University of Oslo

PO Box 1080 Blindern N-0316 Oslo Norway +47 22 84 08 37 safiquli@ifi.uio.no

University of Oslo

PO Box 1080 Blindern N-0316 Oslo Norway +47 22 85 24 20 michawe@ifi.uio.no

University of Oslo

PO Box 1080 Blindern N-0316 Oslo Norway +47 22 85 24 44 steing@ifi.uio.no

Transport RTP Media Congestion Avoidance Techniques (rmcat) tcp When multiple congestion-controlled Real-time Transport Protocol (RTP) sessions traverse the same network bottleneck, combining their controls can improve the total on-the-wire behavior in terms of delay, loss, and fairness. This document describes such a method for flows that have the same sender, in a way that is as flexible and simple as possible while minimizing the number of changes needed to existing RTP applications. This document also specifies how to apply the method for the Network-Assisted Dynamic Adaptation (NADA) congestion control algorithm and provides suggestions on how to apply it to other congestion control algorithms.

Introduction When there is enough data to send, a congestion controller attempts to increase its sending rate until the path's capacity has been reached. Some controllers detect path capacity by increasing the sending rate further, until packets are ECN-marked or dropped, and then decreasing the sending rate until that stops happening. This process inevitably creates undesirable queuing delay when multiple congestion-controlled connections traverse the same network bottleneck, and each connection overshoots the path capacity as it determines its sending rate. The Congestion Manager (CM) couples flows by providing a single congestion controller. It is hard to implement because it requires an additional congestion controller and removes all per-connection congestion control functionality, which is quite a significant change to existing RTP-based applications. This document presents a method to combine the behavior of congestion control mechanisms that is easier to implement than the Congestion Manager and also requires fewer significant changes to existing RTP-based applications. It attempts to roughly approximate the CM behavior by sharing information between existing congestion controllers. It is able to honor user-specified priorities, which is required by WebRTC . The described mechanisms are believed safe to use, but they are experimental and are presented for wider review and operational evaluation.

Definitions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.

Available Bandwidth:

The available bandwidth is the nominal link capacity minus the amount of traffic that traversed the link during a certain time interval, divided by that time interval.

Bottleneck:

The first link with the smallest available bandwidth along the path between a sender and receiver.

Flow:

A flow is the entity that congestion control is operating on. It could, for example, be a transport-layer connection or an RTP stream , regardless of whether or not this RTP stream is multiplexed onto an RTP session with other RTP streams.

Flow Group Identifier (FGI):

A unique identifier for each subset of flows that is limited by a common bottleneck.

Flow State Exchange (FSE):

The entity that maintains information that is exchanged between flows.

Flow Group (FG):

A group of flows having the same FGI.

Shared Bottleneck Detection (SBD):

The entity that determines which flows traverse the same bottleneck in the network or the process of doing so.

Limitations

Sender-side only:

Shared bottlenecks can exist when multiple flows originate from the same sender or when flows from different senders reach the same receiver (see ). Coupled congestion control, as described here, only supports the former case, not the latter, as it operates inside a single host on the sender side.

Shared bottlenecks do not change quickly:

As per the definition above, a bottleneck depends on cross traffic, and since such traffic can heavily fluctuate, bottlenecks can change at a high frequency (e.g., there can be oscillation between two or more links). This means that, when flows are partially routed along different paths, they may quickly change between sharing and not sharing a bottleneck. For simplicity, here it is assumed that a shared bottleneck is valid for a time interval that is significantly longer than the interval at which congestion controllers operate. Note that, for the only SBD mechanism defined in this document (multiplexing on the same five-tuple), the notion of a shared bottleneck stays correct even in the presence of fast traffic fluctuations; since all flows that are assumed to share a bottleneck are routed in the same way, if the bottleneck changes, it will still be shared.

Architectural Overview shows the elements of the architecture for coupled congestion control: the Flow State Exchange (FSE), Shared Bottleneck Detection (SBD), and Flows. The FSE is a storage element that can be implemented in two ways: active and passive. In the active version, it initiates communication with flows and SBD. However, in the passive version, it does not actively initiate communication with flows and SBD; its only active role is internal state maintenance (e.g., an implementation could use soft state to remove a flow's data after long periods of inactivity). Every time a flow's congestion control mechanism would normally update its sending rate, the flow instead updates information in the FSE and performs a query on the FSE, leading to a sending rate that can be different from what the congestion controller originally determined. Using information about/from the currently active flows, SBD updates the FSE with the correct Flow Group Identifiers (FGIs). This document describes both active and passive versions. While the passive algorithm works better for congestion controls with RTT-independent convergence, it can still produce oscillations on short time scales. The passive algorithm, described in , is therefore considered highly experimental and not safe to deploy outside of testbed environments. shows the interaction between flows and the FSE using the variable names defined in .

