZSB0byBzdWZmZXIuDQoJPg0KCT4gSW5zdGVhZCwgcGVvcGxlIGhhdmluZyBp
c3N1ZXMgd2l0aCB0aGUgYWdyZWVtZW50IHNob3VsZCBicmluZyB0aG9zZSBpc3N1ZXMgdG8N
Cgk+IGEgbW9yZSBhcHByb3ByaWF0ZSBpbnN0YW5jZS4NCgk+DQoJPiBSZWdhcmRzLA0KCT4N
Cgk+IENocmlzdGVyDQoJPg0KCT4NCgk+PiAtLS0tLU9yaWdpbmFsIE1lc3NhZ2UtLS0tLQ0K
CT4+IEZyb206IGRpc3BhdGNoLWJvdW5jZXNAaWV0Zi5vcmcNCgk+PiBbbWFpbHRvOmRpc3Bh
dGNoLWJvdW5jZXNAaWV0Zi5vcmddIE9uIEJlaGFsZiBPZiBBbmRyZXcgQWxsZW4NCgk+PiBT
ZW50OiAxMS4gbWFhbGlza3V1dGEgMjAxMSAwOjA0DQoJPj4gVG86IEhlbnJ5IFNpbm5yZWlj
aDsgdGFuYWthaUBudHRkb2NvbW8uY28uanA7IGRpc3BhdGNoQGlldGYub3JnDQoJPj4gU3Vi
amVjdDogUmU6IFtkaXNwYXRjaF0NCgk+PiBSZXZpc2lvbm9mZHJhZnQtYWxsZW4tZGlzcGF0
Y2gtaW1laS11cm4tYXMtaW5zdGFuY2VpZA0KCT4+DQoJPj4gSGkgSGVucnksDQoJPj4NCgk+
PiBJIHRoaW5rIHRoZW4gdGhhdCBpdCdzIGEgZmFpciBzdW1tYXJ5IG9mIHlvdXIgZW1haWwg
dG8gc2F5DQoJPj4gdGhhdCB5b3VyIG9iamVjdGlvbnMgYXJlIG9mIGEgcG9saXRpY2FsIGFu
ZCBwaGlsb3NvcGhpY2FsDQoJPj4gbmF0dXJlIHJlbGF0ZWQgdG8geW91ciB2aWV3IG9mIHdo
YXQgY29uc3RpdHV0ZXMgdGhlIGludGVybmV0DQoJPj4gYW5kIHdoZXRoZXIgdGhlIElFVEYg
c2hvdWxkIGJlIGludm9sdmVkIGluIGVuaGFuY2luZyBJRVRGDQoJPj4gcHJvdG9jb2xzIGZv
ciBzdGFrZWhvbGRlcnMgdGhhdCB1dGlsaXplIHRob3NlIElFVEYgcHJvdG9jb2xzDQoJPj4g
Zm9yIGRlcGxveW1lbnRzIHRoYXQgc29tZSBpbiB0aGUgY29tbXVuaXR5IG1pZ2h0IG5vdCBj
b25zaWRlcg0KCT4+ICJ0aGUgSW50ZXJuZXQiIG9yICJ0aGUgcHVyZSBJbnRlcm5ldCIgb3Ig
dGhlICJvcGVuIGUyZQ0KCT4+IEludGVybmV0IiByYXRoZXIgdGhhbiB0ZWNobmljYWwgaXNz
dWVzIHdpdGggdGhlIGNvbnRlbnRzIG9mDQoJPj4gdGhlIGRyYWZ0cy4NCgk+Pg0KCT4+IFdo
aWxlIEkgZG9uJ3Qgd2FudCB0byBzdGFydCBhIGdlbmVyYWwgZGViYXRlIG9uIHdoYXQgaXMg
InRoZSBJbnRlcm5ldCINCgk+PiBoZXJlIEkgd2lsbCBzdGF0ZSB0aGF0IGluIG15IHZpZXcg
aXQncyB2ZXJ5IGRpZmZpY3VsdCB0bw0KCT4+IGV4Y2x1ZGUgYW55IHBhcnRpY3VsYXIgc3Rh
a2Vob2xkZXJzIGFuZCB0aGVpciBkZXBsb3ltZW50cyBvZg0KCT4+IElFVEYgcHJvdG9jb2xz
IGZyb20gYmVpbmcgcGFydCBvZiB0aGUgSW50ZXJuZXQgd2hlbiB0aG9zZQ0KCT4+IHN0YWtl
aG9sZGVycyBhcmUgaW4gbWFueSBjYXNlcyB0aGUgdmVyeSBzYW1lIGVzdGFibGlzaGVkDQoJ
Pj4gdGVsZWNvbW11bmljYXRpb25zIGNvbXBhbmllcyBvciBJU1BzIHRoYXQgdGhlIHZhc3Qg
bWFqb3JpdHkNCgk+PiBvZiB0aGUgcG9wdWxhdGlvbiBpbiBtb3N0IGNvdW50cmllcyB1c2Ug
dG8gb2J0YWluIGFjY2VzcyB0bw0KCT4+IHRoZSBJbnRlcm5ldC4NCgk+Pg0KCT4+IE15IHVu
ZGVyc3RhbmRpbmcgYWxzbyBpcyB0aGF0IElFVEYgbmVhcmx5IGEgZGVjYWRlIGFnbyBhZ3Jl
ZWQNCgk+PiB0byBjb29wZXJhdGUgd2l0aCAzR1BQIChhbmQgb3RoZXIgc3RhbmRhcmRzIGJv
ZGllcykgaW4NCgk+PiBkZXZlbG9waW5nIGFuZCBlbmhhbmNpbmcgaW50ZXJuZXQgcHJvdG9j
b2xzIGFuZCB0aGF0IElFVEYNCgk+PiBzcGVjaWZpY2FsbHkgcmVxdWVzdGVkIChtb3JlIGxp
a2UgaW5zaXN0ZWQpIHRoYXQgYW55DQoJPj4gZXh0ZW5zaW9ucyBvciBlbmhhbmNlbWVudHMg
dG8gSUVURiBwcm90b2NvbHMgcmVxdWlyZWQgYnkgM0dQUA0KCT4+IGJlIGRvbmUgd2l0aGlu
IElFVEYgYWNjb3JkaW5nIHRvIHRoZSBwcm9jZWR1cmVzIGRlZmluZWQgYnkNCgk+PiBJRVRG
IGZvciBlbmhhbmNpbmcgSUVURiBwcm90b2NvbHMgYW5kIGZvciByZWdpc3RlcmluZyBuYW1l
c3BhY2VzIGV0Yy4NCgk+PiBUaGVzZSBkcmFmdHMgc2ltcGx5IGNvbXBseSB3aXRoIHRoYXQg
YWdyZWVkIHdheSBvZiB3b3JraW5nDQoJPj4gYmV0d2VlbiBJRVRGIGFuZCAzR1BQLiBUbyBt
eSBrbm93bGVkZ2UgdGhpcyB3b3JraW5nIGFncmVlbWVudA0KCT4+IGlzIHN0aWxsIGluIGVm
ZmVjdC4gSSBkb24ndCB0aGluayBpdCBpcyB0aGUgcm9sZSBvZiB0aGUNCgk+PiBkaXNwYXRj
aCBXRyB0byBkZXRlcm1pbmUgd2hpY2ggYXBwbGljYXRpb25zIG9yIGRlcGxveW1lbnRzIG9m
DQoJPj4gU0lQIHRvIHN1cHBvcnQgb3IgY29vcGVyYXRlIHdpdGggY291bnRlciB0byBhZ3Jl
ZW1lbnRzDQoJPj4gcHJldmlvdXNseSBtYWRlIGF0IElFU0cgbGV2ZWwuDQoJPj4NCgk+PiBS
ZWZ1c2luZyBjb29wZXJhdGlvbiBiYXNlZCBvbiB3aGF0IGlzIGNvbnNpZGVyZWQgbm90IHRv
IGJlDQoJPj4gcGFydCBvZiB0aGUgIm9wZW4gZTJlIEludGVybmV0IiBpcyBvbmx5IGxpa2Vs
eSB0byBzZXQgdXAgdHdvDQoJPj4gKG9yIG1vcmUpIGNvbXBldGluZyB2ZXJzaW9ucyBvZiB0
aGUgSW50ZXJuZXQgcmVzdWx0aW5nIGluIHRoZQ0KCT4+IDIxc3QgY2VudHVyeSBlcXVpdmFs
ZW50IG9mIHRoZSBzdGFuZGFyZCByYWlsIGdhdWdlIHZzIGJyb2FkDQoJPj4gcmFpbCBnYXVn
ZSBjb21tdW5pY2F0aW9ucyBpbnRlcm9wZXJhYmlsaXR5IGlzc3VlIGZyb20gdGhlDQoJPj4g
MTl0aCBjZW50dXJ5IC0gZWZmZWN0aXZlbHkga2lsbGluZyB0aGUgZ29hbCBvZiBhbiBvcGVu
IGUyZQ0KCT4+IGludGVybmV0ISBBcyBEYWxlIGluZGljYXRlZCB3aGF0IGlzIGNvbnNpZGVy
ZWQgYSB3YWxsZWQNCgk+PiBnYXJkZW4gdG9kYXkgbWF5IG5vdCBiZSB3YWxsZWQgaW4gdGhl
IGZ1dHVyZSAoaWYgSSByZW1lbWJlcg0KCT4+IHNvbWUgdmVyeSBzaWduaWZpY2FudCB3YWxs
cyBjYW1lIHR1bWJsaW5nIGRvd24gZHVyaW5nIGp1c3QNCgk+PiB0aGUgbGFzdCBkZWNhZGUg
b2YgdGhlIDIwdGggY2VudHVyeSBhbG9uZSkuDQoJPj4NCgk+PiBEaWZmZXJlbnQgYnVzaW5l
c3MgbW9kZWxzIHdpbGwgYmUgcmVzb2x2ZWQgdGhyb3VnaCB0aGUgbWFya2V0DQoJPj4gbm90
IGJ5IGNyZWF0aW5nIGNvbXBldGluZyBpbmNvbXBhdGlibGUgdGVjaG5pY2FsIHNvbHV0aW9u
cy4NCgk+Pg0KCT4+IEluIHRlcm1zIG9mIHRoZSB0aW1lIHVzZWQgb24gZGlzY3Vzc2lvbiB3
aXRoaW4gZGlzcGF0Y2ggSQ0KCT4+IGRvbid0IHRoaW5rIHRoZXJlIGlzIGEgbmVlZCBmb3Ig
bXVjaCBtb3JlIGRpc2N1c3Npb24gaW4NCgk+PiBkaXNwYXRjaC4gVGhlc2UgZHJhZnRzIGNv
dWxkIGdvIGFoZWFkIGFzIEFEIHNwb25zb3JlZCB3aXRob3V0DQoJPj4gZnVydGhlciBXRyBp
bnZvbHZlbWVudC4gQXMgSSB1bmRlcnN0YW5kIGl0IHRoZSBJTUVJIFVSTg0KCT4+IGRlZmlu
aXRpb24gZHJhZnQgZGlkIG5vdCBhcyBwYXJ0IG9mIHRoZSBVUk4gcmVnaXN0cmF0aW9uDQoJ
Pj4gcHJvY2VzcyBuZWVkIHRvIGV2ZW4gYmUgZGlzY3Vzc2VkIGluIGRpc3BhdGNoIGF0IGFs
bA0KCT4+IChhbHRob3VnaCBEYWxlIFdvcmxleSdzIHJldmlldyB3YXMgaW52YWx1YWJsZSBp
biBpbXByb3ZpbmcNCgk+PiB0aGUgZHJhZnQgdGVjaG5pY2FsbHkpIGFzIHRoZSBOSURzIGxp
c3Qgd2FzIHRoZSBmb3JhIGZvciB0aGlzDQoJPj4gZGlzY3Vzc2lvbiBhbmQgd2hlcmUgcHJl
dmlvdXMgdmVyc2lvbnMgb2YgdGhhdCBkcmFmdCB3ZXJlIGRpc2N1c3NlZC4NCgk+Pg0KCT4+
IEl0IHNob3VsZCBiZSBub3RlZCBhbGwgd2UgYXJlIHRhbGtpbmcgYWJvdXQgd2l0aCByZWdh
cmQgdG8NCgk+PiB0aGVzZSBkcmFmdHMgaXMgdGhlIHJlZ2lzdHJhdGlvbiBvZiBhIFVSTiBu
YW1lc3BhY2UgYW5kDQoJPj4gZGVmaW5pdGlvbiBvZiBpdHMgdXNhZ2UgYXMgYW4gSW5zdGFu
Y2UgSUQgYXMgcmVxdWlyZWQgYnkgUkZDDQoJPj4gNTYyNi4gSWYgd2UgY2FuJ3QgYWdyZWUg
dGhhdCB0aGVuIHdoYXQgZnV0dXJlIGV4dGVuc2lvbnMgY2FuIHdlIGFncmVlPw0KCT4+DQoJ
Pj4gQW5kcmV3DQoJPj4NCgk+PiAtLS0tLU9yaWdpbmFsIE1lc3NhZ2UtLS0tLQ0KCT4+IEZy
b206IEhlbnJ5IFNpbm5yZWljaCBbbWFpbHRvOmhlbnJ5LnNpbm5yZWljaEBnbWFpbC5jb21d
DQoJPj4gU2VudDogVGh1cnNkYXksIE1hcmNoIDEwLCAyMDExIDExOjAzIEFNDQoJPj4gVG86
IEFuZHJldyBBbGxlbjsgdGFuYWthaUBudHRkb2NvbW8uY28uanA7IGRpc3BhdGNoQGlldGYu
b3JnDQoJPj4gU3ViamVjdDogUmU6IFtkaXNwYXRjaF0NCgk+PiBSZXZpc2lvbm9mZHJhZnQt
YWxsZW4tZGlzcGF0Y2gtaW1laS11cm4tYXMtaW5zdGFuY2VpZA0KCT4+DQoJPj4gSGkgQW5k
cmV3LA0KCT4+DQoJPj4gSSBoYXZlIGdlbmVyaWMgb2JqZWN0aW9ucyB0byBzZWUgdGhlIHB1
YmxpY2F0aW9uIG9mIHlldA0KCT4+IGFub3RoZXIgaW5mb3JtYXRpb25hbCBSRkMgYWJvdXQg
cHJhY3RpY2VzIGluIGNvbnNvcnRpdW1zIG9mDQoJPj4gY2xvc2VkIG5ldHdvcmtzIGluIHRo
ZSBJRVRGLCBmb3IgdGhlIGZvbGxvd2luZyByZWFzb25zOg0KCT4+DQoJPj4gKiBVc2luZyB0
aGUgUkZDIHByb2Nlc3MgYXMgYSBwdWJsaXNoaW5nIHZlbnVlIGZvcg0KCT4+IG9yZ2FuaXph
dGlvbnMgdGhhdCBkb24ndCBleHBsaWNpdGx5IGRvIHRoYXQgaW4gc3VwcG9ydCBvZiB0aGUN
Cgk+PiBvcGVuIGUyZSBJbnRlcm5ldC4NCgk+Pg0KCT4+ICogVGhlIHJlc291cmNlcyBvZiB0
aGUgSUVURiBhcmUgbm90IHVubGltaXRlZCwgbWVldGluZyB0aW1lDQoJPj4gYW5kIGVtYWls
IHRpbWUgdG8gZGlzY3VzcyBpcyBzY2FyY2UgYW5kIHRoZSBJRVRGIHJlZ2lzdHJpZXMNCgk+
PiBhcmUganVzdCBub3Qgc2NhbGFibGUgZW5vdWdoIHRvIHN1cHBvcnQgcmVnaXN0cmF0aW9u
cyBmb3IgYWxsDQoJPj4gdGhlIGV4dGVuc2lvbnMgdGhhdCBiZW5lZml0IGNvdW50bGVzcyBl
eHRlcm5hbCBvcmdhbml6YXRpb25zDQoJPj4gYW5kIGNvbXBhbmllcyB0aGF0IGRlcGxveSBJ
bnRlcm5ldCB0ZWNobm9sb2d5Lg0KCT4+IE5ldyBidXNpbmVzcyBwbGFucyBhbmQgdGhlIHJl
cXVpcmVkIGV4dGVuc2lvbnMgcG9wIHVwIGluIGFuDQoJPj4gZW5kbGVzcyBmYXNoaW9uLg0K
CT4+IEJpZyBvcmdhbml6YXRpb25zIHN1Y2ggYXMgeW91IG1lbnRpb24gaGF2ZSB0aGVpciBv
d24NCgk+PiByZXNvdXJjZXMgdG8gcHVibGlzaCB0aGVpciBpbnRlcm5hbCB0ZWNobmljYWwg
ZG9jdW1lbnRzIGFuZA0KCT4+IHdoaXRlIHBhcGVycyBpZiB0aGV5IGRlc2lyZSB0byBkbyBz
by4NCgk+Pg0KCT4+IE1lbWJlcnMgb2Ygb3V0c2lkZSBvcmdhbml6YXRpb25zIGFyZSBjZXJ0
YWlubHkgZG9pbmcNCgk+PiBldmVyeWJvZHkgYSBnb29kIHNlcnZpY2UgYnkgaW5mb3JtaW5n
IHZpYSBlbWFpbCBvZiBuZXcNCgk+PiBzdGFuZGFyZHMvZXh0ZW5zaW9ucyBpbiBvdGhlciBv
cmdhbml6YXRpb24sIHN1Y2ggYXMgeW91IGhhdmUNCgk+PiBkb25lIGJlbG93LCBidXQgc3Vj
aCBpbmZvcm1hdGlvbiBuZWVkIG5vdCBvdmVybG9hZCB0aGUNCgk+PiBJbnRlcm5ldCBSRkMg
cHJvY2VzcyBub3Igb3ZlcmxvYWQgdGhlIEludGVybmV0IHJlZ2lzdHJpZXMuDQoJPj4NCgk+
PiAqIEhhcmQgZm9yIFNJUCBkZXZlbG9wZXJzIGFuZCByaWdodCBvdXQgaW1wb3NzaWJsZSBm
b3INCgk+PiBhcHBsaWNhdGlvbiBkZXZlbG9wZXJzLg0KCT4+IFRob3VnaCBTSVAgd29ya3Mg
ZmluZSBpbmRlZWQgYWNyb3NzIHRoZSB3b3JsZDsgU0lQaXQgYW5kIHRoZQ0KCT4+IFNJUCBG
b3J1bSB0YWtlIGNhcmUgb2YgdW5hdm9pZGFibGUgcHJvYmxlbXMgc3RpbGwgZW5jb3VudGVy
ZWQNCgk+PiBieSBvcGVyYXRvcnMsIGFyZ3VhYmx5IGluY3JlYXNpbmcgU0lQIGNvbXBsZXhp
dHkgZm9yDQoJPj4gbm9uLUludGVybmV0IHB1cnBvc2VzIGlzIG5vdCBhIGdvb2QgaWRlYSwg
ZXZlbiBpZiBuZXcgU0lQDQoJPj4gZXh0ZW5zaW9ucyBhcmUgbGFiZWxlZCBhcyBqdXN0ICJp
bmZvcm1hdGlvbmFsIi4gSXQgbWFrZXMgaXQNCgk+PiBldmVuIG1vcmUgZGlmZmljdWx0IGZv
ciBuZXcgZGV2ZWxvcGVycyBlbnRlcmluZyB0aGUgZmllbGQgdG8NCgk+PiBldmVyIHVuZGVy
c3RhbmQsIGxldCBhbG9uZSB3cml0ZSBjb2RlIGZvciBzb21lIG9yIGFsbCB0aGF0DQoJPj4g
aGFzIGJlZW4gcHVibGlzaGVkIGFscmVhZHkgYWJvdXQgU0lQIGZvciBleGFtcGxlLiBDaGVj
a2luZw0KCT4+IHJpZ2h0IG5vdyB0aGVyZSBhcmUgMTk3IFJGQ3MgbWF0Y2hpbmcgU0lQIGFu
ZCA3OCBJLURzIGZvdW5kDQoJPj4gYnkgdGhlIFJGQyBzZWFyY2ggZW5naW5lLg0KCT4+DQoJ
Pj4gSXQgaXMgdGltZSB0byBsZWF2ZSBTSVAgYWxvbmUgYW5kIG1vdmUgb24gdG8gb3RoZXIg
Y3JpdGljYWwNCgk+PiBhcmVhcyBmb3IgdGhlIEludGVybmV0LCBzdWNoIGFzIGZvciBleGFt
cGxlIHRoZSBmdXR1cmUgb2YNCgk+PiBhcHBsaWNhdGlvbnMgdGhhdCB3aWxsIGJlIGRpc2N1
c3NlZCBpbiB0aGUgbmV4dCBJRVRGDQoJPj4gdGVjaG5pY2FsIHBsZW5hcnkgb24gTW9uZGF5
LCBNYXJjaCAyOC4NCgk+Pg0KCT4+IFdpdGggdGhpcyBpbiBtaW5kLCBJIHdvdWxkIHJhdGhl
ciBub3QgZ28gaW50byBkaXNjdXNzaW5nIHlvdXINCgk+PiBJLUQgaW4gZGV0YWlsLCB0aG91
Z2ggYXQgZmlyc3QgZ2xhbmNlIGl0IHNlZW1zIHRlY2huaWNhbGx5DQoJPj4gY29ycmVjdCBh
bmQgaWYgaXQgd2VyZSBkaXNjdXNzZWQgb24gdGhlIGxpc3QsIG1heSBpbmRlZWQNCgk+PiB0
ZWNobmljYWxseSBxdWFsaWZ5IGFzIGFuIGluZm9ybWF0aW9uYWwgUkZDIC0ganVzdCB3aGF0
IEkgYW0NCgk+PiBhcmd1aW5nIGFnYWluc3QuDQoJPj4NCgk+PiBUaGVzZSBhcmUgYWxsIHBl
cnNvbmFsIG9waW5pb25zLiBUaGUgY2hhaXJzIGFuZCBBRHMgbWF5IG9yDQoJPj4gbWF5IG5v
dCBkaXNjdXNzIG9yIGFncmVlL2Rpc2FncmVlIHdpdGguIFlvdXIgSS1EIGlzIGp1c3QgYQ0K
CT4+IGdvb2Qgb3Bwb3J0dW5pdHkgdG8gcmFpc2UgdGhlc2UgaXNzdWVzLg0KCT4+DQoJPj4g
Tm90ZSB0aGF0IEkgYW0gbm90IHN1Z2dlc3RpbmcgYW55IHZvdGluZyBvbiB5b3VyIEktRCwg
aWYgaXQNCgk+PiB3aWxsIGJlIGRpc2N1c3NlZCwgc2luY2UgZm9yIHZhcmlvdXMgcmVhc29u
cywgdGhlIG1ham9yaXR5IG9mDQoJPj4gZm9sa3MgYXR0ZW5kaW5nIHRoZSBESVNQQVRDSCBX
RyBhcmUgZnJvbSB2YXJpb3VzIG91dHNpZGUNCgk+PiBvcmdhbml6YXRpb25zIHdpdGggYnVz
aW5lc3MgcGxhbnMgdGhhdCBtYXkgb3IgbWF5IG5vdCBmdWxseQ0KCT4+IGVtYnJhY2UgdGhl
IG9wZW4gZTJlIEludGVybmV0OyBhIHBlcmZlY3QgbGVnaXRpbWF0ZSBjaG9pY2UNCgk+PiBm
b3IgdGhlbSwgYnV0IG5vdCAgaW4gdGhlIElFVEYgKEkganVzdCBjYW4ndCByZW1lbWJlciB3
aGVuDQoJPj4gZXZlciB0aGVyZSB3YXMgYSB2b3RlIGFnYWluc3QgYW55IG5ldyBJLUQgZm9y
IFNJUCwgc2luY2Ugbm8NCgk+PiBjb3N0IHdhcyBwZXJjZWl2ZWQgYnkgdGhvc2Ugdm90aW5n
KS4NCgk+Pg0KCT4+IFRoYW5rcywNCgk+PiBIZW5yeQ0KCT4+DQoJPj4NCgk+PiBPbiAzLzkv
MTEgNTo0NSBQTSwgIkFuZHJldyBBbGxlbiIgPGFhbGxlbkByaW0uY29tPiB3cm90ZToNCgk+
Pg0KCT4+Pg0KCT4+PiBIZW5yeQ0KCT4+Pg0KCT4+PiBXb3VsZCB5b3UgY2FyZSB0byBlbGFi
b3JhdGUgYW5kIGV4cGxhaW4geW91ciBzcGVjaWZpYyBjb25jZXJucz8gIEhvdw0KCT4+IGlz
IHdoYXQNCgk+Pj4gaXMgZGVmaW5lZCBpbiB0aGVzZSBkcmFmdHMgcG90ZW50aWFsbHkgaW5j
b21wYXRpYmxlIHdpdGgNCgk+PiB0aGUgb3BlbiwgZTJlDQoJPj4+IGFyY2hpdGVjdHVyYWwg
YW5kIGluIGFueSBvdGhlciB3YXkgdGVjaG5pY2FsIGNvbmZsaWN0IHdpdGggdGhlDQoJPj4g
SW50ZXJuZXQ/DQoJPj4+DQoJPj4+IFRoZSB0d28gZHJhZnRzIGRvIHRoZSBmb2xsb3dpbmcg
dGhpbmdzOg0KCT4+Pg0KCT4+PiBPbmUgZHJhZnQgc2ltcGx5IGRlZmluZXMgYW5kIHJlZ2lz
dGVycyB3aXRoIElBTkEgYSBVUk4gbmFtZXNwYWNlDQoJPj4gYWNjb3JkaW5nIHRvDQoJPj4+
IHRoZSBwcm9jZWR1cmVzIGRlZmluZWQgZm9yIGRvaW5nIHRoYXQuDQoJPj4+DQoJPj4+IFRo
ZSBvdGhlciBkcmFmdCBzaW1wbHkgZGVmaW5lcyBob3cgdGhpcyBVUk4gY2FuIGJlIHVzZWQg
YXMNCgk+PiBhbiBpbnN0YW5jZQ0KCT4+IElEIGFzDQoJPj4+IHJlcXVpcmVkIGJ5IFJGQyA1
NjI2Lg0KCT4+Pg0KCT4+PiBTSVAgYW5kIHRoZSBtZWNoYW5pc21zIGluIFJGQyA1NjI2IGFy
ZSBhbHJlYWR5IGJlaW5nIHVzZWQNCgk+PiBpbiBuZXR3b3Jrcw0KCT4+IGRlZmluZWQNCgk+
Pj4gYnkgM0dQUCBhbmQgYXJlIGJlaW5nIGRlcGxveWVkIGJ5IG1lbWJlcnMgb2YgdGhlIEdT
TUEuIDNHUFAgYW5kIEdTTUENCgk+PiBoYXZlDQoJPj4+IGFscmVhZHkgZGVmaW5lZCB0aGlz
IGJlaGF2aW9yIHVzaW5nIHRoZWlyIGV4aXN0aW5nIGRldmljZQ0KCT4+IElkZW50aWZpZXJz
DQoJPj4gZm9yDQoJPj4+IGluc3RhbmNlIElEIGdlbmVyYXRpb24gaW4gdGhlaXIgc3BlY2lm
aWNhdGlvbnMuIFRoZXNlDQoJPj4gZHJhZnRzIGNvbXBsZXRlDQoJPj4gdGhlDQoJPj4+IHBy
b2Nlc3MgZG9pbmcgdGhpbmdzIHRoZSByaWdodCB3YXkgYnkgcmVzZXJ2aW5nIHRoZSBuYW1l
c3BhY2UgYW5kDQoJPj4gZGVmaW5pbmcgdGhlDQoJPj4+IHVzYWdlIG9mIGl0IHVzaW5nIHRo
ZSBJRVRGIGRlZmluZWQgbWVjaGFuaXNtcyBmb3IgZG9pbmcgdGhhdC4NCgk+Pj4NCgk+Pj4g
QW5kcmV3DQoJPj4+DQoJPj4+IC0tLS0tIE9yaWdpbmFsIE1lc3NhZ2UgLS0tLS0NCgk+Pj4g
RnJvbTogSGVucnkgU2lubnJlaWNoIFttYWlsdG86aGVucnkuc2lubnJlaWNoQGdtYWlsLmNv
bV0NCgk+Pj4gU2VudDogV2VkbmVzZGF5LCBNYXJjaCAwOSwgMjAxMSAwMzoxMCBQTQ0KCT4+
PiBUbzogSXRzdW1hIFRBTkFLQSA8dGFuYWthaUBudHRkb2NvbW8uY28uanA+OyBkaXNwYXRj
aEBpZXRmLm9yZw0KCT4+PiA8ZGlzcGF0Y2hAaWV0Zi5vcmc+DQoJPj4+IFN1YmplY3Q6IFJl
OiBbZGlzcGF0Y2hdDQoJPj4gUmV2aXNpb25vZmRyYWZ0LWFsbGVuLWRpc3BhdGNoLWltZWkt
dXJuLWFzLWluc3RhbmNlaWQNCgk+Pj4NCgk+Pj4gU3VjaCB3b3JrIGluIHRoZSBJRVRGIHdv
dWxkIHJlcXVpcmUgSU1PIGFuIGV4cGxhbmF0aW9uIGFib3V0IGl0cw0KCT4+PiBjb21wYXRp
YmlsaXR5IHdpdGggdGhlIG9wZW4sIGUyZSBhcmNoaXRlY3R1cmFsIGFuZCBvdGhlciB0ZWNo
bmljYWwNCgk+Pj4gcHJpbmNpcGxlcyBvZiB0aGUgSW50ZXJuZXQuIEl0IGFjdHVhbGx5IHNl
ZW1zIHRvIGJlIHRoZSBvcHBvc2l0ZS4NCgk+Pj4NCgk+Pj4gRm9yIHRoaXMgcmVhc29uLCBz
dWNoIHdvcmsgbWF5IGJlIGJldHRlciBhY2NvbXBsaXNoZWQgaW4gb3RoZXINCgk+Pj4gb3Jn
YW5pemF0aW9ucywgc28gYXMgdG8gYXZvaWQgYXJjaGl0ZWN0dXJhbCBwcmluY2lwbGUgYW5k
IG90aGVyDQoJPj4gdGVjaG5pY2FsDQoJPj4+IGNvbmZsaWN0cy4gU2ltcGxpY2l0eSBhbmQg
c2NhbGFiaWxpdHkgY29tZXMgdG8gbWluZCBhbW9uZyBvdGhlciwNCgk+PiB0aG91Z2ggdGhp
cw0KCT4+PiBpcyBiZXR0ZXIgZGlzY3Vzc2VkIGVsc2V3aGVyZS4NCgk+Pj4NCgk+Pj4gVGhh
bmtzLA0KCT4+Pg0KCT4+PiBIZW5yeSBTaW5ucmVpY2gNCgk+Pj4NCgk+Pj4NCgk+Pj4gT24g
My85LzExIDg6NTQgQU0sICJJdHN1bWEgVEFOQUtBIiA8dGFuYWthaUBudHRkb2NvbW8uY28u
anA+IHdyb3RlOg0KCT4+Pg0KCT4+Pj4gRGVhciBBbGwsDQoJPj4+Pg0KCT4+Pj4gSSB3b3Jr
IGZvciBOVFQgRE9DT01PLCBJbmMgaW4gSmFwYW4uIGFuZCBhdHRlbmQgdmFyaW91cw0KCT4+
IDNHUFAgbWVldGluZ3MNCgk+PiAoU0EyLA0KCT4+Pj4gQ1QxLCBDVDQsIGFuZCBDVDMpIGFu
ZCBHU00gQXNzb2NpYXRpb24gKEdTTUEpIEludGVyd29ya2luZywgUm9hbWluZw0KCT4+IEV4
cGVydA0KCT4+Pj4gR3JvdXAgKElSRUcpLCBSb2FtaW5nIGluIExURSAoUklMVEUpIGFuZCBQ
YWNrZXQgc3ViZ3JvdXANCgk+PiBtZWV0aW5ncy4gIEkNCgk+PiBzZXJ2ZQ0KCT4+Pj4gYXMg
YSBWaWNlIENoYWlyIG9mIEdTTUEgSVJFRyBSSUxURSBhbmQgUGFja2V0IHN1Ymdyb3Vwcy4N
Cgk+Pj4+DQoJPj4+PiBJIHdvdWxkIGxpa2UgdG8gZXhwcmVzcyBteSBzdXBwb3J0IG9mDQoJ
Pj4gZHJhZnQtbW9udGVtdXJyby1nc21hLWltZWktdXJuDQoJPj4gYW5kIHRoZQ0KCT4+Pj4g
YXNzb2NpYXRlZCBkcmFmdC1hbGxlbi1kaXNwYXRjaC1pbWVpLXVybi1hcy1pbnN0YW5jZWlk
Lg0KCT4+Pj4NCgk+Pj4+IFRoZSByZWFzb25zIGFyZSBleGFjdGx5IHNhbWUgYXMgd2hhdCBS
aWNreSBLYXVyYSAoU2Ftc3VuZykNCgk+PiBleHByZXNzZWQuDQoJPj4gQQ0KCT4+Pj4gdmFy
aW91cyAzR1BQIHNwZWNpZmljYXRpb25zIChJTVMgUmVnaXN0cmF0aW9uIGluIDNHUFANCgk+
PiBUUzI0LjIyOSwgSU1TDQoJPj4+PiBlbWVyZ2VuY3kgY2FsbHMgaW4gVFMyMy4yMzcgYW5k
IFRTMjQuMjM3LCBJQ1MgaW4gVFMyMy4yOTIgYW5kDQoJPj4gVFMyNC4yOTIpLCBhbmQNCgk+
Pj4+IEdTTUEgZG9jdW1lbnQgKEdTTUEgUFJEIElSLjkyLCAiSU1TIFByb2ZpbGUgZm9yIFZv
aWNlIGFuZCBTTVMiKSBhcmUNCgk+PiB1c2luZw0KCT4+Pj4gdGhlc2UgZHJhZnRzLg0KCT4+
Pj4NCgk+Pj4+IFRoZXNlIGludGVybmV0IGRyYWZ0cyBhcmUgdmVyeSBpbXBvcnRhbnQgdG8g
c3VwcG9ydCB2YXJpb3VzIGtleQ0KCT4+IGZlYXR1cmVzIGluDQoJPj4+PiAzR1BQIGFuZCB0
byBlbmFibGUgR1NNQSBWb2ljZSBvdmVyIExURS4gIFRoZXJlZm9yZSwgSSBraW5kbHkgd291
bGQNCgk+PiBsaWtlIGFzaw0KCT4+Pj4gdGhhdCB0aGlzIHdvcmsgaXMgcHJvZ3Jlc3NlZCBh
cyBxdWlja2x5IGFzIHBvc3NpYmxlIGluIElFVEYuDQoJPj4+Pg0KCT4+Pj4NCgk+Pj4+IEJl
c3QgcmVnYXJkcywNCgk+Pj4+DQoJPj4+PiBJdHN1bWEgVGFuYWthDQoJPj4+PiBOVFQgRE9D
T01PDQoJPj4+Pg0KCT4+Pj4NCgk+Pj4+IF9fX19fX19fX19fX19fX19fX19fX19fX19fX19f
X19fX19fX19fX19fX19fX19fDQoJPj4+PiBkaXNwYXRjaCBtYWlsaW5nIGxpc3QNCgk+Pj4+
IGRpc3BhdGNoQGlldGYub3JnDQoJPj4+PiBodHRwczovL3d3dy5pZXRmLm9yZy9tYWlsbWFu
L2xpc3RpbmZvL2Rpc3BhdGNoDQoJPj4+DQoJPj4+DQoJPj4+IF9fX19fX19fX19fX19fX19f
X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fDQoJPj4+IGRpc3BhdGNoIG1haWxpbmcg
bGlzdA0KCT4+PiBkaXNwYXRjaEBpZXRmLm9yZw0KCT4+PiBodHRwczovL3d3dy5pZXRmLm9y
Zy9tYWlsbWFuL2xpc3RpbmZvL2Rpc3BhdGNoDQoJPj4+DQoJPj4+DQoJPj4gLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tDQoJPj4+IFRoaXMgdHJhbnNtaXNzaW9uIChpbmNsdWRpbmcgYW55IGF0dGFjaG1l
bnRzKSBtYXkgY29udGFpbg0KCT4+IGNvbmZpZGVudGlhbA0KCT4+PiBpbmZvcm1hdGlvbiwg
cHJpdmlsZWdlZCBtYXRlcmlhbCAoaW5jbHVkaW5nIG1hdGVyaWFsDQoJPj4gcHJvdGVjdGVk
IGJ5IHRoZQ0KCT4+PiBzb2xpY2l0b3ItY2xpZW50IG9yIG90aGVyIGFwcGxpY2FibGUgcHJp
dmlsZWdlcyksIG9yIGNvbnN0aXR1dGUNCgk+PiBub24tcHVibGljDQoJPj4+IGluZm9ybWF0
aW9uLiBBbnkgdXNlIG9mIHRoaXMgaW5mb3JtYXRpb24gYnkgYW55b25lIG90aGVyIHRoYW4g
dGhlDQoJPj4gaW50ZW5kZWQNCgk+Pj4gcmVjaXBpZW50IGlzIHByb2hpYml0ZWQuIElmIHlv
dSBoYXZlIHJlY2VpdmVkIHRoaXMgdHJhbnNtaXNzaW9uIGluDQoJPj4gZXJyb3IsDQoJPj4+
IHBsZWFzZSBpbW1lZGlhdGVseSByZXBseSB0byB0aGUgc2VuZGVyIGFuZCBkZWxldGUgdGhp
cyBpbmZvcm1hdGlvbg0KCT4+IGZyb20geW91cg0KCT4+PiBzeXN0ZW0uIFVzZSwgZGlzc2Vt
aW5hdGlvbiwgZGlzdHJpYnV0aW9uLCBvciByZXByb2R1Y3Rpb24gb2YgdGhpcw0KCT4+IHRy
YW5zbWlzc2lvbg0KCT4+PiBieSB1bmludGVuZGVkIHJlY2lwaWVudHMgaXMgbm90IGF1dGhv
cml6ZWQgYW5kIG1heSBiZSB1bmxhd2Z1bC4NCgk+Pg0KCT4+DQoJPj4NCgk+Pg0KCT4+DQoJ
Pj4NCgk+PiAtLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0NCgk+PiBUaGlzIHRyYW5zbWlzc2lvbiAoaW5jbHVk
aW5nIGFueSBhdHRhY2htZW50cykgbWF5IGNvbnRhaW4NCgk+PiBjb25maWRlbnRpYWwgaW5m
b3JtYXRpb24sIHByaXZpbGVnZWQgbWF0ZXJpYWwgKGluY2x1ZGluZw0KCT4+IG1hdGVyaWFs
IHByb3RlY3RlZCBieSB0aGUgc29saWNpdG9yLWNsaWVudCBvciBvdGhlcg0KCT4+IGFwcGxp
Y2FibGUgcHJpdmlsZWdlcyksIG9yIGNvbnN0aXR1dGUgbm9uLXB1YmxpYyBpbmZvcm1hdGlv
bi4NCgk+PiBBbnkgdXNlIG9mIHRoaXMgaW5mb3JtYXRpb24gYnkgYW55b25lIG90aGVyIHRo
YW4gdGhlIGludGVuZGVkDQoJPj4gcmVjaXBpZW50IGlzIHByb2hpYml0ZWQuIElmIHlvdSBo
YXZlIHJlY2VpdmVkIHRoaXMNCgk+PiB0cmFuc21pc3Npb24gaW4gZXJyb3IsIHBsZWFzZSBp
bW1lZGlhdGVseSByZXBseSB0byB0aGUgc2VuZGVyDQoJPj4gYW5kIGRlbGV0ZSB0aGlzIGlu
Zm9ybWF0aW9uIGZyb20geW91ciBzeXN0ZW0uIFVzZSwNCgk+PiBkaXNzZW1pbmF0aW9uLCBk
aXN0cmlidXRpb24sIG9yIHJlcHJvZHVjdGlvbiBvZiB0aGlzDQoJPj4gdHJhbnNtaXNzaW9u
IGJ5IHVuaW50ZW5kZWQgcmVjaXBpZW50cyBpcyBub3QgYXV0aG9yaXplZCBhbmQNCgk+PiBt
YXkgYmUgdW5sYXdmdWwuDQoJPj4gX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19f
X19fX19fX19fX19fX18NCgk+PiBkaXNwYXRjaCBtYWlsaW5nIGxpc3QNCgk+PiBkaXNwYXRj
aEBpZXRmLm9yZw0KCT4+IGh0dHBzOi8vd3d3LmlldGYub3JnL21haWxtYW4vbGlzdGluZm8v
ZGlzcGF0Y2gNCgk+Pg0KCQ0KCQ0KCV9fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19f
X19fX19fX19fX19fX19fDQoJZGlzcGF0Y2ggbWFpbGluZyBsaXN0DQoJZGlzcGF0Y2hAaWV0
Zi5vcmcNCglodHRwczovL3d3dy5pZXRmLm9yZy9tYWlsbWFuL2xpc3RpbmZvL2Rpc3BhdGNo
DQoJDQoNCg0KDQotLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0t
LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0NClRoaXMgdHJhbnNtaXNzaW9uIChpbmNsdWRp
bmcgYW55IGF0dGFjaG1lbnRzKSBtYXkgY29udGFpbiBjb25maWRlbnRpYWwgaW5mb3JtYXRp
b24sIHByaXZpbGVnZWQgbWF0ZXJpYWwgKGluY2x1ZGluZyBtYXRlcmlhbCBwcm90ZWN0ZWQg
YnkgdGhlIHNvbGljaXRvci1jbGllbnQgb3Igb3RoZXIgYXBwbGljYWJsZSBwcml2aWxlZ2Vz
KSwgb3IgY29uc3RpdHV0ZSBub24tcHVibGljIGluZm9ybWF0aW9uLiBBbnkgdXNlIG9mIHRo
aXMgaW5mb3JtYXRpb24gYnkgYW55b25lIG90aGVyIHRoYW4gdGhlIGludGVuZGVkIHJlY2lw
aWVudCBpcyBwcm9oaWJpdGVkLiBJZiB5b3UgaGF2ZSByZWNlaXZlZCB0aGlzIHRyYW5zbWlz
c2lvbiBpbiBlcnJvciwgcGxlYXNlIGltbWVkaWF0ZWx5IHJlcGx5IHRvIHRoZSBzZW5kZXIg
YW5kIGRlbGV0ZSB0aGlzIGluZm9ybWF0aW9uIGZyb20geW91ciBzeXN0ZW0uIFVzZSwgZGlz
c2VtaW5hdGlvbiwgZGlzdHJpYnV0aW9uLCBvciByZXByb2R1Y3Rpb24gb2YgdGhpcyB0cmFu
c21pc3Npb24gYnkgdW5pbnRlbmRlZCByZWNpcGllbnRzIGlzIG5vdCBhdXRob3JpemVkIGFu
ZCBtYXkgYmUgdW5sYXdmdWwuDQo=

------_=_NextPart_001_01CBE108.4BD61574
Content-Type: text/html;
	charset="UTF-8"
content-transfer-encoding: base64
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------_=_NextPart_001_01CBE108.4BD61574--

From henry.sinnreich@gmail.com  Sun Mar 13 09:30:40 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5FA1D3A69FA for ; Sun, 13 Mar 2011 09:30:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.443
X-Spam-Level: 
X-Spam-Status: No, score=-2.443 tagged_above=-999 required=5 tests=[AWL=-0.241, BAYES_00=-2.599, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id PeYV25J4wgub for ; Sun, 13 Mar 2011 09:30:37 -0700 (PDT)
Received: from mail-gx0-f172.google.com (mail-gx0-f172.google.com [209.85.161.172]) by core3.amsl.com (Postfix) with ESMTP id AB3033A69DB for ; Sun, 13 Mar 2011 09:30:36 -0700 (PDT)
Received: by gxk19 with SMTP id 19so975198gxk.31 for ; Sun, 13 Mar 2011 09:31:58 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:cc:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type; bh=bJkSZ3HCBW4bY/zu9VZ68d99ws7GS7jIJOKRe2wQDNY=; b=smaQ/hWmZXw49yjv1KMmdYUC35bh7Hqi54dgm6d7GbilXTJA52zqDhckJlJzpJIMqt hg0mKHRoOqsWyTZ5/7ugkIEnJShEe5HtjgRSeMLQfL6M+pdZYpLa0kJiwaFXbOis7OTa RBJZUPS0MeynQmZzP1JBP7KiPsysxjii8lMDc=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type; b=Y63w0CbOEdJEgWRaC1SpnwdYUBkfQU5rLJaU0HI+s05vzKlFLCr0nUGhWV2YcZhAE1 pe0+za/frJZsYcV715knrAUjXRtKNVvipzFzPVCW6sHuZQVIQE0SkzjPkhcGs19yFFvo 32iv3e0YBC6wZDaZC42SJORd5o8tZAiVHF5lk=
Received: by 10.150.114.4 with SMTP id m4mr4890504ybc.442.1300033916854; Sun, 13 Mar 2011 09:31:56 -0700 (PDT)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id l2sm4652132ybn.15.2011.03.13.09.31.53 (version=TLSv1/SSLv3 cipher=OTHER); Sun, 13 Mar 2011 09:31:54 -0700 (PDT)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Sun, 13 Mar 2011 11:31:51 -0500
From: Henry Sinnreich 
To: Andrew Allen , 
Message-ID: 
Thread-Topic: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
Thread-Index: AcvgIl42kPbjR/gjTkGSZWWm9wulvwA5ezXbACT3NHo=
In-Reply-To: 
Mime-version: 1.0
Content-type: multipart/alternative; boundary="B_3382860713_9183333"
Cc: dispatch@ietf.org
Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sun, 13 Mar 2011 16:30:40 -0000

> This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

--B_3382860713_9183333
Content-type: text/plain;
	charset="ISO-8859-1"
Content-transfer-encoding: quoted-printable

Hi Andrew,

Yes, chatting at the IETF would be interesting, though there is nothing
philosophical in the overwhelming hard numbers and trends in the market for
all industry segments driven by applications. And the Web platforms for
these applications.

Henry


On 3/12/11 5:53 PM, "Andrew Allen"  wrote:

>=20
> Henry
>=20
> I think Mary has covered this all very well.
>=20
> I am happy to engage in further philosophical discussions on the future o=
f the
> Internet in Prague.
>=20
> Andrew
> =20
> From: Mary Barnes [mailto:mary.ietf.barnes@gmail.com]
> Sent: Friday, March 11, 2011 01:27 PM
> To: Henry Sinnreich 
> Cc: Christer Holmberg ; Andrew Allen;
> tanakai@nttdocomo.co.jp ; dispatch@ietf.org
> 
> Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instan=
ceid
> =20
> Hi Henry,=A0
>=20
> I'll make just a few comments on this discussion thread, as I don't think=
 it's
> a productive discussion on this mailing list - likely it would be better =
on
> the ietf-discussion list or you might want to approach someone on the IAB
> about this topic, since the IAB is the responsible party in establishing =
and
> appointing liaisons.=A0
>=20
> However, as far as the 3gpp specs that are all online, fairly up-to-date =
as
> best as I can tell:
> http://www.3gpp.org/-specifications-
>=20
> As you puruse the specs, you will find that 3gpp uses a number of the spe=
cs in
> the exact same manner as do "e2e" implementations. =A0So, I don't think the
> situation is as dire as it might appear. =A0We can debate this topic ad nas=
eum,
> but IMS is what it is and is being deployed and nothing any of us say or =
think
> here will change that. =A0The reality is that a large number of contributor=
s to
> the e2e specs are here because the folks that pay them to contribute to t=
he
> IETF make money selling those systems. =A0I will also note that folks that
> aren't interested in IMS can use your document that was published as a
> baseline "e2e" SIP implementation:=A0http://tools.ietf.org/html/rfc5638
>=20
> There is also a spreadsheet of IETF dependencies that is regularly update=
d and
> it includes references to the specs that depend upon core IETF RFCs:
> http://www.3gpp.org/-IETF-Dependencies-and-Priorities-
>=20
> I will note that summaries of issues and actions related to the RAI area
> documents has been posted on the RAI area mailing list in the past:
> http://www.ietf.org/mail-archive/web/rai/current/msg00955.html
> So, I think there has been a reasonable amount of transparency with regar=
ds to
> this working relationship. Since this relationship between 3GPP and IETF =
has
> been in place, there have been a number of documents that have been appro=
ved
> and in general it has been a fruitful relationship on both parts. =A0In som=
e
> cases, core changes were made to SIP that benefited both the e2e model an=
d the
> walled garden. =A0
>=20
> Regards,
> Mary.=A0
>=20
> On Fri, Mar 11, 2011 at 11:28 AM, Henry Sinnreich 
> wrote:
>> Hi Andrew, Christer,
>>=20
>>>> >> My understanding also is that IETF nearly a decade ago agreed to
>>>> >> cooperate with 3GPP (and other standards bodies) in developing and
>>>> >> enhancing internet protocols
>>=20
>> 3GPP etc. documents are best served by having them online, up-to-date, o=
n
>> their respective web sites, where they can be searched and studied.
>> Using IETF RFCs for this seems rather odd, and certainly less effective.
>>=20
>>> >some people in IETF suddently thinks there are more "fun and sexy" stu=
ff to
>>> >work on than SIP...
>>=20
>> As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost al=
l
>> SIP apps are telephony features, while 100,000s of apps are all using HT=
ML,
>> some of them to communicate in new ways - quite different from the SIP
>> telephone call signaling models. RTC-Web may or may not be based on SIP
>> signaling _alone_ , if at all. Communication models and user habits have
>> changed very much (and are still changing) since SIP has been defined.
>>=20
>> Don't you think we have a problem here?
>> If yes, wouldn't listening to the IETF technical plenary presentation on=
 the
>> future of applications be a good start to address the challenge for SIP
>> applications?
>>=20
>> What do you think?
>>=20
>> Henry
>>=20
>>=20
>> On 3/11/11 12:57 AM, "Christer Holmberg" 
>> wrote:
>>=20
>>> >
>>> > Hi,
>>> >
>>> > I agree with Andrew.
>>> >
>>> > People have invested, and are investing, lots of resources into SIP b=
ased
>>> > systems. They are not suddently going to throw everything away just
>>> because
>>> > some people in IETF suddently thinks there are more "fun and sexy" st=
uff
>>> to
>>> > work on than SIP...
>>> >
>>> > Also, regarding the agreement between IETF and 3GPP, if people have i=
ssues
>>> > with that agreement I don't think that people who are working based o=
n
>>> that
>>> > agreement should have to suffer.
>>> >
>>> > Instead, people having issues with the agreement should bring those i=
ssues
>>> to
>>> > a more appropriate instance.
>>> >
>>> > Regards,
>>> >
>>> > Christer
>>> >
>>> >
>>>> >> -----Original Message-----
>>>> >> From: dispatch-bounces@ietf.org
>>>> >> [mailto:dispatch-bounces@ietf.org] On Behalf Of Andrew Allen
>>>> >> Sent: 11. maaliskuuta 2011 0:04
>>>> >> To: Henry Sinnreich; tanakai@nttdocomo.co.jp; dispatch@ietf.org
>>>> >> Subject: Re: [dispatch]
>>>> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>>>> >>
>>>> >> Hi Henry,
>>>> >>
>>>> >> I think then that it's a fair summary of your email to say
>>>> >> that your objections are of a political and philosophical
>>>> >> nature related to your view of what constitutes the internet
>>>> >> and whether the IETF should be involved in enhancing IETF
>>>> >> protocols for stakeholders that utilize those IETF protocols
>>>> >> for deployments that some in the community might not consider
>>>> >> "the Internet" or "the pure Internet" or the "open e2e
>>>> >> Internet" rather than technical issues with the contents of
>>>> >> the drafts.
>>>> >>
>>>> >> While I don't want to start a general debate on what is "the Intern=
et"
>>>> >> here I will state that in my view it's very difficult to
>>>> >> exclude any particular stakeholders and their deployments of
>>>> >> IETF protocols from being part of the Internet when those
>>>> >> stakeholders are in many cases the very same established
>>>> >> telecommunications companies or ISPs that the vast majority
>>>> >> of the population in most countries use to obtain access to
>>>> >> the Internet.
>>>> >>
>>>> >> My understanding also is that IETF nearly a decade ago agreed
>>>> >> to cooperate with 3GPP (and other standards bodies) in
>>>> >> developing and enhancing internet protocols and that IETF
>>>> >> specifically requested (more like insisted) that any
>>>> >> extensions or enhancements to IETF protocols required by 3GPP
>>>> >> be done within IETF according to the procedures defined by
>>>> >> IETF for enhancing IETF protocols and for registering namespaces et=
c.
>>>> >> These drafts simply comply with that agreed way of working
>>>> >> between IETF and 3GPP. To my knowledge this working agreement
>>>> >> is still in effect. I don't think it is the role of the
>>>> >> dispatch WG to determine which applications or deployments of
>>>> >> SIP to support or cooperate with counter to agreements
>>>> >> previously made at IESG level.
>>>> >>
>>>> >> Refusing cooperation based on what is considered not to be
>>>> >> part of the "open e2e Internet" is only likely to set up two
>>>> >> (or more) competing versions of the Internet resulting in the
>>>> >> 21st century equivalent of the standard rail gauge vs broad
>>>> >> rail gauge communications interoperability issue from the
>>>> >> 19th century - effectively killing the goal of an open e2e
>>>> >> internet! As Dale indicated what is considered a walled
>>>> >> garden today may not be walled in the future (if I remember
>>>> >> some very significant walls came tumbling down during just
>>>> >> the last decade of the 20th century alone).
>>>> >>
>>>> >> Different business models will be resolved through the market
>>>> >> not by creating competing incompatible technical solutions.
>>>> >>
>>>> >> In terms of the time used on discussion within dispatch I
>>>> >> don't think there is a need for much more discussion in
>>>> >> dispatch. These drafts could go ahead as AD sponsored without
>>>> >> further WG involvement. As I understand it the IMEI URN
>>>> >> definition draft did not as part of the URN registration
>>>> >> process need to even be discussed in dispatch at all
>>>> >> (although Dale Worley's review was invaluable in improving
>>>> >> the draft technically) as the NIDs list was the fora for this
>>>> >> discussion and where previous versions of that draft were discussed=
.
>>>> >>
>>>> >> It should be noted all we are talking about with regard to
>>>> >> these drafts is the registration of a URN namespace and
>>>> >> definition of its usage as an Instance ID as required by RFC
>>>> >> 5626. If we can't agree that then what future extensions can we agr=
ee?
>>>> >>
>>>> >> Andrew
>>>> >>
>>>> >> -----Original Message-----
>>>> >> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
>>>> >> Sent: Thursday, March 10, 2011 11:03 AM
>>>> >> To: Andrew Allen; tanakai@nttdocomo.co.jp; dispatch@ietf.org
>>>> >> Subject: Re: [dispatch]
>>>> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>>>> >>
>>>> >> Hi Andrew,
>>>> >>
>>>> >> I have generic objections to see the publication of yet
>>>> >> another informational RFC about practices in consortiums of
>>>> >> closed networks in the IETF, for the following reasons:
>>>> >>
>>>> >> * Using the RFC process as a publishing venue for
>>>> >> organizations that don't explicitly do that in support of the
>>>> >> open e2e Internet.
>>>> >>
>>>> >> * The resources of the IETF are not unlimited, meeting time
>>>> >> and email time to discuss is scarce and the IETF registries
>>>> >> are just not scalable enough to support registrations for all
>>>> >> the extensions that benefit countless external organizations
>>>> >> and companies that deploy Internet technology.
>>>> >> New business plans and the required extensions pop up in an
>>>> >> endless fashion.
>>>> >> Big organizations such as you mention have their own
>>>> >> resources to publish their internal technical documents and
>>>> >> white papers if they desire to do so.
>>>> >>
>>>> >> Members of outside organizations are certainly doing
>>>> >> everybody a good service by informing via email of new
>>>> >> standards/extensions in other organization, such as you have
>>>> >> done below, but such information need not overload the
>>>> >> Internet RFC process nor overload the Internet registries.
>>>> >>
>>>> >> * Hard for SIP developers and right out impossible for
>>>> >> application developers.
>>>> >> Though SIP works fine indeed across the world; SIPit and the
>>>> >> SIP Forum take care of unavoidable problems still encountered
>>>> >> by operators, arguably increasing SIP complexity for
>>>> >> non-Internet purposes is not a good idea, even if new SIP
>>>> >> extensions are labeled as just "informational". It makes it
>>>> >> even more difficult for new developers entering the field to
>>>> >> ever understand, let alone write code for some or all that
>>>> >> has been published already about SIP for example. Checking
>>>> >> right now there are 197 RFCs matching SIP and 78 I-Ds found
>>>> >> by the RFC search engine.
>>>> >>
>>>> >> It is time to leave SIP alone and move on to other critical
>>>> >> areas for the Internet, such as for example the future of
>>>> >> applications that will be discussed in the next IETF
>>>> >> technical plenary on Monday, March 28.
>>>> >>
>>>> >> With this in mind, I would rather not go into discussing your
>>>> >> I-D in detail, though at first glance it seems technically
>>>> >> correct and if it were discussed on the list, may indeed
>>>> >> technically qualify as an informational RFC - just what I am
>>>> >> arguing against.
>>>> >>
>>>> >> These are all personal opinions. The chairs and ADs may or
>>>> >> may not discuss or agree/disagree with. Your I-D is just a
>>>> >> good opportunity to raise these issues.
>>>> >>
>>>> >> Note that I am not suggesting any voting on your I-D, if it
>>>> >> will be discussed, since for various reasons, the majority of
>>>> >> folks attending the DISPATCH WG are from various outside
>>>> >> organizations with business plans that may or may not fully
>>>> >> embrace the open e2e Internet; a perfect legitimate choice
>>>> >> for them, but not =A0in the IETF (I just can't remember when
>>>> >> ever there was a vote against any new I-D for SIP, since no
>>>> >> cost was perceived by those voting).
>>>> >>
>>>> >> Thanks,
>>>> >> Henry
>>>> >>
>>>> >>
>>>> >> On 3/9/11 5:45 PM, "Andrew Allen"  wrote:
>>>> >>
>>>>> >>>
>>>>> >>> Henry
>>>>> >>>
>>>>> >>> Would you care to elaborate and explain your specific concerns? =A0=
How
>>>> >> is what
>>>>> >>> is defined in these drafts potentially incompatible with
>>>> >> the open, e2e
>>>>> >>> architectural and in any other way technical conflict with the
>>>> >> Internet?
>>>>> >>>
>>>>> >>> The two drafts do the following things:
>>>>> >>>
>>>>> >>> One draft simply defines and registers with IANA a URN namespace
>>>> >> according to
>>>>> >>> the procedures defined for doing that.
>>>>> >>>
>>>>> >>> The other draft simply defines how this URN can be used as
>>>> >> an instance
>>>> >> ID as
>>>>> >>> required by RFC 5626.
>>>>> >>>
>>>>> >>> SIP and the mechanisms in RFC 5626 are already being used
>>>> >> in networks
>>>> >> defined
>>>>> >>> by 3GPP and are being deployed by members of the GSMA. 3GPP and G=
SMA
>>>> >> have
>>>>> >>> already defined this behavior using their existing device
>>>> >> Identifiers
>>>> >> for
>>>>> >>> instance ID generation in their specifications. These
>>>> >> drafts complete
>>>> >> the
>>>>> >>> process doing things the right way by reserving the namespace and
>>>> >> defining the
>>>>> >>> usage of it using the IETF defined mechanisms for doing that.
>>>>> >>>
>>>>> >>> Andrew
>>>>> >>>
>>>>> >>> ----- Original Message -----
>>>>> >>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]
>>>>> >>> Sent: Wednesday, March 09, 2011 03:10 PM
>>>>> >>> To: Itsuma TANAKA ; dispatch@ietf.org
>>>>> >>> 
>>>>> >>> Subject: Re: [dispatch]
>>>> >> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid
>>>>> >>>
>>>>> >>> Such work in the IETF would require IMO an explanation about its
>>>>> >>> compatibility with the open, e2e architectural and other technica=
l
>>>>> >>> principles of the Internet. It actually seems to be the opposite.
>>>>> >>>
>>>>> >>> For this reason, such work may be better accomplished in other
>>>>> >>> organizations, so as to avoid architectural principle and other
>>>> >> technical
>>>>> >>> conflicts. Simplicity and scalability comes to mind among other,
>>>> >> though this
>>>>> >>> is better discussed elsewhere.
>>>>> >>>
>>>>> >>> Thanks,
>>>>> >>>
>>>>> >>> Henry Sinnreich
>>>>> >>>
>>>>> >>>
>>>>> >>> On 3/9/11 8:54 AM, "Itsuma TANAKA"  wrot=
e:
>>>>> >>>
>>>>>> >>>> Dear All,
>>>>>> >>>>
>>>>>> >>>> I work for NTT DOCOMO, Inc in Japan. and attend various
>>>> >> 3GPP meetings
>>>> >> (SA2,
>>>>>> >>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking, Roa=
ming
>>>> >> Expert
>>>>>> >>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup
>>>> >> meetings. =A0I
>>>> >> serve
>>>>>> >>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>>>>>> >>>>
>>>>>> >>>> I would like to express my support of
>>>> >> draft-montemurro-gsma-imei-urn
>>>> >> and the
>>>>>> >>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>>>>>> >>>>
>>>>>> >>>> The reasons are exactly same as what Ricky Kaura (Samsung)
>>>> >> expressed.
>>>> >> A
>>>>>> >>>> various 3GPP specifications (IMS Registration in 3GPP
>>>> >> TS24.229, IMS
>>>>>> >>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 and
>>>> >> TS24.292), and
>>>>>> >>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice and SMS")=
 are
>>>> >> using
>>>>>> >>>> these drafts.
>>>>>> >>>>
>>>>>> >>>> These internet drafts are very important to support various key
>>>> >> features in
>>>>>> >>>> 3GPP and to enable GSMA Voice over LTE. =A0Therefore, I kindly wo=
uld
>>>> >> like ask
>>>>>> >>>> that this work is progressed as quickly as possible in IETF.
>>>>>> >>>>
>>>>>> >>>>
>>>>>> >>>> Best regards,
>>>>>> >>>>
>>>>>> >>>> Itsuma Tanaka
>>>>>> >>>> NTT DOCOMO
>>>>>> >>>>
>>>>>> >>>>
>>>>>> >>>> _______________________________________________
>>>>>> >>>> dispatch mailing list
>>>>>> >>>> dispatch@ietf.org
>>>>>> >>>> https://www.ietf.org/mailman/listinfo/dispatch
>>>>> >>>
>>>>> >>>
>>>>> >>> _______________________________________________
>>>>> >>> dispatch mailing list
>>>>> >>> dispatch@ietf.org
>>>>> >>> https://www.ietf.org/mailman/listinfo/dispatch
>>>>> >>>
>>>>> >>>
>>>> >> -------------------------------------------------------------------=
--
>>>>> >>> This transmission (including any attachments) may contain
>>>> >> confidential
>>>>> >>> information, privileged material (including material
>>>> >> protected by the
>>>>> >>> solicitor-client or other applicable privileges), or constitute
>>>> >> non-public
>>>>> >>> information. Any use of this information by anyone other than the
>>>> >> intended
>>>>> >>> recipient is prohibited. If you have received this transmission i=
n
>>>> >> error,
>>>>> >>> please immediately reply to the sender and delete this informatio=
n
>>>> >> from your
>>>>> >>> system. Use, dissemination, distribution, or reproduction of this
>>>> >> transmission
>>>>> >>> by unintended recipients is not authorized and may be unlawful.
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >> -------------------------------------------------------------------=
--
>>>> >> This transmission (including any attachments) may contain
>>>> >> confidential information, privileged material (including
>>>> >> material protected by the solicitor-client or other
>>>> >> applicable privileges), or constitute non-public information.
>>>> >> Any use of this information by anyone other than the intended
>>>> >> recipient is prohibited. If you have received this
>>>> >> transmission in error, please immediately reply to the sender
>>>> >> and delete this information from your system. Use,
>>>> >> dissemination, distribution, or reproduction of this
>>>> >> transmission by unintended recipients is not authorized and
>>>> >> may be unlawful.
>>>> >> _______________________________________________
>>>> >> dispatch mailing list
>>>> >> dispatch@ietf.org
>>>> >> https://www.ietf.org/mailman/listinfo/dispatch
>>>> >>
>>=20
>>=20
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>=20
> ---------------------------------------------------------------------
> This transmission (including any attachments) may contain confidential
> information, privileged material (including material protected by the
> solicitor-client or other applicable privileges), or constitute non-publi=
c
> information. Any use of this information by anyone other than the intende=
d
> recipient is prohibited. If you have received this transmission in error,
> please immediately reply to the sender and delete this information from y=
our
> system. Use, dissemination, distribution, or reproduction of this transmi=
ssion
> by unintended recipients is not authorized and may be unlawful.


--B_3382860713_9183333
Content-type: text/html;
	charset="ISO-8859-1"
Content-transfer-encoding: quoted-printable



Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-instanceid=



Hi Andrew,



Yes, chatting at the IETF would be interesting, though there is nothing phi=
losophical in the overwhelming hard numbers and trends in the market for all=
 industry segments driven by applications. And the Web platforms for these a=
pplications. 



Henry





On 3/12/11 5:53 PM, "Andrew Allen" <a=
allen@rim.com> wrote:




<=
FONT SIZE=3D"2">

Henry



I think Mary has covered this all very well.



I am happy to engage in further philosophical discussions on the future of =
the Internet in Prague.



Andrew

 

From: Mary Barnes=
 [mailto:mary.ietf.barnes@gmail.=
com] 

Sent: Friday, March 11, 2011 01:27 PM

To: Henry Sinnreich <henry.si=
nnreich@gmail.com> 

Cc: Christer Holmberg <c=
hrister.holmberg@ericsson.com>; Andrew Allen; tanakai@nttdocomo.co.jp <tanakai@nttdocomo.co.jp>; dispatch@ietf.=
org <dispatch@ietf.org> 

Subject: Re: [dispatch] Revisionofdraft-allen-dispatch-imei-urn-as-i=
nstanceid 

 

Hi Henry,=A0



I'll make just a few comments on this discussion thread, as I don't think i=
t's a productive discussion on this mailing list - likely it would be better=
 on the ietf-discussion list or you might want to approach someone on the IA=
B about this topic, since the IAB is the responsible party in establishing a=
nd appointing liaisons.=A0



However, as far as the 3gpp specs that are all online, fairly up-to-date as=
 best as I can tell:

http://www.3gpp.org/-specifi=
cations-



As you puruse the specs, you will find that 3gpp uses a number of the specs=
 in the exact same manner as do "e2e" implementations. =A0So, I don'=
t think the situation is as dire as it might appear. =A0We can debate this top=
ic ad naseum, but IMS is what it is and is being deployed and nothing any of=
 us say or think here will change that. =A0The reality is that a large number =
of contributors to the e2e specs are here because the folks that pay them to=
 contribute to the IETF make money selling those systems. =A0I will also note =
that folks that aren't interested in IMS can use your document that was publ=
ished as a baseline "e2e" SIP implementation:=A0http://tools.ietf.org/html/rfc5638



There is also a spreadsheet of IETF dependencies that is regularly updated =
and it includes references to the specs that depend upon core IETF RFCs:

http://www=
.3gpp.org/-IETF-Dependencies-and-Priorities-



I will note that summaries of issues and actions related to the RAI area do=
cuments has been posted on the RAI area mailing list in the past:

ht=
tp://www.ietf.org/mail-archive/web/rai/current/msg00955.html

So, I think there has been a reasonable amount of transparency with regards=
 to this working relationship. Since this relationship between 3GPP and IETF=
 has been in place, there have been a number of documents that have been app=
roved and in general it has been a fruitful relationship on both parts. =A0In =
some cases, core changes were made to SIP that benefited both the e2e model =
and the walled garden. =A0



Regards,

Mary.=A0



On Fri, Mar 11, 2011 at 11:28 AM, Henry Sinnreich <henry.sinnreich@gmail.com> wrote:


<=
SPAN STYLE=3D'font-size:13pt'>Hi Andrew, Christer,



>> My understanding also is that IETF nearly a decade ago agreed to
>> cooperate with 3GPP (and other standards bodies) in developing and=


>> enhancing internet protocols



3GPP etc. documents are best served by having them online, up-to-date, on
their respective web sites, where they can be searched and studied.

Using IETF RFCs for this seems rather odd, and certainly less effective.


>some people in IETF suddently thinks there are more "fun and sexy&=
quot; stuff to

>work on than SIP...



As mentioned, after 100's of SIP RFCs and I-Ds over 12+ years, almost all
SIP apps are telephony features, while 100,000s of apps are all using HTML,=


some of them to communicate in new ways - quite different from the SIP

telephone call signaling models. RTC-Web may or may not be based on SIP

signaling _alone_ , if at all. Communication models and user habits have
changed very much (and are still changing) since SIP has been defined.



Don't you think we have a problem here?

If yes, wouldn't listening to the IETF technical plenary presentation on th=
e

future of applications be a good start to address the challenge for SIP

applications?



What do you think?



Henry





On 3/11/11 12:57 AM, "Christer Holmberg" <christer.holmberg@ericsson.com>

wrote:



>

> Hi,

>

> I agree with Andrew.

>

> People have invested, and are investing, lots of resources into SIP ba=
sed

> systems. They are not suddently going to throw everything away just be=
cause

> some people in IETF suddently thinks there are more "fun and sexy=
" stuff to

> work on than SIP...

>

> Also, regarding the agreement between IETF and 3GPP, if people have is=
sues

> with that agreement I don't think that people who are working based on=
 that

> agreement should have to suffer.

>

> Instead, people having issues with the agreement should bring those is=
sues to

> a more appropriate instance.

>

> Regards,

>

> Christer

>

>

>> -----Original Message-----

>> From: dispatch-bounces@ietf.or=
g

>> [mailto:dispatch-bounce=
s@ietf.org] On Behalf Of Andrew Allen

>> Sent: 11. maaliskuuta 2011 0:04

>> To: Henry Sinnreich; tanakai@ntt=
docomo.co.jp; dispatch@ietf.org

>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>

>> Hi Henry,

>>

>> I think then that it's a fair summary of your email to say

>> that your objections are of a political and philosophical

>> nature related to your view of what constitutes the internet

>> and whether the IETF should be involved in enhancing IETF

>> protocols for stakeholders that utilize those IETF protocols

>> for deployments that some in the community might not consider

>> "the Internet" or "the pure Internet" or the &=
quot;open e2e

>> Internet" rather than technical issues with the contents of
>> the drafts.

>>

>> While I don't want to start a general debate on what is "the =
Internet"

>> here I will state that in my view it's very difficult to

>> exclude any particular stakeholders and their deployments of

>> IETF protocols from being part of the Internet when those

>> stakeholders are in many cases the very same established

>> telecommunications companies or ISPs that the vast majority

>> of the population in most countries use to obtain access to

>> the Internet.

>>

>> My understanding also is that IETF nearly a decade ago agreed

>> to cooperate with 3GPP (and other standards bodies) in

>> developing and enhancing internet protocols and that IETF

>> specifically requested (more like insisted) that any

>> extensions or enhancements to IETF protocols required by 3GPP

>> be done within IETF according to the procedures defined by

>> IETF for enhancing IETF protocols and for registering namespaces e=
tc.

>> These drafts simply comply with that agreed way of working

>> between IETF and 3GPP. To my knowledge this working agreement

>> is still in effect. I don't think it is the role of the

>> dispatch WG to determine which applications or deployments of

>> SIP to support or cooperate with counter to agreements

>> previously made at IESG level.

>>

>> Refusing cooperation based on what is considered not to be

>> part of the "open e2e Internet" is only likely to set up=
 two

>> (or more) competing versions of the Internet resulting in the

>> 21st century equivalent of the standard rail gauge vs broad

>> rail gauge communications interoperability issue from the

>> 19th century - effectively killing the goal of an open e2e

>> internet! As Dale indicated what is considered a walled

>> garden today may not be walled in the future (if I remember

>> some very significant walls came tumbling down during just

>> the last decade of the 20th century alone).

>>

>> Different business models will be resolved through the market

>> not by creating competing incompatible technical solutions.

>>

>> In terms of the time used on discussion within dispatch I

>> don't think there is a need for much more discussion in

>> dispatch. These drafts could go ahead as AD sponsored without

>> further WG involvement. As I understand it the IMEI URN

>> definition draft did not as part of the URN registration

>> process need to even be discussed in dispatch at all

>> (although Dale Worley's review was invaluable in improving

>> the draft technically) as the NIDs list was the fora for this

>> discussion and where previous versions of that draft were discusse=
d.

>>

>> It should be noted all we are talking about with regard to

>> these drafts is the registration of a URN namespace and

>> definition of its usage as an Instance ID as required by RFC

>> 5626. If we can't agree that then what future extensions can we ag=
ree?

>>

>> Andrew

>>

>> -----Original Message-----

>> From: Henry Sinnreich [=
mailto:henry.sinnreich@gmail.com]

>> Sent: Thursday, March 10, 2011 11:03 AM

>> To: Andrew Allen; tanakai@nttdoc=
omo.co.jp; dispatch@ietf.org

>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>

>> Hi Andrew,

>>

>> I have generic objections to see the publication of yet

>> another informational RFC about practices in consortiums of

>> closed networks in the IETF, for the following reasons:

>>

>> * Using the RFC process as a publishing venue for

>> organizations that don't explicitly do that in support of the

>> open e2e Internet.

