]> Google LLC

martin.h.duke@gmail.com

University of Aberdeen

gorry@erg.abdn.ac.uk

General CCWG Internet-Draft Introducing new or modified congestion control algorithms in the global Internet has possible ramifications to both the flows using the proposed congestion control algorithms and to flows using a standardized congestion control algorithm. Therefore, the IETF must proceed with caution when evaluating proposals for alternate congestion control. The goal of this document is to provide guidance for considering standardization of a proposed congestion control algorithm at the IETF. It obsoletes RFC5033 to reflect changes in the congestion control landscape. Status information for this document may be found at . Discussion of this document takes place on the Congestion Control Working Group (ccwg) Working Group mailing list (), which is archived at . Subscribe at . Source for this draft and an issue tracker can be found at .

Introduction This document provides guidelines for the IETF to use when evaluating a proposed congestion control algorithm that differs from the general congestion control principles outlined in . The guidance is intended to be useful to authors proposing congestion control algorithms and for the IETF community when evaluating whether a proposal is appropriate for publication in the RFC series and for deployment in the Internet. This document obsoletes , which was published in 2007 as a Best Current Practice to evaluate proposed congestion control algorithms as Experimental or Proposed Standard RFCs. The IETF's standard congestion control algorithms have performance challenges in various environments (e.g., high-speed networks, cellular and WiFi wireless technologies, and long distance satellite links) and also for flOws carrying specific workloads (Voice over IP (VoIP), gaming, and videoconferencing). When was published in 2007, TCP was the dominant consumer of this work, and proposals were typically discussed in the Internet Congestion Control Research Group (ICCRG). The Datagram Congestion Control Protocol (DCCP) was developed as a method for defining new congestion control algorithms for datagram traffic. Since then, many conditions have changed. The set of protocols using these algorithms has spread beyond TCP, Stream Control Transmission Protocol (SCTP) , and DCCP, to include QUIC , RTP Media Congestion Avoidance Techniques (RMCAT) (e.g., ) and beyond. Some proponents of alternative congestion control algorithms now have the opportunity to test and deploy at scale without IETF review. There is more interest in specialized use cases, such as data centers (e.g., ), and in support for a variety of upper layer protocols/applications, e.g., real-time protocols. Finally, the community has gained much more experience with indications of congestion beyond packet loss. Multicast congestion control is a considerably less mature field of study and are not in scope for this document. However, provides additional guidelines for multicast and broadcast usage of UDP. Congestion control algorithms have been developed outside of the IETF, including at least two that saw large scale deployment: Cubic and Bottleneck Bandwidth and Round-trip propagation time (BBR) . Cubic was documented in a research publication in 2007 , and was then adopted as the default congestion control algorithm for the TCP implementation in Linux. It was already used in a significant fraction of TCP connections over the Internet before being documented in an Informational Internet-Draft in 2015, published as an Informational RFC in 2017 and then as a Proposed Standard in 2023 . At the time of writing, BBR is being developed as an internal research project by Google, with the first implementation contributed to Linux kernel 4.19 in 2016. It was described in an IRTF Internet-Draft in 2018, and that Internet- Draft is regularly updated to document the evolving versions of the algorithm . BBR is currently widely used for Google services using either TCP or QUIC, and is also widely deployed outside of Google. We cannot say now whether the original authors of expected that developers would be somehow waiting for IETF review before widely deploying a new congestion control algorithm over the Internet, but the examples of Cubic and BBR teach us that deployment of new algorithms is not in fact gated by the publication of the algorithm as an RFC. Nevertheless, specifying congestion control algorithms has a number of advantages: A specification can help implementers, operators, and other interested parties develop a shared understanding of how the algorithm works and how it is expected to behave in various scenarios and configurations. A specification can help potential contributors understand the algorithm, which can make it easier for them to suggest improvements and/or identify limitations. Furthermore, the specification can help multiple contributors align on a consensus change to the algorithm. A specification that is accessible to anyone can circumvent the issue that some implementers may be unable to read open source reference implementations due to the constraints of some open source licenses. Beyond helping develop specific algorithm proposals, guidelines can also serve as a reminder to potential inventors and developers of the multiple facets of the congestion control problem. The evaluation guidelines in this document are intended to be consistent with the congestion control principles from of preventing congestion collapse, considering fairness, and optimizing a flow's own performance in terms of throughput, delay, and loss. also discusses the goal of avoiding a congestion control "arms race" among competing transport protocols. This document does not give hard-and-fast requirements for an appropriate congestion control algorithm. Rather, the document provides a set of criteria that should be considered and weighed by the developers of alternative algorithms and by the IETF in the context of each proposal. The high-order criterion for advancing any proposal within the IETF is a serious scientific study of the pros and cons that occurs when the proposal is considered for publication by the IETF, or before it is deployed at large scale. After initial studies, authors are encouraged to write a specification of their proposal for publication in the RFC series. This allows others to understand and investigate the wealth of proposals in this space. This document is intended to reduce the barriers to entry for new congestion control work to the IETF. As such, proponents of new congestion control algorithms ought not to interpret these criteria as a checklist of requirements before approaching the IETF. Instead, proponents are encouraged to think about these issues beforehand, and have the willingness to do the work implied by the remainder of this document.

