idnits 2.17.00 (12 Aug 2021) /tmp/idnits41904/draft-ietf-rtcweb-rtp-usage-26.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 17, 2016) is 2249 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: draft-ietf-avtcore-multi-media-rtp-session has been published as RFC 8860 == Outdated reference: draft-ietf-avtcore-rtp-circuit-breakers has been published as RFC 8083 == Outdated reference: draft-ietf-avtcore-rtp-multi-stream has been published as RFC 8108 == Outdated reference: draft-ietf-avtcore-rtp-multi-stream-optimisation has been published as RFC 8861 == Outdated reference: draft-ietf-avtcore-rtp-topologies-update has been published as RFC 7667 ** Downref: Normative reference to an Informational draft: draft-ietf-avtcore-rtp-topologies-update (ref. 'I-D.ietf-avtcore-rtp-topologies-update') == Outdated reference: draft-ietf-mmusic-mux-exclusive has been published as RFC 8858 == Outdated reference: draft-ietf-mmusic-sdp-bundle-negotiation has been published as RFC 8843 == Outdated reference: draft-ietf-rtcweb-audio has been published as RFC 7874 == Outdated reference: draft-ietf-rtcweb-fec has been published as RFC 8854 == Outdated reference: draft-ietf-rtcweb-overview has been published as RFC 8825 == Outdated reference: draft-ietf-rtcweb-security has been published as RFC 8826 == Outdated reference: draft-ietf-rtcweb-security-arch has been published as RFC 8827 == Outdated reference: draft-ietf-rtcweb-video has been published as RFC 7742 ** Obsolete normative reference: RFC 4566 (Obsoleted by RFC 8866) ** Obsolete normative reference: RFC 5285 (Obsoleted by RFC 8285) == Outdated reference: draft-ietf-avtcore-multiplex-guidelines has been published as RFC 8872 == Outdated reference: draft-ietf-avtext-rtp-grouping-taxonomy has been published as RFC 7656 == Outdated reference: draft-ietf-dart-dscp-rtp has been published as RFC 7657 == Outdated reference: draft-ietf-mmusic-msid has been published as RFC 8830 == Outdated reference: draft-ietf-payload-rtp-howto has been published as RFC 8088 == Outdated reference: draft-ietf-rmcat-cc-requirements has been published as RFC 8836 == Outdated reference: draft-ietf-rtcweb-jsep has been published as RFC 8829 == Outdated reference: draft-ietf-tsvwg-rtcweb-qos has been published as RFC 8837 -- Obsolete informational reference (is this intentional?): RFC 5245 (Obsoleted by RFC 8445, RFC 8839) Summary: 3 errors (**), 0 flaws (~~), 22 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 RTCWEB Working Group C. Perkins 3 Internet-Draft University of Glasgow 4 Intended status: Standards Track M. Westerlund 5 Expires: September 18, 2016 Ericsson 6 J. Ott 7 Aalto University 8 March 17, 2016 10 Web Real-Time Communication (WebRTC): Media Transport and Use of RTP 11 draft-ietf-rtcweb-rtp-usage-26 13 Abstract 15 The Web Real-Time Communication (WebRTC) framework provides support 16 for direct interactive rich communication using audio, video, text, 17 collaboration, games, etc. between two peers' web-browsers. This 18 memo describes the media transport aspects of the WebRTC framework. 19 It specifies how the Real-time Transport Protocol (RTP) is used in 20 the WebRTC context, and gives requirements for which RTP features, 21 profiles, and extensions need to be supported. 23 Status of This Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at http://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six months 34 and may be updated, replaced, or obsoleted by other documents at any 35 time. It is inappropriate to use Internet-Drafts as reference 36 material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on September 18, 2016. 40 Copyright Notice 42 Copyright (c) 2016 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents 47 (http://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with respect 50 to this document. Code Components extracted from this document must 51 include Simplified BSD License text as described in Section 4.e of 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 58 2. Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . 4 59 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 60 4. WebRTC Use of RTP: Core Protocols . . . . . . . . . . . . . . 5 61 4.1. RTP and RTCP . . . . . . . . . . . . . . . . . . . . . . 5 62 4.2. Choice of the RTP Profile . . . . . . . . . . . . . . . . 7 63 4.3. Choice of RTP Payload Formats . . . . . . . . . . . . . . 8 64 4.4. Use of RTP Sessions . . . . . . . . . . . . . . . . . . . 10 65 4.5. RTP and RTCP Multiplexing . . . . . . . . . . . . . . . . 10 66 4.6. Reduced Size RTCP . . . . . . . . . . . . . . . . . . . . 11 67 4.7. Symmetric RTP/RTCP . . . . . . . . . . . . . . . . . . . 11 68 4.8. Choice of RTP Synchronisation Source (SSRC) . . . . . . . 12 69 4.9. Generation of the RTCP Canonical Name (CNAME) . . . . . . 12 70 4.10. Handling of Leap Seconds . . . . . . . . . . . . . . . . 13 71 5. WebRTC Use of RTP: Extensions . . . . . . . . . . . . . . . . 13 72 5.1. Conferencing Extensions and Topologies . . . . . . . . . 14 73 5.1.1. Full Intra Request (FIR) . . . . . . . . . . . . . . 15 74 5.1.2. Picture Loss Indication (PLI) . . . . . . . . . . . . 15 75 5.1.3. Slice Loss Indication (SLI) . . . . . . . . . . . . . 16 76 5.1.4. Reference Picture Selection Indication (RPSI) . . . . 16 77 5.1.5. Temporal-Spatial Trade-off Request (TSTR) . . . . . . 16 78 5.1.6. Temporary Maximum Media Stream Bit Rate Request 79 (TMMBR) . . . . . . . . . . . . . . . . . . . . . . . 16 80 5.2. Header Extensions . . . . . . . . . . . . . . . . . . . . 17 81 5.2.1. Rapid Synchronisation . . . . . . . . . . . . . . . . 17 82 5.2.2. Client-to-Mixer Audio Level . . . . . . . . . . . . . 17 83 5.2.3. Mixer-to-Client Audio Level . . . . . . . . . . . . . 18 84 5.2.4. Media Stream Identification . . . . . . . . . . . . . 18 85 5.2.5. Coordination of Video Orientation . . . . . . . . . . 18 86 6. WebRTC Use of RTP: Improving Transport Robustness . . . . . . 19 87 6.1. Negative Acknowledgements and RTP Retransmission . . . . 19 88 6.2. Forward Error Correction (FEC) . . . . . . . . . . . . . 20 89 7. WebRTC Use of RTP: Rate Control and Media Adaptation . . . . 20 90 7.1. Boundary Conditions and Circuit Breakers . . . . . . . . 21 91 7.2. Congestion Control Interoperability and Legacy Systems . 22 92 8. WebRTC Use of RTP: Performance Monitoring . . . . . . . . . . 22 93 9. WebRTC Use of RTP: Future Extensions . . . . . . . . . . . . 23 94 10. Signalling Considerations . . . . . . . . . . . . . . . . . . 23 95 11. WebRTC API Considerations . . . . . . . . . . . . . . . . . . 25 96 12. RTP Implementation Considerations . . . . . . . . . . . . . . 27 97 12.1. Configuration and Use of RTP Sessions . . . . . . . . . 27 98 12.1.1. Use of Multiple Media Sources Within an RTP Session 27 99 12.1.2. Use of Multiple RTP Sessions . . . . . . . . . . . . 28 100 12.1.3. Differentiated Treatment of RTP Streams . . . . . . 33 101 12.2. Media Source, RTP Streams, and Participant 102 Identification . . . . . . . . . . . . . . . . . . . . . 35 103 12.2.1. Media Source Identification . . . . . . . . . . . . 35 104 12.2.2. SSRC Collision Detection . . . . . . . . . . . . . . 36 105 12.2.3. Media Synchronisation Context . . . . . . . . . . . 37 106 13. Security Considerations . . . . . . . . . . . . . . . . . . . 37 107 14. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 108 15. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 39 109 16. References . . . . . . . . . . . . . . . . . . . . . . . . . 39 110 16.1. Normative References . . . . . . . . . . . . . . . . . . 39 111 16.2. Informative References . . . . . . . . . . . . . . . . . 44 112 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 46 114 1. Introduction 116 The Real-time Transport Protocol (RTP) [RFC3550] provides a framework 117 for delivery of audio and video teleconferencing data and other real- 118 time media applications. Previous work has defined the RTP protocol, 119 along with numerous profiles, payload formats, and other extensions. 120 When combined with appropriate signalling, these form the basis for 121 many teleconferencing systems. 123 The Web Real-Time communication (WebRTC) framework provides the 124 protocol building blocks to support direct, interactive, real-time 125 communication using audio, video, collaboration, games, etc., between 126 two peers' web-browsers. This memo describes how the RTP framework 127 is to be used in the WebRTC context. It proposes a baseline set of 128 RTP features that are to be implemented by all WebRTC Endpoints, 129 along with suggested extensions for enhanced functionality. 131 This memo specifies a protocol intended for use within the WebRTC 132 framework, but is not restricted to that context. An overview of the 133 WebRTC framework is given in [I-D.ietf-rtcweb-overview]. 135 The structure of this memo is as follows. Section 2 outlines our 136 rationale in preparing this memo and choosing these RTP features. 137 Section 3 defines terminology. Requirements for core RTP protocols 138 are described in Section 4 and suggested RTP extensions are described 139 in Section 5. Section 6 outlines mechanisms that can increase 140 robustness to network problems, while Section 7 describes congestion 141 control and rate adaptation mechanisms. The discussion of mandated 142 RTP mechanisms concludes in Section 8 with a review of performance 143 monitoring and network management tools. Section 9 gives some 144 guidelines for future incorporation of other RTP and RTP Control 145 Protocol (RTCP) extensions into this framework. Section 10 describes 146 requirements placed on the signalling channel. Section 11 discusses 147 the relationship between features of the RTP framework and the WebRTC 148 application programming interface (API), and Section 12 discusses RTP 149 implementation considerations. The memo concludes with security 150 considerations (Section 13) and IANA considerations (Section 14). 152 2. Rationale 154 The RTP framework comprises the RTP data transfer protocol, the RTP 155 control protocol, and numerous RTP payload formats, profiles, and 156 extensions. This range of add-ons has allowed RTP to meet various 157 needs that were not envisaged by the original protocol designers, and 158 to support many new media encodings, but raises the question of what 159 extensions are to be supported by new implementations. The 160 development of the WebRTC framework provides an opportunity to review 161 the available RTP features and extensions, and to define a common 162 baseline RTP feature set for all WebRTC Endpoints. This builds on 163 the past 20 years of RTP development to mandate the use of extensions 164 that have shown widespread utility, while still remaining compatible 165 with the wide installed base of RTP implementations where possible. 167 RTP and RTCP extensions that are not discussed in this document can 168 be implemented by WebRTC Endpoints if they are beneficial for new use 169 cases. However, they are not necessary to address the WebRTC use 170 cases and requirements identified in [RFC7478]. 172 While the baseline set of RTP features and extensions defined in this 173 memo is targeted at the requirements of the WebRTC framework, it is 174 expected to be broadly useful for other conferencing-related uses of 175 RTP. In particular, it is likely that this set of RTP features and 176 extensions will be appropriate for other desktop or mobile video 177 conferencing systems, or for room-based high-quality telepresence 178 applications. 180 3. Terminology 182 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 183 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 184 document are to be interpreted as described in [RFC2119]. The RFC 185 2119 interpretation of these key words applies only when written in 186 ALL CAPS. Lower- or mixed-case uses of these key words are not to be 187 interpreted as carrying special significance in this memo. 189 We define the following additional terms: 191 WebRTC MediaStream: The MediaStream concept defined by the W3C in 192 the WebRTC API [W3C.WD-mediacapture-streams-20130903]. A 193 MediaStream consists of zero or more MediaStreamTracks. 195 MediaStreamTrack: Part of the MediaStream concept defined by the W3C 196 in the WebRTC API [W3C.WD-mediacapture-streams-20130903]. A 197 MediaStreamTrack is an individual stream of media from any type of 198 media source like a microphone or a camera, but also conceptual 199 sources, like a audio mix or a video composition, are possible. 201 Transport-layer Flow: A uni-directional flow of transport packets 202 that are identified by having a particular 5-tuple of source IP 203 address, source port, destination IP address, destination port, 204 and transport protocol used. 206 Bi-directional Transport-layer Flow: A bi-directional transport- 207 layer flow is a transport-layer flow that is symmetric. That is, 208 the transport-layer flow in the reverse direction has a 5-tuple 209 where the source and destination address and ports are swapped 210 compared to the forward path transport-layer flow, and the 211 transport protocol is the same. 213 This document uses the terminology from 214 [I-D.ietf-avtext-rtp-grouping-taxonomy] and 215 [I-D.ietf-rtcweb-overview]. Other terms are used according to their 216 definitions from the RTP Specification [RFC3550]. Especially note 217 the following frequently used terms: RTP Stream, RTP Session, and 218 Endpoint. 220 4. WebRTC Use of RTP: Core Protocols 222 The following sections describe the core features of RTP and RTCP 223 that need to be implemented, along with the mandated RTP profiles. 224 Also described are the core extensions providing essential features 225 that all WebRTC Endpoints need to implement to function effectively 226 on today's networks. 228 4.1. RTP and RTCP 230 The Real-time Transport Protocol (RTP) [RFC3550] is REQUIRED to be 231 implemented as the media transport protocol for WebRTC. RTP itself 232 comprises two parts: the RTP data transfer protocol, and the RTP 233 control protocol (RTCP). RTCP is a fundamental and integral part of 234 RTP, and MUST be implemented and used in all WebRTC Endpoints. 236 The following RTP and RTCP features are sometimes omitted in limited 237 functionality implementations of RTP, but are REQUIRED in all WebRTC 238 Endpoints: 240 o Support for use of multiple simultaneous SSRC values in a single 241 RTP session, including support for RTP endpoints that send many 242 SSRC values simultaneously, following [RFC3550] and 243 [I-D.ietf-avtcore-rtp-multi-stream]. The RTCP optimisations for 244 multi-SSRC sessions defined in 245 [I-D.ietf-avtcore-rtp-multi-stream-optimisation] MAY be supported; 246 if supported the usage MUST be signalled. 