REGISTER REGISTER <--register-- JOIN #1 #2 CC_R(1) ----UPDATE----> UPDATE (in) #3 NEW RATE <---FSE_R(1)-- UPDATE (out) --FSE_R(2)-> #3 NEW RATE ]]>

Since everything shown in is assumed to operate on a single host (the sender) only, this document only describes aspects that have an influence on the resulting on-the-wire behavior. It does not, for instance, define how many bits must be used to represent FGIs or in which way the entities communicate. Implementations can take various forms; for instance, all the elements in the figure could be implemented within a single application, thereby operating on flows generated by that application only. Another alternative could be to implement both the FSE and SBD together in a separate process that different applications communicate with via some form of Inter-Process Communication (IPC). Such an implementation would extend the scope to flows generated by multiple applications. The FSE and SBD could also be included in the Operating System kernel. However, only one type of coupling algorithm should be used for all flows. Combinations of multiple algorithms at different aggregation levels (e.g., the Operating System coupling application aggregates with one algorithm, and applications coupling their flows with another) have not been tested and are therefore not recommended.

Roles This section gives an overview of the roles of the elements of coupled congestion control and provides an example of how coupled congestion control can operate.

SBD SBD uses knowledge about the flows to determine which flows belong in the same Flow Group (FG) and assigns FGIs accordingly. This knowledge can be derived in three basic ways:

1. From multiplexing: It can be based on the simple assumption that packets sharing the same five-tuple (IP source and destination address, protocol, and transport-layer port number pair) and having the same values for the Differentiated Services Code Point (DSCP) and the ECN field in the IP header are typically treated in the same way along the path. This method is the only one specified in this document; SBD MAY consider all flows that use the same five-tuple, DSCP, and ECN field value to belong to the same FG. This classification applies to certain tunnels or RTP flows that are multiplexed over one transport (cf. ). Such multiplexing is also a recommended usage of RTP in WebRTC .
2. Via configuration: e.g., by assuming that a common wireless uplink is also a shared bottleneck.
3. From measurements: e.g., by considering correlations among measured delay and loss as an indication of a shared bottleneck.

The methods above have some essential trade-offs. For example, multiplexing is a completely reliable measure, but it is limited in scope to two endpoints (i.e., it cannot be applied to couple congestion controllers of one sender talking to multiple receivers). A measurement-based SBD mechanism is described in . Measurements can never be 100% reliable, in particular because they are based on the past, but applying coupled congestion control involves making an assumption about the future; it is therefore recommended to implement cautionary measures, e.g., by disabling coupled congestion control if enabling it causes a significant increase in delay and/or packet loss. Measurements also take time, which entails a certain delay for turning on coupling (refer to for details). When this is possible, it can be more efficient to statically configure shared bottlenecks (e.g., via a system configuration or user input) based on assumptions about the network environment.

FSE The FSE contains a list of all flows that have registered with it. For each flow, the FSE stores the following:

• a unique flow number f to identify the flow.
• the FGI of the FG that it belongs to (based on the definitions in this document, a flow has only one bottleneck and can therefore be in only one FG).
• a priority P(f), which is a number greater than zero.
• The rate used by the flow in bits per second, FSE_R(f).
• The desired rate DR(f) of flow f. This can be smaller than FSE_R(f) if the application feeding into the flow has less data to send than FSE_R(f) would allow or if a maximum value is imposed on the rate. In the absence of such limits, DR(f) must be set to the sending rate provided by the congestion control module of flow f.