Guidelines for Authors about Document Status This document applies to proposed congestion control algorithms proposals that seek Experimental or Standards Track status. Evaluation of both cases involves the same questions, but with different expectations for both the answers and the degree of certainty of those answers. Congestion control algorithms without empirical evidence of Internet-scale deployment SHOULD seek Experimental status. Congestion control algorithms with a record of measured Internet-scale deployment MAY directly seek the Standards Track if there is solid data that reflects that it is safe, and the design is stable, guided by the considerations in . However, the existence of this data does not waive the other considerations in this document. Experimental specifications SHOULD NOT be deployed as a default. They SHOULD only be deployed in situations where they are being actively measured, and where it is possible to deactivate them if there are signs of pathological behavior. Each published congestion control algorithm is REQUIRED to include a statement in the abstract indicating whether or not there is IETF consensus that the proposed congestion control algorithm is considered safe for use on the Internet. Each published algorithm is also required to include a statement in the abstract describing environments where the protocol is not recommended for deployment. There can be environments where the congestion control algorithm is deemed safe for use, but it is still is not recommended for use because it does not perform well for the user. As examples of such statements, specifies HighSpeed TCP and includes a statement in the abstract stating that the proposed congestion control algorithm is Experimental, but may be deployed in the Internet. In contrast, the Quick-Start document includes a paragraph in the abstract stating that the mechanism is only being proposed for use in controlled environments. The abstract specifies environments where the Quick- Start request could give false positives (and therefore would be unsafe for incremental deployment where some routers forward, but do not process the option). The abstract also specifies environments where packets containing the Quick-Start request could be dropped in the network; in such an environment, Quick-Start would not be unsafe to deploy, but deployment is not recommended because it could lead to unnecessary delays for the connections attempting to use Quick-Start. The Quick-Start method is discussed as an example in . Alhough out of the scope of this document, proponents of a new algorithm could alternatively seek publication as an Informational or Experimental RFC via the Internet Research Task Force (IRTF). In general, these algorithms are expected to be less mature than ones that follow the procedures in this document. Authors documenting deployed congestion control algorithms that cannot be changed by IETF or IRTF review are invited to publish as an Informational RFC via the Independent Stream Editor (ISE).

Specifying Algorithms for Use in Controlled Environments Algorithms can be designed for general Internet deployment or for use in controlled environments . An operator can ensure that flows within a controlled environment are isolated from other Internet flows, or they might allow these flows to share resources with other Internet flows. Algorithms that rely on specific functions or configurations in a network need to provide a reference or specification for these functions (an RFC or another stable specification). The IETF will need to assess whether the relevant working group is able to review the proposed new algorithm and whether there is sufficient experience to understand any dependent functions. A data center is an example of a controlled environment, which often deploys fabrics with rich signalling from switches to endpoints. In evaluating a new proposal for use in a controlled environment, the IETF needs to understand the usage, e.g., how the usage is scoped to the controlled environment, whether the algorithm will share resources with Internet traffic, and consider what could happen if used in a protocol that is bridged across an Internet path. Algorithms that are designed to be confined to a controlled environment and are not intended for use in the general Internet, might instead seek real-world data for those environments. In such cases, the evaluation criteria in the remainder of this document might not apply.

Evaluation Criteria As noted above, authors are expected to do a well-rounded evaluation of the pros and cons of congestion control algorithms that are brought to the IETF. The following guidelines are designed to help authors and the IETF community. Concerns that fall outside the scope of these guidelines are certainly possible; these guidelines should not be considered an all-encompassing check-list. When considering a proposed congestion control algorithm, the community MUST consider the following criteria. These criteria will be evaluated in various domains (see and ).

Single Algorithm Behavior The criteria in this section evaluate the congestion control algorithm when one or more flows using that algorithm share a bottleneck link (i.e., with no flows using a differing congestion control algorithm).

Protection Against Congestion Collapse A congestion control algorithm should either stop sending when the packet drop rate exceeds some threshold , or should include some notion of "full backoff". For "full backoff", at some point the algorithm would reduce the sending rate to one packet per round-trip time and then exponentially backoff the time between single packet transmissions if the congestion persists. Exactly when either "full backoff" or a pause in sending comes into play will be algorithm-specific. However, as discussed in and , this requirement is crucial to protect the network in times of extreme (persistent) congestion. If full backoff is used, this test does not require that the mechanism must be identical to that of TCP (, ). For example, this does not preclude full backoff mechanisms that would give flows with different round- trip times comparable capacity during backoff.

Protection Against Bufferbloat A congestion control algorithm should try to avoid maintaining excessive queues in the network. Exactly how the algorithm achieves this is algorithm-specific, but see and for requirements. Bufferbloat refers to the building of excessive queues in the network. Many network routers are configured with very large buffers. The standards-track Reno and Cubic congestion control algorithms send at progressively higher rates until a First-In First-Out (FIFO) buffer completely fills, and packet losses then occur. Every connection passing through that bottleneck experiences increased latency due to the high buffer occupancy. This adds unwanted latency that negatively impacts highly interactive applications such as videoconferencing or games, but it also affects routine web browsing and video playing. This problem has been widely discussed since 2011 , but was not discussed in the Congestion Control Principles published in September 2002 . The Reno and Cubic congestion control algorithms do not address it, but a new congestion control algorithm has the opportunity to improve the state of the art.

Protection Against High Packet Loss A congestion control algorithm should try to avoid causing excessively high rates of packet loss. To accomplish this, it should avoid excessive increases in sending rate, and reduce its sending rate if experiencing high packet loss. The first version of the BBR algorithm failed this requirement. Experimental evaluation showed that it caused a sustained rate of packet loss when multiple BBRv1 flows shared a bottleneck and the buffer size was less than roughly one and a half BDP. This was unsatisfactory, and indeed further versions of BBR fixed it. This requirement does not imply that the algorithm should react to packet losses in exactly the same way as current standards-track congestion control algorithms (e.g., ).

Fairness within the Proposed Congestion Control Algorithm When multiple competing flows all use the same proposed congestion control algorithm, the proposal should explore how the capacity is shared among the competing flows. Capacity fairness can be important when a small number of similar flows compete to fill a bottleneck. However, it can also not be useful, for example, when comparing flows that seek to send at different rates, or if some of the flows do not last sufficiently long to approach asymptotic behavior.

Short Flows A great deal of congestion control analysis concerns the steady-state behavior of long flows. However, many Internet flows are relatively short-lived. If flows never experience a packet loss, many short-lived flows today remain in the "slow start" mode of operation that commonly features exponential congestion window growth. A proposed congestion control algorithm MUST consider how new and short-lived flows affect long-lived flows, and vice versa.

Mixed Algorithm Behavior Mixed algorithm behavior criteria evaluate the interaction of the proposed congestion control algorithm with commonly deployed congestion control algorithms. In contexts where differing congestion control algorithms are used, it is important to understand whether the proposed congestion control algorithm could result in more harm than previous standards-track algorithms (e.g., , , ) to flows sharing a common bottleneck. The measure of harm is not restricted to unequal capacity, but ought also to consider metrics such as the introduced latency, or an increase in packet loss. An evaluation MUST assess the potential to cause starvation, including assurance that a loss of all feedback (e.g., detected by expiry of a retransmission time out) results in backoff.