248 o Random choice of SSRC on joining a session; collision detection 249 and resolution for SSRC values (see also Section 4.8). 251 o Support for reception of RTP data packets containing CSRC lists, 252 as generated by RTP mixers, and RTCP packets relating to CSRCs. 254 o Sending correct synchronisation information in the RTCP Sender 255 Reports, to allow receivers to implement lip-synchronisation; see 256 Section 5.2.1 regarding support for the rapid RTP synchronisation 257 extensions. 259 o Support for multiple synchronisation contexts. Participants that 260 send multiple simultaneous RTP packet streams SHOULD do so as part 261 of a single synchronisation context, using a single RTCP CNAME for 262 all streams and allowing receivers to play the streams out in a 263 synchronised manner. For compatibility with potential future 264 versions of this specification, or for interoperability with non- 265 WebRTC devices through a gateway, receivers MUST support multiple 266 synchronisation contexts, indicated by the use of multiple RTCP 267 CNAMEs in an RTP session. This specification mandates the usage 268 of a single CNAME when sending RTP Streams in some circumstances, 269 see Section 4.9. 271 o Support for sending and receiving RTCP SR, RR, SDES, and BYE 272 packet types. Note that support for other RTCP packet types is 273 OPTIONAL, unless mandated by other parts of this specification. 274 Note that additional RTCP Packet types are used by the RTP/SAVPF 275 Profile (Section 4.2) and the other RTCP extensions (Section 5). 276 WebRTC endpoints that implement the SDP bundle negotiation 277 extension will use the SDP grouping framework 'mid' attribute to 278 identify media streams. Such endpoints MUST implement the RTCP 279 SDES MID item described in 280 [I-D.ietf-mmusic-sdp-bundle-negotiation]. 282 o Support for multiple endpoints in a single RTP session, and for 283 scaling the RTCP transmission interval according to the number of 284 participants in the session; support for randomised RTCP 285 transmission intervals to avoid synchronisation of RTCP reports; 286 support for RTCP timer reconsideration (Section 6.3.6 of 288 [RFC3550]) and reverse reconsideration (Section 6.3.4 of 289 [RFC3550]). 291 o Support for configuring the RTCP bandwidth as a fraction of the 292 media bandwidth, and for configuring the fraction of the RTCP 293 bandwidth allocated to senders, e.g., using the SDP "b=" line 294 [RFC4566][RFC3556]. 296 o Support for the reduced minimum RTCP reporting interval described 297 in Section 6.2 of [RFC3550]. When using the reduced minimum RTCP 298 reporting interval, the fixed (non-reduced) minimum interval MUST 299 be used when calculating the participant timeout interval (see 300 Sections 6.2 and 6.3.5 of [RFC3550]). The delay before sending 301 the initial compound RTCP packet can be set to zero (see 302 Section 6.2 of [RFC3550] as updated by 303 [I-D.ietf-avtcore-rtp-multi-stream]). 305 o Support for discontinuous transmission. RTP allows endpoints to 306 pause and resume transmission at any time. When resuming, the RTP 307 sequence number will increase by one, as usual, while the increase 308 in the RTP timestamp value will depend on the duration of the 309 pause. Discontinuous transmission is most commonly used with some 310 audio payload formats, but is not audio specific, and can be used 311 with any RTP payload format. 313 o Ignore unknown RTCP packet types and RTP header extensions. This 314 is to ensure robust handling of future extensions, middlebox 315 behaviours, etc., that can result in not signalled RTCP packet 316 types or RTP header extensions being received. If a compound RTCP 317 packet is received that contains a mixture of known and unknown 318 RTCP packet types, the known packets types need to be processed as 319 usual, with only the unknown packet types being discarded. 321 It is known that a significant number of legacy RTP implementations, 322 especially those targeted at VoIP-only systems, do not support all of 323 the above features, and in some cases do not support RTCP at all. 324 Implementers are advised to consider the requirements for graceful 325 degradation when interoperating with legacy implementations. 327 Other implementation considerations are discussed in Section 12. 329 4.2. Choice of the RTP Profile 331 The complete specification of RTP for a particular application domain 332 requires the choice of an RTP Profile. For WebRTC use, the Extended 333 Secure RTP Profile for RTCP-Based Feedback (RTP/SAVPF) [RFC5124], as 334 extended by [RFC7007], MUST be implemented. The RTP/SAVPF profile is 335 the combination of basic RTP/AVP profile [RFC3551], the RTP profile 336 for RTCP-based feedback (RTP/AVPF) [RFC4585], and the secure RTP 337 profile (RTP/SAVP) [RFC3711]. 339 The RTCP-based feedback extensions [RFC4585] are needed for the 340 improved RTCP timer model. This allows more flexible transmission of 341 RTCP packets in response to events, rather than strictly according to 342 bandwidth, and is vital for being able to report congestion signals 343 as well as media events. These extensions also allow saving RTCP 344 bandwidth, and an endpoint will commonly only use the full RTCP 345 bandwidth allocation if there are many events that require feedback. 346 The timer rules are also needed to make use of the RTP conferencing 347 extensions discussed in Section 5.1. 349 Note: The enhanced RTCP timer model defined in the RTP/AVPF 350 profile is backwards compatible with legacy systems that implement 351 only the RTP/AVP or RTP/SAVP profile, given some constraints on 352 parameter configuration such as the RTCP bandwidth value and "trr- 353 int" (the most important factor for interworking with RTP/(S)AVP 354 endpoints via a gateway is to set the trr-int parameter to a value 355 representing 4 seconds, see Section 6.1 in 356 [I-D.ietf-avtcore-rtp-multi-stream]). 358 The secure RTP (SRTP) profile extensions [RFC3711] are needed to 359 provide media encryption, integrity protection, replay protection and 360 a limited form of source authentication. WebRTC Endpoints MUST NOT 361 send packets using the basic RTP/AVP profile or the RTP/AVPF profile; 362 they MUST employ the full RTP/SAVPF profile to protect all RTP and 363 RTCP packets that are generated (i.e., implementations MUST use SRTP 364 and SRTCP). The RTP/SAVPF profile MUST be configured using the 365 cipher suites, DTLS-SRTP protection profiles, keying mechanisms, and 366 other parameters described in [I-D.ietf-rtcweb-security-arch]. 368 4.3. Choice of RTP Payload Formats 370 Mandatory to implement audio codecs and RTP payload formats for 371 WebRTC endpoints are defined in [I-D.ietf-rtcweb-audio]. Mandatory 372 to implement video codecs and RTP payload formats for WebRTC 373 endpoints are defined in [I-D.ietf-rtcweb-video]. WebRTC endpoints 374 MAY additionally implement any other codec for which an RTP payload 375 format and associated signalling has been defined. 377 WebRTC Endpoints cannot assume that the other participants in an RTP 378 session understand any RTP payload format, no matter how common. The 379 mapping between RTP payload type numbers and specific configurations 380 of particular RTP payload formats MUST be agreed before those payload 381 types/formats can be used. In an SDP context, this can be done using 382 the "a=rtpmap:" and "a=fmtp:" attributes associated with an "m=" 383 line, along with any other SDP attributes needed to configure the RTP 384 payload format. 386 Endpoints can signal support for multiple RTP payload formats, or 387 multiple configurations of a single RTP payload format, as long as 388 each unique RTP payload format configuration uses a different RTP 389 payload type number. As outlined in Section 4.8, the RTP payload 390 type number is sometimes used to associate an RTP packet stream with 391 a signalling context. This association is possible provided unique 392 RTP payload type numbers are used in each context. For example, an 393 RTP packet stream can be associated with an SDP "m=" line by 394 comparing the RTP payload type numbers used by the RTP packet stream 395 with payload types signalled in the "a=rtpmap:" lines in the media 396 sections of the SDP. This leads to the following considerations: 398 If RTP packet streams are being associated with signalling 399 contexts based on the RTP payload type, then the assignment of RTP 400 payload type numbers MUST be unique across signalling contexts. 402 If the same RTP payload format configuration is used in multiple 403 contexts, then a different RTP payload type number has to be 404 assigned in each context to ensure uniqueness. 406 If the RTP payload type number is not being used to associate RTP 407 packet streams with a signalling context, then the same RTP 408 payload type number can be used to indicate the exact same RTP 409 payload format configuration in multiple contexts. 411 A single RTP payload type number MUST NOT be assigned to different 412 RTP payload formats, or different configurations of the same RTP 413 payload format, within a single RTP session (note that the "m=" lines 414 in an SDP bundle group [I-D.ietf-mmusic-sdp-bundle-negotiation] form 415 a single RTP session). 417 An endpoint that has signalled support for multiple RTP payload 418 formats MUST be able to accept data in any of those payload formats 419 at any time, unless it has previously signalled limitations on its 420 decoding capability. This requirement is constrained if several 421 types of media (e.g., audio and video) are sent in the same RTP 422 session. In such a case, a source (SSRC) is restricted to switching 423 only between the RTP payload formats signalled for the type of media 424 that is being sent by that source; see Section 4.4. To support rapid 425 rate adaptation by changing codec, RTP does not require advance 426 signalling for changes between RTP payload formats used by a single 427 SSRC that were signalled during session set-up. 429 If performing changes between two RTP payload types that use 430 different RTP clock rates, an RTP sender MUST follow the 431 recommendations in Section 4.1 of [RFC7160]. RTP receivers MUST 432 follow the recommendations in Section 4.3 of [RFC7160] in order to 433 support sources that switch between clock rates in an RTP session 434 (these recommendations for receivers are backwards compatible with 435 the case where senders use only a single clock rate). 437 4.4. Use of RTP Sessions 439 An association amongst a set of endpoints communicating using RTP is 440 known as an RTP session [RFC3550]. An endpoint can be involved in 441 several RTP sessions at the same time. In a multimedia session, each 442 type of media has typically been carried in a separate RTP session 443 (e.g., using one RTP session for the audio, and a separate RTP 444 session using a different transport-layer flow for the video). 445 WebRTC Endpoints are REQUIRED to implement support for multimedia 446 sessions in this way, separating each RTP session using different 447 transport-layer flows for compatibility with legacy systems (this is 448 sometimes called session multiplexing). 450 In modern day networks, however, with the widespread use of network 451 address/port translators (NAT/NAPT) and firewalls, it is desirable to 452 reduce the number of transport-layer flows used by RTP applications. 453 This can be done by sending all the RTP packet streams in a single 454 RTP session, which will comprise a single transport-layer flow (this 455 will prevent the use of some quality-of-service mechanisms, as 456 discussed in Section 12.1.3). Implementations are therefore also 457 REQUIRED to support transport of all RTP packet streams, independent 458 of media type, in a single RTP session using a single transport layer 459 flow, according to [I-D.ietf-avtcore-multi-media-rtp-session] (this 460 is sometimes called SSRC multiplexing). If multiple types of media 461 are to be used in a single RTP session, all participants in that RTP 462 session MUST agree to this usage. In an SDP context, 463 [I-D.ietf-mmusic-sdp-bundle-negotiation] can be used to signal such a 464 bundle of RTP packet streams forming a single RTP session. 466 Further discussion about the suitability of different RTP session 467 structures and multiplexing methods to different scenarios can be 468 found in [I-D.ietf-avtcore-multiplex-guidelines]. 470 4.5. RTP and RTCP Multiplexing 472 Historically, RTP and RTCP have been run on separate transport layer 473 flows (e.g., two UDP ports for each RTP session, one port for RTP and 474 one port for RTCP). With the increased use of Network Address/Port 475 Translation (NAT/NAPT) this has become problematic, since maintaining 476 multiple NAT bindings can be costly. It also complicates firewall 477 administration, since multiple ports need to be opened to allow RTP 478 traffic. To reduce these costs and session set-up times, 479 implementations are REQUIRED to support multiplexing RTP data packets 480 and RTCP control packets on a single transport-layer flow [RFC5761]. 481 Such RTP and RTCP multiplexing MUST be negotiated in the signalling 482 channel before it is used. If SDP is used for signalling, this 483 negotiation MUST use the mechanism defined in [RFC5761]. 484 Implementations can also support sending RTP and RTCP on separate 485 transport-layer flows, but this is OPTIONAL to implement. If an 486 implementation does not support RTP and RTCP sent on separate 487 transport-layer flows, it MUST indicate that using the mechanism 488 defined in [I-D.ietf-mmusic-mux-exclusive]. 490 Note that the use of RTP and RTCP multiplexed onto a single 491 transport-layer flow ensures that there is occasional traffic sent on 492 that port, even if there is no active media traffic. This can be 493 useful to keep NAT bindings alive [RFC6263]. 495 4.6. Reduced Size RTCP 497 RTCP packets are usually sent as compound RTCP packets, and [RFC3550] 498 requires that those compound packets start with an Sender Report (SR) 499 or Receiver Report (RR) packet. When using frequent RTCP feedback 500 messages under the RTP/AVPF Profile [RFC4585] these statistics are 501 not needed in every packet, and unnecessarily increase the mean RTCP 502 packet size. This can limit the frequency at which RTCP packets can 503 be sent within the RTCP bandwidth share. 