idnits 2.17.1 draft-ietf-avtcore-rtp-topologies-update-10.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 (July 2, 2015) is 3219 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-12) exists of draft-ietf-avtcore-rtp-multi-stream-optimisation-05 == Outdated reference: A later version (-54) exists of draft-ietf-mmusic-sdp-bundle-negotiation-22 -- Obsolete informational reference (is this intentional?): RFC 5117 (Obsoleted by RFC 7667) -- Obsolete informational reference (is this intentional?): RFC 5285 (Obsoleted by RFC 8285) -- Obsolete informational reference (is this intentional?): RFC 5766 (Obsoleted by RFC 8656) Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Westerlund 3 Internet-Draft Ericsson 4 Obsoletes: 5117 (if approved) S. Wenger 5 Intended status: Informational Vidyo 6 Expires: January 3, 2016 July 2, 2015 8 RTP Topologies 9 draft-ietf-avtcore-rtp-topologies-update-10 11 Abstract 13 This document discusses point to point and multi-endpoint topologies 14 used in Real-time Transport Protocol (RTP)-based environments. In 15 particular, centralized topologies commonly employed in the video 16 conferencing industry are mapped to the RTP terminology. 18 This document is updated with additional topologies and replaces RFC 19 5117. 21 Status of This Memo 23 This Internet-Draft is submitted in full conformance with the 24 provisions of BCP 78 and BCP 79. 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF). Note that other groups may also distribute 28 working documents as Internet-Drafts. The list of current Internet- 29 Drafts is at http://datatracker.ietf.org/drafts/current/. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 This Internet-Draft will expire on January 3, 2016. 38 Copyright Notice 40 Copyright (c) 2015 IETF Trust and the persons identified as the 41 document authors. All rights reserved. 43 This document is subject to BCP 78 and the IETF Trust's Legal 44 Provisions Relating to IETF Documents 45 (http://trustee.ietf.org/license-info) in effect on the date of 46 publication of this document. Please review these documents 47 carefully, as they describe your rights and restrictions with respect 48 to this document. Code Components extracted from this document must 49 include Simplified BSD License text as described in Section 4.e of 50 the Trust Legal Provisions and are provided without warranty as 51 described in the Simplified BSD License. 53 Table of Contents 55 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 56 2. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . 4 57 2.1. Glossary . . . . . . . . . . . . . . . . . . . . . . . . 4 58 2.2. Definitions related to RTP grouping taxonomy . . . . . . 4 59 3. Topologies . . . . . . . . . . . . . . . . . . . . . . . . . 5 60 3.1. Point to Point . . . . . . . . . . . . . . . . . . . . . 5 61 3.2. Point to Point via Middlebox . . . . . . . . . . . . . . 6 62 3.2.1. Translators . . . . . . . . . . . . . . . . . . . . . 6 63 3.2.2. Back to Back RTP sessions . . . . . . . . . . . . . . 10 64 3.3. Point to Multipoint Using Multicast . . . . . . . . . . . 11 65 3.3.1. Any Source Multicast (ASM) . . . . . . . . . . . . . 11 66 3.3.2. Source Specific Multicast (SSM) . . . . . . . . . . . 13 67 3.3.3. SSM with Local Unicast Resources . . . . . . . . . . 15 68 3.4. Point to Multipoint Using Mesh . . . . . . . . . . . . . 16 69 3.5. Point to Multipoint Using the RFC 3550 Translator . . . . 19 70 3.5.1. Relay - Transport Translator . . . . . . . . . . . . 19 71 3.5.2. Media Translator . . . . . . . . . . . . . . . . . . 20 72 3.6. Point to Multipoint Using the RFC 3550 Mixer Model . . . 21 73 3.6.1. Media Mixing Mixer . . . . . . . . . . . . . . . . . 23 74 3.6.2. Media Switching . . . . . . . . . . . . . . . . . . . 26 75 3.7. Selective Forwarding Middlebox . . . . . . . . . . . . . 28 76 3.8. Point to Multipoint Using Video Switching MCUs . . . . . 32 77 3.9. Point to Multipoint Using RTCP-Terminating MCU . . . . . 33 78 3.10. Split Component Terminal . . . . . . . . . . . . . . . . 34 79 3.11. Non-Symmetric Mixer/Translators . . . . . . . . . . . . . 37 80 3.12. Combining Topologies . . . . . . . . . . . . . . . . . . 37 81 4. Topology Properties . . . . . . . . . . . . . . . . . . . . . 38 82 4.1. All to All Media Transmission . . . . . . . . . . . . . . 38 83 4.2. Transport or Media Interoperability . . . . . . . . . . . 39 84 4.3. Per Domain Bit-Rate Adaptation . . . . . . . . . . . . . 39 85 4.4. Aggregation of Media . . . . . . . . . . . . . . . . . . 40 86 4.5. View of All Session Participants . . . . . . . . . . . . 40 87 4.6. Loop Detection . . . . . . . . . . . . . . . . . . . . . 41 88 4.7. Consistency between header extensions and RTCP . . . . . 41 89 5. Comparison of Topologies . . . . . . . . . . . . . . . . . . 41 90 6. Security Considerations . . . . . . . . . . . . . . . . . . . 42 91 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 44 92 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 44 93 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 44 94 9.1. Normative References . . . . . . . . . . . . . . . . . . 44 95 9.2. Informative References . . . . . . . . . . . . . . . . . 45 96 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 46 98 1. Introduction 100 Real-time Transport Protocol (RTP) [RFC3550] topologies describe 101 methods for interconnecting RTP entities and their processing 102 behavior for RTP and RTCP. This document tries to address past and 103 existing confusion, especially with respect to terms not defined in 104 RTP but in common use in the communication industry, such as the 105 Multipoint Control Unit or MCU. 107 When the Audio-Visual Profile with Feedback (AVPF) [RFC4585] was 108 developed the main emphasis lay in the efficient support of point to 109 point and small multipoint scenarios without centralized multipoint 110 control. In practice, however, most multipoint conferences operate 111 utilizing centralized units referred to as MCUs. MCUs may implement 112 Mixer or Translator functionality (in RTP [RFC3550] terminology), and 113 signalling support. They may also contain additional application 114 layer functionality. This document focuses on the media transport 115 aspects of the MCU that can be realized using RTP, as discussed 116 below. Further considered are the properties of Mixers and 117 Translators, and how some types of deployed MCUs deviate from these 118 properties. 120 This document also codifies new multipoint architectures that have 121 recently been introduced and which were not anticipated in RFC 5117, 122 thus this document replaces [RFC5117]. These architectures use 123 scalable video coding and simulcasting, and their associated 124 centralized units are referred to as Selective Forwarding Units 125 (SFU). This codification provides a common information basis for 126 future discussion and specification work. 128 The new topologies are Point to Point via Middlebox (Section 3.2), 129 Source Specific Multicast (Section 3.3.2), SSM with Local Unicast 130 Resources (Section 3.3.3), Point to Multipoint Using Mesh 131 (Section 3.4), Selective Forwarding Middlebox (Section 3.7), and 132 Split Component Terminal (Section 3.10). The Point to Multipoint 133 Using the RFC 3550 Mixer Model (Section 3.6) has been significantly 134 expanded to cover two different versions namely Media Mixing Mixer 135 (Section 3.6.1), and Media Switching (Section 3.6.2). 137 The document's attempt to clarify and explain sections of the Real- 138 time Transport Protocol (RTP) spec [RFC3550] is informal. It is not 139 intended to update or change what is normatively specified within RFC 140 3550. 142 2. Definitions 144 2.1. Glossary 146 ASM: Any Source Multicast 148 AVPF: The Extended RTP Profile for RTCP-based Feedback 150 CSRC: Contributing Source 152 Link: The data transport to the next IP hop 154 Middlebox: A device that is on the Path that media travel between 155 two endpoints 157 MCU: Multipoint Control Unit 159 Path: The concatenation of multiple links, resulting in an end-to- 160 end data transfer. 162 PtM: Point to Multipoint 164 PtP: Point to Point 166 SFU: Selective Forwarding Unit 168 SSM: Source-Specific Multicast 170 SSRC: Synchronization Source 172 2.2. Definitions related to RTP grouping taxonomy 174 [Note to RFC editor: The following definitions have been taken from 175 draft-ietf-avtext-rtp-grouping-taxonomy-02 (taxonomy draft 176 henceforth). It is avtcore working group agreement to not delay the 177 publication of the topologies-update document through a dependency to 178 the taxonomy draft. If, however, the taxonomy draft and this draft 179 are in your work queue at the same time and there would be no 180 significant additional delay (through your schedule, normative 181 reference citations, or similar) in publishing both documents roughly 182 in parallel, it would be preferable to replace the definition 183 language with something like "as in [RFC YYYY]" where YYYY would be 184 the RFC number of the published taxonomy draft.] 186 The following definitions have been taken from draft-ietf-avtext-rtp- 187 grouping-taxonomy-02, and are used in capitalized form throughout the 188 document. 190 Communication Session: A Communication Session is an association 191 among group of participants communicating with each other via a 192 set of Multimedia Sessions. 194 Endpoint: A single addressable entity sending or receiving RTP 195 packets. It may be decomposed into several functional blocks, but 196 as long as it behaves as a single RTP stack entity it is 197 classified as a single "Endpoint". 199 Media Source: A Media Source is the logical source of a reference 200 clock synchronized, time progressing, digital media stream, called 201 a Source Stream. 203 Multimedia Session: A multimedia session is an association among a 204 group of participants engaged in the communication via one or more 205 RTP Sessions. 207 3. Topologies 209 This subsection defines several topologies that are relevant for 210 codec control but also RTP usage in other contexts. The section 211 starts with point to point cases, with or without middleboxes. Then 212 follows a number of different methods for establishing point to 213 multipoint communication. These are structured around the most 214 fundamental enabler, i.e., multicast, a mesh of connections, 215 translators, mixers and finally MCUs and SFUs. The section ends by 216 discussing de-composited terminals, asymmetric middlebox behaviors 217 and combining topologies. 219 The topologies may be referenced in other documents by a shortcut 220 name, indicated by the prefix "Topo-". 222 For each of the RTP-defined topologies, we discuss how RTP, RTCP, and 223 the carried media are handled. With respect to RTCP, we also discuss 224 the handling of RTCP feedback messages as defined in [RFC4585] and 225 [RFC5104]. 227 3.1. Point to Point 229 Shortcut name: Topo-Point-to-Point 231 The Point to Point (PtP) topology (Figure 1) consists of two 232 endpoints, communicating using unicast. Both RTP and RTCP traffic 233 are conveyed endpoint-to-endpoint, using unicast traffic only (even 234 if, in exotic cases, this unicast traffic happens to be conveyed over 235 an IP-multicast address). 237 +---+ +---+ 238 | A |<------->| B | 239 +---+ +---+ 241 Figure 1: Point to Point 243 The main property of this topology is that A sends to B, and only B, 244 while B sends to A, and only A. This avoids all complexities of 245 handling multiple endpoints and combining the requirements stemming 246 from them. Note that an endpoint can still use multiple RTP 247 Synchronization Sources (SSRCs) in an RTP session. The number of RTP 248 sessions in use between A and B can also be of any number, subject 249 only to system level limitations like the number range of ports. 251 RTCP feedback messages for the indicated SSRCs are communicated 252 directly between the endpoints. Therefore, this topology poses 253 minimal (if any) issues for any feedback messages. For RTP sessions 254 which use multiple SSRC per endpoint it can be relevant to implement 255 support for cross-reporting suppression as defined in "Sending 256 Multiple Media Streams in a Single RTP Session" 257 [I-D.ietf-avtcore-rtp-multi-stream-optimisation]. 259 3.2. Point to Point via Middlebox 261 This section discusses cases where two endpoints communicate but have 262 one or more middleboxes involved in the RTP session. 264 3.2.1. Translators 266 Shortcut name: Topo-PtP-Translator 268 Two main categories of Translators can be distinguished; Transport 269 Translators and Media translators. Both Translator types share 270 common attributes that separate them from Mixers. For each RTP 271 stream that the Translator receives, it generates an individual RTP 272 stream in the other domain. A translator keeps the SSRC for an RTP 273 stream across the translation, whereas a Mixer can select a single 274 RTP stream from multiple received RTP streams (in cases like audio/ 275 video switching), or send out an RTP stream composed of multiple 276 mixed media received in multiple RTP streams (in cases like audio 277 mixing or video tiling), but always under its own SSRC, possibly 278 using the CSRC field to indicate the source(s) of the content. 