Existing General-Purpose Transports A proposed congestion control algorithm SHOULD be evaluated when competing using standard IETF congestion controls, e.g. , , . A proposed congestion control algorithm that has a significantly negative impact on flows using standard congestion control might be suspect, and this aspect should be part of the community's decision making with regards to the suitability of the proposed congestion control algorithm. The community should also consider other non-standard congestion control algorithms that are known to be widely deployed. Note that this guideline is not a requirement for strict Reno- or Cubic- friendliness as a prerequisite for a proposed congestion control mechanism to advance to Experimental or Standards Track status. As an example, HighSpeed TCP is a congestion control mechanism specified as Experimental, that is not TCP- friendly in all environments. When a new congestion control algorithm is deployed, the existing major algorithm deployments need to be considered to avoid severe performance degradation. Note that this guideline does not constrain the interaction with non-best-effort flows. As an example from an Experimental RFC, fairness with standard TCP is discussed in Sections 4 and 6 of (HighSpeed TCP) and using spare capacity is discussed in Sections 6, 11.1, and 12 of .

Real-Time Protocols General-purpose protocols need to coexist in the Internet with real-time congestion control algorithms, which, in general, have finite throughput requirements (i.e., do not seek to utilize all available capacity) and more strict latency bounds. provides suggestions for real-time congestion control design and suggests test cases. describes some considerations for the RTP Control Protocol (RTCP). In particular, feedback for real-time flows can be less frequent than the acknowledgements provided by reliable transports. This document does not change the informational status of those RFCs. A proposed congestion control algorithm SHOULD consider coexistence with widely deployed real-time congestion control algorithms. Regrettably, at the time of writing (2024), many algorithms with detailed public specifications are not widely deployed, while many widely deployed real-time congestion control algorithms have incomplete public specifications. It is hoped that this situation will change. To the extent that behavior of widely deployed algorithms is understood, proponents of a proposed congestion control algorithm can analyze and simulate its interaction with those algorithms. To the extent they are not, experiments can be conducted where possible. Real-time flows can be directed into distinct queues via Differentiated Services Code Points (DSCP) or other mechanisms, which can substantially reduce the interplay with other traffic. However, a proposal targeting general Internet use can not assume this is always the case. describes the impact of network transport circuit breaker algorithms. also defines a minimal set of RTP circuit breakers that operate across a path. This identifies conditions under which a sender needs to stop transmitting media data to protect the network from excessive congestion. It is expected that, in the absence of long-lived excessive congestion, RTP applications running on best-effort IP networks will be able to operate without triggering these circuit breakers.

Short and Long Flows The effect on short-lived and long-lived flows using other common congestion control algorithms MUST be evaluated, as in .

Other Criteria

Differences with Congestion Control Principles A proposed congestion control algorithm SHOULD include a clear explanation of any deviations from and .

Incremental Deployment The proposed congestion control algorithm ought to discuss whether it allows for incremental deployment in the targeted environment. For a mechanism targeted for deployment in the current Internet, it would be helpful for the proposal to discuss what is known (if anything) about the correct operation of the mechanisms with some of the equipment in the current Internet, e.g., routers, transparent proxies, WAN optimizers, intrusion detection systems, home routers, and the like. Similarly, if the proposed congestion control algorithm is intended only for specific environments (and not the global Internet), the proposal should consider how this intention is to be realised. The community will have to address the question of whether the scope can be enforced by stating the restrictions, or whether additional protocol mechanisms are required to enforce this scoping. The answer will necessarily depend on the proposed change. As an example from an Experimental RFC, deployment issues are discussed in Sections 10.3 and 10.4 of (Quick-Start).

General Use The criteria in will be evaluated in the following scenarios. Unless a proposed congestion control specification explicitly forbids use on the public Internet, the community MUST find that it meets the criteria in these scenarios for the proposed congestion control algorithm to progress. The evaluation in each scenario SHOULD occur over a representative range of bandwidths, delays, and queue depths. Of course, the set of parameters representative of the public Internet will change over time. These criteria are intended to capture a statistically dominant set of Internet conditions. In the case that a proposed congestion control algorithm has been tested at Internet scale, the results from that deployment are often useful for answering these questions.

Paths with Tail-drop Queues The performance of a congestion control algorithm is affected by the queue discipline applied at the bottleneck link. The drop-tail queue discipline (using a FIFO buffer) MUST be evaluated. See for evaluation of other queue disciplines.

Tunnel Behavior When a proposed congestion control algorithm relies on explicit signals from the path, the proposal MUST consider the effect of traffic passing through a tunnel, where routers may not be aware of the flow.

Wired Paths Wired networks are usually characterized by extremely low rates of packet loss except for those due to queue drops. They tend to have stable aggregate bandwidth, usually higher than other types of links, and low non-queueing delay. Because the properties are relatively simple, wired links are typically used as a "baseline" case even if they are not always the bottleneck link in the modern Internet.

Wireless Paths While the early Internet was dominated by wired links, the properties of wireless links have become extremely important to Internet performance. In particular, a proposed congestion control algorithm should be evaluated in situations where some packet losses are due to radio effects, rather than router queue drops; the link capacity varies over time due to changing link conditions; and media access delays and link-layer retransmission lead to increased jitter in round-trip times. See and Section 16 of for further discussion of wireless properties.

Special Cases The criteria in will be evaluated in the following scenarios, unless the proposed congestion control algorithm specifically excludes its use in a scenario. For these specific use-cases, the community MAY allow a proposal to progress even if the criteria indicate an unsatisfactory result for these scenarios. In general, measurements from Internet-scale deployments might not expose the properties of operation in each of these scenarios, because they are not as ubiquitous as the General Use scenarios.