>>

>> * The resources of the IETF are not unlimited, meeting time

>> and email time to discuss is scarce and the IETF registries

>> are just not scalable enough to support registrations for all

>> the extensions that benefit countless external organizations

>> and companies that deploy Internet technology.

>> New business plans and the required extensions pop up in an

>> endless fashion.

>> Big organizations such as you mention have their own

>> resources to publish their internal technical documents and

>> white papers if they desire to do so.

>>

>> Members of outside organizations are certainly doing

>> everybody a good service by informing via email of new

>> standards/extensions in other organization, such as you have

>> done below, but such information need not overload the

>> Internet RFC process nor overload the Internet registries.

>>

>> * Hard for SIP developers and right out impossible for

>> application developers.

>> Though SIP works fine indeed across the world; SIPit and the

>> SIP Forum take care of unavoidable problems still encountered

>> by operators, arguably increasing SIP complexity for

>> non-Internet purposes is not a good idea, even if new SIP

>> extensions are labeled as just "informational". It makes=
 it

>> even more difficult for new developers entering the field to

>> ever understand, let alone write code for some or all that

>> has been published already about SIP for example. Checking

>> right now there are 197 RFCs matching SIP and 78 I-Ds found

>> by the RFC search engine.

>>

>> It is time to leave SIP alone and move on to other critical

>> areas for the Internet, such as for example the future of

>> applications that will be discussed in the next IETF

>> technical plenary on Monday, March 28.

>>

>> With this in mind, I would rather not go into discussing your

>> I-D in detail, though at first glance it seems technically

>> correct and if it were discussed on the list, may indeed

>> technically qualify as an informational RFC - just what I am

>> arguing against.

>>

>> These are all personal opinions. The chairs and ADs may or

>> may not discuss or agree/disagree with. Your I-D is just a

>> good opportunity to raise these issues.

>>

>> Note that I am not suggesting any voting on your I-D, if it

>> will be discussed, since for various reasons, the majority of

>> folks attending the DISPATCH WG are from various outside

>> organizations with business plans that may or may not fully

>> embrace the open e2e Internet; a perfect legitimate choice

>> for them, but not =A0in the IETF (I just can't remember when

>> ever there was a vote against any new I-D for SIP, since no

>> cost was perceived by those voting).

>>

>> Thanks,

>> Henry

>>

>>

>> On 3/9/11 5:45 PM, "Andrew Allen" <aallen@rim.com> wrote:

>>

>>>

>>> Henry

>>>

>>> Would you care to elaborate and explain your specific concerns=
? =A0How

>> is what

>>> is defined in these drafts potentially incompatible with

>> the open, e2e

>>> architectural and in any other way technical conflict with the=


>> Internet?

>>>

>>> The two drafts do the following things:

>>>

>>> One draft simply defines and registers with IANA a URN namespa=
ce

>> according to

>>> the procedures defined for doing that.

>>>

>>> The other draft simply defines how this URN can be used as

>> an instance

>> ID as

>>> required by RFC 5626.

>>>

>>> SIP and the mechanisms in RFC 5626 are already being used

>> in networks

>> defined

>>> by 3GPP and are being deployed by members of the GSMA. 3GPP an=
d GSMA

>> have

>>> already defined this behavior using their existing device

>> Identifiers

>> for

>>> instance ID generation in their specifications. These

>> drafts complete

>> the

>>> process doing things the right way by reserving the namespace =
and

>> defining the

>>> usage of it using the IETF defined mechanisms for doing that.<=
BR>
>>>

>>> Andrew

>>>

>>> ----- Original Message -----

>>> From: Henry Sinnreich [mailto:henry.sinnreich@gmail.com]

>>> Sent: Wednesday, March 09, 2011 03:10 PM

>>> To: Itsuma TANAKA <tanaka=
i@nttdocomo.co.jp>; dispatch@ietf.org=


>>> <dispatch@ietf.org>

>>> Subject: Re: [dispatch]

>> Revisionofdraft-allen-dispatch-imei-urn-as-instanceid

>>>

>>> Such work in the IETF would require IMO an explanation about i=
ts

>>> compatibility with the open, e2e architectural and other techn=
ical

>>> principles of the Internet. It actually seems to be the opposi=
te.

>>>

>>> For this reason, such work may be better accomplished in other=


>>> organizations, so as to avoid architectural principle and othe=
r

>> technical

>>> conflicts. Simplicity and scalability comes to mind among othe=
r,

>> though this

>>> is better discussed elsewhere.

>>>

>>> Thanks,

>>>

>>> Henry Sinnreich

>>>

>>>

>>> On 3/9/11 8:54 AM, "Itsuma TANAKA" <tanakai@nttdocomo.co.jp> wrote:

>>>

>>>> Dear All,

>>>>

>>>> I work for NTT DOCOMO, Inc in Japan. and attend various
>> 3GPP meetings

>> (SA2,

>>>> CT1, CT4, and CT3) and GSM Association (GSMA) Interworking=
, Roaming

>> Expert

>>>> Group (IREG), Roaming in LTE (RILTE) and Packet subgroup
>> meetings. =A0I

>> serve

>>>> as a Vice Chair of GSMA IREG RILTE and Packet subgroups.
>>>>

>>>> I would like to express my support of

>> draft-montemurro-gsma-imei-urn

>> and the

>>>> associated draft-allen-dispatch-imei-urn-as-instanceid.
>>>>

>>>> The reasons are exactly same as what Ricky Kaura (Samsung)=


>> expressed.

>> A

>>>> various 3GPP specifications (IMS Registration in 3GPP

>> TS24.229, IMS

>>>> emergency calls in TS23.237 and TS24.237, ICS in TS23.292 =
and

>> TS24.292), and

>>>> GSMA document (GSMA PRD IR.92, "IMS Profile for Voice=
 and SMS") are

>> using

>>>> these drafts.

>>>>

>>>> These internet drafts are very important to support variou=
s key

>> features in

>>>> 3GPP and to enable GSMA Voice over LTE. =A0Therefore, I kind=
ly would

>> like ask

>>>> that this work is progressed as quickly as possible in IET=
F.

>>>>

>>>>

>>>> Best regards,

>>>>

>>>> Itsuma Tanaka

>>>> NTT DOCOMO

>>>>

>>>>

>>>> _______________________________________________

>>>> dispatch mailing list

>>>> dispatch@ietf.org

>>>> h=
ttps://www.ietf.org/mailman/listinfo/dispatch

>>>

>>>

>>> _______________________________________________

>>> dispatch mailing list

>>> dispatch@ietf.org

>>> https=
://www.ietf.org/mailman/listinfo/dispatch

>>>

>>>

>> ------------------------------------------------------------------=
---

>>> This transmission (including any attachments) may contain

>> confidential

>>> information, privileged material (including material

>> protected by the

>>> solicitor-client or other applicable privileges), or constitut=
e

>> non-public

>>> information. Any use of this information by anyone other than =
the

>> intended

>>> recipient is prohibited. If you have received this transmissio=
n in

>> error,

>>> please immediately reply to the sender and delete this informa=
tion

>> from your

>>> system. Use, dissemination, distribution, or reproduction of t=
his

>> transmission

>>> by unintended recipients is not authorized and may be unlawful=
.

>>

>>

>>

>>

>>

>>

>> ------------------------------------------------------------------=
---

>> This transmission (including any attachments) may contain

>> confidential information, privileged material (including

>> material protected by the solicitor-client or other

>> applicable privileges), or constitute non-public information.

>> Any use of this information by anyone other than the intended

>> recipient is prohibited. If you have received this

>> transmission in error, please immediately reply to the sender

>> and delete this information from your system. Use,

>> dissemination, distribution, or reproduction of this

>> transmission by unintended recipients is not authorized and

>> may be unlawful.

>> _______________________________________________

>> dispatch mailing list

>> dispatch@ietf.org

>> https://w=
ww.ietf.org/mailman/listinfo/dispatch

>>





_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.o=
rg/mailman/listinfo/dispatch


=


--------------------------------------------------------------------- 

This transmission (including any attachments) may contain confidential info=
rmation, privileged material (including material protected by the solicitor-=
client or other applicable privileges), or constitute non-public information=
. Any use of this information by anyone other than the intended recipient is=
 prohibited. If you have received this transmission in error, please immedia=
tely reply to the sender and delete this information from your system. Use, =
dissemination, distribution, or reproduction of this transmission by uninten=
ded recipients is not authorized and may be unlawful.







--B_3382860713_9183333--



From christer.holmberg@ericsson.com  Mon Mar 14 05:01:46 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7F9573A68BA for ; Mon, 14 Mar 2011 05:01:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.448
X-Spam-Level: 
X-Spam-Status: No, score=-6.448 tagged_above=-999 required=5 tests=[AWL=-0.076, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, SARE_SUB_OBFU_Q1=0.227]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2dv1W6u88zmI for ; Mon, 14 Mar 2011 05:01:45 -0700 (PDT)
Received: from mailgw10.se.ericsson.net (mailgw10.se.ericsson.net [193.180.251.61]) by core3.amsl.com (Postfix) with ESMTP id 1E6253A6AF5 for ; Mon, 14 Mar 2011 05:01:38 -0700 (PDT)
X-AuditID: c1b4fb3d-b7bbbae000005311-77-4d7e03f5ed72
Received: from esessmw0197.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw10.se.ericsson.net (Symantec Mail Security) with SMTP id 1B.29.21265.5F30E7D4; Mon, 14 Mar 2011 13:03:01 +0100 (CET)
Received: from ESESSCMS0356.eemea.ericsson.se ([169.254.1.30]) by esessmw0197.eemea.ericsson.se ([153.88.115.87]) with mapi; Mon, 14 Mar 2011 13:03:01 +0100
From: Christer Holmberg 
To: "rtc-web@alvestrand.no" , "dispatch@ietf.org" 
Date: Mon, 14 Mar 2011 13:03:00 +0100
Thread-Topic: Draft new version: draft-holmberg-rtcweb-ucreqs-01
Thread-Index: AcvhtBoKcB/pe1tiQlOtldNCEQq7+wAi0FAw
Message-ID: <7F2072F1E0DE894DA4B517B93C6A05851948FF3678@ESESSCMS0356.eemea.ericsson.se>
References: ,   <4D7D199F.6000900@alvestrand.no>
In-Reply-To: <4D7D199F.6000900@alvestrand.no>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAA==
Subject: [dispatch] Draft new version: draft-holmberg-rtcweb-ucreqs-01
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 14 Mar 2011 12:01:46 -0000

Hi,

We've submitted a new version of the ucreqs draft.

The new version contains fixes and additions based on comments given on the=
 previous version on the list.

The draft can be found at: http://www.ietf.org/id/draft-holmberg-rtcweb-ucr=
eqs-01.txt

Regards,

Christer=

From Thierry.Bessis@alcatel-lucent.com  Mon Mar 14 13:27:37 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E7D543A6B57 for ; Mon, 14 Mar 2011 13:27:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.097
X-Spam-Level: 
X-Spam-Status: No, score=-6.097 tagged_above=-999 required=5 tests=[AWL=0.500,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ThuaDvXuYGTz for ; Mon, 14 Mar 2011 13:27:36 -0700 (PDT)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 1D7763A6A6D for ; Mon, 14 Mar 2011 13:27:36 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p2EKSwHG023404 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 14 Mar 2011 15:28:58 -0500 (CDT)
Received: from [135.244.0.142] (tbessis.lra.lucent.com [135.244.0.142]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2EKSujt005295; Mon, 14 Mar 2011 15:28:57 -0500 (CDT)
Message-ID: <4D7E7A88.3000600@alcatel-lucent.com>
Date: Mon, 14 Mar 2011 15:28:56 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>	<060701cbdfc6$8fa7ab40$aef701c0$@packetizer.com>	<4D7A50AA.3060805@alcatel-lucent.com>	<074301cbe01a$c2f92330$48eb6990$@packetizer.com>	<4D7ABCAC.2070802@alcatel-lucent.com> <4D7AC073.8010806@cisco.com>
In-Reply-To: <4D7AC073.8010806@cisco.com>
Content-Type: multipart/alternative; boundary="------------030200050804000300050408"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 14 Mar 2011 20:27:38 -0000

This is a multi-part message in MIME format.
--------------030200050804000300050408
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Sorry Paul that it was not apparent from our draft that this was the goal.
This goal has also be chosen because the solution to get to this goal is 
pretty simple.

Cordially,

Thierry



On 11-Mar-2011 18:38, Paul Kyzivat wrote:
>
>
> On 3/11/2011 7:22 PM, Thierry Bessis wrote:
>
> ...
>
>> The requirement we used for our load balancing
>> requirement is based on the need to balance the utilization ratio on
>> each node
>
> Ah!!! Finally I understand what you are up to!
> It never made sense to me up till now.
>
>     Thanks,
>     Paul
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------030200050804000300050408
Content-Type: multipart/related;
 boundary="------------040700070704070302050202"


--------------040700070704070302050202
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
    
  
  
    Sorry Paul that it was not apparent from our draft that this was the
    goal. 

    This goal has also be chosen because the solution to get to this
    goal is pretty simple. 

    

    Cordially, 

    

    Thierry

    

    

    

    On 11-Mar-2011 18:38, Paul Kyzivat wrote:
    

      

      

      On 3/11/2011 7:22 PM, Thierry Bessis wrote:
      

      

      ...
      

      

      
The requirement we used for our load
        balancing
        

        requirement is based on the need to balance the utilization
        ratio on
        

        each node
        

      

      

      Ah!!! Finally I understand what you are up to!
      

      It never made sense to me up till now.
      

      

          Thanks,
      

          Paul
      

      _______________________________________________
      

      dispatch mailing list
      

      dispatch@ietf.org
      

      https://www.ietf.org/mailman/listinfo/dispatch
      

    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------040700070704070302050202
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------040700070704070302050202--

--------------030200050804000300050408--

From dworley@avaya.com  Mon Mar 14 15:20:01 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2946D3A6BD2 for ; Mon, 14 Mar 2011 15:20:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.541
X-Spam-Level: 
X-Spam-Status: No, score=-102.541 tagged_above=-999 required=5 tests=[AWL=0.058, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fUzB9Sq0sNnv for ; Mon, 14 Mar 2011 15:20:00 -0700 (PDT)
Received: from p-us1-iereast-outbound-tmp.us1.avaya.com (nj300815-nj-outbound.net.avaya.com [135.11.29.16]) by core3.amsl.com (Postfix) with ESMTP id 3C0763A6944 for ; Mon, 14 Mar 2011 15:19:57 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEADIxfk2HCzI1/2dsb2JhbACmDXenJAKaBIViBJA6
X-IronPort-AV: E=Sophos;i="4.62,319,1297054800"; d="scan'208";a="63447349"
Received: from unknown (HELO p-us1-erheast.us1.avaya.com) ([135.11.50.53]) by p-us1-iereast-outbound-tmp.us1.avaya.com with ESMTP; 14 Mar 2011 18:21:20 -0400
X-IronPort-AV: E=Sophos;i="4.62,319,1297054800"; d="scan'208";a="620346350"
Received: from dc-us1hcex2.us1.avaya.com (HELO DC-US1HCEX2.global.avaya.com) ([135.11.52.21]) by p-us1-erheast-out.us1.avaya.com with ESMTP; 14 Mar 2011 18:21:12 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX2.global.avaya.com ([::1]) with mapi; Mon, 14 Mar 2011 18:21:12 -0400
From: "Worley, Dale R (Dale)" 
To: Thierry Bessis 
Date: Mon, 14 Mar 2011 18:20:57 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvel8hDOpcAFAaATj+yN/7xg3drLgD/k7V4
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,<4D76C924.7050900@cisco.com> , <4D77E128.8060302@alcatel-lucent.com>
In-Reply-To: <4D77E128.8060302@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 14 Mar 2011 22:20:01 -0000

> From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]=20
>=20
> If you have read my previous comments, we suggest there is no
> algorithm per se: this is a measurement of the resource: If you
> measure the CPU and the Memory (for example), you'll take the
> smallest. It is better to smooth (with a 10s sliding window average
> for instance) the number.

But "If you measure the CPU and the Memory (for example), you'll take
the smallest" *is* an algorithm.  Its name is "minimum".  There are an
infinite number of other algorithms.  And you do not state in the
draft that you've chosen "minimum".

Also, the particular smoothing chosen is not trivial, and should be
stated in the draft.

> This is not what we have in mind: the weight should be your relative
> resource usage. (0 - 100%). Using the proposed algorithm, this insure
> that servers reach full capacity simultaneously (even when they have
> vastly different capacities).

Actually, you want the weight to be (or to be proportional) to the
*unused* portion of your resources, so that an entirely unloaded
system has maximum weight and a totally loaded system has zero weight.

But either way, you don't state this in the draft.

And now I realize you sneaked one past me.  The draft refers to RFC
2782 in a number of places, and I didn't recognize it.  You should say
RFC 3263, which is what people who've implemented SIP recognize.  And
the implication is to replace, in the RFC 3263 process, the "weight"
values in the SRV records with the "weight" values fed back from the
servers.

> Yes, this is true. We are successfully using the simple algorithm
> mentioned with DNS (where we update the weight every
> minute). Otherwise they are algorithms that can be used to improve the
> oscillation in the feedback chain command. We think that since we are
> proposing a default one that works, this should be OK.

I'd like to see some analysis and modeling to show that this
particular updating strategy works in a wide range of operating
situations.  Otherwise ... unpleasant ... surprises may happen in the
field.

> > But I don't see any of the needed analysis in the draft, and I
> > don't see any indication in the draft of a realization that this
> > analysis needs to be done.  Until that is fixed, I predict that a
> > solution will not be found, or worse, a solution that appears to
> > work on paper but does not work in practice.
>=20
> This solution works in practice: we are using it in ALU between
> components and nodes that are in the same NE (we currently use a
> proprietary transport mechanism to return the weights).

You mean it works in your one situation.

Dale

From Thierry.Bessis@alcatel-lucent.com  Mon Mar 14 16:35:40 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B2C143A7029 for ; Mon, 14 Mar 2011 16:35:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.18
X-Spam-Level: 
X-Spam-Status: No, score=-6.18 tagged_above=-999 required=5 tests=[AWL=0.417,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7pOCqwzJPF76 for ; Mon, 14 Mar 2011 16:35:38 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 172F33A701B for ; Mon, 14 Mar 2011 16:35:35 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2ENastB012040 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 14 Mar 2011 18:36:54 -0500 (CDT)
Received: from [135.244.0.142] (tbessis.lra.lucent.com [135.244.0.142]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2ENaqNL008346; Mon, 14 Mar 2011 18:36:53 -0500 (CDT)
Message-ID: <4D7EA694.4000308@alcatel-lucent.com>
Date: Mon, 14 Mar 2011 18:36:52 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Worley, Dale R (Dale)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com>	<4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com>	<4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com>	, <4D77E128.8060302@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090201040702040308050200"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 14 Mar 2011 23:35:40 -0000

This is a multi-part message in MIME format.
--------------090201040702040308050200
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dale,

On 14-Mar-2011 17:20, Worley, Dale R (Dale) wrote:
>> From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]
>>
>> If you have read my previous comments, we suggest there is no
>> algorithm per se: this is a measurement of the resource: If you
>> measure the CPU and the Memory (for example), you'll take the
>> smallest. It is better to smooth (with a 10s sliding window average
>> for instance) the number.
> But "If you measure the CPU and the Memory (for example), you'll take
> the smallest" *is* an algorithm.  Its name is "minimum".  There are an
> infinite number of other algorithms.  And you do not state in the
> draft that you've chosen "minimum".
I agree that we need to clarify.
Memory and CPU are just typical examples of constrain resources. The 
draft should probably say:
<< The availability of the key Resources required to setup and maintain 
SIP sessions are measured in terms of available ratios.
The weight represents the MIN of those ratios. >>
> Also, the particular smoothing chosen is not trivial, and should be
> stated in the draft.
But the issue is that it may depends on the applications and situations 
(which is part of your point below).
For instance, in our case we found that a flat average sliding window 
over the last 10 seconds works well. But would it be always the case ?
May be an ascending weights average sliding window of 20s may work 
better in some other cases ?
>> This is not what we have in mind: the weight should be your relative
>> resource usage. (0 - 100%). Using the proposed algorithm, this insure
>> that servers reach full capacity simultaneously (even when they have
>> vastly different capacities).
> Actually, you want the weight to be (or to be proportional) to the
> *unused* portion of your resources, so that an entirely unloaded
> system has maximum weight and a totally loaded system has zero weight.
Yes, absolutely.
> But either way, you don't state this in the draft.
>
> And now I realize you sneaked one past me.  The draft refers to RFC
> 2782 in a number of places, and I didn't recognize it.  You should say
> RFC 3263, which is what people who've implemented SIP recognize.
Actually RFC 3263 refers to RFC 2782, so those people should be familiar 
with RFC 3263.
>   And
> the implication is to replace, in the RFC 3263 process, the "weight"
> values in the SRV records with the "weight" values fed back from the
> servers.
We can do that, but 3263 says very little about weights. The explanation 
really is in RFC 3263
>> Yes, this is true. We are successfully using the simple algorithm
>> mentioned with DNS (where we update the weight every
>> minute). Otherwise they are algorithms that can be used to improve the
>> oscillation in the feedback chain command. We think that since we are
>> proposing a default one that works, this should be OK.
> I'd like to see some analysis and modeling to show that this
> particular updating strategy works in a wide range of operating
> situations.  Otherwise ... unpleasant ... surprises may happen in the
> field.
Moving forward, we'll try to do some modeling and simulations for 
different types of situations.
Otherwise, what is the "particular" updating strategy you mentioned: The 
fact that we use an existing  SIP message to convey the feedback ?
>>> But I don't see any of the needed analysis in the draft, and I
>>> don't see any indication in the draft of a realization that this
>>> analysis needs to be done.  Until that is fixed, I predict that a
>>> solution will not be found, or worse, a solution that appears to
>>> work on paper but does not work in practice.
>> This solution works in practice: we are using it in ALU between
>> components and nodes that are in the same NE (we currently use a
>> proprietary transport mechanism to return the weights).
> You mean it works in your one situation.
Inside one NE, we have a variable set of different types of SIP servers: 
statefull proxies, stateless proxies, Application servers, servers that 
are exposed to the users, etc.. This may help to build some confidence.
It also works between NEs with DNS, since we are updating the DNS 
weights in SRV the same way.

But yes, we' ve not proved that it could work in any situation.

Cordially,

Thierry
> Dale
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------090201040702040308050200
Content-Type: multipart/related;
 boundary="------------070506090003090604010206"


--------------070506090003090604010206
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Dale, 

    

    On 14-Mar-2011 17:20, Worley, Dale R (Dale) wrote:
    

      

        
From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com] 

If you have read my previous comments, we suggest there is no
algorithm per se: this is a measurement of the resource: If you
measure the CPU and the Memory (for example), you'll take the
smallest. It is better to smooth (with a 10s sliding window average
for instance) the number.


      

      
But "If you measure the CPU and the Memory (for example), you'll take
the smallest" *is* an algorithm.  Its name is "minimum".  There are an
infinite number of other algorithms.  And you do not state in the
draft that you've chosen "minimum".


    

    I agree that we need to clarify. 

    Memory and CPU are just typical examples of constrain resources. The
    draft should probably say: 

    << The availability of the key Resources required to setup and
    maintain SIP sessions are measured in terms of available ratios. 

    The weight represents the MIN of those ratios. >> 

    

      
Also, the particular smoothing chosen is not trivial, and should be
stated in the draft.


    

    But the issue is that it may depends on the applications and
    situations (which is part of your point below). 

    For instance, in our case we found that a flat average sliding
    window over the last 10 seconds works well. But would it be always
    the case ? 

    May be an ascending weights average sliding window of 20s may work
    better in some other cases ? 

    

      

        
This is not what we have in mind: the weight should be your relative
resource usage. (0 - 100%). Using the proposed algorithm, this insure
that servers reach full capacity simultaneously (even when they have
vastly different capacities).


      

      
Actually, you want the weight to be (or to be proportional) to the
*unused* portion of your resources, so that an entirely unloaded
system has maximum weight and a totally loaded system has zero weight.


    

    Yes, absolutely. 

    

      
But either way, you don't state this in the draft.

And now I realize you sneaked one past me.  The draft refers to RFC
2782 in a number of places, and I didn't recognize it.  You should say
RFC 3263, which is what people who've implemented SIP recognize. 

    

    Actually RFC 3263 refers to RFC 2782, so those people should be
    familiar with RFC 3263. 

    

      
 And
the implication is to replace, in the RFC 3263 process, the "weight"
values in the SRV records with the "weight" values fed back from the
servers.


    

    We can do that, but 3263 says very little about weights. The
    explanation really is in RFC 3263
    

      

        
Yes, this is true. We are successfully using the simple algorithm
mentioned with DNS (where we update the weight every
minute). Otherwise they are algorithms that can be used to improve the
oscillation in the feedback chain command. We think that since we are
proposing a default one that works, this should be OK.


      

      
I'd like to see some analysis and modeling to show that this
particular updating strategy works in a wide range of operating
situations.  Otherwise ... unpleasant ... surprises may happen in the
field.


    

    Moving forward, we'll try to do some modeling and simulations for
    different types of situations. 

    Otherwise, what is the "particular" updating strategy you mentioned:
    The fact that we use an existing  SIP message to convey the feedback
    ? 

    

      

        

          
But I don't see any of the needed analysis in the draft, and I
don't see any indication in the draft of a realization that this
analysis needs to be done.  Until that is fixed, I predict that a
solution will not be found, or worse, a solution that appears to
work on paper but does not work in practice.


        

        
This solution works in practice: we are using it in ALU between
components and nodes that are in the same NE (we currently use a
proprietary transport mechanism to return the weights).


      

      
You mean it works in your one situation.


    

    Inside one NE, we have a variable set of different types of SIP
    servers: statefull proxies, stateless proxies, Application servers,
    servers that are exposed to the users, etc.. This may help to build
    some confidence. 

    It also works between NEs with DNS, since we are updating the DNS
    weights in SRV the same way. 

    

    But yes, we' ve not proved that it could work in any situation.  

    

    Cordially, 

    

    Thierry

    

      
Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis   
        - Yahoo IM:    
        thierry_bessis

      

      My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

        My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------070506090003090604010206
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------070506090003090604010206--

--------------090201040702040308050200--

From dean.willis@softarmor.com  Mon Mar 14 21:35:46 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4B8C53A68B5 for ; Mon, 14 Mar 2011 21:35:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -103.585
X-Spam-Level: 
X-Spam-Status: No, score=-103.585 tagged_above=-999 required=5 tests=[AWL=0.014, BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GojW3CvevW99 for ; Mon, 14 Mar 2011 21:35:44 -0700 (PDT)
Received: from mail-yx0-f172.google.com (mail-yx0-f172.google.com [209.85.213.172]) by core3.amsl.com (Postfix) with ESMTP id 2A1413A6950 for ; Mon, 14 Mar 2011 21:35:44 -0700 (PDT)
Received: by yxk30 with SMTP id 30so105855yxk.31 for ; Mon, 14 Mar 2011 21:37:08 -0700 (PDT)
Received: by 10.150.103.10 with SMTP id a10mr6747261ybc.417.1300163828251; Mon, 14 Mar 2011 21:37:08 -0700 (PDT)
Received: from [192.168.2.149] (cpe-66-25-2-124.tx.res.rr.com [66.25.2.124]) by mx.google.com with ESMTPS id q29sm5931597yba.2.2011.03.14.21.37.07 (version=TLSv1/SSLv3 cipher=OTHER); Mon, 14 Mar 2011 21:37:07 -0700 (PDT)
References: <4D4425C8.2080705@jdrosen.net>
In-Reply-To: <4D4425C8.2080705@jdrosen.net>
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
Message-Id: <7CB0D0EE-0D7A-405E-8C9B-0D4774723B43@softarmor.com>
Content-Transfer-Encoding: quoted-printable
From: Dean Willis 
Date: Mon, 14 Mar 2011 23:37:06 -0500
To: Jonathan Rosenberg 
X-Mailer: Apple Mail (2.1082)
Cc: rtc-web@alvestrand.no, 'DISPATCH list' 
Subject: Re: [dispatch] Does RTC-WEB need to pick a signaling protocol?
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 04:35:46 -0000

Oh hell no....

Really, what we need are the primitives that make a web app able to do =
things like slurp data from a microphone and send it someplace useful. =
We might also need interfaces to A/V streams (maybe even RTP), and =
mixers. We need those primitives exposed to things like Javascript. What =
we don't need is an API that turns every browser into a generic SIP UA =
capable of being reused in haphazard fashion, such that everybody and =
his chihuahua reskins their "SIP browser phone" with CSS, and every =
browser swells to the size of the Hindenberg just before the fatal =
spark.

--
Dean


From dean.willis@softarmor.com  Mon Mar 14 21:46:12 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B05403A6AB9 for ; Mon, 14 Mar 2011 21:46:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -103.586
X-Spam-Level: 
X-Spam-Status: No, score=-103.586 tagged_above=-999 required=5 tests=[AWL=0.013, BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RQr159-583tW for ; Mon, 14 Mar 2011 21:46:11 -0700 (PDT)
Received: from mail-gw0-f44.google.com (mail-gw0-f44.google.com [74.125.83.44]) by core3.amsl.com (Postfix) with ESMTP id C41223A6AB4 for ; Mon, 14 Mar 2011 21:46:11 -0700 (PDT)
Received: by gwb20 with SMTP id 20so105973gwb.31 for ; Mon, 14 Mar 2011 21:47:35 -0700 (PDT)
Received: by 10.91.96.2 with SMTP id y2mr4099246agl.179.1300164455779; Mon, 14 Mar 2011 21:47:35 -0700 (PDT)
Received: from [192.168.2.149] (cpe-66-25-2-124.tx.res.rr.com [66.25.2.124]) by mx.google.com with ESMTPS id 35sm1405063ano.11.2011.03.14.21.47.34 (version=TLSv1/SSLv3 cipher=OTHER); Mon, 14 Mar 2011 21:47:35 -0700 (PDT)
References: <4D4425C8.2080705@jdrosen.net> <7CB0D0EE-0D7A-405E-8C9B-0D4774723B43@softarmor.com>
In-Reply-To: <7CB0D0EE-0D7A-405E-8C9B-0D4774723B43@softarmor.com>
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
Message-Id: <57E0260F-2980-49CB-A9DA-8BF0F378932E@softarmor.com>
Content-Transfer-Encoding: quoted-printable
From: Dean Willis 
Date: Mon, 14 Mar 2011 23:47:33 -0500
To: Dean Willis 
X-Mailer: Apple Mail (2.1082)
Cc: rtc-web@alvestrand.no, DISPATCH list 
Subject: Re: [dispatch] Does RTC-WEB need to pick a signaling protocol?
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 04:46:12 -0000

Well, THAT was an interesting time-warp.  Sorry about that. But I guess =
the sentiment still stands even if the mail was stuck in my outbox for a =
while...

On Mar 14, 2011, at 11:37 PM, Dean Willis wrote:

>=20
> Oh hell no....
>=20
> Really, what we need are the primitives that make a web app able to do =
things like slurp data from a microphone and send it someplace useful. =
We might also need interfaces to A/V streams (maybe even RTP), and =
mixers. We need those primitives exposed to things like Javascript. What =
we don't need is an API that turns every browser into a generic SIP UA =
capable of being reused in haphazard fashion, such that everybody and =
his chihuahua reskins their "SIP browser phone" with CSS, and every =
browser swells to the size of the Hindenberg just before the fatal =
spark.
>=20
> --
> Dean
>=20


From bernard_aboba@hotmail.com  Mon Mar 14 23:11:52 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B900A3A6AAB for ; Mon, 14 Mar 2011 23:11:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.461
X-Spam-Level: 
X-Spam-Status: No, score=-102.461 tagged_above=-999 required=5 tests=[AWL=0.137, BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id l6c4TorGucFK for ; Mon, 14 Mar 2011 23:11:33 -0700 (PDT)
Received: from blu0-omc2-s1.blu0.hotmail.com (blu0-omc2-s1.blu0.hotmail.com [65.55.111.76]) by core3.amsl.com (Postfix) with ESMTP id B66AE3A6AEE for ; Mon, 14 Mar 2011 23:11:28 -0700 (PDT)
Received: from BLU152-W21 ([65.55.111.71]) by blu0-omc2-s1.blu0.hotmail.com with Microsoft SMTPSVC(6.0.3790.4675);  Mon, 14 Mar 2011 23:12:43 -0700
Message-ID: 
Content-Type: multipart/alternative; boundary="_8d8344db-a323-4cd0-a2f1-46ae0c52ac4f_"
X-Originating-IP: [98.203.198.61]
From: Bernard Aboba 
To: , 
Date: Mon, 14 Mar 2011 23:12:43 -0700
Importance: Normal
In-Reply-To: <57E0260F-2980-49CB-A9DA-8BF0F378932E@softarmor.com>
References: <4D4425C8.2080705@jdrosen.net>, <7CB0D0EE-0D7A-405E-8C9B-0D4774723B43@softarmor.com>, <57E0260F-2980-49CB-A9DA-8BF0F378932E@softarmor.com>
MIME-Version: 1.0
X-OriginalArrivalTime: 15 Mar 2011 06:12:43.0750 (UTC) FILETIME=[FF31F060:01CBE2D7]
Subject: Re: [dispatch] [RTW] Does RTC-WEB need to pick a signaling protocol?
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 06:11:52 -0000

--_8d8344db-a323-4cd0-a2f1-46ae0c52ac4f_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable


No problem.  The message is indeed timeless.  Feel free to resend it every =
week or so =3B)

> From: dean.willis@softarmor.com
> Date: Mon=2C 14 Mar 2011 23:47:33 -0500
> To: dean.willis@softarmor.com
> CC: jdrosen@jdrosen.net=3B rtc-web@alvestrand.no=3B dispatch@ietf.org
> Subject: Re: [RTW] [dispatch] Does RTC-WEB need to pick a signaling proto=
col?
>=20
>=20
> Well=2C THAT was an interesting time-warp.  Sorry about that. But I guess=
 the sentiment still stands even if the mail was stuck in my outbox for a w=
hile...
>=20
> On Mar 14=2C 2011=2C at 11:37 PM=2C Dean Willis wrote:
>=20
> >=20
> > Oh hell no....
> >=20
> > Really=2C what we need are the primitives that make a web app able to d=
o things like slurp data from a microphone and send it someplace useful. We=
 might also need interfaces to A/V streams (maybe even RTP)=2C and mixers. =
We need those primitives exposed to things like Javascript. What we don't n=
eed is an API that turns every browser into a generic SIP UA capable of bei=
ng reused in haphazard fashion=2C such that everybody and his chihuahua res=
kins their "SIP browser phone" with CSS=2C and every browser swells to the =
size of the Hindenberg just before the fatal spark.
> >=20
> > --
> > Dean
> >=20
>=20
> _______________________________________________
> RTC-Web mailing list
> RTC-Web@alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/rtc-web
 		 	   		  =

--_8d8344db-a323-4cd0-a2f1-46ae0c52ac4f_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable






No problem. =3B The message is indeed timeless. =3B Feel free to re=
send it every week or so =3B)

>=3B From: dean.willis@softarmor.com=

>=3B Date: Mon=2C 14 Mar 2011 23:47:33 -0500
>=3B To: dean.willi=
s@softarmor.com
>=3B CC: jdrosen@jdrosen.net=3B rtc-web@alvestrand.no=
=3B dispatch@ietf.org
>=3B Subject: Re: [RTW] [dispatch] Does RTC-WEB =
need to pick a signaling protocol?
>=3B 
>=3B 
>=3B Well=2C =
THAT was an interesting time-warp.  Sorry about that. But I guess the senti=
ment still stands even if the mail was stuck in my outbox for a while...>=3B 
>=3B On Mar 14=2C 2011=2C at 11:37 PM=2C Dean Willis wrote:>=3B 
>=3B >=3B 
>=3B >=3B Oh hell no....
>=3B >=
=3B 
>=3B >=3B Really=2C what we need are the primitives that make a=
 web app able to do things like slurp data from a microphone and send it so=
meplace useful. We might also need interfaces to A/V streams (maybe even RT=
P)=2C and mixers. We need those primitives exposed to things like Javascrip=
t. What we don't need is an API that turns every browser into a generic SIP=
 UA capable of being reused in haphazard fashion=2C such that everybody and=
 his chihuahua reskins their "SIP browser phone" with CSS=2C and every brow=
ser swells to the size of the Hindenberg just before the fatal spark.
&g=
t=3B >=3B 
>=3B >=3B --
>=3B >=3B Dean
>=3B >=3B >=3B 
>=3B _______________________________________________
>=
=3B RTC-Web mailing list
>=3B RTC-Web@alvestrand.no
>=3B http://w=
ww.alvestrand.no/mailman/listinfo/rtc-web
 		 	   		  
=

--_8d8344db-a323-4cd0-a2f1-46ae0c52ac4f_--

From fluffy@cisco.com  Tue Mar 15 10:41:01 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 446143A6DAF for ; Tue, 15 Mar 2011 10:41:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -108.062
X-Spam-Level: 
X-Spam-Status: No, score=-108.062 tagged_above=-999 required=5 tests=[AWL=2.538, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HFXYsn4cjH1s for ; Tue, 15 Mar 2011 10:41:00 -0700 (PDT)
Received: from sj-iport-3.cisco.com (sj-iport-3.cisco.com [171.71.176.72]) by core3.amsl.com (Postfix) with ESMTP id A0ACA3A6971 for ; Tue, 15 Mar 2011 10:41:00 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=483; q=dns/txt; s=iport; t=1300210946; x=1301420546; h=from:content-transfer-encoding:subject:date:message-id: cc:to:mime-version; bh=8BnuPmMe3DQXMBcPeIiepsAnL8xIpu3K5QCAAXdG8EY=; b=jByWvPZx9/K9dftdvTW2sbyeAd0Gz6jxKjNZTZlzgzpQEx9z9Yq4vBie gxbPIGK5WoanROaZBqc8jZemMfwJNobffFq/xNbiKunlfMmeFXGBKGpKV KjGn/Ltqin9zbtRw81o07wFqGiJeYZYbZHQ5/7wgqUJdMosnk39OhuXQY s=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAElCf02tJV2Z/2dsb2JhbACmCHelcZ0GhWIEhTCHLYNP
X-IronPort-AV: E=Sophos;i="4.63,188,1299456000"; d="scan'208";a="278265398"
Received: from rcdn-core-2.cisco.com ([173.37.93.153]) by sj-iport-3.cisco.com with ESMTP; 15 Mar 2011 17:42:25 +0000
Received: from [192.168.4.100] (rcdn-fluffy-8712.cisco.com [10.99.9.19]) by rcdn-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2FHgPD4016592;  Tue, 15 Mar 2011 17:42:25 GMT
From: Cullen Jennings 
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Date: Tue, 15 Mar 2011 11:45:13 -0600
Message-Id: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com>
To: DISPATCH list 
Mime-Version: 1.0 (Apple Message framework v1082)
X-Mailer: Apple Mail (2.1082)
Subject: [dispatch] DISPATCH Ad-HOC meetings about VIPR and RTCWeb at IETF 80
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 17:41:01 -0000

The dispatch WG will have two ad-hoc meeting.=20

The first is discuss the technical content of the drafts related to the =
proposed VIPR WG.=20

Tentative time: VIPR's is in Tyrolka on Thursday 1300-1400

The second is to discuss the technical contents of drafts related to the =
proposed RTCWeb WG.=20

Tentative time:  RTCWeb's is  Friday 1415-1515 in the Grand Ballroom

Send Agenda requests for time in either of these to Mary and I.=20

Thanks,
Mary & Cullen



From dworley@avaya.com  Tue Mar 15 13:06:11 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5B0A13A6EE4 for ; Tue, 15 Mar 2011 13:06:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.542
X-Spam-Level: 
X-Spam-Status: No, score=-102.542 tagged_above=-999 required=5 tests=[AWL=0.057, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id B6nNu8q2hRAD for ; Tue, 15 Mar 2011 13:06:09 -0700 (PDT)
Received: from de307622-de-outbound.net.avaya.com (de307622-de-outbound.net.avaya.com [198.152.71.100]) by core3.amsl.com (Postfix) with ESMTP id 1CD6F3A6D9A for ; Tue, 15 Mar 2011 13:06:08 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU3GmAcF/2dsb2JhbACmZXSkeQKZFoVhBJAM
X-IronPort-AV: E=Sophos;i="4.63,190,1299474000"; d="scan'208";a="236908125"
Received: from unknown (HELO co300216-co-erhwest.avaya.com) ([198.152.7.5]) by de307622-de-outbound.net.avaya.com with ESMTP; 15 Mar 2011 16:07:33 -0400
X-IronPort-AV: E=Sophos;i="4.63,190,1299474000"; d="scan'208";a="595996714"
Received: from dc-us1hcex2.us1.avaya.com (HELO DC-US1HCEX2.global.avaya.com) ([135.11.52.21]) by co300216-co-erhwest-out.avaya.com with ESMTP; 15 Mar 2011 16:07:32 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX2.global.avaya.com ([::1]) with mapi; Tue, 15 Mar 2011 16:07:32 -0400
From: "Worley, Dale R (Dale)" 
To: Thierry Bessis 
Date: Tue, 15 Mar 2011 16:06:22 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvioLcg/OD66Wo4RA6DOQcn/kXv3wAq72pu
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,<4D76C924.7050900@cisco.com> , <4D77E128.8060302@alcatel-lucent.com> , <4D7EA694.4000308@alcatel-lucent.com>
In-Reply-To: <4D7EA694.4000308@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 20:06:11 -0000

________________________________________
From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]

Inside one NE, we have a variable set of different types of SIP servers: st=
atefull proxies, stateless proxies, Application servers, servers that are e=
xposed to the users, etc.. This may help to build some confidence.
_______________________________________________

Can you describe the benefits you have seen in your cluster (Network Elemen=
t) from load balancing?

Dale

From HKaplan@acmepacket.com  Tue Mar 15 13:33:08 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CF1EC3A6808 for ; Tue, 15 Mar 2011 13:33:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.495
X-Spam-Level: 
X-Spam-Status: No, score=-0.495 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553,  RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 8952I8FQYD1H for ; Tue, 15 Mar 2011 13:33:08 -0700 (PDT)
Received: from ETMail2.acmepacket.com (unknown [216.41.24.9]) by core3.amsl.com (Postfix) with ESMTP id ED1373A691E for ; Tue, 15 Mar 2011 13:33:07 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by ETMail2.acmepacket.com (216.41.24.9) with Microsoft SMTP Server (TLS) id 8.1.240.5; Tue, 15 Mar 2011 16:34:32 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Tue, 15 Mar 2011 16:34:32 -0400
From: Hadriel Kaplan 
To: Thierry Bessis 
Date: Tue, 15 Mar 2011 16:34:30 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvjUGNYKqiHHthoT4a3tpVYt6OVTw==
Message-ID: <4B0458E2-D1D3-4B0D-99E2-E614C4686043@acmepacket.com>
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,	<4D76C924.7050900@cisco.com> , <4D77E128.8060302@alcatel-lucent.com>  <4D7EA694.4000308@alcatel-lucent.com>
In-Reply-To: <4D7EA694.4000308@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAQAAAUA=
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 20:33:09 -0000

On Mar 14, 2011, at 7:36 PM, Thierry Bessis wrote:

> Inside one NE, we have a variable set of different types of SIP servers: =
statefull proxies, stateless proxies, Application servers, servers that are=
 exposed to the users, etc.. This may help to build     some confidence.=20
> It also works between NEs with DNS, since we are updating the DNS weights=
 in SRV the same way.=20
>=20
> But yes, we' ve not proved that it could work in any situation. =20

We've used the same type of solution as well, but it doesn't work in all si=
tuations (or at least not the same way).  In particular, when load-balancin=
g a farm of P-CSCF's, S-CSCF's or subscriber-access SBC's, you have to load=
-balance *subscribers* not SIP messages/transactions/calls.  In other words=
, the same subscriber/endpoint had better have all of its SIP messages go t=
o the same P-CSCF/S-CSCF/SBC, or all hell breaks loose. (there're a couple =
vendors who claim in powerpoint their *-CSCF/SBC's don't need this, but in =
practice you do, if for no other reason than troubleshooting and monitoring=
)  And since subscriber capacity/utilization is a different resource measur=
ement from CPU utilization, both sides have to be in on the game and know w=
hat mode they're in. (not that's that's a big deal)

But on sip-trunks and peering links you can load-balance per session/out-of=
-dialog-message this way.  It's just not the same thing as session/call con=
straints, which is I think what SCE is trying to get instead.

-hadriel


From Thierry.Bessis@alcatel-lucent.com  Tue Mar 15 14:03:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1E1043A6ED1 for ; Tue, 15 Mar 2011 14:03:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oWuySBE2uRGW for ; Tue, 15 Mar 2011 14:03:57 -0700 (PDT)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 679D23A6EC7 for ; Tue, 15 Mar 2011 14:03:57 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p2FL5LCF021245 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 15 Mar 2011 16:05:21 -0500 (CDT)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2FL5KfC025464; Tue, 15 Mar 2011 16:05:20 -0500 (CDT)
Message-ID: <4D7FD490.1060109@alcatel-lucent.com>
Date: Tue, 15 Mar 2011 16:05:20 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Worley, Dale R (Dale)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com>	<4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com>	<4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com>	, <4D77E128.8060302@alcatel-lucent.com> , <4D7EA694.4000308@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------030604060005030208010305"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 21:03:59 -0000

This is a multi-part message in MIME format.
--------------030604060005030208010305
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dale,

What do you mean by "from" load balancing ?

Our main benefit is that our resource utilization (essentially CPU and 
Memory as we discussed) is very good. We see that the system effectively 
load the cluster evenly, and returns to its equilibrium even after some 
"disrupting" events occurred. Such event are for example:  temporary 
loss and then return of a node, or a temporary overload of some nodes 
due to administrative work, or even the dynamic installation of some new 
components on existing hosts. Also, we are dealing with remote nodes 
which are NOT getting (or not willing to use)  the weights and may 
concentrate the traffic unevenly. Then, our mechanism will fix the issue 
using the "good behavior" of nodes which are using the weights to 
re-balance the cluster.

Since all our nodes get overloaded at roughly the same time, we know we 
have roughly the best possible throughput. In addition, based on this 
knowledge, we decide to never re-transmit any message in the same 
cluster when we get the "shedding" signal back: This is a huge advantage.

Does that answer your question ?

Cordially,

Thierry


On 15-Mar-2011 15:06, Worley, Dale R (Dale) wrote:
> ________________________________________
> From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]
>
> Inside one NE, we have a variable set of different types of SIP servers: statefull proxies, stateless proxies, Application servers, servers that are exposed to the users, etc.. This may help to build some confidence.
> _______________________________________________
>
> Can you describe the benefits you have seen in your cluster (Network Element) from load balancing?
>
> Dale

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------030604060005030208010305
Content-Type: multipart/related;
 boundary="------------050306080905070801010305"


--------------050306080905070801010305
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Dale, 

    

    What do you mean by "from" load balancing ? 

    

    Our main benefit is that our resource utilization (essentially CPU
    and Memory as we discussed) is very good. We see that the system
    effectively load the cluster evenly, and returns to its equilibrium
    even after some "disrupting" events occurred. Such event are for
    example:  temporary loss and then return of a node, or a temporary
    overload of some nodes due to administrative work, or even the
    dynamic installation of some new components on existing hosts. Also,
    we are dealing with remote nodes which are NOT getting (or not
    willing to use)  the weights and may concentrate the traffic
    unevenly. Then, our mechanism will fix the issue using the "good
    behavior" of nodes which are using the weights to re-balance the
    cluster. 

    

    Since all our nodes get overloaded at roughly the same time, we know
    we have roughly the best possible throughput. In addition, based on
    this knowledge, we decide to never re-transmit any message in the
    same cluster when we get the "shedding" signal back: This is a huge
    advantage. 

    

    Does that answer your question ?

    

    Cordially, 

    

    Thierry

    

    

    On 15-Mar-2011 15:06, Worley, Dale R (Dale) wrote:
    

      
________________________________________
From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]

Inside one NE, we have a variable set of different types of SIP servers: statefull proxies, stateless proxies, Application servers, servers that are exposed to the users, etc.. This may help to build some confidence.
_______________________________________________

Can you describe the benefits you have seen in your cluster (Network Element) from load balancing?

Dale


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------050306080905070801010305
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------050306080905070801010305--

--------------030604060005030208010305--

From Thierry.Bessis@alcatel-lucent.com  Tue Mar 15 14:43:28 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EC6CB3A6B57 for ; Tue, 15 Mar 2011 14:43:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0bK504zn583x for ; Tue, 15 Mar 2011 14:43:27 -0700 (PDT)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 385BE3A6E4D for ; Tue, 15 Mar 2011 14:43:26 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p2FLiolM006451 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 15 Mar 2011 16:44:50 -0500 (CDT)
Received: from [135.185.234.200] (tbessis-c1.ih.lucent.com [135.185.234.200]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2FLioED025657; Tue, 15 Mar 2011 16:44:50 -0500 (CDT)
Message-ID: <4D7FDDD2.4050005@alcatel-lucent.com>
Date: Tue, 15 Mar 2011 16:44:50 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Hadriel Kaplan 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com>	<4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com>	<4D76A762.6090708@bell-labs.com>, <4D76C924.7050900@cisco.com>	, <4D77E128.8060302@alcatel-lucent.com>		<4D7EA694.4000308@alcatel-lucent.com> <4B0458E2-D1D3-4B0D-99E2-E614C4686043@acmepacket.com>
In-Reply-To: <4B0458E2-D1D3-4B0D-99E2-E614C4686043@acmepacket.com>
Content-Type: multipart/alternative; boundary="------------000402000404080603020006"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 15 Mar 2011 21:43:29 -0000

This is a multi-part message in MIME format.
--------------000402000404080603020006
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Kaplan,

On 15-Mar-2011 15:34, Hadriel Kaplan wrote:
> On Mar 14, 2011, at 7:36 PM, Thierry Bessis wrote:
>
>> Inside one NE, we have a variable set of different types of SIP servers: statefull proxies, stateless proxies, Application servers, servers that are exposed to the users, etc.. This may help to build     some confidence.
>> It also works between NEs with DNS, since we are updating the DNS weights in SRV the same way.
>>
>> But yes, we' ve not proved that it could work in any situation.
> We've used the same type of solution as well, but it doesn't work in all situations (or at least not the same way).  In particular, when load-balancing a farm of P-CSCF's, S-CSCF's or subscriber-access SBC's, you have to load-balance *subscribers* not SIP messages/transactions/calls.  In other words, the same subscriber/endpoint had better have all of its SIP messages go to the same P-CSCF/S-CSCF/SBC, or all hell breaks loose. (there're a couple vendors who claim in powerpoint their *-CSCF/SBC's don't need this, but in practice you do, if for no other reason than troubleshooting and monitoring)  And since subscriber capacity/utilization is a different resource measurement from CPU utilization, both sides have to be in on the game and know what mode they're in. (not that's that's a big deal)
This is a good point. We are doing the same thing as you describe: we 
have a farm of P-CSCF and the subcriber/endpoint MUST go to the same 
P-CSCF node once it is registered. But we still successfully use the 
same principle. We have multiple types of settings: P-CSCF and other 
CSCF (like I-CSCF, ...  S-CSCF) running on the same host (and sharing 
the resources), or hosts that are used only for P-CSCF.
It works in both cases because we use the Memory also in the weight. The 
memory is sensitive to both the number of simultaneous instances, and 
the number of registrations. In components that maintains registries 
(such as P-CSCF), additional memory will be consumed by the 
registrations and the memory will become the leading factor in the weights.

So basically what happens is that the weight in DNS will reflect the 
current number of subscribers, and less subscribers will register to 
P-CSCF that have more registrations. (I'm talking about DNS here, 
because of course our SIP load balancing feedback technique will not 
work with subscribers, nor would SCE ! ).

Some people have argued that this is not optimal, since you may have lot 
of CPU available and very few memory, so you could process lot of 
transactions but still your weight is low, so you are not utilizing you 
CPU wisely. In practice though it is still optimal: statistically call 
volumes and registrations go together: we are using such big capacity 
numbers that this is verified.

Cordially,

Thierry

> But on sip-trunks and peering links you can load-balance per session/out-of-dialog-message this way.  It's just not the same thing as session/call constraints, which is I think what SCE is trying to get instead.
>
> -hadriel
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------000402000404080603020006
Content-Type: multipart/related;
 boundary="------------040204090401030702090403"


--------------040204090401030702090403
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Kaplan, 

    

    On 15-Mar-2011 15:34, Hadriel Kaplan wrote:
    

      
On Mar 14, 2011, at 7:36 PM, Thierry Bessis wrote:



      

        
Inside one NE, we have a variable set of different types of SIP servers: statefull proxies, stateless proxies, Application servers, servers that are exposed to the users, etc.. This may help to build     some confidence. 
It also works between NEs with DNS, since we are updating the DNS weights in SRV the same way. 

But yes, we' ve not proved that it could work in any situation.  


      

      
We've used the same type of solution as well, but it doesn't work in all situations (or at least not the same way).  In particular, when load-balancing a farm of P-CSCF's, S-CSCF's or subscriber-access SBC's, you have to load-balance *subscribers* not SIP messages/transactions/calls.  In other words, the same subscriber/endpoint had better have all of its SIP messages go to the same P-CSCF/S-CSCF/SBC, or all hell breaks loose. (there're a couple vendors who claim in powerpoint their *-CSCF/SBC's don't need this, but in practice you do, if for no other reason than troubleshooting and monitoring)  And since subscriber capacity/utilization is a different resource measurement from CPU utilization, both sides have to be in on the game and know what mode they're in. (not that's that's a big deal)


    

    This is a good point. We are doing the same thing as you describe:
    we have a farm of P-CSCF and the subcriber/endpoint MUST go to the
    same P-CSCF node once it is registered. But we still successfully
    use the same principle. We have multiple types of settings: P-CSCF
    and other CSCF (like I-CSCF, ...  S-CSCF) running on the same host
    (and sharing the resources), or hosts that are used only for P-CSCF.
    

    It works in both cases because we use the Memory also in the weight.
    The memory is sensitive to both the number of simultaneous
    instances, and the number of registrations. In components that
    maintains registries (such as P-CSCF), additional memory will be
    consumed by the registrations and the memory will become the leading
    factor in the weights.

    

    So basically what happens is that the weight in DNS will reflect the
    current number of subscribers, and less subscribers will register to
    P-CSCF that have more registrations. (I'm talking about DNS here,
    because of course our SIP load balancing feedback technique will not
    work with subscribers, nor would SCE ! ).  

    

    Some people have argued that this is not optimal, since you may have
    lot of CPU available and very few memory, so you could process lot
    of transactions but still your weight is low, so you are not
    utilizing you CPU wisely. In practice though it is still optimal:
    statistically call volumes and registrations go together: we are
    using such big capacity numbers that this is verified.

    

    Cordially, 

    

    Thierry

    

    

      
But on sip-trunks and peering links you can load-balance per session/out-of-dialog-message this way.  It's just not the same thing as session/call constraints, which is I think what SCE is trying to get instead.

-hadriel

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------040204090401030702090403
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------040204090401030702090403--

--------------000402000404080603020006--

From partr@cisco.com  Wed Mar 16 06:16:43 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B0AC43A695B for ; Wed, 16 Mar 2011 06:16:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.259
X-Spam-Level: 
X-Spam-Status: No, score=-9.259 tagged_above=-999 required=5 tests=[AWL=-0.262, BAYES_00=-2.599, EXTRA_MPART_TYPE=1, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 86Tee42Cu46G for ; Wed, 16 Mar 2011 06:16:41 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id B693A3A693D for ; Wed, 16 Mar 2011 06:16:39 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=67202; q=dns/txt; s=iport; t=1300281486; x=1301491086; h=mime-version:subject:date:message-id:in-reply-to: references:from:to:cc; bh=rerODS46P8W5Og5kwAMeXd5dapVYMKcG9G+VX75zkzk=; b=DFrchLj6dn09w5ZnWMjuDDjyL043nWZSaPXkAUNskmZTxDcKnwG21p24 EQhZ5wk2wIKdeqe/nKpruBqEOdd7A9RuGH4khbAx3s0K5z1stptPTARPE zHkLq91/uNeJX6/koXdRD/0ZlMFj1K4i8JEk0kE7m0l1qHLB2l9iOzkBg Q=;
X-Files: al_corp_h_3d_16mm.jpg : 32899
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AikCAKNVgE2Q/khNgWdsb2JhbAAwgVuWRo03FAEBFiYlpTWcUYMVgk0EgXWDOIsE
X-IronPort-AV: E=Sophos;i="4.63,194,1299456000";  d="jpg'145?scan'145,208,217,145";a="79448488"
Received: from ams-core-4.cisco.com ([144.254.72.77]) by ams-iport-1.cisco.com with ESMTP; 16 Mar 2011 13:18:05 +0000
Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by ams-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2GDI2nX029854; Wed, 16 Mar 2011 13:18:04 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Wed, 16 Mar 2011 18:48:04 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----_=_NextPart_001_01CBE3DC.94AA21ED"; type="multipart/alternative"
Date: Wed, 16 Mar 2011 18:48:02 +0530
Message-ID: 
In-Reply-To: <4D7A485E.2080602@alcatel-lucent.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvgBjiZj9Yvwr1/RSitdsC3TglPvwD01eyA
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>
From: "Parthasarathi R (partr)" 
To: "Thierry Bessis" 
X-OriginalArrivalTime: 16 Mar 2011 13:18:04.0093 (UTC) FILETIME=[94EAFED0:01CBE3DC]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 13:16:43 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBE3DC.94AA21ED
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_002_01CBE3DC.94AA21ED"


------_=_NextPart_002_01CBE3DC.94AA21ED
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Thierry,
=20
Please read inline.
=20
Thanks
Partha

________________________________

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]=20
Sent: Friday, March 11, 2011 9:36 PM
To: Parthasarathi R (partr)
Cc: Paul E. Jones; dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing


Paul,=20


=09
	For instance, with our approach, it is fundamental that the
weight represents a "Relative" occupation, and not an absolute capacity.

	 The question is how to perform the "relative" operation
without until understanding the capacity of the peer devices. Say Load
balancer has 3 devices: 2 devices with 1x capacity and 1 device with 2x
capacity. Here all devices will be able to provide the "relative" value
without understanding the capacity of the other device. 
=09
	The first reason is that we found that loading continuously all
the servers at the same utilization rate (same percentage) is the
optimal way of loading a cluster. (We can discuss that if not agreed).=20
	 I agree that same utilization rate is the optimal way
but the question is how to achieve it. =20
=09

Thierry>> we are on line on this one ! :-)=20


	Round robin algorithm works in case all devices have same
capacity and same amount of load per call.

Thierry >> Yes I agree it would work in this case. The reason we are not
using it is that in the IMS network (where we are already using those
principles), this is not the case: Devices do not have the same capacity
because they are on different hardware, or (more often) because they are
sharing resources with other devices that are doing different things at
different moments. (which provide a huge resource saving !). The result
is that it's impossible for any device to calculate its own capacity "in
isolation": It would need to know exactly which other devices are
sharing the same resources, their resource usage, and what they are
doing at any time: an operational nightmare which would break all the
software isolation rules ! :-) . In addition, we sometime shutdown some
devices and re-enable them later, (or we grow more devices in the
cluster): Round robin does not work at all in this case.=20



	In reality, there are multiple varying factor like call
duration, different vareity of platform within the cluster (say
Generation-1 10 devices and Generation-2 5 devies), different resource
usage for the call, rate at which calls generation. 

Thierry>>  yes !=20


	=20
	The second reason is that a relative occupation is very easy to
derive from information already available: any system knows where it is
in terms of CPU or Memory usage.
	=20
	 Converting CPU or Memory to weight is not explained
your draft. It is the real algorithm which needs lot of discussion.
Because preparing the mapping between CPU & Memory to weight will be
mind-blogging activity 

Thierry>> We could be more explicit, but should we ? (see next
paragraph). Memory and CPU are just the two most obvious examples or
resources that you need to process sessions. For instance, some of our
servers have a fixed amount of sessions registers (for instance 50 000).
In this case, the weight also reflects that. The algorithm we use for
weights is very simple: this is the smallest remaining (in percentage)
of each resource that is (1) Required to process a call, (2) Scarce
(not infinite). Let's consider the following system:=20
The server has 70% of its memory available for calls, an average of 30%
CPU usage (average over the last 10s for instance), and 40% of its
registers available (remember that those resources may be shared by
other components or devices for any other usage: it does not matter).=20
Then the weight at this moment is MIN (70, 30, 40) =3D 30%. You then =
need
to normalize this weight to match the weight range (an INT ? )=20

Now we are not saying that everybody should calculate their weight like
that. What is important is that the weight reflects your current
relative capacity. In our case, this "algorithm" balances the load of
our servers very well. Now if you use a server that has a fixed amount
of resources and that is dimensioned to process let's say 50 000
sessions, at a pace of 100 session/s, then you could probably derive a
weight that would depend on both your count of sessions left, and your
current rate of session/s. =20
=20
 In case every device calculate its own algorithm for weight,
then weight from one device may not same as other device. Say dev1 takes
Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =
=3D
30% and another device dev2 calculates Min (CPU, Mem) then weight will
be Min (70, 40) =3D 40% even though they are equally loaded. It is very
important to have the common unit to solve this issue  =20

	=20
	 In contrast, deriving a left capacity in terms of remaining
sessions can be difficult: Host can be used by different applications
and perform different duties, which can change over the day, all of them
impacting their true capacity at a certain moment. Beside, as already
discussed, different sessions require different resources, so there is a
whole complication of dealing with a multitude of call profiles, that
(again) are not even constant over the day.
=09
	Also, we do not want to derive some shedding decision from the
weight.
	=20
	  In case you specify, IETF SoC does not allow and so we
want another value, I think that we are in same page.

Thierry>> we are on the same page. I thought you suggested to use the
same value. =20
=20

	If not, I could not understand your argument. As you mentioned,
all system will know CPU or memory usage and it is possible to come up
the static or refreshable value saying that load shed has to happen when
85% of CPU utilization or 90% of memory utilization. The different
between our proposal is that there is no need of two value per message
instead 1 value in each message & 1 value is static or in periodic
manner  

Thierry>> We use a single value: see the example above. =20
 I'm proposing is to use single real time value with units like
number of dialog and exchange the threshold (max dialog) using some
other static interface. Let overload decides how to reduce or when to
shed the load based on this information. The advantage in this mechanism
is that only one real time value is required between entities.
 =20
In your proposed solution, "weight" as one value for load balancing and
"oc" as another value for overload shedding. Pls correct me incase I
misunderstand something here =20


--

--=20
Cordially,=20
Thierry Bessis=20
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE,
PdM
=20
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code:
9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None=20

------_=_NextPart_002_01CBE3DC.94AA21ED
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable







Thierry,


 


Please read inline.


 


Thanks


Partha






From: Thierry Bessis=20
[mailto:Thierry.Bessis@alcatel-lucent.com] 
Sent: Friday, =
March 11,=20
2011 9:36 PM
To: Parthasarathi R (partr)
Cc: Paul E. =
Jones;=20
dispatch@ietf.org
Subject: Re: [dispatch] SIP load=20
balancing



Paul, 


  

For =
instance, with our=20
  approach, it is fundamental that the weight represents a "Relative"=20
  occupation, and not an absolute capacity. 

  
<Partha> The question is how to perform the "relative" =
operation=20
  without until understanding the capacity of the peer devices. Say =
Load=20
  balancer has 3 devices: 2 devices with 1x capacity and 1 device =
with 2x=20
  capacity. Here all devices will be able to provide the "relative" =
value=20
  without understanding the capacity of the other=20
  device. </Partha>

  

The first =
reason is that=20
  we found that loading continuously all the servers at the same =
utilization=20
  rate (same percentage) is the optimal way of loading a cluster. (We =
can=20
  discuss that if not agreed). 

  
<Partha> I agree that same utilization rate is the =
optimal way=20
  but the question is how to achieve it. =20

Thierry>> we are on line on this one ! =
:-) 


  
Round robin algorithm works in case all devices have same =
capacity and=20
  same amount of load per call.
Thierry >> Yes I =
agree it=20
would work in this case. The reason we are not using it is that in the =
IMS=20
network (where we are already using those principles), this is not the =
case:=20
Devices do not have the same capacity because they are on different =
hardware, or=20
(more often) because they are sharing resources with other devices that =
are=20
doing different things at different moments. (which provide a huge =
resource=20
saving !). The result is that it's impossible for any device to =
calculate its=20
own capacity "in isolation": It would need to know exactly which other =
devices=20
are sharing the same resources, their resource usage, and what they are =
doing at=20
any time: an operational nightmare which would break all the software =
isolation=20
rules ! :-) . In =
addition, we=20
sometime shutdown some devices and re-enable them later, (or we grow =
more=20
devices in the cluster): Round robin does not work at all in this case. =




  
In reality, there are multiple varying factor like call =
duration,=20
  different vareity of platform within the cluster (say Generation-1 10 =
devices=20
  and Generation-2 5 devies), different resource usage for the call, =
rate at=20
  which calls generation. =
</Partha>
Thierry>> =20
yes ! 


  
 

  
The second reason is that a relative occupation is very easy to =
derive=20
  from information already available: any system knows where it is in =
terms of=20
  CPU or Memory usage.

  
 

  
<Partha> Converting CPU or Memory to weight is not =
explained your=20
  draft. It is the real algorithm which needs lot of discussion. Because =

  preparing the mapping between CPU & Memory to weight will be=20
  mind-blogging activity </Partha>


Thierry>> We could be more explicit, but should we ? (see =
next=20
paragraph). Memory and CPU are just the two most obvious examples or =
resources=20
that you need to process sessions. For instance, some of our servers =
have a=20
fixed amount of sessions registers (for instance 50 000). In this case, =
the=20
weight also reflects that. The algorithm we use for weights is very =
simple: this=20
is the smallest remaining (in percentage) of each resource that is (1) =
Required=20
to process a call, (2) Scarce  (not infinite). Let's consider the =
following=20
system: 
The server has 70% of its memory available for calls, an =
average of=20
30% CPU usage (average over the last 10s for instance), and 40% of its =
registers=20
available (remember that those resources may be shared by other =
components or=20
devices for any other usage: it does not matter). 
Then the weight at =
this=20
moment is MIN (70, 30, 40) =3D 30%. You then need to normalize this =
weight to=20
match the weight range (an INT ? ) 

Now we are not saying that =
everybody=20
should calculate their weight like that. What is important is that the =
weight=20
reflects your current relative capacity. In our case, this =
"algorithm"=20
balances the load of our servers very well. Now if you use a server that =
has a=20
fixed amount of resources and that is dimensioned to process let's say =
50 000=20
sessions, at a pace of 100 session/s, then you could probably derive a =
weight=20
that would depend on both your count of sessions left, and your current =
rate of=20
session/s.  


 


<Partha> In case every device calculate =
its own=20
algorithm for weight, then weight from one device may not same as other =
device.=20
Say dev1 takes Min(CPU, Mem, DSP) then weight will be =
calculated=20
as Min (70, 40, 30) =3D 30% and another device =
dev2 calculates Min=20
(CPU, Mem) then weight will be Min (70, 40) =3D 40% even =
though they are=20
equally loaded. It is very important to have the common unit to solve =
this issue=20
</Partha>  


  
 

  
 In contrast, deriving a left capacity in terms of remaining =

  sessions can be difficult: Host can be used by different applications =
and=20
  perform different duties, which can change over the day, all of them =
impacting=20
  their true capacity at a certain moment. Beside, as already discussed, =

  different sessions require different resources, so there is a whole=20
  complication of dealing with a multitude of call profiles, that =
(again) are=20
  not even constant over the day.

Also, we do not want to derive =
some=20
  shedding decision from the weight.

  
 

  
<Partha>  In case you specify, IETF SoC does not allow =
and so=20
  we want another value, I think that we are in same =
page.


Thierry>> we are on the same page. I thought you suggested to =
use the=20
same value.  


 


  
If not, I could not understand your argument. As you =
mentioned,=20
  all system will know CPU or memory usage and it is possible to come up =
the=20
  static or refreshable value saying that load shed has to happen when =
85% of=20
  CPU utilization or 90% of memory utilization. The different =
between our=20
  proposal is that there is no need of two value per=20
  message instead 1 value in each message & 1 value is =
static or=20
  in periodic manner  </Partha>


Thierry>> We use a single value: see the example =
above.  


<Partha> I'm =
proposing is to use=20
single real time value with units like number of dialog and =
exchange=20
the threshold (max dialog) using some other static=20
interface. Let =
overload=20
decides how to reduce or when to shed the load based on this =
information. The=20
advantage in this mechanism is that only one real time value is required =
between=20
entities.


  


In=20
your proposed solution, "weight" as one value for load =
balancing and=20
"oc" as another value for overload shedding. Pls correct me incase I=20
misunderstand something here =
</Partha> 



--

-- 
Cordially,=20

Thierry =
Bessis=20



  IMS Solution: =
Architecture=20
& Product Management - ALTA Member  =
Organization:=20
ALU > Carrier group > Wireline BU >  IMS Sol =
Arch,=20
SE, PdM

  1960 Lucent=20
Lane - Room 6L-517
  Naperville, IL  60563  U =
S  =
Tel/Fax: +1=20
630 979 7989
  Corporate IM: =
tbessis    -=20
Yahoo IM:     thierry_bessis

My =
WEB=20
page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My=20
Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =

Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference=20
information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - =
Access=20
Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: =
12 April -=20
14 April 2011
Upcoming planned Vacation:=20
None 


------_=_NextPart_002_01CBE3DC.94AA21ED--

------_=_NextPart_001_01CBE3DC.94AA21ED
Content-Type: image/jpeg;
	name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: <856275612@16032011-2B7C>
Content-Description: al_corp_h_3d_16mm.jpg
Content-Location: al_corp_h_3d_16mm.jpg
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------_=_NextPart_001_01CBE3DC.94AA21ED--

From Thierry.Bessis@alcatel-lucent.com  Wed Mar 16 10:02:01 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 51A163A69ED for ; Wed, 16 Mar 2011 10:02:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.94
X-Spam-Level: 
X-Spam-Status: No, score=-5.94 tagged_above=-999 required=5 tests=[AWL=0.057,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yxHG05PSRqfS for ; Wed, 16 Mar 2011 10:01:59 -0700 (PDT)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 943D33A69EA for ; Wed, 16 Mar 2011 10:01:58 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p2GH3NvW004671 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 16 Mar 2011 12:03:23 -0500 (CDT)
Received: from [135.244.7.43] (tbessis.lra.lucent.com [135.244.7.43]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2GH3KBj006752; Wed, 16 Mar 2011 12:03:20 -0500 (CDT)
Message-ID: <4D80ED57.7030709@alcatel-lucent.com>
Date: Wed, 16 Mar 2011 12:03:19 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Parthasarathi R (partr)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090102040909080707040609"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 17:02:01 -0000

This is a multi-part message in MIME format.
--------------090102040909080707040609
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi Partha,

> Thierry>> We could be more explicit, but should we ? (see next 
> paragraph). Memory and CPU are just the two most obvious examples or 
> resources that you need to process sessions. For instance, some of our 
> servers have a fixed amount of sessions registers (for instance 50 
> 000). In this case, the weight also reflects that. The algorithm we 
> use for weights is very simple: this is the smallest remaining (in 
> percentage) of each resource that is (1) Required to process a call, 
> (2) Scarce  (not infinite). Let's consider the following system:
> The server has 70% of its memory available for calls, an average of 
> 30% CPU usage (average over the last 10s for instance), and 40% of its 
> registers available (remember that those resources may be shared by 
> other components or devices for any other usage: it does not matter).
> Then the weight at this moment is MIN (70, 30, 40) = 30%. You then 
> need to normalize this weight to match the weight range (an INT ? )
>
> Now we are not saying that everybody should calculate their weight 
> like that. What is important is that the weight reflects your current 
> _relative_ capacity. In our case, this "algorithm" balances the load 
> of our servers very well. Now if you use a server that has a fixed 
> amount of resources and that is dimensioned to process let's say 50 
> 000 sessions, at a pace of 100 session/s, then you could probably 
> derive a weight that would depend on both your count of sessions left, 
> and your current rate of session/s.
>  In case every device calculate its own algorithm for weight, 
> then weight from one device may not same as other device. Say dev1 
> takes Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 
> 40, 30) = 30% and another device dev2 calculates Min (CPU, Mem) then 
> weight will be Min (70, 40) = 40% even though they are equally loaded. 
> It is very important to have the common unit to solve this issue 
> 
TB >> I think here you imply that the DSP is used as a finite pool of 
bearer channels resources, and your are probably using the DSP to 
process bearers. May be you are an SBC or so. Then, it is obvious that 
you should also consider this resource as well, since you need it to 
process new sessions. So you can't say that you have 40% resource 
available to process calls if actually one of your resource needed has 
only 30% available left !
Our recommendation is that the occupation (weight) reflects the MIN of 
ANY resource that must be used for call processing and that is constrained.