505 To avoid this problem, [RFC5506] specifies how to reduce the mean 506 RTCP message size and allow for more frequent feedback. Frequent 507 feedback, in turn, is essential to make real-time applications 508 quickly aware of changing network conditions, and to allow them to 509 adapt their transmission and encoding behaviour. Implementations 510 MUST support sending and receiving non-compound RTCP feedback packets 511 [RFC5506]. Use of non-compound RTCP packets MUST be negotiated using 512 the signalling channel. If SDP is used for signalling, this 513 negotiation MUST use the attributes defined in [RFC5506]. For 514 backwards compatibility, implementations are also REQUIRED to support 515 the use of compound RTCP feedback packets if the remote endpoint does 516 not agree to the use of non-compound RTCP in the signalling exchange. 518 4.7. Symmetric RTP/RTCP 520 To ease traversal of NAT and firewall devices, implementations are 521 REQUIRED to implement and use Symmetric RTP [RFC4961]. The reason 522 for using symmetric RTP is primarily to avoid issues with NATs and 523 Firewalls by ensuring that the send and receive RTP packet streams, 524 as well as RTCP, are actually bi-directional transport-layer flows. 525 This will keep alive the NAT and firewall pinholes, and help indicate 526 consent that the receive direction is a transport-layer flow the 527 intended recipient actually wants. In addition, it saves resources, 528 specifically ports at the endpoints, but also in the network as NAT 529 mappings or firewall state is not unnecessary bloated. The amount of 530 per flow QoS state kept in the network is also reduced. 532 4.8. Choice of RTP Synchronisation Source (SSRC) 534 Implementations are REQUIRED to support signalled RTP synchronisation 535 source (SSRC) identifiers. If SDP is used, this MUST be done using 536 the "a=ssrc:" SDP attribute defined in Section 4.1 and Section 5 of 537 [RFC5576] and the "previous-ssrc" source attribute defined in 538 Section 6.2 of [RFC5576]; other per-SSRC attributes defined in 539 [RFC5576] MAY be supported. 541 While support for signalled SSRC identifiers is mandated, their use 542 in an RTP session is OPTIONAL. Implementations MUST be prepared to 543 accept RTP and RTCP packets using SSRCs that have not been explicitly 544 signalled ahead of time. Implementations MUST support random SSRC 545 assignment, and MUST support SSRC collision detection and resolution, 546 according to [RFC3550]. When using signalled SSRC values, collision 547 detection MUST be performed as described in Section 5 of [RFC5576]. 549 It is often desirable to associate an RTP packet stream with a non- 550 RTP context. For users of the WebRTC API a mapping between SSRCs and 551 MediaStreamTracks are provided per Section 11. For gateways or other 552 usages it is possible to associate an RTP packet stream with an "m=" 553 line in a session description formatted using SDP. If SSRCs are 554 signalled this is straightforward (in SDP the "a=ssrc:" line will be 555 at the media level, allowing a direct association with an "m=" line). 556 If SSRCs are not signalled, the RTP payload type numbers used in an 557 RTP packet stream are often sufficient to associate that packet 558 stream with a signalling context (e.g., if RTP payload type numbers 559 are assigned as described in Section 4.3 of this memo, the RTP 560 payload types used by an RTP packet stream can be compared with 561 values in SDP "a=rtpmap:" lines, which are at the media level in SDP, 562 and so map to an "m=" line). 564 4.9. Generation of the RTCP Canonical Name (CNAME) 566 The RTCP Canonical Name (CNAME) provides a persistent transport-level 567 identifier for an RTP endpoint. While the Synchronisation Source 568 (SSRC) identifier for an RTP endpoint can change if a collision is 569 detected, or when the RTP application is restarted, its RTCP CNAME is 570 meant to stay unchanged for the duration of a RTCPeerConnection 571 [W3C.WD-webrtc-20130910], so that RTP endpoints can be uniquely 572 identified and associated with their RTP packet streams within a set 573 of related RTP sessions. 575 Each RTP endpoint MUST have at least one RTCP CNAME, and that RTCP 576 CNAME MUST be unique within the RTCPeerConnection. RTCP CNAMEs 577 identify a particular synchronisation context, i.e., all SSRCs 578 associated with a single RTCP CNAME share a common reference clock. 579 If an endpoint has SSRCs that are associated with several 580 unsynchronised reference clocks, and hence different synchronisation 581 contexts, it will need to use multiple RTCP CNAMEs, one for each 582 synchronisation context. 584 Taking the discussion in Section 11 into account, a WebRTC Endpoint 585 MUST NOT use more than one RTCP CNAME in the RTP sessions belonging 586 to single RTCPeerConnection (that is, an RTCPeerConnection forms a 587 synchronisation context). RTP middleboxes MAY generate RTP packet 588 streams associated with more than one RTCP CNAME, to allow them to 589 avoid having to resynchronize media from multiple different endpoints 590 part of a multi-party RTP session. 592 The RTP specification [RFC3550] includes guidelines for choosing a 593 unique RTP CNAME, but these are not sufficient in the presence of NAT 594 devices. In addition, long-term persistent identifiers can be 595 problematic from a privacy viewpoint (Section 13). Accordingly, a 596 WebRTC Endpoint MUST generate a new, unique, short-term persistent 597 RTCP CNAME for each RTCPeerConnection, following [RFC7022], with a 598 single exception; if explicitly requested at creation an 599 RTCPeerConnection MAY use the same CNAME as as an existing 600 RTCPeerConnection within their common same-origin context. 602 An WebRTC Endpoint MUST support reception of any CNAME that matches 603 the syntax limitations specified by the RTP specification [RFC3550] 604 and cannot assume that any CNAME will be chosen according to the form 605 suggested above. 607 4.10. Handling of Leap Seconds 609 The guidelines regarding handling of leap seconds to limit their 610 impact on RTP media play-out and synchronization given in [RFC7164] 611 SHOULD be followed. 613 5. WebRTC Use of RTP: Extensions 615 There are a number of RTP extensions that are either needed to obtain 616 full functionality, or extremely useful to improve on the baseline 617 performance, in the WebRTC context. One set of these extensions is 618 related to conferencing, while others are more generic in nature. 619 The following subsections describe the various RTP extensions 620 mandated or suggested for use within WebRTC. 622 5.1. Conferencing Extensions and Topologies 624 RTP is a protocol that inherently supports group communication. 625 Groups can be implemented by having each endpoint send its RTP packet 626 streams to an RTP middlebox that redistributes the traffic, by using 627 a mesh of unicast RTP packet streams between endpoints, or by using 628 an IP multicast group to distribute the RTP packet streams. These 629 topologies can be implemented in a number of ways as discussed in 630 [I-D.ietf-avtcore-rtp-topologies-update]. 632 While the use of IP multicast groups is popular in IPTV systems, the 633 topologies based on RTP middleboxes are dominant in interactive video 634 conferencing environments. Topologies based on a mesh of unicast 635 transport-layer flows to create a common RTP session have not seen 636 widespread deployment to date. Accordingly, WebRTC Endpoints are not 637 expected to support topologies based on IP multicast groups or to 638 support mesh-based topologies, such as a point-to-multipoint mesh 639 configured as a single RTP session (Topo-Mesh in the terminology of 640 [I-D.ietf-avtcore-rtp-topologies-update]). However, a point-to- 641 multipoint mesh constructed using several RTP sessions, implemented 642 in WebRTC using independent RTCPeerConnections 643 [W3C.WD-webrtc-20130910], can be expected to be used in WebRTC, and 644 needs to be supported. 646 WebRTC Endpoints implemented according to this memo are expected to 647 support all the topologies described in 648 [I-D.ietf-avtcore-rtp-topologies-update] where the RTP endpoints send 649 and receive unicast RTP packet streams to and from some peer device, 650 provided that peer can participate in performing congestion control 651 on the RTP packet streams. The peer device could be another RTP 652 endpoint, or it could be an RTP middlebox that redistributes the RTP 653 packet streams to other RTP endpoints. This limitation means that 654 some of the RTP middlebox-based topologies are not suitable for use 655 in WebRTC. Specifically: 657 o Video switching MCUs (Topo-Video-switch-MCU) SHOULD NOT be used, 658 since they make the use of RTCP for congestion control and quality 659 of service reports problematic (see Section 3.8 of 660 [I-D.ietf-avtcore-rtp-topologies-update]). 662 o The Relay-Transport Translator (Topo-PtM-Trn-Translator) topology 663 SHOULD NOT be used because its safe use requires a congestion 664 control algorithm or RTP circuit breaker that handles point to 665 multipoint, which has not yet been standardised. 667 The following topology can be used, however it has some issues worth 668 noting: 670 o Content modifying MCUs with RTCP termination (Topo-RTCP- 671 terminating-MCU) MAY be used. Note that in this RTP Topology, RTP 672 loop detection and identification of active senders is the 673 responsibility of the WebRTC application; since the clients are 674 isolated from each other at the RTP layer, RTP cannot assist with 675 these functions (see section 3.9 of 676 [I-D.ietf-avtcore-rtp-topologies-update]). 678 The RTP extensions described in Section 5.1.1 to Section 5.1.6 are 679 designed to be used with centralised conferencing, where an RTP 680 middlebox (e.g., a conference bridge) receives a participant's RTP 681 packet streams and distributes them to the other participants. These 682 extensions are not necessary for interoperability; an RTP endpoint 683 that does not implement these extensions will work correctly, but 684 might offer poor performance. Support for the listed extensions will 685 greatly improve the quality of experience and, to provide a 686 reasonable baseline quality, some of these extensions are mandatory 687 to be supported by WebRTC Endpoints. 689 The RTCP conferencing extensions are defined in Extended RTP Profile 690 for Real-time Transport Control Protocol (RTCP)-Based Feedback (RTP/ 691 AVPF) [RFC4585] and the memo on Codec Control Messages (CCM) in RTP/ 692 AVPF [RFC5104]; they are fully usable by the Secure variant of this 693 profile (RTP/SAVPF) [RFC5124]. 695 5.1.1. Full Intra Request (FIR) 697 The Full Intra Request message is defined in Sections 3.5.1 and 4.3.1 698 of the Codec Control Messages [RFC5104]. It is used to make the 699 mixer request a new Intra picture from a participant in the session. 700 This is used when switching between sources to ensure that the 701 receivers can decode the video or other predictive media encoding 702 with long prediction chains. WebRTC Endpoints that are sending media 703 MUST understand and react to FIR feedback messages they receive, 704 since this greatly improves the user experience when using 705 centralised mixer-based conferencing. Support for sending FIR 706 messages is OPTIONAL. 708 5.1.2. Picture Loss Indication (PLI) 710 The Picture Loss Indication message is defined in Section 6.3.1 of 711 the RTP/AVPF profile [RFC4585]. It is used by a receiver to tell the 712 sending encoder that it lost the decoder context and would like to 713 have it repaired somehow. This is semantically different from the 714 Full Intra Request above as there could be multiple ways to fulfil 715 the request. WebRTC Endpoints that are sending media MUST understand 716 and react to PLI feedback messages as a loss tolerance mechanism. 717 Receivers MAY send PLI messages. 719 5.1.3. Slice Loss Indication (SLI) 721 The Slice Loss Indication message is defined in Section 6.3.2 of the 722 RTP/AVPF profile [RFC4585]. It is used by a receiver to tell the 723 encoder that it has detected the loss or corruption of one or more 724 consecutive macro blocks, and would like to have these repaired 725 somehow. It is RECOMMENDED that receivers generate SLI feedback 726 messages if slices are lost when using a codec that supports the 727 concept of macro blocks. A sender that receives an SLI feedback 728 message SHOULD attempt to repair the lost slice(s). 730 5.1.4. Reference Picture Selection Indication (RPSI) 732 Reference Picture Selection Indication (RPSI) messages are defined in 733 Section 6.3.3 of the RTP/AVPF profile [RFC4585]. Some video encoding 734 standards allow the use of older reference pictures than the most 735 recent one for predictive coding. If such a codec is in use, and if 736 the encoder has learnt that encoder-decoder synchronisation has been 737 lost, then a known as correct reference picture can be used as a base 738 for future coding. The RPSI message allows this to be signalled. 739 Receivers that detect that encoder-decoder synchronisation has been 740 lost SHOULD generate an RPSI feedback message if codec being used 741 supports reference picture selection. A RTP packet stream sender 742 that receives such an RPSI message SHOULD act on that messages to 743 change the reference picture, if it is possible to do so within the 744 available bandwidth constraints, and with the codec being used. 746 5.1.5. Temporal-Spatial Trade-off Request (TSTR) 748 The temporal-spatial trade-off request and notification are defined 749 in Sections 3.5.2 and 4.3.2 of [RFC5104]. This request can be used 750 to ask the video encoder to change the trade-off it makes between 751 temporal and spatial resolution, for example to prefer high spatial 752 image quality but low frame rate. Support for TSTR requests and 753 notifications is OPTIONAL. 755 5.1.6. Temporary Maximum Media Stream Bit Rate Request (TMMBR) 757 The TMMBR feedback message is defined in Sections 3.5.4 and 4.2.1 of 758 the Codec Control Messages [RFC5104]. This request and its 759 notification message are used by a media receiver to inform the 760 sending party that there is a current limitation on the amount of 761 bandwidth available to this receiver. There can be various reasons 762 for this: for example, an RTP mixer can use this message to limit the 763 media rate of the sender being forwarded by the mixer (without doing 764 media transcoding) to fit the bottlenecks existing towards the other 765 session participants. WebRTC Endpoints that are sending media are 766 REQUIRED to implement support for TMMBR messages, and MUST follow 767 bandwidth limitations set by a TMMBR message received for their SSRC. 768 The sending of TMMBR requests is OPTIONAL. 770 5.2. Header Extensions 772 The RTP specification [RFC3550] provides the capability to include 773 RTP header extensions containing in-band data, but the format and 774 semantics of the extensions are poorly specified. The use of header 775 extensions is OPTIONAL in WebRTC, but if they are used, they MUST be 776 formatted and signalled following the general mechanism for RTP 777 header extensions defined in [RFC5285], since this gives well-defined 778 semantics to RTP header extensions. 