Note that the absolute range of priorities does not matter; the algorithm works with a flow's priority portion of the sum of all priority values. For example, if there are two flows, flow 1 with priority 1 and flow 2 with priority 2, the sum of the priorities is 3. Then, flow 1 will be assigned 1/3 of the aggregate sending rate, and flow 2 will be assigned 2/3 of the aggregate sending rate. Priorities can be mapped to the "very-low", "low", "medium", or "high" priority levels described in by simply using the values 1, 2, 4, and 8, respectively. In the FSE, each FG contains one static variable, S_CR, which is the sum of the calculated rates of all flows in the same FG. This value is used to calculate the sending rate. The information listed here is enough to implement the sample flow algorithm given below. FSE implementations could easily be extended to store, e.g., a flow's current sending rate for statistics gathering or future potential optimizations.

Flows Flows register themselves with SBD and FSE when they start, deregister from the FSE when they stop, and carry out an UPDATE function call every time their congestion controller calculates a new sending rate. Via UPDATE, they provide the newly calculated rate and, optionally (if the algorithm supports it), the desired rate. The desired rate is less than the calculated rate in case of application-limited flows; otherwise, it is the same as the calculated rate. Below, two example algorithms are described. While other algorithms could be used instead, the same algorithm must be applied to all flows. Names of variables used in the algorithms are explained below.

CC_R(f)

The rate received from the congestion controller of flow f when it calls UPDATE.

FSE_R(f)

The rate calculated by the FSE for flow f.

DR(f)

The desired rate of flow f.

S_CR

The sum of the calculated rates of all flows in the same FG; this value is used to calculate the sending rate.

FG

A group of flows having the same FGI and hence, sharing the same bottleneck.

P(f)

The priority of flow f, which is received from the flow's congestion controller; the FSE uses this variable for calculating FSE_R(f).

S_P

The sum of all the priorities.

TLO

The total leftover rate; the sum of rates that could not be assigned to flows that were limited by their desired rate.

AR

The aggregate rate that is assigned to flows that are not limited by their desired rate.

Example Algorithm 1 - Active FSE This algorithm was designed to be the simplest possible method to assign rates according to the priorities of flows. Simulation results in indicate that it does not, however, significantly reduce queuing delay and packet loss.

1. When a flow f starts, it registers itself with SBD and the FSE. FSE_R(f) is initialized with the congestion controller's initial rate. SBD will assign the correct FGI. When a flow is assigned an FGI, it adds its FSE_R(f) to S_CR.
2. When a flow f stops or pauses, its entry is removed from the list.
3. Every time the congestion controller of the flow f determines a new sending rate CC_R(f), the flow calls UPDATE, which carries out the tasks listed below to derive the new sending rates for all the flows in the FG. A flow's UPDATE function uses three local (i.e., per-flow) temporary variables: S_P, TLO, and AR.
   1. It updates S_CR.
   2. It calculates the sum of all the priorities, S_P, and initializes FSE_R.
   3. It distributes S_CR among all flows, ensuring that each flow's desired rate is not exceeded. 0 and S_P>0) AR = 0 for all flows i in FG do if FSE_R[i] < DR[i] then if TLO * P[i] / S_P >= DR[i] then TLO = TLO - DR[i] FSE_R[i] = DR[i] S_P = S_P - P[i] else FSE_R[i] = TLO * P[i] / S_P AR = AR + TLO * P[i] / S_P end if end if end for end while ]]>
   4. It distributes FSE_R to all the flows.

Example Algorithm 2 - Conservative Active FSE This algorithm changes algorithm 1 to conservatively emulate the behavior of a single flow by proportionally reducing the aggregate rate on congestion. Simulation results in indicate that it can significantly reduce queuing delay and packet loss. Step (a) of the UPDATE function is changed as described below. This also introduces a local variable DELTA, which is used to calculate the difference between CC_R(f) and the previously stored FSE_R(f). To prevent flows from either ignoring congestion or overreacting, a timer keeps them from changing their rates immediately after the common rate reduction that follows a congestion event. This timer is set to two RTTs of the flow that experienced congestion because it is assumed that a congestion event can persist for up to one RTT of that flow, with another RTT added to compensate for fluctuations in the measured RTT value.

1. It updates S_CR based on DELTA.

Application This section specifies how the FSE can be applied to specific congestion control mechanisms and makes general recommendations that facilitate applying the FSE to future congestion controls.

NADA Network-Assisted Dynamic Adaptation (NADA) is a congestion control scheme for WebRTC. It calculates a reference rate r_ref upon receiving an acknowledgment and then, based on the reference rate, calculates a video target rate r_vin and a sending rate for the flows, r_send. When applying the FSE to NADA, the UPDATE function call described in gives the FSE NADA's reference rate r_ref. The recommended algorithm for NADA is the Active FSE in . In step 3 (d), when the FSE_R(i) is "sent" to the flow i, r_ref (r_vin and r_send) of flow i is updated with the value of FSE_R(i).