I think it is on the contrary an advantage to not specify exactly how it 
is calculated, but to specify exactly what it is, or what it means 
instead (The most constrained resource ratio).
>> If not, I could not understand your argument. As you mentioned, all 
>> system will know CPU or memory usage and it is possible to come up 
>> the static or refreshable value saying that load shed has to happen 
>> when 85% of CPU utilization or 90% of memory utilization. The 
>> different between our proposal is that there is no need of two value 
>> per message instead 1 value in each message & 1 value is static or in 
>> periodic manner 
> Thierry>> We use a single value: see the example above.
>  I'm proposing is to use single real time value with units 
> like number of dialog and exchange the threshold (max dialog) using 
> some other static interface.
TB> This proposal is exactly what we tried to stay away from ! I thought 
I gave lot of arguments in this thread to explain  why working with 
dialogs is at the same time more difficult and not the most 
straightforward information to do resource based load sharing.  
Moreover, using and sharing thresholds and trying to do overload control 
as well from there is also operationally very difficult: We did that 
before and it did not work, because (inside the same company !) nobody 
never agreed about what those threshold were and how to use them. This 
is why we decided to separate load distribution and overload shedding 
(OLC provides a rejection rate, and load distribution provides a 
resource ratio). Our OLC mechanism what the starting point for Vijay's 
effort to lead the IETF OLC draft.
> Let overload decides how to reduce or when to shed the load based on 
> this information. The advantage in this mechanism is that only one 
> real time value is required between entities.
See my comment above: we tried that, and it was very difficult to come 
up with a common approach in all of our (IMS related) products.
Beside, having the destination clearly specifies the requested rejection 
rate is much more convenient and determinist than sharing "session 
available" and threshold information and letting the source node figure 
out what it really means in terms of shedding.
> In your proposed solution, "weight" as one value for load balancing 
> and "oc" as another value for overload shedding. Pls correct me incase 
> I misunderstand something here 
Yes, but the "oc"  valus is the one already proposed in the OLC draft: 
this is not in THIS proposed solution.

Cordially,

Thierry
>
> --
> -- 
> Cordially,
> Thierry Bessis IMS Solution: Architecture & Product Management - ALTA 
> Member
>   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, 
> SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
>   Naperville, IL  60563  U S
>   Tel/Fax: +1 630 979 7989
>   Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
>
> My WEB page: http://ihgpweb.ih.lucent.com/~tbessis/
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 
> 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> Upcoming planned Business Trip: 12 April - 14 April 2011
> Upcoming planned Vacation: None
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------090102040909080707040609
Content-Type: multipart/related;
 boundary="------------060601080601010206080106"


--------------060601080601010206080106
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Hi Partha, 

    

    

      
Thierry>> We could be more explicit, but should we ?
        (see next paragraph). Memory and CPU are just the two most
        obvious examples or resources that you need to process sessions.
        For instance, some of our servers have a fixed amount of
        sessions registers (for instance 50 000). In this case, the
        weight also reflects that. The algorithm we use for weights is
        very simple: this is the smallest remaining (in percentage) of
        each resource that is (1) Required to process a call, (2)
        Scarce  (not infinite). Let's consider the following system: 

        The server has 70% of its memory available for calls, an average
        of 30% CPU usage (average over the last 10s for instance), and
        40% of its registers available (remember that those resources
        may be shared by other components or devices for any other
        usage: it does not matter). 

        Then the weight at this moment is MIN (70, 30, 40) = 30%. You
        then need to normalize this weight to match the weight range (an
        INT ? ) 

        

        Now we are not saying that everybody should calculate their
        weight like that. What is important is that the weight reflects
        your current relative capacity. In our case, this
        "algorithm" balances the load of our servers very well. Now if
        you use a server that has a fixed amount of resources and that
        is dimensioned to process let's say 50 000 sessions, at a pace
        of 100 session/s, then you could probably derive a weight that
        would depend on both your count of sessions left, and your
        current rate of session/s.  

      
 

      
<Partha> In case every
            device calculate its own algorithm for weight, then weight
            from one device may not same as other device. Say dev1 takes
            Min(CPU, Mem, DSP) then weight will be calculated as Min
            (70, 40, 30) = 30% and another device dev2 calculates Min
            (CPU, Mem) then weight will be Min (70, 40) = 40% even
            though they are equally loaded. It is very important to have
            the common unit to solve this issue </Partha>  

          

    

    TB >> I think here you imply that the DSP is used as a finite
    pool of bearer channels resources, and your are probably using the
    DSP to process bearers. May be you are an SBC or so. Then, it is
    obvious that you should also consider this resource as well, since
    you need it to process new sessions. So you can't say that you have
    40% resource available to process calls if actually one of your
    resource needed has only 30% available left ! 

    Our recommendation is that the occupation (weight) reflects the MIN
    of ANY resource that must be used for call processing and that is
    constrained. 

    

    I think it is on the contrary an advantage to not specify exactly
    how it is calculated, but to specify exactly what it is, or what it
    means instead (The most constrained resource ratio). 

    

      

        
If not, I could not understand your argument. As you
          mentioned, all system will know CPU or memory usage and it is
          possible to come up the static or refreshable value saying
          that load shed has to happen when 85% of CPU utilization
          or 90% of memory utilization. The different between our
          proposal is that there is no need of two value per
          message instead 1 value in each message & 1 value is
          static or in periodic manner  </Partha>

      

      
Thierry>> We use a single value: see the example
        above.  

      
<Partha> I'm proposing is to use
              single real time value with units like number of dialog
              and exchange the threshold (max dialog) using some other
              static interface. 

    

    TB> This proposal is exactly what we tried to stay away from ! I
    thought I gave lot of arguments in this thread to explain  why
    working with dialogs is at the same time more difficult and not the
    most straightforward information to do resource based load sharing. 
    Moreover, using and sharing thresholds and trying to do overload
    control as well from there is also operationally very difficult: We
    did that before and it did not work, because (inside the same
    company !) nobody never agreed about what those threshold were and
    how to use them. This is why we decided to separate load
    distribution and overload shedding (OLC provides a rejection rate,
    and load distribution provides a resource ratio). Our OLC mechanism
    what the starting point for Vijay's effort to lead the IETF OLC
    draft. 

    

      
Let overload decides how to reduce or when to
              shed the load based on this information. The advantage in
              this mechanism is that only one real time value is
              required between entities.

    

    See my comment above: we tried that, and it was very difficult to
    come up with a common approach in all of our (IMS related) products.
    

    Beside, having the destination clearly specifies the requested
    rejection rate is much more convenient and determinist than sharing
    "session available" and threshold information and letting the source
    node figure out what it really means in terms of shedding. 

    

      
In your proposed
            solution, "weight" as one value for load balancing and "oc"
            as another value for overload shedding. Pls correct me
            incase I misunderstand something here </Partha>
        

      

    

    Yes, but the "oc"  valus is the one already proposed in the OLC
    draft: this is not in THIS proposed solution. 

    

    Cordially, 

    

    Thierry

    

      


      

      
--

        
        -- 

                Cordially, 

              Thierry Bessis
          IMS Solution: Architecture
          & Product Management - ALTA Member

          Organization: ALU >
          Carrier group > Wireline BU >  IMS Sol
        Arch, SE, PdM

        Alcatel-Lucent

          1960 Lucent Lane - Room 6L-517

          Naperville, IL  60563  U S

          Tel/Fax: +1 630 979 7989

          Corporate IM: tbessis    -
          Yahoo IM:     thierry_bessis

        

        My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/

          My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

          ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

        

        Conference information:

          2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
          Code: 9797989

          others countries see: http://all.alcatel-lucent.com/audio

        

        Upcoming planned Business Trip: 12
              April - 14 April 2011

        Upcoming planned Vacation: None
      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------060601080601010206080106
Content-Type: image/jpeg
Content-Transfer-Encoding: base64
Content-ID: 
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--------------060601080601010206080106
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------060601080601010206080106--

--------------090102040909080707040609--

From henry.sinnreich@gmail.com  Wed Mar 16 11:53:18 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4D0EA3A6A49 for ; Wed, 16 Mar 2011 11:53:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.126
X-Spam-Level: 
X-Spam-Status: No, score=-3.126 tagged_above=-999 required=5 tests=[AWL=0.473,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fPBh++95-uE5 for ; Wed, 16 Mar 2011 11:53:17 -0700 (PDT)
Received: from mail-gx0-f172.google.com (mail-gx0-f172.google.com [209.85.161.172]) by core3.amsl.com (Postfix) with ESMTP id 7797D3A6997 for ; Wed, 16 Mar 2011 11:53:17 -0700 (PDT)
Received: by gxk19 with SMTP id 19so952006gxk.31 for ; Wed, 16 Mar 2011 11:54:44 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=iigAz3lriwbvU89fqWRZ+nOoPOe5y21pg9bscDwLzkQ=; b=pApZBdWXRXP++PQ0AiiobjqgyZyYyCB+h4sGtDuC9UDyuOmHFCurr9OQUmgrToHfnG IIQFWslWcep6wsUQsCZjN780dsqW1LlXgZygH3IhXIFpvhr1lWCF3CZqCvYiMKM1UIZq ECQe0UhJyiYfDhux+6iVtJyDHZD55jPYIBXIE=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=UlXqOLA8MugyLG7jkVCnr3WzB++IhQ7qSblra3Cl7F+jnMykIyW5O5wY6bufPoM/ew 32I0c3JXi8xxg0fNmxe1j/HcXvRufGrV+66S4T4AuXZR+Z+WNwmt322hjCzST0uKez/a +Q6uf1qa/k8Lcmj7cPU5aF9BZJ9Mx6CzWKjPA=
Received: by 10.150.214.8 with SMTP id m8mr423564ybg.121.1300301683871; Wed, 16 Mar 2011 11:54:43 -0700 (PDT)
Received: from [10.0.1.5] (cpe-76-184-225-135.tx.res.rr.com [76.184.225.135]) by mx.google.com with ESMTPS id q20sm3714507ybk.5.2011.03.16.11.54.41 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 16 Mar 2011 11:54:42 -0700 (PDT)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Wed, 16 Mar 2011 13:54:39 -0500
From: Henry Sinnreich 
To: Cullen Jennings , DISPATCH list 
Message-ID: 
Thread-Topic: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
Thread-Index: AcvkC5oFXi2QhtU05k64Wfm2zXK+RA==
In-Reply-To: 
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 18:53:18 -0000

Hi Cullen,

Section 3. Requirements says:
... In addition, there is a strong interest to define a TURN-like protocol
   that looks like HTTP to intermediaries, so that media can be tunneled
   over HTTP. 
...

This is very interesting and innovative.
Can you explain here a little bit?

Thanks, Henry


On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:

> 
> I wrote up the start of a draft on requirements.
> 
> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> 
> I think the key thing for the key thing for IETF working is starting to think
> about the requirements and what on the on the wire protocols are that we want
> to select. I included some API definition because I found it helpful to be
> thinking about that in this draft but that would not be part of the IETF work.
> That would be work I think W3C should do. The draft also touches on the of the
> security issues for media and signaling originating from browsers.
> 
> Cullen 
> 
> 
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch



From adam@nostrum.com  Wed Mar 16 12:54:02 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 238CB3A6AA9 for ; Wed, 16 Mar 2011 12:54:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.588
X-Spam-Level: 
X-Spam-Status: No, score=-102.588 tagged_above=-999 required=5 tests=[AWL=0.012, BAYES_00=-2.599, SPF_PASS=-0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BdlqWUbQEHtC for ; Wed, 16 Mar 2011 12:54:01 -0700 (PDT)
Received: from nostrum.com (nostrum-pt.tunnel.tserv2.fmt.ipv6.he.net [IPv6:2001:470:1f03:267::2]) by core3.amsl.com (Postfix) with ESMTP id C87183A6A43 for ; Wed, 16 Mar 2011 12:54:00 -0700 (PDT)
Received: from dn3-228.estacado.net (vicuna-alt.estacado.net [75.53.54.121]) (authenticated bits=0) by nostrum.com (8.14.3/8.14.3) with ESMTP id p2GJtMhH009446 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Wed, 16 Mar 2011 14:55:22 -0500 (CDT) (envelope-from adam@nostrum.com)
Message-ID: <4D8115AA.9000402@nostrum.com>
Date: Wed, 16 Mar 2011 14:55:22 -0500
From: Adam Roach 
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: Henry Sinnreich 
References: 
In-Reply-To: 
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Received-SPF: pass (nostrum.com: 75.53.54.121 is authenticated by a trusted mechanism)
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 19:54:02 -0000

On 3/16/11 1:54 PM, Henry Sinnreich wrote:
> Hi Cullen,
>
> Section 3. Requirements says:
> ... In addition, there is a strong interest to define a TURN-like protocol
>     that looks like HTTP to intermediaries, so that media can be tunneled
>     over HTTP.
> ...
>
> This is very interesting and innovative.
> Can you explain here a little bit?

There's already been some work in this area. See, for example:

http://tools.ietf.org/html/rfc3093

...and...

http://msdn.microsoft.com/en-us/library/dd358571%28v=prot.10%29.aspx

(I'm becoming increasingly frustrated at the inability of creating 
parodies that last; reality keeps catching up with and then surpassing 
the level of absurdity that parody generally lives at).

/a

From dworley@avaya.com  Wed Mar 16 13:14:16 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 695B13A6997 for ; Wed, 16 Mar 2011 13:14:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.543
X-Spam-Level: 
X-Spam-Status: No, score=-102.543 tagged_above=-999 required=5 tests=[AWL=0.056, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id SOXavxrfyJei for ; Wed, 16 Mar 2011 13:14:15 -0700 (PDT)
Received: from p-us1-iereast-outbound-tmp.us1.avaya.com (p-us1-iereast-outbound-tmp.us1.avaya.com [135.11.29.16]) by core3.amsl.com (Postfix) with ESMTP id 9B3C03A690F for ; Wed, 16 Mar 2011 13:14:15 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAOa2gE2HCzI1/2dsb2JhbACmDHeoJgKaGYVjBJBF
X-IronPort-AV: E=Sophos;i="4.63,195,1299474000"; d="scan'208";a="63818848"
Received: from unknown (HELO p-us1-erheast.us1.avaya.com) ([135.11.50.53]) by p-us1-iereast-outbound-tmp.us1.avaya.com with ESMTP; 16 Mar 2011 16:15:40 -0400
X-IronPort-AV: E=Sophos;i="4.63,195,1299474000"; d="scan'208";a="622919801"
Received: from dc-us1hcex2.us1.avaya.com (HELO DC-US1HCEX2.global.avaya.com) ([135.11.52.21]) by p-us1-erheast-out.us1.avaya.com with ESMTP; 16 Mar 2011 16:15:40 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX2.global.avaya.com ([::1]) with mapi; Wed, 16 Mar 2011 16:15:39 -0400
From: "Worley, Dale R (Dale)" 
To: Thierry Bessis 
Date: Wed, 16 Mar 2011 16:13:08 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvioLcg/OD66Wo4RA6DOQcn/kXv3wBddnY3
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,<4D76C924.7050900@cisco.com> , <4D77E128.8060302@alcatel-lucent.com> , <4D7EA694.4000308@alcatel-lucent.com>
In-Reply-To: <4D7EA694.4000308@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 20:14:16 -0000

________________________________________
From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]

I agree that we need to clarify.
________________________________________

More importantly, you need to conceptualize all of these items as *choices*
that have *consequences*, and be able to say, "We *know* what the consequen=
ces
of these particular choices are."

Dale

From lorenzo@meetecho.com  Wed Mar 16 13:19:39 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DDC613A699F for ; Wed, 16 Mar 2011 13:19:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.719
X-Spam-Level: 
X-Spam-Status: No, score=-0.719 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4isGnOt4n5Z1 for ; Wed, 16 Mar 2011 13:19:38 -0700 (PDT)
Received: from smtplq03.aruba.it (smtplq-out8.aruba.it [62.149.158.28]) by core3.amsl.com (Postfix) with SMTP id 4F7283A694C for ; Wed, 16 Mar 2011 13:19:37 -0700 (PDT)
Received: (qmail 16625 invoked by uid 89); 16 Mar 2011 20:21:00 -0000
Received: from unknown (HELO smtp2.aruba.it) (62.149.158.222) by smtplq03.aruba.it with SMTP; 16 Mar 2011 20:21:00 -0000
Received: (qmail 28973 invoked by uid 89); 16 Mar 2011 20:21:00 -0000
Received: from unknown (HELO rainpc) (lorenzo@meetecho.com@87.1.155.125) by smtp2.ad.aruba.it with SMTP; 16 Mar 2011 20:21:00 -0000
Date: Wed, 16 Mar 2011 21:12:31 +0100
From: Lorenzo Miniero 
To: Henry Sinnreich 
Message-ID: <20110316211231.0f54d962@rainpc>
In-Reply-To: 
References:  
Organization: Meetecho
X-Mailer: Claws Mail 3.7.8 (GTK+ 2.22.0; x86_64-redhat-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Spam-Rating: smtp2.ad.aruba.it 1.6.2 0/1000/N
X-Spam-Rating: smtplq03.aruba.it 1.6.2 0/1000/N
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 20:19:40 -0000

On Wed, 16 Mar 2011 13:54:39 -0500
Henry Sinnreich  wrote:

> Hi Cullen,
> 
> Section 3. Requirements says:
> ... In addition, there is a strong interest to define a TURN-like protocol
>    that looks like HTTP to intermediaries, so that media can be tunneled
>    over HTTP. 
> ...
> 
> This is very interesting and innovative.
> Can you explain here a little bit?
> 
> Thanks, Henry
> 


We implemented something like that for our system, also documented in a paper lately: basically a controller protocol that tunnels every other protocol on top of HTTP to bypass restrictive firewalls. Since it works, we were also thinking about a "lightweight" version of such protocol to only convey RTP specifically for our web interface, considering RTP is the only protocol we couldn't wrap via JavaScript so far. So I agree that such a TURN-like protocol would be a good idea for the WG.

Of course, there would be limitations and issues because of this transport, but since firewalls are there to stay, and few of them like RTP appearing "out of the blue", that's something we probably need to live with.

L.


> 
> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
> 
> > 
> > I wrote up the start of a draft on requirements.
> > 
> > http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> > 
> > I think the key thing for the key thing for IETF working is starting to think
> > about the requirements and what on the on the wire protocols are that we want
> > to select. I included some API definition because I found it helpful to be
> > thinking about that in this draft but that would not be part of the IETF work.
> > That would be work I think W3C should do. The draft also touches on the of the
> > security issues for media and signaling originating from browsers.
> > 
> > Cullen 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > dispatch mailing list
> > dispatch@ietf.org
> > https://www.ietf.org/mailman/listinfo/dispatch
> 
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


-- 
Lorenzo Miniero, COB

Meetecho s.r.l.
Web Conferencing and Collaboration Tools
http://www.meetecho.com

From dworley@avaya.com  Wed Mar 16 13:21:26 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EFB053A69DC for ; Wed, 16 Mar 2011 13:21:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.544
X-Spam-Level: 
X-Spam-Status: No, score=-102.544 tagged_above=-999 required=5 tests=[AWL=0.055, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 32eMXbFW6EDd for ; Wed, 16 Mar 2011 13:21:25 -0700 (PDT)
Received: from co300216-co-outbound.net.avaya.com (co300216-co-outbound.net.avaya.com [198.152.13.100]) by core3.amsl.com (Postfix) with ESMTP id E2B7D3A69D9 for ; Wed, 16 Mar 2011 13:21:24 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU2HCzI1/2dsb2JhbACmZXSkeQKZFoVhBJAM
X-IronPort-AV: E=Sophos;i="4.63,195,1299474000"; d="scan'208";a="269820165"
Received: from unknown (HELO p-us1-erheast.us1.avaya.com) ([135.11.50.53]) by co300216-co-outbound.net.avaya.com with ESMTP; 16 Mar 2011 16:22:51 -0400
X-IronPort-AV: E=Sophos;i="4.63,195,1299474000"; d="scan'208";a="622929082"
Received: from unknown (HELO DC-US1HCEX3.global.avaya.com) ([135.11.52.22]) by p-us1-erheast-out.us1.avaya.com with ESMTP; 16 Mar 2011 16:22:51 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX3.global.avaya.com ([135.11.52.22]) with mapi; Wed, 16 Mar 2011 16:22:50 -0400
From: "Worley, Dale R (Dale)" 
To: Thierry Bessis 
Date: Wed, 16 Mar 2011 16:22:08 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvjVLOrBXOh8I1pQkGg9GMeF5ypagAwx9Pm
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com>	<4D7633F9.7000608@cisco.com> <4D768BEE.2060400@bell-labs.com>	<4D7698FD.4030501@cisco.com> <4D76A762.6090708@bell-labs.com>,<4D76C924.7050900@cisco.com> , <4D77E128.8060302@alcatel-lucent.com> , <4D7EA694.4000308@alcatel-lucent.com> , <4D7FD490.1060109@alcatel-lucent.com>
In-Reply-To: <4D7FD490.1060109@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 20:21:26 -0000

_______________________________________
From: Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]

Our main benefit is that our resource utilization (essentially CPU and Memo=
ry as we discussed) is very good. We see that the system effectively load t=
he cluster evenly, and returns to its equilibrium even after some "disrupti=
ng" events occurred. Such event are for example:  temporary loss and then r=
eturn of a node, or a temporary overload of some nodes due to administrativ=
e work, or even the dynamic installation of some new components on existing=
 hosts. Also, we are dealing with remote nodes which are NOT getting (or no=
t willing to use)  the weights and may concentrate the traffic unevenly. Th=
en, our mechanism will fix the issue using the "good behavior" of nodes whi=
ch are using the weights to re-balance the cluster.

Since all our nodes get overloaded at roughly the same time, we know we hav=
e roughly the best possible throughput. In addition, based on this knowledg=
e, we decide to never re-transmit any message in the same cluster when we g=
et the "shedding" signal back: This is a huge advantage.
________________________________________

How does this compare to the situation that obtains when using static weigh=
ts?

Dale

From petithug@acm.org  Wed Mar 16 14:19:13 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AA3933A6A38 for ; Wed, 16 Mar 2011 14:19:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.149
X-Spam-Level: 
X-Spam-Status: No, score=-102.149 tagged_above=-999 required=5 tests=[AWL=0.116, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3rIzKum6800s for ; Wed, 16 Mar 2011 14:19:12 -0700 (PDT)
Received: from server.implementers.org (server.implementers.org [69.55.225.91]) by core3.amsl.com (Postfix) with ESMTP id 99F4A3A694C for ; Wed, 16 Mar 2011 14:19:12 -0700 (PDT)
Received: by server.implementers.org (Postfix, from userid 1001) id 30D30DBCC044; Wed, 16 Mar 2011 21:20:39 +0000 (UTC)
Received: from [192.168.2.3] (server.implementers.org [127.0.0.1]) by server.implementers.org (Postfix) with ESMTPA id 65F416C9845C; Wed, 16 Mar 2011 21:20:38 +0000 (UTC)
Message-ID: <4D8129A5.40105@acm.org>
Date: Wed, 16 Mar 2011 14:20:37 -0700
From: Marc Petit-Huguenin 
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.16) Gecko/20110303 Iceowl/1.0b1 Icedove/3.0.11
MIME-Version: 1.0
To: Lorenzo Miniero 
References: 	 <20110316211231.0f54d962@rainpc>
In-Reply-To: <20110316211231.0f54d962@rainpc>
X-Enigmail-Version: 1.0.1
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 21:19:13 -0000

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

There is already a tunneling protocol for RTP, and you cannot find more
turn-like than this: It's called TURN.  Just add an HTTP transport between the
client and the TURN server, and that's it.  I wrote an implementation (probably
4 hours of work to add this on an existing TURN server) and even an unpublished
draft of a I-D back in 2008.  The draft is probably 4 pages long without the
boilerplates (counting the Application tag for RFC 5928) - TURN is 67 pages
long, and that probably what will be needed for a "TURN-like protocol that looks
like HTTP to intermediaries".

On 03/16/2011 01:12 PM, Lorenzo Miniero wrote:
> On Wed, 16 Mar 2011 13:54:39 -0500
> Henry Sinnreich  wrote:
> 
>> Hi Cullen,
>>
>> Section 3. Requirements says:
>> ... In addition, there is a strong interest to define a TURN-like protocol
>>    that looks like HTTP to intermediaries, so that media can be tunneled
>>    over HTTP. 
>> ...
>>
>> This is very interesting and innovative.
>> Can you explain here a little bit?
>>
>> Thanks, Henry
>>
> 
> 
> We implemented something like that for our system, also documented in a paper lately: basically a controller protocol that tunnels every other protocol on top of HTTP to bypass restrictive firewalls. Since it works, we were also thinking about a "lightweight" version of such protocol to only convey RTP specifically for our web interface, considering RTP is the only protocol we couldn't wrap via JavaScript so far. So I agree that such a TURN-like protocol would be a good idea for the WG.
> 
> Of course, there would be limitations and issues because of this transport, but since firewalls are there to stay, and few of them like RTP appearing "out of the blue", that's something we probably need to live with.
> 
> L.
> 
> 
>>
>> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
>>
>>>
>>> I wrote up the start of a draft on requirements.
>>>
>>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>>>
>>> I think the key thing for the key thing for IETF working is starting to think
>>> about the requirements and what on the on the wire protocols are that we want
>>> to select. I included some API definition because I found it helpful to be
>>> thinking about that in this draft but that would not be part of the IETF work.
>>> That would be work I think W3C should do. The draft also touches on the of the
>>> security issues for media and signaling originating from browsers.
>>>
>>> Cullen 
>>>

- -- 
Marc Petit-Huguenin
Personal email: marc@petit-huguenin.org
Professional email: petithug@acm.org
Blog: http://blog.marc.petit-huguenin.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAk2BKaMACgkQ9RoMZyVa61fEXgCfXXszvQyai97vQV73hzfa9B4k
BkcAn05Jxb8Gjl2yFDLu3piqQDX+99f4
=ZF4s
-----END PGP SIGNATURE-----

From lconroy@insensate.co.uk  Wed Mar 16 15:48:36 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8E58F3A6A1B for ; Wed, 16 Mar 2011 15:48:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.56
X-Spam-Level: 
X-Spam-Status: No, score=-2.56 tagged_above=-999 required=5 tests=[AWL=0.039,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id seArRqVwrS1H for ; Wed, 16 Mar 2011 15:48:35 -0700 (PDT)
Received: from insensate.co.uk (ghost.insensate.co.uk [213.152.49.121]) by core3.amsl.com (Postfix) with ESMTP id 77C1C3A69B9 for ; Wed, 16 Mar 2011 15:48:35 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by insensate.co.uk (Postfix) with ESMTP id 53C532B81EC; Wed, 16 Mar 2011 22:50:01 +0000 (GMT)
X-Virus-Scanned: amavisd-new at insensate.co.uk
Received: from insensate.co.uk ([127.0.0.1]) by localhost (psyche.insensate.co.uk [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MnfBwZRx9n7V; Wed, 16 Mar 2011 22:50:01 +0000 (GMT)
Received: from [127.0.0.1] (localhost [127.0.0.1]) by insensate.co.uk (Postfix) with ESMTP id F414F2B81E1; Wed, 16 Mar 2011 22:50:00 +0000 (GMT)
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
From: Lawrence Conroy 
In-Reply-To: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com>
Date: Wed, 16 Mar 2011 22:50:00 +0000
Content-Transfer-Encoding: quoted-printable
Message-Id: 
References: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com>
To: Cullen Jennings 
X-Mailer: Apple Mail (2.1082)
Cc: DISPATCH list 
Subject: Re: [dispatch] DISPATCH Ad-HOC meetings about VIPR and RTCWeb at IETF 80
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 22:48:36 -0000

Hi Cullen,
 could you clarify what exactly Dispatch is going to do in Ad-Hoc =
meetings
discussing docs that are targeted at suggested WGs that have their own =
BOFs?

Why do these need their own Ad Hoc meetings in DISPATCH -- are these =
docs so urgent
that they have to be considered in DISPATCH before their intended WGs =
have been created?

Also, the choice of Friday @ 14:15 for the "RTCWEB" chat seems kinda =
sub-optimal,
considering that's BEHAVE's traditional slot.

all the best from a puzzled,
  Lawrence

On 15 Mar 2011, at 17:45, Cullen Jennings wrote:
> The dispatch WG will have two ad-hoc meeting.=20
>=20
> The first is discuss the technical content of the drafts related to =
the proposed VIPR WG.=20
>=20
> Tentative time: VIPR's is in Tyrolka on Thursday 1300-1400
>=20
> The second is to discuss the technical contents of drafts related to =
the proposed RTCWeb WG.=20
>=20
> Tentative time:  RTCWeb's is  Friday 1415-1515 in the Grand Ballroom
>=20
> Send Agenda requests for time in either of these to Mary and I.=20
>=20
> Thanks,
> Mary & Cullen


From mary.ietf.barnes@gmail.com  Wed Mar 16 16:08:45 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 139FE3A6A7A for ; Wed, 16 Mar 2011 16:08:45 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.9
X-Spam-Level: 
X-Spam-Status: No, score=-102.9 tagged_above=-999 required=5 tests=[AWL=-0.302, BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3H9xFszAK3lx for ; Wed, 16 Mar 2011 16:08:44 -0700 (PDT)
Received: from mail-vx0-f172.google.com (mail-vx0-f172.google.com [209.85.220.172]) by core3.amsl.com (Postfix) with ESMTP id BD80B3A69B9 for ; Wed, 16 Mar 2011 16:08:43 -0700 (PDT)
Received: by vxg33 with SMTP id 33so2511554vxg.31 for ; Wed, 16 Mar 2011 16:10:10 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc:content-type; bh=CRi+c5e1l2/ZtO0qQZ4P7VDhQ602flQc1E8jlfQQReQ=; b=nnNH+mX3C7Hf50TT1hw5IiRDXGKMWtH3cjn8+BvgXiLESeUxkfBRMkQraz0Kb5wNPK /SbQodM/ZDRoUoOc7oYSriYR5LV5SMpe8Ug516qyM8w1lTwXURRvpg/rs44xh/SlcbIu /Bmu/9Zi9eqNBIk7gkRWzJNLUuTc76T5CH3/w=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=YDmVSQ7egSSBzfoBaIYJ3g7DTGrXCET4LzuDV1lru1QwLCOkhLibwbLFpznJFaaIhn fIQN6/guoDkuJArkij+lv5hDUpVsST+PinDvV+VmBHHaFmUmhFEa4zRfWhS33kY6m8vP bAz9By8ihdSbLUffIdbq7ZGsicKLMSKzfb5QI=
MIME-Version: 1.0
Received: by 10.52.74.66 with SMTP id r2mr792962vdv.263.1300317010351; Wed, 16 Mar 2011 16:10:10 -0700 (PDT)
Received: by 10.52.167.137 with HTTP; Wed, 16 Mar 2011 16:10:10 -0700 (PDT)
In-Reply-To: 
References: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com> 
Date: Wed, 16 Mar 2011 18:10:10 -0500
Message-ID: 
From: Mary Barnes 
To: Lawrence Conroy 
Content-Type: multipart/alternative; boundary=bcaec50166277f57af049ea1a500
Cc: DISPATCH list 
Subject: Re: [dispatch] DISPATCH Ad-HOC meetings about VIPR and RTCWeb at IETF 80
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 23:08:45 -0000

--bcaec50166277f57af049ea1a500
Content-Type: text/plain; charset=ISO-8859-1

Lawrence,

There's a few reasons why we are doing these as adhocs:

1) Folks want to take advantage of the meeting week to have some additional
f2f time to hash out issues.

2) We have found that having the sessions as adhoc dispatch sessions
facilitates the former as it makes it easier to get a meeting room and adds
some structure to the meeting/discussions and clarity to the fact that these
are covered by the note well, etc.

3) While these topics have been dispatched in terms of basically deciding
how they should move forward, they're still not officially chartered and
thus the dispatching thereof is not yet complete, so it's quite reasonable
to have these topics still under the guise of dispatch until such is
completed.  We have done this for several other topics in the past.

Agreed that the time is not ideal for folks involved in BEHAVE, but really
there is no other slot that wouldn't put the activity against another RAI
WG.  BEHAVE does have two of those afternoon slots, so perhaps folks that
are critical to BEHAVE that want to participate in the adhoc can work the
chairs on arranging the agenda to facilitate such. That's something that's
been necessary for other conficts in the past, as well.

Regards,
Mary.
DISPATCH WG co-chair


On Wed, Mar 16, 2011 at 5:50 PM, Lawrence Conroy wrote:

> Hi Cullen,
>  could you clarify what exactly Dispatch is going to do in Ad-Hoc meetings
> discussing docs that are targeted at suggested WGs that have their own
> BOFs?
>
> Why do these need their own Ad Hoc meetings in DISPATCH -- are these docs
> so urgent
> that they have to be considered in DISPATCH before their intended WGs have
> been created?
>
> Also, the choice of Friday @ 14:15 for the "RTCWEB" chat seems kinda
> sub-optimal,
> considering that's BEHAVE's traditional slot.
>
> all the best from a puzzled,
>  Lawrence
>
> On 15 Mar 2011, at 17:45, Cullen Jennings wrote:
> > The dispatch WG will have two ad-hoc meeting.
> >
> > The first is discuss the technical content of the drafts related to the
> proposed VIPR WG.
> >
> > Tentative time: VIPR's is in Tyrolka on Thursday 1300-1400
> >
> > The second is to discuss the technical contents of drafts related to the
> proposed RTCWeb WG.
> >
> > Tentative time:  RTCWeb's is  Friday 1415-1515 in the Grand Ballroom
> >
> > Send Agenda requests for time in either of these to Mary and I.
> >
> > Thanks,
> > Mary & Cullen
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>

--bcaec50166277f57af049ea1a500
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Lawrence,

There's a few reasons why we are doing the=
se as adhocs:

1) Folks want to take advantage of t=
he meeting week to have some additional f2f time to hash out issues.=A0


2) We have found that having the sessions as adhoc disp=
atch sessions facilitates the former as it makes it easier to get a meeting=
 room and adds some structure to the meeting/discussions and clarity to the=
 fact that these are covered by the note well, etc.=A0



3) While these topics have been dispatched in terms of =
basically deciding how they should move forward, they're still not offi=
cially chartered and thus the dispatching thereof is not yet complete, so i=
t's quite reasonable to have these topics still under the guise of disp=
atch until such is completed. =A0We have done this for several other topics=
 in the past.=A0



Agreed that the time is not ideal for folks involved in=
 BEHAVE, but really there is no other slot that wouldn't put the activi=
ty against another RAI WG. =A0BEHAVE does have two of those afternoon slots=
, so perhaps folks that are critical to BEHAVE that want to participate in =
the adhoc can work the chairs on arranging the agenda to facilitate such. T=
hat's something that's been necessary for other conficts in the pas=
t, as well.=A0



Regards,
Mary.
DISPATCH WG co-chair=



On Wed, Mar 16, 20=
11 at 5:50 PM, Lawrence Conroy <lconroy@insensate.co.uk> wrote:


Hi Cullen,

=A0could you clarify what exactly Dispatch is going to do in Ad-Hoc meeting=
s

discussing docs that are targeted at suggested WGs that have their own BOFs=
?



Why do these need their own Ad Hoc meetings in DISPATCH -- are these docs s=
o urgent

that they have to be considered in DISPATCH before their intended WGs have =
been created?



Also, the choice of Friday @ 14:15 for the "RTCWEB" chat seems ki=
nda sub-optimal,

considering that's BEHAVE's traditional slot.



all the best from a puzzled,

 =A0Lawrence




On 15 Mar 2011, at 17:45, Cullen Jennings wrote:

> The dispatch WG will have two ad-hoc meeting.

>

> The first is discuss the technical content of the drafts related to th=
e proposed VIPR WG.

>

> Tentative time: VIPR's is in Tyrolka on Thursday 1300-1400

>

> The second is to discuss the technical contents of drafts related to t=
he proposed RTCWeb WG.

>

> Tentative time: =A0RTCWeb's is =A0Friday 1415-1515 in the Grand Ba=
llroom

>

> Send Agenda requests for time in either of these to Mary and I.

>

> Thanks,

> Mary & Cullen




_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.org/mailman/listinfo/dispatch





--bcaec50166277f57af049ea1a500--

From partr@cisco.com  Wed Mar 16 16:16:57 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 053973A6AA5 for ; Wed, 16 Mar 2011 16:16:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -8.516
X-Spam-Level: 
X-Spam-Status: No, score=-8.516 tagged_above=-999 required=5 tests=[AWL=-0.915, BAYES_00=-2.599, EXTRA_MPART_TYPE=1, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id b-if7ogoJEd2 for ; Wed, 16 Mar 2011 16:16:54 -0700 (PDT)
Received: from ams-iport-2.cisco.com (ams-iport-2.cisco.com [144.254.224.141]) by core3.amsl.com (Postfix) with ESMTP id 865693A69B9 for ; Wed, 16 Mar 2011 16:16:51 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=115966; q=dns/txt; s=iport; t=1300317499; x=1301527099; h=mime-version:subject:date:message-id:references:from:to: cc; bh=KZAHHxXbt8KXwifGG7TWFCnAnVhk+O4SifeonEDijV8=; b=hRM9w2mssXmO1KSdmhxAnn/FbqTWlxUR2lf5FVdWQED3g40Hs45ykD1Y reYF+5/LszvHAN3Us8MxDecrDcrlgdnZct4xjyeVuSDTcWM80oihaEkI+ +MQxdLwDNrNA0sdS2NNouhbjXQdiVBnGDNQ5jqxZC1DGS0BzbUTnXaTCI s=;
X-Files: ATT19223371.jpg, al_corp_h_3d_16mm.jpg : 32899, 32899
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AjYKANThgE2Q/khLgWdsb2JhbAAwgVtQoiBTFAEBFiYlpVOcWYJ7GoJOBIF1gziLBA
X-IronPort-AV: E=Sophos;i="4.63,195,1299456000";  d="jpg'145?scan'145,208,217,145";a="22019148"
Received: from ams-core-2.cisco.com ([144.254.72.75]) by ams-iport-2.cisco.com with ESMTP; 16 Mar 2011 23:18:17 +0000
Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by ams-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2GNIFE4017842; Wed, 16 Mar 2011 23:18:16 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Thu, 17 Mar 2011 04:48:15 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----_=_NextPart_001_01CBE430.6CCBA05D"; type="multipart/alternative"
Date: Thu, 17 Mar 2011 04:48:14 +0530
Message-ID: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvj/CCgM8QByuttTsmek/rNZ/1g7wAJ8zgC
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com>
From: "Parthasarathi R (partr)" 
To: "Thierry Bessis" 
X-OriginalArrivalTime: 16 Mar 2011 23:18:15.0539 (UTC) FILETIME=[6D699430:01CBE430]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 23:16:57 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBE430.6CCBA05D
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_002_01CBE430.6CCBA05D"


------_=_NextPart_002_01CBE430.6CCBA05D
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi Thierry,
=20
Please read inline.
=20
Thanks
Partha

________________________________

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
Sent: Wed 3/16/2011 10:33 PM
To: Parthasarathi R (partr)
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing


Hi Partha,=20



	Thierry>> We could be more explicit, but should we ? (see next =
paragraph). Memory and CPU are just the two most obvious examples or =
resources that you need to process sessions. For instance, some of our =
servers have a fixed amount of sessions registers (for instance 50 000). =
In this case, the weight also reflects that. The algorithm we use for =
weights is very simple: this is the smallest remaining (in percentage) =
of each resource that is (1) Required to process a call, (2) Scarce  =
(not infinite). Let's consider the following system:=20
	The server has 70% of its memory available for calls, an average of 30% =
CPU usage (average over the last 10s for instance), and 40% of its =
registers available (remember that those resources may be shared by =
other components or devices for any other usage: it does not matter).=20
	Then the weight at this moment is MIN (70, 30, 40) =3D 30%. You then =
need to normalize this weight to match the weight range (an INT ? )=20
=09
	Now we are not saying that everybody should calculate their weight like =
that. What is important is that the weight reflects your current =
relative capacity. In our case, this "algorithm" balances the load of =
our servers very well. Now if you use a server that has a fixed amount =
of resources and that is dimensioned to process let's say 50 000 =
sessions, at a pace of 100 session/s, then you could probably derive a =
weight that would depend on both your count of sessions left, and your =
current rate of session/s. =20
	=20
	 In case every device calculate its own algorithm for weight, =
then weight from one device may not same as other device. Say dev1 takes =
Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =
=3D 30% and another device dev2 calculates Min (CPU, Mem) then weight =
will be Min (70, 40) =3D 40% even though they are equally loaded. It is =
very important to have the common unit to solve this issue  =20
=09

TB >> I think here you imply that the DSP is used as a finite pool of =
bearer channels resources, and your are probably using the DSP to =
process bearers. May be you are an SBC or so.
=20
 Here is the interesting part, dev1 achieve the specific service =
using DSP whereas dev2 achieve the same result using multi =
core/processor architecture or softDSP or ASIC implementation. IMO, =
weight varies based on the system architecture in case algorithm is not =
spelt out correctly or generic units are not defined. 
=20
 Then, it is obvious that you should also consider this resource as =
well, since you need it to process new sessions. So you can't say that =
you have 40% resource available to process calls if actually one of your =
resource needed has only 30% available left !=20
Our recommendation is that the occupation (weight) reflects the MIN of =
ANY resource that must be used for call processing and that is =
constrained.=20

I think it is on the contrary an advantage to not specify exactly how it =
is calculated, but to specify exactly what it is, or what it means =
instead (The most constrained resource ratio).=20


		If not, I could not understand your argument. As you mentioned, all =
system will know CPU or memory usage and it is possible to come up the =
static or refreshable value saying that load shed has to happen when 85% =
of CPU utilization or 90% of memory utilization. The different between =
our proposal is that there is no need of two value per message instead 1 =
value in each message & 1 value is static or in periodic manner  =


	Thierry>> We use a single value: see the example above. =20
	 I'm proposing is to use single real time value with units like =
number of dialog and exchange the threshold (max dialog) using some =
other static interface.=20

TB> This proposal is exactly what we tried to stay away from ! I thought =
I gave lot of arguments in this thread to explain  why working with =
dialogs is at the same time more difficult and not the most =
straightforward information to do resource based load sharing. =20
 dialog is only one of the unit. we will able to work out other =
unit which acceptable to everybody. As Paul Jones mentioned in another =
mail thread, SCE is working well in terms H.323 network for lot of years =
already but it is not working in SIP entity like SBC.   
=20
Moreover, using and sharing thresholds and trying to do overload control =
as well from there is also operationally very difficult: We did that =
before and it did not work, because (inside the same company !) nobody =
never agreed about what those threshold were and how to use them.=20
 I agree that threshold varies from device to device varies =
within the same company and huge difference will be seen in case of =
multi vendor scenario.  It may be tough to manage in case it is =
provisioned statically as the threshold has to be tweaked based on the =
solutions. It is the main reason, I proposed to pass the threshold value =
from device while booting and refresh it whenever required. =20
=20
Hope we both agrees to have that two variables x,y is required to solve =
this issue. Load balancer (upstream entity) will use f(x,y) based on its =
own definition to share the load and shed the load towards single =
device. The open issue is whether both x & y has to be passed every time =
or x shall be passed in real time and y shall be passed once in a while. =

=20
BTW, I have written Media overload control requirement =
(draft-partha-dispatch-sip-media-overload-control-00) draft to identify =
the common problem which will be agreed upon lot of vendors and solve it =
based on the requirement.=20


This is why we decided to separate load distribution and overload =
shedding (OLC provides a rejection rate, and load distribution provides =
a resource ratio). Our OLC mechanism what the starting point for Vijay's =
effort to lead the IETF OLC draft.=20


	Let overload decides how to reduce or when to shed the load based on =
this information. The advantage in this mechanism is that only one real =
time value is required between entities.

See my comment above: we tried that, and it was very difficult to come =
up with a common approach in all of our (IMS related) products.=20
Beside, having the destination clearly specifies the requested rejection =
rate is much more convenient and determinist than sharing "session =
available" and threshold information and letting the source node figure =
out what it really means in terms of shedding.=20


	In your proposed solution, "weight" as one value for load balancing and =
"oc" as another value for overload shedding. Pls correct me incase I =
misunderstand something here =20
=09

Yes, but the "oc"  valus is the one already proposed in the OLC draft: =
this is not in THIS proposed solution.=20

Cordially,=20

Thierry



	--
	--=20
	Cordially,=20
	Thierry Bessis   IMS Solution: Architecture & Product Management - ALTA =
Member
	  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
	=20
	  1960 Lucent Lane - Room 6L-517
	  Naperville, IL  60563  U S
	  Tel/Fax: +1 630 979 7989
	  Corporate IM: tbessis    - Yahoo IM:     thierry_bessis
=09
	My WEB page:   http://ihgpweb.ih.lucent.com/~tbessis/
	My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
	ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
=09
	Conference information:
	2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
	others countries see: http://all.alcatel-lucent.com/audio
=09
	Upcoming planned Business Trip: 12 April - 14 April 2011
	Upcoming planned Vacation: None=20
=09
	_______________________________________________
	dispatch mailing list
	dispatch@ietf.org
	https://www.ietf.org/mailman/listinfo/dispatch


--=20
--=20
Cordially,=20
Thierry Bessis=20
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
=20
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None=20

------_=_NextPart_002_01CBE430.6CCBA05D
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thierry's signature=0A=
=0A=
=0A=
=0A=

=0A=

Hi =
Thierry,
=0A=

 
=0A=

Please read =
inline.
=0A=

 
=0A=

Thanks
=0A=

Partha
=0A=


=0A=

=0A=
From: Thierry Bessis =
[mailto:Thierry.Bessis@alcatel-lucent.com]
Sent: Wed 3/16/2011 =
10:33 PM
To: Parthasarathi R (partr)
Cc: =
dispatch@ietf.org
Subject: Re: [dispatch] SIP load =
balancing

=0A=

Hi Partha, 

=0A=

=0A=

Thierry>> We could be more explicit, but should we ? (see =
next paragraph). Memory and CPU are just the two most obvious examples =
or resources that you need to process sessions. For instance, some of =
our servers have a fixed amount of sessions registers (for instance 50 =
000). In this case, the weight also reflects that. The algorithm we use =
for weights is very simple: this is the smallest remaining (in =
percentage) of each resource that is (1) Required to process a call, (2) =
Scarce  (not infinite). Let's consider the following system: =

The server has 70% of its memory available for calls, an average of =
30% CPU usage (average over the last 10s for instance), and 40% of its =
registers available (remember that those resources may be shared by =
other components or devices for any other usage: it does not matter). =

Then the weight at this moment is MIN (70, 30, 40) =3D 30%. You then =
need to normalize this weight to match the weight range (an INT ? ) =


Now we are not saying that everybody should calculate their =
weight like that. What is important is that the weight reflects your =
current relative capacity. In our case, this "algorithm" balances =
the load of our servers very well. Now if you use a server that has a =
fixed amount of resources and that is dimensioned to process let's say =
50 000 sessions, at a pace of 100 session/s, then you could probably =
derive a weight that would depend on both your count of sessions left, =
and your current rate of session/s.  
=0A=

 
=0A=

<Partha> In case every device calculate =
its own algorithm for weight, then weight from one device may not =
same as other device. Say dev1 takes Min(CPU, Mem, =
DSP) then weight will be calculated as Min (70, 40, 30) =
=3D 30% and another device dev2 calculates Min =
(CPU, Mem) then weight will be Min (70, 40) =3D 40% even =
though they are equally loaded. It is very important to have the common =
unit to solve this issue </Partha>  =

=0A=

TB >> I think here you imply that the DSP is used as a finite =
pool of bearer channels resources, and your are probably using the DSP =
to process bearers. May be you are an SBC or so.
=0A=

 
=0A=

<partha> Here is the interesting part, dev1 achieve the =
specific service using DSP whereas dev2 achieve the same result using =
multi core/processor architecture or softDSP or ASIC implementation. =
IMO, weight varies based on the system architecture in case algorithm is =
not spelt out correctly or generic units are not defined. =
</Partha>
=0A=

 
=0A=

 Then, it is obvious that you should also consider this =
resource as well, since you need it to process new sessions. So you =
can't say that you have 40% resource available to process calls if =
actually one of your resource needed has only 30% available left ! =

Our recommendation is that the occupation (weight) reflects the MIN =
of ANY resource that must be used for call processing and that is =
constrained. 

I think it is on the contrary an advantage to not =
specify exactly how it is calculated, but to specify exactly what it is, =
or what it means instead (The most constrained resource ratio). =

=0A=

=0A=

=0A=

If not, I could not understand your argument. As you =
mentioned, all system will know CPU or memory usage and it is possible =
to come up the static or refreshable value saying that load shed has to =
happen when 85% of CPU utilization or 90% of memory utilization. =
The different between our proposal is that there is no need of =
two value per message instead 1 value in each message & =
1 value is static or in periodic =
manner  </Partha>
=0A=

Thierry>> We use a single value: see the example =
above.  
=0A=

<Partha> I'm proposing is to use =
single real time value with units like number of dialog and =
exchange the threshold (max dialog) using some other static =
interface. 
=0A=

TB> This proposal is exactly what we tried to stay away from ! I =
thought I gave lot of arguments in this thread to explain  why =
working with dialogs is at the same time more difficult and not the most =
straightforward information to do resource based load sharing.  =

=0A=

<Partha> dialog is only one of the unit. we will able to work =
out other unit which acceptable to everybody. As Paul Jones mentioned in =
another mail thread, SCE is working well in terms H.323 =
network for lot of years already but it is not working in =
SIP entity like SBC.   </Partha>
=0A=

 
=0A=

Moreover, using and sharing thresholds and trying to do overload =
control as well from there is also operationally very difficult: We did =
that before and it did not work, because (inside the same company !) =
nobody never agreed about what those threshold were and how to use them. =

=0A=

=0A=

=0A=

<Partha> I agree that threshold =
varies from device to device varies within the same company and =
huge difference will be =
seen in case of multi vendor scenario.  It may be tough =
to manage in case it is provisioned statically as the threshold has to =
be tweaked based on the solutions. It is the main reason, I proposed to =
pass the threshold value from device while booting and refresh it =
whenever required.  
=0A=

 
=0A=

Hope we both agrees to have that =
two variables x,y is required to solve this issue. Load balancer (upstream =
entity) will use f(x,y) based on its own definition to share the =
load and shed the load towards single device. The open issue is whether =
both x & y has to be passed every time or x shall be passed in real =
time and y shall be passed once in a while. 
=0A=

 
=0A=

BTW, I have written Media overload =
control requirement (draft-partha-dispatch-sip-media-overload-control-00) =
draft to identify the common problem which will be agreed =
upon lot of vendors and solve it based on the =
requirement. 
=0A=

</partha>
=0A=

This is why we decided to separate load distribution and overload =
shedding (OLC provides a rejection rate, and load distribution provides =
a resource ratio). Our OLC mechanism what the starting point for Vijay's =
effort to lead the IETF OLC draft. 
=0A=

=0A=

Let overload decides how to reduce or when to shed the load =
based on this information. The advantage in this mechanism is that only =
one real time value is required between =
entities.
=0A=

See my comment above: we tried that, and it was very difficult to =
come up with a common approach in all of our (IMS related) products. =

Beside, having the destination clearly specifies the requested =
rejection rate is much more convenient and determinist than sharing =
"session available" and threshold information and letting the source =
node figure out what it really means in terms of shedding. 
=0A=

=0A=

In your proposed solution, "weight" as one value =
for load balancing and "oc" as another value for overload shedding. Pls =
correct me incase I misunderstand something here </Partha> =

=0A=

Yes, but the "oc"  valus is the one already proposed in the =
OLC draft: this is not in THIS proposed solution. 

Cordially, =


Thierry
=0A=

=0A=


=0A=

--
-- =

Cordially, 
Thierry =
Bessis   IMS Solution: Architecture & Product Management - ALTA =
Member
  Organization: ALU > Carrier group =
> Wireline BU >  IMS Sol Arch, SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis    - Yahoo IM:     =
thierry_bessis

My WEB page:   http://ihgpweb.ih.lucent.=
com/~tbessis/
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference =
information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - =
Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: =
12 April - 14 April 2011
Upcoming planned Vacation: None =

=0A=
_______________________________________________=0A=
dispatch mailing list=0A=
dispatch@ietf.org=0A=
https://www.ietf.=
org/mailman/listinfo/dispatch=0A=

=0A=


=0A=

-- 
-- =

Cordially, 
Thierry =
Bessis =0A=

=0A=

  IMS Solution: =
Architecture & Product Management - ALTA Member
  =
Organization: ALU > Carrier group > Wireline BU =
>  IMS Sol Arch, SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis

Engage: https://engage.=
alcatel-lucent.com/people/tbessis
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference =
information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - =
Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: =
12 April - 14 April 2011
Upcoming planned Vacation: None =


------_=_NextPart_002_01CBE430.6CCBA05D--

------_=_NextPart_001_01CBE430.6CCBA05D
Content-Type: image/jpeg;
	name="ATT19223371.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Description: ATT19223371.jpg
Content-Location: ATT19223371.jpg
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------_=_NextPart_001_01CBE430.6CCBA05D
Content-Type: image/jpeg;
	name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Description: al_corp_h_3d_16mm.jpg
Content-Disposition: inline;
	filename="al_corp_h_3d_16mm.jpg"
Content-Location: al_corp_h_3d_16mm.jpg
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------_=_NextPart_001_01CBE430.6CCBA05D--

From fluffy@cisco.com  Wed Mar 16 16:59:02 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D11CF3A6A2D for ; Wed, 16 Mar 2011 16:59:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -108.569
X-Spam-Level: 
X-Spam-Status: No, score=-108.569 tagged_above=-999 required=5 tests=[AWL=2.030, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ojppp-CTR5IJ for ; Wed, 16 Mar 2011 16:59:01 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id BD0963A6A28 for ; Wed, 16 Mar 2011 16:59:01 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=705; q=dns/txt; s=iport; t=1300320029; x=1301529629; h=subject:mime-version:from:in-reply-to:date:cc: content-transfer-encoding:message-id:references:to; bh=7MVjHi3/j37rwcJudynMKwCQoJjjoqjg8TrC0IbAiiM=; b=G7U9WAy8qz5j3ovGXsNxhCoM14hdFdrZ9iZZFE+9KHXrPKYo6bz9NEoj 5J8yPeZ/8w4nh7x5gDFS8N0tH3j/4awDkLYT4/nAxoSJrM6nsE89Nms4P mFa6PyhVWQUp5yQjlr/B1hM+3x3CS7XV7BmwaoAugtDijilvRN7Go6Ow2 s=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAKLrgE2tJV2a/2dsb2JhbAClTneldJxUhWMEhS+HL4NN
X-IronPort-AV: E=Sophos;i="4.63,196,1299456000"; d="scan'208";a="226350746"
Received: from rcdn-core-3.cisco.com ([173.37.93.154]) by rtp-iport-2.cisco.com with ESMTP; 17 Mar 2011 00:00:28 +0000
Received: from [192.168.4.100] (rcdn-fluffy-8712.cisco.com [10.99.9.19]) by rcdn-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p2H00R0T003967;  Thu, 17 Mar 2011 00:00:27 GMT
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
From: Cullen Jennings 
In-Reply-To: 
Date: Wed, 16 Mar 2011 18:03:16 -0600
Content-Transfer-Encoding: quoted-printable
Message-Id: <37088805-C32B-49D1-874E-02E51BCDA890@cisco.com>
References: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com> 
To: Lawrence Conroy 
X-Mailer: Apple Mail (2.1082)
Cc: DISPATCH list 
Subject: Re: [dispatch] DISPATCH Ad-HOC meetings about VIPR and RTCWeb at IETF 80
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 16 Mar 2011 23:59:02 -0000

On Mar 16, 2011, at 4:50 PM, Lawrence Conroy wrote:

>=20
> Also, the choice of Friday @ 14:15 for the "RTCWEB" chat seems kinda =
sub-optimal,
> considering that's BEHAVE's traditional slot.

I'll note that behave has 4 sessions and also conflicts with SIPCLF, SIP =
Overload, XMPP, and Splices. If there is something of specific interests =
to both the Behave folks and RTCWeb ad-hoc meeting, we can work with the =
behave chairs to try and keep that topic in a behave session that does =
not conflict with RTCWeb.=20

Cullen=20

Note the number of WG in RAI alone is going to ensure we have conflicts =
- I think I mention this every time dispatch HUMS on forming a new WG =
:-)=20



From HKaplan@acmepacket.com  Wed Mar 16 17:37:41 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C2CC53A68DE for ; Wed, 16 Mar 2011 17:37:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.947
X-Spam-Level: 
X-Spam-Status: No, score=-0.947 tagged_above=-999 required=5 tests=[AWL=0.452,  BAYES_00=-2.599, J_CHICKENPOX_33=0.6, J_CHICKENPOX_52=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2WxdLD8vcZHj for ; Wed, 16 Mar 2011 17:37:40 -0700 (PDT)
Received: from etmail.acmepacket.com (etmail.acmepacket.com [216.41.24.6]) by core3.amsl.com (Postfix) with ESMTP id B02AE3A67EE for ; Wed, 16 Mar 2011 17:37:40 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by etmail.acmepacket.com (216.41.24.6) with Microsoft SMTP Server (TLS) id 8.2.254.0; Wed, 16 Mar 2011 20:39:06 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Wed, 16 Mar 2011 20:39:06 -0400
From: Hadriel Kaplan 
To: "Parthasarathi R (partr)" 
Date: Wed, 16 Mar 2011 20:39:05 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvkO7iIglfI7dCCTqK0Hxuq/SXR0Q==
Message-ID: <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com>
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> 
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAQAAAUA=
Cc: "dispatch@ietf.org" , Thierry Bessis 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 00:37:41 -0000

Hey Partha,=20
comments inline...


On Mar 16, 2011, at 7:18 PM, Parthasarathi R (partr) wrote:

>>=20
>>  In case every device calculate its own algorithm for weight, th=
en weight from one device may not same as other device. Say dev1 takes Min(=
CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =3D 30% a=
nd another device dev2 calculates Min (CPU, Mem) then weight will be Min (7=
0, 40) =3D 40% even though they are equally loaded. It is very important to=
 have the common unit to solve this issue 

HK> By definition if dev1 calculates 30% and dev2 calculates 40%, then they=
 are *not* equally loaded.  Unless you're just saying "well what if they fo=
rgot to take a resource bottleneck into account". (in which case all bets a=
re off anyway - it would be a bug)


>> =20
> TB >> I think here you imply that the DSP is used as a finite pool of bea=
rer channels resources, and your are probably using the DSP to process bear=
ers. May be you are an SBC or so.
> =20
>  Here is the interesting part, dev1 achieve the specific service =
using DSP whereas dev2 achieve the same result using multi core/processor a=
rchitecture or softDSP or ASIC implementation. IMO, weight varies based on =
the system architecture in case algorithm is not spelt out correctly or gen=
eric units are not defined. 


HK> But that's the point - we can't ever write a spec to specify all possib=
le inputs into an algorithm for every possible architecture, because each a=
rchitecture has its own limitations/bottlenecks.  For example you've mentio=
ned DSP's, and Thierry has mentioned CPU and memory; but there're also SRTP=
 encryption resources, media recording resources, DTMF interworking resourc=
es, v4-v6 interworking resources, TCP-media resources, and so on.  In some =
architectures the resources are the same, in some not.  Some SIP devices ha=
ve no such resources, some have more.  We can't possibly enumerate them all=
 in a spec.

HK> So instead, the spec will just say: "whatever your limiting factor is a=
t any given time, use it for your numbers".  Only the developers of the sys=
tems know what their limiting factors are, after all.


> =20
>>=20
>>  I'm proposing is to use single real time value with units like =
number of dialog and exchange the threshold (max dialog) using some other s=
tatic interface.=20
>  dialog is only one of the unit. we will able to work out other u=
nit which acceptable to everybody. As Paul Jones mentioned in another mail =
thread, SCE is working well in terms H.323 network for lot of years already=
 but it is not working in SIP entity like SBC.   

HK> It worked in H.323 because everything was a call and consumed equal res=
ources, and you were treating it like a PRI with a fixed/configured number =
of channels, and things rarely changed.  Arguably that's frequently also tr=
ue for SIP across SIP-trunks.  Unfortunately, that's the *only* case it wor=
ks well in, afaict.  Inside the service provider and peering between them t=
hings aren't that "clean" (or restricted, depending on how you view the wor=
ld).  It's much harder to have fixed limits of dialogs/channels/calls betwe=
en a big mesh network of systems, where the members of the mesh (the number=
 of discrete systems/blades) grow and shrink relatively frequently.  And it=
's gotten even worse with SUBSCRIBE dialogs and other non-call/non-register=
 traffic consuming resources.  That's not to say it isn't possible to do, b=
ut it ends up wasting resources because you can't take advantage of statist=
ical multiplexing, and you have to re-configure lot of systems frequently.

-hadriel


From partr@cisco.com  Wed Mar 16 17:43:14 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DFE563A6906; Wed, 16 Mar 2011 17:43:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.292
X-Spam-Level: 
X-Spam-Status: No, score=-9.292 tagged_above=-999 required=5 tests=[AWL=-0.090, BAYES_00=-2.599, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qvkuqOSQsEW0; Wed, 16 Mar 2011 17:43:13 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id 6BF2D3A68DE; Wed, 16 Mar 2011 17:43:13 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=3432; q=dns/txt; s=iport; t=1300322681; x=1301532281; h=mime-version:subject:date:message-id:from:to:cc; bh=q7Z1Mw4k7VXsAbRQJK5n80hCRWnpjLoGvTJDBJ3UT40=; b=Qj1MV8TR7M0V0LQwRDaBLcNxjlSbHtQ4HM5i+vrKeNAZJxJhb93P6ZRu Dq315jTYatZwvu1WmLMr0Qhn7FBBb8oJ9BtGSulKjmnQvuHg5Q7mTF49j Bi/GLiVu6GhNcinzAhp+gSwBYViCHi/xl+7q0YUpPGxO0veLSweMs1o22 Q=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Aq8EAGr2gE2Q/khMgWdsb2JhbACkfFMUAQEWJiWmHpxegnuCaASFLYsE
X-IronPort-AV: E=Sophos;i="4.63,196,1299456000"; d="scan'208,217";a="79530412"
Received: from ams-core-3.cisco.com ([144.254.72.76]) by ams-iport-1.cisco.com with ESMTP; 17 Mar 2011 00:44:39 +0000
Received: from xbh-bgl-412.cisco.com (xbh-bgl-412.cisco.com [72.163.129.202]) by ams-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p2H0icJt008871; Thu, 17 Mar 2011 00:44:39 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-412.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Thu, 17 Mar 2011 06:14:38 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01CBE43C.7E7458A2"
Date: Thu, 17 Mar 2011 06:14:38 +0530
Message-ID: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Bar BOF request for SIP Load balancing & (media) overload handling
Thread-Index: AcvkNQDWIeXRlo1yRRq3Yg+LUWi3ZA==
From: "Parthasarathi R (partr)" 
To: , "Cullen Jennings (fluffy)" , "Mike Hammer (hmmr)" , , , ,  , "Paul Kyzivat (pkyzivat)" , , , , , , 
X-OriginalArrivalTime: 17 Mar 2011 00:44:38.0848 (UTC) FILETIME=[7EE7C400:01CBE43C]
X-Mailman-Approved-At: Wed, 16 Mar 2011 20:32:34 -0700
Cc: dispatch@ietf.org, sip-overload@ietf.org
Subject: [dispatch] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 00:43:15 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBE43C.7E7458A2
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi all,
=20
There is lot of discussion about SIP load balancing and overload =
handling in Dispatch alias. I wish to have discussion about SIP Load =
balancing & (media) overload handling at IETF-80 Prague. I guess that =
Bar BOF is the idle way to start the discussion about this topic to =
common understanding and identify the way to move forward. Please =
suggest me in case any other mechanism exists in IETF.
=20
I have included the folks who has shown the interest in this topic as =
part of SIP Load balancing & overload handling, dispatch WG chairs, CCed =
to Dispatch and SoC WG alias. AFAIK, the couple of drafts are related to =
this problem:
=20
draft-bessis-dispatch-adaptive-load-balancing-00
draft-partha-dispatch-sip-media-overload-control-00.txt
draft-jones-sip-overload-sce-00
draft-partha-dispatch-resource-availability-00=20
draft-sparks-sipping-load-00
=20
Please let me about others interest and opinion to have Bar BOF for SIP =
Load balancing & overload handling during IETF-80.
=20
Thanks
Partha

------_=_NextPart_001_01CBE43C.7E7458A2
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

=0A=
=0A=
=0A=
=0A=

=0A=

Hi all,
=0A=

 
=0A=

There is lot of discussion about SIP =
load balancing and overload handling in Dispatch alias. I wish to have =
discussion about SIP Load balancing & (media) overload handling at =
IETF-80 Prague. I guess that Bar BOF is the idle way to start the =
discussion about this topic to common understanding and identify the way =
to move forward. Please suggest me in case any other mechanism exists in =
IETF.
=0A=

 
=0A=

I have included the folks who has shown =
the interest in this topic as part of SIP Load balancing & overload =
handling, dispatch WG chairs, =
CCed to Dispatch and SoC WG alias. AFAIK, the couple of drafts are =
related to this problem:
=0A=

 
=0A=

draft-bessis-dispatch-adaptive-load-balancing-00
draft-pa=
rtha-dispatch-sip-media-overload-control-00.txt
draft-jones-sip-overlo=
ad-sce-00
draft-partha-dispatch-resource-availability-00 =

draft-sparks-sipping-load-00
=0A=

 
=0A=

Please let me about others interest and =
opinion to have Bar BOF for SIP Load balancing & overload handling =
during IETF-80.
=0A=

 
=0A=

Thanks
=0A=

Partha

------_=_NextPart_001_01CBE43C.7E7458A2--

From fluffy@cisco.com  Wed Mar 16 19:14:19 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9D2BC3A68EF; Wed, 16 Mar 2011 19:14:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -108.754
X-Spam-Level: 
X-Spam-Status: No, score=-108.754 tagged_above=-999 required=5 tests=[AWL=1.845, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3VUlzeEH4EcJ; Wed, 16 Mar 2011 19:14:18 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 689B93A68CE; Wed, 16 Mar 2011 19:14:18 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=1289; q=dns/txt; s=iport; t=1300328146; x=1301537746; h=subject:mime-version:from:in-reply-to:date:cc: content-transfer-encoding:message-id:references:to; bh=JOqP5k9E7v/2Nl3L8TqW3oinbfi4+HSVFB5WljN3NGU=; b=kgxSxF/DiC583a6ArBuYkKv35lnClG/8A0c4hAJmPNVXpE2cOSs2UOX7 UkSp16BC1gshMoZU6nup83QdxhMUkWPxuSJgNNAYdJuIKr3201A6gUokI 3DZM+avy5bdonr4l0uBi0h0zvgk1fcyxHuADeI1WlBaAyR0XFssexrxAO U=;
X-IronPort-AV: E=Sophos;i="4.63,197,1299456000"; d="scan'208";a="226371636"
Received: from rcdn-core-4.cisco.com ([173.37.93.155]) by rtp-iport-2.cisco.com with ESMTP; 17 Mar 2011 02:15:44 +0000
Received: from [192.168.4.100] (rcdn-fluffy-8712.cisco.com [10.99.9.19]) by rcdn-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2H2FgcT001648;  Thu, 17 Mar 2011 02:15:42 GMT
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
From: Cullen Jennings 
In-Reply-To: 
Date: Wed, 16 Mar 2011 20:18:32 -0600
Content-Transfer-Encoding: quoted-printable
Message-Id: 
References: 
To: Parthasarathi R (partr) 
X-Mailer: Apple Mail (2.1082)
X-Mailman-Approved-At: Wed, 16 Mar 2011 20:32:34 -0700
Cc: vpascual@acmepacket.com, sip-overload@ietf.org, keith.drage@alcatel-lucent.com, HKaplan@acmepacket.com, dispatch@ietf.org, Thierry.Bessis@alcatel-lucent.com
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 02:14:19 -0000

Just as FYI ... if you have not seen it, you might want to have a read =
of=20
http://tools.ietf.org/html/draft-eggert-successful-bar-bof


On Mar 16, 2011, at 6:44 PM, Parthasarathi R (partr) wrote:

> Hi all,
> =20
> There is lot of discussion about SIP load balancing and overload =
handling in Dispatch alias. I wish to have discussion about SIP Load =
balancing & (media) overload handling at IETF-80 Prague. I guess that =
Bar BOF is the idle way to start the discussion about this topic to =
common understanding and identify the way to move forward. Please =
suggest me in case any other mechanism exists in IETF.
> =20
> I have included the folks who has shown the interest in this topic as =
part of SIP Load balancing & overload handling, dispatch WG chairs, CCed =
to Dispatch and SoC WG alias. AFAIK, the couple of drafts are related to =
this problem:
> =20
> draft-bessis-dispatch-adaptive-load-balancing-00
> draft-partha-dispatch-sip-media-overload-control-00.txt
> draft-jones-sip-overload-sce-00
> draft-partha-dispatch-resource-availability-00=20
> draft-sparks-sipping-load-00
> =20
> Please let me about others interest and opinion to have Bar BOF for =
SIP Load balancing & overload handling during IETF-80.
> =20
> Thanks
> Partha


From HKaplan@acmepacket.com  Wed Mar 16 19:59:32 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 97B503A6A21; Wed, 16 Mar 2011 19:59:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.773
X-Spam-Level: 
X-Spam-Status: No, score=-1.773 tagged_above=-999 required=5 tests=[AWL=0.826,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CC6NpNwjcFRk; Wed, 16 Mar 2011 19:59:31 -0700 (PDT)
Received: from etmail.acmepacket.com (etmail.acmepacket.com [216.41.24.6]) by core3.amsl.com (Postfix) with ESMTP id A04CC3A6A1F; Wed, 16 Mar 2011 19:59:31 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by etmail.acmepacket.com (216.41.24.6) with Microsoft SMTP Server (TLS) id 8.2.254.0; Wed, 16 Mar 2011 23:00:57 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Wed, 16 Mar 2011 23:00:57 -0400
From: Hadriel Kaplan 
To: Cullen Jennings 
Date: Wed, 16 Mar 2011 23:00:55 -0400
Thread-Topic: Bar BOF request for SIP Load balancing & (media) overload handling
Thread-Index: AcvkT4k3HPU5zNPIQb+aOQOg62ZNCw==
Message-ID: 
References:  
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAQAAAUA=
X-Mailman-Approved-At: Wed, 16 Mar 2011 20:32:34 -0700
Cc: Victor Pascual , "sip-overload@ietf.org" , "keith.drage@alcatel-lucent.com" , "Thierry.Bessis@alcatel-lucent.com" , "dispatch@ietf.org" 
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 02:59:32 -0000

Lars should let Robert be the editor for this draft.
That way it'll be "draft-sparks-successful-bar-bof" which would be a great =
title.
:)


On Mar 16, 2011, at 10:18 PM, Cullen Jennings wrote:

>=20
> Just as FYI ... if you have not seen it, you might want to have a read of=
=20
> http://tools.ietf.org/html/draft-eggert-successful-bar-bof
>=20
>=20
> On Mar 16, 2011, at 6:44 PM, Parthasarathi R (partr) wrote:
>=20
>> Hi all,
>>=20
>> There is lot of discussion about SIP load balancing and overload handlin=
g in Dispatch alias. I wish to have discussion about SIP Load balancing & (=
media) overload handling at IETF-80 Prague. I guess that Bar BOF is the idl=
e way to start the discussion about this topic to common understanding and =
identify the way to move forward. Please suggest me in case any other mecha=
nism exists in IETF.
>>=20
>> I have included the folks who has shown the interest in this topic as pa=
rt of SIP Load balancing & overload handling, dispatch WG chairs, CCed to D=
ispatch and SoC WG alias. AFAIK, the couple of drafts are related to this p=
roblem:
>>=20
>> draft-bessis-dispatch-adaptive-load-balancing-00
>> draft-partha-dispatch-sip-media-overload-control-00.txt
>> draft-jones-sip-overload-sce-00
>> draft-partha-dispatch-resource-availability-00=20
>> draft-sparks-sipping-load-00
>>=20
>> Please let me about others interest and opinion to have Bar BOF for SIP =
Load balancing & overload handling during IETF-80.
>>=20
>> Thanks
>> Partha
>=20


From paulej@packetizer.com  Thu Mar 17 03:18:40 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B55CB3A6839 for ; Thu, 17 Mar 2011 03:18:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.799
X-Spam-Level: 
X-Spam-Status: No, score=-0.799 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, J_CHICKENPOX_21=0.6, J_CHICKENPOX_33=0.6, J_CHICKENPOX_52=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uxXBw0BF1cz1 for ; Thu, 17 Mar 2011 03:18:36 -0700 (PDT)
Received: from dublin.packetizer.com (dublin.packetizer.com [75.101.130.125]) by core3.amsl.com (Postfix) with ESMTP id 750C13A6806 for ; Thu, 17 Mar 2011 03:18:36 -0700 (PDT)
Received: from [156.106.218.97] ([156.106.218.97]) (authenticated bits=0) by dublin.packetizer.com (8.14.4/8.14.4) with ESMTP id p2HAJqJT032166 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Thu, 17 Mar 2011 06:19:58 -0400
DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=packetizer.com; s=dublin; t=1300357199; bh=lwymozB7JCz5WZ7zaocQ6WjdOV4NTeyHoKJZVEVxTNc=; h=Message-ID:Date:From:MIME-Version:To:CC:Subject:References: In-Reply-To:Content-Type:Content-Transfer-Encoding; b=epI+nsnjW9LYybZ95o8bfIjG/2/u7wRr+yilOLpfact7/HQ7DGXqxwhu5AVnbM2Jw Zc15Udl9XT8nnjBqzf9QqZs6posCe0GHN/0l0ZOsYuY/srHicXRm1FKVEnqCPmM0Gs kfGMKUC/zbIBWlh4zpMmRUrebhu3h/3F6Vb8q10g=
Message-ID: <4D81E047.60809@packetizer.com>
Date: Thu, 17 Mar 2011 06:19:51 -0400
From: "Paul E. Jones" 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Hadriel Kaplan 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	 <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com>
In-Reply-To: <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: quoted-printable
Cc: "dispatch@ietf.org" , Thierry Bessis 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 10:18:40 -0000

Hadriel,

SCE did not work for H.323 because H.323 was mostly about call counting=20
and using a fixed number of resources.  SCE was designed with the notion =

that calls use different resources, actually.  That's why the SCE value=20
does not increase or decrease by one, necessarily.  If the limiting=20
resource was DS0s, I'd agree, but that's not the case most of the time.  =

That's true of both H.323 and SIP, of course.