279 Mixers are more common in point to multipoint cases than in PtP. The 280 reason is that in PtP use cases the primary focus of a middlebox is 281 enabling interoperability, between otherwise non-interoperable 282 endpoints, such as transcoding to a codec the receiver supports, 283 which can be done by a media translator. 285 As specified in Section 7.1 of [RFC3550], the SSRC space is common 286 for all participants in the RTP session, independent of on which side 287 of the Translator the session resides. Therefore, it is the 288 responsibility of the endpoints (as the RTP session participants) to 289 run SSRC collision detection, and the SSRC is thus a field the 290 Translator cannot change. Any SDES information associated with a 291 SSRC or CSRC also needs to be forwarded between the domains for any 292 SSRC/CSRC used in the different domains. 294 A Translator commonly does not use an SSRC of its own, and is not 295 visible as an active participant in the RTP session. One reason to 296 have its own SSRC is when a Translator acts as a quality monitor that 297 sends RTCP reports and therefore is required to have an SSRC. 298 Another example is the case when a Translator is prepared to use RTCP 299 feedback messages. This may, for example, occur in a translator 300 configured to detect packet loss of important video packets and wants 301 to trigger repair by the media sending endpoint, by sending feedback 302 messages. While such feedback could use the SSRC of the target for 303 the translator (the receiving endpoint), this in turn would require 304 translation of the targets RTCP reports to make them consistent. It 305 may be simpler to expose an additional SSRC in the session. The only 306 concern is endpoints failing to support the full RTP specification 307 may have issues with multiple SSRCs reporting on the RTP streams sent 308 by that endpoint, as this use case may be viewed as excotic by 309 implementers. 311 In general, a Translator implementation should consider which RTCP 312 feedback messages or codec-control messages it needs to understand in 313 relation to the functionality of the Translator itself. This is 314 completely in line with the requirement to also translate RTCP 315 messages between the domains. 317 3.2.1.1. Transport Relay/Anchoring 319 Shortcut name: Topo-PtP-Relay 321 There exist a number of different types of middleboxes that might be 322 inserted between two endpoints on the transport level, e.g., to 323 perform changes on the IP/UDP headers, and are, therefore, basic 324 transport translators. These middleboxes come in many variations 325 including NAT [RFC3022] traversal by pinning the media path to a 326 public address domain relay, network topologies where the RTP stream 327 is required to pass a particular point for audit by employing 328 relaying, or preserving privacy by hiding each peer's transport 329 addresses to the other party. Other protocols or functionalities 330 that provide this behavior are TURN [RFC5766] servers, Session Border 331 Gateways and Media Processing Nodes with media anchoring 332 functionalities. 334 +---+ +---+ +---+ 335 | A |<------>| T |<------->| B | 336 +---+ +---+ +---+ 338 Figure 2: Point to Point with Translator 340 A common element in these functions is that they are normally 341 transparent at the RTP level, i.e., they perform no changes on any 342 RTP or RTCP packet fields and only affect the lower layers. They may 343 affect, however, the path the RTP and RTCP packets are routed between 344 the endpoints in the RTP session, and thereby indirectly affect the 345 RTP session. For this reason, one could believe that transport 346 translator-type middleboxes do not need to be included in this 347 document. This topology, however, can raise additional requirements 348 in the RTP implementation and its interactions with the signalling 349 solution. Both in signalling and in certain RTCP fields, network 350 addresses other than those of the relay can occur since B has a 351 different network address than the relay (T). Implementations that 352 cannot support this will also not work correctly when endpoints are 353 subject to NAT. 355 The transport relay implementations also have to take into account 356 security considerations. In particular, source address filtering of 357 incoming packets is usually important in relays, to prevent attackers 358 to inject traffic into a session, which one peer may, in the absence 359 fo adequate security in the relay, think it comes from the other 360 peer. 362 3.2.1.2. Transport Translator 364 Shortcut name: Topo-Trn-Translator 366 Transport Translators (Topo-Trn-Translator) do not modify the RTP 367 stream itself, but are concerned with transport parameters. 368 Transport parameters, in the sense of this section, comprise the 369 transport addresses (to bridge different domains such unicast to 370 multicast) and the media packetization to allow other transport 371 protocols to be interconnected to a session (in gateways). 373 Translators that bridge between different protocol worlds need to be 374 concerned about the mapping of the SSRC/CSRC (Contributing Source) 375 concept to the non-RTP protocol. When designing a Translator to a 376 non-RTP-based media transport, an important consideration is how to 377 handle different sources and their identities. This problem space is 378 not discussed henceforth. 380 Of the transport Translators, this memo is primarily interested in 381 those that use RTP on both sides, and this is assumed henceforth. 383 The most basic transport translators that operate below the RTP level 384 were already discussed in Section 3.2.1.1. 386 3.2.1.3. Media Translator 388 Shortcut name: Topo-Media-Translator 390 Media Translators (Topo-Media-Translator) modify the media inside the 391 RTP stream. This process is commonly known as transcoding. The 392 modification of the media can be as small as removing parts of the 393 stream, and it can go all the way to a full decoding and re-encoding 394 (down to the sample level or equivalent) utilizing a different media 395 codec. Media Translators are commonly used to connect endpoints 396 without a common interoperability point in the media encoding. 398 Stand-alone Media Translators are rare. Most commonly, a combination 399 of Transport and Media Translator is used to translate both the media 400 and the transport aspects of the RTP stream carrying the media 401 between two transport domains. 403 When media translation occurs, the Translator's task regarding 404 handling of RTCP traffic becomes substantially more complex. In this 405 case, the Translator needs to rewrite endpoint B's RTCP Receiver 406 Report before forwarding them to endpoint A. The rewriting is needed 407 as the RTP stream received by B is not the same RTP stream as the 408 other participants receive. For example, the number of packets 409 transmitted to B may be lower than what A sends, due to the different 410 media format and data rate. Therefore, if the Receiver Reports were 411 forwarded without changes, the extended highest sequence number would 412 indicate that B were substantially behind in reception, while most 413 likely it would not be. Therefore, the Translator must translate 414 that number to a corresponding sequence number for the stream the 415 Translator received. Similar requirements exists for most other 416 fields in the RTCP Receiver Reports. 418 A media Translator may in some cases act on behalf of the "real" 419 source (the endpoint originally sending the media to the Translator) 420 and respond to RTCP feedback messages. This may occur, for example, 421 when a receiving endpoint requests a bandwidth reduction, and the 422 media Translator has not detected any congestion or other reasons for 423 bandwidth reduction between the sending endpoint and itself. In that 424 case, it is sensible that the media Translator reacts to codec 425 control messages itself, for example, by transcoding to a lower media 426 rate. 428 A variant of translator behaviour worth pointing out is the one 429 depicted in Figure 3 of an endpoint A sending a RTP stream containing 430 media (only) to B. On the path there is a device T that on A's 431 behalf manipulates the RTP streams. One common example is that T 432 adds a second RTP stream containing Forward Error Correction (FEC) 433 information in order to protect A's (non FEC-protected) RTP stream. 434 In this case, T needs to semantically bind the new FEC RTP stream to 435 A's media-carrying RTP stream, for example by using the same CNAME as 436 A. 438 +------+ +------+ +------+ 439 | | | | | | 440 | A |------->| T |-------->| B | 441 | | | |---FEC-->| | 442 +------+ +------+ +------+ 444 Figure 3: Media Translator adding FEC 446 there may also be cases where information is added into the original 447 RTP stream, while leaving most or all of the original RTP packets 448 intact (with the exception of certain RTP header fields, such as the 449 sequence number). One example is the injection of meta-data into the 450 RTP stream, carried in their own RTP packets. 452 Similarly, a Media Translator can sometimes remove information from 453 the RTP stream, while otherwise leaving the remaining RTP packets 454 unchanged (again with the exception of certain RTP header fields). 456 Either type of functionality where T manipulates the RTP stream, or 457 adds an accompanying RTP stream, on behalf of A is also covered under 458 the media translator definition. 460 3.2.2. Back to Back RTP sessions 462 Shortcut name: Topo-Back-To-Back 464 There exist middleboxes that interconnect two endpoints A and B 465 through themselves (MB), but not by being part of a common RTP 466 session. They establish instead two different RTP sessions, one 467 between A and the middlebox and another between the middlebox and B. 468 This topology is called Topo-Back-To-Back 470 |<--Session A-->| |<--Session B-->| 471 +------+ +------+ +------+ 472 | A |------->| MB |-------->| B | 473 +------+ +------+ +------+ 475 Figure 4: Back-to-back RTP sessions through Middlebox 477 The middlebox acts as an application-level gateway and bridges the 478 two RTP sessions. This bridging can be as basic as forwarding the 479 RTP payloads between the sessions, or more complex including media 480 transcoding. The difference of this topology relative to the single 481 RTP session context is the handling of the SSRCs and the other 482 session-related identifiers, such as CNAMEs. With two different RTP 483 sessions these can be freely changed and it becomes the middlebox's 484 respnsibility to maintain the correct relations. 486 The signalling or other above-RTP level functionalities referencing 487 RTP streams may be what is most impacted by using two RTP sessions 488 and changing identifiers. The structure with two RTP sessions also 489 puts a congestion control requirement on the middlebox, because it 490 becomes fully responsible for the media stream it sources into each 491 of the sessions. 493 Adherence to congestion control can be solved locally on each of the 494 two segments, or by bridging statistics from the receiving endpoint 495 through the middlebox to the sending endpoint. From an 496 implementation point, however, the latter requires dealing with a 497 number of inconsistencies. First, packet loss must be detected for 498 an RTP stream sent from A to the middlebox, and that loss must be 499 reported through a skipped sequence number in the RTP stream from the 500 middlebox to B. This coupling and the resulting inconsistencies are 501 conceptually easier to handle when considering the two RTP streams as 502 belonging to a single RTP session. 504 3.3. Point to Multipoint Using Multicast 506 Multicast is an IP layer functionality that is available in some 507 networks. Two main flavors can be distinguished: Any Source 508 Multicast (ASM) [RFC1112] where any multicast group participant can 509 send to the group address and expect the packet to reach all group 510 participants; and Source Specific Multicast (SSM) [RFC3569], where 511 only a particular IP host sends to the multicast group. Each of 512 these models are discussed below in their respective sections. 514 3.3.1. Any Source Multicast (ASM) 516 Shortcut name: Topo-ASM (was Topo-Multicast) 517 +-----+ 518 +---+ / \ +---+ 519 | A |----/ \---| B | 520 +---+ / Multi- \ +---+ 521 + Cast + 522 +---+ \ Network / +---+ 523 | C |----\ /---| D | 524 +---+ \ / +---+ 525 +-----+ 527 Figure 5: Point to Multipoint Using Multicast 529 Point to Multipoint (PtM) is defined here as using a multicast 530 topology as a transmission model, in which traffic from any multicast 531 group participant reaches all the other multicast group participants, 532 except for cases such as: 534 o packet loss, or 536 o when a multicast group participant does not wish to receive the 537 traffic for a specific multicast group and, therefore, has not 538 subscribed to the IP multicast group in question. This scenario 539 can occur, for example, where a multimedia session is distributed 540 using two or more multicast groups and a multicast group 541 participant is subscribed only to a subset of these sessions. 