General Recommendations This section provides general advice for applying the FSE to congestion control mechanisms.

Receiver-side calculations:

When receiver-side calculations make assumptions about the rate of the sender, the calculations need to be synchronized, or the receiver needs to be updated accordingly. This applies to TCP Friendly Rate Control (TFRC) , for example, where simulations showed somewhat less favorable results when using the FSE without a receiver-side change .

Stateful algorithms:

When a congestion control algorithm is stateful (e.g., during the TCP slow start, congestion avoidance, or fast recovery phase), these states should be carefully considered such that the overall state of the aggregate flow is correct. This may require sharing more information in the UPDATE call.

Rate jumps:

The FSE-based coupling algorithms can let a flow quickly increase its rate to its fair share, e.g., when a new flow joins or after a quiescent period. In case of window-based congestion controls, this may produce a burst that should be mitigated in some way. An example of how this could be done without using a timer is presented in , using TCP as an example.

Expected Feedback from Experiments The algorithm described in this memo has so far been evaluated using simulations covering all the tests for more than one flow from (see and ). Experiments should confirm these results using at least the NADA congestion control algorithm with real-life code (e.g., browsers communicating over an emulated network covering the conditions in ). The tests with real-life code should be repeated afterwards in real network environments and monitored. Experiments should investigate cases where the media coder's output rate is below the rate that is calculated by the coupling algorithm (FSE_R(i) in algorithms 1 () and 2 ()). Implementers and testers are invited to document their findings in an Internet-Draft.

IANA Considerations This document has no IANA actions.

Security Considerations In scenarios where the architecture described in this document is applied across applications, various cheating possibilities arise, e.g., supporting wrong values for the calculated rate, desired rate, or priority of a flow. In the worst case, such cheating could either prevent other flows from sending or make them send at a rate that is unreasonably large. The end result would be unfair behavior at the network bottleneck, akin to what could be achieved with any UDP-based application. Hence, since this is no worse than UDP in general, there seems to be no significant harm in using this in the absence of UDP rate limiters. In the case of a single-user system, it should also be in the interest of any application programmer to give the user the best possible experience by using reasonable flow priorities or even letting the user choose them. In a multi-user system, this interest may not be given, and one could imagine the worst case of an "arms race" situation where applications end up setting their priorities to the maximum value. If all applications do this, the end result is a fair allocation in which the priority mechanism is implicitly eliminated and no major harm is done. Implementers should also be aware of the Security Considerations sections of , , and .

References Normative References Informative References ACM SIGCOMM Capacity Sharing Workshop (CSWS 2014) and ACM SIGCOMM CCR 44(4) 2014 IEEE NOMS 2016 RTP Media Congestion Avoidance Techniques (rmcat) Working Group RTP Media Congestion Avoidance Techniques (rmcat) Working Group ACM, IRTF, ISOC Applied Networking Research Workshop 2016 (ANRW 2016)

Application to GCC Google Congestion Control (GCC) is another congestion control scheme for RTP flows that is under development. GCC is not yet finalized, but at the time of this writing, the rate control of GCC employs two parts: controlling the bandwidth estimate based on delay and controlling the bandwidth estimate based on loss. Both are designed to estimate the available bandwidth, A_hat. When applying the FSE to GCC, the UPDATE function call described in gives the FSE GCC's estimate of available bandwidth A_hat. The recommended algorithm for GCC is the Active FSE in . In step 3 (d) of this algorithm, when the FSE_R(i) is "sent" to the flow i, A_hat of flow i is updated with the value of FSE_R(i).

Scheduling When flows originate from the same host, it would be possible to use only one sender-side congestion controller that determines the overall allowed sending rate and then use a local scheduler to assign a proportion of this rate to each RTP session. This way, priorities could also be implemented as a function of the scheduler. The Congestion Manager (CM) also uses such a scheduling function.