Once convergence is reached, SCE values for non-DS0-type calls often=20
increase or decrease by one if the device processing the call does not=20
handle any part of the media.  The reason is that call state information =

often consumes about the same amount of memory, etc.  This is not always =

the case, of course, so SCE was not designed with the preconceived=20
notion that calls did consume an equal amount of resources.  It was=20
designed, though, to consider what the average call consumed.

Paul

On 3/16/2011 8:39 PM, Hadriel Kaplan wrote:
> Hey Partha,
> comments inline...
>
>
> On Mar 16, 2011, at 7:18 PM, Parthasarathi R (partr) wrote:
>
>>>   In case every device calculate its own algorithm for weight=
, then weight from one device may not same as other device. Say dev1 take=
s Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =3D=
 30% and another device dev2 calculates Min (CPU, Mem) then weight will b=
e Min (70, 40) =3D 40% even though they are equally loaded. It is very im=
portant to have the common unit to solve this issue
> HK>  By definition if dev1 calculates 30% and dev2 calculates 40%, then=
 they are *not* equally loaded.  Unless you're just saying "well what if =
they forgot to take a resource bottleneck into account". (in which case a=
ll bets are off anyway - it would be a bug)
>
>
>>>
>> TB>>  I think here you imply that the DSP is used as a finite pool of =
bearer channels resources, and your are probably using the DSP to process=
 bearers. May be you are an SBC or so.
>>
>>   Here is the interesting part, dev1 achieve the specific serv=
ice using DSP whereas dev2 achieve the same result using multi core/proce=
ssor architecture or softDSP or ASIC implementation. IMO, weight varies b=
ased on the system architecture in case algorithm is not spelt out correc=
tly or generic units are not defined.
>
> HK>  But that's the point - we can't ever write a spec to specify all p=
ossible inputs into an algorithm for every possible architecture, because=
 each architecture has its own limitations/bottlenecks.  For example you'=
ve mentioned DSP's, and Thierry has mentioned CPU and memory; but there'r=
e also SRTP encryption resources, media recording resources, DTMF interwo=
rking resources, v4-v6 interworking resources, TCP-media resources, and s=
o on.  In some architectures the resources are the same, in some not.  So=
me SIP devices have no such resources, some have more.  We can't possibly=
 enumerate them all in a spec.
>
> HK>  So instead, the spec will just say: "whatever your limiting factor=
 is at any given time, use it for your numbers".  Only the developers of =
the systems know what their limiting factors are, after all.
>
>
>>
>>>   I'm proposing is to use single real time value with units l=
ike number of dialog and exchange the threshold (max dialog) using some o=
ther static interface.
>>   dialog is only one of the unit. we will able to work out oth=
er unit which acceptable to everybody. As Paul Jones mentioned in another=
 mail thread, SCE is working well in terms H.323 network for lot of years=
 already but it is not working in SIP entity like SBC.
> HK>  It worked in H.323 because everything was a call and consumed equa=
l resources, and you were treating it like a PRI with a fixed/configured =
number of channels, and things rarely changed.  Arguably that's frequentl=
y also true for SIP across SIP-trunks.  Unfortunately, that's the *only* =
case it works well in, afaict.  Inside the service provider and peering b=
etween them things aren't that "clean" (or restricted, depending on how y=
ou view the world).  It's much harder to have fixed limits of dialogs/cha=
nnels/calls between a big mesh network of systems, where the members of t=
he mesh (the number of discrete systems/blades) grow and shrink relativel=
y frequently.  And it's gotten even worse with SUBSCRIBE dialogs and othe=
r non-call/non-register traffic consuming resources.  That's not to say i=
t isn't possible to do, but it ends up wasting resources because you can'=
t take advantage of statistical multiplexing, and you have to re-configur=
e lot of systems frequently.
>
> -hadriel
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


From andrew.hutton@siemens-enterprise.com  Thu Mar 17 03:23:26 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 74C0D3A686C for ; Thu, 17 Mar 2011 03:23:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.95
X-Spam-Level: 
X-Spam-Status: No, score=-1.95 tagged_above=-999 required=5 tests=[AWL=0.422,  BAYES_00=-2.599, SARE_SUB_OBFU_Q1=0.227]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 436uTln3k5fJ for ; Thu, 17 Mar 2011 03:23:22 -0700 (PDT)
Received: from ms03.m0019.fra.mmp.de.bt.com (m0019.fra.mmp.de.bt.com [62.180.227.30]) by core3.amsl.com (Postfix) with ESMTP id 0BDD13A6839 for ; Thu, 17 Mar 2011 03:23:21 -0700 (PDT)
Received: from senmx12-mx ([62.134.46.10] [62.134.46.10]) by ms03.m0020.fra.mmp.de.bt.com with ESMTP id BT-MMP-3767694; Thu, 17 Mar 2011 11:24:46 +0100
Received: from MCHP063A.global-ad.net (unknown [172.29.37.61]) by senmx12-mx (Server) with ESMTP id D9BE723F0278; Thu, 17 Mar 2011 11:24:46 +0100 (CET)
Received: from MCHP058A.global-ad.net ([172.29.37.55]) by MCHP063A.global-ad.net ([172.29.37.61]) with mapi; Thu, 17 Mar 2011 11:24:46 +0100
From: "Hutton, Andrew" 
To: Christer Holmberg , "rtc-web@alvestrand.no" , "dispatch@ietf.org" 
Date: Thu, 17 Mar 2011 11:24:44 +0100
Thread-Topic: Draft new version: draft-holmberg-rtcweb-ucreqs-01 - API Reqs
Thread-Index: AcvhtBoKcB/pe1tiQlOtldNCEQq7+wAi0FAwAJGCbjA=
Message-ID: <101C6067BEC68246B0C3F6843BCCC1E30B61E2F98F@MCHP058A.global-ad.net>
References: ,   <4D7D199F.6000900@alvestrand.no> <7F2072F1E0DE894DA4B517B93C6A05851948FF3678@ESESSCMS0356.eemea.ericsson.se>
In-Reply-To: <7F2072F1E0DE894DA4B517B93C6A05851948FF3678@ESESSCMS0356.eemea.ericsson.se>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [dispatch] Draft new version: draft-holmberg-rtcweb-ucreqs-01 - API Reqs
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 10:23:26 -0000

Hi,

I think this is useful input.

A few comments.

1. A5 "The web application MUST be able to control the media format (codec)=
 to be used for the streams sent to a peer".  =20

It should also be able to control the codec used for the received stream (i=
.e. control what if offered to the peer).

2. A8 "The web application MUST be able to pause/unpause the sending of a s=
tream to a peer".

I think it should be able to set/change the direction of the stream(s) whic=
h in SDP terms means sendonly/recvonly/sendrecv/inactive.=20
=20
3. A12 "The web application MUST be informed when a stream from a peer is n=
o longer received".=20

The web application MUST also be informed when a stream from a peer starts =
to be received. This probably also needs to take account of some early medi=
a scenarios when multiple streams could potentially be received from multip=
le peers. The application needs to be able to control which streams are ren=
dered. This also impacts the browser requirements as obviously the browser =
needs to be able to detect when the stream starts etc.

Regards
Andy






> -----Original Message-----
> From: rtc-web-bounces@alvestrand.no=20
> [mailto:rtc-web-bounces@alvestrand.no] On Behalf Of Christer Holmberg
> Sent: 14 March 2011 12:03
> To: rtc-web@alvestrand.no; dispatch@ietf.org
> Subject: [RTW] Draft new version: draft-holmberg-rtcweb-ucreqs-01
>=20
>=20
> Hi,
>=20
> We've submitted a new version of the ucreqs draft.
>=20
> The new version contains fixes and additions based on=20
> comments given on the previous version on the list.
>=20
> The draft can be found at:=20
> http://www.ietf.org/id/draft-holmberg-rtcweb-ucreqs-01.txt
>=20
> Regards,
>=20
> Christer
> _______________________________________________
> RTC-Web mailing list
> RTC-Web@alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/rtc-web
> =

From jgunn6@csc.com  Thu Mar 17 07:25:17 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D07423A6A08; Thu, 17 Mar 2011 07:25:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.598
X-Spam-Level: 
X-Spam-Status: No, score=-6.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AaQDYgHPr+wg; Thu, 17 Mar 2011 07:25:15 -0700 (PDT)
Received: from mail63.messagelabs.com (mail63.messagelabs.com [216.82.242.83]) by core3.amsl.com (Postfix) with ESMTP id 245143A6827; Thu, 17 Mar 2011 07:25:14 -0700 (PDT)
X-VirusChecked: Checked
X-Env-Sender: jgunn6@csc.com
X-Msg-Ref: server-15.tower-63.messagelabs.com!1300371900!134876145!1
X-StarScan-Version: 6.2.9; banners=-,-,-
X-Originating-IP: [20.137.2.87]
Received: (qmail 28877 invoked from network); 17 Mar 2011 14:25:00 -0000
Received: from amer-mta101.csc.com (HELO amer-mta101.csc.com) (20.137.2.87) by server-15.tower-63.messagelabs.com with DHE-RSA-AES256-SHA encrypted SMTP; 17 Mar 2011 14:25:00 -0000
Received: from amer-gw09.amer.csc.com (amer-gw09.amer.csc.com [20.6.39.245]) by amer-mta101.csc.com (Switch-3.4.3/Switch-3.3.3mp) with ESMTP id p2HEOdCv006438; Thu, 17 Mar 2011 10:24:57 -0400
In-Reply-To: 
References: 
To: "Parthasarathi R (partr)" 
MIME-Version: 1.0
X-KeepSent: CEDB53D9:C5DB51F1-85257856:004EFA4D; type=4; name=$KeepSent
X-Mailer: Lotus Notes Release 8.0.2FP1  CCH2 April 23, 2009
From: Janet P Gunn 
Message-ID: 
Date: Thu, 17 Mar 2011 10:24:49 -0400
X-MIMETrack: Serialize by Router on AMER-GW09/SRV/CSC(Release 8.5.2FP1 HF29|January 09, 2011) at 03/17/2011 10:23:59 AM, Serialize complete at 03/17/2011 10:23:59 AM
Content-Type: multipart/alternative; boundary="=_alternative 004F2DB185257856_="
X-Mailman-Approved-At: Thu, 17 Mar 2011 07:52:46 -0700
Cc: dispatch@ietf.org, sip-overload@ietf.org, keith.drage@alcatel-lucent.com, dispatch-bounces@ietf.org, HKaplan@acmepacket.com, vpascual@acmepacket.com, Thierry.Bessis@alcatel-lucent.com
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media)	overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 14:25:18 -0000

This is a multipart message in MIME format.
--=_alternative 004F2DB185257856_=
Content-Type: text/plain; charset="US-ASCII"

Partha,

I am definitely interested in the subject, but I will, unfortunately, not 
be in Prague.

If it were possible to hold the bar bof somewhere with an audio feed and 
jabber room, I would definitely participate remotely.

Thanks

Janet

This is a PRIVATE message. If you are not the intended recipient, please 
delete without copying and kindly advise us by e-mail of the mistake in 
delivery. 
NOTE: Regardless of content, this e-mail shall not operate to bind CSC to 
any order or other contract unless pursuant to explicit written agreement 
or government initiative expressly permitting the use of e-mail for such 
purpose.



From:
"Parthasarathi R (partr)" 
To:
, "Cullen Jennings (fluffy)" , "Mike 
Hammer (hmmr)" , , 
, , 
, "Paul Kyzivat (pkyzivat)" 
, , , 
, , 
, 
Cc:
dispatch@ietf.org, sip-overload@ietf.org
Date:
03/16/2011 11:34 PM
Subject:
[dispatch] Bar BOF request for SIP Load balancing & (media)     overload 
handling



Hi all,
 
There is lot of discussion about SIP load balancing and overload handling 
in Dispatch alias. I wish to have discussion about SIP Load balancing & 
(media) overload handling at IETF-80 Prague. I guess that Bar BOF is the 
idle way to start the discussion about this topic to common understanding 
and identify the way to move forward. Please suggest me in case any other 
mechanism exists in IETF.
 
I have included the folks who has shown the interest in this topic as part 
of SIP Load balancing & overload handling, dispatch WG chairs, CCed to 
Dispatch and SoC WG alias. AFAIK, the couple of drafts are related to this 
problem:
 
draft-bessis-dispatch-adaptive-load-balancing-00
draft-partha-dispatch-sip-media-overload-control-00.txt
draft-jones-sip-overload-sce-00
draft-partha-dispatch-resource-availability-00 
draft-sparks-sipping-load-00
 
Please let me about others interest and opinion to have Bar BOF for SIP 
Load balancing & overload handling during IETF-80.
 
Thanks
Partha_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch



--=_alternative 004F2DB185257856_=
Content-Type: text/html; charset="US-ASCII"



Partha,



I am definitely interested in the subject,
but I will, unfortunately, not be in Prague.



If it were possible to hold the bar
bof somewhere with an audio feed and jabber room, I would definitely participate
remotely.



Thanks



Janet



This is a PRIVATE message. If you are not the intended recipient, please
delete without copying and kindly advise us by e-mail of the mistake in
delivery. 

NOTE: Regardless of content, this e-mail shall not operate to bind CSC
to any order or other contract unless pursuant to explicit written agreement
or government initiative expressly permitting the use of e-mail for such
purpose.










From:
"Parthasarathi R (partr)"
<partr@cisco.com>


To:
<dworley@avaya.com>, "Cullen
Jennings (fluffy)" <fluffy@cisco.com>, "Mike Hammer (hmmr)"
<hmmr@cisco.com>, <HKaplan@acmepacket.com>, <keith.drage@alcatel-lucent.com>,
<john.elwell@siemens-enterprise.com>, <mary.ietf.barnes@gmail.com>,
"Paul Kyzivat (pkyzivat)" <pkyzivat@cisco.com>, <paulej@packetizer.com>,
<rjsparks@nostrum.com>, <spromano@unina.it>, <Thierry.Bessis@alcatel-lucent.com>,
<vkg@bell-labs.com>, <vpascual@acmepacket.com>


Cc:
dispatch@ietf.org, sip-overload@ietf.org


Date:
03/16/2011 11:34 PM


Subject:
[dispatch] Bar BOF request for SIP Load
balancing & (media)        overload handling










Hi all,

 

There is lot of discussion about SIP load
balancing and overload handling in Dispatch alias. I wish to have discussion
about SIP Load balancing & (media) overload handling at IETF-80 Prague.
I guess that Bar BOF is the idle way to start the discussion about this
topic to common understanding and identify the way to move forward. Please
suggest me in case any other mechanism exists in IETF.

 

I have included the folks who has shown the
interest in this topic as part of SIP Load balancing & overload handling,
dispatch WG chairs, CCed to Dispatch and SoC WG alias. AFAIK, the couple
of drafts are related to this problem:

 

draft-bessis-dispatch-adaptive-load-balancing-00

draft-partha-dispatch-sip-media-overload-control-00.txt

draft-jones-sip-overload-sce-00

draft-partha-dispatch-resource-availability-00 

draft-sparks-sipping-load-00

 

Please let me about others interest and opinion
to have Bar BOF for SIP Load balancing & overload handling during IETF-80.

 

Thanks

Partha_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.org/mailman/listinfo/dispatch






--=_alternative 004F2DB185257856_=--

From dworley@avaya.com  Thu Mar 17 08:21:52 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3AD803A6968; Thu, 17 Mar 2011 08:21:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.545
X-Spam-Level: 
X-Spam-Status: No, score=-102.545 tagged_above=-999 required=5 tests=[AWL=0.054, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3GCL92V0vk29; Thu, 17 Mar 2011 08:21:50 -0700 (PDT)
Received: from co300216-co-outbound.net.avaya.com (co300216-co-outbound.net.avaya.com [198.152.13.100]) by core3.amsl.com (Postfix) with ESMTP id B567D3A6920; Thu, 17 Mar 2011 08:21:50 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAFMtcU2HCzI1/2dsb2JhbACmZXSkeQKZFoVhBJAM
X-IronPort-AV: E=Sophos;i="4.63,199,1299474000"; d="scan'208";a="269962120"
Received: from unknown (HELO p-us1-erheast.us1.avaya.com) ([135.11.50.53]) by co300216-co-outbound.net.avaya.com with ESMTP; 17 Mar 2011 11:23:17 -0400
X-IronPort-AV: E=Sophos;i="4.63,199,1299474000"; d="scan'208";a="623980768"
Received: from unknown (HELO DC-US1HCEX3.global.avaya.com) ([135.11.52.22]) by p-us1-erheast-out.us1.avaya.com with ESMTP; 17 Mar 2011 11:23:17 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX3.global.avaya.com ([135.11.52.22]) with mapi; Thu, 17 Mar 2011 11:23:17 -0400
From: "Worley, Dale R (Dale)" 
To: "Parthasarathi R (partr)" , "Cullen Jennings (fluffy)" , "Mike Hammer (hmmr)" , "HKaplan@acmepacket.com" , "keith.drage@alcatel-lucent.com" , "john.elwell@siemens-enterprise.com" ,  "mary.ietf.barnes@gmail.com" , "Paul Kyzivat (pkyzivat)" , "paulej@packetizer.com" , "rjsparks@nostrum.com" ,  "spromano@unina.it" , "Thierry.Bessis@alcatel-lucent.com" , "vkg@bell-labs.com" ,  "vpascual@acmepacket.com" 
Date: Thu, 17 Mar 2011 11:22:29 -0400
Thread-Topic: Bar BOF request for SIP Load balancing & (media) overload handling
Thread-Index: AcvkNQDWIeXRlo1yRRq3Yg+LUWi3ZAAgh/mh
Message-ID: 
References: 
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Mailman-Approved-At: Thu, 17 Mar 2011 09:41:37 -0700
Cc: "dispatch@ietf.org" , "sip-overload@ietf.org" 
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 15:21:52 -0000

________________________________________
From: Parthasarathi R (partr) [partr@cisco.com]

Please let me about others interest and opinion to have Bar BOF for SIP Loa=
d balancing & overload handling during IETF-80.
________________________________________

Clearly, this is a subject that needs study within the IETF, so the bar BOF=
 is a very good idea.
(I'll be at IETF Prague.)

Dale

From gsalguei@cisco.com  Thu Mar 17 08:27:40 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 212593A6968; Thu, 17 Mar 2011 08:27:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.418
X-Spam-Level: 
X-Spam-Status: No, score=-10.418 tagged_above=-999 required=5 tests=[AWL=0.180, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id aB0kVpJo0eJn; Thu, 17 Mar 2011 08:27:38 -0700 (PDT)
Received: from av-tac-rtp.cisco.com (hen.cisco.com [64.102.19.198]) by core3.amsl.com (Postfix) with ESMTP id 272A93A68FC; Thu, 17 Mar 2011 08:27:38 -0700 (PDT)
X-TACSUNS: Virus Scanned
Received: from rooster.cisco.com (localhost.cisco.com [127.0.0.1]) by av-tac-rtp.cisco.com (8.13.8+Sun/8.13.8) with ESMTP id p2HFSwhX015862; Thu, 17 Mar 2011 11:28:59 -0400 (EDT)
Received: from dhcp-64-102-211-8.cisco.com (dhcp-64-102-211-8.cisco.com [64.102.211.8]) by rooster.cisco.com (8.13.8+Sun/8.13.8) with ESMTP id p2HFSvLl006267;  Thu, 17 Mar 2011 11:28:58 -0400 (EDT)
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: multipart/alternative; boundary=Apple-Mail-72--54009375
From: Gonzalo Salgueiro 
In-Reply-To: 
Date: Thu, 17 Mar 2011 11:28:57 -0400
Message-Id: <74D7A7C9-A12B-4CBB-8186-A4314A4CEDF5@cisco.com>
References:  
To: Janet P Gunn 
X-Mailer: Apple Mail (2.1082)
X-Mailman-Approved-At: Thu, 17 Mar 2011 09:41:37 -0700
Cc: dispatch@ietf.org, sip-overload@ietf.org, keith.drage@alcatel-lucent.com, HKaplan@acmepacket.com, vpascual@acmepacket.com, Thierry.Bessis@alcatel-lucent.com, dispatch-bounces@ietf.org
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media)	overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 15:27:40 -0000

--Apple-Mail-72--54009375
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii

This is my situation as well. I would love to join but will not be in =
attendance.

Gonzalo

On Mar 17, 2011, at 10:24 AM, Janet P Gunn wrote:

>=20
> Partha,=20
>=20
> I am definitely interested in the subject, but I will, unfortunately, =
not be in Prague.=20
>=20
> If it were possible to hold the bar bof somewhere with an audio feed =
and jabber room, I would definitely participate remotely.=20
>=20
> Thanks=20
>=20
> Janet
>=20
> This is a PRIVATE message. If you are not the intended recipient, =
please delete without copying and kindly advise us by e-mail of the =
mistake in delivery.=20
> NOTE: Regardless of content, this e-mail shall not operate to bind CSC =
to any order or other contract unless pursuant to explicit written =
agreement or government initiative expressly permitting the use of =
e-mail for such purpose.=20
>=20
>=20
> From:	"Parthasarathi R (partr)" 
> To:	, "Cullen Jennings (fluffy)" =
, "Mike Hammer (hmmr)" , =
, , =
, , =
"Paul Kyzivat (pkyzivat)" , , =
, , =
, , =

> Cc:	dispatch@ietf.org, sip-overload@ietf.org
> Date:	03/16/2011 11:34 PM
> Subject: 	[dispatch] Bar BOF request for SIP Load balancing & =
(media)        overload handling
>=20
>=20
>=20
>=20
> Hi all,=20
>  =20
> There is lot of discussion about SIP load balancing and overload =
handling in Dispatch alias. I wish to have discussion about SIP Load =
balancing & (media) overload handling at IETF-80 Prague. I guess that =
Bar BOF is the idle way to start the discussion about this topic to =
common understanding and identify the way to move forward. Please =
suggest me in case any other mechanism exists in IETF.=20
>  =20
> I have included the folks who has shown the interest in this topic as =
part of SIP Load balancing & overload handling, dispatch WG chairs, CCed =
to Dispatch and SoC WG alias. AFAIK, the couple of drafts are related to =
this problem:=20
>  =20
> draft-bessis-dispatch-adaptive-load-balancing-00
> draft-partha-dispatch-sip-media-overload-control-00.txt
> draft-jones-sip-overload-sce-00
> draft-partha-dispatch-resource-availability-00=20
> draft-sparks-sipping-load-00=20
>  =20
> Please let me about others interest and opinion to have Bar BOF for =
SIP Load balancing & overload handling during IETF-80.=20
>  =20
> Thanks=20
> Partha_______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>=20
>=20
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


--Apple-Mail-72--54009375
Content-Transfer-Encoding: 7bit
Content-Type: text/html;
	charset=us-ascii

This is my situation as well. I would love to join but will not be in attendance.

Gonzalo

On Mar 17, 2011, at 10:24 AM, Janet P Gunn wrote:



Partha,



I am definitely interested in the subject,
but I will, unfortunately, not be in Prague.



If it were possible to hold the bar
bof somewhere with an audio feed and jabber room, I would definitely participate
remotely.



Thanks



Janet



This is a PRIVATE message. If you are not the intended recipient, please
delete without copying and kindly advise us by e-mail of the mistake in
delivery. 

NOTE: Regardless of content, this e-mail shall not operate to bind CSC
to any order or other contract unless pursuant to explicit written agreement
or government initiative expressly permitting the use of e-mail for such
purpose.










From:
"Parthasarathi R (partr)"
<partr@cisco.com>


To:
<dworley@avaya.com>, "Cullen
Jennings (fluffy)" <fluffy@cisco.com>, "Mike Hammer (hmmr)"
<hmmr@cisco.com>, <HKaplan@acmepacket.com>, <keith.drage@alcatel-lucent.com>,
<john.elwell@siemens-enterprise.com>, <mary.ietf.barnes@gmail.com>,
"Paul Kyzivat (pkyzivat)" <pkyzivat@cisco.com>, <paulej@packetizer.com>,
<rjsparks@nostrum.com>, <spromano@unina.it>, <Thierry.Bessis@alcatel-lucent.com>,
<vkg@bell-labs.com>, <vpascual@acmepacket.com>


Cc:
dispatch@ietf.org, sip-overload@ietf.org


Date:
03/16/2011 11:34 PM


Subject:
[dispatch] Bar BOF request for SIP Load
balancing & (media)        overload handling










Hi all,

 

There is lot of discussion about SIP load
balancing and overload handling in Dispatch alias. I wish to have discussion
about SIP Load balancing & (media) overload handling at IETF-80 Prague.
I guess that Bar BOF is the idle way to start the discussion about this
topic to common understanding and identify the way to move forward. Please
suggest me in case any other mechanism exists in IETF.

 

I have included the folks who has shown the
interest in this topic as part of SIP Load balancing & overload handling,
dispatch WG chairs, CCed to Dispatch and SoC WG alias. AFAIK, the couple
of drafts are related to this problem:

 

draft-bessis-dispatch-adaptive-load-balancing-00

draft-partha-dispatch-sip-media-overload-control-00.txt

draft-jones-sip-overload-sce-00

draft-partha-dispatch-resource-availability-00 

draft-sparks-sipping-load-00

 

Please let me about others interest and opinion
to have Bar BOF for SIP Load balancing & overload handling during IETF-80.

 

Thanks

Partha_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.org/mailman/listinfo/dispatch





_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


--Apple-Mail-72--54009375--

From mary.ietf.barnes@gmail.com  Thu Mar 17 10:28:46 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DAB8C3A6AF4 for ; Thu, 17 Mar 2011 10:28:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.878
X-Spam-Level: 
X-Spam-Status: No, score=-102.878 tagged_above=-999 required=5 tests=[AWL=-0.280, BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9ejLMJV4xcSj for ; Thu, 17 Mar 2011 10:28:40 -0700 (PDT)
Received: from mail-vx0-f172.google.com (mail-vx0-f172.google.com [209.85.220.172]) by core3.amsl.com (Postfix) with ESMTP id 1E3983A6AED for ; Thu, 17 Mar 2011 10:28:39 -0700 (PDT)
Received: by vxg33 with SMTP id 33so3314016vxg.31 for ; Thu, 17 Mar 2011 10:30:07 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:in-reply-to:references:date :message-id:subject:from:to:content-type; bh=CfSGxipRgT/FdJDDc7WO4JxC/FtFh4Co84eKYXIqF6Y=; b=Kjr9gOV1KRe0UM9pX3L6rfLnhqJMHlOLgipZdnabRtGvz5FOb+CLgjKmHSj4cECdG1 ChlA2ms2c4hojkqcG2Uq8aTadywzQ/jZ54EiXvd/mpbXmUYzKoXs0ZZ7S51nUfX9gUbP nsc7YeytUcjljAW9mbUrThTyDzSSEqsSKYoL4=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; b=hTxOUwSCzX+hmv4sxSBvtztBAvmsuN9thEsAh7o7Z0JbjqM0rttncQBHZM2BrvMeLr dWXiKTO/8Pe9VJBA/LBgCkkxjbCYso/+gNvE8pEYTRRi66ShbVbniiUPR46WrEnfoX02 lS4Siw7iZXlAL0Pngp/ciuR8mUwS88FH2F+lI=
MIME-Version: 1.0
Received: by 10.52.0.169 with SMTP id 9mr2132513vdf.303.1300383006091; Thu, 17 Mar 2011 10:30:06 -0700 (PDT)
Received: by 10.52.167.137 with HTTP; Thu, 17 Mar 2011 10:30:05 -0700 (PDT)
In-Reply-To: 
References:   
Date: Thu, 17 Mar 2011 12:30:05 -0500
Message-ID: 
From: Mary Barnes 
To: DISPATCH 
Content-Type: multipart/alternative; boundary=20cf304346d62675c1049eb1030f
Subject: [dispatch] Fwd: Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 17:28:47 -0000

--20cf304346d62675c1049eb1030f
Content-Type: text/plain; charset=ISO-8859-1

Per my email to sip-overload mailing list, please trim the receiver list AND
please keep the discussion on dispatch mailing list only.

Thanks,
Mary.

---------- Forwarded message ----------
From: Mary Barnes 
Date: Thu, Mar 17, 2011 at 12:29 PM
Subject: Re: Bar BOF request for SIP Load balancing & (media) overload
handling
To: sip-overload@ietf.org
Cc: "Parthasarathi R (partr)" 


Can I ask folks to PLEASE just keep this discussion on the DISPATCH mailing
list only and trim the receiver list?  My mailbox is filling up with
moderator actions due to the many recipients and cross posting to multi-WGs
fills up everyone's mailboxes.

Note, I'm sending this email separately to each WG.

Thanks,
Mary.


On Thu, Mar 17, 2011 at 10:22 AM, Worley, Dale R (Dale)
wrote:

> ________________________________________
> From: Parthasarathi R (partr) [partr@cisco.com]
>
> Please let me about others interest and opinion to have Bar BOF for SIP
> Load balancing & overload handling during IETF-80.
> ________________________________________
>

--20cf304346d62675c1049eb1030f
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Per my email to sip-overload mailing list, please trim the receiver list AN=
D please keep the discussion on dispatch mailing list only.

<=
div>Thanks,
Mary.

---------- Fo=
rwarded message ----------

From: Mary Barnes <<=
a href=3D"mailto:mary.ietf.barnes@gmail.com">mary.ietf.barnes@gmail.com=
>
Date: Thu, Mar 17, 2011 at 12:29 PM
Subject: Re: Bar BOF =
request for SIP Load balancing & (media) overload handling

To: sip-overload@ietf.org
C=
c: "Parthasarathi R (partr)" <partr@cisco.com>


Can I ask folks to PLEASE just keep t=
his discussion on the DISPATCH mailing list only and trim the receiver list=
? =A0My mailbox is filling up with moderator actions due to the many recipi=
ents and cross posting to multi-WGs fills up everyone's mailboxes.



Note, I'm sending this email separately to each WG.
<=
div>
Thanks,
Mary.=A0


On Thu, Mar 17, 2011 at 10:22 AM, Worley, Dale R (Da=
le) <dworley@avaya.com> wrote:



________________________________________

From: Parthasarathi R (partr) [partr@cisco.com]




Please let me about others interest and opinion to have Bar BOF for SIP Loa=
d balancing & overload handling during IETF-80.


________________________________________
=






--20cf304346d62675c1049eb1030f--

From volker.hilt@alcatel-lucent.com  Thu Mar 17 10:54:03 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 69C6E3A6ADF for ; Thu, 17 Mar 2011 10:54:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2s75djyVVNX3 for ; Thu, 17 Mar 2011 10:54:02 -0700 (PDT)
Received: from ihemail3.lucent.com (ihemail3.lucent.com [135.245.0.37]) by core3.amsl.com (Postfix) with ESMTP id 47A3F3A6A1C for ; Thu, 17 Mar 2011 10:54:02 -0700 (PDT)
Received: from usnavsmail2.ndc.alcatel-lucent.com (usnavsmail2.ndc.alcatel-lucent.com [135.3.39.10]) by ihemail3.lucent.com (8.13.8/IER-o) with ESMTP id p2HHtT6H023447 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 17 Mar 2011 12:55:30 -0500 (CDT)
Received: from USNAVSXCHHUB02.ndc.alcatel-lucent.com (usnavsxchhub02.ndc.alcatel-lucent.com [135.3.39.111]) by usnavsmail2.ndc.alcatel-lucent.com (8.14.3/8.14.3/GMO) with ESMTP id p2HHtT4s012392 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT) for ; Thu, 17 Mar 2011 12:55:29 -0500
Received: from [135.112.131.41] (135.3.63.242) by USNAVSXCHHUB02.ndc.alcatel-lucent.com (135.3.39.111) with Microsoft SMTP Server (TLS) id 8.3.106.1; Thu, 17 Mar 2011 12:55:28 -0500
Message-ID: <4D824B23.3090806@bell-labs.com>
Date: Thu, 17 Mar 2011 13:55:47 -0400
From: Volker Hilt 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: DISPATCH 
References: 	<14C85D6CCBE92743AF33663BF5D24EBA096F3937@gaalpa1msgusr7e.ugd.att.com> <4D8244AD.40200@ericsson.com>
In-Reply-To: <4D8244AD.40200@ericsson.com>
Content-Type: text/plain; charset="ISO-8859-1"; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.37
X-Scanned-By: MIMEDefang 2.64 on 135.3.39.10
Subject: Re: [dispatch] [sip-overload] Bar BOF request for SIP Load balancing &	(media)overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 17:54:03 -0000

Same here!

Volker

On 3/17/2011 1:28 PM, Salvatore Loreto wrote:
> +1
>
> On 3/17/11 5:46 PM, DOLLY, MARTIN C (ATTSI) wrote:
>>
>> I would be interested. Thank you.
>>
>> *From:*sip-overload-bounces@ietf.org
>> [mailto:sip-overload-bounces@ietf.org] *On Behalf Of *Parthasarathi R
>> (partr)
>> *Sent:* Wednesday, March 16, 2011 8:45 PM
>> *To:* dworley@avaya.com; Cullen Jennings (fluffy); Mike Hammer (hmmr);
>> HKaplan@acmepacket.com; keith.drage@alcatel-lucent.com;
>> john.elwell@siemens-enterprise.com; mary.ietf.barnes@gmail.com; Paul
>> Kyzivat (pkyzivat); paulej@packetizer.com; rjsparks@nostrum.com;
>> spromano@unina.it; Thierry.Bessis@alcatel-lucent.com;
>> vkg@bell-labs.com; vpascual@acmepacket.com
>> *Cc:* dispatch@ietf.org; sip-overload@ietf.org
>> *Subject:* [sip-overload] Bar BOF request for SIP Load balancing &
>> (media)overload handling
>>
>> Hi all,
>>
>> There is lot of discussion about SIP load balancing and overload
>> handling in Dispatch alias. I wish to have discussion about SIP Load
>> balancing & (media) overload handling at IETF-80 Prague. I guess that
>> Bar BOF is the idle way to start the discussion about this topic to
>> common understanding and identify the way to move forward. Please
>> suggest me in case any other mechanism exists in IETF.
>>
>> I have included the folks who has shown the interest in this topic as
>> part of SIP Load balancing & overload handling, dispatch WG chairs,
>> CCed to Dispatch and SoC WG alias. AFAIK, the couple of drafts are
>> related to this problem:
>>
>> draft-bessis-dispatch-adaptive-load-balancing-00
>> draft-partha-dispatch-sip-media-overload-control-00.txt
>> draft-jones-sip-overload-sce-00
>> draft-partha-dispatch-resource-availability-00
>> draft-sparks-sipping-load-00
>>
>> Please let me about others interest and opinion to have Bar BOF for
>> SIP Load balancing & overload handling during IETF-80.
>>
>> Thanks
>>
>> Partha
>>
>
>
> --
> Salvatore Loreto
> www.sloreto.com
>


From salvatore.loreto@ericsson.com  Thu Mar 17 10:26:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CBEDE3A6ADC; Thu, 17 Mar 2011 10:26:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.594
X-Spam-Level: 
X-Spam-Status: No, score=-106.594 tagged_above=-999 required=5 tests=[AWL=0.004, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GoZ4Bw10SIoP; Thu, 17 Mar 2011 10:26:49 -0700 (PDT)
Received: from mailgw9.se.ericsson.net (mailgw9.se.ericsson.net [193.180.251.57]) by core3.amsl.com (Postfix) with ESMTP id 303193A69F3; Thu, 17 Mar 2011 10:26:49 -0700 (PDT)
X-AuditID: c1b4fb39-b7c6dae0000023f2-85-4d8244afdfc3
Received: from esessmw0256.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw9.se.ericsson.net (Symantec Mail Security) with SMTP id 12.4C.09202.FA4428D4; Thu, 17 Mar 2011 18:28:16 +0100 (CET)
Received: from mail.lmf.ericsson.se (153.88.115.8) by esessmw0256.eemea.ericsson.se (153.88.115.97) with Microsoft SMTP Server id 8.2.234.1; Thu, 17 Mar 2011 18:28:15 +0100
Received: from nomadiclab.lmf.ericsson.se (nomadiclab.lmf.ericsson.se [131.160.33.3])	by mail.lmf.ericsson.se (Postfix) with ESMTP id 536D3249B; Thu, 17 Mar 2011 19:28:15 +0200 (EET)
Received: from nomadiclab.lmf.ericsson.se (localhost [127.0.0.1])	by nomadiclab.lmf.ericsson.se (Postfix) with ESMTP id 10E4750A75; Thu, 17 Mar 2011 19:28:15 +0200 (EET)
Received: from Salvatore-Loretos-MacBook-Pro.local (localhost [127.0.0.1])	by nomadiclab.lmf.ericsson.se (Postfix) with ESMTP id 374C350A72; Thu, 17 Mar 2011 19:28:14 +0200 (EET)
Message-ID: <4D8244AD.40200@ericsson.com>
Date: Thu, 17 Mar 2011 19:28:13 +0200
From: Salvatore Loreto 
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: "DOLLY, MARTIN C (ATTSI)" 
References:  <14C85D6CCBE92743AF33663BF5D24EBA096F3937@gaalpa1msgusr7e.ugd.att.com>
In-Reply-To: <14C85D6CCBE92743AF33663BF5D24EBA096F3937@gaalpa1msgusr7e.ugd.att.com>
Content-Type: multipart/alternative; boundary="------------000101040901050704020900"
X-Virus-Scanned: ClamAV using ClamSMTP
X-Brightmail-Tracker: AAAAAA==
X-Mailman-Approved-At: Thu, 17 Mar 2011 13:00:55 -0700
Cc: "vpascual@acmepacket.com" , "sip-overload@ietf.org" , "keith.drage@alcatel-lucent.com" , "HKaplan@acmepacket.com" , "dispatch@ietf.org" , "Thierry.Bessis@alcatel-lucent.com" 
Subject: Re: [dispatch] [sip-overload] Bar BOF request for SIP Load balancing &	(media)overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 17:26:51 -0000

--------------000101040901050704020900
Content-Type: text/plain; charset="ISO-8859-1"; format=flowed
Content-Transfer-Encoding: 7bit

+1

On 3/17/11 5:46 PM, DOLLY, MARTIN C (ATTSI) wrote:
>
> I would be interested. Thank you.
>
> *From:*sip-overload-bounces@ietf.org 
> [mailto:sip-overload-bounces@ietf.org] *On Behalf Of *Parthasarathi R 
> (partr)
> *Sent:* Wednesday, March 16, 2011 8:45 PM
> *To:* dworley@avaya.com; Cullen Jennings (fluffy); Mike Hammer (hmmr); 
> HKaplan@acmepacket.com; keith.drage@alcatel-lucent.com; 
> john.elwell@siemens-enterprise.com; mary.ietf.barnes@gmail.com; Paul 
> Kyzivat (pkyzivat); paulej@packetizer.com; rjsparks@nostrum.com; 
> spromano@unina.it; Thierry.Bessis@alcatel-lucent.com; 
> vkg@bell-labs.com; vpascual@acmepacket.com
> *Cc:* dispatch@ietf.org; sip-overload@ietf.org
> *Subject:* [sip-overload] Bar BOF request for SIP Load balancing & 
> (media)overload handling
>
> Hi all,
>
> There is lot of discussion about SIP load balancing and overload 
> handling in Dispatch alias. I wish to have discussion about SIP Load 
> balancing & (media) overload handling at IETF-80 Prague. I guess that 
> Bar BOF is the idle way to start the discussion about this topic to 
> common understanding and identify the way to move forward. Please 
> suggest me in case any other mechanism exists in IETF.
>
> I have included the folks who has shown the interest in this topic as 
> part of SIP Load balancing & overload handling, dispatch WG chairs, 
> CCed to Dispatch and SoC WG alias. AFAIK, the couple of drafts are 
> related to this problem:
>
> draft-bessis-dispatch-adaptive-load-balancing-00
> draft-partha-dispatch-sip-media-overload-control-00.txt
> draft-jones-sip-overload-sce-00
> draft-partha-dispatch-resource-availability-00
> draft-sparks-sipping-load-00
>
> Please let me about others interest and opinion to have Bar BOF for 
> SIP Load balancing & overload handling during IETF-80.
>
> Thanks
>
> Partha
>


-- 
Salvatore Loreto
www.sloreto.com


--------------000101040901050704020900
Content-Type: text/html; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit



  
    
  
  
    +1

    

    On 3/17/11 5:46 PM, DOLLY, MARTIN C (ATTSI) wrote:
    

      
      
      
      

        
I would be interested. Thank you.

        
 

        

          

            
From:
                sip-overload-bounces@ietf.org
                [mailto:sip-overload-bounces@ietf.org] On Behalf Of
                Parthasarathi R (partr)

                Sent: Wednesday, March 16, 2011 8:45 PM

                To: dworley@avaya.com; Cullen Jennings (fluffy);
                Mike Hammer (hmmr); HKaplan@acmepacket.com;
                keith.drage@alcatel-lucent.com;
                john.elwell@siemens-enterprise.com;
                mary.ietf.barnes@gmail.com; Paul Kyzivat (pkyzivat);
                paulej@packetizer.com; rjsparks@nostrum.com;
                spromano@unina.it; Thierry.Bessis@alcatel-lucent.com;
                vkg@bell-labs.com; vpascual@acmepacket.com

                Cc: dispatch@ietf.org; sip-overload@ietf.org

                Subject: [sip-overload] Bar BOF request for SIP
                Load balancing & (media)overload handling

          

        

        
 

        

          

            
Hi all,

          

          

            
 

          

          

            
There is lot of discussion about SIP load
                balancing and overload handling in Dispatch alias. I
                wish to have discussion about SIP Load balancing &
                (media) overload handling at IETF-80 Prague. I guess
                that Bar BOF is the idle way to start the discussion
                about this topic to common understanding and identify
                the way to move forward. Please suggest me in case any
                other mechanism exists in IETF.

          

          

            
 

          

          

            
I have included the folks who has shown
                the interest in this topic as part of SIP Load balancing
                & overload handling, dispatch WG chairs, CCed to
                Dispatch and SoC WG alias. AFAIK, the couple of drafts
                are related to this problem:

          

          

            
 

          

          

            
draft-bessis-dispatch-adaptive-load-balancing-00

                draft-partha-dispatch-sip-media-overload-control-00.txt

                draft-jones-sip-overload-sce-00

                draft-partha-dispatch-resource-availability-00 

                draft-sparks-sipping-load-00

          

          

            
 

          

          

            
Please let me about others interest and
                opinion to have Bar BOF for SIP Load balancing &
                overload handling during IETF-80.

          

          

            
 

          

          

            
Thanks

          

          

            
Partha

          

        

      

    

    

    

    
-- 
Salvatore Loreto
www.sloreto.com

  


--------------000101040901050704020900--

From vpascual@acmepacket.com  Thu Mar 17 11:15:11 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0BBB23A6A1C; Thu, 17 Mar 2011 11:15:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eIluk9SKdG0X; Thu, 17 Mar 2011 11:15:10 -0700 (PDT)
Received: from etmail.acmepacket.com (etmail.acmepacket.com [216.41.24.6]) by core3.amsl.com (Postfix) with ESMTP id 2DC903A6A04; Thu, 17 Mar 2011 11:15:10 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by etmail.acmepacket.com (216.41.24.6) with Microsoft SMTP Server (TLS) id 8.2.254.0; Thu, 17 Mar 2011 14:16:37 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Thu, 17 Mar 2011 14:16:37 -0400
From: Victor Pascual 
To: "Worley, Dale R (Dale)" , "Parthasarathi R (partr)" , "Cullen Jennings (fluffy)" , "Mike Hammer (hmmr)" , Hadriel Kaplan , "keith.drage@alcatel-lucent.com" , "john.elwell@siemens-enterprise.com" ,  "mary.ietf.barnes@gmail.com" , "Paul Kyzivat (pkyzivat)" , "paulej@packetizer.com" , "rjsparks@nostrum.com" ,  "spromano@unina.it" , "Thierry.Bessis@alcatel-lucent.com" , "vkg@bell-labs.com" 
Date: Thu, 17 Mar 2011 14:16:37 -0400
Thread-Topic: Bar BOF request for SIP Load balancing & (media) overload handling
Thread-Index: AcvkNQDWIeXRlo1yRRq3Yg+LUWi3ZAAgh/mhAAZHhQA=
Message-ID: 
References:  
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAgAAAUAAAAFU
X-Mailman-Approved-At: Thu, 17 Mar 2011 13:00:55 -0700
Cc: "dispatch@ietf.org" , "sip-overload@ietf.org" 
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 18:15:11 -0000

Count me in

-Victor

>-----Original Message-----
>From: Worley, Dale R (Dale) [mailto:dworley@avaya.com]
>Sent: Thursday, March 17, 2011 4:22 PM
>To: Parthasarathi R (partr); Cullen Jennings (fluffy); Mike Hammer
>(hmmr); Hadriel Kaplan; keith.drage@alcatel-lucent.com;
>john.elwell@siemens-enterprise.com; mary.ietf.barnes@gmail.com; Paul
>Kyzivat (pkyzivat); paulej@packetizer.com; rjsparks@nostrum.com;
>spromano@unina.it; Thierry.Bessis@alcatel-lucent.com; vkg@bell-labs.com;
>Victor Pascual
>Cc: dispatch@ietf.org; sip-overload@ietf.org
>Subject: RE: Bar BOF request for SIP Load balancing & (media) overload
>handling
>
>________________________________________
>From: Parthasarathi R (partr) [partr@cisco.com]
>
>Please let me about others interest and opinion to have Bar BOF for SIP
>Load balancing & overload handling during IETF-80.
>________________________________________
>
>Clearly, this is a subject that needs study within the IETF, so the bar
>BOF is a very good idea.
>(I'll be at IETF Prague.)
>
>Dale

From Thierry.Bessis@alcatel-lucent.com  Thu Mar 17 13:18:00 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 26E613A6AF0 for ; Thu, 17 Mar 2011 13:18:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.947
X-Spam-Level: 
X-Spam-Status: No, score=-5.947 tagged_above=-999 required=5 tests=[AWL=0.050,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id onsDjJOvA3vc for ; Thu, 17 Mar 2011 13:17:53 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 166D03A6A2E for ; Thu, 17 Mar 2011 13:17:53 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2HKJKhn020332 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 17 Mar 2011 15:19:20 -0500 (CDT)
Received: from [135.244.40.104] (tbessis.lra.lucent.com [135.244.40.104]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2HKJI6o013820; Thu, 17 Mar 2011 15:19:18 -0500 (CDT)
Message-ID: <4D826CC5.9010502@alcatel-lucent.com>
Date: Thu, 17 Mar 2011 15:19:17 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Parthasarathi R (partr)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------030009070707060808000204"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 20:18:00 -0000

This is a multi-part message in MIME format.
--------------030009070707060808000204
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi Partha,

Thank you for your comments. See my answers..
>>
>> Now we are not saying that everybody should calculate their weight 
>> like that. What is important is that the weight reflects your current 
>> _relative_ capacity. In our case, this "algorithm" balances the load 
>> of our servers very well. Now if you use a server that has a fixed 
>> amount of resources and that is dimensioned to process let's say 50 
>> 000 sessions, at a pace of 100 session/s, then you could probably 
>> derive a weight that would depend on both your count of sessions 
>> left, and your current rate of session/s.
>>  In case every device calculate its own algorithm for weight, 
>> then weight from one device may not same as other device. Say dev1 
>> takes Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 
>> 40, 30) = 30% and another device dev2 calculates Min (CPU, Mem) then 
>> weight will be Min (70, 40) = 40% even though they are equally 
>> loaded. It is very important to have the common unit to solve this 
>> issue 
> TB >> I think here you imply that the DSP is used as a finite pool of 
> bearer channels resources, and your are probably using the DSP to 
> process bearers. May be you are an SBC or so.
>  Here is the interesting part, dev1 achieve the specific 
> service using DSP whereas dev2 achieve the same result using multi 
> core/processor architecture or softDSP or ASIC implementation. IMO, 
> weight varies based on the system architecture in case algorithm is 
> not spelt out correctly or generic units are not defined. 
TB> It seems to me that this comment just reinforces my point (I hope I 
understand well): We need to let the destination come up its own way 
with its percentage of resource left, because there are so many possible 
implementations that it would be cumbersome and restrictive to impose 
any algorithm. Also just to be sure: The proposed weight is a percentage 
(a ratio): it has no unit. Part of the robustness and universality is 
derived from this characteristic.
>  Then, it is obvious that you should also consider this resource as 
> well, since you need it to process new sessions. So you can't say that 
> you have 40% resource available to process calls if actually one of 
> your resource needed has only 30% available left !
> Our recommendation is that the occupation (weight) reflects the MIN of 
> ANY resource that must be used for call processing and that is 
> constrained.
>
> I think it is on the contrary an advantage to not specify exactly how 
> it is calculated, but to specify exactly what it is, or what it means 
> instead (The most constrained resource ratio).
>>> If not, I could not understand your argument. As you mentioned, all 
>>> system will know CPU or memory usage and it is possible to come up 
>>> the static or refreshable value saying that load shed has to happen 
>>> when 85% of CPU utilization or 90% of memory utilization. The 
>>> different between our proposal is that there is no need of two value 
>>> per message instead 1 value in each message & 1 value is static or 
>>> in periodic manner 
>> Thierry>> We use a single value: see the example above.
>>  I'm proposing is to use single real time value with units 
>> like number of dialog and exchange the threshold (max dialog) using 
>> some other static interface.
>  I agree that threshold varies from device to device varies 
> within the same company and huge differencewill be seen in case of 
> multi vendor scenario.  It may be tough to manage in case it is 
> provisioned statically as the threshold has to be tweaked based on the 
> solutions. It is the main reason, I proposed to pass the threshold 
> value from device while booting and refresh it whenever required.
> Hope we both agrees to have that two variables x,y is required to 
> solve this issue.
Not at all ! We don't need any threshold to balance the load. We just 
need a weight: What you want is to balance the weights so that they tend 
to be identical everywhere: why do we need a threshold to do that ?
> Load balancer (upstream entity) will use f(x,y) based on its own 
> definition to share the load and shed the load towards single device.
TB> Again: We are not interested to redefine overload control and load 
shedding in this proposal: The new OLC draft has just be defined to do 
that.  I've explained (and we should probably make it clear in the 
proposal) that both mechanisms work very well together while each has 
its own independent algorithm.
> The open issue is whether both x & y has to be passed every time or x 
> shall be passed in real time and y shall be passed once in a while.
> BTW, I have written Media overload control requirement 
> (draft-partha-dispatch-sip-media-overload-control-00) draft to 
> identify the common problem which will be agreed upon lot of 
> vendors and solve it based on the requirement.
TB> This requirement was an entry for the OLC draft, is that your point ?

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------030009070707060808000204
Content-Type: multipart/related;
 boundary="------------060805000806090004060500"


--------------060805000806090004060500
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
    
  
  
    Hi Partha, 

    

    Thank you for your comments. See my answers.. 

    

      

        


          Now we are not saying that everybody should calculate their
          weight like that. What is important is that the weight
          reflects your current relative capacity. In our case,
          this "algorithm" balances the load of our servers very well.
          Now if you use a server that has a fixed amount of resources
          and that is dimensioned to process let's say 50 000 sessions,
          at a pace of 100 session/s, then you could probably derive a
          weight that would depend on both your count of sessions left,
          and your current rate of session/s.  

        
 

        
<Partha> In case every
              device calculate its own algorithm for weight, then weight
              from one device may not same as other device. Say dev1
              takes Min(CPU, Mem, DSP) then weight will be calculated
              as Min (70, 40, 30) = 30% and another device
              dev2 calculates Min (CPU, Mem) then weight will be Min
              (70, 40) = 40% even though they are equally loaded. It is
              very important to have the common unit to solve this issue
              </Partha>  

            

      

      
TB >> I think here you imply that the DSP is used as a
        finite pool of bearer channels resources, and your are probably
        using the DSP to process bearers. May be you are an SBC or so.

      
 

      
<partha> Here is the interesting part, dev1 achieve the
        specific service using DSP whereas dev2 achieve the same result
        using multi core/processor architecture or softDSP or ASIC
        implementation. IMO, weight varies based on the system
        architecture in case algorithm is not spelt out correctly or
        generic units are not defined. </Partha>

    

    TB> It seems to me that this comment just reinforces my point (I
    hope I understand well): We need to let the destination come up its
    own way with its percentage of resource left, because there are so
    many possible implementations that it would be cumbersome and
    restrictive to impose any algorithm. Also just to be sure: The
    proposed weight is a percentage (a ratio): it has no unit. Part of
    the robustness and universality is derived from this characteristic.
    

    

      
 

      
 Then, it is obvious that you should also consider this
        resource as well, since you need it to process new sessions. So
        you can't say that you have 40% resource available to process
        calls if actually one of your resource needed has only 30%
        available left ! 

        Our recommendation is that the occupation (weight) reflects the
        MIN of ANY resource that must be used for call processing and
        that is constrained. 

        

        I think it is on the contrary an advantage to not specify
        exactly how it is calculated, but to specify exactly what it is,
        or what it means instead (The most constrained resource ratio).
        

      

      

        

          
If not, I could not understand your argument. As you
            mentioned, all system will know CPU or memory usage and it
            is possible to come up the static or refreshable value
            saying that load shed has to happen when 85% of CPU
            utilization or 90% of memory utilization. The different
            between our proposal is that there is no need of two value
            per message instead 1 value in each message & 1 value is
            static or in periodic manner  </Partha>

        

        
Thierry>> We use a single value: see the example
          above.  

        
<Partha> I'm proposing is to use
                single real time value with units like number of dialog
                and exchange the threshold (max dialog) using some other
                static interface. 

            

      

      

        

          
<Partha> I agree that
              threshold varies from device to device varies within the
              same company and huge difference will be seen in case
              of multi vendor scenario.  It may be tough to manage in
              case it is provisioned statically as the threshold has to
              be tweaked based on the solutions. It is the main reason,
              I proposed to pass the threshold value from device while
              booting and refresh it whenever required.  

          
 

          
Hope we both agrees to
              have that two variables x,y is required to solve this issue. 

        

      

    

    Not at all ! We don't need any threshold to balance the load. We
    just need a weight: What you want is to balance the weights so that
    they tend to be identical everywhere: why do we need a threshold to
    do that ? 

    

      

        

          
Load balancer (upstream
              entity) will use f(x,y) based on its own definition to
              share the load and shed the load towards single device. 

        

      

    

    TB> Again: We are not interested to redefine overload control and
    load shedding in this proposal: The new OLC draft has just be
    defined to do that.  I've explained (and we should probably make it
    clear in the proposal) that both mechanisms work very well together
    while each has its own independent algorithm. 

    

      

        

          
The open issue is whether
              both x & y has to be passed every time or x shall be
              passed in real time and y shall be passed once in a while.
            

          
 

          
BTW, I have written Media
              overload control requirement (draft-partha-dispatch-sip-media-overload-control-00)
                draft to identify the common problem which will be
                agreed upon lot of vendors and solve it based on the
                requirement. 

              

        

      

    

    TB> This requirement was an entry for the OLC draft, is that your
    point ? 

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------060805000806090004060500
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"