Active Queue Management (AQM) The proposed congestion control algorithm SHOULD be evaluated under a variety of bottleneck queue disciplines. The effect of an AQM discipline can be hard to detect by Internet evaluation. At a minimum, a proposal should reason about an algorithm's response to various AQM disciplines. Simulation or empirical results are, of course, valuable. Among the AQM techniques that might have an impact on a proposed congestion control algorithm are FQ-CoDel ; Proportional Integral Controller Enhanced (PIE) ; and Low Latency, Low Loss, and Scalable Throughput (L4S) . A proposed congestion control algorithm that sets one of the two Explicit Congestion Transport (ECT) codepoints in the IP header can gain the benefits of receiving Explicit Congestion Notification (ECN) Congestion Experienced (CE) signals from an on-path AQM . Use of ECN , requires the congestion control algorithm to react when it receives a packet with an ECN-CE marking. This reaction needs to be evaluated to confirm that the algorithm conforms with the requirements of the ECT codepoint that was used. Note that evaluation of AQM techniques -- as opposed to their impact on a specific proposed congestion control algorithm -- is out of scope of this document. describes design considerations for AQMs.

Operation with the Envelope set by Network Circuit Breakers Some equipment in the network uses an automatic mechanism to continuously monitor the use of resources by a flow or aggregate set of flows . Such a network transport circuit breaker can automatically detect excessive congestion, and when detected, it can terminate (or significantly reduce the rate of) the flow(s). A well-designed congestion control algorithm ought to react before the flow uses excessive resources, and therefore will operate within the envelope set by network transport circuit breaker algorithms.

Paths with Varying Delay An Internet Path can include simple links, where the minimum delay is the propagation delay, and any additional delay can be attributed to link buffering. This cannot be assumed. An Internet Path can also include complex subnetworks where the minimum delay changes over various time scales, resulting in a non- stationary minimum delay. This occurs when a subnet changes the forwarding path to optimise capacity, resilience, etc. It could also arise when a subnet uses a capacity management method where the available resource is periodically distributed among the active nodes. A node might then have to buffer data until an assigned transmission opportunity or until the physical path changes (e.g., when the length of a wireless path changes, or the physical layer changes its mode of operation). Variation also arises when traffic with a higher priority DSCP pre-empts transmission of traffic with a lower class. In these cases, the delay varies as a function of external factors, and attempting to infer congestion from an increase in the delay results in reduced throughput. This variation in the delay over short timescales (jitter) might not be distinguishable from jitter that results from other effects. A proposed congestion control algorithm SHOULD be evaluated to ensure its operation is robust when there is a significant change in the minimum delay.

Internet of Things The "Internet of Things" (IoT) is a broad concept, but when evaluating a proposed congestion control algorithm, it is often associated with unique characteristics. IoT nodes might be more constrained in power, CPU, or other parameters than conventional Internet hosts. This might place limits on the complexity of any given algorithm. These power and radio constraints might make the volume of control packets in a given algorithm a key evaluation metric. Furthermore, many IoT applications do not a have a human in the loop, and therefore can have weaker latency constraints because they do not relate to a user experience. They mightt still need to share the path with other flows with different constraints. Extremely low-power links can lead to very low throughput and a low bandwidth- delay product, well below the standard operating range of most Internet flows.

Paths with High Delay A proposed congestion control algorithm ought not to presume that all general Internet paths have a low delay. Some paths include links that contibute much more delay than for a typical Internet path. Satellite links often have delays longer than typical for wired paths and high delay bandwidth products . They can also have variable delay as described in .

Misbehaving Nodes A proposed congestion control algorithm should explore how the algorithm performs with non-compliant senders, receivers, or routers. In addition, the proposal should explore how a proposed congestion control algorithm performs with outside attackers. This can be particularly important for proposed congestion control algorithms that involve explicit feedback from routers along the path. As an example from an Experimental RFC, performance with misbehaving nodes and outside attackers is discussed in Sections 9.4, 9.5, and 9.6 of . This includes discussion of misbehaving senders and receivers; collusion between misbehaving routers; misbehaving middleboxes; and the potential use of Quick- Start to attack routers or to tie up available Quick-Start bandwidth.

Extreme Packet Reordering A proposed congestion control algorithm ought not to presume that all general Internet paths reliably deliver packets in order. discusses the effect of extreme packet reordering.

Transient Events A proposed congestion control algorithm SHOULD consider how the proposed congestion control algorithm would perform in the presence of transient events such as sudden onset of congestion, a routing change, or a mobility event. Routing changes, link disconnections, intermittent link connectivity, and mobility are discussed in more detail in Section 16 of . As an example from an Experimental RFC, response to transient events is discussed in .

Sudden changes in the Path An IETF transport is not tied to a specific Internet path or type of path. The set of routers that form a path can and do change with time. This will cause the properties of the path to change with respect to time. A proposed congestion control algorithm MUST evaluate the impact of changes in the path, and be robust to changes in path characteristics on the interval of common Internet re-routing intervals.

Multipath Transport Multipath transport protocols permit more than one path to be differentiated and used by a single connection at the sender. A multipath sender can schedule which packets travel on which of its active paths. This enables a tradeoff in timeliness and reliability. There are various ways that multipath techniques can be used. One example use is to provide fail-over from one path to another when the original path is no longer viable, or provides inferior performance. Designs need to independently track the congestion state of each path, and demonstrate independent congestion control for each path being used. Authors of a proposed multipath congestion control algorithm that implements path fail-over MUST evaluate the harm resulting from a change in the path, and show that this does not result in flow starvation. Synchronisation of failover (e.g., where multiple flows change their path on similar timeframes) can also contribute to harm and/or reduce fairness. These effects also ought to be evaluated. Another example use is concurrent multipath, where the transport protocol simultaneously schedules a flow to aggregate the capacity across multiple paths. The Internet provides no guarantee that different paths (e.g., using different endpoint addresses) are disjoint. This has additional implications. A congestion control algorithm proposal MUST evaluate the potential harm to other flows when the multiple paths share a common congested bottleneck or share resources that are coupled between different paths, such as an overall capacity limit). It SHOULD consider the potential for harm to other flows. Synchronisation of congestion control mechanisms (e.g., where multiple flows change their behaviour on similar timeframes) can also contribute to harm and/or reduce fairness. These effects also ought to be evaluated. At the time of writing (2024), there are currently no Standards Track RFCs for concurrent multipath, but there is an Experimental RFC that specifies a concurrent multipath congestion control algorithm for MPTCP .