543 In the above context, "traffic" encompasses both RTP and RTCP 544 traffic. The number of multicast group participants can vary between 545 one and many, as RTP and RTCP scale to very large multicast groups 546 (the theoretical limit of the number of participants in a single RTP 547 session is in the range of billions). The above can be realized 548 using Any Source Multicast (ASM). 550 For feedback usage, it is useful to define a "small multicast group" 551 as a group where the number of multicast group participants is so low 552 (and other factors such as the connectivity is so good) that it 553 allows the participants to use early or immediate feedback, as 554 defined in AVPF [RFC4585]. Even when the environment would allow for 555 the use of a small multicast group, some applications may still want 556 to use the more limited options for RTCP feedback available to large 557 multicast groups, for example when there is a likelihood that the 558 threshold of the small multicast group (in terms of multicast group 559 participants) may be exceeded during the lifetime of a session. 561 RTCP feedback messages in multicast reach, like media data, every 562 subscriber (subject to packet losses and multicast group 563 subscription). Therefore, the feedback suppression mechanism 564 discussed in [RFC4585] is typically required. Each individual 565 endpoint that is a multicast group participant needs to process every 566 feedback message it receives, not only to determine if it is affected 567 or if the feedback message applies only to some other endpoint, but 568 also to derive timing restrictions for the sending of its own 569 feedback messages, if any. 571 3.3.2. Source Specific Multicast (SSM) 573 Shortcut name: Topo-SSM 575 In Any Source Multicast, any of the multicast group participants can 576 send to all the other multicast group participants, by sending a 577 packet to the multicast group. In contrast, Source Specific 578 Multicast [RFC3569][RFC4607] refers to scenarios where only a single 579 source (Distribution Source) can send to the multicast group, 580 creating a topology that looks like the one below: 582 +--------+ +-----+ 583 |Media | | | Source-specific 584 |Sender 1|<----->| D S | Multicast 585 +--------+ | I O | +--+----------------> R(1) 586 | S U | | | | 587 +--------+ | T R | | +-----------> R(2) | 588 |Media |<----->| R C |->+ | : | | 589 |Sender 2| | I E | | +------> R(n-1) | | 590 +--------+ | B | | | | | | 591 : | U | +--+--> R(n) | | | 592 : | T +-| | | | | 593 : | I | |<---------+ | | | 594 +--------+ | O |F|<---------------+ | | 595 |Media | | N |T|<--------------------+ | 596 |Sender M|<----->| | |<-------------------------+ 597 +--------+ +-----+ RTCP Unicast 599 FT = Feedback Target 600 Transport from the Feedback Target to the Distribution 601 Source is via unicast or multicast RTCP if they are not 602 co-located. 604 Figure 6: Point to Multipoint using Source Specific Multicast 606 In the SSM topology (Figure 6) a number of RTP sending endpoints (RTP 607 sources henceforth) (1 to M) are allowed to send media to the SSM 608 group. These sources send media to a dedicated distribution source, 609 which forwards the RTP streams to the multicast group on behalf of 610 the original RTP sources. The RTP streams reach the receiving 611 endpoints (Receivers henceforth) (R(1) to R(n)). The Receivers' RTCP 612 messages cannot be sent to the multicast group, as the SSM multicast 613 group by definition has only a single IP sender. To support RTCP, an 614 RTP extension for SSM [RFC5760] was defined. It uses unicast 615 transmission to send RTCP from each of the receivers to one or more 616 Feedback Targets (FT). The feedback targets relay the RTCP 617 unmodified, or provide a summary of the participants RTCP reports 618 towards the whole group by forwarding the RTCP traffic to the 619 distribution source. Figure 6 only shows a single feedback target 620 integrated in the distribution source, but for scalability the FT can 621 be distributed and each instance can have responsibility for sub- 622 groups of the receivers. For summary reports, however, there 623 typically must be a single feedback target aggregating all the 624 summaries to a common message to the whole receiver group. 626 The RTP extension for SSM specifies how feedback (both reception 627 information and specific feedback events) are handled. The more 628 general problems associated with the use of multicast, where everyone 629 receives what the distribution source sends needs to be accounted 630 for. 632 Aforementioned situation results in common behavior for RTP 633 multicast: 635 1. Multicast applications often use a group of RTP sessions, not 636 one. Each endpoint needs to be a member of most or all of these 637 RTP sessions in order to perform well. 639 2. Within each RTP session, the number of media sinks is likely to 640 be much larger than the number of RTP sources. 642 3. Multicast applications need signalling functions to identify the 643 relationships between RTP sessions. 645 4. Multicast applications need signalling functions to identify the 646 relationships between SSRCs in different RTP sessions. 648 All multicast configurations share a signalling requirement: all of 649 the endpoints need to have the same RTP and payload type 650 configuration. Otherwise, endpoint A could, for example, be using 651 payload type 97 to identify the video codec H.264, while endpoint B 652 would identify it as MPEG-2, with unpredicatble but almost certainly 653 not visually pleasing results. 655 Security solutions for this type of group communications are also 656 challenging. First, the key-management and the security protocol 657 must support group communication. Source authentication becomes more 658 difficult and requires specialized solutions. For more discussion on 659 this please review Options for Securing RTP Sessions [RFC7201]. 661 3.3.3. SSM with Local Unicast Resources 663 Shortcut name: Topo-SSM-RAMS 665 "Unicast-Based Rapid Acquisition of Multicast RTP Sessions" [RFC6285] 666 results in additional extensions to SSM Topology. 668 ----------- -------------- 669 | |------------------------------------>| | 670 | |.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.->| | 671 | | | | 672 | Multicast | ---------------- | | 673 | Source | | Retransmission | | | 674 | |-------->| Server (RS) | | | 675 | |.-.-.-.->| | | | 676 | | | ------------ | | | 677 ----------- | | Feedback | |<.=.=.=.=.| | 678 | | Target (FT)| |<~~~~~~~~~| RTP Receiver | 679 PRIMARY MULTICAST | ------------ | | (RTP_Rx) | 680 RTP SESSION with | | | | 681 UNICAST FEEDBACK | | | | 682 | | | | 683 - - - - - - - - - - - |- - - - - - - - |- - - - - |- - - - - - - |- - 684 | | | | 685 UNICAST BURST | ------------ | | | 686 (or RETRANSMISSION) | | Burst/ | |<~~~~~~~~>| | 687 RTP SESSION | | Retrans. | |.........>| | 688 | |Source (BRS)| |<.=.=.=.=>| | 689 | ------------ | | | 690 | | | | 691 ---------------- -------------- 693 -------> Multicast RTP Stream 694 .-.-.-.> Multicast RTCP Stream 695 .=.=.=.> Unicast RTCP Reports 696 ~~~~~~~> Unicast RTCP Feedback Messages 697 .......> Unicast RTP Stream 699 Figure 7: SSM with Local Unicast Resources (RAMS) 701 The Rapid acquisition extension allows an endpoint joining an SSM 702 multicast session to request media starting with the last sync-point 703 (from where media can be decoded without requiring context 704 established by the decoding of prior packets) to be sent at high 705 speed until such time where, after decoding of these burst-delivered 706 media packets, the correct media timing is established, i.e. media 707 packets are received within adequate buffer intervals for this 708 application. This is accomplished by first establishing a unicast 709 PtP RTP session between the Burst/Retransmission Source (BRS, 710 Figure 7) and the RTP Receiver. The unicast session is used to 711 transmit cached packets from the multicast group at higher then 712 normal speed in order to synchronize the receiver to the ongoing 713 multicast RTP stream. Once the RTP receiver and its decoder have 714 caught up with the multicast session's current delivery, the receiver 715 switches over to receiving directly from the multicast group. In 716 many deployed application, the (still existing) PtP RTP session is 717 used as a repair channel, i.e., for RTP Retransmission traffic of 718 those packets that were not received from the multicast group. 720 3.4. Point to Multipoint Using Mesh 722 Shortcut name: Topo-Mesh 724 +---+ +---+ 725 | A |<---->| B | 726 +---+ +---+ 727 ^ ^ 728 \ / 729 \ / 730 v v 731 +---+ 732 | C | 733 +---+ 735 Figure 8: Point to Multi-Point using Mesh 737 Based on the RTP session definition, it is clearly possible to have a 738 joint RTP session involving three or more endpoints over multiple 739 unicast transport flows, like the joint three endpoint session 740 depicted above. In this case, A needs to send its RTP streams and 741 RTCP packets to both B and C over their respective transport flows. 742 As long as all endpoints do the same, everyone will have a joint view 743 of the RTP session. 745 This topology does not create any additional requirements beyond the 746 need to have multiple transport flows associated with a single RTP 747 session. Note that an endpoint may use a single local port to 748 receive all these transport flows (in which case the sending port, IP 749 address, or SSRC can be used to demultiplex), or it might have 750 separate local reception ports for each of the endpoints. 752 +-A--------------------+ 753 |+---+ | 754 ||CAM| | +-B-----------+ 755 |+---+ +-UDP1------| |-UDP1------+ | 756 | | | +-RTP1----| |-RTP1----+ | | 757 | V | | +-Video-| |-Video-+ | | | 758 |+----+ | | | |<----------------|BV1 | | | | 759 ||ENC |----+-+-+--->AV1|---------------->| | | | | 760 |+----+ | | +-------| |-------+ | | | 761 | | | +---------| |---------+ | | 762 | | +-----------| |-----------+ | 763 | | | +-------------+ 764 | | | 765 | | | +-C-----------+ 766 | | +-UDP2------| |-UDP2------+ | 767 | | | +-RTP1----| |-RTP1----+ | | 768 | | | | +-Video-| |-Video-+ | | | 769 | +-------+-+-+--->AV1|---------------->| | | | | 770 | | | | |<----------------|CV1 | | | | 771 | | | +-------| |-------+ | | | 772 | | +---------| |---------+ | | 773 | +-----------| |-----------+ | 774 +----------------------+ +-------------+ 776 Figure 9: An Multi-unicast Mesh with a joint RTP session 778 Figure 9 depicts endpoints A's view of using a common RTP session 779 when establishing the mesh as shown in Figure 8. There is only one 780 RTP session (RTP1) but two transport flows (UDP1 and UDP2). The 781 Media Source (CAM) is encoded and transmitted over the SSRC (AV1) 782 across both transport layers. However, as this is a joint RTP 783 session, the two streams must be the same. Thus, an congestion 784 control adaptation needed for the paths A to B and A to C needs to 785 use the most restricting path's properties. 787 An alternative structure for establishing the above topology is to 788 use independent RTP sessions between each pair of peers, i.e., three 789 different RTP sessions. In some scenarios, the same RTP stream may 790 be sent from the transmitting endpoint, however it also supports 791 local adaptation taking place in one or more of the RTP streams, 792 rendering them non-identical. 794 +-A----------------------+ +-B-----------+ 795 |+---+ | | | 796 ||MIC| +-UDP1------| |-UDP1------+ | 797 |+---+ | +-RTP1----| |-RTP1----+ | | 798 | | +----+ | | +-Audio-| |-Audio-+ | | | 799 | +->|ENC1|--+-+-+--->AA1|------------->| | | | | 800 | | +----+ | | | |<-------------|BA1 | | | | 801 | | | | +-------| |-------+ | | | 802 | | | +---------| |---------+ | | 803 | | +-----------| |-----------+ | 804 | | ------------| |-------------| 805 | | | |-------------+ 806 | | | 807 | | | +-C-----------+ 808 | | | | | 809 | | +-UDP2------| |-UDP2------+ | 810 | | | +-RTP2----| |-RTP2----+ | | 811 | | +----+ | | +-Audio-| |-Audio-+ | | | 812 | +->|ENC2|--+-+-+--->AA2|------------->| | | | | 813 | +----+ | | | |<-------------|CA1 | | | | 814 | | | +-------| |-------+ | | | 815 | | +---------| |---------+ | | 816 | +-----------| |-----------+ | 817 +------------------------+ +-------------+ 819 Figure 10: An Multi-unicast Mesh with independent RTP session 821 Lets review the topology when independent RTP sessions are used, from 822 A's perspective in Figure 10 by considering both how the media is 823 handled and the RTP sessions that are set-up in Figure 10. A's 824 microphone is captured and the audio is fed into two different 825 encoder instances, each with a different independent RTP session, 826 i.e. RTP1 and RTP2 respectively. The SSRCs (AA1 and AA2) in each RTP 827 session are completely independent and the media bit-rate produced by 828 the encoders can also be tuned differently to address any congestion 829 control requirements differing for the paths A to B compared to A to 830 C. 832 From a topologies viewpoint, an important difference exists in the 833 behavior around RTCP. First, when a single RTP session spans all 834 three endpoints A, B, and C, and their connecting RTP streams, a 835 common RTCP bandwidth is calculated and used for this single joint 836 session. In contrast, when there are multiple independent RTP 837 sessions, each RTP session has its local RTCP bandwidth allocation. 