780 As noted in [RFC5285], the requirement from the RTP specification 781 that header extensions are "designed so that the header extension may 782 be ignored" [RFC3550] stands. To be specific, header extensions MUST 783 only be used for data that can safely be ignored by the recipient 784 without affecting interoperability, and MUST NOT be used when the 785 presence of the extension has changed the form or nature of the rest 786 of the packet in a way that is not compatible with the way the stream 787 is signalled (e.g., as defined by the payload type). Valid examples 788 of RTP header extensions might include metadata that is additional to 789 the usual RTP information, but that can safely be ignored without 790 compromising interoperability. 792 5.2.1. Rapid Synchronisation 794 Many RTP sessions require synchronisation between audio, video, and 795 other content. This synchronisation is performed by receivers, using 796 information contained in RTCP SR packets, as described in the RTP 797 specification [RFC3550]. This basic mechanism can be slow, however, 798 so it is RECOMMENDED that the rapid RTP synchronisation extensions 799 described in [RFC6051] be implemented in addition to RTCP SR-based 800 synchronisation. 802 This header extension uses the [RFC5285] generic header extension 803 framework, and so needs to be negotiated before it can be used. 805 5.2.2. Client-to-Mixer Audio Level 807 The Client to Mixer Audio Level extension [RFC6464] is an RTP header 808 extension used by an endpoint to inform a mixer about the level of 809 audio activity in the packet to which the header is attached. This 810 enables an RTP middlebox to make mixing or selection decisions 811 without decoding or detailed inspection of the payload, reducing the 812 complexity in some types of mixers. It can also save decoding 813 resources in receivers, which can choose to decode only the most 814 relevant RTP packet streams based on audio activity levels. 816 The Client-to-Mixer Audio Level [RFC6464] header extension MUST be 817 implemented. It is REQUIRED that implementations are capable of 818 encrypting the header extension according to [RFC6904] since the 819 information contained in these header extensions can be considered 820 sensitive. The use of this encryption is RECOMMENDED, however usage 821 of the encryption can be explicitly disabled through API or 822 signalling. 824 This header extension uses the [RFC5285] generic header extension 825 framework, and so needs to be negotiated before it can be used. 827 5.2.3. Mixer-to-Client Audio Level 829 The Mixer to Client Audio Level header extension [RFC6465] provides 830 an endpoint with the audio level of the different sources mixed into 831 a common source stream by a RTP mixer. This enables a user interface 832 to indicate the relative activity level of each session participant, 833 rather than just being included or not based on the CSRC field. This 834 is a pure optimisation of non critical functions, and is hence 835 OPTIONAL to implement. If this header extension is implemented, it 836 is REQUIRED that implementations are capable of encrypting the header 837 extension according to [RFC6904] since the information contained in 838 these header extensions can be considered sensitive. It is further 839 RECOMMENDED that this encryption is used, unless the encryption has 840 been explicitly disabled through API or signalling. 842 This header extension uses the [RFC5285] generic header extension 843 framework, and so needs to be negotiated before it can be used. 845 5.2.4. Media Stream Identification 847 WebRTC endpoints that implement the SDP bundle negotiation extension 848 will use the SDP grouping framework 'mid' attribute to identify media 849 streams. Such endpoints MUST implement the RTP MID header extension 850 described in [I-D.ietf-mmusic-sdp-bundle-negotiation]. 852 This header extension uses the [RFC5285] generic header extension 853 framework, and so needs to be negotiated before it can be used. 855 5.2.5. Coordination of Video Orientation 857 WebRTC endpoints that send or receive video MUST implement the 858 coordination of video orientation (CVO) RTP header extension as 859 described in Section 4 of [I-D.ietf-rtcweb-video]. 861 This header extension uses the [RFC5285] generic header extension 862 framework, and so needs to be negotiated before it can be used. 864 6. WebRTC Use of RTP: Improving Transport Robustness 866 There are tools that can make RTP packet streams robust against 867 packet loss and reduce the impact of loss on media quality. However, 868 they generally add some overhead compared to a non-robust stream. 869 The overhead needs to be considered, and the aggregate bit-rate MUST 870 be rate controlled to avoid causing network congestion (see 871 Section 7). As a result, improving robustness might require a lower 872 base encoding quality, but has the potential to deliver that quality 873 with fewer errors. The mechanisms described in the following sub- 874 sections can be used to improve tolerance to packet loss. 876 6.1. Negative Acknowledgements and RTP Retransmission 878 As a consequence of supporting the RTP/SAVPF profile, implementations 879 can send negative acknowledgements (NACKs) for RTP data packets 880 [RFC4585]. This feedback can be used to inform a sender of the loss 881 of particular RTP packets, subject to the capacity limitations of the 882 RTCP feedback channel. A sender can use this information to optimise 883 the user experience by adapting the media encoding to compensate for 884 known lost packets. 886 RTP packet stream senders are REQUIRED to understand the Generic NACK 887 message defined in Section 6.2.1 of [RFC4585], but MAY choose to 888 ignore some or all of this feedback (following Section 4.2 of 889 [RFC4585]). Receivers MAY send NACKs for missing RTP packets. 890 Guidelines on when to send NACKs are provided in [RFC4585]. It is 891 not expected that a receiver will send a NACK for every lost RTP 892 packet, rather it needs to consider the cost of sending NACK 893 feedback, and the importance of the lost packet, to make an informed 894 decision on whether it is worth telling the sender about a packet 895 loss event. 897 The RTP Retransmission Payload Format [RFC4588] offers the ability to 898 retransmit lost packets based on NACK feedback. Retransmission needs 899 to be used with care in interactive real-time applications to ensure 900 that the retransmitted packet arrives in time to be useful, but can 901 be effective in environments with relatively low network RTT (an RTP 902 sender can estimate the RTT to the receivers using the information in 903 RTCP SR and RR packets, as described at the end of Section 6.4.1 of 904 [RFC3550]). The use of retransmissions can also increase the forward 905 RTP bandwidth, and can potentially caused increased packet loss if 906 the original packet loss was caused by network congestion. Note, 907 however, that retransmission of an important lost packet to repair 908 decoder state can have lower cost than sending a full intra frame. 909 It is not appropriate to blindly retransmit RTP packets in response 910 to a NACK. The importance of lost packets and the likelihood of them 911 arriving in time to be useful needs to be considered before RTP 912 retransmission is used. 914 Receivers are REQUIRED to implement support for RTP retransmission 915 packets [RFC4588] sent using SSRC multiplexing, and MAY also support 916 RTP retransmission packets sent using session multiplexing. Senders 917 MAY send RTP retransmission packets in response to NACKs if support 918 for the RTP retransmission payload format has been negotiated, and if 919 the sender believes it is useful to send a retransmission of the 920 packet(s) referenced in the NACK. Senders do not need to retransmit 921 every NACKed packet. 923 6.2. Forward Error Correction (FEC) 925 The use of Forward Error Correction (FEC) can provide an effective 926 protection against some degree of packet loss, at the cost of steady 927 bandwidth overhead. There are several FEC schemes that are defined 928 for use with RTP. Some of these schemes are specific to a particular 929 RTP payload format, others operate across RTP packets and can be used 930 with any payload format. It needs to be noted that using redundant 931 encoding or FEC will lead to increased play out delay, which needs to 932 be considered when choosing FEC schemes and their parameters. 934 WebRTC endpoints MUST follow the recommendations for FEC use given in 935 [I-D.ietf-rtcweb-fec]. WebRTC endpoints MAY support other types of 936 FEC, but these MUST be negotiated before they are used. 938 7. WebRTC Use of RTP: Rate Control and Media Adaptation 940 WebRTC will be used in heterogeneous network environments using a 941 variety of link technologies, including both wired and wireless 942 links, to interconnect potentially large groups of users around the 943 world. As a result, the network paths between users can have widely 944 varying one-way delays, available bit-rates, load levels, and traffic 945 mixtures. Individual endpoints can send one or more RTP packet 946 streams to each participant, and there can be several participants. 947 Each of these RTP packet streams can contain different types of 948 media, and the type of media, bit rate, and number of RTP packet 949 streams as well as transport-layer flows can be highly asymmetric. 950 Non-RTP traffic can share the network paths with RTP transport-layer 951 flows. Since the network environment is not predictable or stable, 952 WebRTC Endpoints MUST ensure that the RTP traffic they generate can 953 adapt to match changes in the available network capacity. 955 The quality of experience for users of WebRTC is very dependent on 956 effective adaptation of the media to the limitations of the network. 957 Endpoints have to be designed so they do not transmit significantly 958 more data than the network path can support, except for very short 959 time periods, otherwise high levels of network packet loss or delay 960 spikes will occur, causing media quality degradation. The limiting 961 factor on the capacity of the network path might be the link 962 bandwidth, or it might be competition with other traffic on the link 963 (this can be non-WebRTC traffic, traffic due to other WebRTC flows, 964 or even competition with other WebRTC flows in the same session). 966 An effective media congestion control algorithm is therefore an 967 essential part of the WebRTC framework. However, at the time of this 968 writing, there is no standard congestion control algorithm that can 969 be used for interactive media applications such as WebRTC's flows. 970 Some requirements for congestion control algorithms for 971 RTCPeerConnections are discussed in [I-D.ietf-rmcat-cc-requirements]. 972 If a standardized congestion control algorithm that satisfies these 973 requirements is developed in the future, this memo will need to be be 974 updated to mandate its use. 976 7.1. Boundary Conditions and Circuit Breakers 978 WebRTC Endpoints MUST implement the RTP circuit breaker algorithm 979 that is described in [I-D.ietf-avtcore-rtp-circuit-breakers]. The 980 RTP circuit breaker is designed to enable applications to recognise 981 and react to situations of extreme network congestion. However, 982 since the RTP circuit breaker might not be triggered until congestion 983 becomes extreme, it cannot be considered a substitute for congestion 984 control, and applications MUST also implement congestion control to 985 allow them to adapt to changes in network capacity. The congestion 986 control algorithm will have to be proprietary until a standardized 987 congestion control algorithm is available. Any future RTP congestion 988 control algorithms are expected to operate within the envelope 989 allowed by the circuit breaker. 991 The session establishment signalling will also necessarily establish 992 boundaries to which the media bit-rate will conform. The choice of 993 media codecs provides upper- and lower-bounds on the supported bit- 994 rates that the application can utilise to provide useful quality, and 995 the packetisation choices that exist. In addition, the signalling 996 channel can establish maximum media bit-rate boundaries using, for 997 example, the SDP "b=AS:" or "b=CT:" lines and the RTP/AVPF Temporary 998 Maximum Media Stream Bit Rate (TMMBR) Requests (see Section 5.1.6 of 999 this memo). Signalled bandwidth limitations, such as SDP "b=AS:" or 1000 "b=CT:" lines received from the peer, MUST be followed when sending 1001 RTP packet streams. A WebRTC Endpoint receiving media SHOULD signal 1002 its bandwidth limitations. These limitations have to be based on 1003 known bandwidth limitations, for example the capacity of the edge 1004 links. 1006 7.2. Congestion Control Interoperability and Legacy Systems 1008 All endpoints that wish to interwork with WebRTC MUST implement RTCP 1009 and provide congestion feedback via the defined RTCP reporting 1010 mechanisms. 1012 When interworking with legacy implementations that support RTCP using 1013 the RTP/AVP profile [RFC3551], congestion feedback is provided in 1014 RTCP RR packets every few seconds. Implementations that have to 1015 interwork with such endpoints MUST ensure that they keep within the 1016 RTP circuit breaker [I-D.ietf-avtcore-rtp-circuit-breakers] 1017 constraints to limit the congestion they can cause. 1019 If a legacy endpoint supports RTP/AVPF, this enables negotiation of 1020 important parameters for frequent reporting, such as the "trr-int" 1021 parameter, and the possibility that the endpoint supports some useful 1022 feedback format for congestion control purpose such as TMMBR 1023 [RFC5104]. Implementations that have to interwork with such 1024 endpoints MUST ensure that they stay within the RTP circuit breaker 1025 [I-D.ietf-avtcore-rtp-circuit-breakers] constraints to limit the 1026 congestion they can cause, but might find that they can achieve 1027 better congestion response depending on the amount of feedback that 1028 is available. 