Congestion Control Requirements for Interactive Real-Time Media
draft-ietf-rmcat-cc-requirements-08

Abstract

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 real-time media congestion avoidance technique.

Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. The terms are presented in many cases using lowercase for readability.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

This Internet-Draft will expire on May 15, 2015.

Copyright Notice

Copyright (c) 2014 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

• 1. Introduction
• 2. Requirements
• 3. Deficiencies of existing mechanisms
• 4. IANA Considerations
• 5. Security Considerations
• 6. Acknowledgements
• 7. References
• 7.1. Normative References
• 7.2. Informative References
• Authors' Addresses

1. Introduction

Most of today's TCP congestion control schemes were developed with a focus on an use of the Internet for reliable bulk transfer of non-time-critical data, such as transfer of large files. They have also been used successfully to govern the reliable transfer of smaller chunks of data in as short a time as possible, such as when fetching Web pages.

These algorithms have also been used for transfer of media streams that are viewed in a non-interactive manner, such as "streaming" video, where having the data ready when the viewer wants it is important, but the exact timing of the delivery is not.

When doing real-time interactive media, the requirements are different; one needs to provide the data continuously, within a very limited time window (no more than 100s of milliseconds end-to-end delay), the sources of data may be able to adapt the amount of data that needs sending within fairly wide margins but can be rate limited by the application- even not always have data to send, and may tolerate some amount of packet loss, but since the data is generated in real-time, sending "future" data is impossible, and since it's consumed in real-time, data delivered late is commonly useless.

While the requirements for real-time interactive media differ from the requirements for the other flow types, these other flow types will be present in the network. The congestion control algorithm for real-time interactive media must work properly when these other flow types are present as cross traffic on the network.

One particular protocol portfolio being developed for this use case is WebRTC [I-D.ietf-rtcweb-overview], where one envisions sending multiple flows using the Real-time Transport Protocol (RTP) [RFC3550] between two peers, in conjunction with data flows, all at the same time, without having special arrangements with the intervening service providers. As RTP does not provide any congestion control mechanism; a set of circuit breakers, such as [I-D.ietf-avtcore-rtp-circuit-breakers], are required to protect the network from excessive congestion caused by the non-congestion controlled flows. When the real-time interactive media is congestion controlled, it is recommended that the congestion control mechanism operates within the constraints defined by these circuit breakers when circuit breaker is present and that it should not cause congestion collapse when circuit breaker is not implemented.

Given that this use case is the focus of this document, use cases involving non-interactive media such as video streaming, and use cases using multicast/broadcast-type technologies, are out of scope.

The terminology defined in [I-D.ietf-rtcweb-overview] is used in this memo.