839 Further, when multiple sessions are used, endpoints not directly 840 involved in a session do not have any awareness of the conditions in 841 those sessions. For example, in the case of the three endpoint 842 configuration in Figure 8, endpoint A has no awareness of the 843 conditions occurring in the session between endpoints B and C 844 (whereas, if a single RTP session were used, it would have such 845 awareness). 847 Loop detection is also affected. With independent RTP sessions, the 848 SSRC/CSRC cannot be used to determine when an endpoint receives its 849 own media stream, or a mixed media stream including its own media 850 stream (a condition known as a loop). The identification of loops 851 and, in most cases, their avoidance, has to be achieved by other 852 means, for example through signaling or the use of an RTP external 853 name space binding SSRC/CSRC among any communicating RTP sessions in 854 the mesh. 856 3.5. Point to Multipoint Using the RFC 3550 Translator 858 This section discusses some additional usages related to point to 859 multipoint of Translators compared to the point to point only cases 860 in Section 3.2.1. 862 3.5.1. Relay - Transport Translator 864 Shortcut name: Topo-PtM-Trn-Translator 866 This section discusses Transport Translator only usages to enable 867 multipoint sessions. 869 +-----+ 870 +---+ / \ +------------+ +---+ 871 | A |<---/ \ | |<---->| B | 872 +---+ / Multi- \ | | +---+ 873 + cast +->| Translator | 874 +---+ \ Network / | | +---+ 875 | C |<---\ / | |<---->| D | 876 +---+ \ / +------------+ +---+ 877 +-----+ 879 Figure 11: Point to Multipoint Using Multicast 881 Figure 11 depicts an example of a Transport Translator performing at 882 least IP address translation. It allows the (non-multicast-capable) 883 endpoints B and D to take part in an any source multicast session 884 involving endpoints A and C, by having the Translator forward their 885 unicast traffic to the multicast addresses in use, and vice versa. 886 It must also forward B's traffic to D, and vice versa, to provide 887 each of B and D with a complete view of the session. 889 +---+ +------------+ +---+ 890 | A |<---->| |<---->| B | 891 +---+ | | +---+ 892 | Translator | 893 +---+ | | +---+ 894 | C |<---->| |<---->| D | 895 +---+ +------------+ +---+ 897 Figure 12: RTP Translator (Relay) with Only Unicast Paths 899 Another Translator scenario is depicted in Figure 12. The Translator 900 in this case connects multiple endpoints through unicast. This can 901 be implemented using a very simple transport Translator which, in 902 this document, is called a relay. The relay forwards all traffic it 903 receives, both RTP and RTCP, to all other endpoints. In doing so, a 904 multicast network is emulated without relying on a multicast-capable 905 network infrastructure. 907 For RTCP feedback this results in a similar set of considerations to 908 those described in the ASM RTP topology. It also puts some 909 additional signalling requirements onto the session establishment; 910 for example, a common configuration of RTP payload types is required. 912 Transport translators and relays should always consider implementing 913 source address filtering, to prevent attackers to inject traffic 914 using the listening ports on the translator. The translator can, 915 however, go one step further, and especially if explicit SSRC 916 signalling is used, prevent endpoints to send SSRCs other than its 917 own (that are, for example, used by other participants in the 918 session). This can improve the security properties of the session, 919 despite the use of group keys that on cryptographic level allows 920 anyone to impersonate another in the same RTP session. 922 A Translator that doesn't change the RTP/RTCP packets content can be 923 operated without the requiring it to have access to the security 924 contexts used to protect the RTP/RTCP traffic between the 925 participants. 927 3.5.2. Media Translator 929 In the context of multipoint communications a Media Translator is not 930 providing new mechanisms to establish a multipoint session. It is 931 more of an enabler, or facilitator, that ensures a given endpoint or 932 a defined sub-set of endpoints can participate in the session. 934 If endpoint B in Figure 11 were behind a limited network path, the 935 Translator may perform media transcoding to allow the traffic 936 received from the other endpoints to reach B without overloading the 937 path. This transcoding can help the other endpoints in the multicast 938 part of the session, by not requiring the quality transmitted by A to 939 be lowered to the bitrates that B is actually capable of receiving 940 (and vice versa). 942 3.6. Point to Multipoint Using the RFC 3550 Mixer Model 944 Shortcut name: Topo-Mixer 946 A Mixer is a middlebox that aggregates multiple RTP streams that are 947 part of a session by generating one or more new RTP streams and, in 948 most cases, by manipulating the media data. One common application 949 for a Mixer is to allow a participant to receive a session with a 950 reduced amount of resources. 952 +-----+ 953 +---+ / \ +-----------+ +---+ 954 | A |<---/ \ | |<---->| B | 955 +---+ / Multi- \ | | +---+ 956 + cast +->| Mixer | 957 +---+ \ Network / | | +---+ 958 | C |<---\ / | |<---->| D | 959 +---+ \ / +-----------+ +---+ 960 +-----+ 962 Figure 13: Point to Multipoint Using the RFC 3550 Mixer Model 964 A Mixer can be viewed as a device terminating the RTP streams 965 received from other endpoints in the same RTP session. Using the 966 media data carried in the received RTP streams, a Mixer generates 967 derived RTP streams that are sent to the receiving endpoints. 969 The content that the Mixer provides is the mixed aggregate of what 970 the Mixer receives over the PtP or PtM paths, which are part of the 971 same Communication Session. 973 The Mixer creates the Media Source and the source RTP stream just 974 like an endpoint, as it mixes the content (often in the uncompressed 975 domain) and then encodes and packetizes it for transmission to a 976 receiving endpoint. The CSRC Count (CC) and CSRC fields in the RTP 977 header can be used to indicate the contributors to the newly 978 generated RTP stream. The SSRCs of the to-be-mixed streams on the 979 Mixer input appear as the CSRCs at the Mixer output. That output 980 stream uses a unique SSRC that identifies the Mixer's stream. The 981 CSRC should be forwarded between the different endpoints to allow for 982 loop detection and identification of sources that are part of the 983 Communication Session. Note that Section 7.1 of RFC 3550 requires 984 the SSRC space to be shared between domains for these reasons. This 985 also implies that any SDES information normally needs to be forwarded 986 across the mixer. 988 The Mixer is responsible for generating RTCP packets in accordance 989 with its role. It is an RTP receiver and should therefore send RTCP 990 receiver reports for the RTP streams it receives and terminates. In 991 its role as an RTP sender, it should also generate RTCP sender 992 reports for those RTP streams it sends. As specified in Section 7.3 993 of RFC 3550, a Mixer must not forward RTCP unaltered between the two 994 domains. 996 The Mixer depicted in Figure 13 is involved in three domains that 997 need to be separated: the any source multicast network (including 998 endpoints A and C), endpoint B, and endpoint D. Assuming all four 999 endpoints in the conference are interested in receiving content from 1000 all other endpoints, the Mixer produces different mixed RTP streams 1001 for B and D, as the one to B may contain content received from D, and 1002 vice versa. However, the Mixer may only need one SSRC per media type 1003 in each domain where it is the receiving entity and transmitter of 1004 mixed content. 1006 In the multicast domain, a Mixer still needs to provide a mixed view 1007 of the other domains. This makes the Mixer simpler to implement and 1008 avoids any issues with advanced RTCP handling or loop detection, 1009 which would be problematic if the Mixer were providing non-symmetric 1010 behavior. Please see Section 3.11 for more discussion on this topic. 1011 The mixing operation, however, in each domain could potentially be 1012 different. 1014 A Mixer is responsible for receiving RTCP feedback messages and 1015 handling them appropriately. The definition of "appropriate" depends 1016 on the message itself and the context. In some cases, the reception 1017 of a codec-control message by the Mixer may result in the generation 1018 and transmission of RTCP feedback messages by the Mixer to the 1019 endpoints in the other domain(s). In other cases, a message is 1020 handled by the Mixer locally and therefore not forwarded to any other 1021 domain. 1023 When replacing the multicast network in Figure 13 (to the left of the 1024 Mixer) with individual unicast paths as depicted in Figure 14, the 1025 Mixer model is very similar to the one discussed in Section 3.9 1026 below. Please see the discussion in Section 3.9 about the 1027 differences between these two models. 1029 +---+ +------------+ +---+ 1030 | A |<---->| |<---->| B | 1031 +---+ | | +---+ 1032 | Mixer | 1033 +---+ | | +---+ 1034 | C |<---->| |<---->| D | 1035 +---+ +------------+ +---+ 1037 Figure 14: RTP Mixer with Only Unicast Paths 1039 We now discuss in more detail the different mixing operations that a 1040 mixer can perform and how they can affect RTP and RTCP behavior. 1042 3.6.1. Media Mixing Mixer 1044 The media mixing mixer is likely the one that most think of when they 1045 hear the term "mixer". Its basic mode of operation is that it 1046 receives RTP streams from several endpoints and selects the stream(s) 1047 to be included in a media-domain mix. The selection can be through 1048 static configuration or by dynamic, content dependent means such as 1049 voice activation. The mixer then creates a single outgoing RTP 1050 stream from this mix. 1052 The most commonly deployed media mixer is probably the audio mixer, 1053 used in voice conferencing, where the output consists of a mixture of 1054 all the input audio signals; this needs minimal signalling to be 1055 successfully set up. From a signal processing viewpoint, audio 1056 mixing is relatively straightforward and commonly possible for a 1057 reasonable number of endpoints. Assume, for example, that one wants 1058 to mix N streams from N different endpoints. The mixer needs to 1059 decode those N streams, typically into the sample domain, and then 1060 produce N or N+1 mixes. Different mixes are needed so that each 1061 contributing source gets a mix of all other sources except its own, 1062 as this would result in an echo. When N is lower than the number of 1063 all endpoints, one may produce a mix of all N streams for the group 1064 that are currently not included in the mix, thus N+1 mixes. These 1065 audio streams are then encoded again, RTP packetized and sent out. 1066 In many cases, audio level normalization, noise suppression, and 1067 similar signal processing steps are also required or desirable before 1068 the actual mixing process commences. 1070 In video, the term "mixing" has a different interpretation than 1071 audio. It is commonly used to refer to the process of spatially 1072 combining contributed video streams, which is also known as "tiling". 1073 The reconstructed, appropriately scaled down videos can be spatially 1074 arranged in a set of tiles, each tile containing the video from an 1075 endpoint (typically showing a human participant). Tiles can be of 1076 different sizes, so that, for example, a particularly important 1077 participant, or the loudest speaker, is being shown on in larger tile 1078 than other participants. A self-view picture can be included in the 1079 tiling, which can either be locally produced or be a feedback from a 1080 mixer-received and reconstructed video image. Such remote loopback 1081 allows for confidence monitoring, i.e., it enables the participant to 1082 see himself/herself in the same quality as other participants see 1083 him/her. The tiling normally operates on reconstructed video in the 1084 sample domain. The tiled image is encoded, packetized, and sent by 1085 the mixer to the receiving endpoints. It is possible that a 1086 middlebox with media mixing duties contains only a single mixer of 1087 the aforementioned type, in which case all participants necessarily 1088 see the same tiled video, even if it is being sent over different RTP 1089 streams. More common, however, are mixing arrangement where an 1090 individual mixer is available for each outgoing port of the 1091 middlebox, allowing individual compositions for each receiving 1092 endpoint (a feature commonly referred to as personalized layout). 