1030 With proprietary congestion control algorithms issues can arise when 1031 different algorithms and implementations interact in a communication 1032 session. If the different implementations have made different 1033 choices in regards to the type of adaptation, for example one sender 1034 based, and one receiver based, then one could end up in situation 1035 where one direction is dual controlled, when the other direction is 1036 not controlled. This memo cannot mandate behaviour for proprietary 1037 congestion control algorithms, but implementations that use such 1038 algorithms ought to be aware of this issue, and try to ensure that 1039 effective congestion control is negotiated for media flowing in both 1040 directions. If the IETF were to standardise both sender- and 1041 receiver-based congestion control algorithms for WebRTC traffic in 1042 the future, the issues of interoperability, control, and ensuring 1043 that both directions of media flow are congestion controlled would 1044 also need to be considered. 1046 8. WebRTC Use of RTP: Performance Monitoring 1048 As described in Section 4.1, implementations are REQUIRED to generate 1049 RTCP Sender Report (SR) and Reception Report (RR) packets relating to 1050 the RTP packet streams they send and receive. These RTCP reports can 1051 be used for performance monitoring purposes, since they include basic 1052 packet loss and jitter statistics. 1054 A large number of additional performance metrics are supported by the 1055 RTCP Extended Reports (XR) framework, see [RFC3611][RFC6792]. At the 1056 time of this writing, it is not clear what extended metrics are 1057 suitable for use in WebRTC, so there is no requirement that 1058 implementations generate RTCP XR packets. However, implementations 1059 that can use detailed performance monitoring data MAY generate RTCP 1060 XR packets as appropriate. The use of RTCP XR packets SHOULD be 1061 signalled; implementations MUST ignore RTCP XR packets that are 1062 unexpected or not understood. 1064 9. WebRTC Use of RTP: Future Extensions 1066 It is possible that the core set of RTP protocols and RTP extensions 1067 specified in this memo will prove insufficient for the future needs 1068 of WebRTC. In this case, future updates to this memo have to be made 1069 following the Guidelines for Writers of RTP Payload Format 1070 Specifications [RFC2736], How to Write an RTP Payload Format 1071 [I-D.ietf-payload-rtp-howto] and Guidelines for Extending the RTP 1072 Control Protocol [RFC5968], and SHOULD take into account any future 1073 guidelines for extending RTP and related protocols that have been 1074 developed. 1076 Authors of future extensions are urged to consider the wide range of 1077 environments in which RTP is used when recommending extensions, since 1078 extensions that are applicable in some scenarios can be problematic 1079 in others. Where possible, the WebRTC framework will adopt RTP 1080 extensions that are of general utility, to enable easy implementation 1081 of a gateway to other applications using RTP, rather than adopt 1082 mechanisms that are narrowly targeted at specific WebRTC use cases. 1084 10. Signalling Considerations 1086 RTP is built with the assumption that an external signalling channel 1087 exists, and can be used to configure RTP sessions and their features. 1088 The basic configuration of an RTP session consists of the following 1089 parameters: 1091 RTP Profile: The name of the RTP profile to be used in session. The 1092 RTP/AVP [RFC3551] and RTP/AVPF [RFC4585] profiles can interoperate 1093 on basic level, as can their secure variants RTP/SAVP [RFC3711] 1094 and RTP/SAVPF [RFC5124]. The secure variants of the profiles do 1095 not directly interoperate with the non-secure variants, due to the 1096 presence of additional header fields for authentication in SRTP 1097 packets and cryptographic transformation of the payload. WebRTC 1098 requires the use of the RTP/SAVPF profile, and this MUST be 1099 signalled. Interworking functions might transform this into the 1100 RTP/SAVP profile for a legacy use case, by indicating to the 1101 WebRTC Endpoint that the RTP/SAVPF is used and configuring a trr- 1102 int value of 4 seconds. 1104 Transport Information: Source and destination IP address(s) and 1105 ports for RTP and RTCP MUST be signalled for each RTP session. In 1106 WebRTC these transport addresses will be provided by ICE [RFC5245] 1107 that signals candidates and arrives at nominated candidate address 1108 pairs. If RTP and RTCP multiplexing [RFC5761] is to be used, such 1109 that a single port, i.e. transport-layer flow, is used for RTP and 1110 RTCP flows, this MUST be signalled (see Section 4.5). 1112 RTP Payload Types, media formats, and format parameters: The mapping 1113 between media type names (and hence the RTP payload formats to be 1114 used), and the RTP payload type numbers MUST be signalled. Each 1115 media type MAY also have a number of media type parameters that 1116 MUST also be signalled to configure the codec and RTP payload 1117 format (the "a=fmtp:" line from SDP). Section 4.3 of this memo 1118 discusses requirements for uniqueness of payload types. 1120 RTP Extensions: The use of any additional RTP header extensions and 1121 RTCP packet types, including any necessary parameters, MUST be 1122 signalled. This signalling is to ensure that a WebRTC Endpoint's 1123 behaviour, especially when sending, of any extensions is 1124 predictable and consistent. For robustness, and for compatibility 1125 with non-WebRTC systems that might be connected to a WebRTC 1126 session via a gateway, implementations are REQUIRED to ignore 1127 unknown RTCP packets and RTP header extensions (see also 1128 Section 4.1). 1130 RTCP Bandwidth: Support for exchanging RTCP Bandwidth values to the 1131 endpoints will be necessary. This SHALL be done as described in 1132 "Session Description Protocol (SDP) Bandwidth Modifiers for RTP 1133 Control Protocol (RTCP) Bandwidth" [RFC3556] if using SDP, or 1134 something semantically equivalent. This also ensures that the 1135 endpoints have a common view of the RTCP bandwidth. A common view 1136 of the RTCP bandwidth among different endpoints is important, to 1137 prevent differences in RTCP packet timing and timeout intervals 1138 causing interoperability problems. 1140 These parameters are often expressed in SDP messages conveyed within 1141 an offer/answer exchange. RTP does not depend on SDP or on the 1142 offer/answer model, but does require all the necessary parameters to 1143 be agreed upon, and provided to the RTP implementation. Note that in 1144 WebRTC it will depend on the signalling model and API how these 1145 parameters need to be configured but they will be need to either be 1146 set in the API or explicitly signalled between the peers. 1148 11. WebRTC API Considerations 1150 The WebRTC API [W3C.WD-webrtc-20130910] and the Media Capture and 1151 Streams API [W3C.WD-mediacapture-streams-20130903] defines and uses 1152 the concept of a MediaStream that consists of zero or more 1153 MediaStreamTracks. A MediaStreamTrack is an individual stream of 1154 media from any type of media source like a microphone or a camera, 1155 but also conceptual sources, like a audio mix or a video composition, 1156 are possible. The MediaStreamTracks within a MediaStream might need 1157 to be synchronized during play back. 1159 A MediaStreamTrack's realisation in RTP in the context of an 1160 RTCPeerConnection consists of a source packet stream identified with 1161 an SSRC within an RTP session part of the RTCPeerConnection. The 1162 MediaStreamTrack can also result in additional packet streams, and 1163 thus SSRCs, in the same RTP session. These can be dependent packet 1164 streams from scalable encoding of the source stream associated with 1165 the MediaStreamTrack, if such a media encoder is used. They can also 1166 be redundancy packet streams, these are created when applying Forward 1167 Error Correction (Section 6.2) or RTP retransmission (Section 6.1) to 1168 the source packet stream. 1170 It is important to note that the same media source can be feeding 1171 multiple MediaStreamTracks. As different sets of constraints or 1172 other parameters can be applied to the MediaStreamTrack, each 1173 MediaStreamTrack instance added to a RTCPeerConnection SHALL result 1174 in an independent source packet stream, with its own set of 1175 associated packet streams, and thus different SSRC(s). It will 1176 depend on applied constraints and parameters if the source stream and 1177 the encoding configuration will be identical between different 1178 MediaStreamTracks sharing the same media source. If the encoding 1179 parameters and constraints are the same, an implementation could 1180 choose to use only one encoded stream to create the different RTP 1181 packet streams. Note that such optimisations would need to take into 1182 account that the constraints for one of the MediaStreamTracks can at 1183 any moment change, meaning that the encoding configurations might no 1184 longer be identical and two different encoder instances would then be 1185 needed. 1187 The same MediaStreamTrack can also be included in multiple 1188 MediaStreams, thus multiple sets of MediaStreams can implicitly need 1189 to use the same synchronisation base. To ensure that this works in 1190 all cases, and does not force an endpoint to disrupt the media by 1191 changing synchronisation base and CNAME during delivery of any 1192 ongoing packet streams, all MediaStreamTracks and their associated 1193 SSRCs originating from the same endpoint need to be sent using the 1194 same CNAME within one RTCPeerConnection. This is motivating the use 1195 of a single CNAME in Section 4.9. 1197 The requirement on using the same CNAME for all SSRCs that 1198 originate from the same endpoint, does not require a middlebox 1199 that forwards traffic from multiple endpoints to only use a single 1200 CNAME. 1202 Different CNAMEs normally need to be used for different 1203 RTCPeerConnection instances, as specified in Section 4.9. Having two 1204 communication sessions with the same CNAME could enable tracking of a 1205 user or device across different services (see Section 4.4.1 of 1206 [I-D.ietf-rtcweb-security] for details). A web application can 1207 request that the CNAMEs used in different RTCPeerConnections (within 1208 a same-orign context) be the same, this allows for synchronization of 1209 the endpoint's RTP packet streams across the different 1210 RTCPeerConnections. 1212 Note: this doesn't result in a tracking issue, since the creation 1213 of matching CNAMEs depends on existing tracking within a single 1214 origin. 1216 The above will currently force a WebRTC Endpoint that receives a 1217 MediaStreamTrack on one RTCPeerConnection and adds it as an outgoing 1218 on any RTCPeerConnection to perform resynchronisation of the stream. 1219 Since the sending party needs to change the CNAME to the one it uses, 1220 this implies it has to use a local system clock as timebase for the 1221 synchronisation. Thus, the relative relation between the timebase of 1222 the incoming stream and the system sending out needs to be defined. 1223 This relation also needs monitoring for clock drift and likely 1224 adjustments of the synchronisation. The sending entity is also 1225 responsible for congestion control for its sent streams. In cases of 1226 packet loss the loss of incoming data also needs to be handled. This 1227 leads to the observation that the method that is least likely to 1228 cause issues or interruptions in the outgoing source packet stream is 1229 a model of full decoding, including repair etc., followed by encoding 1230 of the media again into the outgoing packet stream. Optimisations of 1231 this method are clearly possible and implementation specific. 1233 A WebRTC Endpoint MUST support receiving multiple MediaStreamTracks, 1234 where each of the different MediaStreamTracks (and their sets of 1235 associated packet streams) uses different CNAMEs. However, 1236 MediaStreamTracks that are received with different CNAMEs have no 1237 defined synchronisation. 1239 Note: The motivation for supporting reception of multiple CNAMEs 1240 is to allow for forward compatibility with any future changes that 1241 enable more efficient stream handling when endpoints relay/forward 1242 streams. It also ensures that endpoints can interoperate with 1243 certain types of multi-stream middleboxes or endpoints that are 1244 not WebRTC. 1246 Javascript Session Establishment Protocol [I-D.ietf-rtcweb-jsep] 1247 specifies that the binding between the WebRTC MediaStreams, 1248 MediaStreamTracks and the SSRC is done as specified in "Cross Session 1249 Stream Identification in the Session Description Protocol" 1250 [I-D.ietf-mmusic-msid]. The MSID document [I-D.ietf-mmusic-msid] 1251 also defines, in section 4.1, how to map unknown source packet stream 1252 SSRCs to MediaStreamTracks and MediaStreams. This later is relevant 1253 to handle some cases of legacy interoperability. Commonly the RTP 1254 Payload Type of any incoming packets will reveal if the packet stream 1255 is a source stream or a redundancy or dependent packet stream. The 1256 association to the correct source packet stream depends on the 1257 payload format in use for the packet stream. 1259 Finally this specification puts a requirement on the WebRTC API to 1260 realize a method for determining the CSRC list (Section 4.1) as well 1261 as the Mixer-to-Client audio levels (Section 5.2.3) (when supported) 1262 and the basic requirements for this is further discussed in 1263 Section 12.2.1. 1265 12. RTP Implementation Considerations 1267 The following discussion provides some guidance on the implementation 1268 of the RTP features described in this memo. The focus is on a WebRTC 1269 Endpoint implementation perspective, and while some mention is made 1270 of the behaviour of middleboxes, that is not the focus of this memo. 1272 12.1. Configuration and Use of RTP Sessions 1274 A WebRTC Endpoint will be a simultaneous participant in one or more 1275 RTP sessions. Each RTP session can convey multiple media sources, 1276 and can include media data from multiple endpoints. In the 1277 following, some ways in which WebRTC Endpoints can configure and use 1278 RTP sessions are outlined. 1280 12.1.1. Use of Multiple Media Sources Within an RTP Session 1282 RTP is a group communication protocol, and every RTP session can 1283 potentially contain multiple RTP packet streams. There are several 1284 reasons why this might be desirable: 1286 Multiple media types: Outside of WebRTC, it is common to use one RTP 1287 session for each type of media source (e.g., one RTP session for 1288 audio sources and one for video sources, each sent over different 1289 transport layer flows). However, to reduce the number of UDP 1290 ports used, the default in WebRTC is to send all types of media in 1291 a single RTP session, as described in Section 4.4, using RTP and 1292 RTCP multiplexing (Section 4.5) to further reduce the number of 1293 UDP ports needed. This RTP session then uses only one bi- 1294 directional transport-layer flow, but will contain multiple RTP 1295 packet streams, each containing a different type of media. A 1296 common example might be an endpoint with a camera and microphone 1297 that sends two RTP packet streams, one video and one audio, into a 1298 single RTP session. 