Data Centers Data centers are characterized by very low latencies (< 2 ms). Many workloads involve bursty traffic where many nodes complete a task at the same time. As a controlled environment, data centers often deploy fabrics that employ rich signalling from switches to endpoints. Furthermore, the operator can often limit the number of operating congestion control algorithms. For these reasons, data center congestion controls are often distinct from those running elsewhere on the Internet (see . A proposed congestion control need not coexist well with all other algorithms if it is intended for data centers, but the proposal SHOULD indicate which are expected to safely coexist with it.

Security Considerations This document does not represent a change to any aspect of the TCP/IP protocol suite and therefore does not directly impact Internet security. The implementation of various facets of the Internet's current congestion control algorithms do have security implications (e.g., as outlined in ). The IETF process that results in publication needs to ensure that these security implications are considered. Proposed congestion control algorithms therefore ought to be mindful of pitfalls, and should examine any potential security issues that may arise.

IANA Considerations This document has no IANA actions.

The goal of this document is to explain the need for congestion control in the Internet, and to discuss what constitutes correct congestion control. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. The User Datagram Protocol (UDP) provides a minimal message-passing transport that has no inherent congestion control mechanisms. This document provides guidelines on the use of UDP for the designers of applications, tunnels, and other protocols that use UDP. Congestion control guidelines are a primary focus, but the document also provides guidance on other topics, including message sizes, reliability, checksums, middlebox traversal, the use of Explicit Congestion Notification (ECN), Differentiated Services Code Points (DSCPs), and ports. Because congestion control is critical to the stable operation of the Internet, applications and other protocols that choose to use UDP as an Internet transport must employ mechanisms to prevent congestion collapse and to establish some degree of fairness with concurrent traffic. They may also need to implement additional mechanisms, depending on how they use UDP. Some guidance is also applicable to the design of other protocols (e.g., protocols layered directly on IP or via IP-based tunnels), especially when these protocols do not themselves provide congestion control. This document obsoletes RFC 5405 and adds guidelines for multicast UDP usage. CUBIC is a standard TCP congestion control algorithm that uses a cubic function instead of a linear congestion window increase function to improve scalability and stability over fast and long-distance networks. CUBIC has been adopted as the default TCP congestion control algorithm by the Linux, Windows, and Apple stacks. This document updates the specification of CUBIC to include algorithmic improvements based on these implementations and recent academic work. Based on the extensive deployment experience with CUBIC, this document also moves the specification to the Standards Track and obsoletes RFC 8312. This document also updates RFC 5681, to allow for CUBIC's occasionally more aggressive sending behavior. Many protocols must detect packet loss for various reasons (e.g., to ensure reliability using retransmissions or to understand the level of congestion along a network path). While many mechanisms have been designed to detect loss, ultimately, protocols can only count on the passage of time without delivery confirmation to declare a packet "lost". Each implementation of a time-based loss detection mechanism represents a balance between correctness and timeliness; therefore, no implementation suits all situations. This document provides high-level requirements for time-based loss detectors appropriate for general use in unicast communication across the Internet. Within the requirements, implementations have latitude to define particulars that best address each situation. This document defines TCP's four intertwined congestion control algorithms: slow start, congestion avoidance, fast retransmit, and fast recovery. In addition, the document specifies how TCP should begin transmission after a relatively long idle period, as well as discussing various acknowledgment generation methods. This document obsoletes RFC 2581. [STANDARDS-TRACK] This document describes loss detection and congestion control mechanisms for QUIC. The Real-time Transport Protocol (RTP) is widely used in telephony, video conferencing, and telepresence applications. Such applications are often run on best-effort UDP/IP networks. If congestion control is not implemented in these applications, then network congestion can lead to uncontrolled packet loss and a resulting deterioration of the user's multimedia experience. The congestion control algorithm acts as a safety measure by stopping RTP flows from using excessive resources and protecting the network from overload. At the time of this writing, however, while there are several proprietary solutions, there is no standard algorithm for congestion control of interactive RTP flows. This document does not propose a congestion control algorithm. It instead defines a minimal set of RTP circuit breakers: conditions under which an RTP sender needs to stop transmitting media data to protect the network from excessive congestion. It is expected that, in the absence of long-lived excessive congestion, RTP applications running on best-effort IP networks will be able to operate without triggering these circuit breakers. To avoid triggering the RTP circuit breaker, any Standards Track congestion control algorithms defined for RTP will need to operate within the envelope set by these RTP circuit breaker algorithms. This document provides recommendations of best current practice for dropping or marking packets using any active queue management (AQM) algorithm, including Random Early Detection (RED), BLUE, Pre- Congestion Notification (PCN), and newer schemes such as CoDel (Controlled Delay) and PIE (Proportional Integral controller Enhanced). We give three strong recommendations: (1) packet size should be taken into account when transports detect and respond to congestion indications, (2) packet size should not be taken into account when network equipment creates congestion signals (marking, dropping), and therefore (3) in the specific case of RED, the byte- mode packet drop variant that drops fewer small packets should not be used. This memo updates RFC 2309 to deprecate deliberate preferential treatment of small packets in AQM algorithms. This document explains what is meant by the term "network transport Circuit Breaker". It describes the need for Circuit Breakers (CBs) for network tunnels and applications when using non-congestion- controlled traffic and explains where CBs are, and are not, needed. It also defines requirements for building a CB and the expected outcomes of using a CB within the Internet. Google Google Google Google Google This document specifies the BBR congestion control algorithm. BBR ("Bottleneck Bandwidth and Round-trip propagation time") uses recent measurements of a transport connection's delivery rate, round-trip time, and packet loss rate to build an explicit model of the network path. BBR then uses this model to control both how fast it sends data and the maximum volume of data it allows in flight in the network at any time. Relative to loss-based congestion control algorithms such as Reno [RFC5681] or CUBIC [RFC8312], BBR offers substantially higher throughput for bottlenecks with shallow buffers or random losses, and substantially lower queueing delays for bottlenecks with deep buffers (avoiding "bufferbloat"). BBR can be implemented in any transport protocol that supports packet-delivery acknowledgment. Thus far, open source implementations are available for TCP [RFC793] and QUIC [RFC9000]. This document specifies version 2 of the BBR algorithm, also sometimes referred to as BBRv2 or bbr2. Google, Inc Google, Inc Google, Inc Google, Inc This document specifies the BBR congestion control algorithm. BBR uses recent measurements of a transport connection's delivery rate and round-trip time to build an explicit model that includes both the maximum recent bandwidth available to that connection, and its minimum recent round-trip delay. BBR then uses this model to control both how fast it sends data and the maximum amount of data it allows in flight in the network at any time. Relative to loss-based congestion control algorithms such as Reno [RFC5681] or CUBIC [draft-ietf-tcpm-cubic], BBR offers substantially higher throughput for bottlenecks with shallow buffers or random losses, and substantially lower queueing delays for bottlenecks with deep buffers (avoiding "bufferbloat"). This algorithm can be implemented in any transport protocol that supports packet-delivery acknowledgment (thus far, open source implementations are available for TCP [RFC793] and QUIC [draft-ietf-quic-transport-00]). The IETF's standard congestion control schemes have been widely shown to be inadequate for various environments (e.g., high-speed networks). Recent research has yielded many alternate congestion control schemes that significantly differ from the IETF's congestion control principles. Using these new congestion control schemes in the global Internet has possible ramifications to both the traffic using the new congestion control and to traffic using the currently standardized congestion control. Therefore, the IETF must proceed with caution when dealing with alternate congestion control proposals. The goal of this document is to provide guidance for considering alternate congestion control algorithms within the IETF. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. This document specifies the Transmission Control Protocol (TCP). TCP is an important transport-layer protocol in the Internet protocol stack, and it has continuously evolved over decades of use and growth of the Internet. Over this time, a number of changes have been made to TCP as it was specified in RFC 793, though these have only been documented in a piecemeal fashion. This document collects and brings those changes together with the protocol specification from RFC 793. This document obsoletes RFC 793, as well as RFCs 879, 2873, 6093, 6429, 6528, and 6691 that updated parts of RFC 793. It updates RFCs 1011 and 1122, and it should be considered as a replacement for the portions of those documents dealing with TCP requirements. It also updates RFC 5961 by adding a small clarification in reset handling while in the SYN-RECEIVED state. The TCP header control bits from RFC 793 have also been updated based on RFC 3168. The Datagram Congestion Control Protocol (DCCP) is a transport protocol that provides bidirectional unicast connections of congestion-controlled unreliable datagrams. DCCP is suitable for applications that transfer fairly large amounts of data and that can benefit from control over the tradeoff between timeliness and reliability. [STANDARDS-TRACK] This document describes the Stream Control Transmission Protocol (SCTP) and obsoletes RFC 4960. It incorporates the specification of the chunk flags registry from RFC 6096 and the specification of the I bit of DATA chunks from RFC 7053. Therefore, RFCs 6096 and 7053 are also obsoleted by this document. In addition, RFCs 4460 and 8540, which describe errata for SCTP, are obsoleted by this document. SCTP was originally designed to transport Public Switched Telephone Network (PSTN) signaling messages over IP networks. It is also suited to be used for other applications, for example, WebRTC. SCTP is a reliable transport protocol operating on top of a connectionless packet network, such as IP. It offers the following services to its users: The design of SCTP includes appropriate congestion avoidance behavior and resistance to flooding and masquerade attacks. This document defines the core of the QUIC transport protocol. QUIC provides applications with flow-controlled streams for structured communication, low-latency connection establishment, and network path migration. QUIC includes security measures that ensure confidentiality, integrity, and availability in a range of deployment circumstances. Accompanying documents describe the integration of TLS for key negotiation, loss detection, and an exemplary congestion control algorithm. Congestion control is needed for all data transported across the Internet, in order to promote fair usage and prevent congestion collapse. The requirements for interactive, point-to-point real-time multimedia, which needs low-delay, semi-reliable data delivery, are different from the requirements for bulk transfer like FTP or bursty transfers like web pages. Due to an increasing amount of RTP-based real-time media traffic on the Internet (e.g., with the introduction of the Web Real-Time Communication (WebRTC)), it is especially important to ensure that this kind of traffic is congestion controlled. This document describes a set of requirements that can be used to evaluate other congestion control mechanisms in order to figure out their fitness for this purpose, and in particular to provide a set of possible requirements for a real-time media congestion avoidance technique. This Informational RFC describes Data Center TCP (DCTCP): a TCP congestion control scheme for data-center traffic. DCTCP extends the Explicit Congestion Notification (ECN) processing to estimate the fraction of bytes that encounter congestion rather than simply detecting that some congestion has occurred. DCTCP then scales the TCP congestion window based on this estimate. This method achieves high-burst tolerance, low latency, and high throughput with shallow- buffered switches. This memo also discusses deployment issues related to the coexistence of DCTCP and conventional TCP, discusses the lack of a negotiating mechanism between sender and receiver, and presents some possible mitigations. This memo documents DCTCP as currently implemented by several major operating systems. DCTCP, as described in this specification, is applicable to deployments in controlled environments like data centers, but it must not be deployed over the public Internet without additional measures. CUBIC is an extension to the current TCP standards. It differs from the current TCP standards only in the congestion control algorithm on the sender side. In particular, it uses a cubic function instead of a linear window increase function of the current TCP standards to improve scalability and stability under fast and long-distance networks. CUBIC and its predecessor algorithm have been adopted as defaults by Linux and have been used for many years. This document provides a specification of CUBIC to enable third-party implementations and to solicit community feedback through experimentation on the performance of CUBIC. The proposals in this document are experimental. While they may be deployed in the current Internet, they do not represent a consensus that this is the best method for high-speed congestion control. In particular, we note that alternative experimental proposals are likely to be forthcoming, and it is not well understood how the proposals in this document will interact with such alternative proposals. This document proposes HighSpeed TCP, a modification to TCP's congestion control mechanism for use with TCP connections with large congestion windows. The congestion control mechanisms of the current Standard TCP constrains the congestion windows that can be achieved by TCP in realistic environments. For example, for a Standard TCP connection with 1500-byte packets and a 100 ms round-trip time, achieving a steady-state throughput of 10 Gbps would require an average congestion window of 83,333 segments, and a packet drop rate of at most one congestion event every 5,000,000,000 packets (or equivalently, at most one congestion event every 1 2/3 hours). This is widely acknowledged as an unrealistic constraint. To address his limitation of TCP, this document proposes HighSpeed TCP, and solicits experimentation and feedback from the wider community. This document specifies an optional Quick-Start mechanism for transport protocols, in cooperation with routers, to determine an allowed sending rate at the start and, at times, in the middle of a data transfer (e.g., after an idle period). While Quick-Start is designed to be used by a range of transport protocols, in this document we only specify its use with TCP. Quick-Start is designed to allow connections to use higher sending rates when there is significant unused bandwidth along the path, and the sender and all of the routers along the path approve the Quick-Start Request. This document describes many paths where Quick-Start Requests would not be approved. These paths include all paths containing routers, IP tunnels, MPLS paths, and the like that do not support Quick- Start. These paths also include paths with routers or middleboxes that drop packets containing IP options. Quick-Start Requests could be difficult to approve over paths that include multi-access layer- two networks. This document also describes environments where the Quick-Start process could fail with false positives, with the sender incorrectly assuming that the Quick-Start Request had been approved by all of the routers along the path. As a result of these concerns, and as a result of the difficulties and seeming absence of motivation for routers, such as core routers to deploy Quick-Start, Quick-Start is being proposed as a mechanism that could be of use in controlled environments, and not as a mechanism that would be intended or appropriate for ubiquitous deployment in the global Internet. This memo defines an Experimental Protocol for the Internet community. This document is a product of the Path Aware Networking Research Group (PANRG). At the first meeting of the PANRG, the Research Group agreed to catalog and analyze past efforts to develop and deploy Path Aware techniques, most of which were unsuccessful or at most partially successful, in order to extract insights and lessons for Path Aware networking researchers. This document contains that catalog and analysis. There is a noticeable trend towards network behaviors and semantics that are specific to a particular set of requirements applied within a limited region of the Internet. Policies, default parameters, the options supported, the style of network management, and security requirements may vary between such limited regions. This document reviews examples of such limited domains (also known as controlled environments), notes emerging solutions, and includes a related taxonomy. It then briefly discusses the standardization of protocols for limited domains. Finally, it shows the need for a precise definition of "limited domain membership" and for mechanisms to allow nodes to join a domain securely and to find other members, including boundary nodes. This document is the product of the research of the authors. It has been produced through discussions and consultation within the IETF but is not the product of IETF consensus. This document discusses IAB concerns about effective end-to-end congestion control for best-effort voice traffic in the Internet. These concerns have to do with fairness, user quality, and with the dangers of congestion collapse. The concerns are particularly relevant in light of the absence of a widespread Quality of Service (QoS) deployment in the Internet, and the likelihood that this situation will not change much in the near term. This document is not making any recommendations about deployment paths for Voice over Internet Protocol (VoIP) in terms of QoS support, and is not claiming that best-effort service can be relied upon to give acceptable performance for VoIP. We are merely observing that voice traffic is occasionally deployed as best-effort traffic over some links in the Internet, that we expect this occasional deployment to continue, and that we have concerns about the lack of effective end-to-end congestion control for this best-effort voice traffic. This memo provides information for the Internet community. This document defines the standard algorithm that Transmission Control Protocol (TCP) senders are required to use to compute and manage their retransmission timer. [STANDARDS-TRACK] The Real-Time Transport Protocol (RTP) is used to transmit media in telephony and video conferencing applications. This document describes the guidelines to evaluate new congestion control algorithms for interactive point-to-point real-time media. The Real-time Transport Protocol (RTP) is used to transmit media in multimedia telephony applications. These applications are typically required to implement congestion control. This document describes the test cases to be used in the performance evaluation of such congestion control algorithms in a controlled environment. This memo discusses the rate at which congestion control feedback can be sent using the RTP Control Protocol (RTCP) and the suitability of RTCP for implementing congestion control for unicast multimedia applications. This document provides advice to the designers of digital communication equipment, link-layer protocols, and packet-switched local networks (collectively referred to as subnetworks), who wish to support the Internet protocols but may be unfamiliar with the Internet architecture and the implications of their design choices on the performance and efficiency of the Internet. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. This memo presents the FQ-CoDel hybrid packet scheduler and Active Queue Management (AQM) algorithm, a powerful tool for fighting bufferbloat and reducing latency. FQ-CoDel mixes packets from multiple flows and reduces the impact of head-of-line blocking from bursty traffic. It provides isolation for low-rate traffic such as DNS, web, and videoconferencing traffic. It improves utilisation across the networking fabric, especially for bidirectional traffic, by keeping queue lengths short, and it can be implemented in a memory- and CPU-efficient fashion across a wide range of hardware. Bufferbloat is a phenomenon in which excess buffers in the network cause high latency and latency variation. As more and more interactive applications (e.g., voice over IP, real-time video streaming, and financial transactions) run in the Internet, high latency and latency variation degrade application performance. There is a pressing need to design intelligent queue management schemes that can control latency and latency variation, and hence provide desirable quality of service to users. This document presents a lightweight active queue management design called "PIE" (Proportional Integral controller Enhanced) that can effectively control the average queuing latency to a target value. Simulation results, theoretical analysis, and Linux testbed results have shown that PIE can ensure low latency and achieve high link utilization under various congestion situations. The design does not require per-packet timestamps, so it incurs very little overhead and is simple enough to implement in both hardware and software. This specification defines a framework for coupling the Active Queue Management (AQM) algorithms in two queues intended for flows with different responses to congestion. This provides a way for the Internet to transition from the scaling problems of standard TCP-Reno-friendly ('Classic') congestion controls to the family of 'Scalable' congestion controls. These are designed for consistently very low queuing latency, very low congestion loss, and scaling of per-flow throughput by using Explicit Congestion Notification (ECN) in a modified way. Until the Coupled Dual Queue (DualQ), these Scalable L4S congestion controls could only be deployed where a clean-slate environment could be arranged, such as in private data centres. This specification first explains how a Coupled DualQ works. It then gives the normative requirements that are necessary for it to work well. All this is independent of which two AQMs are used, but pseudocode examples of specific AQMs are given in appendices. The goal of this document is to describe the potential benefits of applications using a transport that enables Explicit Congestion Notification (ECN). The document outlines the principal gains in terms of increased throughput, reduced delay, and other benefits when ECN is used over a network path that includes equipment that supports Congestion Experienced (CE) marking. It also discusses challenges for successful deployment of ECN. It does not propose new algorithms to use ECN nor does it describe the details of implementation of ECN in endpoint devices (Internet hosts), routers, or other network devices. This memo specifies the incorporation of ECN (Explicit Congestion Notification) to TCP and IP, including ECN's use of two bits in the IP header. [STANDARDS-TRACK] This memo presents recommendations to the Internet community concerning measures to improve and preserve Internet performance. It presents a strong recommendation for testing, standardization, and widespread deployment of active queue management (AQM) in network devices to improve the performance of today's Internet. It also urges a concerted effort of research, measurement, and ultimate deployment of AQM mechanisms to protect the Internet from flows that are not sufficiently responsive to congestion notification. Based on 15 years of experience and new research, this document replaces the recommendations of RFC 2309. The Transmission Control Protocol (TCP) provides reliable delivery of data across any network path, including network paths containing satellite channels. While TCP works over satellite channels there are several IETF standardized mechanisms that enable TCP to more effectively utilize the available capacity of the network path. This document outlines some of these TCP mitigations. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. This document specifies Non-Congestion Robustness (NCR) for TCP. In the absence of explicit congestion notification from the network, TCP uses loss as an indication of congestion. One of the ways TCP detects loss is using the arrival of three duplicate acknowledgments. However, this heuristic is not always correct, notably in the case when network paths reorder segments (for whatever reason), resulting in degraded performance. TCP-NCR is designed to mitigate this degraded performance by increasing the number of duplicate acknowledgments required to trigger loss recovery, based on the current state of the connection, in an effort to better disambiguate true segment loss from segment reordering. This document specifies the changes to TCP, as well as the costs and benefits of these modifications. This memo defines an Experimental Protocol for the Internet community. Often endpoints are connected by multiple paths, but communications are usually restricted to a single path per connection. Resource usage within the network would be more efficient were it possible for these multiple paths to be used concurrently. Multipath TCP is a proposal to achieve multipath transport in TCP. New congestion control algorithms are needed for multipath transport protocols such as Multipath TCP, as single path algorithms have a series of issues in the multipath context. One of the prominent problems is that running existing algorithms such as standard TCP independently on each path would give the multipath flow more than its fair share at a bottleneck link traversed by more than one of its subflows. Further, it is desirable that a source with multiple paths available will transfer more traffic using the least congested of the paths, achieving a property called "resource pooling" where a bundle of links effectively behaves like one shared link with bigger capacity. This would increase the overall efficiency of the network and also its robustness to failure. This document presents a congestion control algorithm that couples the congestion control algorithms running on different subflows by linking their increase functions, and dynamically controls the overall aggressiveness of the multipath flow. The result is a practical algorithm that is fair to TCP at bottlenecks while moving traffic away from congested links. This document defines an Experimental Protocol for the Internet community. TCP/IP communication is currently restricted to a single path per connection, yet multiple paths often exist between peers. The simultaneous use of these multiple paths for a TCP/IP session would improve resource usage within the network and thus improve user experience through higher throughput and improved resilience to network failure. Multipath TCP provides the ability to simultaneously use multiple paths between peers. This document presents a set of extensions to traditional TCP to support multipath operation. The protocol offers the same type of service to applications as TCP (i.e., a reliable bytestream), and it provides the components necessary to establish and use multiple TCP flows across potentially disjoint paths. This document specifies v1 of Multipath TCP, obsoleting v0 as specified in RFC 6824, through clarifications and modifications primarily driven by deployment experience. This document discusses the metrics to be considered in an evaluation of new or modified congestion control mechanisms for the Internet. These include metrics for the evaluation of new transport protocols, of proposed modifications to TCP, of application-level congestion control, and of Active Queue Management (AQM) mechanisms in the router. This document is the first in a series of documents aimed at improving the models that we use in the evaluation of transport protocols. This document is a product of the Transport Modeling Research Group (TMRG), and has received detailed feedback from many members of the Research Group (RG). As the document tries to make clear, there is not necessarily a consensus within the research community (or the IETF community, the vendor community, the operations community, or any other community) about the metrics that congestion control mechanisms should be designed to optimize, in terms of trade-offs between throughput and delay, fairness between competing flows, and the like. However, we believe that there is a clear consensus that congestion control mechanisms should be evaluated in terms of trade-offs between a range of metrics, rather than in terms of optimizing for a single metric. This memo provides information for the Internet community.