1094 One problem with media mixing is that it consumes both large amounts 1095 of media processing resources (for the decoding and mixing process in 1096 the uncompressed domain) and encoding resources (for the encoding of 1097 the mixed signal). Another problem is the quality degradation 1098 created by decoding and re-encoding the media, which is the result of 1099 the lossy nature of most commonly used media codecs. A third problem 1100 is the latency introduced by the media mixing, which can be 1101 substantial and annoyingly noticeable in case of video, or in case of 1102 audio if that mixed audio is lip-sychronized with high latency video. 1103 The advantage of media mixing is that it is straightforward for the 1104 endpoints to handle the single media stream (which includes the mixed 1105 aggregate of many sources), as they don't need to handle multiple 1106 decodings, local mixing and composition. In fact, mixers were 1107 introduced in pre-RTP times so that legacy, single stream receiving 1108 endpoints (that, in some protocol environments, actually didn't need 1109 to be aware of the multipoint nature of the conference) could 1110 successfully participate in what a user would recognize as a 1111 multiparty video conference. 1113 +-A---------+ +-MIXER----------------------+ 1114 | +-RTP1----| |-RTP1------+ +-----+ | 1115 | | +-Audio-| |-Audio---+ | +---+ | | | 1116 | | | AA1|--------->|---------+-+-|DEC|->| | | 1117 | | | |<---------|MA1 <----+ | +---+ | | | 1118 | | | | |(BA1+CA1)|\| +---+ | | | 1119 | | +-------| |---------+ +-|ENC|<-| B+C | | 1120 | +---------| |-----------+ +---+ | | | 1121 +-----------+ | | | | 1122 | | M | | 1123 +-B---------+ | | E | | 1124 | +-RTP2----| |-RTP2------+ | D | | 1125 | | +-Audio-| |-Audio---+ | +---+ | I | | 1126 | | | BA1|--------->|---------+-+-|DEC|->| A | | 1127 | | | |<---------|MA2 <----+ | +---+ | | | 1128 | | +-------| |(AA1+CA1)|\| +---+ | | | 1129 | +---------| |---------+ +-|ENC|<-| A+C | | 1130 +-----------+ |-----------+ +---+ | | | 1131 | | M | | 1132 +-C---------+ | | I | | 1133 | +-RTP3----| |-RTP3------+ | X | | 1134 | | +-Audio-| |-Audio---+ | +---+ | E | | 1135 | | | CA1|--------->|---------+-+-|DEC|->| R | | 1136 | | | |<---------|MA3 <----+ | +---+ | | | 1137 | | +-------| |(AA1+BA1)|\| +---+ | | | 1138 | +---------| |---------+ +-|ENC|<-| A+B | | 1139 +-----------+ |-----------+ +---+ +-----+ | 1140 +----------------------------+ 1142 Figure 15: Session and SSRC details for Media Mixer 1144 From an RTP perspective media mixing can be a very simple process, as 1145 can be seen in Figure 15. The mixer presents one SSRC towards the 1146 receiving endpoint, e.g., MA1 to Peer A, where the associated stream 1147 is the media mix of the other endpoints. As each peer, in this 1148 example, receives a different version of a mix from the mixer, there 1149 is no actual relation between the different RTP sessions in terms of 1150 actual media or transport level information. There are, however, 1151 common relationships between RTP1-RTP3, namely SSRC space and 1152 identity information. When A receives the MA1 stream which is a 1153 combination of BA1 and CA1 streams, the mixer may include CSRC 1154 information in the MA1 stream to identify the contributing source BA1 1155 and CA1, allowing the receiver to identify the contributing sources 1156 even if this were not possible through the media itself or through 1157 other signaling means. 1159 The CSRC has, in turn, utility in RTP extensions, like the Mixer to 1160 Client audio levels RTP header extension [RFC6465]. If the SSRCs 1161 from the endpoint to mixer paths are used as CSRCs in another RTP 1162 session, then RTP1, RTP2 and RTP3 become one joint session as they 1163 have a common SSRC space. At this stage, the mixer also needs to 1164 consider which RTCP information it needs to expose in the different 1165 paths. In the above scenario, a mixer would normally expose nothing 1166 more than the Source Description (SDES) information and RTCP BYE for 1167 a CSRC leaving the session. The main goal would be to enable the 1168 correct binding against the application logic and other information 1169 sources. This also enables loop detection in the RTP session. 1171 3.6.2. Media Switching 1173 Media switching mixers are used in limited functionality scenarios 1174 where no, or only very limited, concurrent presentation of multiple 1175 sources is required by the application, to more complex multi-stream 1176 usages with receiver mixing or tiling, including combined with 1177 simulcast and/or scalability between source and mixer. An RTP Mixer 1178 based on media switching avoids the media decoding and encoding 1179 operations in the mixer, as it conceptually forwards the encoded 1180 media stream as it was being sent to the mixer. It does not avoid, 1181 however, the decryption and re-encryption cycle as it rewrites RTP 1182 headers. Forwarding media (in contrast to reconstructing-mixing- 1183 encoding media) reduces the amount of computational resources needed 1184 in the mixer and increases the media quality (both in terms of 1185 fidelity and reduced latency). 1187 A media switching mixer maintains a pool of SSRCs representing 1188 conceptual or functional RTP streams that the mixer can produce. 1189 These RTP streams are created by selecting media from one of the RTP 1190 streams received by the mixer and forwarded to the peer using the 1191 mixer's own SSRCs. The mixer can switch between available sources if 1192 that is required by the concept for the source, like the currently 1193 active speaker. Note that the mixer, in most cases, still needs to 1194 perform a certain amount of media processing, as many media formats 1195 do not allow to "tune into" the stream at arbitrary points in their 1196 bitstream. 1198 To achieve a coherent RTP stream from the mixer's SSRC, the mixer 1199 needs to rewrite the incoming RTP packet's header. First the SSRC 1200 field must be set to the value of the Mixer's SSRC. Second, the 1201 sequence number must be the next in the sequence of outgoing packets 1202 it sent. Third, the RTP timestamp value needs to be adjusted using 1203 an offset that changes each time one switches media source. Finally, 1204 depending on the negotiation of the RTP payload type, the value 1205 representing this particular RTP payload configuration may have to be 1206 changed if the different endpoint-to-mixer paths have not arrived on 1207 the same numbering for a given configuration. This also requires 1208 that the different endpoints support a common set of codecs, 1209 otherwise media transcoding for codec compatibility would still be 1210 required. 1212 We now consider the operation of a media switching mixer that 1213 supports a video conference with six participating endpoints (A-F) 1214 where the two most recent speakers in the conference are shown to 1215 each receiving endpoint. The mixer has thus two SSRCs sending video 1216 to each peer, and each peer is capable of locally handling two video 1217 streams simultaneously. 1219 +-A---------+ +-MIXER----------------------+ 1220 | +-RTP1----| |-RTP1------+ +-----+ | 1221 | | +-Video-| |-Video---+ | | | | 1222 | | | AV1|------------>|---------+-+------->| S | | 1223 | | | |<------------|MV1 <----+-+-BV1----| W | | 1224 | | | |<------------|MV2 <----+-+-EV1----| I | | 1225 | | +-------| |---------+ | | T | | 1226 | +---------| |-----------+ | C | | 1227 +-----------+ | | H | | 1228 | | | | 1229 +-B---------+ | | M | | 1230 | +-RTP2----| |-RTP2------+ | A | | 1231 | | +-Video-| |-Video---+ | | T | | 1232 | | | BV1|------------>|---------+-+------->| R | | 1233 | | | |<------------|MV3 <----+-+-AV1----| I | | 1234 | | | |<------------|MV4 <----+-+-EV1----| X | | 1235 | | +-------| |---------+ | | | | 1236 | +---------| |-----------+ | | | 1237 +-----------+ | | | | 1238 : : : : 1239 : : : : 1240 +-F---------+ | | | | 1241 | +-RTP6----| |-RTP6------+ | | | 1242 | | +-Video-| |-Video---+ | | | | 1243 | | | FV1|------------>|---------+-+------->| | | 1244 | | | |<------------|MV11 <---+-+-AV1----| | | 1245 | | | |<------------|MV12 <---+-+-EV1----| | | 1246 | | +-------| |---------+ | | | | 1247 | +---------| |-----------+ +-----+ | 1248 +-----------+ +----------------------------+ 1250 Figure 16: Media Switching RTP Mixer 1252 The Media Switching RTP mixer can, similarly to the Media Mixing 1253 Mixer, reduce the bit-rate required for media transmission towards 1254 the different peers by selecting and forwarding only a sub-set of RTP 1255 streams it receives from the sending endpoints. In cases the mixer 1256 receives simulcast transmissions or a scalable encoding of the media 1257 source, the mixer has more degrees of freedom to select streams or 1258 sub-sets of stream to forward to a receiving endpoint, both based on 1259 transport or endpoint restrictions as well as application logic. 1261 To ensure that a media receiver in an endpoint can correctly decode 1262 the media in the RTP stream after a switch, a codec that uses 1263 temporal prediction needs to start its decoding from independent 1264 refresh points, or points in the bitstream offering similar 1265 functionality (like "dirty refresh points"). For some codecs, for 1266 example frame based speech and audio codecs, this is easily achieved 1267 by starting the decoding at RTP packet boundaries, as each packet 1268 boundary provides a refresh point (assuming proper packetization on 1269 the encoder side). For other codecs, particularly in video, refresh 1270 points are less common in the bitstream or may not be present at all 1271 without an explicit request to the respective encoder. The Full 1272 Intra Request [RFC5104] RTCP codec control message has been defined 1273 for this purpose. 1275 In this type of mixer one could consider to fully terminate the RTP 1276 sessions between the different endpoint and mixer paths. The same 1277 arguments and considerations as discussed in Section 3.9 need to be 1278 taken into consideration and apply here. 1280 3.7. Selective Forwarding Middlebox 1282 Another method for handling media in the RTP mixer is to "project", 1283 or make available, all potential RTP sources (SSRCs) into a per- 1284 endpoint, independent RTP session. The middlebox can select which of 1285 the potential sources that are currently actively transmitting media 1286 will be sent to each of the endpoints. This is similar to the media 1287 switching Mixer but has some important differences in RTP details. 1289 +-A---------+ +-Middlebox-----------------+ 1290 | +-RTP1----| |-RTP1------+ +-----+ | 1291 | | +-Video-| |-Video---+ | | | | 1292 | | | AV1|------------>|---------+-+------>| | | 1293 | | | |<------------|BV1 <----+-+-------| S | | 1294 | | | |<------------|CV1 <----+-+-------| W | | 1295 | | | |<------------|DV1 <----+-+-------| I | | 1296 | | | |<------------|EV1 <----+-+-------| T | | 1297 | | | |<------------|FV1 <----+-+-------| C | | 1298 | | +-------| |---------+ | | H | | 1299 | +---------| |-----------+ | | | 1300 +-----------+ | | M | | 1301 | | A | | 1302 +-B---------+ | | T | | 1303 | +-RTP2----| |-RTP2------+ | R | | 1304 | | +-Video-| |-Video---+ | | I | | 1305 | | | BV1|------------>|---------+-+------>| X | | 1306 | | | |<------------|AV1 <----+-+-------| | | 1307 | | | |<------------|CV1 <----+-+-------| | | 1308 | | | | : : : |: : : : : : : : :| | | 1309 | | | |<------------|FV1 <----+-+-------| | | 1310 | | +-------| |---------+ | | | | 1311 | +---------| |-----------+ | | | 1312 +-----------+ | | | | 1313 : : : : 1314 : : : : 1315 +-F---------+ | | | | 1316 | +-RTP6----| |-RTP6------+ | | | 1317 | | +-Video-| |-Video---+ | | | | 1318 | | | FV1|------------>|---------+-+------>| | | 1319 | | | |<------------|AV1 <----+-+-------| | | 1320 | | | | : : : |: : : : : : : : :| | | 1321 | | | |<------------|EV1 <----+-+-------| | | 1322 | | +-------| |---------+ | | | | 1323 | +---------| |-----------+ +-----+ | 1324 +-----------+ +---------------------------+ 1326 Figure 17: Selective Forwarding Middlebox 1328 In the six endpoint conference depicted above in (Figure 17) one can 1329 see that endpoint A is aware of five incoming SSRCs, BV1-FV1. If 1330 this middlebox intends to have a similar behavior as in Section 3.6.2 1331 where the mixer provides the endpoints with the two latest speaking 1332 endpoints, then only two out of these five SSRCs need concurrently 1333 transmit media to A. As the middlebox selects the source in the 1334 different RTP sessions that transmit media to the endpoints, each RTP 1335 stream requires rewriting of certain RTP header fields when being 1336 projected from one session into another. In particular, the sequence 1337 number needs to be consecutively incremented based on the packet 1338 actually being transmitted in each RTP session. Therefore, the RTP 1339 sequence number offset will change each time a source is turned on in 1340 a RTP session. The timestamp (possibly offset) stays the same. 1342 The RTP sessions can be considered independent, resulting in that the 1343 SSRC numbers used can also be handled independently. This simplifies 1344 the SSRC collision detection and avoidance, but require tools such as 1345 remapping tables between the RTP sessions. Using independent RTP 1346 sessions are not required, as the switching behavior is possible to 1347 perform also with a common SSRC space. However, in this case 1348 collision detection and handling becomes a different problem. It is 1349 up to the implementation to use a single common SSRC space or 1350 separate ones. 1352 Using separate SSRC spaces has some implications. For example, the 1353 RTP stream that is being sent by endpoint B to the middlebox (BV1) 1354 may use an SSRC value of 12345678. When that RTP stream is sent to 1355 endpoint F by the middlebox, it can use any SSRC value, e.g. 1356 87654321. As a result, each endpoint may have a different view of 1357 the application usage of a particular SSRC. Any RTP level identity 1358 information, such as SDES items also needs to update the SSRC 1359 referenced, if the included SDES items are intended to be global. 