/9j/4AAQSkZJRgABAgEAYABgAAD/4QreRXhpZgAATU0AKgAAAAgABwESAAMAAAABAAEAAAEa
AAUAAAABAAAAYgEbAAUAAAABAAAAagEoAAMAAAABAAIAAAExAAIAAAAdAAAAcgEyAAIAAAAU
AAAAj4dpAAQAAAABAAAApAAAANAAAABgAAAAAQAAAGAAAAABQWRvYmUgUGhvdG9zaG9wIENT
IE1hY2ludG9zaAAyMDA4OjAyOjEyIDExOjE5OjU0AAAAA6ABAAMAAAAB//8AAKACAAQAAAAB
AAABK6ADAAQAAAABAAAAaAAAAAAAAAAGAQMAAwAAAAEABgAAARoABQAAAAEAAAEeARsABQAA
AAEAAAEmASgAAwAAAAEAAgAAAgEABAAAAAEAAAEuAgIABAAAAAEAAAmoAAAAAAAAAEgAAAAB
AAAASAAAAAH/2P/gABBKRklGAAECAQBIAEgAAP/tAAxBZG9iZV9DTQAB/+4ADkFkb2JlAGSA
AAAAAf/bAIQADAgICAkIDAkJDBELCgsRFQ8MDA8VGBMTFRMTGBEMDAwMDAwRDAwMDAwMDAwM
DAwMDAwMDAwMDAwMDAwMDAwMDAENCwsNDg0QDg4QFA4ODhQUDg4ODhQRDAwMDAwREQwMDAwM
DBEMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwM/8AAEQgAOACgAwEiAAIRAQMRAf/dAAQA
Cv/EAT8AAAEFAQEBAQEBAAAAAAAAAAMAAQIEBQYHCAkKCwEAAQUBAQEBAQEAAAAAAAAAAQAC
AwQFBgcICQoLEAABBAEDAgQCBQcGCAUDDDMBAAIRAwQhEjEFQVFhEyJxgTIGFJGhsUIjJBVS
wWIzNHKC0UMHJZJT8OHxY3M1FqKygyZEk1RkRcKjdDYX0lXiZfKzhMPTdePzRieUpIW0lcTU
5PSltcXV5fVWZnaGlqa2xtbm9jdHV2d3h5ent8fX5/cRAAICAQIEBAMEBQYHBwYFNQEAAhED
ITESBEFRYXEiEwUygZEUobFCI8FS0fAzJGLhcoKSQ1MVY3M08SUGFqKygwcmNcLSRJNUoxdk
RVU2dGXi8rOEw9N14/NGlKSFtJXE1OT0pbXF1eX1VmZ2hpamtsbW5vYnN0dXZ3eHl6e3x//a
AAwDAQACEQMRAD8A9VSSSSUpJJJJSkkkklKSSSSUpJJJJSkkkklKSSSSUpJJCycrGxKHZGVa
yiivV9tjgxo/rPfDUlJUlj1fW76vXE+jleqBy9lVrm/9uNqNf/SWljZeLl1C7FtZfUTG+twc
JHLZb+ciYyG4I8wpMkkkgp//0PVUkkLKyacTGtyr3bKaGOtteeAxgL3u/stakppYv1h6Vl9Z
yuiU27s/CY2y6uDG12z6D/ov2erV6n9dWOn9Sxeo1WW4xcW03WY79wLffS4028/m72rzfoPV
MijL6F1W7p+ZiOzMvJOflOqa3Gtb1V3qY+y/e+zZVczB9P8AR1I1xrczCwup2GjoGV1vqjOq
P3mppe197sHGvtYW/oLrt/qbvZ7ElPpiq5PUsPGy8TCuftyM9z2YzIJ3Gtjr7f6rWVsXIdVw
vq4bum9A6ZhYeS0nJvx/tV7xh1uDmDMYymt1rMzL/S+zE/7S+/8AmGLHqwsDLp+rfTOsGuzp
1Wd1XFcPUeKgyoZDMaqvJss9f0mbK2Y2+/1PT9NJT6iqmN1PEys7MwaS439PdW3IBaQAbWC+
va/6L/0bl54/EyOqZnXbHnDr6ri5b24eflZtmPfiY9Ya/plmNi14t1bML0rPW9Rt3696tvrL
osH9onrf1uHTjUM/fieibZ9Pf9kq+lt9yQFmlPQ39SxaOoYvTrC77RmNtfSACQRTs9Xc783+
eYra5TqeW3C+sn1cu6vfRj2jEzRkWF4rq9Utw94qdcW+3fu9NLF6x0ln1wzsp2djjGyOm476
bzaz03tqtzG3vrt3em70Xfznu9iSnq0l5V9XaMfIw/q1036yONfR7sXKvoque6uu/LOU5+PX
kat9X9TsbfjV2fzm/wDsLS6jh9NyOu0dGbRh5vSG9PaOk4+VmW00E77qst+JbVTnfaMupnp1
/wA4y/Gq99KSn0NVsvqGLhvx673EPzLRj47QC4usLXWR7R7Wtrqsse9/sXEYnQ7M/qXQunde
yh1A19NynXPoue6rIY2/H+y132tFD8qttZpfbvb+mvo/SeoqDcbp9vTsTpXUCHdKxPrLk4Nd
dtjg1lDK8puPj+sX+p7d36P9Ikp9LvvpxqLMi9wrppa6yx54a1o3vcf6rQvHOp9W6n9auqst
c19gtfs6bgDRrAf5v2/Q+0OZ+kycl/8AN/yMepei/Wqiuj6i59HTobj04JZSGEuHoMYBDX+7
cz7O1cx9UB0/o3R7frTnRZba52LgUsIL54e1v7l1763fS/msWn1P8JYrGCgJSq5XwxHiUhu9
GweldMyMrB2/bczAxnZHU81riGVWR+gwcNk/T+m97/8Atz+c9PH0MZ9ttTc7HpOF1ypgdZSf
azKa3+cot2/orf3a7m+/Ht/7bWI/Nud9TczLsYyn9o5TcSmqoQxtLHera2T+kusts+3PvyLv
0t9lqycDPODhWX02Bp+1VtrqaRILa7X5N3p/ms2Px6N/+Fs/4hTe2ZCRJ1B4f++/wU0X1fDy
qszFpy6STVextjJ0MOG6HD81zfzkZY31Qu9f6v49wENsfc9n9V11rmrZVKQqRHYkLX//0fVV
C6mnIpfRfW22m1pZZW8BzXNOjmPY72ua5TSSUhtxMS6gY11NdlDdsUvaHMGwh9X6Nw2fo3Ma
6v8AcUXdPwH49uK/Gqdj3uc+6ksaWPc8+pa+2vbssdY/3v3qwkkpoP6B0J+IzCf07FdiVOL6
8c0VmtrjO57Kdnpse7cpO6J0Z2E3p7sDGOEx29uKaazUHane2jZ6W73O/NV1JJTRyeh9FzHV
vy+n42Q6hoZS62mt5Y0fRZXvY7Yxv8lWa8bHqtturqYy28g3WNaA55aPTYbXj3WbGDYzeipJ
Kamd0npXUtn7RwsfM9KfT+0VMt27o37PVa/bu2NQ7OgdBtxqsS3puI/Gxy40UOorNbC87rPS
qLNle9309ivpJKa+VgYObj/ZszHqycfQ+jaxr2e36P6OwOZ7UK7o3R8jFqw78HGtxaI9HHfU
x1bIG0elU5vp1+0/mK6kkpBVgYNBqdTj1VOorNNJYxrSys7Saato/R1exn6NihZ0rpduPbi2
YdD8fIebb6XVMLLLCfUddbWW7LLXPG/1Hq0kkpCcPEOGcH0WDENfo+gAAz09vp+i1jfa2v0/
ZsXjHWOi5f1X6g7DymuOI536lmH6FrPzWucPa3KrYPTur+n+ez9EvbULJxsbLofj5VTL6LBD
6rWh7HDn3sfua5SYspxmx9QkGnzTF619WLfqvidN6r9pfdhXvvGPjgt9Uudc5gOR/NtqdXk+
/wDTUXf9/wAyinqP1t6uMfApZjV1MZSPSbNOHjNLixjn+z1bvda9rXfpMm//AEWP/Negn/F7
9Tjb6v7MYDM7WvsDP+2G2ej/AOBrbw8LDwaG42FRXjUN+jVU0MaJ59rAFKeYiLMQeI2fVtHi
34UmSsHDowMOjCxm7aMattVbefa0bWyf3kdJJVlr/9L1VJfKqSSn6qSXyqkkp+qkl8qpJKfq
pJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn/9n/7S24
UGhvdG9zaG9wIDMuMAA4QklNBCUAAAAAABAAAAAAAAAAAAAAAAAAAAAAOEJJTQPqAAAAAB2t
PD94bWwgdmVyc2lvbj0iMS4wIiBlbmNvZGluZz0iVVRGLTgiPz4KPCFET0NUWVBFIHBsaXN0
IFBVQkxJQyAiLS8vQXBwbGUgQ29tcHV0ZXIvL0RURCBQTElTVCAxLjAvL0VOIiAiaHR0cDov
L3d3dy5hcHBsZS5jb20vRFREcy9Qcm9wZXJ0eUxpc3QtMS4wLmR0ZCI+CjxwbGlzdCB2ZXJz
aW9uPSIxLjAiPgo8ZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1I
b3Jpem9udGFsUmVzPC9rZXk+Cgk8ZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNr
ZXQuY3JlYXRvcjwva2V5PgoJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwv
c3RyaW5nPgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5pdGVtQXJyYXk8L2tleT4K
CQk8YXJyYXk+CgkJCTxkaWN0PgoJCQkJPGtleT5jb20uYXBwbGUucHJpbnQuUGFnZUZvcm1h
dC5QTUhvcml6b250YWxSZXM8L2tleT4KCQkJCTxyZWFsPjcyPC9yZWFsPgoJCQkJPGtleT5j
b20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5PgoJCQkJPHN0cmluZz5jb20uYXBw
bGUucHJpbnRpbmdtYW5hZ2VyPC9zdHJpbmc+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50
aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0OFo8L2Rh
dGU+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuc3RhdGVGbGFnPC9rZXk+CgkJ
CQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQk8L2RpY3Q+CgkJPC9hcnJheT4KCTwvZGljdD4K
CTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1PcmllbnRhdGlvbjwva2V5PgoJ
PGRpY3Q+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8
c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5h
cHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4K
CQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1PcmllbnRhdGlvbjwva2V5
PgoJCQkJPGludGVnZXI+MTwvaW50ZWdlcj4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp
Y2tldC5jbGllbnQ8L2tleT4KCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdl
cjwvc3RyaW5nPgoJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0Lm1vZERhdGU8L2tl
eT4KCQkJCTxkYXRlPjIwMDgtMDItMTJUMTY6MTg6NDhaPC9kYXRlPgoJCQkJPGtleT5jb20u
YXBwbGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJPGludGVnZXI+MDwvaW50
ZWdlcj4KCQkJPC9kaWN0PgoJCTwvYXJyYXk+Cgk8L2RpY3Q+Cgk8a2V5PmNvbS5hcHBsZS5w
cmludC5QYWdlRm9ybWF0LlBNU2NhbGluZzwva2V5PgoJPGRpY3Q+CgkJPGtleT5jb20uYXBw
bGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmlu
dGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRl
bUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4KCQkJCTxrZXk+Y29tLmFwcGxlLnBy
aW50LlBhZ2VGb3JtYXQuUE1TY2FsaW5nPC9rZXk+CgkJCQk8cmVhbD4xPC9yZWFsPgoJCQkJ
PGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5PgoJCQkJPHN0cmluZz5j
b20uYXBwbGUucHJpbnRpbmdtYW5hZ2VyPC9zdHJpbmc+CgkJCQk8a2V5PmNvbS5hcHBsZS5w
cmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0
OFo8L2RhdGU+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuc3RhdGVGbGFnPC9r
ZXk+CgkJCQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQk8L2RpY3Q+CgkJPC9hcnJheT4KCTwv
ZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFJlczwv
a2V5PgoJPGRpY3Q+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tl
eT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5
PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8
ZGljdD4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFJl
czwva2V5PgoJCQkJPHJlYWw+NzI8L3JlYWw+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50
aWNrZXQuY2xpZW50PC9rZXk+CgkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFn
ZXI8L3N0cmluZz4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9r
ZXk+CgkJCQk8ZGF0ZT4yMDA4LTAyLTEyVDE2OjE4OjQ4WjwvZGF0ZT4KCQkJCTxrZXk+Y29t
LmFwcGxlLnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCTxpbnRlZ2VyPjA8L2lu
dGVnZXI+CgkJCTwvZGljdD4KCQk8L2FycmF5PgoJPC9kaWN0PgoJPGtleT5jb20uYXBwbGUu
cHJpbnQuUGFnZUZvcm1hdC5QTVZlcnRpY2FsU2NhbGluZzwva2V5PgoJPGRpY3Q+CgkJPGtl
eT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQk8c3RyaW5nPmNvbS5h
cHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC50
aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJPGFycmF5PgoJCQk8ZGljdD4KCQkJCTxrZXk+Y29t
LmFwcGxlLnByaW50LlBhZ2VGb3JtYXQuUE1WZXJ0aWNhbFNjYWxpbmc8L2tleT4KCQkJCTxy
ZWFsPjE8L3JlYWw+CgkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuY2xpZW50PC9r
ZXk+CgkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJ
CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJCQk8ZGF0ZT4y
MDA4LTAyLTEyVDE2OjE4OjQ4WjwvZGF0ZT4KCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp
Y2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCTxpbnRlZ2VyPjA8L2ludGVnZXI+CgkJCTwvZGlj
dD4KCQk8L2FycmF5PgoJPC9kaWN0PgoJPGtleT5jb20uYXBwbGUucHJpbnQuc3ViVGlja2V0
LnBhcGVyX2luZm9fdGlja2V0PC9rZXk+Cgk8ZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmlu
dC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYWdlUmVjdDwva2V5PgoJCTxkaWN0PgoJCQk8a2V5
PmNvbS5hcHBsZS5wcmludC50aWNrZXQuY3JlYXRvcjwva2V5PgoJCQk8c3RyaW5nPmNvbS5h
cHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJPGtleT5jb20uYXBwbGUucHJpbnQu
dGlja2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4KCQkJCQk8a2V5
PmNvbS5hcHBsZS5wcmludC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYWdlUmVjdDwva2V5PgoJ
CQkJCTxhcnJheT4KCQkJCQkJPHJlYWw+MC4wPC9yZWFsPgoJCQkJCQk8cmVhbD4wLjA8L3Jl
YWw+CgkJCQkJCTxyZWFsPjczNDwvcmVhbD4KCQkJCQkJPHJlYWw+NTc2PC9yZWFsPgoJCQkJ
CTwvYXJyYXk+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwva2V5
PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwvc3RyaW5nPgoJCQkJ
CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJCQkJPGRhdGU+
MjAwOC0wMi0xMlQxNjoxODo0OVo8L2RhdGU+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQu
dGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJCTxpbnRlZ2VyPjA8L2ludGVnZXI+CgkJCQk8
L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0PgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBh
Z2VGb3JtYXQuUE1BZGp1c3RlZFBhcGVyUmVjdDwva2V5PgoJCTxkaWN0PgoJCQk8a2V5PmNv
bS5hcHBsZS5wcmludC50aWNrZXQuY3JlYXRvcjwva2V5PgoJCQk8c3RyaW5nPmNvbS5hcHBs
ZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlj
a2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4KCQkJCQk8a2V5PmNv
bS5hcHBsZS5wcmludC5QYWdlRm9ybWF0LlBNQWRqdXN0ZWRQYXBlclJlY3Q8L2tleT4KCQkJ
CQk8YXJyYXk+CgkJCQkJCTxyZWFsPi0xODwvcmVhbD4KCQkJCQkJPHJlYWw+LTE4PC9yZWFs
PgoJCQkJCQk8cmVhbD43NzQ8L3JlYWw+CgkJCQkJCTxyZWFsPjU5NDwvcmVhbD4KCQkJCQk8
L2FycmF5PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5jbGllbnQ8L2tleT4K
CQkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8L3N0cmluZz4KCQkJCQk8
a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJCTxkYXRlPjIw
MDgtMDItMTJUMTY6MTg6NDlaPC9kYXRlPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRp
Y2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4wPC9pbnRlZ2VyPgoJCQkJPC9k
aWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5wcmludC5QYXBl
ckluZm8uUE1QYXBlck5hbWU8L2tleT4KCQk8ZGljdD4KCQkJPGtleT5jb20uYXBwbGUucHJp
bnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmluZz5jb20uYXBwbGUucHJpbnQucG0u
UG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQuaXRl
bUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0PgoJCQkJCTxrZXk+Y29tLmFwcGxl
LnByaW50LlBhcGVySW5mby5QTVBhcGVyTmFtZTwva2V5PgoJCQkJCTxzdHJpbmc+bmEtbGV0
dGVyPC9zdHJpbmc+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNsaWVudDwv
a2V5PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50LnBtLlBvc3RTY3JpcHQ8L3N0cmlu
Zz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5PgoJCQkJ
CTxkYXRlPjIwMDAtMDctMjhUMjI6NTc6MDRaPC9kYXRlPgoJCQkJCTxrZXk+Y29tLmFwcGxl
LnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4xPC9pbnRlZ2Vy
PgoJCQkJPC9kaWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5hcHBsZS5w
cmludC5QYXBlckluZm8uUE1VbmFkanVzdGVkUGFnZVJlY3Q8L2tleT4KCQk8ZGljdD4KCQkJ
PGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmluZz5j
b20uYXBwbGUucHJpbnQucG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5hcHBs
ZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0PgoJ
CQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYWdlUmVj
dDwva2V5PgoJCQkJCTxhcnJheT4KCQkJCQkJPHJlYWw+MC4wPC9yZWFsPgoJCQkJCQk8cmVh
bD4wLjA8L3JlYWw+CgkJCQkJCTxyZWFsPjczNDwvcmVhbD4KCQkJCQkJPHJlYWw+NTc2PC9y
ZWFsPgoJCQkJCTwvYXJyYXk+CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNs
aWVudDwva2V5PgoJCQkJCTxzdHJpbmc+Y29tLmFwcGxlLnByaW50aW5nbWFuYWdlcjwvc3Ry
aW5nPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5tb2REYXRlPC9rZXk+CgkJ
CQkJPGRhdGU+MjAwOC0wMi0xMlQxNjoxODo0OFo8L2RhdGU+CgkJCQkJPGtleT5jb20uYXBw
bGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJCTxpbnRlZ2VyPjA8L2ludGVn
ZXI+CgkJCQk8L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0PgoJCTxrZXk+Y29tLmFwcGxl
LnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYXBlclJlY3Q8L2tleT4KCQk8ZGljdD4K
CQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LmNyZWF0b3I8L2tleT4KCQkJPHN0cmlu
Zz5jb20uYXBwbGUucHJpbnQucG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQk8a2V5PmNvbS5h
cHBsZS5wcmludC50aWNrZXQuaXRlbUFycmF5PC9rZXk+CgkJCTxhcnJheT4KCQkJCTxkaWN0
PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LlBhcGVySW5mby5QTVVuYWRqdXN0ZWRQYXBl
clJlY3Q8L2tleT4KCQkJCQk8YXJyYXk+CgkJCQkJCTxyZWFsPi0xODwvcmVhbD4KCQkJCQkJ
PHJlYWw+LTE4PC9yZWFsPgoJCQkJCQk8cmVhbD43NzQ8L3JlYWw+CgkJCQkJCTxyZWFsPjU5
NDwvcmVhbD4KCQkJCQk8L2FycmF5PgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tl
dC5jbGllbnQ8L2tleT4KCQkJCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludGluZ21hbmFnZXI8
L3N0cmluZz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC50aWNrZXQubW9kRGF0ZTwva2V5
PgoJCQkJCTxkYXRlPjIwMDgtMDItMTJUMTY6MTg6NDhaPC9kYXRlPgoJCQkJCTxrZXk+Y29t
LmFwcGxlLnByaW50LnRpY2tldC5zdGF0ZUZsYWc8L2tleT4KCQkJCQk8aW50ZWdlcj4wPC9p
bnRlZ2VyPgoJCQkJPC9kaWN0PgoJCQk8L2FycmF5PgoJCTwvZGljdD4KCQk8a2V5PmNvbS5h
cHBsZS5wcmludC5QYXBlckluZm8ucHBkLlBNUGFwZXJOYW1lPC9rZXk+CgkJPGRpY3Q+CgkJ
CTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5jcmVhdG9yPC9rZXk+CgkJCTxzdHJpbmc+
Y29tLmFwcGxlLnByaW50LnBtLlBvc3RTY3JpcHQ8L3N0cmluZz4KCQkJPGtleT5jb20uYXBw
bGUucHJpbnQudGlja2V0Lml0ZW1BcnJheTwva2V5PgoJCQk8YXJyYXk+CgkJCQk8ZGljdD4K
CQkJCQk8a2V5PmNvbS5hcHBsZS5wcmludC5QYXBlckluZm8ucHBkLlBNUGFwZXJOYW1lPC9r
ZXk+CgkJCQkJPHN0cmluZz5MZXR0ZXI8L3N0cmluZz4KCQkJCQk8a2V5PmNvbS5hcHBsZS5w
cmludC50aWNrZXQuY2xpZW50PC9rZXk+CgkJCQkJPHN0cmluZz5jb20uYXBwbGUucHJpbnQu
cG0uUG9zdFNjcmlwdDwvc3RyaW5nPgoJCQkJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tl
dC5tb2REYXRlPC9rZXk+CgkJCQkJPGRhdGU+MjAwMC0wNy0yOFQyMjo1NzowNFo8L2RhdGU+
CgkJCQkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0LnN0YXRlRmxhZzwva2V5PgoJCQkJ
CTxpbnRlZ2VyPjE8L2ludGVnZXI+CgkJCQk8L2RpY3Q+CgkJCTwvYXJyYXk+CgkJPC9kaWN0
PgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5BUElWZXJzaW9uPC9rZXk+CgkJPHN0
cmluZz4wMC4yMDwvc3RyaW5nPgoJCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5wcml2
YXRlTG9jazwva2V5PgoJCTxmYWxzZS8+CgkJPGtleT5jb20uYXBwbGUucHJpbnQudGlja2V0
LnR5cGU8L2tleT4KCQk8c3RyaW5nPmNvbS5hcHBsZS5wcmludC5QYXBlckluZm9UaWNrZXQ8
L3N0cmluZz4KCTwvZGljdD4KCTxrZXk+Y29tLmFwcGxlLnByaW50LnRpY2tldC5BUElWZXJz
aW9uPC9rZXk+Cgk8c3RyaW5nPjAwLjIwPC9zdHJpbmc+Cgk8a2V5PmNvbS5hcHBsZS5wcmlu
dC50aWNrZXQucHJpdmF0ZUxvY2s8L2tleT4KCTxmYWxzZS8+Cgk8a2V5PmNvbS5hcHBsZS5w
cmludC50aWNrZXQudHlwZTwva2V5PgoJPHN0cmluZz5jb20uYXBwbGUucHJpbnQuUGFnZUZv
cm1hdFRpY2tldDwvc3RyaW5nPgo8L2RpY3Q+CjwvcGxpc3Q+CgA4QklNA+kAAAAAAHgAAwAA
AEgASAAAAAAC3gJA/+7/7gMGAlIDZwUoA/wAAgAAAEgASAAAAAAC2AIoAAEAAABkAAAAAQAD
AwMAAAABf/8AAQABAAAAAAAAAAAAAAAAaAgAGQGQAAAAAAAgAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAA4QklNA+0AAAAAABAAYAAAAAEAAQBgAAAAAQABOEJJTQQmAAAAAAAOAAAAAAAA
AAAAAD+AAAA4QklNBA0AAAAAAAQAAAB4OEJJTQQZAAAAAAAEAAAAHjhCSU0D8wAAAAAACQAA
AAAAAAAAAQA4QklNBAoAAAAAAAEAADhCSU0nEAAAAAAACgABAAAAAAAAAAE4QklNA/UAAAAA
AEgAL2ZmAAEAbGZmAAYAAAAAAAEAL2ZmAAEAoZmaAAYAAAAAAAEAMgAAAAEAWgAAAAYAAAAA
AAEANQAAAAEALQAAAAYAAAAAAAE4QklNA/gAAAAAAHAAAP//////////////////////////
//8D6AAAAAD/////////////////////////////A+gAAAAA////////////////////////
/////wPoAAAAAP////////////////////////////8D6AAAOEJJTQQIAAAAAAAQAAAAAQAA
AkAAAAJAAAAAADhCSU0EHgAAAAAABAAAAAA4QklNBBoAAAAAA0kAAAAGAAAAAAAAAAAAAABo
AAABKwAAAAoAVQBuAHQAaQB0AGwAZQBkAC0AMQAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAQAA
AAAAAAAAAAABKwAAAGgAAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAQAAAA
AQAAAAAAAG51bGwAAAACAAAABmJvdW5kc09iamMAAAABAAAAAAAAUmN0MQAAAAQAAAAAVG9w
IGxvbmcAAAAAAAAAAExlZnRsb25nAAAAAAAAAABCdG9tbG9uZwAAAGgAAAAAUmdodGxvbmcA
AAErAAAABnNsaWNlc1ZsTHMAAAABT2JqYwAAAAEAAAAAAAVzbGljZQAAABIAAAAHc2xpY2VJ
RGxvbmcAAAAAAAAAB2dyb3VwSURsb25nAAAAAAAAAAZvcmlnaW5lbnVtAAAADEVTbGljZU9y
aWdpbgAAAA1hdXRvR2VuZXJhdGVkAAAAAFR5cGVlbnVtAAAACkVTbGljZVR5cGUAAAAASW1n
IAAAAAZib3VuZHNPYmpjAAAAAQAAAAAAAFJjdDEAAAAEAAAAAFRvcCBsb25nAAAAAAAAAABM
ZWZ0bG9uZwAAAAAAAAAAQnRvbWxvbmcAAABoAAAAAFJnaHRsb25nAAABKwAAAAN1cmxURVhU
AAAAAQAAAAAAAG51bGxURVhUAAAAAQAAAAAAAE1zZ2VURVhUAAAAAQAAAAAABmFsdFRhZ1RF
WFQAAAABAAAAAAAOY2VsbFRleHRJc0hUTUxib29sAQAAAAhjZWxsVGV4dFRFWFQAAAABAAAA
AAAJaG9yekFsaWduZW51bQAAAA9FU2xpY2VIb3J6QWxpZ24AAAAHZGVmYXVsdAAAAAl2ZXJ0
QWxpZ25lbnVtAAAAD0VTbGljZVZlcnRBbGlnbgAAAAdkZWZhdWx0AAAAC2JnQ29sb3JUeXBl
ZW51bQAAABFFU2xpY2VCR0NvbG9yVHlwZQAAAABOb25lAAAACXRvcE91dHNldGxvbmcAAAAA
AAAACmxlZnRPdXRzZXRsb25nAAAAAAAAAAxib3R0b21PdXRzZXRsb25nAAAAAAAAAAtyaWdo
dE91dHNldGxvbmcAAAAAADhCSU0EKAAAAAAADAAAAAE/8AAAAAAAADhCSU0EFAAAAAAABAAA
AAM4QklNBAwAAAAACcQAAAABAAAAoAAAADgAAAHgAABpAAAACagAGAAB/9j/4AAQSkZJRgAB
AgEASABIAAD/7QAMQWRvYmVfQ00AAf/uAA5BZG9iZQBkgAAAAAH/2wCEAAwICAgJCAwJCQwR
CwoLERUPDAwPFRgTExUTExgRDAwMDAwMEQwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwB
DQsLDQ4NEA4OEBQODg4UFA4ODg4UEQwMDAwMEREMDAwMDAwRDAwMDAwMDAwMDAwMDAwMDAwM
DAwMDAwMDAwMDP/AABEIADgAoAMBIgACEQEDEQH/3QAEAAr/xAE/AAABBQEBAQEBAQAAAAAA
AAADAAECBAUGBwgJCgsBAAEFAQEBAQEBAAAAAAAAAAEAAgMEBQYHCAkKCxAAAQQBAwIEAgUH
BggFAwwzAQACEQMEIRIxBUFRYRMicYEyBhSRobFCIyQVUsFiMzRygtFDByWSU/Dh8WNzNRai
soMmRJNUZEXCo3Q2F9JV4mXys4TD03Xj80YnlKSFtJXE1OT0pbXF1eX1VmZ2hpamtsbW5vY3
R1dnd4eXp7fH1+f3EQACAgECBAQDBAUGBwcGBTUBAAIRAyExEgRBUWFxIhMFMoGRFKGxQiPB
UtHwMyRi4XKCkkNTFWNzNPElBhaisoMHJjXC0kSTVKMXZEVVNnRl4vKzhMPTdePzRpSkhbSV
xNTk9KW1xdXl9VZmdoaWprbG1ub2JzdHV2d3h5ent8f/2gAMAwEAAhEDEQA/APVUkkklKSSS
SUpJJJJSkkkklKSSSSUpJJJJSkkkklKSSQsnKxsSh2RlWsoor1fbY4MaP6z3w1JSVJY9X1u+
r1xPo5XqgcvZVa5v/bjajX/0lpY2Xi5dQuxbWX1ExvrcHCRy2W/nImMhuCPMKTJJJIKf/9D1
VJJCysmnExrcq92ymhjrbXngMYC97v7LWpKaWL9YelZfWcrolNu7PwmNsurgxtds+g/6L9nq
1ep/XVjp/UsXqNVluMXFtN1mO/cC330uNNvP5u9q836D1TIoy+hdVu6fmYjszLyTn5Tqmtxr
W9Vd6mPsv3vs2VXMwfT/AEdSNca3MwsLqdho6Bldb6ozqj95qaXtfe7Bxr7WFv6C67f6m72e
xJT6YquT1LDxsvEwrn7cjPc9mMyCdxrY6+3+q1lbFyHVcL6uG7pvQOmYWHktJyb8f7Ve8Ydb
g5gzGMprdazMy/0vsxP+0vv/AJhix6sLAy6fq30zrBrs6dVndVxXD1HioMqGQzGqrybLPX9J
mytmNvv9T0/TSU+oqpjdTxMrOzMGkuN/T3VtyAWkAG1gvr2v+i/9G5eePxMjqmZ12x5w6+q4
uW9uHn5WbZj34mPWGv6ZZjYteLdWzC9Kz1vUbd+verb6y6LB/aJ639bh041DP34nom2fT3/Z
KvpbfckBZpT0N/UsWjqGL06wu+0ZjbX0gAkEU7PV3O/N/nmK2uU6nltwvrJ9XLur30Y9oxM0
ZFheK6vVLcPeKnXFvt37vTSxesdJZ9cM7KdnY4xsjpuO+m82s9N7arcxt767d3pu9F3857vY
kp6tJeVfV2jHyMP6tdN+sjjX0e7Fyr6KrnurrvyzlOfj15GrfV/U7G341dn85v8A7C0uo4fT
cjrtHRm0Yeb0hvT2jpOPlZltNBO+6rLfiW1U532jLqZ6df8AOMvxqvfSkp9DVbL6hi4b8eu9
xD8y0Y+O0AuLrC11ke0e1ra6rLHvf7FxGJ0OzP6l0Lp3XsodQNfTcp1z6LnuqyGNvx/std9r
RQ/KrbWaX272/pr6P0nqKg3G6fb07E6V1Ah3SsT6y5ODXXbY4NZQyvKbj4/rF/qe3d+j/SJK
fS776caizIvcK6aWusseeGtaN73H+q0LxzqfVup/WrqrLXNfYLX7Om4A0awH+b9v0PtDmfpM
nJf/ADf8jHqXov1qoro+oufR06G49OCWUhhLh6DGAQ1/u3M+ztXMfVAdP6N0e36050WW2udi
4FLCC+eHtb+5de+t30v5rFp9T/CWKxgoCUquV8MR4lIbvRsHpXTMjKwdv23MwMZ2R1PNa4hl
VkfoMHDZP0/pve//ALc/nPTx9DGfbbU3Ox6ThdcqYHWUn2symt/nKLdv6K392u5vvx7f+21i
PzbnfU3My7GMp/aOU3EpqqEMbSx3q2tk/pLrLbPtz78i79LfZasnAzzg4Vl9NgaftVba6mkS
C2u1+Td6f5rNj8ejf/hbP+IU3tmQkSdQeH/vv8FNF9Xw8qrMxacukk1XsbYydDDhuhw/Nc38
5GWN9ULvX+r+PcBDbH3PZ/Vdda5q2VSkKkR2JC1//9H1VQuppyKX0X1ttptaWWVvAc1zTo5j
2O9rmuU0klIbcTEuoGNdTXZQ3bFL2hzBsIfV+jcNn6NzGur/AHFF3T8B+PbivxqnY97nPupL
Glj3PPqWvtr27LHWP9796sJJKaD+gdCfiMwn9OxXYlTi+vHNFZra4zueynZ6bHu3KTuidGdh
N6e7AxjhMdvbimms1B2p3to2elu9zvzVdSSU0cnofRcx1b8vp+NkOoaGUutpreWNH0WV72O2
Mb/JVmvGx6rbbq6mMtvIN1jWgOeWj02G1491mxg2M3oqSSmpndJ6V1LZ+0cLHzPSn0/tFTLd
u6N+z1Wv27tjUOzoHQbcarEt6biPxscuNFDqKzWwvO6z0qizZXvd9PYr6SSmvlYGDm4/2bMx
6snH0Po2sa9nt+j+jsDme1Cu6N0fIxasO/BxrcWiPRx31MdWyBtHpVOb6dftP5iupJKQVYGD
QanU49VTqKzTSWMa0srO0mmraP0dXsZ+jYoWdK6Xbj24tmHQ/HyHm2+l1TCyywn1HXW1luyy
1zxv9R6tJJKQnDxDhnB9FgxDX6PoAAM9Pb6fotY32tr9P2bF4x1jouX9V+oOw8prjiOd+pZh
+haz81rnD2tyq2D07q/p/ns/RL21CycbGy6H4+VUy+iwQ+q1oexw597H7muUmLKcZsfUJBp8
0xetfVi36r4nTeq/aX3YV77xj44LfVLnXOYDkfzbanV5Pv8A01F3/f8AMop6j9berjHwKWY1
dTGUj0mzTh4zS4sY5/s9W73Wva136TJv/wBFj/zXoJ/xe/U42+r+zGAzO1r7Az/thtno/wDg
a28PCw8GhuNhUV41Dfo1VNDGiefawBSnmIizEHiNn1bR4t+FJkrBw6MDDowsZu2jGrbVW3n2
tG1sn95HSSVZa//S9VSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp+qkl8qpJKfqpJfKqSSn
6qSXyqkkp+qkl8qpJKfqpJfKqSSn6qSXyqkkp//ZOEJJTQQhAAAAAABTAAAAAQEAAAAPAEEA
ZABvAGIAZQAgAFAAaABvAHQAbwBzAGgAbwBwAAAAEgBBAGQAbwBiAGUAIABQAGgAbwB0AG8A
cwBoAG8AcAAgAEMAUwAAAAEAOEJJTQQGAAAAAAAHAAgAAAABAQD/4RgFaHR0cDovL25zLmFk
b2JlLmNvbS94YXAvMS4wLwA8P3hwYWNrZXQgYmVnaW49J++7vycgaWQ9J1c1TTBNcENlaGlI
enJlU3pOVGN6a2M5ZCc/Pgo8eDp4bXBtZXRhIHhtbG5zOng9J2Fkb2JlOm5zOm1ldGEvJyB4
OnhtcHRrPSdYTVAgdG9vbGtpdCAzLjAtMjgsIGZyYW1ld29yayAxLjYnPgo8cmRmOlJERiB4
bWxuczpyZGY9J2h0dHA6Ly93d3cudzMub3JnLzE5OTkvMDIvMjItcmRmLXN5bnRheC1ucyMn
IHhtbG5zOmlYPSdodHRwOi8vbnMuYWRvYmUuY29tL2lYLzEuMC8nPgoKIDxyZGY6RGVzY3Jp
cHRpb24gcmRmOmFib3V0PSd1dWlkOmE0NDZmNTA1LWRhZWQtMTFkYy05N2EzLWRjOWRjYjU2
NWRiYycKICB4bWxuczpleGlmPSdodHRwOi8vbnMuYWRvYmUuY29tL2V4aWYvMS4wLyc+CiAg
PGV4aWY6Q29sb3JTcGFjZT40Mjk0OTY3Mjk1PC9leGlmOkNvbG9yU3BhY2U+CiAgPGV4aWY6
UGl4ZWxYRGltZW5zaW9uPjI5OTwvZXhpZjpQaXhlbFhEaW1lbnNpb24+CiAgPGV4aWY6UGl4
ZWxZRGltZW5zaW9uPjEwNDwvZXhpZjpQaXhlbFlEaW1lbnNpb24+CiA8L3JkZjpEZXNjcmlw
dGlvbj4KCiA8cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0ndXVpZDphNDQ2ZjUwNS1kYWVk
LTExZGMtOTdhMy1kYzlkY2I1NjVkYmMnCiAgeG1sbnM6cGRmPSdodHRwOi8vbnMuYWRvYmUu
Y29tL3BkZi8xLjMvJz4KIDwvcmRmOkRlc2NyaXB0aW9uPgoKIDxyZGY6RGVzY3JpcHRpb24g
cmRmOmFib3V0PSd1dWlkOmE0NDZmNTA1LWRhZWQtMTFkYy05N2EzLWRjOWRjYjU2NWRiYycK
ICB4bWxuczpwaG90b3Nob3A9J2h0dHA6Ly9ucy5hZG9iZS5jb20vcGhvdG9zaG9wLzEuMC8n
PgogIDxwaG90b3Nob3A6SGlzdG9yeT48L3Bob3Rvc2hvcDpIaXN0b3J5PgogPC9yZGY6RGVz
Y3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUt
ZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnRpZmY9J2h0dHA6Ly9ucy5h
ZG9iZS5jb20vdGlmZi8xLjAvJz4KICA8dGlmZjpPcmllbnRhdGlvbj4xPC90aWZmOk9yaWVu
dGF0aW9uPgogIDx0aWZmOlhSZXNvbHV0aW9uPjk2LzE8L3RpZmY6WFJlc29sdXRpb24+CiAg
PHRpZmY6WVJlc29sdXRpb24+OTYvMTwvdGlmZjpZUmVzb2x1dGlvbj4KICA8dGlmZjpSZXNv
bHV0aW9uVW5pdD4yPC90aWZmOlJlc29sdXRpb25Vbml0PgogPC9yZGY6RGVzY3JpcHRpb24+
CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUtZGFlZC0xMWRj
LTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnhhcD0naHR0cDovL25zLmFkb2JlLmNvbS94
YXAvMS4wLyc+CiAgPHhhcDpDcmVhdGVEYXRlPjIwMDgtMDItMTJUMTE6MTk6NTQtMDU6MDA8
L3hhcDpDcmVhdGVEYXRlPgogIDx4YXA6TW9kaWZ5RGF0ZT4yMDA4LTAyLTEyVDExOjE5OjU0
LTA1OjAwPC94YXA6TW9kaWZ5RGF0ZT4KICA8eGFwOk1ldGFkYXRhRGF0ZT4yMDA4LTAyLTEy
VDExOjE5OjU0LTA1OjAwPC94YXA6TWV0YWRhdGFEYXRlPgogIDx4YXA6Q3JlYXRvclRvb2w+
QWRvYmUgUGhvdG9zaG9wIENTIE1hY2ludG9zaDwveGFwOkNyZWF0b3JUb29sPgogPC9yZGY6
RGVzY3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1
MDUtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjJwogIHhtbG5zOnhhcE1NPSdodHRwOi8v
bnMuYWRvYmUuY29tL3hhcC8xLjAvbW0vJz4KICA8eGFwTU06RG9jdW1lbnRJRD5hZG9iZTpk
b2NpZDpwaG90b3Nob3A6YTQ0NmY1MDQtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJjPC94
YXBNTTpEb2N1bWVudElEPgogPC9yZGY6RGVzY3JpcHRpb24+CgogPHJkZjpEZXNjcmlwdGlv
biByZGY6YWJvdXQ9J3V1aWQ6YTQ0NmY1MDUtZGFlZC0xMWRjLTk3YTMtZGM5ZGNiNTY1ZGJj
JwogIHhtbG5zOmRjPSdodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyc+CiAgPGRj
OmZvcm1hdD5pbWFnZS9qcGVnPC9kYzpmb3JtYXQ+CiA8L3JkZjpEZXNjcmlwdGlvbj4KCjwv
cmRmOlJERj4KPC94OnhtcG1ldGE+CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAog
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
IAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAK
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgIAo8P3hwYWNrZXQgZW5kPSd3Jz8+/+ICOElDQ19QUk9GSUxFAAEBAAAC
KEFEQkUCEAAAbW50clJHQiBYWVogB88ABgADAAAAAAAAYWNzcEFQUEwAAAAAbm9uZQAAAAAA
AAAAAAAAAAAAAAAAAPbWAAEAAAAA0y1BREJFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAKY3BydAAAAPwAAAAyZGVzYwAAATAAAABkd3RwdAAAAZQA
AAAUYmtwdAAAAagAAAAUclRSQwAAAbwAAAAOZ1RSQwAAAcwAAAAOYlRSQwAAAdwAAAAOclhZ
WgAAAewAAAAUZ1hZWgAAAgAAAAAUYlhZWgAAAhQAAAAUdGV4dAAAAABDb3B5cmlnaHQgMTk5
OSBBZG9iZSBTeXN0ZW1zIEluY29ycG9yYXRlZAAAAGRlc2MAAAAAAAAACkFwcGxlIFJHQgAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAABYWVogAAAAAAAA81EAAQAAAAEWzFhZWiAAAAAAAAAA
AAAAAAAAAAAAY3VydgAAAAAAAAABAc0AAGN1cnYAAAAAAAAAAQHNAABjdXJ2AAAAAAAAAAEB
zQAAWFlaIAAAAAAAAHm9AABBUgAABLlYWVogAAAAAAAAVvgAAKwvAAAdA1hZWiAAAAAAAAAm
IgAAEn8AALFw/+4ADkFkb2JlAGRAAAAAAf/bAIQAAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEB
AQEBAQEBAQEBAQEBAQEBAQEBAQICAgICAgICAgICAwMDAwMDAwMDAwEBAQEBAQEBAQEBAgIB
AgIDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMD/8AA
EQgAaAErAwERAAIRAQMRAf/dAAQAJv/EAaIAAAAGAgMBAAAAAAAAAAAAAAcIBgUECQMKAgEA
CwEAAAYDAQEBAAAAAAAAAAAABgUEAwcCCAEJAAoLEAACAQMEAQMDAgMDAwIGCXUBAgMEEQUS
BiEHEyIACDEUQTIjFQlRQhZhJDMXUnGBGGKRJUOhsfAmNHIKGcHRNSfhUzaC8ZKiRFRzRUY3
R2MoVVZXGrLC0uLyZIN0k4Rlo7PD0+MpOGbzdSo5OkhJSlhZWmdoaWp2d3h5eoWGh4iJipSV
lpeYmZqkpaanqKmqtLW2t7i5usTFxsfIycrU1dbX2Nna5OXm5+jp6vT19vf4+foRAAIBAwIE
BAMFBAQEBgYFbQECAxEEIRIFMQYAIhNBUQcyYRRxCEKBI5EVUqFiFjMJsSTB0UNy8BfhgjQl
klMYY0TxorImNRlUNkVkJwpzg5NGdMLS4vJVZXVWN4SFo7PD0+PzKRqUpLTE1OT0laW1xdXl
9ShHV2Y4doaWprbG1ub2Z3eHl6e3x9fn90hYaHiImKi4yNjo+DlJWWl5iZmpucnZ6fkqOkpa
anqKmqq6ytrq+v/aAAwDAQACEQMRAD8A3+Pfuvde9+691737r3Xvfuvde9+691737r3Xvfuv
de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+
691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917
37r3Xvfuvde9+691737r3Xvfuvde9+691//Q3+Pfuvde9+691737r3Xvfuvde9+691737r3X
vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv
de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+
691737r3Xvfuvde9+691737r3Xvfuvde9+691//R3+Pfuvde9+691737r3Xvfuvde9+69173
7r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3X
vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv
de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691//S3+Pfuvde9+691737r3Xvfuvde9+6
91737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+69173
7r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3XvfuvdNG4NwYHaeCzO
6N05vEba2zt3F12b3BuLcGSo8NgsFhcXTS1uTy+Zy+RmpsfjMXjqOF5p6ieRIoYkZnYKCfd4
45JpEhhjZ5WICqASSTgAAZJJwAOPXgCTQcetWb5p/wDCnzqPr/NZPYPwj66g73zeOqKiiq+3
ewTmdudTR1dO3jvtjbdC2N3rvuiMoZWqJKjAU50h6d6qJ1k9zVyz7N3t+sdzzFdm2hND4cdG
lp/SY1RD8qOfUAinS1LNiKyGg9PPqr3bf8+D+Yx2buCN63vzDdaU1TK9sPsjp7p6fFU6zMpV
Ipd/7I31mWSnAITXWMxBOosbESintLyVbQf8kdp3A4vNMCf+ccka/wAunPAhXBX+Z6td6K+e
f8yLN09HuDb3bvV/e1OtPF93tPsnq/bWAqayMa2mkpMp1PH17LS5KZX0o7JLToVUmBxqVwXv
HI/I8JaGbb7izNcPFKzAfaJfEqP2H59NPHAMUI6s7+Pn80/YO89y4rqr5NbFyvxb7fytemJw
y7lr/wCM9T7wyE8vix9Jt3sf7PGR4nL5Ij0UWZpaBTK6QU9TVzOFMd757dX1lBJuOxXS7hti
irFBSVB5loqmoH8SFsVLKo6aaAgakOpf5/s6ta9xz0x1737r3Xvfuvde9+691737r3X/09/j
37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3WKSoghenjmmiikqpW
p6VJJER6mdYJqpoadWYNNKtNTSSFVuRHGzWspI917rL7917r3v3Xuve/de697917r3v3Xuve
/de697917r3v3Xuve/de697917r3v3Xuve/de697917r3v3XuscksUKh5pI4kMkUQaR1RTLP
KkEEYZiAZJppFRB9WZgByR7917rJ7917r3v3Xuvnkfz4P5ue7vmJ3Ruz4odFbqqcb8UOo9yV
G3c/WYGtlhTvnsHbtaafM5/M1dLL4cp11tvN0rwYGkQyUla8H8VkaUy0aUeTPtpyRFtdpDvF
/DXdplqtR/ZIeAHo7DLniK6MUbUcWltoUSMP1D/If5+qKNq7fqayqocdj6Oor8hXVFPR0VFR
08tTV1lZVSpBTUlLTQI81RU1EzqkaICzsQACT7myKNY0LuwCAVJPAAcST8vM9LCmMjrcm+Iv
8tv4w/yzfjinzf8A5mGLxG6Oz5aWCfYnReYpsfuDGbczmTpGrNu7PptqVh/h2++4MhHTyPOK
ppMRgoxJKwBpJK+OB99505g563scpchyNHY1pJcKSpZQaM5cZjhFRSlHkNB+IISmSRp5DHB8
Pr/q8urEfmF2/wBa7F+FHx3+Wk3Te0Olew+1dxdZVeCwG34MXHlsZh9+bbzu8YNtZ3LYnEbe
bNfabRohLUJNShKatBVEQ3JDvJ22bhe84b7ywN1lu9vto5gzMW0lo3WPWqszaauaChyvEnpO
AdbpWoHSu2Vjfjt/ME6Vm23ubH4WXc9RiZIoaiSKmM7SPDpJBseWbgE3IPtjcRvvI27C4gZ/
o9WRnrys0bVB6kfAXvjsLpXt6t/l+fIXO1udqaDGZHKfGXsDO1ktXltw7YwFPJV5jqbMV9Tq
qMpltp4Onevw9S7ySS4mnqaeRlFFT+ZBzpstluW2pznskISNmAuo1FArsaCVQMAO3a48nKkf
EaWlRWXxU/Mf5erofcVdJ+ve/de697917r3v3Xuv/9Tf49+691737r3XvfuvdNuZzGL29h8r
n83WwYzC4PG12Yy+Sqm0U2PxeMpZa2vral7HRBSUkDyObcKp9+6918s3Ed59yYz5c7d/4UC1
xy1F1Fn/AOahuDr6tpp6KaqyFFs6bbtBvWq2VBURkvUQH4+5eu2/ABHqY0D6GZ45Au+nfLT8
ut6b/hQBW0eR/kz/ADWyGPq6avoK/rzr2toa6iniqqOto6ruDrSemq6Spgd4aimqIXV45EYq
6kEEg+9dUX4h1D/lP90dXfHb+R98SO7e6t6Ybrzq3rf43UO4t47vz07Q4/E42HPZaniVY4km
q6/JZKvqIaShoqaOarr62eKmp45Z5Y429145Y9EL/wCgrz4l5bcGbr9ifEb5r9gdJ7XyFTQb
l7o251/tOXHY2OjSCafJtg23g0FJj/tJvuQuRyWOrFpjG8kEbOUT3W9B9erPPnh/OD+O/wAC
63prae6etvkF3Z2t37tebePWPU3Q3XUO8d25Pb0DUsUmRyTZDOYTG0kaz1QjMNLLXVupSfB4
7OfdVCk9Fe+J/wDwoa+OHf3yR2Z8S+5/j98lvhl3h2ZW0mO61w3yE2ZTYLB7uyuWllg25gly
Bq6PPYjN7pqIvBjRVYpKCsqyKdKszvDHL7rZXFQeqtv5sn8wfKfDX/hQx8NN8dsVnd3Yfxx6
E+Ped3zTdH9S01DuXJnefbXWHyU6qrN2bZ2Nmty7Q23ktxO+YoPvayprYqiHE0soidgohk31
tRVT69Xxw/zovhtjfgPiP5ifZCdr9MdNbk3JuLZe1tkdn7NxOP7q3TvPb+Yz2Gj2rgdkbZ3V
unGVmWzr7bqqmkJyiUyUMZqKqWmiWRk11XSa06Il1t/wp1+KOY331/g++fjT8uvid1l3BkKW
j6s767u68xeL6uztHWpRPT5zM5Wizcz4/AWr4mlrsZ/G6OlhlSeolhpy8qe63pPr0V7/AIVG
VVNXdhfydq2iqYKyjrPlNuCqpKulmjqKaqpqjN9DywVNNPEzxTwTxOGR1JVlIIJB9+9etpwb
rb39+6p1W9/OF/7dY/zAP/FUu5f/AHjsl791scR0VH/hNd/25U+F/wD5cX/8Fj3t791tviPR
tfk7/M76E+KHy2+Kfw07E2j29mez/mBlqDD9aZ7ZeA2ZkdhYOpyO6qbaEEm+cpnN/bc3DjIF
yVUrucfi8mwgBYKXsh91oAkE9Cv87/m11V/L0+Ne8PlN3Rt/sHc/X2ysttHD5XD9X4rbma3j
UVO9NzYzauLkx+P3XuvZWElggyOVjeoMmQiZIVYorsAh914Cpp0kuwP5kPxY6i+EXXfz97f3
dkOsOjO0+rusu0dm0G6qCkk7Fy8Xbm0MZvTZmxaDaW3snnUzPYdTisoBPQ46qraen8FRO1R9
nTy1S+69Q1p1UP0f/wAKevjj3d2rsDaGP+H3zM231d2l2NtPrDZfeOW2TtKq2iu4N67potnY
Kp3SuO3XPi8Ni/43koUqWo8nk6iEFtMMjroPut6TTj0fX+Yl/OY+MH8u/eOz+mdx7d7R75+S
/YePpMvs348dEbYXdO+ajEZGpr6TG5jOz1NVRY/DUWTqcXUpTU8RrcrUGIvHRNBeZfdeCk/Z
0kPgJ/PC+Mvzl7jy/wAYsp193P8AFj5TYfFyZdOi/kXtem2nuDc1JS4tM3k12dVxVsz5HIYr
CyLXS0NfTYvIy0Gqqp6eanhqJYfdeKkZ8uk/8mv+FAnwR+Ivyg+Q3xP7ux/d23+wPjnsPb+9
s7nqHZ+y8ps3f9bu3aXW28NsbA6ukg7HXd+f35mqHtChiSCtw+MoIJKWtlqKyGjpzVN7r2kk
A9A18O/+FJvxA+UvyT2v8Vt59Sd+/GHs/sXN4/b3WLd0YHb9LgN053OeMbW23kqnFZyfL7Y3
NvCWoiixcU1DLQVc8qQrWCWSBJfdeKkCvRDf583zq3Z8TP5uf8sqr3dne3NxfGvrHbNB31u7
pLquKjyVdu7eCb07H2qM9SbVrs1tjE7r3NR4mgp4aFcjXJFRRiZqcxNPOZfdbUVU+vWxX/L1
/mCdVfzI+l9z949Qdd92dZ7a2r2fmuqK7A987S25szd9XnMHtTZW76nL43F7Y3pvqgn23PQb
6poYah6uOZqqnqEMKqiPJ7qpFMHoI/50XygynxG/lr/JvtLbFecdvrMbToup+v6uGpNLkKHd
fb2ZoOvos7iJgQVy20MPnazNwfX143kEXHsTcnbam7cybXaSrWHxNbDyKoNVD8mICn7enraP
xJo18q1/Z18wHamORIYdK86FH0/w95rWUICKadCaNK463Kv5Av8ALY21tnbEv8yT5O4+gxG2
Np0GZ3D0Rj91pHT4rEYzbFPWTbk75zkVV+3DR4aOinTAvMoERgmySL6cfUe4T92ec55Zl5I2
Fi1xIVWcpxYuRpt1p5tUeJTjUIfxr0WbhOdQtIcseNP5D/P+zrh8xeo/mH/Ms+TnQ3Z3aW0s
l038T98967S6U+NvWm+q2uwHY259gZWWu3Lv7taDrpIP43ja/JbB2pX53IVmTTGmCgioqOne
oWOOaQ05U3PlfkLYN5sNvuVuuZIbN57qWMBokkWixwmWukgSOsaqmurF2YLUgWi+ntYpUVg0
yqSxHAHgBX7TTHz6Hf8An2fI7rTY/c/xt6D3T1fQ9n7L2H1luPsSv6/TeW49iYaDMbwy9Ls7
YtVU12znp8lHJtXCbFyy09IkkIMeVV39ARXQeyew7hd7TzBvdtuTW13NcJEJfDSRtMamSQAS
VHe0iVahyhAzWiK1gaSOSQGhrT/L0Rj4jfJ3rjI5PMz9JYDdPSvZGz8DnN7J1fX77ruw9jdi
7P2hja3cW8E2Rm8ti8fuza28Nq7VoKnJNjsjU5amy1DRVBhqaapjipasd8zbHfxxQpvU8V5t
0zrH4wjEUkUkhCJ4iqSjxu5Ca1CFGZaqykslJIiuGyOjnfKnv6DtDrjYHyY67qY07e6J3Ftz
sjAVNLP9vUVNbtaup8o1BPIscxehycMD0tVG6SJLTTOjo6MysCdl2H93Xd/sF4v+6y7R4yDk
AMCKj5g5HoQDx69EpUlCO0inW01sbeGG7C2Ts/f23ZXm2/vja2394YKaVDHLLhtzYmkzWMlk
jPMbyUVahKn6E29423dtLZXVzZzik0UjI3+mUlT/ADHSQggkHiOlT7T9a697917r3v3Xuv/V
3+Pfuvde9+691737r3VJv/Chf5Rf7K1/Km+SGUx9Z9pu3u3G0Hxs2Zap+0klru31rMXu4086
sJ1q6Dqyh3BWQ+IF/LTL+ldTr7qyipHWo9uT5s/yvKr/AITw4P8Al5Y3tnOy/KzD0OI7njx7
dTdjxY5u/ans+Tfeews265dtHbb/AG2zcxWbXirlqFpWhjjkEhX1Nvq1DqrTHVvG4vlAflb/
AMJIt87wyWQSu3h1v0tsjoHfY8z1FXFnelu7Ot9l4epylRKzyT5bcOw6HDZeodiWeTIknk+/
dapR+q3P5oO8N84X/hNb/KG2nhZq2l2JvXcWE/v1LTKyU9dkNubW7IyezsLkJ1/VTTzy1tak
DHTJPj0ktqhUjXWx8R63yvjX090x0V0F1T1J0Dg9v4bpraWxcBjtjU+3oKP+GZbAzYyCoh3H
PU0kax5rJbqE5yFbXya5sjVVMlRK7ySMx91Q8eqff5hH813ePQ3y+6/+AvwI+KuG+UXz639s
nG1ta+4a6i2j131RsGSPIbixdJvHP08uMy2TpKXFUc2VqKJsjhcbj6Sqp5xVy1E4pffutgVF
Scda/X81aP8AmfUvzy/kt70/mKf7Jxg8/lPlVSS9NUHxCXs+DN7WGF7n+MVXu6g7DzO/0qJJ
/sqvI4mfAnHZTILBJJkGd0Z01b6sKUanVhfyI/7i+fgf/wCKpbw/+B/+ZnvXl1ofAege/wCF
M+e7ByP8w3+UR19ieuaTvDbw3lJuXanQm5914jYew+5Oy63tzrvFR7D3FuzdEkO0sfBuiGgx
uKnnrmEdJSZGQFo1qWLe68vBs9Cn/MmxX87b+ZH8St6/FjfP8lLYuxkz+W2fuPaPYtD85/i7
ufM9dbl2huTG5eLNYDE1+7MZA0+VwUFbhanTUQMcdlKhVYEj3vrwoDXV0VD+cp132x1F8Zv+
E4nVfesEtH3F1xuPY+yOxsfUZrGbjqsXurbGG+N+GymIrNw4WsyWIztbiJaP7aaspaqrp6mW
JpI551YSvr162v4utjH+cdjP5gWV6L2BT/AWo3jBnhv8v2eOs8nSYjsiTbgxcn8BGFraiopK
tcEuZ1HJLRSrVMft7hqb7ke5R9qpOSI95vW51WIw+B+j4oLRaq92oUI1afg1Cnxfi09L9rNk
Jn+tpTTivCvn+fp0HHywpfkFRfyB/kfSfKmoap+QEHwa7ZTsmWafG1Vd/FBtjPHHR5mqw4/h
lVuGDAGkTJSwM6SV6zMHkvrYI82tsb8y7w3La02QzHwuIFKCukNkKW1FAeC0wOHSW6MBupTb
f2GrH+r0rw6h/wDCa7/typ8L/wDy4v8A+Cx729hzphviPVbv86NlT+fJ/IzLsqg772ooLEKC
z95YVEW5sNTuwAH5Jt7969bX4W6PH/wqDqIIf5O3fEc00UUlVv8A6Fp6VJJER6mde3tp1TQ0
6swaaVaamkkKrciONmtZSR7rS/EOtfr+Ydj6LtbEf8JWvjV3RXVuO+Mm/fjv8PqfftE0tXis
blZd3Yf417I3xHkclFUUyQVNBs2SGljqifJh4spPMDaZg2+rD8Z6378Ds3aO1tq4bY22tr7e
2/srbmJx2B2/tHDYfH43bWEwmIhhp8ViMVg6OnhxtBjcbBTxpBBFGscSooUAAe9dN9aHXRnb
fzEwX/ChL+aP2d8b/hjtv5x9+bVzO+9o4jCb2706+6OyXVvVWH3VtTYdDura9X2XXUcGV8u2
MVhsJ5KH96DH1hXUIatg++nDTSM9G/7h+M/84r5s/wAz3+XJ8w+xf5am0/iFR/Gft7ryLtzs
bbfyt+O/ZmS3J03S9mbez+6KfNQYHelPuaugwmz2z1JTUdJQ109QuYljXQCLe60CACK9MmxO
mth9vf8ACwj5N1m/cDi9xw9Q9R9e9wbVx2Zx9PlMfDvfBfFT4ybY25mjSVYemGQ23NvB8jQT
lHekyFLBPFomjjkT3Xq0QdLP/hUTtPb2M+RX8n7tLGYumxvYR+R+X2xJu/Hr9lnZ8Hid7dHZ
7C4+oyNN46meLBZirqKih1MTSS1U7RaTNJq169eTg3Si/mnf9xJ/8nH/AMMXB/8AvxO5PfvL
rw+A9beHv3VOtbD/AIVUUuQn/lhYmpow/wBrjPlD0/W5llqI4VTHSYPsLHQtLG8sbVSHMV9I
ojQOwYh9OlCyyH7Ysq80R6vOF6fbVT/gB6WWJpMf9Kf8nWpt/KJ+D2T+f3ym2N1TPFX03WWB
j/v13Tn6FZo2xPXuCqKf7vGU9dHYUeZ3jkJoMTROCXhkqmqArpTyAZJc280x8p8s3G5gqb5/
04VPnIwwaeaoKs3qBpqKjo7mult4GkH9pwA+f+xx63y/5inz+6F/lrdDYXb8WA2zuDsKu25R
YDonoOijpoMauP21DBjsJms5QUwvt3rfZ70EKK6KklTLTrS0dnWSWnxu5G5K3jnzeJZfGkSx
Vy1xcGpNWyyqT8Ur1PyAOpsUDFNhYzbhMTqIjB7m/wA3qT/xfVIX8k7f/dvzz/mM9nfLb5D7
trd7ZbpzqbMptpHX7bb2xcz2Xk121gtvbLwSMaHbeBg2fHuCMRwjzTuTLUSTTyTSvLvuzZbR
yZyJt/LWx2whju7ldfm0ixDWzSNxZtfh8cDgoAAAON4ihsrGO2gSmps+ppk1Pma06qk/nAdx
S9t/zHPk9lxVmbHbM3unU2Kg9Xix8XVOKx+xctSwh7sBNufC19TJyQZqhytlKgSX7X7cu1ch
8vRaaPLD4zfPxmMik/7RlA+QHTlpAEsYKjJWv7c/4D0g/wCXjg63N9+Zffk0lTR7M6F6V717
t7GzcAUjDbX2h1Xumlols5VZ6jcG7svjMVBACJJ565UTk+1XPN6kWyx2S0N1eXdvBEv8TvMh
P5Kis5PABa9F94gC0pkkAft6RfT3yWr8ThsntHK1ryUFfQ1FHJHK90eKaFo7EMSPz7d3XbIp
5EnRRrU16ThaN19BX+XvUZKq+B/wznyyRpVv8XeidJjlM/noU6z20mKrJJWklZqiuxawzS3N
xLIwIFrDCnnIRrzdzOIz2/vC4/b4rVH2A1A+XRbL/aSf6Y/4ejgew11Tr3v3Xuve/de6/9bf
49+691737r3XvfuvdUE/zqv5WnyO/mj75+FuzNo726d2p8Xum+zKrfHyCxG89z77xfYW76bM
5famKrl2JiNudd7lwNZksD15RZqKgNblsWs1dlikjxRr5h7qykCvr1YTUfyuP5Z1TBNTyfy8
fg6sdRFJC7U/xR6JpZ1SVGRmhqqXYkNTTShW9MkbpIhsVIIB9+61U+vVAnQP8hf5tdHfy+f5
n/wAh7a+OWb2L8p957I3b8XcjNvbtJF2nVYDfONm3NW9vUx6ZvhclufrvZu3YycN/G1TIUDx
EtCUnHurahUHqzrDfyhNtdsfye+ov5ZPyuy+Cqty9fddYTGRdkdW1NfmaHZPam2MhmK/be/d
i1O6cFtjI5OCgGVemqYKuioWrsdVVdKTEJvKvutau6o6rD6V/l+/8KWviXsCH4s9CfOP4k7j
6FwEX93OsexOx6LL5rfXW2zfJUQ0lPh6LcnTG68pQNi6Nl+1xNTV7ix+NRYqajmSBNK+63VT
mnRh/mZ/KF+dJ+ZHTf8AMn/l/fJDqzb/AMuNsdO7W6y7uwvfFHmX2F2rV7a2RSbHyedx1Ti9
q7sZo96YSEQVeOrKWj8E9NBW01fT1HMfutAilDw6LT3V/LG/ni/Pv5N/CDtr5q7t+AG2Nn/D
PuPH9mYmPqXM9sY7L5bE5LffVW597Y6LH1OzNzRZfN1dF1ZRJQLPVY2libya5fWLb62CoBAr
0cL+an/Kh+Y/dHzf+Pn8yv8Al090dYdZ/JrpTYp68yWA7fXJptzMYiI72oYsthauHau+cdUV
eY2x2TlcPlMdWUVLTSUIjlhqEqdWvXWgRQg8OuPfP8o/5YfzMPg3gdjfzIO5eptsfOnq3tzd
HYvx77v+PWGr6vZ/W+28thtpUa7EzuKlwnX+RyFFuPI7feoyMtJK1VS1dNjqqCsmWmeik916
oBwMdBrtHpj/AIVR7H2jTdRU/wAlf5em86bGQRYPFd974p965jsFMZBTpBR5fJE9SRUOay0E
Cr5ZsjgchVVFQGaaSoJMre69VfTox383P+Vx8mP5gc/8uqv677G6fgzXxP7OG/O4c32hlN07
Wn3oZn6tmyNds3HbE633VjXymRq9lV0zU06YmkiM8SowQsI/deUgV6v19+6r0Ur559Dbv+Uf
wv8AlB8c+v8AJbbw+9u6+k9/9bbVym8azKY/auPzm6sBV4rHVe4K7C4fcGXpcTDUVAaaSmoa
uZUBKxOeD7rYwQetYr41/wArX/hTf8QelNl/Hb47fzDfgx150515/eP+5+z/AOBU+7f4R/e3
dme3zuD/AH8G+fgbubdOQ/iG6dzV1V/lVdP4vP449EKRxpvqxKnJHRs/lT/Jn+d/zm+Inxcz
fyO+W/WsP8074kdh7+3z198g9i4mXEdTZ+iz+9MZuHau3si2z+q+rsntnIbYh2lgamjzNHta
pmoa2inX7ar+7eoTXWgwBNBjotPbP8pf+ez/ADIIesej/wCZN8yPjlgvi3sfdGF3XvKm6SxX
m332BlcLFXUcOUfE4rqzYeHyGdfHVtRDC9bX0mJx8tStYuNqp4hH731uqitBnq47+Zn/ACee
jP5hvxJ66+OFNkpOo9zfHrG46m+NPY9DQy5+TrqPEbdxm2BtfK4+avoqrN7M3DhcFQQZCFam
Kp8tDS1KSGSnCvrrQahr1Xn0H8Xf+FN+xsl1n1F2D81viPmuk9mby2iu5e0quhrN2dubh6s2
1mMVNk9rw19f0nistmcluHb9DJSNVZCWny7NO7PlkfTOPdeqvp0LP8xD+TJ33v35h4f+ZT/L
L+ReA+MPzJhxdPiewMTvahqper+14aHD0e3oMhlp8dgN2pSVtftzHw0GToK/B5fGZb7elqLU
VVTvUVHuvBhSh4dDX8Qdrfz/AGr+RPVmW+bXZPwYxfxk2rBug9o7X6Pxu55+yOzJ6vr7deI2
glJVZ/ZNXQ4v+F9g1eKytc1JX4QPBSNHGssbvTv7rx00NK16gdVfyxO+9jfz3Pkt/M+y27uo
ajoPuXpPA9bbY2jjs/vObt+gzmL6v6A2VUVeewFTsGk2ZS4l8r1VkZEkp9wVUxp5qdjEHeWO
H3Xq9oHXv5z38sTvv+Yru/4H5/pPd3UO16P4vd2Zjsnf8Xaef3nhKnL4PIZPrWthpNnptLYO
94shlli2bVBo618fCGkitKQzlPdeUgV6S/8AOV/lS/Ir5n9pfFf5e/CfuTZPUXyy+JGSrKna
Z7JhrY9pblx0edx27MC8WWoNt7xWhzO383RVSrR1+KrMXlock8VS9NHGzS+68rUqDw6P7/Lx
wf8AMbwvWG9G/mVby6I3b2tkN9PUbFj6Cpaym27g+v4sDiKVKDNms2rtlps7U7ihram6NWKK
aWIeRCDEnutGnl0if5x3xWy/zJ/lt/KTpPamLkzPYEmxU7C6zxtKjvkcn2B1Tlsf2Lt3AYlU
4OT3fLtx8NFrshORIYqCWB/yxuQ2nfduvXakQfSx/osNJJ+QrX8unrd/DmRjwrQ/njr5938p
L+ax25/Li3tunP7D23tXfeyuycficZ2X15uaCPGVGck2yMs22Mhid5UlDU7g23lNv1WaqwqA
VNBPDVTLNSvL4J6fKDeOVdr57261tr25khuYSTHIudOqmoFCQrBgB6MCBRqVBNp4EmAVyRTg
epfyf+UvY/zL7y3z8gu08iKjce9MgXocNTVFTLhdm7ZpC0W3tmbahqZJHpMFt+gtHGP1zzGS
omL1E8sjypyzy/t3LWy2Wz7alLeJcsQNTufid6cWY5+QooooAAhsFiigSKLgP5n1+3rbK/4T
C7GgoPjt8l+zFhAqN2907e2JLUeKINLD13sai3BTwmcSGaQQSdoysEZVVPISpJZguN/3hLwy
b7sG317YrRpKfOWQqcfPwR+zoj5ik1T28fohP7TT/J1qpUPXPdvzD+TW9ttdO7A3N2R2V2T2
PvPcZweExlTFNTNnd1VtdkMtn58oKOm2rhaGpyINXWZSWkpqIN+/JHz7yBl3LauWdgtJtyvI
4LGCCNdTEZ0oAAoFS7GmFQEt5A9G8zxW8Cl2ARVA/l/P8ujbfO/pXOfyu6Kj+MO0PkRT7y7C
+Q/RO3D8yNibdpo/4LtTJ0G/6PfGysFicytPT5CLH1lDQRIIqhkrqqkSeqnhiostS08QM5Z3
yHntm3652bwrOyu2+ikY9zgxlHZlrSoJOR2g0UEtGzElSQXVZTHRVbtPr5f6v9jqrPpfYW++
/O6usekOsaWXIb67S3phNl7fhjjlligqczWR00+Ur/ArvFiMJRmWsrZraYKSnkkYhVJAq3Xd
7barC83G8eltDGXb5gCtB8ycAeZIHVJKKC7cB19YzYuz8R17snZ2wdvpJHgdj7V29s/CRylD
KmI21iKTC41JDGkcZkWiokB0qq3+gA494GXdzJeXVzeTH9aWRnb7WJY/zPRISSSTxPSq9p+t
de9+691737r3X//X3+Pfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfu
vde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9
+691737r3Xvfuvde9+691737r3XvfuvdfOm/4UIfyft3/Dbundnzd+OW1Mrmvir3HuOv3V2j
g9v42srY/j32buHJGqzk2Rgo4pYsX1TvjM1xqMZVEpS4vIVEmNZYIv4cJ579u+dRJHDtd7LS
8jACEn+0UcP9so4jiQK5zQ3tZxKojc/qDh8x/n6oS2b2DDPBETOCpUcFhyDb+p495FbZvKSI
oZujCKZomBB620P5Dn82v4p/EHp/tvor5N7wyfW2OznZUnbOzt6Q7P3ZvPDV9VmtqbW2jntu
5Ck2Jg9ybix9fTR7MoqimkajelmSWYNLE6Ks0P8Au9yPvfNW5bbvOwwLO6W/gumtEYBXd1YF
2VSDrYEVqKDBBwxuUEl68c0IBIWhFQPMmuft6HH5Sf8ACgT4kfHvae79hfyu+kNoUW8N6Vdf
kc13AeqsR1dsGiy+SlqHn3PSbIXFYXdG/t0o8rvEc3SY+kp3ZGdayMPTew/s3tVv+7T213z3
u0jW8QAWHxTLIQKdperLGnroLE5+E0bpmOwmlKteSkgeVan9vkPs609Oze2d3dk7w3Lvzfm5
szvPfO88zW53cu5c7Wz5LNZ3NZOczVVbW1czPLPUTyvwBwosqgKABOSfS7fbRWdpEsdrGoVV
UUVVHAAenS5gqgKooo63Wf8AhO7/ACldydE4+n+d3yU23X4Dt7eu3K7G9G9dZ2lWnynXWw9y
0scWS37uKhqYfvsRvfe2Id6SkpWMUuPwlTOtQpmrWipcb/dHnlN2c8vbXKGsUcGVwcO68EB4
FEOSfNwKYWpJbycOfDQ9o49bW/uGOkPXvfuvde9+691737r3X//Q3+Pfuvde9+691737r3Xv
fuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvd
e9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3XvfuvdNuaw
uH3Hh8tt7cOJxuewGextdhc5g81Q0uUw+Zw+UpZaHJ4nLYyuinosjjcjRTvDPBMjxTROyOpU
ke7KzIyujEODUEYII4EH168CQajj1qe/Or/hKZ8fu289kex/g12Ufipuivmq6/JdT7jxeV3x
0jla+od5rbckTKxbv6vhkqJWZ4qc5rGwxhYaSgpYwLSVsPuXue2hYtxjNxEODA6XA+eKN+dD
5knpfFfMBSVdQ9fP/Z6oq3r/AMJt/wCcHsKsrqPb/W3UXb9PRPVpS5TrvvDZuLo8rHTLC0El
HB2xL1lkIf4h5SsQqYYCrRt5PGNDPJln7ubEyL480sZ9GRjT/eNXSpb6EGoYj8v81ehC6l/4
TRfzU+w8nSU3YeJ6R6Iw7zIMjlN6dq0G7a+npTFFLNLQ4jquj3zDkKtS5jSKWrpY3lQ3lRCs
h3ee8OxRI3gCaZ6YCppH5lyv+A9WfcYaYqT9n+frZs/l2/8ACev4mfCXO4TtfsrI1Pyk7+wd
TT5HB7v3xgKXC9fbGylOUmp8lsbrEV+cpI85RVKh4cnlq3J1NPLHHNRrRSAkxVzL7kb3v6SW
sP8AitgwoVQksw9GfBofNVCg8DqHRfNeSSjSO1P9Xn1f37jvpJ1737r3Xvfuvde9+691737r
3X//0d/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+
691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917
37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3
Xvfuvdf/0t/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv
de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+
691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+6917
37r3Xvfuvdf/09/j37r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3X
vfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuv
de9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+691737r3Xvfuvde9+
691737r3Xvfuvdf/2Q==
--------------060805000806090004060500--

--------------030009070707060808000204--

From Thierry.Bessis@alcatel-lucent.com  Thu Mar 17 13:21:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 79B213A6AFE for ; Thu, 17 Mar 2011 13:21:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.653
X-Spam-Level: 
X-Spam-Status: No, score=-5.653 tagged_above=-999 required=5 tests=[AWL=-0.256, BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, J_CHICKENPOX_52=0.6, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id YCHbgC6Y+JVC for ; Thu, 17 Mar 2011 13:21:44 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 14C803A6AEB for ; Thu, 17 Mar 2011 13:21:43 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2HKN99V022916 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 17 Mar 2011 15:23:09 -0500 (CDT)
Received: from [135.244.40.104] (tbessis.lra.lucent.com [135.244.40.104]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2HKN6rw016717; Thu, 17 Mar 2011 15:23:07 -0500 (CDT)
Message-ID: <4D826DAA.7020704@alcatel-lucent.com>
Date: Thu, 17 Mar 2011 15:23:06 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Hadriel Kaplan 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com>  <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com>
In-Reply-To: <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com>
Content-Type: multipart/alternative; boundary="------------020202060102030106030500"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 20:21:51 -0000

This is a multi-part message in MIME format.
--------------020202060102030106030500
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: quoted-printable

Hi Kaplan,

Thank you for your comment: that's exactly the way I see it ! :-)

Cordially,

Thierry

>>>   In case every device calculate its own algorithm for weight=
, then weight from one device may not same as other device. Say dev1 take=
s Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =3D=
 30% and another device dev2 calculates Min (CPU, Mem) then weight will b=
e Min (70, 40) =3D 40% even though they are equally loaded. It is very im=
portant to have the common unit to solve this issue
> HK>  By definition if dev1 calculates 30% and dev2 calculates 40%, then=
 they are *not* equally loaded.  Unless you're just saying "well what if =
they forgot to take a resource bottleneck into account". (in which case a=
ll bets are off anyway - it would be a bug)
>
>
>>>
>> TB>>  I think here you imply that the DSP is used as a finite pool of =
bearer channels resources, and your are probably using the DSP to process=
 bearers. May be you are an SBC or so.
>>
>>   Here is the interesting part, dev1 achieve the specific serv=
ice using DSP whereas dev2 achieve the same result using multi core/proce=
ssor architecture or softDSP or ASIC implementation. IMO, weight varies b=
ased on the system architecture in case algorithm is not spelt out correc=
tly or generic units are not defined.
>
> HK>  But that's the point - we can't ever write a spec to specify all p=
ossible inputs into an algorithm for every possible architecture, because=
 each architecture has its own limitations/bottlenecks.  For example you'=
ve mentioned DSP's, and Thierry has mentioned CPU and memory; but there'r=
e also SRTP encryption resources, media recording resources, DTMF interwo=
rking resources, v4-v6 interworking resources, TCP-media resources, and s=
o on.  In some architectures the resources are the same, in some not.  So=
me SIP devices have no such resources, some have more.  We can't possibly=
 enumerate them all in a spec.
>
> HK>  So instead, the spec will just say: "whatever your limiting factor=
 is at any given time, use it for your numbers".  Only the developers of =
the systems know what their limiting factors are, after all.
>
>
>>
>>>   I'm proposing is to use single real time value with units l=
ike number of dialog and exchange the threshold (max dialog) using some o=
ther static interface.
>>   dialog is only one of the unit. we will able to work out oth=
er unit which acceptable to everybody. As Paul Jones mentioned in another=
 mail thread, SCE is working well in terms H.323 network for lot of years=
 already but it is not working in SIP entity like SBC.
> HK>  It worked in H.323 because everything was a call and consumed equa=
l resources, and you were treating it like a PRI with a fixed/configured =
number of channels, and things rarely changed.  Arguably that's frequentl=
y also true for SIP across SIP-trunks.  Unfortunately, that's the *only* =
case it works well in, afaict.  Inside the service provider and peering b=
etween them things aren't that "clean" (or restricted, depending on how y=
ou view the world).  It's much harder to have fixed limits of dialogs/cha=
nnels/calls between a big mesh network of systems, where the members of t=
he mesh (the number of discrete systems/blades) grow and shrink relativel=
y frequently.  And it's gotten even worse with SUBSCRIBE dialogs and othe=
r non-call/non-register traffic consuming resources.  That's not to say i=
t isn't possible to do, but it ends up wasting resources because you can'=
t take advantage of statistical multiplexing, and you have to re-configur=
e lot of systems frequently.
>
> -hadriel
>

--=20
--=20
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, =
PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr=

ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 97979=
89
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------020202060102030106030500
Content-Type: multipart/related;
 boundary="------------030402000006080901000302"


--------------030402000006080901000302
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Hi Kaplan, 

    

    Thank you for your comment: that's exactly the way I see it !  :-) 

    

    Cordially, 

    

    Thierry

    

    

      

        

          
<Partha> In case every device calculate its own algorithm for weight, then weight from one device may not same as other device. Say dev1 takes Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) = 30% and another device dev2 calculates Min (CPU, Mem) then weight will be Min (70, 40) = 40% even though they are equally loaded. It is very important to have the common unit to solve this issue </Partha>


        

      

      
HK> By definition if dev1 calculates 30% and dev2 calculates 40%, then they are *not* equally loaded.  Unless you're just saying "well what if they forgot to take a resource bottleneck into account". (in which case all bets are off anyway - it would be a bug)




      

        

          
 


        

        
TB >> I think here you imply that the DSP is used as a finite pool of bearer channels resources, and your are probably using the DSP to process bearers. May be you are an SBC or so.
 
<partha> Here is the interesting part, dev1 achieve the specific service using DSP whereas dev2 achieve the same result using multi core/processor architecture or softDSP or ASIC implementation. IMO, weight varies based on the system architecture in case algorithm is not spelt out correctly or generic units are not defined. </Partha>


      

      

HK> But that's the point - we can't ever write a spec to specify all possible inputs into an algorithm for every possible architecture, because each architecture has its own limitations/bottlenecks.  For example you've mentioned DSP's, and Thierry has mentioned CPU and memory; but there're also SRTP encryption resources, media recording resources, DTMF interworking resources, v4-v6 interworking resources, TCP-media resources, and so on.  In some architectures the resources are the same, in some not.  Some SIP devices have no such resources, some have more.  We can't possibly enumerate them all in a spec.

HK> So instead, the spec will just say: "whatever your limiting factor is at any given time, use it for your numbers".  Only the developers of the systems know what their limiting factors are, after all.




      

        
 


        

          
<Partha> I'm proposing is to use single real time value with units like number of dialog and exchange the threshold (max dialog) using some other static interface. 


        

        
<Partha> dialog is only one of the unit. we will able to work out other unit which acceptable to everybody. As Paul Jones mentioned in another mail thread, SCE is working well in terms H.323 network for lot of years already but it is not working in SIP entity like SBC.   </Partha>


      

      
HK> It worked in H.323 because everything was a call and consumed equal resources, and you were treating it like a PRI with a fixed/configured number of channels, and things rarely changed.  Arguably that's frequently also true for SIP across SIP-trunks.  Unfortunately, that's the *only* case it works well in, afaict.  Inside the service provider and peering between them things aren't that "clean" (or restricted, depending on how you view the world).  It's much harder to have fixed limits of dialogs/channels/calls between a big mesh network of systems, where the members of the mesh (the number of discrete systems/blades) grow and shrink relatively frequently.  And it's gotten even worse with SUBSCRIBE dialogs and other non-call/non-register traffic consuming resources.  That's not to say it isn't possible to do, but it ends up wasting resources because you can't take advantage of statistical multiplexing, and you have to re-configure lot of systems frequently.

-hadriel



    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------030402000006080901000302
Content-Type: image/jpeg;
 name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Disposition: inline;
 filename="al_corp_h_3d_16mm.jpg"
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--------------030402000006080901000302--

--------------020202060102030106030500--

From partr@cisco.com  Thu Mar 17 15:05:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 02A683A6A58 for ; Thu, 17 Mar 2011 15:05:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -8.206
X-Spam-Level: 
X-Spam-Status: No, score=-8.206 tagged_above=-999 required=5 tests=[AWL=-1.205, BAYES_00=-2.599, EXTRA_MPART_TYPE=1, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_33=0.6, J_CHICKENPOX_52=0.6, MIME_QP_LONG_LINE=1.396, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5lWEqbA+YvAO for ; Thu, 17 Mar 2011 15:05:49 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id A5BB23A69E5 for ; Thu, 17 Mar 2011 15:05:46 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=60939; q=dns/txt; s=iport; t=1300399635; x=1301609235; h=mime-version:subject:date:message-id:references:from:to: cc; bh=oVzPqoXaB59b2JDbFf0Yxjno0xH/MUbJalrsHQE1+pQ=; b=aie9DunyB9i+ooNygm95pvUtykqyclQYZURAsEOmlcsF5BDchCOkX1NB A3DpfnRKR59PNNM1TJ5JPU1XfLitgF9px3uAV+5uDi+Dc2956q/B7xsEt XisEJrBk2TBC+iURl20KXIn3St1Onvt4ohZ9ol+YroT9PeVbIXjcphVhz U=;
X-Files: al_corp_h_3d_16mm.jpg : 32899
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AjIKAC4igk2Q/khMgWdsb2JhbACCC6J6UBQBARYmJacInCuDDgeCTgSBdYM4iwQ
X-IronPort-AV: E=Sophos;i="4.63,201,1299456000";  d="jpg'145?scan'145,208,217,145";a="79698628"
Received: from ams-core-3.cisco.com ([144.254.72.76]) by ams-iport-1.cisco.com with ESMTP; 17 Mar 2011 22:07:14 +0000
Received: from xbh-bgl-411.cisco.com (xbh-bgl-411.cisco.com [72.163.129.201]) by ams-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p2HM7CTB029124; Thu, 17 Mar 2011 22:07:13 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-411.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 18 Mar 2011 03:37:12 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/related; boundary="----_=_NextPart_001_01CBE4EF.AA97EFBC"; type="multipart/alternative"
Date: Fri, 18 Mar 2011 03:37:12 +0530
Message-ID: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvk4S2+aGwW2VO3SL+oU0y6/QUlKgADb/Q9
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com>  <1D1713B6-18D6-4278-A387-9BB0E784C1DF@acmepacket.com> <4D826DAA.7020704@alcatel-lucent.com>
From: "Parthasarathi R (partr)" 
To: "Thierry Bessis" , "Hadriel Kaplan" 
X-OriginalArrivalTime: 17 Mar 2011 22:07:12.0920 (UTC) FILETIME=[AB1B4980:01CBE4EF]
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 22:05:51 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CBE4EF.AA97EFBC
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_002_01CBE4EF.AA97EFBC"


------_=_NextPart_002_01CBE4EF.AA97EFBC
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thierry/Hadriel,

My understanding about Thierry proposal from the mail thread is that =
(downstream) system load is calculated in terms of percentage, =
percentage is the unit and passed across to load balancer. The =
normalization to weight is a optional process.=20

Please read inline.

Thanks

Partha=20


________________________________

From: Thierry Bessis [mailto:Thierry.Bessis@alcatel-lucent.com]
Sent: Fri 3/18/2011 1:53 AM
To: Hadriel Kaplan
Cc: Parthasarathi R (partr); dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing


Hi Kaplan,=20

Thank you for your comment: that's exactly the way I see it ! :-)=20

Cordially,=20

Thierry



			 In case every device calculate its own algorithm for weight, =
then weight from one device may not same as other device. Say dev1 takes =
Min(CPU, Mem, DSP) then weight will be calculated as Min (70, 40, 30) =
=3D 30% and another device dev2 calculates Min (CPU, Mem) then weight =
will be Min (70, 40) =3D 40% even though they are equally loaded. It is =
very important to have the common unit to solve this issue 

	HK> By definition if dev1 calculates 30% and dev2 calculates 40%, then =
they are *not* equally loaded.  Unless you're just saying "well what if =
they forgot to take a resource bottleneck into account". (in which case =
all bets are off anyway - it would be a bug)
	 It is not required to be bug in the system but may be =
misinterpretation of the unit termed as "weight" between two systems. =
I'll explain with real time example:

	In case Person X report his weight as 150 then I'll assume that person =
X is overweight(may be looks like giant) because IMO, weight above 100 =
for a person is high. Actually, Person X is not overweight in case he =
mentioned his weight in terms of pounds(lb) but I assume with my =
standard metric of Kilogram(kg) for weight. Hope you agree here that =
unit has to be spelt out explicitly or else, it possibly leads to =
misinterpretation.=20

	This may happen in the system implementation as well because of the =
multiple resource usage. Each implementator will come up with his own =
unit under the name of weight which confuses load balancer in case =
metric is not spelt out properly. =20

		TB >> I think here you imply that the DSP is used as a finite pool of =
bearer channels resources, and your are probably using the DSP to =
process bearers. May be you are an SBC or so.
		=20
		 Here is the interesting part, dev1 achieve the specific =
service using DSP whereas dev2 achieve the same result using multi =
core/processor architecture or softDSP or ASIC implementation. IMO, =
weight varies based on the system architecture in case algorithm is not =
spelt out correctly or generic units are not defined. 