1300 Multiple Capture Devices: A WebRTC Endpoint might have multiple 1301 cameras, microphones, or other media capture devices, and so might 1302 want to generate several RTP packet streams of the same media 1303 type. Alternatively, it might want to send media from a single 1304 capture device in several different formats or quality settings at 1305 once. Both can result in a single endpoint sending multiple RTP 1306 packet streams of the same media type into a single RTP session at 1307 the same time. 1309 Associated Repair Data: An endpoint might send a RTP packet stream 1310 that is somehow associated with another stream. For example, it 1311 might send an RTP packet stream that contains FEC or 1312 retransmission data relating to another stream. Some RTP payload 1313 formats send this sort of associated repair data as part of the 1314 source packet stream, while others send it as a separate packet 1315 stream. 1317 Layered or Multiple Description Coding: An endpoint can use a 1318 layered media codec, for example H.264 SVC, or a multiple 1319 description codec, that generates multiple RTP packet streams, 1320 each with a distinct RTP SSRC, within a single RTP session. 1322 RTP Mixers, Translators, and Other Middleboxes: An RTP session, in 1323 the WebRTC context, is a point-to-point association between an 1324 endpoint and some other peer device, where those devices share a 1325 common SSRC space. The peer device might be another WebRTC 1326 Endpoint, or it might be an RTP mixer, translator, or some other 1327 form of media processing middlebox. In the latter cases, the 1328 middlebox might send mixed or relayed RTP streams from several 1329 participants, that the WebRTC Endpoint will need to render. Thus, 1330 even though a WebRTC Endpoint might only be a member of a single 1331 RTP session, the peer device might be extending that RTP session 1332 to incorporate other endpoints. WebRTC is a group communication 1333 environment and endpoints need to be capable of receiving, 1334 decoding, and playing out multiple RTP packet streams at once, 1335 even in a single RTP session. 1337 12.1.2. Use of Multiple RTP Sessions 1339 In addition to sending and receiving multiple RTP packet streams 1340 within a single RTP session, a WebRTC Endpoint might participate in 1341 multiple RTP sessions. There are several reasons why a WebRTC 1342 Endpoint might choose to do this: 1344 To interoperate with legacy devices: The common practice in the non- 1345 WebRTC world is to send different types of media in separate RTP 1346 sessions, for example using one RTP session for audio and another 1347 RTP session, on a separate transport layer flow, for video. All 1348 WebRTC Endpoints need to support the option of sending different 1349 types of media on different RTP sessions, so they can interwork 1350 with such legacy devices. This is discussed further in 1351 Section 4.4. 1353 To provide enhanced quality of service: Some network-based quality 1354 of service mechanisms operate on the granularity of transport 1355 layer flows. If it is desired to use these mechanisms to provide 1356 differentiated quality of service for some RTP packet streams, 1357 then those RTP packet streams need to be sent in a separate RTP 1358 session using a different transport-layer flow, and with 1359 appropriate quality of service marking. This is discussed further 1360 in Section 12.1.3. 1362 To separate media with different purposes: An endpoint might want to 1363 send RTP packet streams that have different purposes on different 1364 RTP sessions, to make it easy for the peer device to distinguish 1365 them. For example, some centralised multiparty conferencing 1366 systems display the active speaker in high resolution, but show 1367 low resolution "thumbnails" of other participants. Such systems 1368 might configure the endpoints to send simulcast high- and low- 1369 resolution versions of their video using separate RTP sessions, to 1370 simplify the operation of the RTP middlebox. In the WebRTC 1371 context this is currently possible by establishing multiple WebRTC 1372 MediaStreamTracks that have the same media source in one (or more) 1373 RTCPeerConnection. Each MediaStreamTrack is then configured to 1374 deliver a particular media quality and thus media bit-rate, and 1375 will produce an independently encoded version with the codec 1376 parameters agreed specifically in the context of that 1377 RTCPeerConnection. The RTP middlebox can distinguish packets 1378 corresponding to the low- and high-resolution streams by 1379 inspecting their SSRC, RTP payload type, or some other information 1380 contained in RTP payload, RTP header extension or RTCP packets, 1381 but it can be easier to distinguish the RTP packet streams if they 1382 arrive on separate RTP sessions on separate transport-layer flows. 1384 To directly connect with multiple peers: A multi-party conference 1385 does not need to use an RTP middlebox. Rather, a multi-unicast 1386 mesh can be created, comprising several distinct RTP sessions, 1387 with each participant sending RTP traffic over a separate RTP 1388 session (that is, using an independent RTCPeerConnection object) 1389 to every other participant, as shown in Figure 1. This topology 1390 has the benefit of not requiring an RTP middlebox node that is 1391 trusted to access and manipulate the media data. The downside is 1392 that it increases the used bandwidth at each sender by requiring 1393 one copy of the RTP packet streams for each participant that are 1394 part of the same session beyond the sender itself. 1396 +---+ +---+ 1397 | A |<--->| B | 1398 +---+ +---+ 1399 ^ ^ 1400 \ / 1401 \ / 1402 v v 1403 +---+ 1404 | C | 1405 +---+ 1407 Figure 1: Multi-unicast using several RTP sessions 1409 The multi-unicast topology could also be implemented as a single 1410 RTP session, spanning multiple peer-to-peer transport layer 1411 connections, or as several pairwise RTP sessions, one between each 1412 pair of peers. To maintain a coherent mapping of the relationship 1413 between RTP sessions and RTCPeerConnection objects it is recommend 1414 that this is implemented as several individual RTP sessions. The 1415 only downside is that endpoint A will not learn of the quality of 1416 any transmission happening between B and C, since it will not see 1417 RTCP reports for the RTP session between B and C, whereas it would 1418 if all three participants were part of a single RTP session. 1419 Experience with the Mbone tools (experimental RTP-based multicast 1420 conferencing tools from the late 1990s) has showed that RTCP 1421 reception quality reports for third parties can be presented to 1422 users in a way that helps them understand asymmetric network 1423 problems, and the approach of using separate RTP sessions prevents 1424 this. However, an advantage of using separate RTP sessions is 1425 that it enables using different media bit-rates and RTP session 1426 configurations between the different peers, thus not forcing B to 1427 endure the same quality reductions if there are limitations in the 1428 transport from A to C as C will. It is believed that these 1429 advantages outweigh the limitations in debugging power. 1431 To indirectly connect with multiple peers: A common scenario in 1432 multi-party conferencing is to create indirect connections to 1433 multiple peers, using an RTP mixer, translator, or some other type 1434 of RTP middlebox. Figure 2 outlines a simple topology that might 1435 be used in a four-person centralised conference. The middlebox 1436 acts to optimise the transmission of RTP packet streams from 1437 certain perspectives, either by only sending some of the received 1438 RTP packet stream to any given receiver, or by providing a 1439 combined RTP packet stream out of a set of contributing streams. 1441 +---+ +-------------+ +---+ 1442 | A |<---->| |<---->| B | 1443 +---+ | RTP mixer, | +---+ 1444 | translator, | 1445 | or other | 1446 +---+ | middlebox | +---+ 1447 | C |<---->| |<---->| D | 1448 +---+ +-------------+ +---+ 1450 Figure 2: RTP mixer with only unicast paths 1452 There are various methods of implementation for the middlebox. If 1453 implemented as a standard RTP mixer or translator, a single RTP 1454 session will extend across the middlebox and encompass all the 1455 endpoints in one multi-party session. Other types of middlebox 1456 might use separate RTP sessions between each endpoint and the 1457 middlebox. A common aspect is that these RTP middleboxes can use 1458 a number of tools to control the media encoding provided by a 1459 WebRTC Endpoint. This includes functions like requesting the 1460 breaking of the encoding chain and have the encoder produce a so 1461 called Intra frame. Another is limiting the bit-rate of a given 1462 stream to better suit the mixer view of the multiple down-streams. 1463 Others are controlling the most suitable frame-rate, picture 1464 resolution, the trade-off between frame-rate and spatial quality. 1465 The middlebox has the responsibility to correctly perform 1466 congestion control, source identification, manage synchronisation 1467 while providing the application with suitable media optimisations. 1468 The middlebox also has to be a trusted node when it comes to 1469 security, since it manipulates either the RTP header or the media 1470 itself (or both) received from one endpoint, before sending it on 1471 towards the endpoint(s), thus they need to be able to decrypt and 1472 then re-encrypt the RTP packet stream before sending it out. 1474 RTP Mixers can create a situation where an endpoint experiences a 1475 situation in-between a session with only two endpoints and 1476 multiple RTP sessions. Mixers are expected to not forward RTCP 1477 reports regarding RTP packet streams across themselves. This is 1478 due to the difference in the RTP packet streams provided to the 1479 different endpoints. The original media source lacks information 1480 about a mixer's manipulations prior to sending it the different 1481 receivers. This scenario also results in that an endpoint's 1482 feedback or requests go to the mixer. When the mixer can't act on 1483 this by itself, it is forced to go to the original media source to 1484 fulfil the receivers request. This will not necessarily be 1485 explicitly visible to any RTP and RTCP traffic, but the 1486 interactions and the time to complete them will indicate such 1487 dependencies. 1489 Providing source authentication in multi-party scenarios is a 1490 challenge. In the mixer-based topologies, endpoints source 1491 authentication is based on, firstly, verifying that media comes 1492 from the mixer by cryptographic verification and, secondly, trust 1493 in the mixer to correctly identify any source towards the 1494 endpoint. In RTP sessions where multiple endpoints are directly 1495 visible to an endpoint, all endpoints will have knowledge about 1496 each others' master keys, and can thus inject packets claimed to 1497 come from another endpoint in the session. Any node performing 1498 relay can perform non-cryptographic mitigation by preventing 1499 forwarding of packets that have SSRC fields that came from other 1500 endpoints before. For cryptographic verification of the source, 1501 SRTP would require additional security mechanisms, for example 1502 TESLA for SRTP [RFC4383], that are not part of the base WebRTC 1503 standards. 1505 To forward media between multiple peers: It is sometimes desirable 1506 for an endpoint that receives an RTP packet stream to be able to 1507 forward that RTP packet stream to a third party. The are some 1508 obvious security and privacy implications in supporting this, but 1509 also potential uses. This is supported in the W3C API by taking 1510 the received and decoded media and using it as media source that 1511 is re-encoding and transmitted as a new stream. 1513 At the RTP layer, media forwarding acts as a back-to-back RTP 1514 receiver and RTP sender. The receiving side terminates the RTP 1515 session and decodes the media, while the sender side re-encodes 1516 and transmits the media using an entirely separate RTP session. 1517 The original sender will only see a single receiver of the media, 1518 and will not be able to tell that forwarding is happening based on 1519 RTP-layer information since the RTP session that is used to send 1520 the forwarded media is not connected to the RTP session on which 1521 the media was received by the node doing the forwarding. 1523 The endpoint that is performing the forwarding is responsible for 1524 producing an RTP packet stream suitable for onwards transmission. 1525 The outgoing RTP session that is used to send the forwarded media 1526 is entirely separate to the RTP session on which the media was 1527 received. This will require media transcoding for congestion 1528 control purpose to produce a suitable bit-rate for the outgoing 1529 RTP session, reducing media quality and forcing the forwarding 1530 endpoint to spend the resource on the transcoding. The media 1531 transcoding does result in a separation of the two different legs 1532 removing almost all dependencies, and allowing the forwarding 1533 endpoint to optimise its media transcoding operation. The cost is 1534 greatly increased computational complexity on the forwarding node. 1535 Receivers of the forwarded stream will see the forwarding device 1536 as the sender of the stream, and will not be able to tell from the 1537 RTP layer that they are receiving a forwarded stream rather than 1538 an entirely new RTP packet stream generated by the forwarding 1539 device. 1541 12.1.3. Differentiated Treatment of RTP Streams 1543 There are use cases for differentiated treatment of RTP packet 1544 streams. Such differentiation can happen at several places in the 1545 system. First of all is the prioritization within the endpoint 1546 sending the media, which controls, both which RTP packet streams that 1547 will be sent, and their allocation of bit-rate out of the current 1548 available aggregate as determined by the congestion control. 