1360 Thus the application must not use SSRC as references to RTP streams 1361 when communicating with other peers directly. This also affects loop 1362 detection which will fail to work, as there is no common namespace 1363 and identities across the different legs in the communication session 1364 on RTP level. Instead this responsibility falls onto higher layers. 1366 The middlebox is also responsible for receiving any RTCP codec 1367 control requests coming from an endpoint, and decide if it can act on 1368 the request locally or needs to translate the request into the RTP 1369 session/transport leg that contains the media source. Both endpoints 1370 and the middlebox need to implement conference related codec control 1371 functionalities to provide a good experience. Commonly used are Full 1372 Intra Request to request from the media source to provide switching 1373 points between the sources, and Temporary Maximum Media Bit-rate 1374 Request (TMMBR) to enable the middlebox to aggregate congestion 1375 control responses towards the media source so to enable it to adjust 1376 its bit-rate (obviously only in case the limitation is not in the 1377 source to middlebox link). 1379 The selective forwarding middlebox has been introduced in recently 1380 developed videoconferencing systems in conjunction with, and to 1381 capitalize on, scalable video coding as well as simulcasting. An 1382 example of scalable video coding is Annex G of H.264, but other 1383 codecs, including H.264 AVC and VP8 also exhibit scalability, albeit 1384 only in the temporal dimension. In both scalable coding and 1385 simulcast cases the video signal is represented by a set of two or 1386 more bitstreams, providing a corresponding number of distinct 1387 fidelity points. The middlebox selects which parts of a scalable 1388 bitstream (or which bitstream, in the case of simulcasting) to 1389 forward to each of the receiving endpoints. The decision may be 1390 driven by a number of factors, such as available bit rate, desired 1391 layout, etc. Contrary to transcoding MCUs, these "Selective 1392 Forwarding Units" (SFUs) have extremely low delay, and provide 1393 features that are typically associated with high-end systems 1394 (personalized layout, error localization) without any signal 1395 processing at the middlebox. They are also capable of scaling to a 1396 large number of concurrent users, and--due to their very low delay-- 1397 can also be cascaded. 1399 This version of the middlebox also puts different requirements on the 1400 endpoint when it comes to decoder instances and handling of the RTP 1401 streams providing media. As each projected SSRC can, at any time, 1402 provide media, the endpoint either needs to be able to handle as many 1403 decoder instances as the middlebox received, or have efficient 1404 switching of decoder contexts in a more limited set of actual decoder 1405 instances to cope with the switches. The application also gets more 1406 responsibility to update how the media provided is to be presented to 1407 the user. 1409 Note that this topology could potentially be seen as a media 1410 translator which include an on/off logic as part of its media 1411 translation. The topology has the property that all SSRCs present in 1412 the session are visible to an endpoint. It also has mixer aspects, 1413 as the streams it provides are not basically translated version, but 1414 instead they have conceptual property assigned to them and can be 1415 both turned on/off as well as being fully or partially delivered. 1416 Thus this topology appears to be some hybrid between the translator 1417 and mixer model. 1419 The differences between selective forwarding middlebox and a 1420 switching mixer (Section 3.6.2) are minor, and they share most 1421 properties. The above requirement on having a large number of 1422 decoding instances or requiring efficient switching of decoder 1423 contexts, are one point of difference. The other is how the 1424 identification is performed, where the Mixer uses CSRC to provide 1425 information on what is included in a particular RTP stream that 1426 represent a particular concept. Selective forwarding gets the source 1427 information through the SSRC, and instead uses other mechanisms to 1428 indicate the streams intended usage, if needed. 1430 3.8. Point to Multipoint Using Video Switching MCUs 1432 Shortcut name: Topo-Video-switch-MCU 1434 +---+ +------------+ +---+ 1435 | A |------| Multipoint |------| B | 1436 +---+ | Control | +---+ 1437 | Unit | 1438 +---+ | (MCU) | +---+ 1439 | C |------| |------| D | 1440 +---+ +------------+ +---+ 1442 Figure 18: Point to Multipoint Using a Video Switching MCU 1444 This PtM topology was popular in early implementations of multipoint 1445 videoconferencing systems due to its simplicity, and the 1446 corresponding middlebox design has been known as a "video switching 1447 MCU". The more complex RTCP-terminating MCUs, discussed in the next 1448 section, became the norm, however, when technology allowed 1449 implementations at acceptable costs. 1451 A video switching MCU forwards to a participant a single media 1452 stream, selected from the available streams. The criteria for 1453 selection are often based on voice activity in the audio-visual 1454 conference, but other conference management mechanisms (like 1455 presentation mode or explicit floor control) are known to exist as 1456 well. 1458 The video switching MCU may also perform media translation to modify 1459 the content in bit-rate, encoding, or resolution. However, it still 1460 may indicate the original sender of the content through the SSRC. In 1461 this case, the values of the CC and CSRC fields are retained. 1463 If not terminating RTP, the RTCP Sender Reports are forwarded for the 1464 currently selected sender. All RTCP Receiver Reports are freely 1465 forwarded between the endpoints. In addition, the MCU may also 1466 originate RTCP control traffic in order to control the session and/or 1467 report on status from its viewpoint. 1469 The video switching MCU has most of the attributes of a Translator. 1470 However, its stream selection is a mixing behavior. This behavior 1471 has some RTP and RTCP issues associated with it. The suppression of 1472 all but one RTP stream results in most participants seeing only a 1473 subset of the sent RTP streams at any given time, often a single RTP 1474 stream per conference. Therefore, RTCP Receiver Reports only report 1475 on these RTP streams. Consequently, the endpoints emitting RTP 1476 streams that are not currently forwarded receive a view of the 1477 session that indicates their RTP streams disappear somewhere en 1478 route. This makes the use of RTCP for congestion control, or any 1479 type of quality reporting, very problematic. 1481 To avoid the aforementioned issues, the MCU needs to implement two 1482 features. First, it needs to act as a Mixer (see Section 3.6) and 1483 forward the selected RTP stream under its own SSRC and with the 1484 appropriate CSRC values. Second, the MCU needs to modify the RTCP 1485 RRs it forwards between the domains. As a result, it is recommended 1486 that one implement a centralized video switching conference using a 1487 Mixer according to RFC 3550, instead of the shortcut implementation 1488 described here. 1490 3.9. Point to Multipoint Using RTCP-Terminating MCU 1492 Shortcut name: Topo-RTCP-terminating-MCU 1494 +---+ +------------+ +---+ 1495 | A |<---->| Multipoint |<---->| B | 1496 +---+ | Control | +---+ 1497 | Unit | 1498 +---+ | (MCU) | +---+ 1499 | C |<---->| |<---->| D | 1500 +---+ +------------+ +---+ 1502 Figure 19: Point to Multipoint Using Content Modifying MCUs 1504 In this PtM scenario, each endpoint runs an RTP point-to-point 1505 session between itself and the MCU. This is a very commonly deployed 1506 topology in multipoint video conferencing. The content that the MCU 1507 provides to each participant is either: 1509 a. a selection of the content received from the other endpoints, or 1511 b. the mixed aggregate of what the MCU receives from the other PtP 1512 paths, which are part of the same Communication Session. 1514 In case (a), the MCU may modify the content in terms of bit-rate, 1515 encoding format, or resolution. No explicit RTP mechanism is used to 1516 establish the relationship between the original RTP stream of the 1517 media being sent RTP stream the MCU sends. In other words, the 1518 outgoing RTP streams typically use a different SSRC, and may well use 1519 a different payload type (PT), even if this different PT happens to 1520 be mapped to the same media type. This is a result of the 1521 individually negotiated RTP session for each endpoint. 1523 In case (b), the MCU is the Media Source and generates the Source RTP 1524 Stream as it mixes the received content and then encodes and 1525 packetizes it for transmission to an endpoint. According to RTP 1527 [RFC3550], the SSRC of the contributors are to be signalled using the 1528 CSRC/CC mechanism. In practice, today, most deployed MCUs do not 1529 implement this feature. Instead, the identification of the endpoints 1530 whose content is included in the Mixer's output is not indicated 1531 through any explicit RTP mechanism. That is, most deployed MCUs set 1532 the CSRC Count (CC) field in the RTP header to zero, thereby 1533 indicating no available CSRC information, even if they could identify 1534 the original sending endpoints as suggested in RTP. 1536 The main feature that sets this topology apart from what RFC 3550 1537 describes is the breaking of the common RTP session across the 1538 centralized device, such as the MCU. This results in the loss of 1539 explicit RTP-level indication of all participants. If one were using 1540 the mechanisms available in RTP and RTCP to signal this explicitly, 1541 the topology would follow the approach of an RTP Mixer. The lack of 1542 explicit indication has at least the following potential problems: 1544 1. Loop detection cannot be performed on the RTP level. When 1545 carelessly connecting two misconfigured MCUs, a loop could be 1546 generated. 1548 2. There is no information about active media senders available in 1549 the RTP packet. As this information is missing, receivers cannot 1550 use it. It also deprives the client of information related to 1551 currently active senders in a machine-usable way, thus preventing 1552 clients from indicating currently active speakers in user 1553 interfaces, etc. 1555 Note that many/most deployed MCUs (and video conferencing endpoints) 1556 rely on signalling layer mechanisms for the identification of the 1557 contributing sources, for example, a SIP conferencing package 1558 [RFC4575]. This alleviates, to some extent, the aforementioned 1559 issues resulting from ignoring RTP's CSRC mechanism. 1561 3.10. Split Component Terminal 1563 Shortcut name: Topo-Split-Terminal 1565 In some applications, for example in some telepresence systems, 1566 terminals may be not integrated into a single functional unit, but 1567 composed of more than one subunits. For example, a telepresence room 1568 terminal employing multiple cameras and monitors may consist of 1569 multiple video conferencing subunits, each capable of handling a 1570 single camera and monitor. Another example would be a video 1571 conferencing terminal in which audio is handled by one subunit, and 1572 video by another. Each of these subunits uses its own physical 1573 network interface (for example: Ethernet jack) and network address. 1575 The various (media processing) subunits need (logically and 1576 physically) to be interconnected by control functionality, but their 1577 media plane functionality may be split. This type of terminals is 1578 referred to as split component terminals. Historically, the earliest 1579 split component terminals were perhaps the independent audio and 1580 video conference software tools used over the MBONE in the late 1581 1990s. 1583 An example for such a split component terminal is depicted in 1584 Figure 20. Within split component terminal A, at least audio and 1585 video subunits are addressed by their own network addresses. In some 1586 of these systems, the control stack subunit may also have its own 1587 network address. 1589 From an RTP viewpoint, each of the subunits terminates RTP, and acts 1590 as an endpoint in the sense that each subunit includes its own, 1591 independent RTP stack. However, as the subunits are semantically 1592 part of the same terminal, it is appropriate that this semantic 1593 relationship is expressed in RTCP protocol elements, namely in the 1594 CNAME. 