	HK> But that's the point - we can't ever write a spec to specify all =
possible inputs into an algorithm for every possible architecture, =
because each architecture has its own limitations/bottlenecks.  For =
example you've mentioned DSP's, and Thierry has mentioned CPU and =
memory; but there're also SRTP encryption resources, media recording =
resources, DTMF interworking resources, v4-v6 interworking resources, =
TCP-media resources, and so on.  In some architectures the resources are =
the same, in some not.  Some SIP devices have no such resources, some =
have more.  We can't possibly enumerate them all in a spec.
=09
	HK> So instead, the spec will just say: "whatever your limiting factor =
is at any given time, use it for your numbers".  Only the developers of =
the systems know what their limiting factors are, after all.
=09
=09

		=20

			 I'm proposing is to use single real time value with units =
like number of dialog and exchange the threshold (max dialog) using some =
other static interface.=20

		 dialog is only one of the unit. we will able to work out =
other unit which acceptable to everybody. As Paul Jones mentioned in =
another mail thread, SCE is working well in terms H.323 network for lot =
of years already but it is not working in SIP entity like SBC.   =


	HK> It worked in H.323 because everything was a call and consumed equal =
resources, and you were treating it like a PRI with a fixed/configured =
number of channels, and things rarely changed.  Arguably that's =
frequently also true for SIP across SIP-trunks.  Unfortunately, that's =
the *only* case it works well in, afaict.  Inside the service provider =
and peering between them things aren't that "clean" (or restricted, =
depending on how you view the world).  It's much harder to have fixed =
limits of dialogs/channels/calls between a big mesh network of systems, =
where the members of the mesh (the number of discrete systems/blades) =
grow and shrink relatively frequently.  And it's gotten even worse with =
SUBSCRIBE dialogs and other non-call/non-register traffic consuming =
resources.  That's not to say it isn't possible to do, but it ends up =
wasting resources because you can't take advantage of statistical =
multiplexing, and you have to re-configure lot of systems frequently.
=09
	-hadriel
=09


--=20
--=20
Cordially,=20
Thierry Bessis=20
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, =
PdM
=20
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: =
9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None=20

------_=_NextPart_002_01CBE4EF.AA97EFBC
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thierry's signature=0A=
=0A=
=0A=
=0A=

=0A=

Thierry/Hadriel,
=0A=

=0A=

My understanding about Thierry =
proposal from the mail thread is that (downstream) system load is =
calculated in terms of percentage, percentage is the unit and passed =
across to load balancer. The normalization to weight is a optional =
process. 
=0A=

Please read inline.
=0A=

Thanks
=0A=

Partha 
=0A=


=0A=

=0A=
From: Thierry Bessis =
[mailto:Thierry.Bessis@alcatel-lucent.com]
Sent: Fri 3/18/2011 =
1:53 AM
To: Hadriel Kaplan
Cc: Parthasarathi R =
(partr); dispatch@ietf.org
Subject: Re: [dispatch] SIP load =
balancing

=0A=

Hi Kaplan, 

Thank you for your comment: that's exactly the =
way I see it ! :-) =


Cordially, 

Thierry

=0A=

=0A=

=0A=

<Partha> In case every device =
calculate its own algorithm for weight, then weight from one device may =
not same as other device. Say dev1 takes Min(CPU, Mem, DSP) then weight =
will be calculated as Min (70, 40, 30) =3D 30% and another device dev2 =
calculates Min (CPU, Mem) then weight will be Min (70, 40) =3D 40% even =
though they are equally loaded. It is very important to have the common =
unit to solve this issue </Partha>=0A=

HK> By definition if dev1 =
calculates 30% and dev2 calculates 40%, then they are *not* equally =
loaded.  Unless you're just saying "well what if they forgot to take a =
resource bottleneck into account". (in which case all bets are off =
anyway - it would be a bug)
<partha> =
It is not required to be bug in the system but may be misinterpretation =
of the unit termed as "weight" between two systems. I'll explain with =
real time example:
In case Person X report his weight as 150 then =
I'll assume that person X is overweight(may be looks like giant) because =
IMO, weight above 100 for a person is high. Actually, Person X is not =
overweight in case he mentioned his weight in terms of pounds(lb) but I =
assume with my standard metric of Kilogram(kg) for weight. Hope you =
agree here that unit has to be spelt out explicitly or else, it possibly =
leads to misinterpretation. 
This may happen in the system =
implementation as well because of the multiple resource usage. Each =
implementator will come up with his own unit under the name of weight =
which confuses load balancer in case metric is not spelt out properly. =
</Partha> =0A=

=0A=

TB >> I think here you imply that =
the DSP is used as a finite pool of bearer channels resources, and your =
are probably using the DSP to process bearers. May be you are an SBC or =
so.=0A=
 =0A=
<partha> Here is the interesting part, dev1 achieve the specific =
service using DSP whereas dev2 achieve the same result using multi =
core/processor architecture or softDSP or ASIC implementation. IMO, =
weight varies based on the system architecture in case algorithm is not =
spelt out correctly or generic units are not defined. </Partha>=0A=

HK> But that's the point - we can't ever =
write a spec to specify all possible inputs into an algorithm for every =
possible architecture, because each architecture has its own =
limitations/bottlenecks.  For example you've mentioned DSP's, and =
Thierry has mentioned CPU and memory; but there're also SRTP encryption =
resources, media recording resources, DTMF interworking resources, v4-v6 =
interworking resources, TCP-media resources, and so on.  In some =
architectures the resources are the same, in some not.  Some SIP devices =
have no such resources, some have more.  We can't possibly enumerate =
them all in a spec.=0A=
=0A=
HK> So instead, the spec will just say: "whatever your limiting =
factor is at any given time, use it for your numbers".  Only the =
developers of the systems know what their limiting factors are, after =
all.=0A=
=0A=
=0A=

=0A=

 =0A=

=0A=

<Partha> I'm proposing is to use =
single real time value with units like number of dialog and exchange the =
threshold (max dialog) using some other static interface. =0A=

<Partha> dialog is only one of the unit. =
we will able to work out other unit which acceptable to everybody. As =
Paul Jones mentioned in another mail thread, SCE is working well in =
terms H.323 network for lot of years already but it is not working in =
SIP entity like SBC.   </Partha>=0A=

HK> It worked in H.323 because everything was =
a call and consumed equal resources, and you were treating it like a PRI =
with a fixed/configured number of channels, and things rarely changed.  =
Arguably that's frequently also true for SIP across SIP-trunks.  =
Unfortunately, that's the *only* case it works well in, afaict.  Inside =
the service provider and peering between them things aren't that "clean" =
(or restricted, depending on how you view the world).  It's much harder =
to have fixed limits of dialogs/channels/calls between a big mesh =
network of systems, where the members of the mesh (the number of =
discrete systems/blades) grow and shrink relatively frequently.  And =
it's gotten even worse with SUBSCRIBE dialogs and other =
non-call/non-register traffic consuming resources.  That's not to say it =
isn't possible to do, but it ends up wasting resources because you can't =
take advantage of statistical multiplexing, and you have to re-configure =
lot of systems frequently.=0A=
=0A=
-hadriel=0A=
=0A=


=0A=

-- 
-- =

Cordially, 
Thierry =
Bessis =0A=

=0A=

  IMS Solution: =
Architecture & Product Management - ALTA Member
  =
Organization: ALU > Carrier group > Wireline BU =
>  IMS Sol Arch, SE, PdM
3DAlcatel-Lucent
  1960 Lucent Lane - Room =
6L-517
  Naperville, IL  60563  U =
S
  Tel/Fax: +1 630 979 7989
  Corporate IM: =
tbessis

Engage: https://engage.=
alcatel-lucent.com/people/tbessis
My Calendar:   =
www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA =
Hot Line: http://alta.all.alcat=
el-lucent.com/hotline

Conference =
information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - =
Access Code: 9797989
others countries see: http://all.alcatel-lucent.co=
m/audio

Upcoming planned Business Trip: =
12 April - 14 April 2011
Upcoming planned Vacation: None =


------_=_NextPart_002_01CBE4EF.AA97EFBC--

------_=_NextPart_001_01CBE4EF.AA97EFBC
Content-Type: image/jpeg;
	name="al_corp_h_3d_16mm.jpg"
Content-Transfer-Encoding: base64
Content-ID: 
Content-Description: al_corp_h_3d_16mm.jpg
Content-Disposition: inline;
	filename="al_corp_h_3d_16mm.jpg"
Content-Location: al_corp_h_3d_16mm.jpg
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------_=_NextPart_001_01CBE4EF.AA97EFBC--

From spromano@unina.it  Thu Mar 17 16:46:07 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9152D3A6B4F; Thu, 17 Mar 2011 16:46:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -99.322
X-Spam-Level: 
X-Spam-Status: No, score=-99.322 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, HTML_MESSAGE=0.001, MIME_QP_LONG_LINE=1.396, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KF3IB4RoNI8D; Thu, 17 Mar 2011 16:46:06 -0700 (PDT)
Received: from smtp1.unina.it (smtp1.unina.it [192.132.34.61]) by core3.amsl.com (Postfix) with ESMTP id DAA733A6B4D; Thu, 17 Mar 2011 16:46:05 -0700 (PDT)
Received: from [1.98.191.99] ([94.167.55.204]) (authenticated bits=0) by smtp1.unina.it (8.14.4/8.14.4) with ESMTP id p2HNlAYX028798 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Fri, 18 Mar 2011 00:47:14 +0100
References:   <74D7A7C9-A12B-4CBB-8186-A4314A4CEDF5@cisco.com>
In-Reply-To: <74D7A7C9-A12B-4CBB-8186-A4314A4CEDF5@cisco.com>
Mime-Version: 1.0 (iPhone Mail 8A306)
Content-Type: multipart/alternative; boundary=Apple-Mail-5--24150602
Message-Id: 
Content-Transfer-Encoding: 7bit
X-Mailer: iPhone Mail (8A306)
From: Simon Pietro Romano 
Date: Fri, 18 Mar 2011 00:46:24 +0100
To: Gonzalo Salgueiro 
Cc: "vpascual@acmepacket.com" , "sip-overload@ietf.org" , "keith.drage@alcatel-lucent.com" , "HKaplan@acmepacket.com" , "dispatch@ietf.org" , "Thierry.Bessis@alcatel-lucent.com" , "dispatch-bounces@ietf.org" 
Subject: Re: [dispatch] Bar BOF request for SIP Load balancing & (media)	overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 17 Mar 2011 23:46:07 -0000

--Apple-Mail-5--24150602
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii

Dear Partha,

thank you for proposing this interesting bar bof. I'll be glad to participat=
e and I'm also willing to support remote participants in the framework of th=
e IETF Meetecho experiment, if needed.

Simon

Il giorno 17/mar/2011, alle ore 16:28, Gonzalo Salgueiro  ha scritto:

> This is my situation as well. I would love to join but will not be in atte=
ndance.
>=20
> Gonzalo
>=20
> On Mar 17, 2011, at 10:24 AM, Janet P Gunn wrote:
>=20
>>=20
>> Partha,=20
>>=20
>> I am definitely interested in the subject, but I will, unfortunately, not=
 be in Prague.=20
>>=20
>> If it were possible to hold the bar bof somewhere with an audio feed and j=
abber room, I would definitely participate remotely.=20
>>=20
>> Thanks=20
>>=20
>> Janet
>>=20
>> This is a PRIVATE message. If you are not the intended recipient, please d=
elete without copying and kindly advise us by e-mail of the mistake in deliv=
ery.=20
>> NOTE: Regardless of content, this e-mail shall not operate to bind CSC to=
 any order or other contract unless pursuant to explicit written agreement o=
r government initiative expressly permitting the use of e-mail for such purp=
ose.=20
>>=20
>>=20
>> From:	"Parthasarathi R (partr)" 
>> To:	, "Cullen Jennings (fluffy)" ,=
 "Mike Hammer (hmmr)" , , , , , "Paul Kyzivat (pkyzivat)" , , , , , , 
>> Cc:	dispatch@ietf.org, sip-overload@ietf.org
>> Date:	03/16/2011 11:34 PM
>> Subject: 	[dispatch] Bar BOF request for SIP Load balancing & (media)=
        overload handling
>>=20
>>=20
>>=20
>>=20
>> Hi all,=20
>>  =20
>> There is lot of discussion about SIP load balancing and overload handling=
 in Dispatch alias. I wish to have discussion about SIP Load balancing & (me=
dia) overload handling at IETF-80 Prague. I guess that Bar BOF is the idle w=
ay to start the discussion about this topic to common understanding and iden=
tify the way to move forward. Please suggest me in case any other mechanism e=
xists in IETF.=20
>>  =20
>> I have included the folks who has shown the interest in this topic as par=
t of SIP Load balancing & overload handling, dispatch WG chairs, CCed to Dis=
patch and SoC WG alias. AFAIK, the couple of drafts are related to this prob=
lem:=20
>>  =20
>> draft-bessis-dispatch-adaptive-load-balancing-00
>> draft-partha-dispatch-sip-media-overload-control-00.txt
>> draft-jones-sip-overload-sce-00
>> draft-partha-dispatch-resource-availability-00=20
>> draft-sparks-sipping-load-00=20
>>  =20
>> Please let me about others interest and opinion to have Bar BOF for SIP L=
oad balancing & overload handling during IETF-80.=20
>>  =20
>> Thanks=20
>> Partha_______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>>=20
>>=20
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>=20
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

--Apple-Mail-5--24150602
Content-Transfer-Encoding: 7bit
Content-Type: text/html;
	charset=utf-8


Dear Partha,

thank you for proposing this interesting bar bof. I'll be glad to participate and I'm also willing to support remote participants in the framework of the IETF Meetecho experiment, if needed.

Simon

Il giorno 17/mar/2011, alle ore 16:28, Gonzalo Salgueiro <gsalguei@cisco.com> ha scritto:

This is my situation as well. I would love to join but will not be in attendance.

Gonzalo

On Mar 17, 2011, at 10:24 AM, Janet P Gunn wrote:



Partha,



I am definitely interested in the subject,
but I will, unfortunately, not be in Prague.



If it were possible to hold the bar
bof somewhere with an audio feed and jabber room, I would definitely participate
remotely.



Thanks



Janet



This is a PRIVATE message. If you are not the intended recipient, please
delete without copying and kindly advise us by e-mail of the mistake in
delivery. 

NOTE: Regardless of content, this e-mail shall not operate to bind CSC
to any order or other contract unless pursuant to explicit written agreement
or government initiative expressly permitting the use of e-mail for such
purpose.










From:
"Parthasarathi R (partr)"
<partr@cisco.com>


To:
<dworley@avaya.com>, "Cullen
Jennings (fluffy)" <fluffy@cisco.com>, "Mike Hammer (hmmr)"
<hmmr@cisco.com>, <HKaplan@acmepacket.com>, <keith.drage@alcatel-lucent.com>,
<john.elwell@siemens-enterprise.com>, <mary.ietf.barnes@gmail.com>,
"Paul Kyzivat (pkyzivat)" <pkyzivat@cisco.com>, <paulej@packetizer.com>,
<rjsparks@nostrum.com>, <spromano@unina.it>, <Thierry.Bessis@alcatel-lucent.com>,
<vkg@bell-labs.com>, <vpascual@acmepacket.com>


Cc:
dispatch@ietf.org, sip-overload@ietf.org


Date:
03/16/2011 11:34 PM


Subject:
[dispatch] Bar BOF request for SIP Load
balancing & (media)        overload handling










Hi all,

 

There is lot of discussion about SIP load
balancing and overload handling in Dispatch alias. I wish to have discussion
about SIP Load balancing & (media) overload handling at IETF-80 Prague.
I guess that Bar BOF is the idle way to start the discussion about this
topic to common understanding and identify the way to move forward. Please
suggest me in case any other mechanism exists in IETF.

 

I have included the folks who has shown the
interest in this topic as part of SIP Load balancing & overload handling,
dispatch WG chairs, CCed to Dispatch and SoC WG alias. AFAIK, the couple
of drafts are related to this problem:

 

draft-bessis-dispatch-adaptive-load-balancing-00

draft-partha-dispatch-sip-media-overload-control-00.txt

draft-jones-sip-overload-sce-00

draft-partha-dispatch-resource-availability-00 

draft-sparks-sipping-load-00

 

Please let me about others interest and opinion
to have Bar BOF for SIP Load balancing & overload handling during IETF-80.

 

Thanks

Partha_______________________________________________

dispatch mailing list

dispatch@ietf.org

https://www.ietf.org/mailman/listinfo/dispatch





_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch

--Apple-Mail-5--24150602--

From dean.willis@softarmor.com  Fri Mar 18 20:32:41 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BBD073A6A63 for ; Fri, 18 Mar 2011 20:32:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -103.332
X-Spam-Level: 
X-Spam-Status: No, score=-103.332 tagged_above=-999 required=5 tests=[AWL=-0.334, BAYES_00=-2.599, HTML_MESSAGE=0.001, J_CHICKENPOX_44=0.6, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id NCvUKBLWg6XK for ; Fri, 18 Mar 2011 20:32:39 -0700 (PDT)
Received: from mail-yi0-f44.google.com (mail-yi0-f44.google.com [209.85.218.44]) by core3.amsl.com (Postfix) with ESMTP id 16F433A6975 for ; Fri, 18 Mar 2011 20:32:38 -0700 (PDT)
Received: by yic13 with SMTP id 13so2153976yic.31 for ; Fri, 18 Mar 2011 20:34:09 -0700 (PDT)
Received: by 10.90.18.23 with SMTP id 23mr1897658agr.34.1300505648893; Fri, 18 Mar 2011 20:34:08 -0700 (PDT)
Received: from [192.168.89.100] (cpe-66-25-8-214.tx.res.rr.com [66.25.8.214]) by mx.google.com with ESMTPS id c18sm772822ana.27.2011.03.18.20.34.07 (version=TLSv1/SSLv3 cipher=OTHER); Fri, 18 Mar 2011 20:34:08 -0700 (PDT)
References:   <20110316211231.0f54d962@rainpc>
In-Reply-To: <20110316211231.0f54d962@rainpc>
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: multipart/alternative; boundary=Apple-Mail-30-75899506
Message-Id: 
From: Dean Willis 
Date: Fri, 18 Mar 2011 22:34:06 -0500
To: Lorenzo Miniero 
X-Mailer: Apple Mail (2.1082)
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sat, 19 Mar 2011 03:32:42 -0000

--Apple-Mail-30-75899506
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii


On Mar 16, 2011, at 3:12 PM, Lorenzo Miniero wrote:
>=20
> We implemented something like that for our system, also documented in =
a paper lately: basically a controller protocol that tunnels every other =
protocol on top of HTTP to bypass restrictive firewalls. Since it works, =
we were also thinking about a "lightweight" version of such protocol to =
only convey RTP specifically for our web interface, considering RTP is =
the only protocol we couldn't wrap via JavaScript so far. So I agree =
that such a TURN-like protocol would be a good idea for the WG.


I've recently been considering the merits of a VPN that looks like =
MobileIP over HTTPS over Tor. It would let one run any kind of =
communications traffic (albeit with odd time-domain distributions), =
while one would appear to be simply running HTTP traffic to various =
innocuous websites. The problem, of course, is that traffic that is not =
Web surfing has a very different time.size fingerprint, and could easily =
be detected. However, the approach would seem to work very well for =
letting people do actual Web traffic obscurely.=20

If one took the Tor layer out of my idea-space, then one might get =
something very much like your approach. Can you share a link to your =
paper?


--
Dean=

--Apple-Mail-30-75899506
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=us-ascii



We implemented =
something like that for our system, also documented in a paper lately: =
basically a controller protocol that tunnels every other protocol on top =
of HTTP to bypass restrictive firewalls. Since it works, we were also =
thinking about a "lightweight" version of such protocol to only convey =
RTP specifically for our web interface, considering RTP is the only =
protocol we couldn't wrap via JavaScript so far. So I agree that such a =
TURN-like protocol would be a good idea for the =
WG.


I've recently =
been considering the merits of a VPN that looks like MobileIP over HTTPS =
over Tor. It would let one run any kind of communications traffic =
(albeit with odd time-domain distributions), while one would appear to =
be simply running HTTP traffic to various innocuous websites. The =
problem, of course, is that traffic that is not Web surfing has a very =
different time.size fingerprint, and could easily be detected. However, =
the approach would seem to work very well for letting people do actual =
Web traffic obscurely. 

If one took the =
Tor layer out of my idea-space, then one might get something very much =
like your approach. Can you share a link to your =
paper?


--
Dean
=

--Apple-Mail-30-75899506--

From lorenzo@meetecho.com  Sun Mar 20 05:41:59 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D08943A6BB5 for ; Sun, 20 Mar 2011 05:41:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.033
X-Spam-Level: 
X-Spam-Status: No, score=0.033 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, J_CHICKENPOX_44=0.6, SARE_SUB_ENC_UTF8=0.152]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id TNVyUW5IiCDl for ; Sun, 20 Mar 2011 05:41:58 -0700 (PDT)
Received: from smtpw2.aruba.it (smtpa1.aruba.it [62.149.128.210]) by core3.amsl.com (Postfix) with SMTP id 2D4303A6A28 for ; Sun, 20 Mar 2011 05:41:57 -0700 (PDT)
Received: (qmail 30367 invoked by uid 89); 20 Mar 2011 12:43:23 -0000
Received: from unknown (HELO aruba.it) (62.149.158.91) by smtpw2.ad.aruba.it with SMTP; 20 Mar 2011 12:43:23 -0000
Date: Sun, 20 Mar 2011 13:43:23 +0100
Message-Id: 
MIME-Version: 1.0
X-Sensitivity: 3
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
From: "Lorenzo Miniero" 
To: dean.willis@softarmor.com
X-XaM3-API-Version: V3(R2)
X-SenderIP: 78.105.50.229
X-Spam-Rating: smtpw2.ad.aruba.it 1.6.2 0/1000/N
Cc: dispatch@ietf.org
Subject: Re: [dispatch] =?utf-8?q?New_draft_on_RTCWeb_Requirements_and_API_dra?= =?utf-8?q?ft-jennings-rtcweb-api?=
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sun, 20 Mar 2011 12:41:59 -0000

---------- Header Originale -----------=0A=0ADa      : "Dean Willis" dean=
.willis@softarmor.com=0AA          : "Lorenzo Miniero" lorenzo@meetecho.c=
om=0ACc          : "Henry Sinnreich" henry.sinnreich@gmail.com,"DISPATCH =
list" dispatch@ietf.org=0AData      : Fri, 18 Mar 2011 22:34:06 -0500=0AO=
ggetto : Re: [dispatch] New draft on RTCWeb Requirements and API draft-je=
nnings-rtcweb-api=0A=0A> =0A> On Mar 16, 2011, at 3:12 PM, Lorenzo Minier=
o wrote:=0A> > =0A> > We implemented something like that for our system, =
also documented in a paper lately: basically a controller protocol that t=
unnels every other protocol on top of HTTP to bypass restrictive firewall=
s. Since it works, we were also thinking about a "lightweight" version of=
 such protocol to only convey RTP specifically for our web interface, con=
sidering RTP is the only protocol we couldn't wrap via JavaScript so far.=
 So I agree that such a TURN-like protocol would be a good idea for the W=
G.=0A> =0A> =0A> I've recently been considering the merits of a VPN that =
looks like MobileIP over HTTPS over Tor. It would let one run any kind of=
 communications traffic (albeit with odd time-domain distributions), whil=
e one would appear to be simply running HTTP traffic to various innocuous=
 websites. The problem, of course, is that traffic that is not Web surfin=
g has a very different time.size fingerprint, and could easily be detecte=
d. However, the approach would seem to work very well for letting people =
do actual Web traffic obscurely. =0A> =0A> If one took the Tor layer out =
of my idea-space, then one might get something very much like your approa=
ch. Can you share a link to your paper?=0A> =0A=0A=0AHi Dean,=0Ayou can f=
ind the paper here:=0A=0Ahttp://www.springerlink.com/content/610773726011=
1804/=0A=0Alet me know if you've problems getting the paper from there.=0A=
Our approach is less "generic" than the one you've devised, though, since=
 we had a specific use case in mind.=0A=0A=0AL.=0A=0A=0A=0A> =0A> --=0A> =
Dean=0A=0A--=0ALorenzo Miniero=0Ahttp://www.meetecho.com


From henry.sinnreich@gmail.com  Wed Mar 23 09:23:38 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 67D213A67AF for ; Wed, 23 Mar 2011 09:23:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.154
X-Spam-Level: 
X-Spam-Status: No, score=-3.154 tagged_above=-999 required=5 tests=[AWL=0.445,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Y7F07hmy9XoK for ; Wed, 23 Mar 2011 09:23:37 -0700 (PDT)
Received: from mail-gy0-f172.google.com (mail-gy0-f172.google.com [209.85.160.172]) by core3.amsl.com (Postfix) with ESMTP id 1C4BE3A67A4 for ; Wed, 23 Mar 2011 09:23:37 -0700 (PDT)
Received: by gyf3 with SMTP id 3so3381024gyf.31 for ; Wed, 23 Mar 2011 09:25:11 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:cc:message-id :thread-topic:thread-index:mime-version:content-type :content-transfer-encoding; bh=wqQi+/9OHKN3BFAsoW8A2hBsK5tC/SjiM6jTsXJWIn0=; b=nq2lgrgUDWRffaV/0rkdzvdUh2NH+Irk/2MgeEmrTTlRc11QKPrMnZe7IIBfsuF6Oi 8FWW2VkMIWsCORVEAkm41CcGrjiog6Lz7boCY2mFZ/PvSNrJGL4fFAzGV+wtAalOH0W4 J6WbudCNexXvrRbnBsKLxbbAdRiBz5O+y6INs=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:mime-version:content-type:content-transfer-encoding; b=nO1FdEUqDwteSPCgm3O0i0No9P1zPKN+fNV6d/j287R9U6upXPEwGf99VB0dJOW5H/ QtznLHVlfG7IKg/pD6FGLN7X2ktDGuNy70HPpOlizbIfk1+SqV9b5HxAgu7sFMeRW1YN /7fOKt2mnMECrisDirHmV8eAsVM8PYWFxxUBY=
Received: by 10.100.147.14 with SMTP id u14mr4863765and.101.1300897510818; Wed, 23 Mar 2011 09:25:10 -0700 (PDT)
Received: from [10.0.1.5] (cpe-76-184-227-164.tx.res.rr.com [76.184.227.164]) by mx.google.com with ESMTPS id x37sm9991484ana.8.2011.03.23.09.25.08 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 23 Mar 2011 09:25:09 -0700 (PDT)
User-Agent: Microsoft-Entourage/12.28.0.101117
Date: Wed, 23 Mar 2011 11:25:07 -0500
From: Henry Sinnreich 
To: Lorenzo Miniero 
Message-ID: 
Thread-Topic: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
Thread-Index: Acvpdt8uB1+rzAa91E+7mhshPuePIA==
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 23 Mar 2011 16:23:38 -0000

Hi Lorenzo, Cullen,

> We implemented something like that for our system, also documented in a paper
> lately: basically a controller protocol that tunnels every other protocol on
> top of HTTP to bypass restrictive firewalls.

Having explained HTTP tunneling flavors in the paper and also shown
measurement results, besides a working product, should really make IMO your
approach or something similar to it an option for the RTC-Web media
solution. Such an option should also be reflected in the API part, on which
the draft-jennings-rtcweb-api is focused.

RT media over HTTP may be quite effective in some usage scenarios, such as
CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
networks with plenty of bandwidth, smaller delay and arguably without all
the other SIP, etc. intermediaries found inside and between most service
provider networks.

I know profiles in standards are not desirable, but hindering innovations
such as yours and ignoring CDN and other Web application scenarios is even
less desirable. 
This is a challenge both the IETF and the W3C will hopefully manage.

Thanks for sharing,

Henry


On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:

> On Wed, 16 Mar 2011 13:54:39 -0500
> Henry Sinnreich  wrote:
> 
>> Hi Cullen,
>> 
>> Section 3. Requirements says:
>> ... In addition, there is a strong interest to define a TURN-like protocol
>>    that looks like HTTP to intermediaries, so that media can be tunneled
>>    over HTTP. 
>> ...
>> 
>> This is very interesting and innovative.
>> Can you explain here a little bit?
>> 
>> Thanks, Henry
>> 
> 
> 
> We implemented something like that for our system, also documented in a paper
> lately: basically a controller protocol that tunnels every other protocol on
> top of HTTP to bypass restrictive firewalls. Since it works, we were also
> thinking about a "lightweight" version of such protocol to only convey RTP
> specifically for our web interface, considering RTP is the only protocol we
> couldn't wrap via JavaScript so far. So I agree that such a TURN-like protocol
> would be a good idea for the WG.
> 
> Of course, there would be limitations and issues because of this transport,
> but since firewalls are there to stay, and few of them like RTP appearing "out
> of the blue", that's something we probably need to live with.
> 
> L.
> 
> 
>> 
>> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
>> 
>>> 
>>> I wrote up the start of a draft on requirements.
>>> 
>>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>>> 
>>> I think the key thing for the key thing for IETF working is starting to
>>> think
>>> about the requirements and what on the on the wire protocols are that we
>>> want
>>> to select. I included some API definition because I found it helpful to be
>>> thinking about that in this draft but that would not be part of the IETF
>>> work.
>>> That would be work I think W3C should do. The draft also touches on the of
>>> the
>>> security issues for media and signaling originating from browsers.
>>> 
>>> Cullen 
>>> 
>>> 
>>> 
>>> 
>>> _______________________________________________
>>> dispatch mailing list
>>> dispatch@ietf.org
>>> https://www.ietf.org/mailman/listinfo/dispatch
>> 
>> 
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
> 




From gonzalo.camarillo@ericsson.com  Thu Mar 24 08:55:54 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E199028C0DE for ; Thu, 24 Mar 2011 08:55:53 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.601
X-Spam-Level: 
X-Spam-Status: No, score=-106.601 tagged_above=-999 required=5 tests=[AWL=-0.002, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yxwK1NChOyOe for ; Thu, 24 Mar 2011 08:55:53 -0700 (PDT)
Received: from mailgw10.se.ericsson.net (mailgw10.se.ericsson.net [193.180.251.61]) by core3.amsl.com (Postfix) with ESMTP id AD60F28C0CE for ; Thu, 24 Mar 2011 08:55:52 -0700 (PDT)
X-AuditID: c1b4fb3d-b7bbbae000005311-89-4d8b69e6890f
Received: from esessmw0197.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw10.se.ericsson.net (Symantec Mail Security) with SMTP id FF.C1.21265.6E96B8D4; Thu, 24 Mar 2011 16:57:26 +0100 (CET)
Received: from mail.lmf.ericsson.se (153.88.115.8) by esessmw0197.eemea.ericsson.se (153.88.115.88) with Microsoft SMTP Server id 8.3.137.0; Thu, 24 Mar 2011 16:57:26 +0100
Received: from [131.160.126.144] (rvi2-126-144.lmf.ericsson.se [131.160.126.144])	by mail.lmf.ericsson.se (Postfix) with ESMTP id 33562245F; Thu, 24 Mar 2011 17:57:26 +0200 (EET)
Message-ID: <4D8B69E6.8010707@ericsson.com>
Date: Thu, 24 Mar 2011 17:57:26 +0200
From: Gonzalo Camarillo 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: Cullen Jennings 
References: <70BB7A39-E179-4235-BAA7-A61D692BAB03@cisco.com>	 <37088805-C32B-49D1-874E-02E51BCDA890@cisco.com>
In-Reply-To: <37088805-C32B-49D1-874E-02E51BCDA890@cisco.com>
X-Enigmail-Version: 1.1.1
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
X-Brightmail-Tracker: AAAAAA==
Cc: DISPATCH list 
Subject: Re: [dispatch] DISPATCH Ad-HOC meetings about VIPR and RTCWeb at IETF 80
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 24 Mar 2011 15:55:54 -0000

Hi,

> Note the number of WG in RAI alone is going to ensure we have conflicts - I think I mention this every time dispatch HUMS on forming a new WG :-) 

we are not only creating new WGs. We are also closing some of them ;o)

We have already closed AVT and will close ENUM during the meeting in
Prague next week. SPEEMINT will also be closed shortly and we intend to
close a few more by the next IETF meeting.

Cheers,

Gonzalo

From lorenzo@meetecho.com  Thu Mar 24 09:55:30 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4CB2428C139 for ; Thu, 24 Mar 2011 09:55:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.178
X-Spam-Level: 
X-Spam-Status: No, score=-0.178 tagged_above=-999 required=5 tests=[AWL=-0.541, BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, URIBL_RHS_DOB=1.083]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WNG8lNb2YyVB for ; Thu, 24 Mar 2011 09:55:29 -0700 (PDT)
Received: from smtplq02.aruba.it (smtplq-out15.aruba.it [62.149.158.35]) by core3.amsl.com (Postfix) with SMTP id A2F4328C0E1 for ; Thu, 24 Mar 2011 09:55:28 -0700 (PDT)
Received: (qmail 13099 invoked by uid 89); 24 Mar 2011 16:57:00 -0000
Received: from unknown (HELO smtp2.aruba.it) (62.149.158.222) by smtplq02.aruba.it with SMTP; 24 Mar 2011 16:57:00 -0000
Received: (qmail 10547 invoked by uid 89); 24 Mar 2011 16:57:00 -0000
Received: from unknown (HELO lminiero-acer) (lorenzo@meetecho.com@143.225.229.189) by smtp2.ad.aruba.it with SMTP; 24 Mar 2011 16:57:00 -0000
Date: Thu, 24 Mar 2011 17:52:17 +0100
From: Lorenzo Miniero 
To: Henry Sinnreich 
Message-ID: <20110324175217.71812238@lminiero-acer>
In-Reply-To: 
References: 
Organization: Meetecho
X-Mailer: Claws Mail 3.7.8 (GTK+ 2.22.0; i386-redhat-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Spam-Rating: smtp2.ad.aruba.it 1.6.2 0/1000/N
X-Spam-Rating: smtplq02.aruba.it 1.6.2 0/1000/N
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 24 Mar 2011 16:55:30 -0000

Hi Henry,


On Wed, 23 Mar 2011 11:25:07 -0500
Henry Sinnreich  wrote:

> Hi Lorenzo, Cullen,
> 
> > We implemented something like that for our system, also documented in a paper
> > lately: basically a controller protocol that tunnels every other protocol on
> > top of HTTP to bypass restrictive firewalls.
> 
> Having explained HTTP tunneling flavors in the paper and also shown
> measurement results, besides a working product, should really make IMO your
> approach or something similar to it an option for the RTC-Web media
> solution. Such an option should also be reflected in the API part, on which
> the draft-jennings-rtcweb-api is focused.
> 


I'd be glad to contribute to the definition of a related solution, should such a work be started within the IETF.

I'll also try and make available, hopefully before the meeting, a publicly reachable prototype version of our current approach for you all to play with, just to have some live experience to foster discussion.

L.


> RT media over HTTP may be quite effective in some usage scenarios, such as
> CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
> networks with plenty of bandwidth, smaller delay and arguably without all
> the other SIP, etc. intermediaries found inside and between most service
> provider networks.
> 
> I know profiles in standards are not desirable, but hindering innovations
> such as yours and ignoring CDN and other Web application scenarios is even
> less desirable. 
> This is a challenge both the IETF and the W3C will hopefully manage.
> 
> Thanks for sharing,
> 
> Henry
> 
> 
> On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
> 
> > On Wed, 16 Mar 2011 13:54:39 -0500
> > Henry Sinnreich  wrote:
> > 
> >> Hi Cullen,
> >> 
> >> Section 3. Requirements says:
> >> ... In addition, there is a strong interest to define a TURN-like protocol
> >>    that looks like HTTP to intermediaries, so that media can be tunneled
> >>    over HTTP. 
> >> ...
> >> 
> >> This is very interesting and innovative.
> >> Can you explain here a little bit?
> >> 
> >> Thanks, Henry
> >> 
> > 
> > 
> > We implemented something like that for our system, also documented in a paper
> > lately: basically a controller protocol that tunnels every other protocol on
> > top of HTTP to bypass restrictive firewalls. Since it works, we were also
> > thinking about a "lightweight" version of such protocol to only convey RTP
> > specifically for our web interface, considering RTP is the only protocol we
> > couldn't wrap via JavaScript so far. So I agree that such a TURN-like protocol
> > would be a good idea for the WG.
> > 
> > Of course, there would be limitations and issues because of this transport,
> > but since firewalls are there to stay, and few of them like RTP appearing "out
> > of the blue", that's something we probably need to live with.
> > 
> > L.
> > 
> > 
> >> 
> >> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
> >> 
> >>> 
> >>> I wrote up the start of a draft on requirements.
> >>> 
> >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> >>> 
> >>> I think the key thing for the key thing for IETF working is starting to
> >>> think
> >>> about the requirements and what on the on the wire protocols are that we
> >>> want
> >>> to select. I included some API definition because I found it helpful to be
> >>> thinking about that in this draft but that would not be part of the IETF
> >>> work.
> >>> That would be work I think W3C should do. The draft also touches on the of
> >>> the
> >>> security issues for media and signaling originating from browsers.
> >>> 
> >>> Cullen 
> >>> 
> >>> 
> >>> 
> >>> 
> >>> _______________________________________________
> >>> dispatch mailing list
> >>> dispatch@ietf.org
> >>> https://www.ietf.org/mailman/listinfo/dispatch
> >> 
> >> 
> >> _______________________________________________
> >> dispatch mailing list
> >> dispatch@ietf.org
> >> https://www.ietf.org/mailman/listinfo/dispatch
> > 
> 
> 
> 


-- 
Lorenzo Miniero, COB

Meetecho s.r.l.
Web Conferencing and Collaboration Tools
http://www.meetecho.com

From dworley@avaya.com  Thu Mar 24 12:55:42 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0604028C119 for ; Thu, 24 Mar 2011 12:55:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -103.054
X-Spam-Level: 
X-Spam-Status: No, score=-103.054 tagged_above=-999 required=5 tests=[AWL=0.545, BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id opVZRxDu45jd for ; Thu, 24 Mar 2011 12:55:41 -0700 (PDT)
Received: from p-us1-iereast-outbound-tmp.us1.avaya.com (p-us1-iereast-outbound-tmp.us1.avaya.com [135.11.29.16]) by core3.amsl.com (Postfix) with ESMTP id 3C8AA28C114 for ; Thu, 24 Mar 2011 12:55:41 -0700 (PDT)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAHM+i03GmAcF/2dsb2JhbAClSneqBQKaGYVpBJBb
X-IronPort-AV: E=Sophos;i="4.63,238,1299474000"; d="scan'208";a="65162268"
Received: from unknown (HELO co300216-co-erhwest.avaya.com) ([198.152.7.5]) by p-us1-iereast-outbound-tmp.us1.avaya.com with ESMTP; 24 Mar 2011 15:57:15 -0400
X-IronPort-AV: E=Sophos;i="4.63,238,1299474000"; d="scan'208";a="600221940"
Received: from dc-us1hcex2.us1.avaya.com (HELO DC-US1HCEX2.global.avaya.com) ([135.11.52.21]) by co300216-co-erhwest-out.avaya.com with ESMTP; 24 Mar 2011 15:57:15 -0400
Received: from DC-US1MBEX4.global.avaya.com ([169.254.2.201]) by DC-US1HCEX2.global.avaya.com ([::1]) with mapi; Thu, 24 Mar 2011 15:57:14 -0400
From: "Worley, Dale R (Dale)" 
To: Thierry Bessis , "Parthasarathi R (partr)" 
Date: Thu, 24 Mar 2011 15:54:59 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvk4KnUVhbbtaVoRv2Nny4PrKFSiwFfLEYi
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> , <4D826CC5.9010502@alcatel-lucent.com>
In-Reply-To: <4D826CC5.9010502@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 24 Mar 2011 19:55:42 -0000

________________________________________
From: dispatch-bounces@ietf.org [dispatch-bounces@ietf.org] On Behalf Of Th=
ierry Bessis [Thierry.Bessis@alcatel-lucent.com]

TB> It seems to me that this comment just reinforces my point (I hope I und=
erstand well): We need to let the destination come up its own way with its =
percentage of resource left, because there are so many possible implementat=
ions that it would be cumbersome and restrictive to impose any algorithm. A=
lso just to be sure: The proposed weight is a percentage (a ratio): it has =
no unit. Part of the robustness and universality is derived from this chara=
cteristic.
________________________________________

I haven't done an analysis, but I strongly expect that the overall system b=
ehaves better if all systems report their available capacity
in the same amount of traffic ("number of calls") rather than percentage of=
 capacity that is unused.

The reason is that if two servers have significantly different capacity, sa=
y by a factor of 10, when both servers report 1% of capacity
free, one server will be able to accept 10 times as much traffic as the oth=
er.

Dale

From gonzalo.camarillo@ericsson.com  Thu Mar 24 08:44:28 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7D40528C0DE; Thu, 24 Mar 2011 08:44:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.601
X-Spam-Level: 
X-Spam-Status: No, score=-106.601 tagged_above=-999 required=5 tests=[AWL=-0.002, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uO1FC8rZsgVD; Thu, 24 Mar 2011 08:44:27 -0700 (PDT)
Received: from mailgw9.se.ericsson.net (mailgw9.se.ericsson.net [193.180.251.57]) by core3.amsl.com (Postfix) with ESMTP id 1903728C0D0; Thu, 24 Mar 2011 08:44:26 -0700 (PDT)
X-AuditID: c1b4fb39-b7c6dae0000023f2-a7-4d8b673832d9
Received: from esessmw0191.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw9.se.ericsson.net (Symantec Mail Security) with SMTP id 26.58.09202.8376B8D4; Thu, 24 Mar 2011 16:46:00 +0100 (CET)
Received: from mail.lmf.ericsson.se (153.88.115.8) by esessmw0191.eemea.ericsson.se (153.88.115.85) with Microsoft SMTP Server id 8.3.137.0; Thu, 24 Mar 2011 16:46:00 +0100
Received: from [131.160.126.144] (rvi2-126-144.lmf.ericsson.se [131.160.126.144])	by mail.lmf.ericsson.se (Postfix) with ESMTP id 0E5A2245F; Thu, 24 Mar 2011 17:46:00 +0200 (EET)
Message-ID: <4D8B6737.8050200@ericsson.com>
Date: Thu, 24 Mar 2011 17:45:59 +0200
From: Gonzalo Camarillo 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: "Paul E. Jones" 
References: 	 <4D8200AC.5070307@packetizer.com>
In-Reply-To: <4D8200AC.5070307@packetizer.com>
X-Enigmail-Version: 1.1.1
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
X-Brightmail-Tracker: AAAAAA==
X-Mailman-Approved-At: Thu, 24 Mar 2011 17:11:06 -0700
Cc: "vpascual@acmepacket.com" , "sip-overload@ietf.org" , "keith.drage@alcatel-lucent.com" , "HKaplan@acmepacket.com" , "dispatch@ietf.org" , "Thierry.Bessis@alcatel-lucent.com" 
Subject: Re: [dispatch] [sip-overload] Bar BOF request for SIP Load balancing & (media) overload handling
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 24 Mar 2011 15:44:28 -0000

Hi,

thanks for the feedback. We will address that issue in the next revision
of the draft ;o)

Jokes aside, one of the main points of the draft is that the term "bar
BOF" confuses many people (specially because it contains the word "BOF"
and people end up confusing bar BOFs with actual BOFs). So, we recommend
to call them "side meetings" instead.

Cheers,

Gonzalo

On 17/03/2011 2:38 PM, Paul E. Jones wrote:
> This is not entirely helpful.  It does not go into explaining what kind 
> or how much beer Partha is going to be required to buy everyone who 
> attends ;-)
> 
> On 3/16/2011 10:18 PM, Cullen Jennings wrote:
>> Just as FYI ... if you have not seen it, you might want to have a read of
>> http://tools.ietf.org/html/draft-eggert-successful-bar-bof
>>
>>
>> On Mar 16, 2011, at 6:44 PM, Parthasarathi R (partr) wrote:
>>
>>> Hi all,
>>>
>>> There is lot of discussion about SIP load balancing and overload handling in Dispatch alias. I wish to have discussion about SIP Load balancing&  (media) overload handling at IETF-80 Prague. I guess that Bar BOF is the idle way to start the discussion about this topic to common understanding and identify the way to move forward. Please suggest me in case any other mechanism exists in IETF.
>>>
>>> I have included the folks who has shown the interest in this topic as part of SIP Load balancing&  overload handling, dispatch WG chairs, CCed to Dispatch and SoC WG alias. AFAIK, the couple of drafts are related to this problem:
>>>
>>> draft-bessis-dispatch-adaptive-load-balancing-00
>>> draft-partha-dispatch-sip-media-overload-control-00.txt
>>> draft-jones-sip-overload-sce-00
>>> draft-partha-dispatch-resource-availability-00
>>> draft-sparks-sipping-load-00
>>>
>>> Please let me about others interest and opinion to have Bar BOF for SIP Load balancing&  overload handling during IETF-80.
>>>
>>> Thanks
>>> Partha
> _______________________________________________
> sip-overload mailing list
> sip-overload@ietf.org
> https://www.ietf.org/mailman/listinfo/sip-overload
> 


From lorenzo@meetecho.com  Sat Mar 26 09:48:16 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 93FB03A67B2 for ; Sat, 26 Mar 2011 09:48:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.718
X-Spam-Level: 
X-Spam-Status: No, score=-0.718 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, NORMAL_HTTP_TO_IP=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id I0bu7oL48DK1 for ; Sat, 26 Mar 2011 09:48:15 -0700 (PDT)
Received: from smtplq02.aruba.it (smtplq-out15.aruba.it [62.149.158.35]) by core3.amsl.com (Postfix) with SMTP id 972933A67A8 for ; Sat, 26 Mar 2011 09:48:13 -0700 (PDT)
Received: (qmail 27802 invoked by uid 89); 26 Mar 2011 16:49:47 -0000
Received: from unknown (HELO smtp6.aruba.it) (62.149.158.226) by smtplq02.aruba.it with SMTP; 26 Mar 2011 16:49:47 -0000
Received: (qmail 22126 invoked by uid 89); 26 Mar 2011 16:49:47 -0000
Received: from unknown (HELO lminiero-acer) (lorenzo@meetecho.com@130.129.21.22) by smtp6.ad.aruba.it with SMTP; 26 Mar 2011 16:49:47 -0000
Date: Sat, 26 Mar 2011 17:45:04 +0100
From: Lorenzo Miniero 
To: Henry Sinnreich 
Message-ID: <20110326174504.56061fab@lminiero-acer>
In-Reply-To: <20110324175217.71812238@lminiero-acer>
References:  <20110324175217.71812238@lminiero-acer>
Organization: Meetecho
X-Mailer: Claws Mail 3.7.8 (GTK+ 2.22.0; i386-redhat-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Spam-Rating: smtp6.ad.aruba.it 1.6.2 0/1000/N
X-Spam-Rating: smtplq02.aruba.it 1.6.2 0/1000/N
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Sat, 26 Mar 2011 16:48:16 -0000

Dear all,

as anticipated, I've just made available a prototype version of our RTP-over-HTTP tunneling approach:

	http://143.225.229.136/WebLite/echotest.jsp

It is a very simple Echo Test: the page uses JavaScript to setup a server-originated SIP call to a URI associated with an RTP echo application. Then, a Java applet is started to take care of the the RTP tunneling and encoding/decoding process.

To play with it, just enable the "Enable RTP tunneling" checkbox and then click the button to launch the test. If everything goes fine, a new box should appear on the left with the remote audio and video streams: a voice prompt with a video logo will precede the actual echo feature. All the RTP packets will be tunneled over HTTP.

Before you start, nevertheless, please have a quick look at a few prerequisites and instructions here:

	http://ietf.conf.meetecho.com/index.php/Web_Client

for what concerns the applet-based Audio/Video functionality (last section in the page: our Web Client uses the same applet). In fact, applets can give a few problems at times (especially for multimedia-driven applications) in some configurations, and so checking a few things before delving in the demo may spare you from headache! :)

Let us know if and how it works. Of course please beware that this is just a very dumb prototype, so rough that it can give you the creeps if you just skim over it! :)
Nevertheless, the basic concepts are there, so it may be useful to foster some discussion on the matter. If you're interested in some more details, I'm already in Prague.

Lorenzo



On Thu, 24 Mar 2011 17:52:17 +0100
Lorenzo Miniero  wrote:

> Hi Henry,
> 
> 
> On Wed, 23 Mar 2011 11:25:07 -0500
> Henry Sinnreich  wrote:
> 
> > Hi Lorenzo, Cullen,
> > 
> > > We implemented something like that for our system, also documented in a paper
> > > lately: basically a controller protocol that tunnels every other protocol on
> > > top of HTTP to bypass restrictive firewalls.
> > 
> > Having explained HTTP tunneling flavors in the paper and also shown
> > measurement results, besides a working product, should really make IMO your
> > approach or something similar to it an option for the RTC-Web media
> > solution. Such an option should also be reflected in the API part, on which
> > the draft-jennings-rtcweb-api is focused.
> > 
> 
> 
> I'd be glad to contribute to the definition of a related solution, should such a work be started within the IETF.
> 
> I'll also try and make available, hopefully before the meeting, a publicly reachable prototype version of our current approach for you all to play with, just to have some live experience to foster discussion.
> 
> L.
> 
> 
> > RT media over HTTP may be quite effective in some usage scenarios, such as
> > CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
> > networks with plenty of bandwidth, smaller delay and arguably without all
> > the other SIP, etc. intermediaries found inside and between most service
> > provider networks.
> > 
> > I know profiles in standards are not desirable, but hindering innovations
> > such as yours and ignoring CDN and other Web application scenarios is even
> > less desirable. 
> > This is a challenge both the IETF and the W3C will hopefully manage.
> > 
> > Thanks for sharing,
> > 
> > Henry
> > 
> > 
> > On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
> > 
> > > On Wed, 16 Mar 2011 13:54:39 -0500
> > > Henry Sinnreich  wrote:
> > > 
> > >> Hi Cullen,
> > >> 
> > >> Section 3. Requirements says:
> > >> ... In addition, there is a strong interest to define a TURN-like protocol
> > >>    that looks like HTTP to intermediaries, so that media can be tunneled
> > >>    over HTTP. 
> > >> ...
> > >> 
> > >> This is very interesting and innovative.
> > >> Can you explain here a little bit?
> > >> 
> > >> Thanks, Henry
> > >> 
> > > 
> > > 
> > > We implemented something like that for our system, also documented in a paper
> > > lately: basically a controller protocol that tunnels every other protocol on
> > > top of HTTP to bypass restrictive firewalls. Since it works, we were also
> > > thinking about a "lightweight" version of such protocol to only convey RTP
> > > specifically for our web interface, considering RTP is the only protocol we
> > > couldn't wrap via JavaScript so far. So I agree that such a TURN-like protocol
> > > would be a good idea for the WG.
> > > 
> > > Of course, there would be limitations and issues because of this transport,
> > > but since firewalls are there to stay, and few of them like RTP appearing "out
> > > of the blue", that's something we probably need to live with.
> > > 
> > > L.
> > > 
> > > 
> > >> 
> > >> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
> > >> 
> > >>> 
> > >>> I wrote up the start of a draft on requirements.
> > >>> 
> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> > >>> 
> > >>> I think the key thing for the key thing for IETF working is starting to
> > >>> think
> > >>> about the requirements and what on the on the wire protocols are that we
> > >>> want
> > >>> to select. I included some API definition because I found it helpful to be
> > >>> thinking about that in this draft but that would not be part of the IETF
> > >>> work.
> > >>> That would be work I think W3C should do. The draft also touches on the of
> > >>> the
> > >>> security issues for media and signaling originating from browsers.
> > >>> 
> > >>> Cullen 
> > >>> 
> > >>> 
> > >>> 
> > >>> 
> > >>> _______________________________________________
> > >>> dispatch mailing list
> > >>> dispatch@ietf.org
> > >>> https://www.ietf.org/mailman/listinfo/dispatch
> > >> 
> > >> 
> > >> _______________________________________________
> > >> dispatch mailing list
> > >> dispatch@ietf.org
> > >> https://www.ietf.org/mailman/listinfo/dispatch
> > > 
> > 
> > 
> > 
> 
> 
> -- 
> Lorenzo Miniero, COB
> 
> Meetecho s.r.l.
> Web Conferencing and Collaboration Tools
> http://www.meetecho.com
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


-- 
Lorenzo Miniero, COB

Meetecho s.r.l.
Web Conferencing and Collaboration Tools
http://www.meetecho.com

From henry.sinnreich@gmail.com  Sun Mar 27 21:57:54 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3A5F23A6804 for ; Sun, 27 Mar 2011 21:57:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.598
X-Spam-Level: 
X-Spam-Status: No, score=-3.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, NORMAL_HTTP_TO_IP=0.001, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id vNT7M92T9Pe8 for ; Sun, 27 Mar 2011 21:57:51 -0700 (PDT)
Received: from mail-bw0-f44.google.com (mail-bw0-f44.google.com [209.85.214.44]) by core3.amsl.com (Postfix) with ESMTP id 2D7453A67E2 for ; Sun, 27 Mar 2011 21:57:51 -0700 (PDT)
Received: by bwz13 with SMTP id 13so2325965bwz.31 for ; Sun, 27 Mar 2011 21:59:27 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:user-agent:date:subject:from:to:cc:message-id :thread-topic:thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; bh=//M+eS+BtfJmDZCrzMCKIxNOD7Ob83ZevsmL/8aqopA=; b=L2meaUEq78vqQ3ZbBhNsoOW/ofNQXc1yBK6Qzc283SjSeu0PqxaxHsH/usWFNv+m8q Zny7lFODxZSZCGX0fq/W8HRFG0p0w6RWx0WagBE39cyRjB2eBs2NmOR19FCpScBr6/Vi xks1DPqj7vnUaT+Zb2Jaen9T19LHpp/QBRcmM=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=uJy9e/jsy24/mI7iUxKCj7ZmaFoADv2j7Ve79xWU4eVBKtb8GBlOODPjtyapapM8KW dwXuu/HGGO+7vXoDkG2KC2Cc+5ZgOFQ+TbL7gbGkOLuTecpgW+vEc+yMAQdCVYWjJQd5 dp+b3A7jXea2G4vkKSEmhC0PilGNr+zRo8PVM=
Received: by 10.204.75.26 with SMTP id w26mr3138528bkj.154.1301288367556; Sun, 27 Mar 2011 21:59:27 -0700 (PDT)
Received: from [192.168.178.31] (a40-prg1-29-125.static.adsl.vol.cz [88.146.87.125]) by mx.google.com with ESMTPS id q18sm2348585bka.15.2011.03.27.21.59.25 (version=TLSv1/SSLv3 cipher=OTHER); Sun, 27 Mar 2011 21:59:26 -0700 (PDT)
User-Agent: Microsoft-Entourage/12.27.0.100913
Date: Mon, 28 Mar 2011 06:59:22 +0200
From: Heinrich Sinnreich 
To: Lorenzo Miniero 
Message-ID: 
Thread-Topic: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
Thread-Index: AcvtBObqoh2Pq+KsRUy1RLp0J8Rq0A==
In-Reply-To: <20110326174504.56061fab@lminiero-acer>
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 04:57:54 -0000

Thanks for the demo yesterday, Sunday outside the welcome reception!
For those who may have missed it, the RTP-over-HTTP tunneling demo really
works, though some of us will try again, after flying home.

Thanks again,
Henry

On 3/26/11 6:45 PM, "Lorenzo Miniero"  wrote:

> Dear all,
> 
> as anticipated, I've just made available a prototype version of our
> RTP-over-HTTP tunneling approach:
> 
> http://143.225.229.136/WebLite/echotest.jsp
> 
> It is a very simple Echo Test: the page uses JavaScript to setup a
> server-originated SIP call to a URI associated with an RTP echo application.
> Then, a Java applet is started to take care of the the RTP tunneling and
> encoding/decoding process.
> 
> To play with it, just enable the "Enable RTP tunneling" checkbox and then
> click the button to launch the test. If everything goes fine, a new box should
> appear on the left with the remote audio and video streams: a voice prompt
> with a video logo will precede the actual echo feature. All the RTP packets
> will be tunneled over HTTP.
> 
> Before you start, nevertheless, please have a quick look at a few
> prerequisites and instructions here:
> 
> http://ietf.conf.meetecho.com/index.php/Web_Client
> 
> for what concerns the applet-based Audio/Video functionality (last section in
> the page: our Web Client uses the same applet). In fact, applets can give a
> few problems at times (especially for multimedia-driven applications) in some
> configurations, and so checking a few things before delving in the demo may
> spare you from headache! :)
> 
> Let us know if and how it works. Of course please beware that this is just a
> very dumb prototype, so rough that it can give you the creeps if you just skim
> over it! :)
> Nevertheless, the basic concepts are there, so it may be useful to foster some
> discussion on the matter. If you're interested in some more details, I'm
> already in Prague.
> 
> Lorenzo
> 
> 
> 
> On Thu, 24 Mar 2011 17:52:17 +0100
> Lorenzo Miniero  wrote:
> 
>> Hi Henry,
>> 
>> 
>> On Wed, 23 Mar 2011 11:25:07 -0500
>> Henry Sinnreich  wrote:
>> 
>>> Hi Lorenzo, Cullen,
>>> 
>>>> We implemented something like that for our system, also documented in a
>>>> paper
>>>> lately: basically a controller protocol that tunnels every other protocol
>>>> on
>>>> top of HTTP to bypass restrictive firewalls.
>>> 
>>> Having explained HTTP tunneling flavors in the paper and also shown
>>> measurement results, besides a working product, should really make IMO your
>>> approach or something similar to it an option for the RTC-Web media
>>> solution. Such an option should also be reflected in the API part, on which
>>> the draft-jennings-rtcweb-api is focused.
>>> 
>> 
>> 
>> I'd be glad to contribute to the definition of a related solution, should
>> such a work be started within the IETF.
>> 
>> I'll also try and make available, hopefully before the meeting, a publicly
>> reachable prototype version of our current approach for you all to play with,
>> just to have some live experience to foster discussion.
>> 
>> L.
>> 
>> 
>>> RT media over HTTP may be quite effective in some usage scenarios, such as
>>> CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
>>> networks with plenty of bandwidth, smaller delay and arguably without all
>>> the other SIP, etc. intermediaries found inside and between most service
>>> provider networks.
>>> 
>>> I know profiles in standards are not desirable, but hindering innovations
>>> such as yours and ignoring CDN and other Web application scenarios is even
>>> less desirable.
>>> This is a challenge both the IETF and the W3C will hopefully manage.
>>> 
>>> Thanks for sharing,
>>> 
>>> Henry
>>> 
>>> 
>>> On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
>>> 
>>>> On Wed, 16 Mar 2011 13:54:39 -0500
>>>> Henry Sinnreich  wrote:
>>>> 
>>>>> Hi Cullen,
>>>>> 
>>>>> Section 3. Requirements says:
>>>>> ... In addition, there is a strong interest to define a TURN-like protocol
>>>>>    that looks like HTTP to intermediaries, so that media can be tunneled
>>>>>    over HTTP.
>>>>> ...
>>>>> 
>>>>> This is very interesting and innovative.
>>>>> Can you explain here a little bit?
>>>>> 
>>>>> Thanks, Henry
>>>>> 
>>>> 
>>>> 
>>>> We implemented something like that for our system, also documented in a
>>>> paper
>>>> lately: basically a controller protocol that tunnels every other protocol
>>>> on
>>>> top of HTTP to bypass restrictive firewalls. Since it works, we were also
>>>> thinking about a "lightweight" version of such protocol to only convey RTP
>>>> specifically for our web interface, considering RTP is the only protocol we
>>>> couldn't wrap via JavaScript so far. So I agree that such a TURN-like
>>>> protocol
>>>> would be a good idea for the WG.
>>>> 
>>>> Of course, there would be limitations and issues because of this transport,
>>>> but since firewalls are there to stay, and few of them like RTP appearing
>>>> "out
>>>> of the blue", that's something we probably need to live with.
>>>> 
>>>> L.
>>>> 
>>>> 
>>>>> 
>>>>> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
>>>>> 
>>>>>> 
>>>>>> I wrote up the start of a draft on requirements.
>>>>>> 
>>>>>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>>>>>> 
>>>>>> I think the key thing for the key thing for IETF working is starting to
>>>>>> think
>>>>>> about the requirements and what on the on the wire protocols are that we
>>>>>> want
>>>>>> to select. I included some API definition because I found it helpful to
>>>>>> be
>>>>>> thinking about that in this draft but that would not be part of the IETF
>>>>>> work.
>>>>>> That would be work I think W3C should do. The draft also touches on the
>>>>>> of
>>>>>> the
>>>>>> security issues for media and signaling originating from browsers.
>>>>>> 
>>>>>> Cullen 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> _______________________________________________
>>>>>> dispatch mailing list
>>>>>> dispatch@ietf.org
>>>>>> https://www.ietf.org/mailman/listinfo/dispatch
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> dispatch mailing list
>>>>> dispatch@ietf.org
>>>>> https://www.ietf.org/mailman/listinfo/dispatch
>>>> 
>>> 
>>> 
>>> 
>> 
>> 
>> -- 
>> Lorenzo Miniero, COB
>> 
>> Meetecho s.r.l.
>> Web Conferencing and Collaboration Tools
>> http://www.meetecho.com
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
> 



From dean.willis@softarmor.com  Mon Mar 28 00:31:50 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A9BE33A68D7 for ; Mon, 28 Mar 2011 00:31:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.975
X-Spam-Level: 
X-Spam-Status: No, score=-102.975 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, NORMAL_HTTP_TO_IP=0.001, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yNVVzzabva9c for ; Mon, 28 Mar 2011 00:31:48 -0700 (PDT)
Received: from mail-pw0-f44.google.com (mail-pw0-f44.google.com [209.85.160.44]) by core3.amsl.com (Postfix) with ESMTP id AC99A3A68D5 for ; Mon, 28 Mar 2011 00:31:48 -0700 (PDT)
Received: by pwi5 with SMTP id 5so662519pwi.31 for ; Mon, 28 Mar 2011 00:33:26 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.142.174.5 with SMTP id w5mr3307857wfe.222.1301297605648; Mon, 28 Mar 2011 00:33:25 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 00:33:25 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 00:33:25 -0700 (PDT)
In-Reply-To: <20110326174504.56061fab@lminiero-acer>
References:  <20110324175217.71812238@lminiero-acer> <20110326174504.56061fab@lminiero-acer>
Date: Mon, 28 Mar 2011 02:33:25 -0500
Message-ID: 
From: Dean Willis 
To: Lorenzo Miniero 
Content-Type: multipart/alternative; boundary=000e0cd2bf9c88201a049f85f5ef
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 07:31:50 -0000

--000e0cd2bf9c88201a049f85f5ef
Content-Type: text/plain; charset=ISO-8859-1

I suggest trying it from a pc tethered to an EDGE mobile modem.
On Mar 26, 2011 4:49 PM, "Lorenzo Miniero"  wrote:
> Dear all,
>
> as anticipated, I've just made available a prototype version of our
RTP-over-HTTP tunneling approach:
>
> http://143.225.229.136/WebLite/echotest.jsp
>
> It is a very simple Echo Test: the page uses JavaScript to setup a
server-originated SIP call to a URI associated with an RTP echo application.
Then, a Java applet is started to take care of the the RTP tunneling and
encoding/decoding process.
>
> To play with it, just enable the "Enable RTP tunneling" checkbox and then
click the button to launch the test. If everything goes fine, a new box
should appear on the left with the remote audio and video streams: a voice
prompt with a video logo will precede the actual echo feature. All the RTP
packets will be tunneled over HTTP.
>
> Before you start, nevertheless, please have a quick look at a few
prerequisites and instructions here:
>
> http://ietf.conf.meetecho.com/index.php/Web_Client
>
> for what concerns the applet-based Audio/Video functionality (last section
in the page: our Web Client uses the same applet). In fact, applets can give
a few problems at times (especially for multimedia-driven applications) in
some configurations, and so checking a few things before delving in the demo
may spare you from headache! :)
>
> Let us know if and how it works. Of course please beware that this is just
a very dumb prototype, so rough that it can give you the creeps if you just
skim over it! :)
> Nevertheless, the basic concepts are there, so it may be useful to foster
some discussion on the matter. If you're interested in some more details,
I'm already in Prague.
>
> Lorenzo
>
>
>
> On Thu, 24 Mar 2011 17:52:17 +0100
> Lorenzo Miniero  wrote:
>
>> Hi Henry,
>>
>>
>> On Wed, 23 Mar 2011 11:25:07 -0500
>> Henry Sinnreich  wrote:
>>
>> > Hi Lorenzo, Cullen,
>> >
>> > > We implemented something like that for our system, also documented in
a paper
>> > > lately: basically a controller protocol that tunnels every other
protocol on
>> > > top of HTTP to bypass restrictive firewalls.
>> >
>> > Having explained HTTP tunneling flavors in the paper and also shown
>> > measurement results, besides a working product, should really make IMO
your
>> > approach or something similar to it an option for the RTC-Web media
>> > solution. Such an option should also be reflected in the API part, on
which
>> > the draft-jennings-rtcweb-api is focused.
>> >
>>
>>
>> I'd be glad to contribute to the definition of a related solution, should
such a work be started within the IETF.
>>
>> I'll also try and make available, hopefully before the meeting, a
publicly reachable prototype version of our current approach for you all to
play with, just to have some live experience to foster discussion.
>>
>> L.
>>
>>
>> > RT media over HTTP may be quite effective in some usage scenarios, such
as
>> > CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
>> > networks with plenty of bandwidth, smaller delay and arguably without
all
>> > the other SIP, etc. intermediaries found inside and between most
service
>> > provider networks.
>> >
>> > I know profiles in standards are not desirable, but hindering
innovations
>> > such as yours and ignoring CDN and other Web application scenarios is
even
>> > less desirable.
>> > This is a challenge both the IETF and the W3C will hopefully manage.
>> >
>> > Thanks for sharing,
>> >
>> > Henry
>> >
>> >
>> > On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
>> >
>> > > On Wed, 16 Mar 2011 13:54:39 -0500
>> > > Henry Sinnreich  wrote:
>> > >
>> > >> Hi Cullen,
>> > >>
>> > >> Section 3. Requirements says:
>> > >> ... In addition, there is a strong interest to define a TURN-like
protocol
>> > >> that looks like HTTP to intermediaries, so that media can be
tunneled
>> > >> over HTTP.
>> > >> ...
>> > >>
>> > >> This is very interesting and innovative.
>> > >> Can you explain here a little bit?
>> > >>
>> > >> Thanks, Henry
>> > >>
>> > >
>> > >
>> > > We implemented something like that for our system, also documented in
a paper
>> > > lately: basically a controller protocol that tunnels every other
protocol on
>> > > top of HTTP to bypass restrictive firewalls. Since it works, we were
also
>> > > thinking about a "lightweight" version of such protocol to only
convey RTP
>> > > specifically for our web interface, considering RTP is the only
protocol we
>> > > couldn't wrap via JavaScript so far. So I agree that such a TURN-like
protocol
>> > > would be a good idea for the WG.
>> > >
>> > > Of course, there would be limitations and issues because of this
transport,
>> > > but since firewalls are there to stay, and few of them like RTP
appearing "out
>> > > of the blue", that's something we probably need to live with.
>> > >
>> > > L.
>> > >
>> > >
>> > >>
>> > >> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
>> > >>
>> > >>>
>> > >>> I wrote up the start of a draft on requirements.
>> > >>>
>> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>> > >>>
>> > >>> I think the key thing for the key thing for IETF working is
starting to
>> > >>> think
>> > >>> about the requirements and what on the on the wire protocols are
that we
>> > >>> want
>> > >>> to select. I included some API definition because I found it
helpful to be
>> > >>> thinking about that in this draft but that would not be part of the
IETF
>> > >>> work.
>> > >>> That would be work I think W3C should do. The draft also touches on
the of
>> > >>> the
>> > >>> security issues for media and signaling originating from browsers.
>> > >>>
>> > >>> Cullen
>> > >>>
>> > >>>
>> > >>>
>> > >>>
>> > >>> _______________________________________________
>> > >>> dispatch mailing list
>> > >>> dispatch@ietf.org
>> > >>> https://www.ietf.org/mailman/listinfo/dispatch
>> > >>
>> > >>
>> > >> _______________________________________________
>> > >> dispatch mailing list
>> > >> dispatch@ietf.org
>> > >> https://www.ietf.org/mailman/listinfo/dispatch
>> > >
>> >
>> >
>> >
>>
>>
>> --
>> Lorenzo Miniero, COB
>>
>> Meetecho s.r.l.
>> Web Conferencing and Collaboration Tools
>> http://www.meetecho.com
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>
>
> --
> Lorenzo Miniero, COB
>
> Meetecho s.r.l.
> Web Conferencing and Collaboration Tools
> http://www.meetecho.com
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

--000e0cd2bf9c88201a049f85f5ef
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable


I suggest trying it from a pc tethered to an EDGE mobile modem.


On Mar 26, 2011 4:49 PM, "Lorenzo Miniero&q=
uot; <lorenzo@meetecho.com&g=
t; wrote:
> Dear all,
> 
> as antici=
pated, I've just made available a prototype version of our RTP-over-HTT=
P tunneling approach:

> 
> 	http=
://143.225.229.136/WebLite/echotest.jsp
> 
> It is a very s=
imple Echo Test: the page uses JavaScript to setup a server-originated SIP =
call to a URI associated with an RTP echo application. Then, a Java applet =
is started to take care of the the RTP tunneling and encoding/decoding proc=
ess.

> 
> To play with it, just enable the "Enable RTP tunneling&q=
uot; checkbox and then click the button to launch the test. If everything g=
oes fine, a new box should appear on the left with the remote audio and vid=
eo streams: a voice prompt with a video logo will precede the actual echo f=
eature. All the RTP packets will be tunneled over HTTP.

> 
> Before you start, nevertheless, please have a quick look at a=
 few prerequisites and instructions here:
> 
> 	http://ietf.conf.meetecho.c=
om/index.php/Web_Client

> 
> for what concerns the applet-based Audio/Video functionality =
(last section in the page: our Web Client uses the same applet). In fact, a=
pplets can give a few problems at times (especially for multimedia-driven a=
pplications) in some configurations, and so checking a few things before de=
lving in the demo may spare you from headache! :)

> 
> Let us know if and how it works. Of course please beware that=
 this is just a very dumb prototype, so rough that it can give you the cree=
ps if you just skim over it! :)
> Nevertheless, the basic concepts ar=
e there, so it may be useful to foster some discussion on the matter. If yo=
u're interested in some more details, I'm already in Prague.

> 
> Lorenzo
> 
> 
> 
> On Thu, 24 Mar 201=
1 17:52:17 +0100
> Lorenzo Miniero <lorenzo@meetecho.com> wrote:
> 
>> Hi Henry=
,

>> 
>> 
>> On Wed, 23 Mar 2011 11:25:07 -0500
&g=
t;> Henry Sinnreich <hen=
ry.sinnreich@gmail.com> wrote:
>> 
>> > Hi Lore=
nzo, Cullen,

>> > 
>> > > We implemented something like that for=
 our system, also documented in a paper
>> > > lately: basic=
ally a controller protocol that tunnels every other protocol on
>>=
 > > top of HTTP to bypass restrictive firewalls.

>> > 
>> > Having explained HTTP tunneling flavors in =
the paper and also shown
>> > measurement results, besides a wo=
rking product, should really make IMO your
>> > approach or som=
ething similar to it an option for the RTC-Web media

>> > solution. Such an option should also be reflected in the API =
part, on which
>> > the draft-jennings-rtcweb-api is focused.>> > 
>> 
>> 
>> I'd be glad to c=
ontribute to the definition of a related solution, should such a work be st=
arted within the IETF.

>> 
>> I'll also try and make available, hopefully befor=
e the meeting, a publicly reachable prototype version of our current approa=
ch for you all to play with, just to have some live experience to foster di=
scussion.

>> 
>> L.
>> 
>> 
>> > RT medi=
a over HTTP may be quite effective in some usage scenarios, such as
>=
> > CDN with >100k HTPP proxies around the globe, collocated in 1,=
000s of

>> > networks with plenty of bandwidth, smaller delay and arguably=
 without all
>> > the other SIP, etc. intermediaries found insi=
de and between most service
>> > provider networks.
>>=
 > 

>> > I know profiles in standards are not desirable, but hindering=
 innovations
>> > such as yours and ignoring CDN and other Web =
application scenarios is even
>> > less desirable. 
>>=
 > This is a challenge both the IETF and the W3C will hopefully manage.<=
br>
>> > 
>> > Thanks for sharing,
>> > 
&g=
t;> > Henry
>> > 
>> > 
>> > On 3=
/16/11 3:12 PM, "Lorenzo Miniero" <lorenzo@meetecho.com> wrote:

>> > 
>> > > On Wed, 16 Mar 2011 13:54:39 -0500
=
>> > > Henry Sinnreich <henry.sinnreich@gmail.com> wrote:
>> > > 
>> > >> Hi Cullen,
>> > >> 
>> &g=
t; >> Section 3. Requirements says:
>> > >> ... In =
addition, there is a strong interest to define a TURN-like protocol

>> > >>    that looks like HTTP to intermediaries, so that m=
edia can be tunneled
>> > >>    over HTTP. 
>> &=
gt; >> ...
>> > >> 
>> > >> This =
is very interesting and innovative.

>> > >> Can you explain here a little bit?
>> > =
>> 
>> > >> Thanks, Henry
>> > >>=
 
>> > > 
>> > > 
>> > > We im=
plemented something like that for our system, also documented in a paper
>> > > lately: basically a controller protocol that tunnels eve=
ry other protocol on
>> > > top of HTTP to bypass restrictiv=
e firewalls. Since it works, we were also
>> > > thinking ab=
out a "lightweight" version of such protocol to only convey RTP
>> > > specifically for our web interface, considering RTP is t=
he only protocol we
>> > > couldn't wrap via JavaScript =
so far. So I agree that such a TURN-like protocol
>> > > wou=
ld be a good idea for the WG.

>> > > 
>> > > Of course, there would be limitat=
ions and issues because of this transport,
>> > > but since =
firewalls are there to stay, and few of them like RTP appearing "out
>> > > of the blue", that's something we probably need=
 to live with.
>> > > 
>> > > L.
>> =
> > 
>> > > 
>> > >> 
>> &g=
t; >> On 3/10/11 2:20 PM, "Cullen Jennings" <fluffy@cisco.com> wrote:

>> > >> 
>> > >>> 
>> > >=
;>> I wrote up the start of a draft on requirements.
>> >=
 >>> 
>> > >>> http://tools.ietf.org/html/draft-jenn=
ings-rtcweb-api-00

>> > >>> 
>> > >>> I think the key t=
hing for the key thing for IETF working is starting to
>> > >=
;>> think
>> > >>> about the requirements and wh=
at on the on the wire protocols are that we

>> > >>> want
>> > >>> to select. I =
included some API definition because I found it helpful to be
>> &=
gt; >>> thinking about that in this draft but that would not be pa=
rt of the IETF

>> > >>> work.
>> > >>> That would b=
e work I think W3C should do. The draft also touches on the of
>> =
> >>> the
>> > >>> security issues for med=
ia and signaling originating from browsers.

>> > >>> 
>> > >>> Cullen 
>&g=
t; > >>> 
>> > >>> 
>> > >&=
gt;> 
>> > >>> 
>> > >>> _____=
__________________________________________

>> > >>> dispatch mailing list
>> > >>&=
gt; dispatch@ietf.org
>> =
> >>> https://www.ietf.org/mailman/listinfo/dispatch

>> > >> 
>> > >> 
>> > >>=
; _______________________________________________
>> > >>=
 dispatch mailing list
>> > >> dispatch@ietf.org

>> > >> https://www.ietf.org/mailman/listinfo/dispatch
>> > =
> 
>> > 
>> > 
>> > 
>> 
>> 
>> -- 
>> Lorenzo Miniero, COB
>> 
=
>> Meetecho s.r.l.
>> Web Conferencing and Collaboration Too=
ls
>> http://www.meetecho.com<=
/a>

>> _______________________________________________
>> dispat=
ch mailing list
>> dispatch@i=
etf.org
>> https://www.ietf.org/mailman/listinfo/dispatch

> 
> 
> -- 
> Lorenzo Miniero, COB
> 
> Me=
etecho s.r.l.
> Web Conferencing and Collaboration Tools
> http://www.meetecho.com
> _______=
________________________________________

> dispatch mailing list
> dis=
patch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch


--000e0cd2bf9c88201a049f85f5ef--

From lorenzo@meetecho.com  Mon Mar 28 05:06:52 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 54E883A68A7 for ; Mon, 28 Mar 2011 05:06:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.718
X-Spam-Level: 
X-Spam-Status: No, score=-0.718 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, NORMAL_HTTP_TO_IP=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rgIrEjNtSRnT for ; Mon, 28 Mar 2011 05:06:51 -0700 (PDT)
Received: from smtplq03.aruba.it (smtplq-out8.aruba.it [62.149.158.28]) by core3.amsl.com (Postfix) with SMTP id 492ED3A6845 for ; Mon, 28 Mar 2011 05:06:49 -0700 (PDT)
Received: (qmail 6628 invoked by uid 89); 28 Mar 2011 12:08:23 -0000
Received: from unknown (HELO smtp8.aruba.it) (62.149.158.228) by smtplq03.aruba.it with SMTP; 28 Mar 2011 12:08:23 -0000
Received: (qmail 11253 invoked by uid 89); 28 Mar 2011 12:08:24 -0000
Received: from unknown (HELO lminiero-acer) (lorenzo@meetecho.com@130.129.21.22) by smtp8.ad.aruba.it with SMTP; 28 Mar 2011 12:08:24 -0000
Date: Mon, 28 Mar 2011 14:03:35 +0200
From: Lorenzo Miniero 
To: Dean Willis 
Message-ID: <20110328140335.4ff34ddb@lminiero-acer>
In-Reply-To: 
References:  <20110324175217.71812238@lminiero-acer> <20110326174504.56061fab@lminiero-acer> 
Organization: Meetecho
X-Mailer: Claws Mail 3.7.8 (GTK+ 2.22.0; i386-redhat-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Spam-Rating: smtp8.ad.aruba.it 1.6.2 0/1000/N
X-Spam-Rating: smtplq03.aruba.it 1.6.2 0/1000/N
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 12:06:52 -0000

Hi Dean,


On Mon, 28 Mar 2011 02:33:25 -0500
Dean Willis  wrote:

> I suggest trying it from a pc tethered to an EDGE mobile modem.