2. Requirements

1. The congestion control algorithm must attempt to provide as-low-as-possible-delay transit for interactive real-time traffic while still providing a useful amount of bandwidth. There may be lower limits on the amount of bandwidth that is useful, but this is largely application-specific and the application may be able to modify or remove flows in order allow some useful flows to get enough bandwidth. (Example: not enough bandwidth for low-latency video+audio, but enough for audio-only.)
   1. Jitter (variation in the bitrate over short timescales) also is relevant, though moderate amounts of jitter will be absorbed by jitter buffers. Transit delay should be considered to track the short-term maximums of delay including jitter.
   2. It should provide this as-low-as-possible-delay transit and minimize self-induced latency even when faced with intermediate bottlenecks and competing flows. Competing flows may limit what's possible to achieve.
   3. It should handle the effect of routing changes which may alter or remove bottlenecks or change the bandwidth available especially if there is a reduction in available bandwidth or increase in observed delay. It is expected that the mechanism reacts quickly to the routing change events to avoid delay buildup. In the context of this memo, a ‘quick’ reaction is on the order of a few RTTs, subject to the constraints of the media codec, but is likely within a second. Reaction on the next RTT is explicitly not required, since many codecs cannot adapt their sending rate that quickly, but equally response cannot be arbitrarily delayed.
   4. It should react quickly to handle both local and remote interface changes (WLAN to 3G data, etc) which may radically change the bandwidth available or bottlenecks, especially if there is a reduction in available bandwidth or increase in bottleneck delay. It is assumed that an interface change can generate a notification to the algorithm.
   5. The algorithm must consider the case where offered loads are less than the available bandwidth at any given moment, and may vary dramatically over time, including dropping to no load and then resuming a high load, such as in a mute/unmute operation. Note that the reaction time between a change in the bandwidth available from the algorithm and a change in the offered load is variable, and may be different when increasing versus decreasing.
   6. The algorithm requires to avoid building up queues when competing with short-term bursts of traffic (for example, traffic generated by web-browsing) which can quickly saturate a local-bottleneck router or link, but also clear quickly. The algorithm should also react quickly to regain its previous share of the bandwidth when the local-bottleneck or link is cleared.
   7. Similarly periodic bursty flows such as MPEG DASH [MPEG_DASH] or proprietary media streaming algorithms may compete in bursts with the algorithm, and may not be adaptive within a burst. They are often layered on top of TCP but use TCP in a bursty manner that can interact poorly with competing flows during the bursts. The algorithm must not increase the already existing delay buildup during those bursts. Note that this competing traffic may be on a shared access link, or the traffic burst may cause a shift in the location of the bottleneck for the duration of the burst.
2. The algorithm must be fair to other flows, both real-time flows (such as other instances of itself), and TCP flows, both long-lived and bursts such as the traffic generated by a typical web browsing session. Note that 'fair' is a rather hard-to-define term. It should be fair with itself, giving fair share of the bandwidth to multiple flows with similar RTTs, and if possible to multiple flows with different RTTs.
   1. Existing flows at a bottleneck must also be fair to new flows to that bottleneck, and must allow new flows to ramp up to a useful share of the bottleneck bandwidth as quickly as possible. A useful share will depend on the media types involved, total bandwidth available and the user experience requirements of a particular service. Note that relative RTTs may affect the rate new flows can ramp up to a reasonable share.
3. The algorithm should not starve competing TCP flows, and should as best as possible avoid starvation by TCP flows.
   1. The congestion control should prioritise achieving a useful share of the bandwidth depending on the media types and total available bandwidth over achieving as low as possible transit delay, when these two requirements are in conflict.
4. The algorithm should as quickly as possible adapt to initial network conditions at the start of a flow. This should occur both if the initial bandwidth is above or below the bottleneck bandwidth.
   1. The algorithm should allow different modes of adaptation for example,the startup adaptation may be faster than adaptation later in a flow. It should allow for both slow-start operation (adapt up) and history-based startup (start at a point expected to be at or below channel bandwidth from historical information, which may need to adapt down quickly if the initial guess is wrong). Starting too low and/or adapting up too slowly can cause a critical point in a personal communication to be poor ("Hello!"). Starting over-bandwidth causes other problems for user experience, so there's a tension here. Alternative methods to help startup like probing during setup with dummy data may be useful in some applications; in some cases there will be a considerable gap in time between flow creation and the initial flow of data. Again, A flow may need to change adaptation rates due to network conditions or changes in the provided flows (such as un-muting or sending data after a gap).
5. The algorithm should be stable if the RTP streams are halted or discontinuous (for example - Voice Activity Detection).
   1. After stream resumption, the algorithm should attempt to rapidly regain its previous share of the bandwidth; the aggressiveness with which this is done will decay with the length of the pause.
6. The algorithm should where possible merge information across multiple RTP streams sent between two endpoints, when those RTP streams share a common bottleneck, whether or not those streams are multiplexed onto the same ports, in order to allow congestion control of the set of streams together instead of as multiple independent streams. This allows better overall bandwidth management, faster response to changing conditions, and fairer sharing of bandwidth with other network users.
   1. The algorithm should also share information and adaptation with other non-RTP flows between the same endpoints, such as a WebRTC DataChannel [I-D.ietf-rtcweb-data-channel], when possible.
   2. When there are multiple streams across the same 5-tuple coordinating their bandwidth use and congestion control, the algorithm should allow the application to control the relative split of available bandwidth.The most correlated bandwidth usage would be with other flows on the same 5-tuple, but there may be use in coordinating measurement and control of the local link(s). Use of information about previous flows, especially on the same 5-tuple, may be useful input to the algorithm, especially to startup performance of a new flow.
7. The algorithm should not require any special support from network elements to convey congestion related information to be functional. As much as possible, it should leverage available information about the incoming flow to provide feedback to the sender. Examples of this information are the packet arrival times, acknowledgements and feedback, packet timestamps, and packet losses, Explicit Congestion Notification (ECN) [RFC3168]; all of these can provide information about the state of the path and any bottlenecks. However, the use of available information is algorithm dependent.
   1. Extra information could be added to the packets to provide more detailed information on actual send times (as opposed to sampling times), but should not be required.
8. Since the assumption here is a set of RTP streams, the backchannel typically should be done via RTCP[RFC3550]; one alternative would be to include it instead in a reverse RTP channel using header extensions.
   1. In order to react sufficiently quickly when using RTCP for a backchannel, an RTP profile such as RTP/AVPF [RFC4585] or RTP/SAVPF [RFC5124] that allows sufficiently frequent feedback must be used. Note that in some cases, backchannel messages may be delayed until the RTCP channel can be allocated enough bandwidth, even under AVPF rules. This may also imply negotiating a higher maximum percentage for RTCP data or allowing solutions to violate or modify the rules specified for AVPF.
   2. Bandwidth for the feedback messages should be minimized (such as via RFC 5506 [RFC5506]to allow RTCP without Sender Reports/Receiver Reports)
   3. Backchannel data should be minimized to avoid taking too much reverse-channel bandwidth (since this will often be used in a bidirectional set of flows). In areas of stability, backchannel data may be sent more infrequently so long as algorithm stability and fairness are maintained. When the channel is unstable or has not yet reached equilibrium after a change, backchannel feedback may be more frequent and use more reverse-channel bandwidth. This is an area with considerable flexibility of design, and different approaches to backchannel messages and frequency are expected to be evaluated.
9. Flows managed by this algorithm and flows competing against at a bottleneck may have different DSCP[RFC5865] markings depending on the type of traffic, or may be subject to flow-based QoS. A particular bottleneck or section of the network path may or may not honor DSCP markings. The algorithm should attempt to leverage DSCP markings when they're available.
   1. In WebRTC, a division of packets into 4 classes is envisioned in order of priority: faster-than-audio, audio, video, best-effort, and bulk-transfer. Typically the flows managed by this algorithm would be audio or video in that hierarchy, and feedback flows would be faster-than-audio.
10. The algorithm should sense the unexpected lack of backchannel information as a possible indication of a channel overuse problem and react accordingly to avoid burst events causing a congestion collapse.
11. The algorithm should be stable and maintain low-delay when faced with Active Queue Management (AQM) algorithms. Also note that these algorithms may apply across multiple queues in the bottleneck, or to a single queue