1596 +---------------------+ 1597 | Endpoint A | 1598 | Local Area Network | 1599 | +------------+ | 1600 | +->| Audio |<+-RTP---\ 1601 | | +------------+ | \ +------+ 1602 | | +------------+ | +-->| | 1603 | +->| Video |<+-RTP-------->| B | 1604 | | +------------+ | +-->| | 1605 | | +------------+ | / +------+ 1606 | +->| Control |<+-SIP---/ 1607 | +------------+ | 1608 +---------------------+ 1610 Figure 20: Split Component Terminal 1612 It is further sensible that the subunits share a common clock from 1613 which RTP and RTCP clocks are derived, to facilitate synchronization 1614 and avoid clock drift. 1616 To indicate that audio and video Source Streams generated by 1617 different sub-units share a common clock, and can be synchronized, 1618 the RTP streams generated from those Source Streams need to include 1619 the same CNAME in their RTCP SDES packets. The use of a common CNAME 1620 for RTP flows carried in different transport-layer flows is entirely 1621 normal for RTP and RTCP senders, and fully compliant RTP endpoints, 1622 middle-boxes, and other tools should have no problem with this. 1624 However, outside of the split component terminal scenario (and 1625 perhaps a multi-homed endpoint scenario, which is not further 1626 discussed herein), the use of a common CNAME in RTP streams sent from 1627 separate endpoints (as opposed to a common CNAME for RTP streams sent 1628 on different transport layer flows between two endpoints) is rare. 1629 It has been reported that at least some third party tools like some 1630 network monitors do not handle endpoints that use of a common CNAME 1631 across multiple transport layer flows gracefully: they report an 1632 error condition that two separate endpoints are using the same CNAME. 1633 Depending on the sophistication of the support staff, such erroneous 1634 reports can lead to support issues. 1636 Aforementioned support issue can sometimes be avoided if each of the 1637 subunits of a split component terminal is configured to use a 1638 different CNAME, with the synchronization between the RTP streams 1639 being indicated by some non-RTP signaling channel rather than using a 1640 common CNAME sent in RTCP. This complicates the signaling, 1641 especially in cases where there are multiple SSRCs in use with 1642 complex synchronization requirements, as is the same in many current 1643 telepresence systems. Unless one uses RTCP terminating topologies 1644 such as Topo-RTCP-terminating-MCU, sessions involving more than one 1645 video subunit with a common CNAME are close to unavoidable. 1647 The different RTP streams comprising a split terminal system can form 1648 a single RTP session or they can form multiple RTP sessions, 1649 depending on the visibility of their SSRC values in RTCP reports. If 1650 the receiver of the RTP streams sent by the split terminal sends 1651 reports relating to all of the RTP flows (i.e., to each SSRC) in each 1652 RTCP report then a single RTP session is formed. Alternatively, if 1653 the receiver of the RTP streams sent by the split terminal does not 1654 send cross-reports in RTCP, then the audio and video form separate 1655 RTP sessions. 1657 For example, in the Figure 20, B will send RTCP reports to each of 1658 the sub-units of A. If the RTCP packets that B sends to the audio 1659 sub-unit of A include reports on the reception quality of the video 1660 as well as the audio, and similarly if the RTCP packets that B sends 1661 to the video sub-unit of A include reports on the reception quality 1662 of the audio as well as video, then a single RTP session is formed. 1663 However, if the RTCP packets B sends to the audio sub-unit of A only 1664 report on the received audio, and the RTCP packet B sends to the 1665 video sub-unit of A only report on the received video, then there are 1666 two separate RTP sessions. 1668 Forming a single RTP session across the RTP streams sent by the 1669 different sub-units of a split terminal gives each sub-unit 1670 visibility into reception quality of RTP streams sent by the other 1671 sub-units. This information can help diagnose reception quality 1672 problems, but at the cost of increased RTCP bandwidth use. 1674 RTP streams sent by the sub-units of a split terminal need to use the 1675 same CNAME in their RTCP packets if they are to be synchronized, 1676 irrespective of whether a single RTP session is formed or not. 1678 3.11. Non-Symmetric Mixer/Translators 1680 Shortcut name: Topo-Asymmetric 1682 It is theoretically possible to construct an MCU that is a Mixer in 1683 one direction and a Translator in another. The main reason to 1684 consider this would be to allow topologies similar to Figure 13, 1685 where the Mixer does not need to mix in the direction from B or D 1686 towards the multicast domains with A and C. Instead, the RTP streams 1687 from B and D are forwarded without changes. Avoiding this mixing 1688 would save media processing resources that perform the mixing in 1689 cases where it isn't needed. However, there would still be a need to 1690 mix B's media towards D. Only in the direction B -> multicast domain 1691 or D -> multicast domain would it be possible to work as a 1692 Translator. In all other directions, it would function as a Mixer. 1694 The Mixer/Translator would still need to process and change the RTCP 1695 before forwarding it in the directions of B or D to the multicast 1696 domain. One issue is that A and C do not know about the mixed-media 1697 stream the Mixer sends to either B or D. Therefore, any reports 1698 related to these streams must be removed. Also, receiver reports 1699 related to A and C's RTP streams would be missing. To avoid A and C 1700 thinking that B and D aren't receiving A and C at all, the Mixer 1701 needs to insert locally generated reports reflecting the situation 1702 for the streams from A and C into B and D's Sender Reports. In the 1703 opposite direction, the Receiver Reports from A and C about B's and 1704 D's stream also need to be aggregated into the Mixer's Receiver 1705 Reports sent to B and D. Since B and D only have the Mixer as source 1706 for the stream, all RTCP from A and C must be suppressed by the 1707 Mixer. 1709 This topology is so problematic and it is so easy to get the RTCP 1710 processing wrong, that it is not recommended for implementation. 1712 3.12. Combining Topologies 1714 Topologies can be combined and linked to each other using Mixers or 1715 Translators. However, care must be taken in handling the SSRC/CSRC 1716 space. A Mixer does not forward RTCP from sources in other domains, 1717 but instead generates its own RTCP packets for each domain it mixes 1718 into, including the necessary Source Description (SDES) information 1719 for both the CSRCs and the SSRCs. Thus, in a mixed domain, the only 1720 SSRCs seen will be the ones present in the domain, while there can be 1721 CSRCs from all the domains connected together with a combination of 1722 Mixers and Translators. The combined SSRC and CSRC space is common 1723 over any Translator or Mixer. It is important to facilitate loop 1724 detection, something that is likely to be even more important in 1725 combined topologies due to the mixed behavior between the domains. 1726 Any hybrid, like the Topo-Video-switch-MCU or Topo-Asymmetric, 1727 requires considerable thought on how RTCP is dealt with. 1729 4. Topology Properties 1731 The topologies discussed in Section 3 have different properties. 1732 This section describes these properties. Note that, even if a 1733 certain property is supported within a particular topology concept, 1734 the necessary functionality may be optional to implement. 1736 4.1. All to All Media Transmission 1738 To recapitulate, multicast, and in particular Any Source Multicast 1739 (ASM), provides the functionality that everyone may send to, or 1740 receive from, everyone else within the session. Source-specific 1741 Multicast (SSM) can provide a similar functionality by having anyone 1742 intending to participate as sender to send its media to the SSM 1743 distribution source. The SSM distribution source forwards the media 1744 to all receivers subscribed to the multicast group. Mesh, MCUs, 1745 Mixers, SFMs and Translators may all provide that functionality at 1746 least on some basic level. However, there are some differences in 1747 which type of reachability they provide. 1749 The topologies that comes closest to emulating Any Source IP 1750 Multicast, with all-to-all transmission capabilities, are the 1751 transport Translator function called "relay" in Section 3.5, as well 1752 as the Mesh with joint RTP sessions (Section 3.4). Media 1753 Translators, Mesh with independent RTP Sessions, Mixers, SFUs and the 1754 MCU variants do not provide a fully meshed forwarding on the 1755 transport level; instead, they only allow limited forwarding of 1756 content from the other session participants. 1758 The "all to all media transmission" requires that any media 1759 transmitting endpoint considers the path to the least capable 1760 receiving endpoint. Otherwise, the media transmissions may overload 1761 that path. Therefore, a sending endpoint needs to monitor the path 1762 from itself to any of the receiving endpoints, to detect the 1763 currently least capable receiver, and adapt its sending rate 1764 accordingly. As multiple endpoints may send simultaneously, the 1765 available resources may vary. RTCP's Receiver Reports help 1766 performing this monitoring, at least on a medium time scale. 1768 The resource consumption for performing all to all transmission 1769 varies depending with the topology. Both ASM and SSM have the 1770 benefit that only one copy of each packet traverses a particular 1771 link. Using a relay causes the transmission of one copy of a packet 1772 per endpoint-to-relay path and packet transmitted. However, in most 1773 cases the links carrying the multiple copies will be the ones close 1774 to the relay (which can be assumed to be part of the network 1775 infrastructure with good connectivity to the backbone), rather than 1776 the endpoints (which may be behind slower access links). The Mesh 1777 causes N-1 streams of transmitted packets to traverse the first hop 1778 link from the endpoint, in an N endpoint mesh. How long the 1779 different paths are common, is highly situation dependent. 1781 The transmission of RTCP by design adapts to any changes in the 1782 number of participants due to the transmission algorithm, defined in 1783 the RTP specification [RFC3550], and the extensions in AVPF [RFC4585] 1784 (when applicable). That way, the resources utilized for RTCP stay 1785 within the bounds configured for the session. 1787 4.2. Transport or Media Interoperability 1789 All Translators, Mixers, and RTCP-terminating MCU, and Mesh with 1790 individual RTP sessions, allow changing the media encoding or the 1791 transport to other properties of the other domain, thereby providing 1792 extended interoperability in cases where the endpoints lack a common 1793 set of media codecs and/or transport protocols. Selective Forwarding 1794 Middleboxes can adopt the transport, and (at least) selectively 1795 forward the encoded streams that match a receiving endpoint's 1796 capability. It requires an additional translator to change the media 1797 encoding if the encoded streams do not match the receiving endpoint's 1798 capabilities. 1800 4.3. Per Domain Bit-Rate Adaptation 1802 Endpoints are often connected to each other with a heterogeneous set 1803 of paths. This makes congestion control in a Point to Multipoint set 1804 problematic. For the ASM, SSM, Mesh with common RTP session, and 1805 Transport Relay scenario, each individual sending endpoint has to 1806 adapt to the receiving endpoint behind the least capable path, 1807 yielding suboptimal quality for the endpoints behind the more capable 1808 paths. This is no longer an issue when Media Translators, Mixers, 1809 SFM or MCUs are involved, as each endpoint only needs to adapt to the 1810 slowest path within its own domain. The Translator, Mixer, SFM, or 1811 MCU topologies all require their respective outgoing RTP streams to 1812 adjust the bit-rate, packet-rate, etc., to adapt to the least capable 1813 path in each of the other domains. That way one can avoid lowering 1814 the quality to the least-capable endpoint in all the domains at the 1815 cost (complexity, delay, equipment) of the Mixer, SFM or Translator, 1816 and potentially media sender (multicast/layered encoding and sending 1817 the different representations). 