Example Algorithm - Passive FSE Active algorithms calculate the rates for all the flows in the FG and actively distribute them. In a passive algorithm, UPDATE returns a rate that should be used instead of the rate that the congestion controller has determined. This can make a passive algorithm easier to implement; however, when round-trip times of flows are unequal, flows with shorter RTTs may (depending on the congestion control algorithm) update and react to the overall FSE state more often than flows with longer RTTs, which can produce unwanted side effects. This problem is more significant when the congestion control convergence depends on the RTT. While the passive algorithm works better for congestion controls with RTT-independent convergence, it can still produce oscillations on short time scales. The algorithm described below is therefore considered highly experimental and not safe to deploy outside of testbed environments. Results of a simplified passive FSE algorithm with both NADA and GCC can be found in . In the passive version of the FSE, TLO (Total Leftover Rate) is a static variable per FG that is initialized to 0. Additionally, S_CR is limited to increase or decrease as conservatively as a flow's congestion controller decides in order to prohibit sudden rate jumps.

1. When a flow f starts, it registers itself with SBD and the FSE. FSE_R(f) and DR(f) are initialized with the congestion controller's initial rate. SBD will assign the correct FGI. When a flow is assigned an FGI, it adds its FSE_R(f) to S_CR.
2. When a flow f stops or pauses, it sets its DR(f) to 0 and sets P(f) to -1.
3. Every time the congestion controller of the flow f determines a new sending rate CC_R(f), assuming the flow's new desired rate new_DR(f) to be "infinity" in case of a bulk data transfer with an unknown maximum rate, the flow calls UPDATE, which carries out the tasks listed below to derive the flow's new sending rate, Rate(f). A flow's UPDATE function uses a few local (i.e., per-flow) temporary variables, which are all initialized to 0: DELTA, new_S_CR, and S_P.
   1. For all the flows in its FG (including itself), it calculates the sum of all the calculated rates, new_S_CR. Then, it calculates DELTA: the difference between FSE_R(f) and CC_R(f).
   2. It updates S_CR, FSE_R(f), and DR(f). 0 then // the flow's rate has increased S_CR = S_CR + DELTA else if DELTA < 0 then S_CR = new_S_CR + DELTA end if DR(f) = min(new_DR(f),FSE_R(f)) ]]>
   3. It calculates the leftover rate TLO, removes the terminated flows from the FSE, and calculates the sum of all the priorities, S_P.
   4. It calculates the sending rate, Rate(f). 0 then TLO = 0 // f has 'taken' TLO end if ]]>
   5. It updates DR(f) and FSE_R(f) with Rate(f). DR(f) then DR(f) = Rate(f) end if FSE_R(f) = Rate(f) ]]>

The goals of the flow algorithm are to achieve prioritization, improve network utilization in the face of application-limited flows, and impose limits on the increase behavior such that the negative impact of multiple flows trying to increase their rate together is minimized. It does that by assigning a flow a sending rate that may not be what the flow's congestion controller expected. It therefore builds on the assumption that no significant inefficiencies arise from temporary application-limited behavior or from quickly jumping to a rate that is higher than the congestion controller intended. How problematic these issues really are depends on the controllers in use and requires careful per-controller experimentation. The coupled congestion control mechanism described here also does not require all controllers to be equal; effects of heterogeneous controllers, or homogeneous controllers being in different states, are also subject to experimentation. This algorithm gives the leftover rate of application-limited flows to the first flow that updates its sending rate, provided that this flow needs it all (otherwise, its own leftover rate can be taken by the next flow that updates its rate). Other policies could be applied, e.g., to divide the leftover rate of a flow equally among all other flows in the FGI.

Example Operation (Passive) In order to illustrate the operation of the passive coupled congestion control algorithm, this section presents a toy example of two flows that use it. Let us assume that both flows traverse a common 10 Mbit/s bottleneck and use a simplistic congestion controller that starts out with 1 Mbit/s, increases its rate by 1 Mbit/s in the absence of congestion, and decreases it by 2 Mbit/s in the presence of congestion. For simplicity, flows are assumed to always operate in a round-robin fashion. Rate numbers below without units are assumed to be in Mbit/s. For illustration purposes, the actual sending rate is also shown for every flow in FSE diagrams even though it is not really stored in the FSE. Flow #1 begins. It is a bulk data transfer and considers itself to have top priority. This is the FSE after the flow algorithm's step 1: Its congestion controller gradually increases its rate. Eventually, at some point, the FSE should look like this: Now, another flow joins. It is also a bulk data transfer and has a lower priority (0.5): Now, assume that the first flow updates its rate to 8, because the total sending rate of 11 exceeds the total capacity. Let us take a closer look at what happens in step 3 of the flow algorithm. CC_R(1) = 8. new_DR(1) = infinity.