1550 It is expected that the WebRTC API [W3C.WD-webrtc-20130910] will 1551 allow the application to indicate relative priorities for different 1552 MediaStreamTracks. These priorities can then be used to influence 1553 the local RTP processing, especially when it comes to congestion 1554 control response in how to divide the available bandwidth between the 1555 RTP packet streams. Any changes in relative priority will also need 1556 to be considered for RTP packet streams that are associated with the 1557 main RTP packet streams, such as redundant streams for RTP 1558 retransmission and FEC. The importance of such redundant RTP packet 1559 streams is dependent on the media type and codec used, in regards to 1560 how robust that codec is to packet loss. However, a default policy 1561 might to be to use the same priority for redundant RTP packet stream 1562 as for the source RTP packet stream. 1564 Secondly, the network can prioritize transport-layer flows and sub- 1565 flows, including RTP packet streams. Typically, differential 1566 treatment includes two steps, the first being identifying whether an 1567 IP packet belongs to a class that has to be treated differently, the 1568 second consisting of the actual mechanism to prioritize packets. 1569 Three common methods for classifying IP packets are: 1571 DiffServ: The endpoint marks a packet with a DiffServ code point to 1572 indicate to the network that the packet belongs to a particular 1573 class. 1575 Flow based: Packets that need to be given a particular treatment are 1576 identified using a combination of IP and port address. 1578 Deep Packet Inspection: A network classifier (DPI) inspects the 1579 packet and tries to determine if the packet represents a 1580 particular application and type that is to be prioritized. 1582 Flow-based differentiation will provide the same treatment to all 1583 packets within a transport-layer flow, i.e., relative prioritization 1584 is not possible. Moreover, if the resources are limited it might not 1585 be possible to provide differential treatment compared to best-effort 1586 for all the RTP packet streams used in a WebRTC session. The use of 1587 flow-based differentiation needs to be coordinated between the WebRTC 1588 system and the network(s). The WebRTC endpoint needs to know that 1589 flow-based differentiation might be used to provide the separation of 1590 the RTP packet streams onto different UDP flows to enable a more 1591 granular usage of flow based differentiation. The used flows, their 1592 5-tuples and prioritization will need to be communicated to the 1593 network so that it can identify the flows correctly to enable 1594 prioritization. No specific protocol support for this is specified. 1596 DiffServ assumes that either the endpoint or a classifier can mark 1597 the packets with an appropriate DSCP so that the packets are treated 1598 according to that marking. If the endpoint is to mark the traffic 1599 two requirements arise in the WebRTC context: 1) The WebRTC Endpoint 1600 has to know which DSCP to use and that it can use them on some set of 1601 RTP packet streams. 2) The information needs to be propagated to the 1602 operating system when transmitting the packet. Details of this 1603 process are outside the scope of this memo and are further discussed 1604 in "DSCP and other packet markings for RTCWeb QoS" 1605 [I-D.ietf-tsvwg-rtcweb-qos]. 1607 Deep Packet Inspectors will, despite the SRTP media encryption, still 1608 be fairly capable at classifying the RTP streams. The reason is that 1609 SRTP leaves the first 12 bytes of the RTP header unencrypted. This 1610 enables easy RTP stream identification using the SSRC and provides 1611 the classifier with useful information that can be correlated to 1612 determine for example the stream's media type. Using packet sizes, 1613 reception times, packet inter-spacing, RTP timestamp increments and 1614 sequence numbers, fairly reliable classifications are achieved. 1616 For packet based marking schemes it might be possible to mark 1617 individual RTP packets differently based on the relative priority of 1618 the RTP payload. For example video codecs that have I, P, and B 1619 pictures could prioritise any payloads carrying only B frames less, 1620 as these are less damaging to loose. However, depending on the QoS 1621 mechanism and what markings that are applied, this can result in not 1622 only different packet drop probabilities but also packet reordering, 1623 see [I-D.ietf-tsvwg-rtcweb-qos] and [I-D.ietf-dart-dscp-rtp] for 1624 further discussion. As a default policy all RTP packets related to a 1625 RTP packet stream ought to be provided with the same prioritization; 1626 per-packet prioritization is outside the scope of this memo, but 1627 might be specified elsewhere in future. 1629 It is also important to consider how RTCP packets associated with a 1630 particular RTP packet stream need to be marked. RTCP compound 1631 packets with Sender Reports (SR), ought to be marked with the same 1632 priority as the RTP packet stream itself, so the RTCP-based round- 1633 trip time (RTT) measurements are done using the same transport-layer 1634 flow priority as the RTP packet stream experiences. RTCP compound 1635 packets containing RR packet ought to be sent with the priority used 1636 by the majority of the RTP packet streams reported on. RTCP packets 1637 containing time-critical feedback packets can use higher priority to 1638 improve the timeliness and likelihood of delivery of such feedback. 1640 12.2. Media Source, RTP Streams, and Participant Identification 1642 12.2.1. Media Source Identification 1644 Each RTP packet stream is identified by a unique synchronisation 1645 source (SSRC) identifier. The SSRC identifier is carried in each of 1646 the RTP packets comprising a RTP packet stream, and is also used to 1647 identify that stream in the corresponding RTCP reports. The SSRC is 1648 chosen as discussed in Section 4.8. The first stage in 1649 demultiplexing RTP and RTCP packets received on a single transport 1650 layer flow at a WebRTC Endpoint is to separate the RTP packet streams 1651 based on their SSRC value; once that is done, additional 1652 demultiplexing steps can determine how and where to render the media. 1654 RTP allows a mixer, or other RTP-layer middlebox, to combine encoded 1655 streams from multiple media sources to form a new encoded stream from 1656 a new media source (the mixer). The RTP packets in that new RTP 1657 packet stream can include a Contributing Source (CSRC) list, 1658 indicating which original SSRCs contributed to the combined source 1659 stream. As described in Section 4.1, implementations need to support 1660 reception of RTP data packets containing a CSRC list and RTCP packets 1661 that relate to sources present in the CSRC list. The CSRC list can 1662 change on a packet-by-packet basis, depending on the mixing operation 1663 being performed. Knowledge of what media sources contributed to a 1664 particular RTP packet can be important if the user interface 1665 indicates which participants are active in the session. Changes in 1666 the CSRC list included in packets needs to be exposed to the WebRTC 1667 application using some API, if the application is to be able to track 1668 changes in session participation. It is desirable to map CSRC values 1669 back into WebRTC MediaStream identities as they cross this API, to 1670 avoid exposing the SSRC/CSRC name space to WebRTC applications. 1672 If the mixer-to-client audio level extension [RFC6465] is being used 1673 in the session (see Section 5.2.3), the information in the CSRC list 1674 is augmented by audio level information for each contributing source. 1675 It is desirable to expose this information to the WebRTC application 1676 using some API, after mapping the CSRC values to WebRTC MediaStream 1677 identities, so it can be exposed in the user interface. 1679 12.2.2. SSRC Collision Detection 1681 The RTP standard requires RTP implementations to have support for 1682 detecting and handling SSRC collisions, i.e., resolve the conflict 1683 when two different endpoints use the same SSRC value (see section 8.2 1684 of [RFC3550]). This requirement also applies to WebRTC Endpoints. 1685 There are several scenarios where SSRC collisions can occur: 1687 o In a point-to-point session where each SSRC is associated with 1688 either of the two endpoints and where the main media carrying SSRC 1689 identifier will be announced in the signalling channel, a 1690 collision is less likely to occur due to the information about 1691 used SSRCs. If SDP is used, this information is provided by 1692 Source-Specific SDP Attributes [RFC5576]. Still, collisions can 1693 occur if both endpoints start using a new SSRC identifier prior to 1694 having signalled it to the peer and received acknowledgement on 1695 the signalling message. The Source-Specific SDP Attributes 1696 [RFC5576] contains a mechanism to signal how the endpoint resolved 1697 the SSRC collision. 1699 o SSRC values that have not been signalled could also appear in an 1700 RTP session. This is more likely than it appears, since some RTP 1701 functions use extra SSRCs to provide their functionality. For 1702 example, retransmission data might be transmitted using a separate 1703 RTP packet stream that requires its own SSRC, separate to the SSRC 1704 of the source RTP packet stream [RFC4588]. In those cases, an 1705 endpoint can create a new SSRC that strictly doesn't need to be 1706 announced over the signalling channel to function correctly on 1707 both RTP and RTCPeerConnection level. 1709 o Multiple endpoints in a multiparty conference can create new 1710 sources and signal those towards the RTP middlebox. In cases 1711 where the SSRC/CSRC are propagated between the different endpoints 1712 from the RTP middlebox collisions can occur. 1714 o An RTP middlebox could connect an endpoint's RTCPeerConnection to 1715 another RTCPeerConnection from the same endpoint, thus forming a 1716 loop where the endpoint will receive its own traffic. While it is 1717 clearly considered a bug, it is important that the endpoint is 1718 able to recognise and handle the case when it occurs. This case 1719 becomes even more problematic when media mixers, and so on, are 1720 involved, where the stream received is a different stream but 1721 still contains this client's input. 1723 These SSRC/CSRC collisions can only be handled on RTP level as long 1724 as the same RTP session is extended across multiple 1725 RTCPeerConnections by a RTP middlebox. To resolve the more generic 1726 case where multiple RTCPeerConnections are interconnected, 1727 identification of the media source(s) part of a MediaStreamTrack 1728 being propagated across multiple interconnected RTCPeerConnection 1729 needs to be preserved across these interconnections. 1731 12.2.3. Media Synchronisation Context 1733 When an endpoint sends media from more than one media source, it 1734 needs to consider if (and which of) these media sources are to be 1735 synchronized. In RTP/RTCP, synchronisation is provided by having a 1736 set of RTP packet streams be indicated as coming from the same 1737 synchronisation context and logical endpoint by using the same RTCP 1738 CNAME identifier. 1740 The next provision is that the internal clocks of all media sources, 1741 i.e., what drives the RTP timestamp, can be correlated to a system 1742 clock that is provided in RTCP Sender Reports encoded in an NTP 1743 format. By correlating all RTP timestamps to a common system clock 1744 for all sources, the timing relation of the different RTP packet 1745 streams, also across multiple RTP sessions can be derived at the 1746 receiver and, if desired, the streams can be synchronized. The 1747 requirement is for the media sender to provide the correlation 1748 information; it is up to the receiver to use it or not. 1750 13. Security Considerations 1752 The overall security architecture for WebRTC is described in 1753 [I-D.ietf-rtcweb-security-arch], and security considerations for the 1754 WebRTC framework are described in [I-D.ietf-rtcweb-security]. These 1755 considerations also apply to this memo. 1757 The security considerations of the RTP specification, the RTP/SAVPF 1758 profile, and the various RTP/RTCP extensions and RTP payload formats 1759 that form the complete protocol suite described in this memo apply. 1760 It is not believed there are any new security considerations 1761 resulting from the combination of these various protocol extensions. 1763 The Extended Secure RTP Profile for Real-time Transport Control 1764 Protocol (RTCP)-Based Feedback [RFC5124] (RTP/SAVPF) provides 1765 handling of fundamental issues by offering confidentiality, integrity 1766 and partial source authentication. A mandatory to implement and use 1767 media security solution is created by combining this secured RTP 1768 profile and DTLS-SRTP keying [RFC5764] as defined by Section 5.5 of 1769 [I-D.ietf-rtcweb-security-arch]. 1771 RTCP packets convey a Canonical Name (CNAME) identifier that is used 1772 to associate RTP packet streams that need to be synchronised across 1773 related RTP sessions. Inappropriate choice of CNAME values can be a 1774 privacy concern, since long-term persistent CNAME identifiers can be 1775 used to track users across multiple WebRTC calls. Section 4.9 of 1776 this memo mandates generation of short-term persistent RTCP CNAMES, 1777 as specified in RFC7022, resulting in untraceable CNAME values that 1778 alleviate this risk. 1780 Some potential denial of service attacks exist if the RTCP reporting 1781 interval is configured to an inappropriate value. This could be done 1782 by configuring the RTCP bandwidth fraction to an excessively large or 1783 small value using the SDP "b=RR:" or "b=RS:" lines [RFC3556], or some 1784 similar mechanism, or by choosing an excessively large or small value 1785 for the RTP/AVPF minimal receiver report interval (if using SDP, this 1786 is the "a=rtcp-fb:... trr-int" parameter) [RFC4585]. The risks are 1787 as follows: 1789 1. the RTCP bandwidth could be configured to make the regular 1790 reporting interval so large that effective congestion control 1791 cannot be maintained, potentially leading to denial of service 1792 due to congestion caused by the media traffic; 1794 2. the RTCP interval could be configured to a very small value, 1795 causing endpoints to generate high rate RTCP traffic, potentially 1796 leading to denial of service due to the non-congestion controlled 1797 RTCP traffic; and 1799 3. RTCP parameters could be configured differently for each 1800 endpoint, with some of the endpoints using a large reporting 1801 interval and some using a smaller interval, leading to denial of 1802 service due to premature participant timeouts due to mismatched 1803 timeout periods which are based on the reporting interval (this 1804 is a particular concern if endpoints use a small but non-zero 1805 value for the RTP/AVPF minimal receiver report interval (trr-int) 1806 [RFC4585], as discussed in Section 6.1 of 1807 [I-D.ietf-avtcore-rtp-multi-stream]). 