Or with no network at all :)

Jokes apart, it's clear that the QoE can't benefit from such limited bandwidth/conditions, and that would be true for pure RTP as well. I can only say that we've used this and older versions of the tunnel on limited networks (home adsls, umts, etc.) for our purposes and it seemed fine, though I agree that every attempt at addressing this feature should take care of tackling such issues.

L.



> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero"  wrote:
> > Dear all,
> >
> > as anticipated, I've just made available a prototype version of our
> RTP-over-HTTP tunneling approach:
> >
> > http://143.225.229.136/WebLite/echotest.jsp
> >
> > It is a very simple Echo Test: the page uses JavaScript to setup a
> server-originated SIP call to a URI associated with an RTP echo application.
> Then, a Java applet is started to take care of the the RTP tunneling and
> encoding/decoding process.
> >
> > To play with it, just enable the "Enable RTP tunneling" checkbox and then
> click the button to launch the test. If everything goes fine, a new box
> should appear on the left with the remote audio and video streams: a voice
> prompt with a video logo will precede the actual echo feature. All the RTP
> packets will be tunneled over HTTP.
> >
> > Before you start, nevertheless, please have a quick look at a few
> prerequisites and instructions here:
> >
> > http://ietf.conf.meetecho.com/index.php/Web_Client
> >
> > for what concerns the applet-based Audio/Video functionality (last section
> in the page: our Web Client uses the same applet). In fact, applets can give
> a few problems at times (especially for multimedia-driven applications) in
> some configurations, and so checking a few things before delving in the demo
> may spare you from headache! :)
> >
> > Let us know if and how it works. Of course please beware that this is just
> a very dumb prototype, so rough that it can give you the creeps if you just
> skim over it! :)
> > Nevertheless, the basic concepts are there, so it may be useful to foster
> some discussion on the matter. If you're interested in some more details,
> I'm already in Prague.
> >
> > Lorenzo
> >
> >
> >
> > On Thu, 24 Mar 2011 17:52:17 +0100
> > Lorenzo Miniero  wrote:
> >
> >> Hi Henry,
> >>
> >>
> >> On Wed, 23 Mar 2011 11:25:07 -0500
> >> Henry Sinnreich  wrote:
> >>
> >> > Hi Lorenzo, Cullen,
> >> >
> >> > > We implemented something like that for our system, also documented in
> a paper
> >> > > lately: basically a controller protocol that tunnels every other
> protocol on
> >> > > top of HTTP to bypass restrictive firewalls.
> >> >
> >> > Having explained HTTP tunneling flavors in the paper and also shown
> >> > measurement results, besides a working product, should really make IMO
> your
> >> > approach or something similar to it an option for the RTC-Web media
> >> > solution. Such an option should also be reflected in the API part, on
> which
> >> > the draft-jennings-rtcweb-api is focused.
> >> >
> >>
> >>
> >> I'd be glad to contribute to the definition of a related solution, should
> such a work be started within the IETF.
> >>
> >> I'll also try and make available, hopefully before the meeting, a
> publicly reachable prototype version of our current approach for you all to
> play with, just to have some live experience to foster discussion.
> >>
> >> L.
> >>
> >>
> >> > RT media over HTTP may be quite effective in some usage scenarios, such
> as
> >> > CDN with >100k HTPP proxies around the globe, collocated in 1,000s of
> >> > networks with plenty of bandwidth, smaller delay and arguably without
> all
> >> > the other SIP, etc. intermediaries found inside and between most
> service
> >> > provider networks.
> >> >
> >> > I know profiles in standards are not desirable, but hindering
> innovations
> >> > such as yours and ignoring CDN and other Web application scenarios is
> even
> >> > less desirable.
> >> > This is a challenge both the IETF and the W3C will hopefully manage.
> >> >
> >> > Thanks for sharing,
> >> >
> >> > Henry
> >> >
> >> >
> >> > On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
> >> >
> >> > > On Wed, 16 Mar 2011 13:54:39 -0500
> >> > > Henry Sinnreich  wrote:
> >> > >
> >> > >> Hi Cullen,
> >> > >>
> >> > >> Section 3. Requirements says:
> >> > >> ... In addition, there is a strong interest to define a TURN-like
> protocol
> >> > >> that looks like HTTP to intermediaries, so that media can be
> tunneled
> >> > >> over HTTP.
> >> > >> ...
> >> > >>
> >> > >> This is very interesting and innovative.
> >> > >> Can you explain here a little bit?
> >> > >>
> >> > >> Thanks, Henry
> >> > >>
> >> > >
> >> > >
> >> > > We implemented something like that for our system, also documented in
> a paper
> >> > > lately: basically a controller protocol that tunnels every other
> protocol on
> >> > > top of HTTP to bypass restrictive firewalls. Since it works, we were
> also
> >> > > thinking about a "lightweight" version of such protocol to only
> convey RTP
> >> > > specifically for our web interface, considering RTP is the only
> protocol we
> >> > > couldn't wrap via JavaScript so far. So I agree that such a TURN-like
> protocol
> >> > > would be a good idea for the WG.
> >> > >
> >> > > Of course, there would be limitations and issues because of this
> transport,
> >> > > but since firewalls are there to stay, and few of them like RTP
> appearing "out
> >> > > of the blue", that's something we probably need to live with.
> >> > >
> >> > > L.
> >> > >
> >> > >
> >> > >>
> >> > >> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
> >> > >>
> >> > >>>
> >> > >>> I wrote up the start of a draft on requirements.
> >> > >>>
> >> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> >> > >>>
> >> > >>> I think the key thing for the key thing for IETF working is
> starting to
> >> > >>> think
> >> > >>> about the requirements and what on the on the wire protocols are
> that we
> >> > >>> want
> >> > >>> to select. I included some API definition because I found it
> helpful to be
> >> > >>> thinking about that in this draft but that would not be part of the
> IETF
> >> > >>> work.
> >> > >>> That would be work I think W3C should do. The draft also touches on
> the of
> >> > >>> the
> >> > >>> security issues for media and signaling originating from browsers.
> >> > >>>
> >> > >>> Cullen
> >> > >>>
> >> > >>>
> >> > >>>
> >> > >>>
> >> > >>> _______________________________________________
> >> > >>> dispatch mailing list
> >> > >>> dispatch@ietf.org
> >> > >>> https://www.ietf.org/mailman/listinfo/dispatch
> >> > >>
> >> > >>
> >> > >> _______________________________________________
> >> > >> dispatch mailing list
> >> > >> dispatch@ietf.org
> >> > >> https://www.ietf.org/mailman/listinfo/dispatch
> >> > >
> >> >
> >> >
> >> >
> >>
> >>
> >> --
> >> Lorenzo Miniero, COB
> >>
> >> Meetecho s.r.l.
> >> Web Conferencing and Collaboration Tools
> >> http://www.meetecho.com
> >> _______________________________________________
> >> dispatch mailing list
> >> dispatch@ietf.org
> >> https://www.ietf.org/mailman/listinfo/dispatch
> >
> >
> > --
> > Lorenzo Miniero, COB
> >
> > Meetecho s.r.l.
> > Web Conferencing and Collaboration Tools
> > http://www.meetecho.com
> > _______________________________________________
> > dispatch mailing list
> > dispatch@ietf.org
> > https://www.ietf.org/mailman/listinfo/dispatch


-- 
Lorenzo Miniero, COB

Meetecho s.r.l.
Web Conferencing and Collaboration Tools
http://www.meetecho.com

From dean.willis@softarmor.com  Mon Mar 28 05:12:53 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 061CF3A68B8 for ; Mon, 28 Mar 2011 05:12:53 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.975
X-Spam-Level: 
X-Spam-Status: No, score=-102.975 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, NORMAL_HTTP_TO_IP=0.001, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IPddRttBod3X for ; Mon, 28 Mar 2011 05:12:51 -0700 (PDT)
Received: from mail-pz0-f44.google.com (mail-pz0-f44.google.com [209.85.210.44]) by core3.amsl.com (Postfix) with ESMTP id 3E83A3A67DF for ; Mon, 28 Mar 2011 05:12:51 -0700 (PDT)
Received: by pzk30 with SMTP id 30so685483pzk.31 for ; Mon, 28 Mar 2011 05:14:28 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.143.178.10 with SMTP id f10mr3741341wfp.108.1301314468752; Mon, 28 Mar 2011 05:14:28 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 05:14:28 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 05:14:28 -0700 (PDT)
In-Reply-To: <20110328140335.4ff34ddb@lminiero-acer>
References:  <20110324175217.71812238@lminiero-acer> <20110326174504.56061fab@lminiero-acer>  <20110328140335.4ff34ddb@lminiero-acer>
Date: Mon, 28 Mar 2011 07:14:28 -0500
Message-ID: 
From: Dean Willis 
To: Lorenzo Miniero 
Content-Type: multipart/alternative; boundary=000e0cd5ccbaa6b270049f89e2c5
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 12:12:53 -0000

--000e0cd5ccbaa6b270049f89e2c5
Content-Type: text/plain; charset=ISO-8859-1

It is not the bandwidth that worries me. It is packet loss and resulting tcp
retransmit jitter.
On Mar 28, 2011 2:08 PM, "Lorenzo Miniero"  wrote:
> Hi Dean,
>
>
> On Mon, 28 Mar 2011 02:33:25 -0500
> Dean Willis  wrote:
>
>> I suggest trying it from a pc tethered to an EDGE mobile modem.
>
>
> Or with no network at all :)
>
> Jokes apart, it's clear that the QoE can't benefit from such limited
bandwidth/conditions, and that would be true for pure RTP as well. I can
only say that we've used this and older versions of the tunnel on limited
networks (home adsls, umts, etc.) for our purposes and it seemed fine,
though I agree that every attempt at addressing this feature should take
care of tackling such issues.
>
> L.
>
>
>
>> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero"  wrote:
>> > Dear all,
>> >
>> > as anticipated, I've just made available a prototype version of our
>> RTP-over-HTTP tunneling approach:
>> >
>> > http://143.225.229.136/WebLite/echotest.jsp
>> >
>> > It is a very simple Echo Test: the page uses JavaScript to setup a
>> server-originated SIP call to a URI associated with an RTP echo
application.
>> Then, a Java applet is started to take care of the the RTP tunneling and
>> encoding/decoding process.
>> >
>> > To play with it, just enable the "Enable RTP tunneling" checkbox and
then
>> click the button to launch the test. If everything goes fine, a new box
>> should appear on the left with the remote audio and video streams: a
voice
>> prompt with a video logo will precede the actual echo feature. All the
RTP
>> packets will be tunneled over HTTP.
>> >
>> > Before you start, nevertheless, please have a quick look at a few
>> prerequisites and instructions here:
>> >
>> > http://ietf.conf.meetecho.com/index.php/Web_Client
>> >
>> > for what concerns the applet-based Audio/Video functionality (last
section
>> in the page: our Web Client uses the same applet). In fact, applets can
give
>> a few problems at times (especially for multimedia-driven applications)
in
>> some configurations, and so checking a few things before delving in the
demo
>> may spare you from headache! :)
>> >
>> > Let us know if and how it works. Of course please beware that this is
just
>> a very dumb prototype, so rough that it can give you the creeps if you
just
>> skim over it! :)
>> > Nevertheless, the basic concepts are there, so it may be useful to
foster
>> some discussion on the matter. If you're interested in some more details,
>> I'm already in Prague.
>> >
>> > Lorenzo
>> >
>> >
>> >
>> > On Thu, 24 Mar 2011 17:52:17 +0100
>> > Lorenzo Miniero  wrote:
>> >
>> >> Hi Henry,
>> >>
>> >>
>> >> On Wed, 23 Mar 2011 11:25:07 -0500
>> >> Henry Sinnreich  wrote:
>> >>
>> >> > Hi Lorenzo, Cullen,
>> >> >
>> >> > > We implemented something like that for our system, also documented
in
>> a paper
>> >> > > lately: basically a controller protocol that tunnels every other
>> protocol on
>> >> > > top of HTTP to bypass restrictive firewalls.
>> >> >
>> >> > Having explained HTTP tunneling flavors in the paper and also shown
>> >> > measurement results, besides a working product, should really make
IMO
>> your
>> >> > approach or something similar to it an option for the RTC-Web media
>> >> > solution. Such an option should also be reflected in the API part,
on
>> which
>> >> > the draft-jennings-rtcweb-api is focused.
>> >> >
>> >>
>> >>
>> >> I'd be glad to contribute to the definition of a related solution,
should
>> such a work be started within the IETF.
>> >>
>> >> I'll also try and make available, hopefully before the meeting, a
>> publicly reachable prototype version of our current approach for you all
to
>> play with, just to have some live experience to foster discussion.
>> >>
>> >> L.
>> >>
>> >>
>> >> > RT media over HTTP may be quite effective in some usage scenarios,
such
>> as
>> >> > CDN with >100k HTPP proxies around the globe, collocated in 1,000s
of
>> >> > networks with plenty of bandwidth, smaller delay and arguably
without
>> all
>> >> > the other SIP, etc. intermediaries found inside and between most
>> service
>> >> > provider networks.
>> >> >
>> >> > I know profiles in standards are not desirable, but hindering
>> innovations
>> >> > such as yours and ignoring CDN and other Web application scenarios
is
>> even
>> >> > less desirable.
>> >> > This is a challenge both the IETF and the W3C will hopefully manage.
>> >> >
>> >> > Thanks for sharing,
>> >> >
>> >> > Henry
>> >> >
>> >> >
>> >> > On 3/16/11 3:12 PM, "Lorenzo Miniero"  wrote:
>> >> >
>> >> > > On Wed, 16 Mar 2011 13:54:39 -0500
>> >> > > Henry Sinnreich  wrote:
>> >> > >
>> >> > >> Hi Cullen,
>> >> > >>
>> >> > >> Section 3. Requirements says:
>> >> > >> ... In addition, there is a strong interest to define a TURN-like
>> protocol
>> >> > >> that looks like HTTP to intermediaries, so that media can be
>> tunneled
>> >> > >> over HTTP.
>> >> > >> ...
>> >> > >>
>> >> > >> This is very interesting and innovative.
>> >> > >> Can you explain here a little bit?
>> >> > >>
>> >> > >> Thanks, Henry
>> >> > >>
>> >> > >
>> >> > >
>> >> > > We implemented something like that for our system, also documented
in
>> a paper
>> >> > > lately: basically a controller protocol that tunnels every other
>> protocol on
>> >> > > top of HTTP to bypass restrictive firewalls. Since it works, we
were
>> also
>> >> > > thinking about a "lightweight" version of such protocol to only
>> convey RTP
>> >> > > specifically for our web interface, considering RTP is the only
>> protocol we
>> >> > > couldn't wrap via JavaScript so far. So I agree that such a
TURN-like
>> protocol
>> >> > > would be a good idea for the WG.
>> >> > >
>> >> > > Of course, there would be limitations and issues because of this
>> transport,
>> >> > > but since firewalls are there to stay, and few of them like RTP
>> appearing "out
>> >> > > of the blue", that's something we probably need to live with.
>> >> > >
>> >> > > L.
>> >> > >
>> >> > >
>> >> > >>
>> >> > >> On 3/10/11 2:20 PM, "Cullen Jennings"  wrote:
>> >> > >>
>> >> > >>>
>> >> > >>> I wrote up the start of a draft on requirements.
>> >> > >>>
>> >> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>> >> > >>>
>> >> > >>> I think the key thing for the key thing for IETF working is
>> starting to
>> >> > >>> think
>> >> > >>> about the requirements and what on the on the wire protocols are
>> that we
>> >> > >>> want
>> >> > >>> to select. I included some API definition because I found it
>> helpful to be
>> >> > >>> thinking about that in this draft but that would not be part of
the
>> IETF
>> >> > >>> work.
>> >> > >>> That would be work I think W3C should do. The draft also touches
on
>> the of
>> >> > >>> the
>> >> > >>> security issues for media and signaling originating from
browsers.
>> >> > >>>
>> >> > >>> Cullen
>> >> > >>>
>> >> > >>>
>> >> > >>>
>> >> > >>>
>> >> > >>> _______________________________________________
>> >> > >>> dispatch mailing list
>> >> > >>> dispatch@ietf.org
>> >> > >>> https://www.ietf.org/mailman/listinfo/dispatch
>> >> > >>
>> >> > >>
>> >> > >> _______________________________________________
>> >> > >> dispatch mailing list
>> >> > >> dispatch@ietf.org
>> >> > >> https://www.ietf.org/mailman/listinfo/dispatch
>> >> > >
>> >> >
>> >> >
>> >> >
>> >>
>> >>
>> >> --
>> >> Lorenzo Miniero, COB
>> >>
>> >> Meetecho s.r.l.
>> >> Web Conferencing and Collaboration Tools
>> >> http://www.meetecho.com
>> >> _______________________________________________
>> >> dispatch mailing list
>> >> dispatch@ietf.org
>> >> https://www.ietf.org/mailman/listinfo/dispatch
>> >
>> >
>> > --
>> > Lorenzo Miniero, COB
>> >
>> > Meetecho s.r.l.
>> > Web Conferencing and Collaboration Tools
>> > http://www.meetecho.com
>> > _______________________________________________
>> > dispatch mailing list
>> > dispatch@ietf.org
>> > https://www.ietf.org/mailman/listinfo/dispatch
>
>
> --
> Lorenzo Miniero, COB
>
> Meetecho s.r.l.
> Web Conferencing and Collaboration Tools
> http://www.meetecho.com

--000e0cd5ccbaa6b270049f89e2c5
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable


It is not the bandwidth that worries me. It is packet loss and resulting=
 tcp retransmit jitter.


On Mar 28, 2011 2:08 PM, "Lorenzo Miniero&q=
uot; <lorenzo@meetecho.com&g=
t; wrote:
> Hi Dean,
> 
> 
> O=
n Mon, 28 Mar 2011 02:33:25 -0500

> Dean Willis <dean.will=
is@softarmor.com> wrote:
> 
>> I suggest trying it fr=
om a pc tethered to an EDGE mobile modem.
> 
> 
> Or with=
 no network at all :)

> 
> Jokes apart, it's clear that the QoE can't benefit fr=
om such limited bandwidth/conditions, and that would be true for pure RTP a=
s well. I can only say that we've used this and older versions of the t=
unnel on limited networks (home adsls, umts, etc.) for our purposes and it =
seemed fine, though I agree that every attempt at addressing this feature s=
hould take care of tackling such issues.

> 
> L.
> 
> 
> 
>> On Mar 26, 2011 4:4=
9 PM, "Lorenzo Miniero" <lorenzo@meetecho.com> wrote:
>> > Dear all,
>&g=
t; >

>> > as anticipated, I've just made available a prototype vers=
ion of our
>> RTP-over-HTTP tunneling approach:
>> >>> > htt=
p://143.225.229.136/WebLite/echotest.jsp

>> >
>> > It is a very simple Echo Test: the page uses=
 JavaScript to setup a
>> server-originated SIP call to a URI asso=
ciated with an RTP echo application.
>> Then, a Java applet is sta=
rted to take care of the the RTP tunneling and

>> encoding/decoding process.
>> >
>> > To pl=
ay with it, just enable the "Enable RTP tunneling" checkbox and t=
hen
>> click the button to launch the test. If everything goes fin=
e, a new box

>> should appear on the left with the remote audio and video streams:=
 a voice
>> prompt with a video logo will precede the actual echo =
feature. All the RTP
>> packets will be tunneled over HTTP.

>> >
>> > Before you start, nevertheless, please have =
a quick look at a few
>> prerequisites and instructions here:
&=
gt;> >
>> > http://ietf.conf.meetecho.com/index.php/Web_Client
>> >
>> > for what concerns the applet-based Audio/Vid=
eo functionality (last section
>> in the page: our Web Client uses=
 the same applet). In fact, applets can give
>> a few problems at =
times (especially for multimedia-driven applications) in

>> some configurations, and so checking a few things before delving i=
n the demo
>> may spare you from headache! :)
>> >
=
>> > Let us know if and how it works. Of course please beware that=
 this is just

>> a very dumb prototype, so rough that it can give you the creeps if=
 you just
>> skim over it! :)
>> > Nevertheless, the b=
asic concepts are there, so it may be useful to foster
>> some dis=
cussion on the matter. If you're interested in some more details,

>> I'm already in Prague.
>> >
>> > Loren=
zo
>> >
>> >
>> >
>> > On T=
hu, 24 Mar 2011 17:52:17 +0100
>> > Lorenzo Miniero <lorenzo@meetecho.com> wrote:

>> >
>> >> Hi Henry,
>> >>
>&g=
t; >>
>> >> On Wed, 23 Mar 2011 11:25:07 -0500
>=
> >> Henry Sinnreich <henry.sinnreich@gmail.com> wrote:

>> >>
>> >> > Hi Lorenzo, Cullen,
>>=
 >> >
>> >> > > We implemented something like=
 that for our system, also documented in
>> a paper
>> &g=
t;> > > lately: basically a controller protocol that tunnels every=
 other

>> protocol on
>> >> > > top of HTTP to bypass r=
estrictive firewalls.
>> >> >
>> >> > H=
aving explained HTTP tunneling flavors in the paper and also shown
>&=
gt; >> > measurement results, besides a working product, should re=
ally make IMO

>> your
>> >> > approach or something similar to it=
 an option for the RTC-Web media
>> >> > solution. Such a=
n option should also be reflected in the API part, on
>> which

>> >> > the draft-jennings-rtcweb-api is focused.
>>=
; >> >
>> >>
>> >>
>> >&=
gt; I'd be glad to contribute to the definition of a related solution, =
should

>> such a work be started within the IETF.
>> >>
&g=
t;> >> I'll also try and make available, hopefully before the =
meeting, a
>> publicly reachable prototype version of our current =
approach for you all to

>> play with, just to have some live experience to foster discussion.=

>> >>
>> >> L.
>> >>
>&=
gt; >>
>> >> > RT media over HTTP may be quite effe=
ctive in some usage scenarios, such

>> as
>> >> > CDN with >100k HTPP proxies around=
 the globe, collocated in 1,000s of
>> >> > networks with=
 plenty of bandwidth, smaller delay and arguably without
>> all
>> >> > the other SIP, etc. intermediaries found inside and =
between most
>> service
>> >> > provider network=
s.
>> >> >
>> >> > I know profiles in s=
tandards are not desirable, but hindering

>> innovations
>> >> > such as yours and ignoring C=
DN and other Web application scenarios is
>> even
>> >=
> > less desirable.
>> >> > This is a challenge bot=
h the IETF and the W3C will hopefully manage.

>> >> >
>> >> > Thanks for sharing,
>=
;> >> >
>> >> > Henry
>> >> &g=
t;
>> >> >
>> >> > On 3/16/11 3:12 PM, =
"Lorenzo Miniero" <lor=
enzo@meetecho.com> wrote:

>> >> >
>> >> > > On Wed, 16 Mar 2011 1=
3:54:39 -0500
>> >> > > Henry Sinnreich <henry.sinnreich@gmail.com> wrote:=


>> >> > >
>> >> > >> Hi Cullen,>> >> > >>
>> >> > >> Sectio=
n 3. Requirements says:
>> >> > >> ... In addition,=
 there is a strong interest to define a TURN-like

>> protocol
>> >> > >> that looks like HTTP t=
o intermediaries, so that media can be
>> tunneled
>> >=
;> > >> over HTTP.
>> >> > >> ...

>> >> > >>
>> >> > >> This is =
very interesting and innovative.
>> >> > >> Can you=
 explain here a little bit?
>> >> > >>
>> =
>> > >> Thanks, Henry

>> >> > >>
>> >> > >
>> =
>> > >
>> >> > > We implemented something =
like that for our system, also documented in
>> a paper
>>=
; >> > > lately: basically a controller protocol that tunnels e=
very other

>> protocol on
>> >> > > top of HTTP to bypass r=
estrictive firewalls. Since it works, we were
>> also
>> =
>> > > thinking about a "lightweight" version of such=
 protocol to only

>> convey RTP
>> >> > > specifically for our web=
 interface, considering RTP is the only
>> protocol we
>>=
 >> > > couldn't wrap via JavaScript so far. So I agree tha=
t such a TURN-like

>> protocol
>> >> > > would be a good idea for t=
he WG.
>> >> > >
>> >> > > Of cou=
rse, there would be limitations and issues because of this
>> tran=
sport,

>> >> > > but since firewalls are there to stay, and few =
of them like RTP
>> appearing "out
>> >> > =
> of the blue", that's something we probably need to live with.=


>> >> > >
>> >> > > L.
>> &=
gt;> > >
>> >> > >
>> >> > =
>>
>> >> > >> On 3/10/11 2:20 PM, "Culle=
n Jennings" <fluffy@cisco.com> wrote:

>> >> > >>
>> >> > >>>
&=
gt;> >> > >>> I wrote up the start of a draft on requi=
rements.
>> >> > >>>
>> >> > &=
gt;>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00

>> >> > >>>
>> >> > >>> =
I think the key thing for the key thing for IETF working is
>> sta=
rting to
>> >> > >>> think
>> >> =
> >>> about the requirements and what on the on the wire protoc=
ols are

>> that we
>> >> > >>> want
>> &g=
t;> > >>> to select. I included some API definition because =
I found it
>> helpful to be
>> >> > >>>=
 thinking about that in this draft but that would not be part of the

>> IETF
>> >> > >>> work.
>> >=
> > >>> That would be work I think W3C should do. The draft =
also touches on
>> the of
>> >> > >>> t=
he

>> >> > >>> security issues for media and signaling=
 originating from browsers.
>> >> > >>>
>&=
gt; >> > >>> Cullen
>> >> > >>>=
;

>> >> > >>>
>> >> > >>><=
br>>> >> > >>>
>> >> > >>&g=
t; _______________________________________________
>> >> >=
; >>> dispatch mailing list

>> >> > >>> di=
spatch@ietf.org
>> >> > >>> https://www.ietf.org/mailman/lis=
tinfo/dispatch

>> >> > >>
>> >> > >>
>&=
gt; >> > >> _______________________________________________<=
br>>> >> > >> dispatch mailing list
>> >&g=
t; > >> dispatch@ietf.org=


>> >> > >> https://www.ietf.org/mailman/listinfo/dispatch
>&=
gt; >> > >
>> >> >
>> >> ><=
br>
>> >> >
>> >>
>> >>
>>=
; >> --
>> >> Lorenzo Miniero, COB
>> >>=
;
>> >> Meetecho s.r.l.
>> >> Web Conferencin=
g and Collaboration Tools

>> >> http://www.meetecho.c=
om
>> >> _______________________________________________=

>> >> dispatch mailing list
>> >> dispatch@ietf.org

>> >> https://www.ietf.org/mailman/listinfo/dispatch
>> >
&g=
t;> >
>> > --
>> > Lorenzo Miniero, COB

>> >
>> > Meetecho s.r.l.
>> > Web Confere=
ncing and Collaboration Tools
>> > http://www.meetecho.com
>> > ______________________=
_________________________

>> > dispatch mailing list
>> > dispatch@ietf.org
>> > https://www.ietf.org/mailman/listinfo=
/dispatch

> 
> 
> -- 
> Lorenzo Miniero, COB
> 
> Me=
etecho s.r.l.
> Web Conferencing and Collaboration Tools
> http://www.meetecho.com


--000e0cd5ccbaa6b270049f89e2c5--

From spromano@unina.it  Mon Mar 28 05:27:10 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A8E4F3A691A for ; Mon, 28 Mar 2011 05:27:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -100.717
X-Spam-Level: 
X-Spam-Status: No, score=-100.717 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245, HTML_MESSAGE=0.001, NORMAL_HTTP_TO_IP=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 402HKFd4oDMQ for ; Mon, 28 Mar 2011 05:27:08 -0700 (PDT)
Received: from smtp1.unina.it (smtp1.unina.it [192.132.34.61]) by core3.amsl.com (Postfix) with ESMTP id A5C4A3A687D for ; Mon, 28 Mar 2011 05:27:07 -0700 (PDT)
Received: from [130.129.23.172] (dhcp-17ac.meeting.ietf.org [130.129.23.172]) (authenticated bits=0) by smtp1.unina.it (8.14.4/8.14.4) with ESMTP id p2SCSgtG003056 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Mon, 28 Mar 2011 14:28:43 +0200
Message-ID: <4D907EF3.9040807@unina.it>
Date: Mon, 28 Mar 2011 14:28:35 +0200
From: Simon Pietro Romano 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; it; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Dean Willis 
References: 	<20110324175217.71812238@lminiero-acer>	<20110326174504.56061fab@lminiero-acer>		<20110328140335.4ff34ddb@lminiero-acer> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------090003080406020300060403"
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API	draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 12:27:10 -0000

This is a multi-part message in MIME format.
--------------090003080406020300060403
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit

Hi Dean,

we made a number of tests with an "off-the-shelf" ADSL connection (both 
in an un-congested and in a congested network). Jitter was between 80 
and 100 msec in the former case; in the latter case (congested network) 
it was much more variable, with an average of around 105 msec and peaks 
between 60 and 160 msec (which clearly show a dependence on TCP 
congestion control mechanisms).

Simon

In the former case (no congestion in the network), jitter

Il 28/03/2011 14:14, Dean Willis ha scritto:
>
> It is not the bandwidth that worries me. It is packet loss and 
> resulting tcp retransmit jitter.
>
> On Mar 28, 2011 2:08 PM, "Lorenzo Miniero"  > wrote:
> > Hi Dean,
> >
> >
> > On Mon, 28 Mar 2011 02:33:25 -0500
> > Dean Willis  > wrote:
> >
> >> I suggest trying it from a pc tethered to an EDGE mobile modem.
> >
> >
> > Or with no network at all :)
> >
> > Jokes apart, it's clear that the QoE can't benefit from such limited 
> bandwidth/conditions, and that would be true for pure RTP as well. I 
> can only say that we've used this and older versions of the tunnel on 
> limited networks (home adsls, umts, etc.) for our purposes and it 
> seemed fine, though I agree that every attempt at addressing this 
> feature should take care of tackling such issues.
> >
> > L.
> >
> >
> >
> >> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero"  > wrote:
> >> > Dear all,
> >> >
> >> > as anticipated, I've just made available a prototype version of our
> >> RTP-over-HTTP tunneling approach:
> >> >
> >> > http://143.225.229.136/WebLite/echotest.jsp
> >> >
> >> > It is a very simple Echo Test: the page uses JavaScript to setup a
> >> server-originated SIP call to a URI associated with an RTP echo 
> application.
> >> Then, a Java applet is started to take care of the the RTP 
> tunneling and
> >> encoding/decoding process.
> >> >
> >> > To play with it, just enable the "Enable RTP tunneling" checkbox 
> and then
> >> click the button to launch the test. If everything goes fine, a new box
> >> should appear on the left with the remote audio and video streams: 
> a voice
> >> prompt with a video logo will precede the actual echo feature. All 
> the RTP
> >> packets will be tunneled over HTTP.
> >> >
> >> > Before you start, nevertheless, please have a quick look at a few
> >> prerequisites and instructions here:
> >> >
> >> > http://ietf.conf.meetecho.com/index.php/Web_Client
> >> >
> >> > for what concerns the applet-based Audio/Video functionality 
> (last section
> >> in the page: our Web Client uses the same applet). In fact, applets 
> can give
> >> a few problems at times (especially for multimedia-driven 
> applications) in
> >> some configurations, and so checking a few things before delving in 
> the demo
> >> may spare you from headache! :)
> >> >
> >> > Let us know if and how it works. Of course please beware that 
> this is just
> >> a very dumb prototype, so rough that it can give you the creeps if 
> you just
> >> skim over it! :)
> >> > Nevertheless, the basic concepts are there, so it may be useful 
> to foster
> >> some discussion on the matter. If you're interested in some more 
> details,
> >> I'm already in Prague.
> >> >
> >> > Lorenzo
> >> >
> >> >
> >> >
> >> > On Thu, 24 Mar 2011 17:52:17 +0100
> >> > Lorenzo Miniero  > wrote:
> >> >
> >> >> Hi Henry,
> >> >>
> >> >>
> >> >> On Wed, 23 Mar 2011 11:25:07 -0500
> >> >> Henry Sinnreich  > wrote:
> >> >>
> >> >> > Hi Lorenzo, Cullen,
> >> >> >
> >> >> > > We implemented something like that for our system, also 
> documented in
> >> a paper
> >> >> > > lately: basically a controller protocol that tunnels every other
> >> protocol on
> >> >> > > top of HTTP to bypass restrictive firewalls.
> >> >> >
> >> >> > Having explained HTTP tunneling flavors in the paper and also 
> shown
> >> >> > measurement results, besides a working product, should really 
> make IMO
> >> your
> >> >> > approach or something similar to it an option for the RTC-Web 
> media
> >> >> > solution. Such an option should also be reflected in the API 
> part, on
> >> which
> >> >> > the draft-jennings-rtcweb-api is focused.
> >> >> >
> >> >>
> >> >>
> >> >> I'd be glad to contribute to the definition of a related 
> solution, should
> >> such a work be started within the IETF.
> >> >>
> >> >> I'll also try and make available, hopefully before the meeting, a
> >> publicly reachable prototype version of our current approach for 
> you all to
> >> play with, just to have some live experience to foster discussion.
> >> >>
> >> >> L.
> >> >>
> >> >>
> >> >> > RT media over HTTP may be quite effective in some usage 
> scenarios, such
> >> as
> >> >> > CDN with >100k HTPP proxies around the globe, collocated in 
> 1,000s of
> >> >> > networks with plenty of bandwidth, smaller delay and arguably 
> without
> >> all
> >> >> > the other SIP, etc. intermediaries found inside and between most
> >> service
> >> >> > provider networks.
> >> >> >
> >> >> > I know profiles in standards are not desirable, but hindering
> >> innovations
> >> >> > such as yours and ignoring CDN and other Web application 
> scenarios is
> >> even
> >> >> > less desirable.
> >> >> > This is a challenge both the IETF and the W3C will hopefully 
> manage.
> >> >> >
> >> >> > Thanks for sharing,
> >> >> >
> >> >> > Henry
> >> >> >
> >> >> >
> >> >> > On 3/16/11 3:12 PM, "Lorenzo Miniero"  > wrote:
> >> >> >
> >> >> > > On Wed, 16 Mar 2011 13:54:39 -0500
> >> >> > > Henry Sinnreich  > wrote:
> >> >> > >
> >> >> > >> Hi Cullen,
> >> >> > >>
> >> >> > >> Section 3. Requirements says:
> >> >> > >> ... In addition, there is a strong interest to define a 
> TURN-like
> >> protocol
> >> >> > >> that looks like HTTP to intermediaries, so that media can be
> >> tunneled
> >> >> > >> over HTTP.
> >> >> > >> ...
> >> >> > >>
> >> >> > >> This is very interesting and innovative.
> >> >> > >> Can you explain here a little bit?
> >> >> > >>
> >> >> > >> Thanks, Henry
> >> >> > >>
> >> >> > >
> >> >> > >
> >> >> > > We implemented something like that for our system, also 
> documented in
> >> a paper
> >> >> > > lately: basically a controller protocol that tunnels every other
> >> protocol on
> >> >> > > top of HTTP to bypass restrictive firewalls. Since it works, 
> we were
> >> also
> >> >> > > thinking about a "lightweight" version of such protocol to only
> >> convey RTP
> >> >> > > specifically for our web interface, considering RTP is the only
> >> protocol we
> >> >> > > couldn't wrap via JavaScript so far. So I agree that such a 
> TURN-like
> >> protocol
> >> >> > > would be a good idea for the WG.
> >> >> > >
> >> >> > > Of course, there would be limitations and issues because of this
> >> transport,
> >> >> > > but since firewalls are there to stay, and few of them like RTP
> >> appearing "out
> >> >> > > of the blue", that's something we probably need to live with.
> >> >> > >
> >> >> > > L.
> >> >> > >
> >> >> > >
> >> >> > >>
> >> >> > >> On 3/10/11 2:20 PM, "Cullen Jennings"  > wrote:
> >> >> > >>
> >> >> > >>>
> >> >> > >>> I wrote up the start of a draft on requirements.
> >> >> > >>>
> >> >> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
> >> >> > >>>
> >> >> > >>> I think the key thing for the key thing for IETF working is
> >> starting to
> >> >> > >>> think
> >> >> > >>> about the requirements and what on the on the wire 
> protocols are
> >> that we
> >> >> > >>> want
> >> >> > >>> to select. I included some API definition because I found it
> >> helpful to be
> >> >> > >>> thinking about that in this draft but that would not be 
> part of the
> >> IETF
> >> >> > >>> work.
> >> >> > >>> That would be work I think W3C should do. The draft also 
> touches on
> >> the of
> >> >> > >>> the
> >> >> > >>> security issues for media and signaling originating from 
> browsers.
> >> >> > >>>
> >> >> > >>> Cullen
> >> >> > >>>
> >> >> > >>>
> >> >> > >>>
> >> >> > >>>
> >> >> > >>> _______________________________________________
> >> >> > >>> dispatch mailing list
> >> >> > >>> dispatch@ietf.org 
> >> >> > >>> https://www.ietf.org/mailman/listinfo/dispatch
> >> >> > >>
> >> >> > >>
> >> >> > >> _______________________________________________
> >> >> > >> dispatch mailing list
> >> >> > >> dispatch@ietf.org 
> >> >> > >> https://www.ietf.org/mailman/listinfo/dispatch
> >> >> > >
> >> >> >
> >> >> >
> >> >> >
> >> >>
> >> >>
> >> >> --
> >> >> Lorenzo Miniero, COB
> >> >>
> >> >> Meetecho s.r.l.
> >> >> Web Conferencing and Collaboration Tools
> >> >> http://www.meetecho.com
> >> >> _______________________________________________
> >> >> dispatch mailing list
> >> >> dispatch@ietf.org 
> >> >> https://www.ietf.org/mailman/listinfo/dispatch
> >> >
> >> >
> >> > --
> >> > Lorenzo Miniero, COB
> >> >
> >> > Meetecho s.r.l.
> >> > Web Conferencing and Collaboration Tools
> >> > http://www.meetecho.com
> >> > _______________________________________________
> >> > dispatch mailing list
> >> > dispatch@ietf.org 
> >> > https://www.ietf.org/mailman/listinfo/dispatch
> >
> >
> > --
> > Lorenzo Miniero, COB
> >
> > Meetecho s.r.l.
> > Web Conferencing and Collaboration Tools
> > http://www.meetecho.com
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
                             _\\|//_
                             ( O-O )
    ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~
                     Simon Pietro Romano
               Universita' di Napoli Federico II
                  Computer Science Department
         Phone: +39 081 7683823 -- Fax: +39 081 7684219
                 e-mail: spromano@unina.it
           http://www.comics.unina.it/simonpietro.romano

     <>. Magritte.
                          oooO
    ~~~~~~~~~~~~~~~~~~~~~~(   )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
                           \ (    (   )
                            \_)    ) /
                                  (_/



--------------090003080406020300060403
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Hi Dean,

    

    we made a number of tests with an "off-the-shelf" ADSL connection
    (both in an un-congested and in a congested network). Jitter was
    between 80 and 100 msec in the former case; in the latter case
    (congested network) it was much more variable, with an average of
    around 105 msec and peaks between 60 and 160 msec (which clearly
    show a dependence on TCP congestion control mechanisms). 

    

    Simon

    

    In the former case (no congestion in the network), jitter 

    

    Il 28/03/2011 14:14, Dean Willis ha scritto:
    

      
It is not the bandwidth that worries me. It is packet loss and
        resulting tcp retransmit jitter.

      
On Mar 28, 2011 2:08 PM, "Lorenzo
        Miniero" <lorenzo@meetecho.com>
        wrote:

        > Hi Dean,

        > 

        > 

        > On Mon, 28 Mar 2011 02:33:25 -0500

        > Dean Willis <dean.willis@softarmor.com>
        wrote:

        > 

        >> I suggest trying it from a pc tethered to an EDGE
        mobile modem.

        > 

        > 

        > Or with no network at all :)

        > 

        > Jokes apart, it's clear that the QoE can't benefit from
        such limited bandwidth/conditions, and that would be true for
        pure RTP as well. I can only say that we've used this and older
        versions of the tunnel on limited networks (home adsls, umts,
        etc.) for our purposes and it seemed fine, though I agree that
        every attempt at addressing this feature should take care of
        tackling such issues.

        > 

        > L.

        > 

        > 

        > 

        >> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero" <lorenzo@meetecho.com>
        wrote:

        >> > Dear all,

        >> >

        >> > as anticipated, I've just made available a
        prototype version of our

        >> RTP-over-HTTP tunneling approach:

        >> >

        >> > http://143.225.229.136/WebLite/echotest.jsp

        >> >

        >> > It is a very simple Echo Test: the page uses
        JavaScript to setup a

        >> server-originated SIP call to a URI associated with an
        RTP echo application.

        >> Then, a Java applet is started to take care of the the
        RTP tunneling and

        >> encoding/decoding process.

        >> >

        >> > To play with it, just enable the "Enable RTP
        tunneling" checkbox and then

        >> click the button to launch the test. If everything goes
        fine, a new box

        >> should appear on the left with the remote audio and
        video streams: a voice

        >> prompt with a video logo will precede the actual echo
        feature. All the RTP

        >> packets will be tunneled over HTTP.

        >> >

        >> > Before you start, nevertheless, please have a
        quick look at a few

        >> prerequisites and instructions here:

        >> >

        >> > http://ietf.conf.meetecho.com/index.php/Web_Client

        >> >

        >> > for what concerns the applet-based Audio/Video
        functionality (last section

        >> in the page: our Web Client uses the same applet). In
        fact, applets can give

        >> a few problems at times (especially for
        multimedia-driven applications) in

        >> some configurations, and so checking a few things
        before delving in the demo

        >> may spare you from headache! :)

        >> >

        >> > Let us know if and how it works. Of course please
        beware that this is just

        >> a very dumb prototype, so rough that it can give you
        the creeps if you just

        >> skim over it! :)

        >> > Nevertheless, the basic concepts are there, so it
        may be useful to foster

        >> some discussion on the matter. If you're interested in
        some more details,

        >> I'm already in Prague.

        >> >

        >> > Lorenzo

        >> >

        >> >

        >> >

        >> > On Thu, 24 Mar 2011 17:52:17 +0100

        >> > Lorenzo Miniero <lorenzo@meetecho.com>
        wrote:

        >> >

        >> >> Hi Henry,

        >> >>

        >> >>

        >> >> On Wed, 23 Mar 2011 11:25:07 -0500

        >> >> Henry Sinnreich <henry.sinnreich@gmail.com>
        wrote:

        >> >>

        >> >> > Hi Lorenzo, Cullen,

        >> >> >

        >> >> > > We implemented something like that
        for our system, also documented in

        >> a paper

        >> >> > > lately: basically a controller
        protocol that tunnels every other

        >> protocol on

        >> >> > > top of HTTP to bypass restrictive
        firewalls.

        >> >> >

        >> >> > Having explained HTTP tunneling flavors
        in the paper and also shown

        >> >> > measurement results, besides a working
        product, should really make IMO

        >> your

        >> >> > approach or something similar to it an
        option for the RTC-Web media

        >> >> > solution. Such an option should also be
        reflected in the API part, on

        >> which

        >> >> > the draft-jennings-rtcweb-api is focused.

        >> >> >

        >> >>

        >> >>

        >> >> I'd be glad to contribute to the definition of
        a related solution, should

        >> such a work be started within the IETF.

        >> >>

        >> >> I'll also try and make available, hopefully
        before the meeting, a

        >> publicly reachable prototype version of our current
        approach for you all to

        >> play with, just to have some live experience to foster
        discussion.

        >> >>

        >> >> L.

        >> >>

        >> >>

        >> >> > RT media over HTTP may be quite effective
        in some usage scenarios, such

        >> as

        >> >> > CDN with >100k HTPP proxies around the
        globe, collocated in 1,000s of

        >> >> > networks with plenty of bandwidth,
        smaller delay and arguably without

        >> all

        >> >> > the other SIP, etc. intermediaries found
        inside and between most

        >> service

        >> >> > provider networks.

        >> >> >

        >> >> > I know profiles in standards are not
        desirable, but hindering

        >> innovations

        >> >> > such as yours and ignoring CDN and other
        Web application scenarios is

        >> even

        >> >> > less desirable.

        >> >> > This is a challenge both the IETF and the
        W3C will hopefully manage.

        >> >> >

        >> >> > Thanks for sharing,

        >> >> >

        >> >> > Henry

        >> >> >

        >> >> >

        >> >> > On 3/16/11 3:12 PM, "Lorenzo Miniero"
        <lorenzo@meetecho.com>
        wrote:

        >> >> >

        >> >> > > On Wed, 16 Mar 2011 13:54:39 -0500

        >> >> > > Henry Sinnreich <henry.sinnreich@gmail.com>
        wrote:

        >> >> > >

        >> >> > >> Hi Cullen,

        >> >> > >>

        >> >> > >> Section 3. Requirements says:

        >> >> > >> ... In addition, there is a
        strong interest to define a TURN-like

        >> protocol

        >> >> > >> that looks like HTTP to
        intermediaries, so that media can be

        >> tunneled

        >> >> > >> over HTTP.

        >> >> > >> ...

        >> >> > >>

        >> >> > >> This is very interesting and
        innovative.

        >> >> > >> Can you explain here a little
        bit?

        >> >> > >>

        >> >> > >> Thanks, Henry

        >> >> > >>

        >> >> > >

        >> >> > >

        >> >> > > We implemented something like that
        for our system, also documented in

        >> a paper

        >> >> > > lately: basically a controller
        protocol that tunnels every other

        >> protocol on

        >> >> > > top of HTTP to bypass restrictive
        firewalls. Since it works, we were

        >> also

        >> >> > > thinking about a "lightweight"
        version of such protocol to only

        >> convey RTP

        >> >> > > specifically for our web interface,
        considering RTP is the only

        >> protocol we

        >> >> > > couldn't wrap via JavaScript so far.
        So I agree that such a TURN-like

        >> protocol

        >> >> > > would be a good idea for the WG.

        >> >> > >

        >> >> > > Of course, there would be
        limitations and issues because of this

        >> transport,

        >> >> > > but since firewalls are there to
        stay, and few of them like RTP

        >> appearing "out

        >> >> > > of the blue", that's something we
        probably need to live with.

        >> >> > >

        >> >> > > L.

        >> >> > >

        >> >> > >

        >> >> > >>

        >> >> > >> On 3/10/11 2:20 PM, "Cullen
        Jennings" <fluffy@cisco.com> wrote:

        >> >> > >>

        >> >> > >>>

        >> >> > >>> I wrote up the start of a
        draft on requirements.

        >> >> > >>>

        >> >> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00

        >> >> > >>>

        >> >> > >>> I think the key thing for
        the key thing for IETF working is

        >> starting to

        >> >> > >>> think

        >> >> > >>> about the requirements and
        what on the on the wire protocols are

        >> that we

        >> >> > >>> want

        >> >> > >>> to select. I included some
        API definition because I found it

        >> helpful to be

        >> >> > >>> thinking about that in this
        draft but that would not be part of the

        >> IETF

        >> >> > >>> work.

        >> >> > >>> That would be work I think
        W3C should do. The draft also touches on

        >> the of

        >> >> > >>> the

        >> >> > >>> security issues for media
        and signaling originating from browsers.

        >> >> > >>>

        >> >> > >>> Cullen

        >> >> > >>>

        >> >> > >>>

        >> >> > >>>

        >> >> > >>>

        >> >> > >>>
        _______________________________________________

        >> >> > >>> dispatch mailing list

        >> >> > >>> dispatch@ietf.org

        >> >> > >>> https://www.ietf.org/mailman/listinfo/dispatch

        >> >> > >>

        >> >> > >>

        >> >> > >>
        _______________________________________________

        >> >> > >> dispatch mailing list

        >> >> > >> dispatch@ietf.org

        >> >> > >> https://www.ietf.org/mailman/listinfo/dispatch

        >> >> > >

        >> >> >

        >> >> >

        >> >> >

        >> >>

        >> >>

        >> >> --

        >> >> Lorenzo Miniero, COB

        >> >>

        >> >> Meetecho s.r.l.

        >> >> Web Conferencing and Collaboration Tools

        >> >> http://www.meetecho.com

        >> >>
        _______________________________________________

        >> >> dispatch mailing list

        >> >> dispatch@ietf.org

        >> >> https://www.ietf.org/mailman/listinfo/dispatch

        >> >

        >> >

        >> > --

        >> > Lorenzo Miniero, COB

        >> >

        >> > Meetecho s.r.l.

        >> > Web Conferencing and Collaboration Tools

        >> > http://www.meetecho.com

        >> > _______________________________________________

        >> > dispatch mailing list

        >> > dispatch@ietf.org

        >> > https://www.ietf.org/mailman/listinfo/dispatch

        > 

        > 

        > -- 

        > Lorenzo Miniero, COB

        > 

        > Meetecho s.r.l.

        > Web Conferencing and Collaboration Tools

        > http://www.meetecho.com

      

      

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 
                            _\\|//_
                            ( O-O )
   ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~
                    Simon Pietro Romano
              Universita' di Napoli Federico II
                 Computer Science Department 
        Phone: +39 081 7683823 -- Fax: +39 081 7684219
                e-mail: spromano@unina.it
          http://www.comics.unina.it/simonpietro.romano