1. new_S_CR = 11; DELTA = 8 - 10 = -2.
2. FSE_R(1) = 8. DELTA is negative, hence S_CR = 9; DR(1) = 8
3. S_P = 1.5.
4. new sending rate Rate(1) = min(infinity, 1/1.5 * 9 + 0) = 6.
5. FSE_R(1) = 6.

The resulting FSE looks as follows: The effect is that flow #1 is sending with 6 Mbit/s instead of the 8 Mbit/s that the congestion controller derived. Let us now assume that flow #2 updates its rate. Its congestion controller detects that the network is not fully saturated (the actual total sending rate is 6+1=7) and increases its rate. CC_R(2) = 2. new_DR(2) = infinity.

1. new_S_CR = 7; DELTA = 2 - 1 = 1.
2. FSE_R(2) = 2. DELTA is positive, hence S_CR = 9 + 1 = 10; DR(2) = 2.
3. S_P = 1.5.
4. Rate(2) = min(infinity, 0.5/1.5 * 10 + 0) = 3.33.
5. DR(2) = FSE_R(2) = 3.33.

The resulting FSE looks as follows: The effect is that flow #2 is now sending with 3.33 Mbit/s, which is close to half of the rate of flow #1 and leads to a total utilization of 6(#1) + 3.33(#2) = 9.33 Mbit/s. Flow #2's congestion controller has increased its rate faster than the controller actually expected. Now, flow #1 updates its rate. Its congestion controller detects that the network is not fully saturated and increases its rate. Additionally, the application feeding into flow #1 limits the flow's sending rate to at most 2 Mbit/s. CC_R(1) = 7. new_DR(1) = 2.

1. new_S_CR = 9.33; DELTA = 1.
2. FSE_R(1) = 7, DELTA is positive, hence S_CR = 10 + 1 = 11; DR(1) = min(2, 7) = 2.
3. S_P = 1.5; DR(1) < FSE_R(1), hence TLO = 1/1.5 * 11 - 2 = 5.33.
4. Rate(1) = min(2, 1/1.5 * 11 + 5.33) = 2.
5. FSE_R(1) = 2.

The resulting FSE looks as follows: Now, the total rate of the two flows is 2 + 3.33 = 5.33 Mbit/s, i.e., the network is significantly underutilized due to the limitation of flow #1. Flow #2 updates its rate. Its congestion controller detects that the network is not fully saturated and increases its rate. CC_R(2) = 4.33. new_DR(2) = infinity.

1. new_S_CR = 5.33; DELTA = 1.
2. FSE_R(2) = 4.33. DELTA is positive, hence S_CR = 12; DR(2) = 4.33.
3. S_P = 1.5.
4. Rate(2) = min(infinity, 0.5/1.5 * 12 + 5.33 ) = 9.33.
5. FSE_R(2) = 9.33, DR(2) = 9.33.

The resulting FSE looks as follows: Now, the total rate of the two flows is 2 + 9.33 = 11.33 Mbit/s. Finally, flow #1 terminates. It sets P(1) to -1 and DR(1) to 0. Let us assume that it terminated late enough for flow #2 to still experience the network in a congested state, i.e., flow #2 decreases its rate in the next iteration. CC_R(2) = 7.33. new_DR(2) = infinity.

1. new_S_CR = 11.33; DELTA = -2.
2. FSE_R(2) = 7.33. DELTA is negative, hence S_CR = 9.33; DR(2) = 7.33.
3. Flow 1 has P(1) = -1, hence it is deleted from the FSE. S_P = 0.5.
4. Rate(2) = min(infinity, 0.5/0.5*9.33 + 0) = 9.33.
5. FSE_R(2) = DR(2) = 9.33.

The resulting FSE looks as follows:

Acknowledgements This document benefited from discussions with and feedback from , , , , (who also gave the FSE its name), , , , , , , , , and . The authors would like to especially thank and for helping with NADA and GCC, and as well as for helping us correct the active algorithm for the case of application-limited flows. This work was partially funded by the European Community under its Seventh Framework Program through the Reducing Internet Transport Latency (RITE) project (ICT-317700).