3. Deficiencies of existing mechanisms

Among the existing congestion control mechanisms TCP Friendly Rate Control (TFRC) [RFC5348] is the one which claims to be suitable for real-time interactive media. TFRC is, an equation based, congestion control mechanism which provides reasonably fair share of the bandwidth when competing with TCP flows and offers much lower throughput variations than TCP. This is achieved by a slower response to the available bandwidth change than TCP. TFRC is designed to perform best with applications which has fixed packet size and does not have fixed period between sending packets.

TFRC operates on detecting loss events and reacts to loss caused by congestion by reducing its sending rate. It allows applications to increase the sending rate until loss is observed in the flows. As it is noted in IAB/IRTF report [RFC7295] large buffers are available in the network elements which introduces additional delay in the communication, it becomes important to take all possible congestion indications into considerations. Looking at the current Internet deployment, TFRC's only consideration of loss events as congestion indication can be considered as biggest lacking.

A typical real-time interactive communication includes live encoded audio and video flow(s). In such a communication scenario an audio source typically needs fixed interval between packets, needs to vary their segment size instead of their packet rate in response to congestion and sends smaller packets, a variant of TFRC , Small-Packet TFRC (TFRC-SP) [RFC4828] addresses the issues related to such kind of sources ; a video source generally varies video frame sizes, can produce large frames which need to be further fragmented to fit into path Maximum Transmission Unit (MTU) size, and have almost fixed interval between producing frames under a certain frame rate, TFRC is known to be less optimal when using with such video sources.

There are also some mismatches between TFRC's design assumptions and how the media sources in a typical real-time interactive application works. TFRC is design to maintain smooth sending rate however media sources can change rates in steps for both rate increase and rate decrease. TFRC can operate in two modes - i) Bytes per second and ii) packets per second, where typical real-time interactive media sources operates on bit per second. There are also limitations on how quickly the media sources can adapt to specific sending rates. The modern video encoders can operate on a mode where they can vary the output bitrate a lot depending on the way there are configured, the current scene it is encoding and more. Therefore, it is possible that the video source does not always output at a bitrate they are allowed to. TFRC tries to raise its sending rate when transmitting at maximum allowed rate and increases only twice the current transmission rate hence it may create issues when the video source vary their bitrates.

Moreover, there are number of studies on TFRC which shows it's limitations which includes TFRC's unfairness on low statistically multiplexed links, oscillatory behavior, performance issue in highly dynamic loss rates conditions and more [CH09].

Looking at all these deficiencies it can be concluded that the requirements of congestion control mechanism for real-time interactive media cannot be met by TFRC as defined in the standard.

4. IANA Considerations

This document makes no request of IANA.

Note to RFC Editor: this section may be removed on publication as an RFC.

5. Security Considerations

An attacker with the ability to delete, delay or insert messages in the flow can fake congestion signals, unless they are passed on a tamper-proof path. Since some possible algorithms depend on the timing of packet arrival, even a traditional protected channel does not fully mitigate such attacks.

An attack that reduces bandwidth is not necessarily significant, since an on-path attacker could break the connection by discarding all packets. Attacks that increase the perceived available bandwidth are conceivable, and need to be evaluated.

Algorithm designers should consider the possibility of malicious on-path attackers.

6. Acknowledgements

This document is the result of discussions in various fora of the WebRTC effort, in particular on the rtp-congestion@alvestrand.no mailing list. Many people contributed their thoughts to this.

7. References

7.1. Normative References

7.2. Informative References

Authors' Addresses