    <<Molti mi dicono che lo scoraggiamento è l'alibi degli 
   idioti. Ci rifletto un istante; e mi scoraggio>>. Magritte.
                         oooO
   ~~~~~~~~~~~~~~~~~~~~~~(   )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
                          \ (    (   )
                           \_)    ) /
                                 (_/



  


--------------090003080406020300060403--

From dean.willis@softarmor.com  Mon Mar 28 07:47:57 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C7F953A68F3 for ; Mon, 28 Mar 2011 07:47:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.975
X-Spam-Level: 
X-Spam-Status: No, score=-102.975 tagged_above=-999 required=5 tests=[AWL=-0.000, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, NORMAL_HTTP_TO_IP=0.001, RCVD_IN_DNSWL_LOW=-1, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id st0g3YqhpRn1 for ; Mon, 28 Mar 2011 07:47:55 -0700 (PDT)
Received: from mail-pv0-f172.google.com (mail-pv0-f172.google.com [74.125.83.172]) by core3.amsl.com (Postfix) with ESMTP id 6A9203A6943 for ; Mon, 28 Mar 2011 07:47:55 -0700 (PDT)
Received: by pve39 with SMTP id 39so712044pve.31 for ; Mon, 28 Mar 2011 07:49:32 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.143.27.13 with SMTP id e13mr3582610wfj.336.1301323772607; Mon, 28 Mar 2011 07:49:32 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 07:49:32 -0700 (PDT)
Received: by 10.142.134.3 with HTTP; Mon, 28 Mar 2011 07:49:32 -0700 (PDT)
In-Reply-To: <4D907EF3.9040807@unina.it>
References:  <20110324175217.71812238@lminiero-acer> <20110326174504.56061fab@lminiero-acer>  <20110328140335.4ff34ddb@lminiero-acer>  <4D907EF3.9040807@unina.it>
Date: Mon, 28 Mar 2011 09:49:32 -0500
Message-ID: 
From: Dean Willis 
To: Simon Pietro Romano 
Content-Type: multipart/alternative; boundary=00504502c5c6343cf0049f8c0db0
Cc: DISPATCH list 
Subject: Re: [dispatch] New draft on RTCWeb Requirements and API draft-jennings-rtcweb-api
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 14:47:57 -0000

--00504502c5c6343cf0049f8c0db0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

What was the packet loss rate? Were they randomly distributed single-packet
losses or contiguos burst losses? RTP/UDP with simple CODECs is surprisingl=
y
useful at 1% losses with burst losses up to 3 sequential packets.
On Mar 28, 2011 2:28 PM, "Simon Pietro Romano"  wrote:
> Hi Dean,
>
> we made a number of tests with an "off-the-shelf" ADSL connection (both
> in an un-congested and in a congested network). Jitter was between 80
> and 100 msec in the former case; in the latter case (congested network)
> it was much more variable, with an average of around 105 msec and peaks
> between 60 and 160 msec (which clearly show a dependence on TCP
> congestion control mechanisms).
>
> Simon
>
> In the former case (no congestion in the network), jitter
>
> Il 28/03/2011 14:14, Dean Willis ha scritto:
>>
>> It is not the bandwidth that worries me. It is packet loss and
>> resulting tcp retransmit jitter.
>>
>> On Mar 28, 2011 2:08 PM, "Lorenzo Miniero" > > wrote:
>> > Hi Dean,
>> >
>> >
>> > On Mon, 28 Mar 2011 02:33:25 -0500
>> > Dean Willis > > wrote:
>> >
>> >> I suggest trying it from a pc tethered to an EDGE mobile modem.
>> >
>> >
>> > Or with no network at all :)
>> >
>> > Jokes apart, it's clear that the QoE can't benefit from such limited
>> bandwidth/conditions, and that would be true for pure RTP as well. I
>> can only say that we've used this and older versions of the tunnel on
>> limited networks (home adsls, umts, etc.) for our purposes and it
>> seemed fine, though I agree that every attempt at addressing this
>> feature should take care of tackling such issues.
>> >
>> > L.
>> >
>> >
>> >
>> >> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero" > > wrote:
>> >> > Dear all,
>> >> >
>> >> > as anticipated, I've just made available a prototype version of our
>> >> RTP-over-HTTP tunneling approach:
>> >> >
>> >> > http://143.225.229.136/WebLite/echotest.jsp
>> >> >
>> >> > It is a very simple Echo Test: the page uses JavaScript to setup a
>> >> server-originated SIP call to a URI associated with an RTP echo
>> application.
>> >> Then, a Java applet is started to take care of the the RTP
>> tunneling and
>> >> encoding/decoding process.
>> >> >
>> >> > To play with it, just enable the "Enable RTP tunneling" checkbox
>> and then
>> >> click the button to launch the test. If everything goes fine, a new
box
>> >> should appear on the left with the remote audio and video streams:
>> a voice
>> >> prompt with a video logo will precede the actual echo feature. All
>> the RTP
>> >> packets will be tunneled over HTTP.
>> >> >
>> >> > Before you start, nevertheless, please have a quick look at a few
>> >> prerequisites and instructions here:
>> >> >
>> >> > http://ietf.conf.meetecho.com/index.php/Web_Client
>> >> >
>> >> > for what concerns the applet-based Audio/Video functionality
>> (last section
>> >> in the page: our Web Client uses the same applet). In fact, applets
>> can give
>> >> a few problems at times (especially for multimedia-driven
>> applications) in
>> >> some configurations, and so checking a few things before delving in
>> the demo
>> >> may spare you from headache! :)
>> >> >
>> >> > Let us know if and how it works. Of course please beware that
>> this is just
>> >> a very dumb prototype, so rough that it can give you the creeps if
>> you just
>> >> skim over it! :)
>> >> > Nevertheless, the basic concepts are there, so it may be useful
>> to foster
>> >> some discussion on the matter. If you're interested in some more
>> details,
>> >> I'm already in Prague.
>> >> >
>> >> > Lorenzo
>> >> >
>> >> >
>> >> >
>> >> > On Thu, 24 Mar 2011 17:52:17 +0100
>> >> > Lorenzo Miniero > > wrote:
>> >> >
>> >> >> Hi Henry,
>> >> >>
>> >> >>
>> >> >> On Wed, 23 Mar 2011 11:25:07 -0500
>> >> >> Henry Sinnreich > > wrote:
>> >> >>
>> >> >> > Hi Lorenzo, Cullen,
>> >> >> >
>> >> >> > > We implemented something like that for our system, also
>> documented in
>> >> a paper
>> >> >> > > lately: basically a controller protocol that tunnels every
other
>> >> protocol on
>> >> >> > > top of HTTP to bypass restrictive firewalls.
>> >> >> >
>> >> >> > Having explained HTTP tunneling flavors in the paper and also
>> shown
>> >> >> > measurement results, besides a working product, should really
>> make IMO
>> >> your
>> >> >> > approach or something similar to it an option for the RTC-Web
>> media
>> >> >> > solution. Such an option should also be reflected in the API
>> part, on
>> >> which
>> >> >> > the draft-jennings-rtcweb-api is focused.
>> >> >> >
>> >> >>
>> >> >>
>> >> >> I'd be glad to contribute to the definition of a related
>> solution, should
>> >> such a work be started within the IETF.
>> >> >>
>> >> >> I'll also try and make available, hopefully before the meeting, a
>> >> publicly reachable prototype version of our current approach for
>> you all to
>> >> play with, just to have some live experience to foster discussion.
>> >> >>
>> >> >> L.
>> >> >>
>> >> >>
>> >> >> > RT media over HTTP may be quite effective in some usage
>> scenarios, such
>> >> as
>> >> >> > CDN with >100k HTPP proxies around the globe, collocated in
>> 1,000s of
>> >> >> > networks with plenty of bandwidth, smaller delay and arguably
>> without
>> >> all
>> >> >> > the other SIP, etc. intermediaries found inside and between most
>> >> service
>> >> >> > provider networks.
>> >> >> >
>> >> >> > I know profiles in standards are not desirable, but hindering
>> >> innovations
>> >> >> > such as yours and ignoring CDN and other Web application
>> scenarios is
>> >> even
>> >> >> > less desirable.
>> >> >> > This is a challenge both the IETF and the W3C will hopefully
>> manage.
>> >> >> >
>> >> >> > Thanks for sharing,
>> >> >> >
>> >> >> > Henry
>> >> >> >
>> >> >> >
>> >> >> > On 3/16/11 3:12 PM, "Lorenzo Miniero" > > wrote:
>> >> >> >
>> >> >> > > On Wed, 16 Mar 2011 13:54:39 -0500
>> >> >> > > Henry Sinnreich > > wrote:
>> >> >> > >
>> >> >> > >> Hi Cullen,
>> >> >> > >>
>> >> >> > >> Section 3. Requirements says:
>> >> >> > >> ... In addition, there is a strong interest to define a
>> TURN-like
>> >> protocol
>> >> >> > >> that looks like HTTP to intermediaries, so that media can be
>> >> tunneled
>> >> >> > >> over HTTP.
>> >> >> > >> ...
>> >> >> > >>
>> >> >> > >> This is very interesting and innovative.
>> >> >> > >> Can you explain here a little bit?
>> >> >> > >>
>> >> >> > >> Thanks, Henry
>> >> >> > >>
>> >> >> > >
>> >> >> > >
>> >> >> > > We implemented something like that for our system, also
>> documented in
>> >> a paper
>> >> >> > > lately: basically a controller protocol that tunnels every
other
>> >> protocol on
>> >> >> > > top of HTTP to bypass restrictive firewalls. Since it works,
>> we were
>> >> also
>> >> >> > > thinking about a "lightweight" version of such protocol to onl=
y
>> >> convey RTP
>> >> >> > > specifically for our web interface, considering RTP is the onl=
y
>> >> protocol we
>> >> >> > > couldn't wrap via JavaScript so far. So I agree that such a
>> TURN-like
>> >> protocol
>> >> >> > > would be a good idea for the WG.
>> >> >> > >
>> >> >> > > Of course, there would be limitations and issues because of
this
>> >> transport,
>> >> >> > > but since firewalls are there to stay, and few of them like RT=
P
>> >> appearing "out
>> >> >> > > of the blue", that's something we probably need to live with.
>> >> >> > >
>> >> >> > > L.
>> >> >> > >
>> >> >> > >
>> >> >> > >>
>> >> >> > >> On 3/10/11 2:20 PM, "Cullen Jennings" > > wrote:
>> >> >> > >>
>> >> >> > >>>
>> >> >> > >>> I wrote up the start of a draft on requirements.
>> >> >> > >>>
>> >> >> > >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00
>> >> >> > >>>
>> >> >> > >>> I think the key thing for the key thing for IETF working is
>> >> starting to
>> >> >> > >>> think
>> >> >> > >>> about the requirements and what on the on the wire
>> protocols are
>> >> that we
>> >> >> > >>> want
>> >> >> > >>> to select. I included some API definition because I found it
>> >> helpful to be
>> >> >> > >>> thinking about that in this draft but that would not be
>> part of the
>> >> IETF
>> >> >> > >>> work.
>> >> >> > >>> That would be work I think W3C should do. The draft also
>> touches on
>> >> the of
>> >> >> > >>> the
>> >> >> > >>> security issues for media and signaling originating from
>> browsers.
>> >> >> > >>>
>> >> >> > >>> Cullen
>> >> >> > >>>
>> >> >> > >>>
>> >> >> > >>>
>> >> >> > >>>
>> >> >> > >>> _______________________________________________
>> >> >> > >>> dispatch mailing list
>> >> >> > >>> dispatch@ietf.org 
>> >> >> > >>> https://www.ietf.org/mailman/listinfo/dispatch
>> >> >> > >>
>> >> >> > >>
>> >> >> > >> _______________________________________________
>> >> >> > >> dispatch mailing list
>> >> >> > >> dispatch@ietf.org 
>> >> >> > >> https://www.ietf.org/mailman/listinfo/dispatch
>> >> >> > >
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >>
>> >> >>
>> >> >> --
>> >> >> Lorenzo Miniero, COB
>> >> >>
>> >> >> Meetecho s.r.l.
>> >> >> Web Conferencing and Collaboration Tools
>> >> >> http://www.meetecho.com
>> >> >> _______________________________________________
>> >> >> dispatch mailing list
>> >> >> dispatch@ietf.org 
>> >> >> https://www.ietf.org/mailman/listinfo/dispatch
>> >> >
>> >> >
>> >> > --
>> >> > Lorenzo Miniero, COB
>> >> >
>> >> > Meetecho s.r.l.
>> >> > Web Conferencing and Collaboration Tools
>> >> > http://www.meetecho.com
>> >> > _______________________________________________
>> >> > dispatch mailing list
>> >> > dispatch@ietf.org 
>> >> > https://www.ietf.org/mailman/listinfo/dispatch
>> >
>> >
>> > --
>> > Lorenzo Miniero, COB
>> >
>> > Meetecho s.r.l.
>> > Web Conferencing and Collaboration Tools
>> > http://www.meetecho.com
>>
>>
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>
> --
> _\\|//_
> ( O-O )
> ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~
> Simon Pietro Romano
> Universita' di Napoli Federico II
> Computer Science Department
> Phone: +39 081 7683823 -- Fax: +39 081 7684219
> e-mail: spromano@unina.it
> http://www.comics.unina.it/simonpietro.romano
>
> < idioti. Ci rifletto un istante; e mi scoraggio>>. Magritte.
> oooO
> ~~~~~~~~~~~~~~~~~~~~~~( )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
> \ ( ( )
> \_) ) /
> (_/
>
>

--00504502c5c6343cf0049f8c0db0
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable


What was the packet loss rate? Were they randomly distributed single-pac=
ket losses or contiguos burst losses? RTP/UDP with simple CODECs is surpris=
ingly useful at 1% losses with burst losses up to 3 sequential packets.



On Mar 28, 2011 2:28 PM, "Simon Pietro Roma=
no" <spromano@unina.it>=
 wrote:
> Hi Dean,
> 
> we made a nu=
mber of tests with an "off-the-shelf" ADSL connection (both 

> in an un-congested and in a congested network). Jitter was between 80 =

> and 100 msec in the former case; in the latter case (congested net=
work) 
> it was much more variable, with an average of around 105 mse=
c and peaks 

> between 60 and 160 msec (which clearly show a dependence on TCP 
&g=
t; congestion control mechanisms).
> 
> Simon
> 
> =
In the former case (no congestion in the network), jitter
> 
> =
Il 28/03/2011 14:14, Dean Willis ha scritto:

>>
>> It is not the bandwidth that worries me. It is packet =
loss and 
>> resulting tcp retransmit jitter.
>>
>&=
gt; On Mar 28, 2011 2:08 PM, "Lorenzo Miniero" <lorenzo@meetecho.com 

>> <mailto:lorenzo@meetech=
o.com>> wrote:
>> > Hi Dean,
>> >
>=
> >
>> > On Mon, 28 Mar 2011 02:33:25 -0500
>> &=
gt; Dean Willis <dean.willi=
s@softarmor.com 

>> <mailto:dean.willi=
s@softarmor.com>> wrote:
>> >
>> >> I =
suggest trying it from a pc tethered to an EDGE mobile modem.
>> &=
gt;

>> >
>> > Or with no network at all :)
>> >=
;
>> > Jokes apart, it's clear that the QoE can't benef=
it from such limited 
>> bandwidth/conditions, and that would be t=
rue for pure RTP as well. I 

>> can only say that we've used this and older versions of the tu=
nnel on 
>> limited networks (home adsls, umts, etc.) for our purp=
oses and it 
>> seemed fine, though I agree that every attempt at =
addressing this 

>> feature should take care of tackling such issues.
>> >=

>> > L.
>> >
>> >
>> >
=
>> >> On Mar 26, 2011 4:49 PM, "Lorenzo Miniero" <=
lorenzo@meetecho.com 

>> <mailto:lorenzo@meetech=
o.com>> wrote:
>> >> > Dear all,
>> &g=
t;> >
>> >> > as anticipated, I've just made av=
ailable a prototype version of our

>> >> RTP-over-HTTP tunneling approach:
>> >> &g=
t;
>> >> > http://143.225.229.136/WebLite/echotest.jsp
>> >&=
gt; >

>> >> > It is a very simple Echo Test: the page uses JavaScr=
ipt to setup a
>> >> server-originated SIP call to a URI ass=
ociated with an RTP echo 
>> application.
>> >> The=
n, a Java applet is started to take care of the the RTP 

>> tunneling and
>> >> encoding/decoding process.
&=
gt;> >> >
>> >> > To play with it, just enabl=
e the "Enable RTP tunneling" checkbox 
>> and then

>> >> click the button to launch the test. If everything goes f=
ine, a new box
>> >> should appear on the left with the remo=
te audio and video streams: 
>> a voice
>> >> promp=
t with a video logo will precede the actual echo feature. All 

>> the RTP
>> >> packets will be tunneled over HTTP.>> >> >
>> >> > Before you start, nevert=
heless, please have a quick look at a few
>> >> prerequisite=
s and instructions here:

>> >> >
>> >> > http://ietf.conf.meetecho.com/index.ph=
p/Web_Client
>> >> >
>> >> > for wh=
at concerns the applet-based Audio/Video functionality 

>> (last section
>> >> in the page: our Web Client use=
s the same applet). In fact, applets 
>> can give
>> >=
> a few problems at times (especially for multimedia-driven 
>>=
 applications) in

>> >> some configurations, and so checking a few things before =
delving in 
>> the demo
>> >> may spare you from he=
adache! :)
>> >> >
>> >> > Let us know =
if and how it works. Of course please beware that 

>> this is just
>> >> a very dumb prototype, so rough =
that it can give you the creeps if 
>> you just
>> >&g=
t; skim over it! :)
>> >> > Nevertheless, the basic conce=
pts are there, so it may be useful 

>> to foster
>> >> some discussion on the matter. If y=
ou're interested in some more 
>> details,
>> >>=
; I'm already in Prague.
>> >> >
>> >>=
 > Lorenzo

>> >> >
>> >> >
>> >> ><=
br>>> >> > On Thu, 24 Mar 2011 17:52:17 +0100
>> &g=
t;> > Lorenzo Miniero <lor=
enzo@meetecho.com 

>> <mailto:lorenzo@meetech=
o.com>> wrote:
>> >> >
>> >> >=
;> Hi Henry,
>> >> >>
>> >> >>=


>> >> >> On Wed, 23 Mar 2011 11:25:07 -0500
>> &=
gt;> >> Henry Sinnreich <henry.sinnreich@gmail.com 
>> <mailto:henry.sinnreich@gmail.com>> wrote:=


>> >> >>
>> >> >> > Hi Lorenzo, C=
ullen,
>> >> >> >
>> >> >> >=
; > We implemented something like that for our system, also 
>>=
 documented in

>> >> a paper
>> >> >> > > lately: b=
asically a controller protocol that tunnels every other
>> >>=
; protocol on
>> >> >> > > top of HTTP to bypass=
 restrictive firewalls.

>> >> >> >
>> >> >> > Having e=
xplained HTTP tunneling flavors in the paper and also 
>> shown>> >> >> > measurement results, besides a working pro=
duct, should really 

>> make IMO
>> >> your
>> >> >> &=
gt; approach or something similar to it an option for the RTC-Web 
>&=
gt; media
>> >> >> > solution. Such an option shoul=
d also be reflected in the API 

>> part, on
>> >> which
>> >> >> =
> the draft-jennings-rtcweb-api is focused.
>> >> >>=
; >
>> >> >>
>> >> >>
>&=
gt; >> >> I'd be glad to contribute to the definition of a =
related 

>> solution, should
>> >> such a work be started withi=
n the IETF.
>> >> >>
>> >> >> I&#=
39;ll also try and make available, hopefully before the meeting, a
>&=
gt; >> publicly reachable prototype version of our current approach f=
or 

>> you all to
>> >> play with, just to have some live =
experience to foster discussion.
>> >> >>
>> =
>> >> L.
>> >> >>
>> >> >=
;>

>> >> >> > RT media over HTTP may be quite effective i=
n some usage 
>> scenarios, such
>> >> as
>&g=
t; >> >> > CDN with >100k HTPP proxies around the globe, =
collocated in 

>> 1,000s of
>> >> >> > networks with plenty =
of bandwidth, smaller delay and arguably 
>> without
>> &=
gt;> all
>> >> >> > the other SIP, etc. intermed=
iaries found inside and between most

>> >> service
>> >> >> > provider netwo=
rks.
>> >> >> >
>> >> >> > =
I know profiles in standards are not desirable, but hindering
>> &=
gt;> innovations

>> >> >> > such as yours and ignoring CDN and other We=
b application 
>> scenarios is
>> >> even
>&g=
t; >> >> > less desirable.
>> >> >> >=
; This is a challenge both the IETF and the W3C will hopefully 

>> manage.
>> >> >> >
>> >> &g=
t;> > Thanks for sharing,
>> >> >> >
>&=
gt; >> >> > Henry
>> >> >> >

>> >> >> >
>> >> >> > On 3/16/=
11 3:12 PM, "Lorenzo Miniero" <lorenzo@meetecho.com 
>> <mailto:lorenzo@meetecho.com>> wrote:

>> >> >> >
>> >> >> > > On =
Wed, 16 Mar 2011 13:54:39 -0500
>> >> >> > > Hen=
ry Sinnreich <henry.sinnrei=
ch@gmail.com 

>> <mailto:henry.sinn=
reich@gmail.com>> wrote:
>> >> >> > ><=
br>>> >> >> > >> Hi Cullen,
>> >>=
 >> > >>

>> >> >> > >> Section 3. Requirements says:
&=
gt;> >> >> > >> ... In addition, there is a strong =
interest to define a 
>> TURN-like
>> >> protocol
>> >> >> > >> that looks like HTTP to intermedia=
ries, so that media can be
>> >> tunneled
>> >&g=
t; >> > >> over HTTP.
>> >> >> > >=
;> ...

>> >> >> > >>
>> >> >> >=
 >> This is very interesting and innovative.
>> >> >=
;> > >> Can you explain here a little bit?
>> >>=
 >> > >>

>> >> >> > >> Thanks, Henry
>> >>=
 >> > >>
>> >> >> > >
>>=
 >> >> > >
>> >> >> > > We imp=
lemented something like that for our system, also 

>> documented in
>> >> a paper
>> >> &g=
t;> > > lately: basically a controller protocol that tunnels every=
 other
>> >> protocol on
>> >> >> > =
> top of HTTP to bypass restrictive firewalls. Since it works, 

>> we were
>> >> also
>> >> >> &g=
t; > thinking about a "lightweight" version of such protocol t=
o only
>> >> convey RTP
>> >> >> > &=
gt; specifically for our web interface, considering RTP is the only

>> >> protocol we
>> >> >> > > could=
n't wrap via JavaScript so far. So I agree that such a 
>> TUR=
N-like
>> >> protocol
>> >> >> > >=
; would be a good idea for the WG.

>> >> >> > >
>> >> >> > >=
; Of course, there would be limitations and issues because of this
>&=
gt; >> transport,
>> >> >> > > but since f=
irewalls are there to stay, and few of them like RTP

>> >> appearing "out
>> >> >> > &g=
t; of the blue", that's something we probably need to live with.>> >> >> > >
>> >> >> > &=
gt; L.

>> >> >> > >
>> >> >> > >=
;
>> >> >> > >>
>> >> >>=
 > >> On 3/10/11 2:20 PM, "Cullen Jennings" <fluffy@cisco.com 

>> <mailto:fluffy@cisco.com>> wrote:
>> >> >> > >>
>> &g=
t;> >> > >>>
>> >> >> > >&g=
t;> I wrote up the start of a draft on requirements.

>> >> >> > >>>
>> >> >> =
> >>> http://tools.ietf.org/html/draft-jennings-rtcweb-api-00

>> >> >> > >>>
>> >> >> =
> >>> I think the key thing for the key thing for IETF working =
is
>> >> starting to
>> >> >> > >=
>> think

>> >> >> > >>> about the requirements and wha=
t on the on the wire 
>> protocols are
>> >> that w=
e
>> >> >> > >>> want
>> >>=
 >> > >>> to select. I included some API definition becau=
se I found it

>> >> helpful to be
>> >> >> > >>=
> thinking about that in this draft but that would not be 
>> p=
art of the
>> >> IETF
>> >> >> > >=
;>> work.

>> >> >> > >>> That would be work I think W3C=
 should do. The draft also 
>> touches on
>> >> the=
 of
>> >> >> > >>> the
>> >>=
; >> > >>> security issues for media and signaling origin=
ating from 

>> browsers.
>> >> >> > >>>
>&=
gt; >> >> > >>> Cullen
>> >> >>=
; > >>>
>> >> >> > >>>

>> >> >> > >>>
>> >> >> =
> >>>
>> >> >> > >>> _________=
______________________________________
>> >> >> > &=
gt;>> dispatch mailing list

>> >> >> > >>> dispatch@ietf.org <mailto:dispatch@ietf.org>
>> >> >> > >>>=
 https://www.iet=
f.org/mailman/listinfo/dispatch

>> >> >> > >>
>> >> >> >=
 >>
>> >> >> > >> _____________________=
__________________________
>> >> >> > >> disp=
atch mailing list

>> >> >> > >> dispatch@ietf.org <mailto:di=
spatch@ietf.org>
>> >> >> > >> https://www.ietf.org/ma=
ilman/listinfo/dispatch

>> >> >> > >
>> >> >> >
=
>> >> >> >
>> >> >> >
>&=
gt; >> >>
>> >> >>
>> >> &g=
t;> --

>> >> >> Lorenzo Miniero, COB
>> >> >&g=
t;
>> >> >> Meetecho s.r.l.
>> >> >&=
gt; Web Conferencing and Collaboration Tools
>> >> >> =
http://www.meetecho.com

>> >> >> _______________________________________________<=
br>>> >> >> dispatch mailing list
>> >> &g=
t;> dispatch@ietf.org <mailt=
o:dispatch@ietf.org>

>> >> >> https://www.ietf.org/mailman/listinfo/dispatch
>> &=
gt;> >
>> >> >
>> >> > --
>=
> >> > Lorenzo Miniero, COB

>> >> >
>> >> > Meetecho s.r.l.
>>=
; >> > Web Conferencing and Collaboration Tools
>> >&g=
t; > http://www.meetecho.com

>> >> > _______________________________________________
&=
gt;> >> > dispatch mailing list
>> >> > dispatch@ietf.org <mailto:dispatch@ietf.org>

>> >> > https://www.ietf.org/mailman/listinfo/dispatch
>> ><=
br>>> >
>> > --
>> > Lorenzo Miniero, COB<=
br>
>> >
>> > Meetecho s.r.l.
>> > Web Confere=
ncing and Collaboration Tools
>> > http://www.meetecho.com
>>
>>
>> ___=
____________________________________________

>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
> 
> -- 
>                              _\\|//_
>     =
                         ( O-O )
>     ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)=
~~00o~~~~~~~~~~~~~~~~~~~~~~~~
>                      Simon Pietro Rom=
ano

>                Universita' di Napoli Federico II
>          =
         Computer Science Department
>          Phone: +39 081 768382=
3 -- Fax: +39 081 7684219
>                  e-mail: spromano@unina.it

>            h=
ttp://www.comics.unina.it/simonpietro.romano
> 
>      <=
<Molti mi dicono che lo scoraggiamento =E8 l'alibi degli
>    =
 idioti. Ci rifletto un istante; e mi scoraggio>>. Magritte.

>                           oooO
>     ~~~~~~~~~~~~~~~~~~~~~~(   )=
~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
>                            \ (    =
(   )
>                             \_)    ) /
>               =
                    (_/

> 
> 


--00504502c5c6343cf0049f8c0db0--

From Thierry.Bessis@alcatel-lucent.com  Mon Mar 28 14:03:20 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 756AF28C0F0 for ; Mon, 28 Mar 2011 14:03:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9LqMnjKC1hZj for ; Mon, 28 Mar 2011 14:03:18 -0700 (PDT)
Received: from ihemail2.lucent.com (ihemail2.lucent.com [135.245.0.35]) by core3.amsl.com (Postfix) with ESMTP id 27AA43A6873 for ; Mon, 28 Mar 2011 14:03:17 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail2.lucent.com (8.13.8/IER-o) with ESMTP id p2SL4puA019292 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 28 Mar 2011 16:04:51 -0500 (CDT)
Received: from [135.185.234.174] (USNAVN0L053562.ih.lucent.com [135.185.234.174]) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2SL4p7p007886; Mon, 28 Mar 2011 16:04:51 -0500 (CDT)
Message-ID: <4D90F7F2.2040300@alcatel-lucent.com>
Date: Mon, 28 Mar 2011 16:04:50 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: "Worley, Dale R (Dale)" 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com> 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------080006080303030500080703"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.35
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 21:03:20 -0000

This is a multi-part message in MIME format.
--------------080006080303030500080703
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dale,

This reasoning makes a lot of sense, but the best strategy is described 
in one particular scenario, using one information that we would not 
have, and another that would be approximate at best.

At best you could estimate the number of sessions left (the other 
proposal), but then, we have no ideas of the percentage of resources 
left (Still the other proposal). So if we have an node that has (or 
think it has  ! ...) 10 times more capacity , starting from idle, the 
other proposal will basically not use the smallest node until the 
biggest gets close to overload, which is something that we (and at least 
some operators) do not like.

  Now in the situation that you describe, you are correct that one 
server would have 10 times more capacity, so we are going to an 
unbalanced situation. But this is how you got there in the first place: 
the next weight in the smallest node will be zero, while it would be 
still 1% in the biggest one, so this will (again) tend to balance.

Also, I believe we do not need to be optimal in extreme cases such as 
this one (99% utilization and one node 10 times the capacity of the 
other). This is especially true when you consider that nodes could have 
different views of their "sessions left" so the theoretical "optimal" 
might not be optimal at all.

Cordially,

Thierry

On 24-Mar-2011 14:54, Worley, Dale R (Dale) wrote:
> ________________________________________
> From: dispatch-bounces@ietf.org [dispatch-bounces@ietf.org] On Behalf Of Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]
>
> TB>  It seems to me that this comment just reinforces my point (I hope I understand well): We need to let the destination come up its own way with its percentage of resource left, because there are so many possible implementations that it would be cumbersome and restrictive to impose any algorithm. Also just to be sure: The proposed weight is a percentage (a ratio): it has no unit. Part of the robustness and universality is derived from this characteristic.
> ________________________________________
>
> I haven't done an analysis, but I strongly expect that the overall system behaves better if all systems report their available capacity
> in the same amount of traffic ("number of calls") rather than percentage of capacity that is unused.
>
> The reason is that if two servers have significantly different capacity, say by a factor of 10, when both servers report 1% of capacity
> free, one server will be able to accept 10 times as much traffic as the other.
>
> Dale
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------080006080303030500080703
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    Dale,

    

    This reasoning makes a lot of sense, but the best strategy is
    described in one particular scenario, using one information that we
    would not have, and another that would be approximate at best. 

    

    At best you could estimate the number of sessions left (the other
    proposal), but then, we have no ideas of the percentage of resources
    left (Still the other proposal). So if we have an node that has (or
    think it has  ! ...) 10 times more capacity , starting from idle,
    the other proposal will basically not use the smallest node until
    the biggest gets close to overload, which is something that we (and
    at least some operators) do not like. 

    

     Now in the situation that you describe, you are correct that one
    server would have 10 times more capacity, so we are going to an
    unbalanced situation. But this is how you got there in the first
    place: the next weight in the smallest node will be zero, while it
    would be still 1% in the biggest one, so this will (again) tend to
    balance.

    

    Also, I believe we do not need to be optimal in extreme cases such
    as this one (99% utilization and one node 10 times the capacity of
    the other). This is especially true when you consider that nodes
    could have different views of their "sessions left" so the
    theoretical "optimal" might not be optimal at all. 

    

    Cordially, 

    

    Thierry

    

    On 24-Mar-2011 14:54, Worley, Dale R (Dale) wrote:
    

      
________________________________________
From: dispatch-bounces@ietf.org [dispatch-bounces@ietf.org] On Behalf Of Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]

TB> It seems to me that this comment just reinforces my point (I hope I understand well): We need to let the destination come up its own way with its percentage of resource left, because there are so many possible implementations that it would be cumbersome and restrictive to impose any algorithm. Also just to be sure: The proposed weight is a percentage (a ratio): it has no unit. Part of the robustness and universality is derived from this characteristic.
________________________________________

I haven't done an analysis, but I strongly expect that the overall system behaves better if all systems report their available capacity
in the same amount of traffic ("number of calls") rather than percentage of capacity that is unused.

The reason is that if two servers have significantly different capacity, say by a factor of 10, when both servers report 1% of capacity
free, one server will be able to accept 10 times as much traffic as the other.

Dale
_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------080006080303030500080703--

From pkyzivat@cisco.com  Mon Mar 28 14:34:05 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0FDCD3A695B for ; Mon, 28 Mar 2011 14:34:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.599
X-Spam-Level: 
X-Spam-Status: No, score=-110.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uHhiH339vYiF for ; Mon, 28 Mar 2011 14:34:03 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id AD23E3A6873 for ; Mon, 28 Mar 2011 14:34:02 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=4660; q=dns/txt; s=iport; t=1301348140; x=1302557740; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=9tAYD7jME+xc4X8ytmMYMdYM7SL4s3YrbvnfSQU3qUU=; b=ciMptIeZQMf8AqNSH+BB/wsMgLpgdCX+2RmyUgkKzNFEUIz38l8PtN85 wRccFsWNH57PsLBAX1+uaHHAxly/hcVv+P+O528/C3mpPe47Gq4aUqea5 grGs/waUyOTbmZpFDaW3t64yfIpm3kaANz/la1Ye+HedJYfrxG9RKSmSj g=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AjkEAJ3+kE2Q/khMgWdsb2JhbAClRBQBARYmJakMnCyDFoJTBIx3g1Q
X-IronPort-AV: E=Sophos;i="4.63,257,1299456000"; d="scan'208";a="81169211"
Received: from ams-core-3.cisco.com ([144.254.72.76]) by ams-iport-1.cisco.com with ESMTP; 28 Mar 2011 21:35:39 +0000
Received: from [10.55.90.214] (dhcp-10-55-90-214.cisco.com [10.55.90.214]) by ams-core-3.cisco.com (8.14.3/8.14.3) with ESMTP id p2SLZd9c031070 for ; Mon, 28 Mar 2011 21:35:39 GMT
Message-ID: <4D90FF2B.6090102@cisco.com>
Date: Mon, 28 Mar 2011 17:35:39 -0400
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: dispatch@ietf.org
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>	 <4D90F7F2.2040300@alcatel-lucent.com>
In-Reply-To: <4D90F7F2.2040300@alcatel-lucent.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Mon, 28 Mar 2011 21:34:05 -0000

ISTM that it doesn't matter much what algorithm you use until at least 
one node starts to approach full capacity. (And this includes not 
mattering if all are at the same percentage utilization. Is there 
anything wrong with the 10x box getting all the traffic for awhile 
before the 1x box gets any, as long as all the traffic is being served?)

What matters is what happens when you start to approach full capacity.

Even then, maybe it doesn't matter whether you get to the point where 
either all the boxes are at 99% capacity, or all boxes have room for one 
more call of any type.

But it does matter if 1% available capacity is enough for 100 calls, or 
if 1% capacity is insufficient for even one more call.

I'm starting to suspect that an algorithm that works for boxes that 
handle 10k calls may not work so well for boxes that handle 100 calls.

	Thanks,
	Paul

On 3/28/2011 5:04 PM, Thierry Bessis wrote:
> Dale,
>
> This reasoning makes a lot of sense, but the best strategy is described
> in one particular scenario, using one information that we would not
> have, and another that would be approximate at best.
>
> At best you could estimate the number of sessions left (the other
> proposal), but then, we have no ideas of the percentage of resources
> left (Still the other proposal). So if we have an node that has (or
> think it has ! ...) 10 times more capacity , starting from idle, the
> other proposal will basically not use the smallest node until the
> biggest gets close to overload, which is something that we (and at least
> some operators) do not like.
>
> Now in the situation that you describe, you are correct that one server
> would have 10 times more capacity, so we are going to an unbalanced
> situation. But this is how you got there in the first place: the next
> weight in the smallest node will be zero, while it would be still 1% in
> the biggest one, so this will (again) tend to balance.
>
> Also, I believe we do not need to be optimal in extreme cases such as
> this one (99% utilization and one node 10 times the capacity of the
> other). This is especially true when you consider that nodes could have
> different views of their "sessions left" so the theoretical "optimal"
> might not be optimal at all.
>
> Cordially,
>
> Thierry
>
> On 24-Mar-2011 14:54, Worley, Dale R (Dale) wrote:
>> ________________________________________
>> From:dispatch-bounces@ietf.org  [dispatch-bounces@ietf.org] On Behalf Of Thierry Bessis [Thierry.Bessis@alcatel-lucent.com]
>>
>> TB>  It seems to me that this comment just reinforces my point (I hope I understand well): We need to let the destination come up its own way with its percentage of resource left, because there are so many possible implementations that it would be cumbersome and restrictive to impose any algorithm. Also just to be sure: The proposed weight is a percentage (a ratio): it has no unit. Part of the robustness and universality is derived from this characteristic.
>> ________________________________________
>>
>> I haven't done an analysis, but I strongly expect that the overall system behaves better if all systems report their available capacity
>> in the same amount of traffic ("number of calls") rather than percentage of capacity that is unused.
>>
>> The reason is that if two servers have significantly different capacity, say by a factor of 10, when both servers report 1% of capacity
>> free, one server will be able to accept 10 times as much traffic as the other.
>>
>> Dale
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>
> --
> --
> Cordially,
> Thierry Bessis
> IMS Solution: Architecture & Product Management - ALTA Member
> Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
> Alcatel-Lucent
> 1960 Lucent Lane - Room 6L-517
> Naperville, IL 60563 U S
> Tel/Fax: +1 630 979 7989
> Corporate IM: tbessis
>
> Engage: https://engage.alcatel-lucent.com/people/tbessis
> My Calendar: www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
> ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline
>
> Conference information:
> 2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
> others countries see: http://all.alcatel-lucent.com/audio
>
> Upcoming planned Business Trip: 12 April - 14 April 2011
> Upcoming planned Vacation: None
>
>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

From HKaplan@acmepacket.com  Tue Mar 29 00:36:05 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2640F3A6889 for ; Tue, 29 Mar 2011 00:36:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.048
X-Spam-Level: 
X-Spam-Status: No, score=-2.048 tagged_above=-999 required=5 tests=[AWL=0.551,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oeZDvCi0Gz78 for ; Tue, 29 Mar 2011 00:36:00 -0700 (PDT)
Received: from etmail.acmepacket.com (etmail.acmepacket.com [216.41.24.6]) by core3.amsl.com (Postfix) with ESMTP id 128653A67AC for ; Tue, 29 Mar 2011 00:35:59 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by etmail.acmepacket.com (216.41.24.6) with Microsoft SMTP Server (TLS) id 8.2.254.0; Tue, 29 Mar 2011 03:37:37 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Tue, 29 Mar 2011 03:37:37 -0400
From: Hadriel Kaplan 
To: "Worley, Dale R (Dale)" 
Date: Tue, 29 Mar 2011 03:37:32 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvt5Cy6Zfr9CgROTqW/M/+5PnKGFw==
Message-ID: 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> , <4D826CC5.9010502@alcatel-lucent.com> 
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAgAAAUAAAAFT
Cc: "dispatch@ietf.org" , Thierry Bessis 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 07:36:05 -0000

On Mar 24, 2011, at 3:54 PM, Worley, Dale R (Dale) wrote:

> I haven't done an analysis, but I strongly expect that the overall system=
 behaves better if all systems report their available capacity
> in the same amount of traffic ("number of calls") rather than percentage =
of capacity that is unused.
>=20
> The reason is that if two servers have significantly different capacity, =
say by a factor of 10, when both servers report 1% of capacity
> free, one server will be able to accept 10 times as much traffic as the o=
ther.

I don't think anyone's proposing to advertise a literal percentage of capac=
ity, for obvious reasons. (ie, the device isn't advertising the numbers 0-1=
00 to mean 0-100%)

But anyway, although in theory the algorithm auto-adjusts in such a case re=
gardless - in practice my guess is the operators of the systems would confi=
gure the literal starting/max-number based on the capacity relative to othe=
r systems, so that you wouldn't really have a case of two systems of very d=
ifferent capacity starting out advertising the same values.

-hadriel


From partr@cisco.com  Tue Mar 29 02:05:29 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 29D3F3A68B7 for ; Tue, 29 Mar 2011 02:05:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.06
X-Spam-Level: 
X-Spam-Status: No, score=-10.06 tagged_above=-999 required=5 tests=[AWL=0.539,  BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id AHlH++uEPENm for ; Tue, 29 Mar 2011 02:05:28 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id C35BF3A67B7 for ; Tue, 29 Mar 2011 02:05:27 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=partr@cisco.com; l=1989; q=dns/txt; s=iport; t=1301389626; x=1302599226; h=mime-version:content-transfer-encoding:subject:date: message-id:in-reply-to:references:from:to:cc; bh=cHvLC5WpUv86oquqEZvrBnhk1WKk4WbRDnDvAFm+SXA=; b=Mipf+PCdlr0jR2jC4sfZ6qxvtxmC+zCH017A4SDIYstGAld21H0H6MKp VgMXrulphpfao3Wdkt8hoZ27JqoSwnFwUYuYtfiUp/huQ6ZXExCrHpeTE SVpYK5ZoH2ZODbJ6hBd8t0SGPneTnXK4ubm6JcECqikdK2478GcGO+xsV E=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Au0AAHOgkU2Q/khNgWdsb2JhbACYD403FAEBCwsmJYh5nlCcRYVqBIU6iyI
X-IronPort-AV: E=Sophos;i="4.63,260,1299456000"; d="scan'208";a="81244588"
Received: from ams-core-4.cisco.com ([144.254.72.77]) by ams-iport-1.cisco.com with ESMTP; 29 Mar 2011 09:07:05 +0000
Received: from xbh-bgl-412.cisco.com (xbh-bgl-412.cisco.com [72.163.129.202]) by ams-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2T96wqG016378; Tue, 29 Mar 2011 09:07:04 GMT
Received: from xmb-bgl-411.cisco.com ([72.163.129.207]) by xbh-bgl-412.cisco.com with Microsoft SMTPSVC(6.0.3790.4675);  Tue, 29 Mar 2011 14:37:00 +0530
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Date: Tue, 29 Mar 2011 14:37:01 +0530
Message-ID: 
In-Reply-To: 
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvt5Cy6Zfr9CgROTqW/M/+5PnKGFwACqBqg
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> , <4D826CC5.9010502@alcatel-lucent.com>  
From: "Parthasarathi R (partr)" 
To: "Hadriel Kaplan" , "Worley, Dale R (Dale)" 
X-OriginalArrivalTime: 29 Mar 2011 09:07:00.0421 (UTC) FILETIME=[A9A3CF50:01CBEDF0]
Cc: dispatch@ietf.org, Thierry Bessis 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 09:05:29 -0000

Hadriel,

In your proposed algorithm, Two information is required for load
balancing:

1) Max or total capacity is exchanged by non-sip mechanism (may be
configuration) one time or frequently in case capacity of the device
varies=20
2) Current capacity of the system is passed in each message.=20

It is the place where I'm looking for units like "number of calls" or
"number of media packets" to be clearly spelt out in the mechanism
because percentage as a unit will not serve the purpose and finding the
relation between two or more system capacity is really tough (to
configure or manage).

Thanks
Partha=20

-----Original Message-----
From: Hadriel Kaplan [mailto:HKaplan@acmepacket.com]=20
Sent: Tuesday, March 29, 2011 1:08 PM
To: Worley, Dale R (Dale)
Cc: Thierry Bessis; Parthasarathi R (partr); dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing


On Mar 24, 2011, at 3:54 PM, Worley, Dale R (Dale) wrote:

> I haven't done an analysis, but I strongly expect that the overall=20
> system behaves better if all systems report their available capacity
in the same amount of traffic ("number of calls") rather than percentage
of capacity that is unused.
>=20
> The reason is that if two servers have significantly different=20
> capacity, say by a factor of 10, when both servers report 1% of
capacity free, one server will be able to accept 10 times as much
traffic as the other.

I don't think anyone's proposing to advertise a literal percentage of
capacity, for obvious reasons. (ie, the device isn't advertising the
numbers 0-100 to mean 0-100%)

But anyway, although in theory the algorithm auto-adjusts in such a case
regardless - in practice my guess is the operators of the systems would
configure the literal starting/max-number based on the capacity relative
to other systems, so that you wouldn't really have a case of two systems
of very different capacity starting out advertising the same values.

-hadriel


From HKaplan@acmepacket.com  Tue Mar 29 02:33:35 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 10FCF3A6906 for ; Tue, 29 Mar 2011 02:33:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.186
X-Spam-Level: 
X-Spam-Status: No, score=-2.186 tagged_above=-999 required=5 tests=[AWL=0.413,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CBMTiKu1qrxe for ; Tue, 29 Mar 2011 02:33:33 -0700 (PDT)
Received: from etmail.acmepacket.com (etmail.acmepacket.com [216.41.24.6]) by core3.amsl.com (Postfix) with ESMTP id 5F3713A690E for ; Tue, 29 Mar 2011 02:33:33 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by etmail.acmepacket.com (216.41.24.6) with Microsoft SMTP Server (TLS) id 8.2.254.0; Tue, 29 Mar 2011 05:35:10 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Tue, 29 Mar 2011 05:35:10 -0400
From: Hadriel Kaplan 
To: "Parthasarathi R (partr)" 
Date: Tue, 29 Mar 2011 05:35:08 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: Acvt9JiymvXl9ABjSa20Gr2eq5RDQw==
Message-ID: <5DEA150B-8FB6-4029-996C-1F9CA0533300@acmepacket.com>
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> , <4D826CC5.9010502@alcatel-lucent.com>   
In-Reply-To: 
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAgAAAUAAAAFR
Cc: "dispatch@ietf.org" , Thierry Bessis 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 09:33:35 -0000

On Mar 29, 2011, at 5:07 AM, Parthasarathi R (partr) wrote:

> Hadriel,
>=20
> In your proposed algorithm, Two information is required for load
> balancing:
>=20
> 1) Max or total capacity is exchanged by non-sip mechanism (may be
> configuration) one time or frequently in case capacity of the device
> varies=20

Nope - it's not signaled separately.  I was talking about the number used a=
s the baseline/reference/100% number value.  So think of it like the call-c=
apacity: if we signaled a call-capacity, all I'm saying is that the max cal=
l-capactiy (the capacity when the system has no current calls) would be con=
figured by the admin to begin with, rather than hardcoded in the system.  B=
ut in this case we're not talking about a call-capacity... we're talking ab=
out a "resource load" or some such. (which may happen to be a call capacity=
 in someone's system, but doesn't have to be)


> 2) Current capacity of the system is passed in each message.=20
>=20
> It is the place where I'm looking for units like "number of calls" or
> "number of media packets" to be clearly spelt out in the mechanism

And I'm saying there's no way we could ever agree on all the possible set o=
f constraints which affect system load - so we could never send a unit like=
 "number of calls" or "number of media packets", because there are just way=
 too many such units which have different impacts on different system types=
.


> because percentage as a unit will not serve the purpose and finding the
> relation between two or more system capacity is really tough (to
> configure or manage).

Just to be clear we're on the same page - it's not really a "percentage uni=
t".  A percentage unit would be something like 73.4%.  No one's proposing s=
ignaling "73.4%". =20

-hadriel


From pkyzivat@cisco.com  Tue Mar 29 04:40:22 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D757828C15C for ; Tue, 29 Mar 2011 04:40:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.599
X-Spam-Level: 
X-Spam-Status: No, score=-110.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id knFRkXF5Rsng for ; Tue, 29 Mar 2011 04:40:18 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id EC54528C148 for ; Tue, 29 Mar 2011 04:40:17 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=1776; q=dns/txt; s=iport; t=1301398916; x=1302608516; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=ii49QUWz/ca7Ljk4iq96wmuEEieAmYTIEnWAJG7DVQI=; b=RGkczX7Nwlk2ONiz+PdCNwd31ultOhWXzVpime161EzRJkyL87RrViXz BAM7ql1I0NejO3o43CpB3AfYPU8SxG+002fewRTENfSGnMCMzhb6jLIJH BGdOqvf5qnDF+0XKesOwYW+PXATVEKezq0Ep7MkWrH0cAnmXGLiiMeJQ+ U=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AnIHALPEkU2Q/khLgWdsb2JhbACYTox3FAEBFiYlqFicTIVqBI0Cg1Q
X-IronPort-AV: E=Sophos;i="4.63,261,1299456000"; d="scan'208";a="81280239"
Received: from ams-core-2.cisco.com ([144.254.72.75]) by ams-iport-1.cisco.com with ESMTP; 29 Mar 2011 11:41:53 +0000
Received: from [10.61.85.87] (ams3-vpn-dhcp5464.cisco.com [10.61.85.87]) by ams-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2TBfrtP007564 for ; Tue, 29 Mar 2011 11:41:53 GMT
Message-ID: <4D91C581.4040802@cisco.com>
Date: Tue, 29 Mar 2011 07:41:53 -0400
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: dispatch@ietf.org
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>	 
In-Reply-To: 
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 11:40:22 -0000

On 3/29/2011 3:37 AM, Hadriel Kaplan wrote:
>
> On Mar 24, 2011, at 3:54 PM, Worley, Dale R (Dale) wrote:
>
>> I haven't done an analysis, but I strongly expect that the overall system behaves better if all systems report their available capacity
>> in the same amount of traffic ("number of calls") rather than percentage of capacity that is unused.
>>
>> The reason is that if two servers have significantly different capacity, say by a factor of 10, when both servers report 1% of capacity
>> free, one server will be able to accept 10 times as much traffic as the other.
>
> I don't think anyone's proposing to advertise a literal percentage of capacity, for obvious reasons. (ie, the device isn't advertising the numbers 0-100 to mean 0-100%)
>
> But anyway, although in theory the algorithm auto-adjusts in such a case regardless - in practice my guess is the operators of the systems would configure the literal starting/max-number based on the capacity relative to other systems, so that you wouldn't really have a case of two systems of very different capacity starting out advertising the same values.

Its possible to abuse anything. But we're in trouble if we start out 
assuming that this be used contrary to the design intent.

If people scale their use of percent capacity so that the big box maxes 
out at 100 and the smaller box maxes out at 50, then it has been 
transformed into something like the SCE mechanism, that have quite 
different properties. If that is what we expect people to use then we 
should design and analyze it on that basis.

	Thanks,
	Paul

> -hadriel
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>

From pkyzivat@cisco.com  Tue Mar 29 05:15:09 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 168623A68F5 for ; Tue, 29 Mar 2011 05:15:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.599
X-Spam-Level: 
X-Spam-Status: No, score=-110.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rOuH5qzW3nPQ for ; Tue, 29 Mar 2011 05:15:08 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id B42353A68FD for ; Tue, 29 Mar 2011 05:15:07 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=pkyzivat@cisco.com; l=2032; q=dns/txt; s=iport; t=1301401006; x=1302610606; h=message-id:date:from:mime-version:to:subject:references: in-reply-to:content-transfer-encoding; bh=vytyT83Bx11gj2YhvSBnU0ZlGILnQcRH+PJZ7B6aNM8=; b=fwe9+tYR/8j4QkoS7SKTgFZdEWE33fM7h6jbbS0kN017CA2VfYySuvcC tczbtxb8NPPTGnRdiLKcXDp12xvP0H3ucryMC+r+TXAVDBRTQf/Ajm6fY FaphGLkTXMfI3G/mE1vGLmlTbEjXz5/LIml0CO7XQ0YznJCkLNALAAkQb 0=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AnIHAN7MkU2Q/khLgWdsb2JhbACYTox3FAEBFiYlqGCcT4VqBI0Cg1Q
X-IronPort-AV: E=Sophos;i="4.63,262,1299456000"; d="scan'208";a="81289322"
Received: from ams-core-2.cisco.com ([144.254.72.75]) by ams-iport-1.cisco.com with ESMTP; 29 Mar 2011 12:16:45 +0000
Received: from [10.61.85.87] (ams3-vpn-dhcp5464.cisco.com [10.61.85.87]) by ams-core-2.cisco.com (8.14.3/8.14.3) with ESMTP id p2TCGiTM022819 for ; Tue, 29 Mar 2011 12:16:44 GMT
Message-ID: <4D91CDAC.1030302@cisco.com>
Date: Tue, 29 Mar 2011 08:16:44 -0400
From: Paul Kyzivat 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.15) Gecko/20110303 Thunderbird/3.1.9
MIME-Version: 1.0
To: dispatch@ietf.org
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>			 <5DEA150B-8FB6-4029-996C-1F9CA0533300@acmepacket.com>
In-Reply-To: <5DEA150B-8FB6-4029-996C-1F9CA0533300@acmepacket.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 12:15:09 -0000

inline

On 3/29/2011 5:35 AM, Hadriel Kaplan wrote:
>
>
>
>
> On Mar 29, 2011, at 5:07 AM, Parthasarathi R (partr) wrote:
>
>> Hadriel,
>>
>> In your proposed algorithm, Two information is required for load
>> balancing:
>>
>> 1) Max or total capacity is exchanged by non-sip mechanism (may be
>> configuration) one time or frequently in case capacity of the device
>> varies
>
> Nope - it's not signaled separately.  I was talking about the number used as the baseline/reference/100% number value.  So think of it like the call-capacity: if we signaled a call-capacity, all I'm saying is that the max call-capactiy (the capacity when the system has no current calls) would be configured by the admin to begin with, rather than hardcoded in the system.  But in this case we're not talking about a call-capacity... we're talking about a "resource load" or some such. (which may happen to be a call capacity in someone's system, but doesn't have to be)
>
>
>> 2) Current capacity of the system is passed in each message.
>>
>> It is the place where I'm looking for units like "number of calls" or
>> "number of media packets" to be clearly spelt out in the mechanism
>
> And I'm saying there's no way we could ever agree on all the possible set of constraints which affect system load - so we could never send a unit like "number of calls" or "number of media packets", because there are just way too many such units which have different impacts on different system types.
>
>
>> because percentage as a unit will not serve the purpose and finding the
>> relation between two or more system capacity is really tough (to
>> configure or manage).
>
> Just to be clear we're on the same page - it's not really a "percentage unit".  A percentage unit would be something like 73.4%.  No one's proposing signaling "73.4%".

Well...

IIUC, Thierry *is* talking about a "percentage unit", but rescaled from 
[0..100) to [0..2^16-1].

I'm sure he'll correct me if I have that wrong.

	Thanks,
	Paul

From Thierry.Bessis@alcatel-lucent.com  Tue Mar 29 08:14:30 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 893013A6835 for ; Tue, 29 Mar 2011 08:14:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id G5VTSOoSHq-R for ; Tue, 29 Mar 2011 08:14:29 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 3B64D3A6927 for ; Tue, 29 Mar 2011 08:14:29 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2TFG1W1017489 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 29 Mar 2011 10:16:01 -0500 (CDT)
Received: from [135.185.234.174] (USNAVN0L053562.ih.lucent.com [135.185.234.174] (may be forged)) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2TFG0Ca005168; Tue, 29 Mar 2011 10:16:00 -0500 (CDT)
Message-ID: <4D91F7B0.1000006@alcatel-lucent.com>
Date: Tue, 29 Mar 2011 10:16:00 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Hadriel Kaplan 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>	 
In-Reply-To: 
Content-Type: multipart/alternative; boundary="------------050809040204080100090705"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 15:14:30 -0000

This is a multi-part message in MIME format.
--------------050809040204080100090705
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Kaplan,
> I don't think anyone's proposing to advertise a literal percentage of capacity, for obvious reasons. (ie, the device isn't advertising the numbers 0-100 to mean 0-100%)
Actually we do !  This is what we are doing in our system. (we really 
are using a byte [0..255] but it means the exact same thing).
We are dealing with cases where some nodes may have 2-3 times the 
capacity of others due to hardware generation gap. (but not 10 times ... 
). we are also dealing with growing/degrowing, and temporary partial 
capacity loss: it works well.
> But anyway, although in theory the algorithm auto-adjusts in such a case regardless - in practice my guess is the operators of the systems would configure the literal starting/max-number based on the capacity relative to other systems, so that you wouldn't really have a case of two systems of very different capacity starting out advertising the same values.
This is also what we are trying to avoid: having the operator need to 
compare nodes capacity (possibly from different vendors), and come up 
with complicated engineering rules to configure the weights (rules that 
additionally would need to change each time something is modfied in the 
network).  All our nodes advertise about 100% availability when they 
start, whatever their capacity (unless they were also used for other tasks)

Cordially,

Thierry
> -hadriel
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------050809040204080100090705
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
    
  
  
    Kaplan, 

    

      
I don't think anyone's proposing to advertise a literal percentage of capacity, for obvious reasons. (ie, the device isn't advertising the numbers 0-100 to mean 0-100%)


    

    Actually we do !  This is what we are doing in our system. (we
    really are using a byte [0..255] but it means the exact same thing).

    We are dealing with cases where some nodes may have 2-3 times the
    capacity of others due to hardware generation gap. (but not 10 times
    ... ). we are also dealing with growing/degrowing, and temporary
    partial capacity loss: it works well.
    

      
But anyway, although in theory the algorithm auto-adjusts in such a case regardless - in practice my guess is the operators of the systems would configure the literal starting/max-number based on the capacity relative to other systems, so that you wouldn't really have a case of two systems of very different capacity starting out advertising the same values.


    

    This is also what we are trying to avoid: having the operator need
    to compare nodes capacity (possibly from different vendors), and
    come up with complicated engineering rules to configure the weights
    (rules that additionally would need to change each time something is
    modfied in the network).  All our nodes advertise about 100%
    availability when they start, whatever their capacity (unless they
    were also used for other tasks)  

    

    Cordially, 

    

    Thierry

    

      
-hadriel

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------050809040204080100090705--

From Thierry.Bessis@alcatel-lucent.com  Tue Mar 29 08:27:27 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E3C3F3A6931 for ; Tue, 29 Mar 2011 08:27:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0PZFIn7RqsBZ for ; Tue, 29 Mar 2011 08:27:26 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 9B4023A6932 for ; Tue, 29 Mar 2011 08:27:23 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2TFT0BW026696 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 29 Mar 2011 10:29:00 -0500 (CDT)
Received: from [135.185.234.174] (USNAVN0L053562.ih.lucent.com [135.185.234.174] (may be forged)) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2TFT0LC017622; Tue, 29 Mar 2011 10:29:00 -0500 (CDT)
Message-ID: <4D91FABC.10206@alcatel-lucent.com>
Date: Tue, 29 Mar 2011 10:29:00 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>				<5DEA150B-8FB6-4029-996C-1F9CA0533300@acmepacket.com> <4D91CDAC.1030302@cisco.com>
In-Reply-To: <4D91CDAC.1030302@cisco.com>
Content-Type: multipart/alternative; boundary="------------070108040209070508000808"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 15:27:28 -0000

This is a multi-part message in MIME format.
--------------070108040209070508000808
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

All,
>> Just to be clear we're on the same page - it's not really a 
>> "percentage unit".  A percentage unit would be something like 73.4%.  
>> No one's proposing signaling "73.4%".
>
> Well...
>
> IIUC, Thierry *is* talking about a "percentage unit", but rescaled 
> from [0..100) to [0..2^16-1].
> I'm sure he'll correct me if I have that wrong.
You completely correct Paul. Whatever the scale we decide to use, I 
proposed a percentage. It is the ratio of the number of available 
resources on the total number of resources.
Vijay will make it clear in the next draft, and we will probably propose 
to use directly a percentage [0 ..100] to make things easier.
We think that a direct percentage should be enough in terms of 
precision, because the resource usage precision is also in this order of 
magnitude.

cordially,

Thierry
>
>     Thanks,
>     Paul
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------070108040209070508000808
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    All, 

    

      
Just to be clear we're on the same page -
        it's not really a "percentage unit".  A percentage unit would be
        something like 73.4%.  No one's proposing signaling "73.4%".
        

      

      

      Well...
      

      

      IIUC, Thierry *is* talking about a "percentage unit", but rescaled
      from [0..100) to [0..2^16-1].
      

      I'm sure he'll correct me if I have that wrong.
      

    

    You completely correct Paul. Whatever the scale we decide to use, I
    proposed a percentage. It is the ratio of the number of available
    resources on the total number of resources. 

    Vijay will make it clear in the next draft, and we will probably
    propose to use directly a percentage [0 ..100] to make things
    easier. 

    We think that a direct percentage should be enough in terms of
    precision, because the resource usage precision is also in this
    order of magnitude. 

    

    cordially, 

    

    Thierry

    

      

          Thanks,
      

          Paul
      

      _______________________________________________
      

      dispatch mailing list
      

      dispatch@ietf.org
      

      https://www.ietf.org/mailman/listinfo/dispatch
      

    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------070108040209070508000808--

From Thierry.Bessis@alcatel-lucent.com  Tue Mar 29 08:40:29 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 57BE83A692D for ; Tue, 29 Mar 2011 08:40:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.597
X-Spam-Level: 
X-Spam-Status: No, score=-6.597 tagged_above=-999 required=5 tests=[AWL=0.000,  BAYES_00=-2.599, HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GfJxZR9FGTTw for ; Tue, 29 Mar 2011 08:40:27 -0700 (PDT)
Received: from ihemail1.lucent.com (ihemail1.lucent.com [135.245.0.33]) by core3.amsl.com (Postfix) with ESMTP id 4B73F3A687E for ; Tue, 29 Mar 2011 08:40:27 -0700 (PDT)
Received: from umail.lucent.com (h135-3-40-63.lucent.com [135.3.40.63]) by ihemail1.lucent.com (8.13.8/IER-o) with ESMTP id p2TFg4C6005685 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 29 Mar 2011 10:42:04 -0500 (CDT)
Received: from [135.185.234.174] (USNAVN0L053562.ih.lucent.com [135.185.234.174] (may be forged)) by umail.lucent.com (8.13.8/TPES) with ESMTP id p2TFg3wG000989; Tue, 29 Mar 2011 10:42:03 -0500 (CDT)
Message-ID: <4D91FDCB.7070008@alcatel-lucent.com>
Date: Tue, 29 Mar 2011 10:42:03 -0500
From: Thierry Bessis 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: Paul Kyzivat 
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com>	<4D756C47.1000509@bell-labs.com>	<086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com>	<4D7668B9.1020105@alcatel-lucent.com>	<004601cbddd2$15b11ab0$41135010$@packetizer.com>	<4D77B831.7000404@alcatel-lucent.com>	<026f01cbded9$be15f290$3a41d7b0$@packetizer.com>	<4D791595.2010703@alcatel-lucent.com>		<4D7A485E.2080602@alcatel-lucent.com>		<4D80ED57.7030709@alcatel-lucent.com>	, <4D826CC5.9010502@alcatel-lucent.com>		 <4D91C581.4040802@cisco.com>
In-Reply-To: <4D91C581.4040802@cisco.com>
Content-Type: multipart/alternative; boundary="------------030006090302000606070308"
X-Scanned-By: MIMEDefang 2.57 on 135.245.2.33
Cc: dispatch@ietf.org
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 15:40:29 -0000

This is a multi-part message in MIME format.
--------------030006090302000606070308
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


>> But anyway, although in theory the algorithm auto-adjusts in such a 
>> case regardless - in practice my guess is the operators of the 
>> systems would configure the literal starting/max-number based on the 
>> capacity relative to other systems, so that you wouldn't really have 
>> a case of two systems of very different capacity starting out 
>> advertising the same values.
>
> Its possible to abuse anything. But we're in trouble if we start out 
> assuming that this be used contrary to the design intent.
>
> If people scale their use of percent capacity so that the big box 
> maxes out at 100 and the smaller box maxes out at 50, then it has been 
> transformed into something like the SCE mechanism, that have quite 
> different properties. If that is what we expect people to use then we 
> should design and analyze it on that basis.
Yes, I agree with Paul's point.
The whole point of our proposal is two folds: (1) simplicity 
(straightforward weight calculation) and (2) NO dependency (No network  
engineering or topology considerations to setup weights: each node 
provides its weight independently).

Cordially,

Thierry

>
>     Thanks,
>     Paul
>
>> -hadriel
>>
>> _______________________________________________
>> dispatch mailing list
>> dispatch@ietf.org
>> https://www.ietf.org/mailman/listinfo/dispatch
>>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch

-- 
-- 
Cordially,
Thierry Bessis
IMS Solution: Architecture & Product Management - ALTA Member
   Organization: ALU > Carrier group > Wireline BU > **IMS Sol Arch, SE, PdM
Alcatel-Lucent
1960 Lucent Lane - Room 6L-517
   Naperville, IL  60563  U S
   Tel/Fax: +1 630 979 7989
   Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None

--------------030006090302000606070308
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit



  
    
  
  
    

    

      
But anyway, although in theory the
        algorithm auto-adjusts in such a case regardless - in practice
        my guess is the operators of the systems would configure the
        literal starting/max-number based on the capacity relative to
        other systems, so that you wouldn't really have a case of two
        systems of very different capacity starting out advertising the
        same values.
        

      

      

      Its possible to abuse anything. But we're in trouble if we start
      out assuming that this be used contrary to the design intent.
      

      

      If people scale their use of percent capacity so that the big box
      maxes out at 100 and the smaller box maxes out at 50, then it has
      been transformed into something like the SCE mechanism, that have
      quite different properties. If that is what we expect people to
      use then we should design and analyze it on that basis.
      

    

    Yes, I agree with Paul's point. 

    The whole point of our proposal is two folds: (1) simplicity
    (straightforward weight calculation) and (2) NO dependency (No
    network  engineering or topology considerations to setup weights:
    each node provides its weight independently). 

    

    Cordially, 

    

    Thierry

    

    

      

          Thanks,
      

          Paul
      

      

      
-hadriel
        

        

        _______________________________________________
        

        dispatch mailing list
        

        dispatch@ietf.org
        

        https://www.ietf.org/mailman/listinfo/dispatch
        

        

      

      _______________________________________________
      

      dispatch mailing list
      

      dispatch@ietf.org
      

      https://www.ietf.org/mailman/listinfo/dispatch
      

    

    

    
-- 

      
      Thierry's signature
      
      -- 

              Cordially, 

            Thierry Bessis
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Carrier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      Alcatel-Lucent

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/people/tbessis

      My Calendar:  
        www.google.com/calendar/embed?src=new_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None
    

  


--------------030006090302000606070308--

From spromano@unina.it  Tue Mar 29 14:17:30 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DB84C3A68FF for ; Tue, 29 Mar 2011 14:17:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -100.285
X-Spam-Level: 
X-Spam-Status: No, score=-100.285 tagged_above=-999 required=5 tests=[AWL=0.434, BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245,  USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3ZJY+HByf5A1 for ; Tue, 29 Mar 2011 14:17:29 -0700 (PDT)
Received: from smtp2.unina.it (smtp2.unina.it [192.132.34.62]) by core3.amsl.com (Postfix) with ESMTP id 4A42B3A6825 for ; Tue, 29 Mar 2011 14:17:29 -0700 (PDT)
Received: from [130.129.23.172] (dhcp-17ac.meeting.ietf.org [130.129.23.172]) (authenticated bits=0) by smtp2.unina.it (8.14.4/8.14.4) with ESMTP id p2TLJ2bC019465 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO); Tue, 29 Mar 2011 23:19:03 +0200
Message-ID: <4D924CB7.30909@unina.it>
Date: Tue, 29 Mar 2011 23:18:47 +0200
From: Simon Pietro Romano 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; it; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: DISPATCH list 
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8bit
Subject: [dispatch] RTCWEB session recording available
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 21:17:31 -0000

Dear all,

for those interested, the full recording (synchronized video, audio, 
slides and jabber room) of today's BOF on RTCWEB is available at the 
following URL:

http://ietf.conf.meetecho.com

Please be advised that the file is quite big (~200M), since it contains 
the entire session. In case of problems with the playout, just drop an 
e-mail to ietf-support@meetecho.com.

Cheers,

Simon


-- 
                             _\\|//_
                             ( O-O )
    ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~
                     Simon Pietro Romano
               Universita' di Napoli Federico II
                  Computer Science Department
         Phone: +39 081 7683823 -- Fax: +39 081 7684219
                 e-mail: spromano@unina.it
           http://www.comics.unina.it/simonpietro.romano

     <>. Magritte.
                          oooO
    ~~~~~~~~~~~~~~~~~~~~~~(   )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
                           \ (    (   )
                            \_)    ) /
                                  (_/



From HKaplan@acmepacket.com  Tue Mar 29 15:56:43 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3CE7728C0EC for ; Tue, 29 Mar 2011 15:56:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.16
X-Spam-Level: 
X-Spam-Status: No, score=-1.16 tagged_above=-999 required=5 tests=[AWL=-0.667,  BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553,  HTML_FONT_SIZE_LARGE=0.001, HTML_MESSAGE=0.001, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MDj95Uxg0RHk for ; Tue, 29 Mar 2011 15:56:41 -0700 (PDT)
Received: from ETMail2.acmepacket.com (unknown [216.41.24.9]) by core3.amsl.com (Postfix) with ESMTP id D22E13A6AC5 for ; Tue, 29 Mar 2011 15:56:40 -0700 (PDT)
Received: from mail.acmepacket.com (216.41.24.7) by ETMail2.acmepacket.com (216.41.24.9) with Microsoft SMTP Server (TLS) id 8.1.240.5; Tue, 29 Mar 2011 18:58:18 -0400
Received: from mailbox1.acmepacket.com ([216.41.24.12]) by mail ([127.0.0.1]) with mapi; Tue, 29 Mar 2011 18:58:18 -0400
From: Hadriel Kaplan 
To: Thierry Bessis 
Date: Tue, 29 Mar 2011 18:58:15 -0400
Thread-Topic: [dispatch] SIP load balancing
Thread-Index: AcvuZMsf18UHsggMTSGmM1FBrQfYMA==
Message-ID: <8219AFE4-B787-4CDD-B286-657565B7129E@acmepacket.com>
References: <4D755B9E.6030008@bell-labs.com>	<4D7566C1.40605@cisco.com> <4D756C47.1000509@bell-labs.com> <086e01cbdd4e$6b8188f0$42849ad0$@packetizer.com> <4D7668B9.1020105@alcatel-lucent.com> <004601cbddd2$15b11ab0$41135010$@packetizer.com> <4D77B831.7000404@alcatel-lucent.com> <026f01cbded9$be15f290$3a41d7b0$@packetizer.com> <4D791595.2010703@alcatel-lucent.com>  <4D7A485E.2080602@alcatel-lucent.com>  <4D80ED57.7030709@alcatel-lucent.com> , <4D826CC5.9010502@alcatel-lucent.com>   <4D91F7B0.1000006@alcatel-lucent.com>
In-Reply-To: <4D91F7B0.1000006@alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: multipart/alternative; boundary="_000_8219AFE4B7874CDDB286657565B7129Eacmepacketcom_"
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAQAAAUA=
Cc: "dispatch@ietf.org" 
Subject: Re: [dispatch] SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Tue, 29 Mar 2011 22:56:43 -0000

--_000_8219AFE4B7874CDDB286657565B7129Eacmepacketcom_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable


No you aren't.  It's a scaled number.  That's a big difference.
Why?  Because it allows the operator to do it either way - they can let all=
 devices start at any number and converge using the algorithm dynamically, =
or they can provision it.

And if they provision it: (1) the number range is large enough that they ca=
n provision it once, and (2) even if they get it wrong or not quite right i=
t converges anyway.  By "provision it once", I mean they don't have to chan=
ge the setting for installed systems when they add new systems of larger ca=
pacities, or if their installed systems decrease in capacity as SIP gets mo=
re chatty (ie, as a single session/call now takes more SIP messages than it=
 did when they first installed the systems, which is a common problem).

-hadriel


On Mar 29, 2011, at 11:16 AM, Thierry Bessis wrote:

Kaplan,

I don't think anyone's proposing to advertise a literal percentage of capac=
ity, for obvious reasons. (ie, the device isn't advertising the numbers 0-1=
00 to mean 0-100%)


Actually we do !  This is what we are doing in our system. (we really are u=
sing a byte [0..255] but it means the exact same thing).
We are dealing with cases where some nodes may have 2-3 times the capacity =
of others due to hardware generation gap. (but not 10 times ... ). we are a=
lso dealing with growing/degrowing, and temporary partial capacity loss: it=
 works well.

But anyway, although in theory the algorithm auto-adjusts in such a case re=
gardless - in practice my guess is the operators of the systems would confi=
gure the literal starting/max-number based on the capacity relative to othe=
r systems, so that you wouldn't really have a case of two systems of very d=
ifferent capacity starting out advertising the same values.


This is also what we are trying to avoid: having the operator need to compa=
re nodes capacity (possibly from different vendors), and come up with compl=
icated engineering rules to configure the weights (rules that additionally =
would need to change each time something is modfied in the network).  All o=
ur nodes advertise about 100% availability when they start, whatever their =
capacity (unless they were also used for other tasks)

Cordially,

Thierry

-hadriel

_______________________________________________
dispatch mailing list
dispatch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


--
--
Cordially,
Thierry Bessis
  IMS Solution: Architecture & Product Management - ALTA Member
  Organization: ALU > Carrier group > Wireline BU >  IMS Sol Arch, SE, PdM
[file:///C:/Documents%20and%20Settings/tbessis/My%20Documents/al_corp_h_3d_=
16mm.jpg]
  1960 Lucent Lane - Room 6L-517
  Naperville, IL  60563  U S
  Tel/Fax: +1 630 979 7989
  Corporate IM: tbessis

Engage: https://engage.alcatel-lucent.com/people/tbessis
My Calendar:   www.google.com/calendar/embed?src=3Dnew_frs3175%40yahoo.fr
ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotline

Conference information:
2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access Code: 9797989
others countries see: http://all.alcatel-lucent.com/audio

Upcoming planned Business Trip: 12 April - 14 April 2011
Upcoming planned Vacation: None


--_000_8219AFE4B7874CDDB286657565B7129Eacmepacketcom_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable



No you=
 aren't.  It's a scaled number.  That's a big difference.
Why?  Because it allows the operator to do it either way - they can=
 let all devices start at any number and converge using the algorithm dynam=
ically, or they can provision it.  

And if th=
ey provision it: (1) the number range is large enough that they c=
an provision it once, and (2) even if they get it wrong or not quite right =
it converges anyway.  By "provision it once", I mean they don't have t=
o change the setting for installed systems when they add new systems of lar=
ger capacities, or if their installed systems decrease in capacity as SIP g=
ets more chatty (ie, as a single session/call now takes more SIP messages t=
han it did when they first installed the systems, which is a common problem=
).

-hadriel


On M=
ar 29, 2011, at 11:16 AM, Thierry Bessis wrote:




    Kaplan, 

    

      
I don't think anyone's proposing to advertise a litera=
l percentage of capacity, for obvious reasons. (ie, the device isn't advert=
ising the numbers 0-100 to mean 0-100%)


    

    Actually we do !  This is what we are doing in our system. (we
    really are using a byte [0..255] but it means the exact same thing).
    We are dealing with cases where some nodes may have 2-3 times the
    capacity of others due to hardware generation gap. (but not 10 times
    ... ). we are also dealing with growing/degrowing, and temporary
    partial capacity loss: it works well.
    

      
But anyway, although in theory the algorithm auto-adju=
sts in such a case regardless - in practice my guess is the operators of th=
e systems would configure the literal starting/max-number based on the capa=
city relative to other systems, so that you wouldn't really have a case of =
two systems of very different capacity starting out advertising the same va=
lues.


    

    This is also what we are trying to avoid: having the operator need
    to compare nodes capacity (possibly from different vendors), and
    come up with complicated engineering rules to configure the weights
    (rules that additionally would need to change each time something is
    modfied in the network).  All our nodes advertise about 100%
    availability when they start, whatever their capacity (unless they
    were also used for other tasks)  

    

    Cordially, 

    

    Thierry

    

      
-hadriel

_______________________________________________
dispatch mailing list
dis=
patch@ietf.org
https://www.ietf.org/mailman/listinfo/dispatch


    

    

    
-- 

     =20
     =20
     =20
      -- 

              Cordially, 

            Thierry Bessis<=
/big>
      
 

      
 

       
      IMS Solution: Architecture
        & Product Management - ALTA Member

        Organization: ALU > Car=
rier
        group >
        Wireline BU >  IMS Sol Arch, SE, PdM

      3D"Alcatel-Lucent"

       
      1960 Lucent Lane - Room 6L-517

        Naperville, IL 
        60563  U
        S

        Tel/Fax: +1 630 979 7989

        Corporate IM:
        tbessis

      

      Engage: https://engage.alcatel-lucent.com/p=
eople/tbessis

      My Calendar:  
        www.google.com/calendar/embed=
?src=3Dnew_frs3175%40yahoo.fr

        ALTA Hot Line: http://alta.all.alcatel-lucent.com/hotli=
ne

      

      Conference information:

        2801 2801 (US):+1 800 771 8734 - (F):+33 800 941 674 - Access
        Code:
        9797989

        others countries see: http://all.alcatel-lucent.com/audio<=
/small>

      

      Upcoming planned Business Trip:
            12 April - 14 April 2011

      Upcoming planned Vacation: None=

    

  




=

--_000_8219AFE4B7874CDDB286657565B7129Eacmepacketcom_--

From spromano@unina.it  Wed Mar 30 09:07:58 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 778553A6A48 for ; Wed, 30 Mar 2011 09:07:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -100.459
X-Spam-Level: 
X-Spam-Status: No, score=-100.459 tagged_above=-999 required=5 tests=[AWL=0.260, BAYES_00=-2.599, HELO_EQ_IT=0.635, HOST_EQ_IT=1.245,  USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id yZhMKrpRTkE8 for ; Wed, 30 Mar 2011 09:07:57 -0700 (PDT)
Received: from smtp1.unina.it (smtp1.unina.it [192.132.34.61]) by core3.amsl.com (Postfix) with ESMTP id 1E5123A699E for ; Wed, 30 Mar 2011 09:07:56 -0700 (PDT)
Received: from [130.129.23.172] (dhcp-17ac.meeting.ietf.org [130.129.23.172]) (authenticated bits=0) by smtp1.unina.it (8.14.4/8.14.4) with ESMTP id p2UG9TID005352 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO) for ; Wed, 30 Mar 2011 18:09:35 +0200
Message-ID: <4D9355AF.6040801@unina.it>
Date: Wed, 30 Mar 2011 18:09:19 +0200
From: Simon Pietro Romano 
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; it; rv:1.9.2.15) Gecko/20110303 Lightning/1.0b2 Thunderbird/3.1.9
MIME-Version: 1.0
To: DISPATCH 
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8bit
Subject: [dispatch] CLUE recording available
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Wed, 30 Mar 2011 16:07:58 -0000

Dear all,

the full recording (synchronized video, audio, slides and jabber room) 
of today's CLUE session is available at the following URL:

http://ietf.conf.meetecho.com

In case of problems with the playout, just drop an e-mail to 
ietf-support@meetecho.com.

Cheers,

Simon

-- 
                             _\\|//_
                             ( O-O )
    ~~~~~~~~~~~~~~~~~~~~~~o00~~(_)~~00o~~~~~~~~~~~~~~~~~~~~~~~~
                     Simon Pietro Romano
               Universita' di Napoli Federico II
                  Computer Science Department
         Phone: +39 081 7683823 -- Fax: +39 081 7684219
                 e-mail: spromano@unina.it
           http://www.comics.unina.it/simonpietro.romano

     <>. Magritte.
                          oooO
    ~~~~~~~~~~~~~~~~~~~~~~(   )~~ Oooo~~~~~~~~~~~~~~~~~~~~~~~~~
                           \ (    (   )
                            \_)    ) /
                                  (_/



From rjsparks@nostrum.com  Thu Mar 31 02:01:51 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B234E28C165 for ; Thu, 31 Mar 2011 02:01:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.585
X-Spam-Level: 
X-Spam-Status: No, score=-102.585 tagged_above=-999 required=5 tests=[AWL=0.014, BAYES_00=-2.599, HTML_MESSAGE=0.001, SPF_PASS=-0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id F2CAIb6pDG8X for ; Thu, 31 Mar 2011 02:01:50 -0700 (PDT)
Received: from nostrum.com (nostrum-pt.tunnel.tserv2.fmt.ipv6.he.net [IPv6:2001:470:1f03:267::2]) by core3.amsl.com (Postfix) with ESMTP id 77E7128C15E for ; Thu, 31 Mar 2011 02:01:50 -0700 (PDT)
Received: from dhcp-12eb.meeting.ietf.org (dhcp-12eb.meeting.ietf.org [130.129.18.235]) (authenticated bits=0) by nostrum.com (8.14.3/8.14.3) with ESMTP id p2V93R5e002069 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO) for ; Thu, 31 Mar 2011 04:03:28 -0500 (CDT) (envelope-from rjsparks@nostrum.com)
From: Robert Sparks 
Content-Type: multipart/alternative; boundary=Apple-Mail-66--1015024056
Date: Thu, 31 Mar 2011 11:03:26 +0200
References: 
To: DISPATCH list 
Message-Id: <1A95337B-C4C8-4119-B99E-0C95FA52EA63@nostrum.com>
Mime-Version: 1.0 (Apple Message framework v1082)
X-Mailer: Apple Mail (2.1082)
Received-SPF: pass (nostrum.com: 130.129.18.235 is authenticated by a trusted mechanism)
Subject: [dispatch] Fwd: [sixpac] SIXPAC group not chartered - next steps for the sixpac list
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 31 Mar 2011 09:01:51 -0000

--Apple-Mail-66--1015024056
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii

fyi -

Begin forwarded message:

> From: Robert Sparks 
> Date: March 31, 2011 10:12:24 AM GMT+02:00
> To: sixpac@ietf.org
> Subject: [sixpac] SIXPAC group not chartered - next steps for the =
sixpac list
>=20
> Folks -
>=20
> Based on the list activity for the last few months and the discussions =
this week, we are not going
> to move forward with chartering a working group for this effort. This =
list will remain open for those
> wishing to continue to discuss these ideas and work on individual =
drafts.
>=20
> RjS
> _______________________________________________
> sixpac mailing list
> sixpac@ietf.org
> https://www.ietf.org/mailman/listinfo/sixpac


--Apple-Mail-66--1015024056
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=us-ascii

fyi =
-

Begin forwarded message:

From: Robert Sparks =
<rjsparks@nostrum.com>
Date: March 31, 2011 =
10:12:24 AM GMT+02:00
Subject: [sixpac] SIXPAC =
group not chartered - next steps for the sixpac =
list

Folks -

Based on the list =
activity for the last few months and the discussions this week, we are =
not going
to move forward with chartering a working group for this =
effort. This list will remain open for those
wishing to continue to =
discuss these ideas and work on individual =
drafts.

RjS
_______________________________________________
s=
ixpac mailing list
sixpac@ietf.org
https://www.ietf.or=
g/mailman/listinfo/sixpac

=

--Apple-Mail-66--1015024056--

From christer.holmberg@ericsson.com  Thu Mar 31 06:29:20 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1D7623A6AB2 for ; Thu, 31 Mar 2011 06:29:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.564
X-Spam-Level: 
X-Spam-Status: No, score=-6.564 tagged_above=-999 required=5 tests=[AWL=0.035,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BXMWjSxwjZdw for ; Thu, 31 Mar 2011 06:29:19 -0700 (PDT)
Received: from mailgw9.se.ericsson.net (mailgw9.se.ericsson.net [193.180.251.57]) by core3.amsl.com (Postfix) with ESMTP id D54553A67E9 for ; Thu, 31 Mar 2011 06:29:18 -0700 (PDT)
X-AuditID: c1b4fb39-b7c6dae0000023f2-9e-4d94821190d2
Received: from esessmw0247.eemea.ericsson.se (Unknown_Domain [153.88.253.125]) by mailgw9.se.ericsson.net (Symantec Mail Security) with SMTP id 92.30.09202.112849D4; Thu, 31 Mar 2011 15:30:57 +0200 (CEST)
Received: from ESESSCMS0356.eemea.ericsson.se ([169.254.1.30]) by esessmw0247.eemea.ericsson.se ([10.2.3.116]) with mapi; Thu, 31 Mar 2011 15:30:57 +0200
From: Christer Holmberg 
To: "rtc-web@alvestrand.no" , "dispatch@ietf.org" 
Date: Thu, 31 Mar 2011 15:30:55 +0200
Thread-Topic: IETF#80: RTCWEB BoF notes
Thread-Index: AQHL76fc8hk0DpQf60Wyk1+RgOxqpg==
Message-ID: <7F2072F1E0DE894DA4B517B93C6A05851948D7E0B2@ESESSCMS0356.eemea.ericsson.se>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-Brightmail-Tracker: AAAAAA==
Subject: [dispatch] IETF#80: RTCWEB BoF notes
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 31 Mar 2011 13:29:20 -0000

Hi,

Below are my notes from the RTCWEB BoF.

Regards,

Christer

---------------

Collaboration with W3C:
--------------------------

- It was noted that while IETF has the network expertise, W3C has the API e=
xpertise, and that collaboration between the group is important.
- It was indicated, as there are people active in both IETF and W3C, it is =
better to rely on active collaboration driven by individuals, rather than s=
ending liaisons between the SDOs.
- Is was questioned whether IETF and W3C have different IPR policies.
- It was indicated that the security model also needs to be defined in coll=
aboration with W3C, as there are web security aspects that IETF might not h=
ave good knowledge of.
- There was a question whether document produced by IETF and W3C needs to b=
e approved by both SDOs. It was indicated that hopefully there will not be =
a need for both groups to formally approve documents of the other group, bu=
t each SDO should follow and review the work of the other group.
- It was commented that we might need to establish a collaboration also wit=
h the WHAT WG, rather than relying on W3C for the WHAT WG collaboration.


Use-cases and requirements:
-----------------------------

Presenter: Christer Holmberg

- The presenter indicated that we need to, in order to produce API requirem=
ents, agree on the functional split between the browser and the web-app.
- The presenter indicated that NAT/FW traversal also contains a mechanism t=
o perform media fallback (e.g. HTTP fallback).


RTC-Web Framework:
-------------------------

Presenter: Jonathan Rosenberg

- The presenter indicated that, as the browser and web-app in most cases ar=
e produced by different organizations, we should look at the API between th=
e browser and web-app as a protocol.
- It was commented that we need to be careful with terminology, as an API d=
oes not dictate a protocol.
- It was commented that the API should not be too complex for web applicati=
on. It was suggested that the API might have different levels of complexity=
.
- It was indicated that the browser application itself might not implement =
the features it offers to the web-app. Some features might also be offered =
by the OS, where the broswer simply provides access to those features to th=
e web-app.


Web Security:
---------------

Presenter: Eric Rescorla

- The following security areas, related to rtc-web, were identified:
-- Media remote peer verification.
-- Access to local device.
-- Communication security.
- There was a comment that identity also needs to be covered. However, in a=
ll use-cases identity might not be needed, or even desired.
- It was questioned how we can prevent an application, claiming to be a bro=
wser, from sending date before getting consent. It was indicated that such =
scenario is outside the scope of the WG.


Negotiation and Extensibility
----------------------------

Presenter: Cullen Jennings

- The presenter indicated that the solution must be extendible, and it must=
 provide a mechanism which allows the negotiation of different features.
- It was indicated that full legacy interoperability might not be possible,=
 mostly due to security constraints, without intermediary functions on the =
media plane.
- It was indicated that the WG needs to decide on the level on legacy inter=
operability.


Charter:
--------
- A large number of individuals had read the proposed charter text.
- There was a question whether document produced by IETF and W3C needs to b=
e approved by both SDOs. It was indicated that hopefully there will not be =
a need for both groups to formally approve documents of the other group, bu=
t each SDO should follow and review the work of the other group.
- It was commented that the charter does not talk about legacy interoperabi=
lity.
- It was commented that none of the presentations have described the handli=
ng of non-RTP connections. It was indicated that the MG might want to consi=
der sending also non-audio/video data over RTP.
- It was commented that the charter should not contain a list of features t=
hat might be added to the charter at a later point.

Poll:
-----

- Willingness to review documents: approx 50-60 individuals.
- Willingness to write and provide text to documents: approx 24 individuals=
.=

From R.Jesske@telekom.de  Thu Mar 31 09:19:32 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 666BB3A684E for ; Thu, 31 Mar 2011 09:19:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.249
X-Spam-Level: 
X-Spam-Status: No, score=-3.249 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_DE=0.35, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zAV+vnIIKi8h for ; Thu, 31 Mar 2011 09:19:31 -0700 (PDT)
Received: from tcmail73.telekom.de (tcmail73.telekom.de [217.243.239.135]) by core3.amsl.com (Postfix) with ESMTP id 4EB1728C12B for ; Thu, 31 Mar 2011 09:19:31 -0700 (PDT)
Received: from he111631.emea1.cds.t-internal.com ([10.134.93.23]) by tcmail71.telekom.de with ESMTP/TLS/AES128-SHA; 31 Mar 2011 18:21:06 +0200
Received: from HE111648.emea1.cds.t-internal.com ([169.254.5.70]) by HE111631.emea1.cds.t-internal.com ([::1]) with mapi; Thu, 31 Mar 2011 18:21:06 +0200
From: 
To: 
Date: Thu, 31 Mar 2011 18:21:05 +0200
Thread-Topic: New Version of draft-jesske-dispatch-update3326-reason-responses-02 
Thread-Index: AcvvvwHNRNVWxh9jQh2I8gtnbCDbVgAAChRA
Message-ID: <580BEA5E3B99744AB1F5BFF5E9A3C67D017102BE24@HE111648.emea1.cds.t-internal.com>
Accept-Language: en-US, de-DE
Content-Language: de-DE
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US, de-DE
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: [dispatch] New Version of draft-jesske-dispatch-update3326-reason-responses-02
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 31 Mar 2011 16:19:32 -0000

Dear all,
please be informed that I have uploaded a new version of draft-jesske-dispa=
tch-update3326-reason-responses-02

http://www.ietf.org/id/draft-jesske-dispatch-update3326-reason-responses-02=
.txt

Comments are welcome

Best Regards

Roland



> -----Urspr=FCngliche Nachricht-----
> Von: IETF I-D Submission Tool [mailto:idsubmission@ietf.org]
> Gesendet: Donnerstag, 31. M=E4rz 2011 18:15
> An: Jesske, Roland
> Cc: Laura.Liess@telekom.de
> Betreff: New Version Notification for
> draft-jesske-dispatch-update3326-reason-responses-02
>
>
> A new version of I-D,
> draft-jesske-dispatch-update3326-reason-responses-02.txt has
> been successfully submitted by Roland Jesske and posted to
> the IETF repository.
>
> Filename:      draft-jesske-dispatch-update3326-reason-responses
> Revision:      02
> Title:                 Reason header filed in Session
> Initiation Protocol (SIP) responses
> Creation_date:         2011-03-31
> WG ID:                 Independent Submission
> Number_of_pages: 5
>
> Abstract:
> This document updates the Reason mechanism defined within RFC3326.
>
>
>
>
> The IETF Secretariat.
>
>
>

From fluffy@cisco.com  Thu Mar 31 09:45:24 2011
Return-Path: 
X-Original-To: dispatch@core3.amsl.com
Delivered-To: dispatch@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5A04C3A699E for ; Thu, 31 Mar 2011 09:45:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.599
X-Spam-Level: 
X-Spam-Status: No, score=-110.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pRPxe3830Tzc for ; Thu, 31 Mar 2011 09:45:23 -0700 (PDT)
Received: from ams-iport-1.cisco.com (ams-iport-1.cisco.com [144.254.224.140]) by core3.amsl.com (Postfix) with ESMTP id 0D94A3A684E for ; Thu, 31 Mar 2011 09:45:21 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=fluffy@cisco.com; l=29154; q=dns/txt; s=iport; t=1301590022; x=1302799622; h=from:subject:date:references:to:message-id:mime-version; bh=no60zD26toNc9KaiihujoiqfhyZrnLMGBAc7S02ELfg=; b=VTaM2OCgT5X/1Xif3mfuNP/y4ImFnGjUjnjc5hPUtxqKj9CJQFoLDk4k xPCqaKfSxgVU9YTYRk5VqKzZkpFQP8EqJcN1Q15/mFtwHves1tQRIzOLV Ti1+Pw/Ym+CeTrADm+yXQrAnt/pzvkJReE6rgCX26IW6ZgtL5NHZ1OIFz c=;
X-Files: load1.ods : 18964
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AtYAAEmvlE2Q/khNgWdsb2JhbAAul2SNQBQBARYmJYh5mk2cMYVrBI0R
X-IronPort-AV: E=Sophos;i="4.63,276,1299456000";  d="xml'?ods'72,48?png'72,48,150?scan'72,48,150,72,48,208,150"; a="81690519"
Received: from ams-core-4.cisco.com ([144.254.72.77]) by ams-iport-1.cisco.com with ESMTP; 31 Mar 2011 16:47:01 +0000
Received: from dhcp-10-61-105-191.cisco.com (dhcp-10-61-105-191.cisco.com [10.61.105.191]) by ams-core-4.cisco.com (8.14.3/8.14.3) with ESMTP id p2VGl0gK018601 for ; Thu, 31 Mar 2011 16:47:00 GMT
From: Cullen Jennings 
Content-Type: multipart/mixed; boundary=Apple-Mail-2--987210070
Date: Thu, 31 Mar 2011 18:47:00 +0200
References: <201103311630.p2VGUrVt000845@eye.ariadne.com>
To: DISPATCH list 
Message-Id: 
Mime-Version: 1.0 (Apple Message framework v1084)
X-Mailer: Apple Mail (2.1084)
Subject: [dispatch] Fwd:  SIP load balancing
X-BeenThere: dispatch@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: DISPATCH Working Group Mail List 
List-Unsubscribe: , 
List-Archive: 
List-Post: 
List-Help: 
List-Subscribe: , 
X-List-Received-Date: Thu, 31 Mar 2011 16:45:24 -0000

--Apple-Mail-2--987210070
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=us-ascii


Forwarding this on behalf of Dale. 

Begin forwarded message:

> From: Dale Worley 
> Date: March 31, 2011 6:30:53 PM GMT+02:00
> To: Cullen Jennings 
> Subject: RE: [dispatch] SIP load balancing
> 
> I think there has been a problem sending this message, so I am resending it
> for a second time.
> 
> Dale
> ________________________________________
> From: Worley, Dale R (Dale)
> Sent: Wednesday, March 30, 2011 4:21 AM
> To: dispatch@ietf.org
> Subject: RE: [dispatch] SIP load balancing
> 
> I have been arguing that load balancing algorithms have more
> complicated behavior than people have been allowing for, and that
> therefore the analysis needed before adopting a load balancing system
> is more extensive than people think.
> 
> As a demonstration, I have written a spreadsheet that simulates the
> behavior of the current proposed scheme and adjusted the parameters
> for a situation where the offered load is 90% of capacity:
> 
>    There are two servers, one that can handle 100 calls and one that can
>    handle 10 calls.
> 
>    The offered load is 5 calls per time unit.  The calls have lifetimes
>    that have an exponential distribution with an average lifetime of 20
>    time units, and so in any time unit, 1/20 of the calls in the system
>    are cleared.  Thus, the steady-state of the system has 100 calls in
>    the system.
> 
>    In steady-state, the system operates at 90% of total capacity.
> 
>    Calls are treated as infinitely divisible, so there is no statistical
>    fluctuation in the loads.
> 
> The spreadsheet is attached to this e-mail, and is also available at
> http://theworld.com/~worley/load1.ods
> It is in ODF format, which is an open standard, and can be read by
> both OpenOffice and Excel.
> 
> The computations involved are incorporated into the spreadsheet, and
> you can change the spreadsheet to calculate the results under
> different assumptions.
> 
> In the selected case, one server has 10 times the capacity of the
> other server, and the system is offered 90% of the load that it can
> process.  Naively, one expects that a good load balancing scheme would
> split the load so that both servers are at about 90% of capacity, and
> that the load is split in approximately a 90%-10% proportion.
> 
> As you can see from the simulation, the steady-state has the load
> split 87.8-9.9, which is close to 90%-10%, but it does not load both
> servers at about 90% capacity -- server 1 is at 88% capacity but
> server 2 is at 99% capacity.
> 
> Dale

--Apple-Mail-2--987210070
Content-Disposition: attachment;
	filename=load1.ods
Content-Type: application/vnd.oasis.opendocument.spreadsheet;
	name="load1.ods"
Content-Transfer-Encoding: base64
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--Apple-Mail-2--987210070
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=us-ascii



Cullen Jennings
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/index.html



--Apple-Mail-2--987210070--