1809 Premature participant timeout can be avoided by using the fixed (non- 1810 reduced) minimum interval when calculating the participant timeout 1811 (see Section 4.1 of this memo and Section 6.1 of 1812 [I-D.ietf-avtcore-rtp-multi-stream]). To address the other concerns, 1813 endpoints SHOULD ignore parameters that configure the RTCP reporting 1814 interval to be significantly longer than the default five second 1815 interval specified in [RFC3550] (unless the media data rate is so low 1816 that the longer reporting interval roughly corresponds to 5% of the 1817 media data rate), or that configure the RTCP reporting interval small 1818 enough that the RTCP bandwidth would exceed the media bandwidth. 1820 The guidelines in [RFC6562] apply when using variable bit rate (VBR) 1821 audio codecs such as Opus (see Section 4.3 for discussion of mandated 1822 audio codecs). The guidelines in [RFC6562] also apply, but are of 1823 lesser importance, when using the client-to-mixer audio level header 1824 extensions (Section 5.2.2) or the mixer-to-client audio level header 1825 extensions (Section 5.2.3). The use of the encryption of the header 1826 extensions are RECOMMENDED, unless there are known reasons, like RTP 1827 middleboxes performing voice activity based source selection or third 1828 party monitoring that will greatly benefit from the information, and 1829 this has been expressed using API or signalling. If further evidence 1830 are produced to show that information leakage is significant from 1831 audio level indications, then use of encryption needs to be mandated 1832 at that time. 1834 In multi-party communication scenarios using RTP Middleboxes, a lot 1835 of trust is placed on these middleboxes to preserve the sessions 1836 security. The middlebox needs to maintain the confidentiality, 1837 integrity and perform source authentication. As discussed in 1838 Section 12.1.1 the middlebox can perform checks that prevents any 1839 endpoint participating in a conference to impersonate another. Some 1840 additional security considerations regarding multi-party topologies 1841 can be found in [I-D.ietf-avtcore-rtp-topologies-update]. 1843 14. IANA Considerations 1845 This memo makes no request of IANA. 1847 Note to RFC Editor: this section is to be removed on publication as 1848 an RFC. 1850 15. Acknowledgements 1852 The authors would like to thank Bernard Aboba, Harald Alvestrand, 1853 Cary Bran, Ben Campbell, Alissa Cooper, Spencer Dawkins, Charles 1854 Eckel, Alex Eleftheriadis, Christian Groves, Chris Inacio, Cullen 1855 Jennings, Olle Johansson, Suhas Nandakumar, Dan Romascanu, Jim 1856 Spring, Martin Thomson, and the other members of the IETF RTCWEB 1857 working group for their valuable feedback. 1859 16. References 1861 16.1. Normative References 1863 [I-D.ietf-avtcore-multi-media-rtp-session] 1864 Westerlund, M., Perkins, C., and J. Lennox, "Sending 1865 Multiple Types of Media in a Single RTP Session", draft- 1866 ietf-avtcore-multi-media-rtp-session-13 (work in 1867 progress), December 2015. 1869 [I-D.ietf-avtcore-rtp-circuit-breakers] 1870 Perkins, C. and V. Varun, "Multimedia Congestion Control: 1871 Circuit Breakers for Unicast RTP Sessions", draft-ietf- 1872 avtcore-rtp-circuit-breakers-13 (work in progress), 1873 February 2016. 1875 [I-D.ietf-avtcore-rtp-multi-stream] 1876 Lennox, J., Westerlund, M., Wu, Q., and C. Perkins, 1877 "Sending Multiple RTP Streams in a Single RTP Session", 1878 draft-ietf-avtcore-rtp-multi-stream-11 (work in progress), 1879 December 2015. 1881 [I-D.ietf-avtcore-rtp-multi-stream-optimisation] 1882 Lennox, J., Westerlund, M., Wu, Q., and C. Perkins, 1883 "Sending Multiple RTP Streams in a Single RTP Session: 1884 Grouping RTCP Reception Statistics and Other Feedback", 1885 draft-ietf-avtcore-rtp-multi-stream-optimisation-12 (work 1886 in progress), March 2016. 1888 [I-D.ietf-avtcore-rtp-topologies-update] 1889 Westerlund, M. and S. Wenger, "RTP Topologies", draft- 1890 ietf-avtcore-rtp-topologies-update-10 (work in progress), 1891 July 2015. 1893 [I-D.ietf-mmusic-mux-exclusive] 1894 Holmberg, C., "Indicating Exclusive Support of RTP/RTCP 1895 Multiplexing using SDP", draft-ietf-mmusic-mux- 1896 exclusive-03 (work in progress), February 2016. 1898 [I-D.ietf-mmusic-sdp-bundle-negotiation] 1899 Holmberg, C., Alvestrand, H., and C. Jennings, 1900 "Negotiating Media Multiplexing Using the Session 1901 Description Protocol (SDP)", draft-ietf-mmusic-sdp-bundle- 1902 negotiation-27 (work in progress), February 2016. 1904 [I-D.ietf-rtcweb-audio] 1905 Valin, J. and C. Bran, "WebRTC Audio Codec and Processing 1906 Requirements", draft-ietf-rtcweb-audio-10 (work in 1907 progress), February 2016. 1909 [I-D.ietf-rtcweb-fec] 1910 Uberti, J., "WebRTC Forward Error Correction 1911 Requirements", draft-ietf-rtcweb-fec-02 (work in 1912 progress), October 2015. 1914 [I-D.ietf-rtcweb-overview] 1915 Alvestrand, H., "Overview: Real Time Protocols for 1916 Browser-based Applications", draft-ietf-rtcweb-overview-15 1917 (work in progress), January 2016. 1919 [I-D.ietf-rtcweb-security] 1920 Rescorla, E., "Security Considerations for WebRTC", draft- 1921 ietf-rtcweb-security-08 (work in progress), February 2015. 1923 [I-D.ietf-rtcweb-security-arch] 1924 Rescorla, E., "WebRTC Security Architecture", draft-ietf- 1925 rtcweb-security-arch-11 (work in progress), March 2015. 1927 [I-D.ietf-rtcweb-video] 1928 Roach, A., "WebRTC Video Processing and Codec 1929 Requirements", draft-ietf-rtcweb-video-06 (work in 1930 progress), June 2015. 1932 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1933 Requirement Levels", BCP 14, RFC 2119, 1934 DOI 10.17487/RFC2119, March 1997, 1935 . 1937 [RFC2736] Handley, M. and C. Perkins, "Guidelines for Writers of RTP 1938 Payload Format Specifications", BCP 36, RFC 2736, 1939 DOI 10.17487/RFC2736, December 1999, 1940 . 1942 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 1943 Jacobson, "RTP: A Transport Protocol for Real-Time 1944 Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, 1945 July 2003, . 1947 [RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and 1948 Video Conferences with Minimal Control", STD 65, RFC 3551, 1949 DOI 10.17487/RFC3551, July 2003, 1950 . 1952 [RFC3556] Casner, S., "Session Description Protocol (SDP) Bandwidth 1953 Modifiers for RTP Control Protocol (RTCP) Bandwidth", 1954 RFC 3556, DOI 10.17487/RFC3556, July 2003, 1955 . 1957 [RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K. 1958 Norrman, "The Secure Real-time Transport Protocol (SRTP)", 1959 RFC 3711, DOI 10.17487/RFC3711, March 2004, 1960 . 1962 [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session 1963 Description Protocol", RFC 4566, DOI 10.17487/RFC4566, 1964 July 2006, . 1966 [RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 1967 "Extended RTP Profile for Real-time Transport Control 1968 Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, 1969 DOI 10.17487/RFC4585, July 2006, 1970 . 1972 [RFC4588] Rey, J., Leon, D., Miyazaki, A., Varsa, V., and R. 1973 Hakenberg, "RTP Retransmission Payload Format", RFC 4588, 1974 DOI 10.17487/RFC4588, July 2006, 1975 . 1977 [RFC4961] Wing, D., "Symmetric RTP / RTP Control Protocol (RTCP)", 1978 BCP 131, RFC 4961, DOI 10.17487/RFC4961, July 2007, 1979 . 1981 [RFC5104] Wenger, S., Chandra, U., Westerlund, M., and B. Burman, 1982 "Codec Control Messages in the RTP Audio-Visual Profile 1983 with Feedback (AVPF)", RFC 5104, DOI 10.17487/RFC5104, 1984 February 2008, . 1986 [RFC5124] Ott, J. and E. Carrara, "Extended Secure RTP Profile for 1987 Real-time Transport Control Protocol (RTCP)-Based Feedback 1988 (RTP/SAVPF)", RFC 5124, DOI 10.17487/RFC5124, February 1989 2008, . 1991 [RFC5285] Singer, D. and H. Desineni, "A General Mechanism for RTP 1992 Header Extensions", RFC 5285, DOI 10.17487/RFC5285, July 1993 2008, . 1995 [RFC5506] Johansson, I. and M. Westerlund, "Support for Reduced-Size 1996 Real-Time Transport Control Protocol (RTCP): Opportunities 1997 and Consequences", RFC 5506, DOI 10.17487/RFC5506, April 1998 2009, . 2000 [RFC5761] Perkins, C. and M. Westerlund, "Multiplexing RTP Data and 2001 Control Packets on a Single Port", RFC 5761, 2002 DOI 10.17487/RFC5761, April 2010, 2003 . 2005 [RFC5764] McGrew, D. and E. Rescorla, "Datagram Transport Layer 2006 Security (DTLS) Extension to Establish Keys for the Secure 2007 Real-time Transport Protocol (SRTP)", RFC 5764, 2008 DOI 10.17487/RFC5764, May 2010, 2009 . 2011 [RFC6051] Perkins, C. and T. Schierl, "Rapid Synchronisation of RTP 2012 Flows", RFC 6051, DOI 10.17487/RFC6051, November 2010, 2013 . 2015 [RFC6464] Lennox, J., Ed., Ivov, E., and E. Marocco, "A Real-time 2016 Transport Protocol (RTP) Header Extension for Client-to- 2017 Mixer Audio Level Indication", RFC 6464, 2018 DOI 10.17487/RFC6464, December 2011, 2019 . 2021 [RFC6465] Ivov, E., Ed., Marocco, E., Ed., and J. Lennox, "A Real- 2022 time Transport Protocol (RTP) Header Extension for Mixer- 2023 to-Client Audio Level Indication", RFC 6465, 2024 DOI 10.17487/RFC6465, December 2011, 2025 . 2027 [RFC6562] Perkins, C. and JM. Valin, "Guidelines for the Use of 2028 Variable Bit Rate Audio with Secure RTP", RFC 6562, 2029 DOI 10.17487/RFC6562, March 2012, 2030 . 2032 [RFC6904] Lennox, J., "Encryption of Header Extensions in the Secure 2033 Real-time Transport Protocol (SRTP)", RFC 6904, 2034 DOI 10.17487/RFC6904, April 2013, 2035 . 2037 [RFC7007] Terriberry, T., "Update to Remove DVI4 from the 2038 Recommended Codecs for the RTP Profile for Audio and Video 2039 Conferences with Minimal Control (RTP/AVP)", RFC 7007, 2040 DOI 10.17487/RFC7007, August 2013, 2041 . 2043 [RFC7022] Begen, A., Perkins, C., Wing, D., and E. Rescorla, 2044 "Guidelines for Choosing RTP Control Protocol (RTCP) 2045 Canonical Names (CNAMEs)", RFC 7022, DOI 10.17487/RFC7022, 2046 September 2013, . 2048 [RFC7160] Petit-Huguenin, M. and G. Zorn, Ed., "Support for Multiple 2049 Clock Rates in an RTP Session", RFC 7160, 2050 DOI 10.17487/RFC7160, April 2014, 2051 . 2053 [RFC7164] Gross, K. and R. Brandenburg, "RTP and Leap Seconds", 2054 RFC 7164, DOI 10.17487/RFC7164, March 2014, 2055 . 2057 [W3C.WD-mediacapture-streams-20130903] 2058 Burnett, D., Bergkvist, A., Jennings, C., and A. 2059 Narayanan, "Media Capture and Streams", World Wide Web 2060 Consortium WD WD-mediacapture-streams-20130903, September 2061 2013, . 2064 [W3C.WD-webrtc-20130910] 2065 Bergkvist, A., Burnett, D., Jennings, C., and A. 2066 Narayanan, "WebRTC 1.0: Real-time Communication Between 2067 Browsers", World Wide Web Consortium WD WD-webrtc- 2068 20130910, September 2013, 2069 . 2071 16.2. Informative References 2073 [I-D.ietf-avtcore-multiplex-guidelines] 2074 Westerlund, M., Perkins, C., and H. Alvestrand, 2075 "Guidelines for using the Multiplexing Features of RTP to 2076 Support Multiple Media Streams", draft-ietf-avtcore- 2077 multiplex-guidelines-03 (work in progress), October 2014. 2079 [I-D.ietf-avtext-rtp-grouping-taxonomy] 2080 Lennox, J., Gross, K., Nandakumar, S., Salgueiro, G., and 2081 B. Burman, "A Taxonomy of Semantics and Mechanisms for 2082 Real-Time Transport Protocol (RTP) Sources", draft-ietf- 2083 avtext-rtp-grouping-taxonomy-08 (work in progress), July 2084 2015. 2086 [I-D.ietf-dart-dscp-rtp] 2087 Black, D. and P. Jones, "Differentiated Services 2088 (DiffServ) and Real-time Communication", draft-ietf-dart- 2089 dscp-rtp-10 (work in progress), November 2014. 2091 [I-D.ietf-mmusic-msid] 2092 Alvestrand, H., "WebRTC MediaStream Identification in the 2093 Session Description Protocol", draft-ietf-mmusic-msid-11 2094 (work in progress), October 2015. 2096 [I-D.ietf-payload-rtp-howto] 2097 Westerlund, M., "How to Write an RTP Payload Format", 2098 draft-ietf-payload-rtp-howto-14 (work in progress), May 2099 2015. 2101 [I-D.ietf-rmcat-cc-requirements] 2102 Jesup, R. and Z. Sarker, "Congestion Control Requirements 2103 for Interactive Real-Time Media", draft-ietf-rmcat-cc- 2104 requirements-09 (work in progress), December 2014. 2106 [I-D.ietf-rtcweb-jsep] 2107 Uberti, J., Jennings, C., and E. Rescorla, "Javascript 2108 Session Establishment Protocol", draft-ietf-rtcweb-jsep-13 2109 (work in progress), March 2016. 2111 [I-D.ietf-tsvwg-rtcweb-qos] 2112 Jones, P., Dhesikan, S., Jennings, C., and D. Druta, "DSCP 2113 and other packet markings for WebRTC QoS", draft-ietf- 2114 tsvwg-rtcweb-qos-14 (work in progress), March 2016. 2116 [RFC3611] Friedman, T., Ed., Caceres, R., Ed., and A. Clark, Ed., 2117 "RTP Control Protocol Extended Reports (RTCP XR)", 2118 RFC 3611, DOI 10.17487/RFC3611, November 2003, 2119 . 2121 [RFC4383] Baugher, M. and E. Carrara, "The Use of Timed Efficient 2122 Stream Loss-Tolerant Authentication (TESLA) in the Secure 2123 Real-time Transport Protocol (SRTP)", RFC 4383, 2124 DOI 10.17487/RFC4383, February 2006, 2125 . 2127 [RFC5245] Rosenberg, J., "Interactive Connectivity Establishment 2128 (ICE): A Protocol for Network Address Translator (NAT) 2129 Traversal for Offer/Answer Protocols", RFC 5245, 2130 DOI 10.17487/RFC5245, April 2010, 2131 . 2133 [RFC5576] Lennox, J., Ott, J., and T. Schierl, "Source-Specific 2134 Media Attributes in the Session Description Protocol 2135 (SDP)", RFC 5576, DOI 10.17487/RFC5576, June 2009, 2136 . 2138 [RFC5968] Ott, J. and C. Perkins, "Guidelines for Extending the RTP 2139 Control Protocol (RTCP)", RFC 5968, DOI 10.17487/RFC5968, 2140 September 2010, . 2142 [RFC6263] Marjou, X. and A. Sollaud, "Application Mechanism for 2143 Keeping Alive the NAT Mappings Associated with RTP / RTP 2144 Control Protocol (RTCP) Flows", RFC 6263, 2145 DOI 10.17487/RFC6263, June 2011, 2146 . 2148 [RFC6792] Wu, Q., Ed., Hunt, G., and P. Arden, "Guidelines for Use 2149 of the RTP Monitoring Framework", RFC 6792, 2150 DOI 10.17487/RFC6792, November 2012, 2151 . 2153 [RFC7478] Holmberg, C., Hakansson, S., and G. Eriksson, "Web Real- 2154 Time Communication Use Cases and Requirements", RFC 7478, 2155 DOI 10.17487/RFC7478, March 2015, 2156 . 2158 Authors' Addresses 2160 Colin Perkins 2161 University of Glasgow 2162 School of Computing Science 2163 Glasgow G12 8QQ 2164 United Kingdom 2166 Email: csp@csperkins.org 2167 URI: https://csperkins.org/ 2169 Magnus Westerlund 2170 Ericsson 2171 Farogatan 6 2172 SE-164 80 Kista 2173 Sweden 2175 Phone: +46 10 714 82 87 2176 Email: magnus.westerlund@ericsson.com 2178 Joerg Ott 2179 Aalto University 2180 School of Electrical Engineering 2181 Espoo 02150 2182 Finland 2184 Email: jorg.ott@aalto.fi