Acknowledgments Sally Floyd and Mark Allman were the authors of this document's predecessor, , which served the community well for over a decade. Thanks to Richard Scheffenegger for helping to get this revision process started. The editors would like to thank Mohamed Boucadair, Neal Cardwell, Reese Enghardt, Jonathan Lennox, Matt Mathis, Dave Taht, Michael Welzl, Magnus Westerlund, and Greg White for suggesting improvements to this document. Discussions with Lars Eggert and Aaron Falk seeded the original RFC5033. Bob Briscoe, Gorry Fairhurst, Doug Leith, Jitendra Padhye, Colin Perkins, Pekka Savola, members of TSVWG, and participants at the TCP Workshop at Microsoft Research all provided feedback and contributions to that document. It also drew from .

Evolution of RFC5033bis

Since draft-ietf-ccwg-rfc5033bis-03 Harmonised the "proposed congestion control algorithm" addressed issues.

Since draft-ietf-ccwg-rfc5033bis-02 Added discussion of real-time protocols Added discussion of short flows Listed properties of wired networks Added IoT section Added discussion of AQM response Rewrote the "Document Status" section Adding improved first sentence of abstract and intro. Added section on Multicast, noting this is out of scope Editorial changes

Since draft-ietf-ccwg-rfc5033bis-01 Added discussion of multipath transports Totally reorganized central sections of the draft

Since draft-ietf-ccwg-rfc5033bis-00 Added QUIC, other congestion control standards Added wireless environments Aligned motivation for this work with the CCWG charter Refined discussion of QuickStart

Since draft-scheffenegger-congress-rfc5033bis-00 Renamed file to reflect WG adpotion Updated authorship and acknowledgements. Include updated text suggested by Dave Taht Added criterion for bufferbloat Mentioned Cubic and BBR as motivation Include section to track updates between revisions Update references

Since RFC5033 converted to Markdown and xml2rfc v3 various formatting changes

Contributors Private Octopus, Inc.

huitema@huitema.net