1819 4.4. Aggregation of Media 1821 In the all-to-all media property mentioned above and provided by ASM, 1822 SSM, Mesh with common RTP session, and relay, all simultaneous media 1823 transmissions share the available bit-rate. For endpoints with 1824 limited reception capabilities, this may result in a situation where 1825 even a minimal acceptable media quality cannot be accomplished, 1826 because multiple RTP streams need to share the same resources. One 1827 solution to this problem is to provide for a Mixer, or MCU to 1828 aggregate the multiple RTP streams into a single one, where the 1829 single RTP stream takes up less resources in terms of bit-rate. This 1830 aggregation can be performed according to different methods. Mixing 1831 or selection are two common methods. Selection is almost always 1832 possible and easy to implement. Mixing requires resources in the 1833 mixer, and may be relatively easy and not impairing the quality too 1834 badly (audio) or quite difficult (video tiling, which is not only 1835 computationally complex but also reduces the pixel count per stream, 1836 with corresponding loss in perceptual quality). 1838 4.5. View of All Session Participants 1840 The RTP protocol includes functionality to identify the session 1841 participants through the use of the SSRC and CSRC fields. In 1842 addition, it is capable of carrying some further identity information 1843 about these participants using the RTCP Source Descriptors (SDES). 1844 In topologies that provide a full all-to-all functionality, i.e. ASM, 1845 Mesh with common RTP session, Relay a compliant RTP implementation 1846 offers the functionality directly as specified in RTP. In topologies 1847 that do not offer all-to-all communication, it is necessary that RTCP 1848 is handled correctly in domain bridging function. RTP includes 1849 explicit specification text for Translators and Mixers, and for SFMs 1850 the required functionality can be derived from that text. However, 1851 the MCU described in Section 3.8 cannot offer the full functionality 1852 for session participant identification through RTP means. The 1853 topologies that create independent RTP sessions per endpoint or pair 1854 of endpoints, like Back-to-Back RTP session, MESH with independent 1855 RTP sessions, and the RTCP terminating MCU RTCP terminating MCU 1856 (Section 3.9), with an exception of SFM, do not support RTP based 1857 identification of session participants. In all those cases, other 1858 non-RTP based mechanisms need to be implemented if such knowledge is 1859 required or desirable. When it comes to SFM the SSRC name space is 1860 not necessarily joint, instead identification will require knowledge 1861 of SSRC/CSRC mappings that the SFM performed, see Section 3.7. 1863 4.6. Loop Detection 1865 In complex topologies with multiple interconnected domains, it is 1866 possible to unintentionally form media loops. RTP and RTCP support 1867 detecting such loops, as long as the SSRC and CSRC identities are 1868 maintained and correctly set in forwarded packets. Loop detection 1869 will work in ASM, SSM, Mesh with joint RTP session, and Relay. It is 1870 likely that loop detection works for the video switching MCU 1871 Section 3.8, at least as long as it forwards the RTCP between the 1872 endpoints. However, the Back-to-Back RTP sessions, Mesh with 1873 independent RTP sessions, SFM, will definitely break the loop 1874 detection mechanism. 1876 4.7. Consistency between header extensions and RTCP 1878 Some RTP header extensions have relevance not only end-to-end, but 1879 also hop-to-hop, meaning at least some of the middleboxes in the path 1880 are aware of their potential presence through signaling, intercept 1881 and interpret such header extensions and potentially also rewrite or 1882 generate them. Modern header extensions generally follow "A General 1883 Mechanism for RTP Header Extensions" [RFC5285], which allows for all 1884 of the above. Examples for such header extensions include the mid 1885 (media ID) in [I-D.ietf-mmusic-sdp-bundle-negotiation]. At the time 1886 of writing there was also a proposal for how to include any SDES into 1887 an RTP header extension [I-D.westerlund-avtext-sdes-hdr-ext]. 1889 When such header extensions are in use, any middlebox that 1890 understands it must ensure consistency between the extensions it sees 1891 and/or generates, and the RTCP it receives and generates. For 1892 example, the mid of bundle is sent in an RTP header extension and 1893 also in an RTCP SDES message. This apparent redundancy was 1894 introduced as unaware middleboxes may choose to discard RTP header 1895 extensions. Obviously, inconsistency between the media ID sent in 1896 the RTP header extension and in the RTCP SDES message could lead to 1897 undesirable results, and, therefore, consistency is needed. 1898 Middleboxes unaware of the nature of a header extension, as specified 1899 in [RFC5285], are free to forward or discard header extensions. 1901 5. Comparison of Topologies 1903 The table below attempts to summarize the properties of the different 1904 topologies. The legend to the topology abbreviations are: Topo- 1905 Point-to-Point (PtP), Topo-ASM (ASM), Topo-SSM (SSM), Topo-Trns- 1906 Translator (TT), Topo-Media-Translator (including Transport 1907 Translator) (MT), Topo-Mesh with joint session (MJS), Topo-Mesh with 1908 individual sessions (MIS), Topo-Mixer (Mix), Topo-Asymmetric (ASY), 1909 Topo-Video-switch-MCU (VSM), and Topo-RTCP-terminating-MCU (RTM), 1910 Selective Forwarding Middlebox (SFM). In the table below, Y 1911 indicates Yes or full support, N indicates No support, (Y) indicates 1912 partial support, and N/A indicates not applicable. 1914 Property PtP ASM SSM TT MT MJS MIS Mix ASY VSM RTM SFM 1915 --------------------------------------------------------------------- 1916 All to All media N Y (Y) Y Y Y (Y) (Y) (Y) (Y) (Y) (Y) 1917 Interoperability N/A N N Y Y Y Y Y Y N Y Y 1918 Per Domain Adaptation N/A N N N Y N Y Y Y N Y Y 1919 Aggregation of media N N N N N N N Y (Y) Y Y N 1920 Full Session View Y Y Y Y Y Y N Y Y (Y) N Y 1921 Loop Detection Y Y Y Y Y Y N Y Y (Y) N N 1923 Please note that the Media Translator also includes the transport 1924 Translator functionality. 1926 6. Security Considerations 1928 The use of Mixers, SFMs and Translators has impact on security and 1929 the security functions used. The primary issue is that both Mixers, 1930 SFMs and Translators modify packets, thus preventing the use of 1931 integrity and source authentication, unless they are trusted devices 1932 that take part in the security context, e.g., the device can send 1933 Secure Realtime Transport Protocol (SRTP) and Secure Realtime 1934 Transport Control Protocol (SRTCP) [RFC3711] packets to endpoints in 1935 the Communication Session. If encryption is employed, the media 1936 Translator, SFM and Mixer need to be able to decrypt the media to 1937 perform its function. A transport Translator may be used without 1938 access to the encrypted payload in cases where it translates parts 1939 that are not included in the encryption and integrity protection, for 1940 example, IP address and UDP port numbers in a media stream using SRTP 1941 [RFC3711]. However, in general, the Translator, SFM or Mixer needs 1942 to be part of the signalling context and get the necessary security 1943 associations (e.g., SRTP crypto contexts) established with its RTP 1944 session participants. 1946 Including the Mixer, SFM and Translator in the security context 1947 allows the entity, if subverted or misbehaving, to perform a number 1948 of very serious attacks as it has full access. It can perform all 1949 the attacks possible (see RFC 3550 and any applicable profiles) as if 1950 the media session were not protected at all, while giving the 1951 impression to the human session participants that they are protected. 1953 Transport Translators have no interactions with cryptography that 1954 works above the transport layer, such as SRTP, since that sort of 1955 Translator leaves the RTP header and payload unaltered. Media 1956 Translators, on the other hand, have strong interactions with 1957 cryptography, since they alter the RTP payload. A media Translator 1958 in a session that uses cryptographic protection needs to perform 1959 cryptographic processing to both inbound and outbound packets. 1961 A media Translator may need to use different cryptographic keys for 1962 the inbound and outbound processing. For SRTP, different keys are 1963 required, because an RFC 3550 media Translator leaves the SSRC 1964 unchanged during its packet processing, and SRTP key sharing is only 1965 allowed when distinct SSRCs can be used to protect distinct packet 1966 streams. 1968 When the media Translator uses different keys to process inbound and 1969 outbound packets, each session participant needs to be provided with 1970 the appropriate key, depending on whether they are listening to the 1971 Translator or the original source. (Note that there is an 1972 architectural difference between RTP media translation, in which 1973 participants can rely on the RTP Payload Type field of a packet to 1974 determine appropriate processing, and cryptographically protected 1975 media translation, in which participants must use information that is 1976 not carried in the packet.) 1978 When using security mechanisms with Translators, SFMs and Mixers, it 1979 is possible that the Translator, SFM or Mixer could create different 1980 security associations for the different domains they are working in. 1981 Doing so has some implications: 1983 First, it might weaken security if the Mixer/Translator accepts a 1984 weaker algorithm or key in one domain than in another. Therefore, 1985 care should be taken that appropriately strong security parameters 1986 are negotiated in all domains. In many cases, "appropriate" 1987 translates to "similar" strength. If a key management system does 1988 allow the negotiation of security parameters resulting in a different 1989 strength of the security, then this system should notify the 1990 participants in the other domains about this. 1992 Second, the number of crypto contexts (keys and security related 1993 state) needed (for example, in SRTP [RFC3711]) may vary between 1994 Mixers, SFMs and Translators. A Mixer normally needs to represent 1995 only a single SSRCs per domain and therefore needs to create only one 1996 security association (SRTP crypto context) per domain. In contrast, 1997 a Translator needs one security association per participant it 1998 translates towards, in the opposite domain. Considering Figure 11, 1999 the Translator needs two security associations towards the multicast 2000 domain, one for B and one for D. It may be forced to maintain a set 2001 of totally independent security associations between itself and B and 2002 D respectively, so as to avoid two-time pad occurrences. These 2003 contexts must also be capable of handling all the sources present in 2004 the other domains. Hence, using completely independent security 2005 associations (for certain keying mechanisms) may force a Translator 2006 to handle N*DM keys and related state; where N is the total number of 2007 SSRCs used over all domains and DM is the total number of domains. 2009 The multicast based (ASM and SSM), Relay and Mesh with common RTP 2010 session are all topologies with multiple endpoints that require 2011 shared knowledge about the different crypto contexts for the 2012 endpoints. These multi-party topologies have special requirements on 2013 the key-management as well as the security functions. Specifically 2014 source-authentication in these environments has special requirements. 2016 There exist a number of different mechanisms to provide keys to the 2017 different participants. One example is the choice between group keys 2018 and unique keys per SSRC. The appropriate keying model is impacted 2019 by the topologies one intends to use. The final security properties 2020 are dependent on both the topologies in use and the keying 2021 mechanisms' properties, and need to be considered by the application. 2022 Exactly which mechanisms are used is outside of the scope of this 2023 document. Please review RTP Security Options [RFC7201] to get a 2024 better understanding of most of the available options. 2026 7. IANA Considerations 2028 This document makes no request of IANA. 2030 Note to RFC Editor: this section may be removed on publication as an 2031 RFC. 2033 8. Acknowledgements 2035 The authors would like to thank Mark Baugher, Bo Burman, Ben 2036 Campbell, Umesh Chandra, Alex Eleftheriadis, Roni Even, Ladan Gharai, 2037 Geoff Hunt, Suresh Krishnan, Keith Lantz, Jonathan Lennox, Scarlet 2038 Liuyan, Suhas Nandakumar, Colin Perkins, and Dan Wing for their help 2039 in reviewing and improving this document. 2041 9. References 2043 9.1. Normative References 2045 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 2046 Jacobson, "RTP: A Transport Protocol for Real-Time 2047 Applications", STD 64, RFC 3550, July 2003. 2049 [RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 2050 "Extended RTP Profile for Real-time Transport Control 2051 Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July 2052 2006. 2054 9.2. Informative References 2056 [I-D.ietf-avtcore-rtp-multi-stream-optimisation] 2057 Lennox, J., Westerlund, M., Wu, W., and C. Perkins, 2058 "Sending Multiple Media Streams in a Single RTP Session: 2059 Grouping RTCP Reception Statistics and Other Feedback", 2060 draft-ietf-avtcore-rtp-multi-stream-optimisation-05 (work 2061 in progress), February 2015. 2063 [I-D.ietf-mmusic-sdp-bundle-negotiation] 2064 Holmberg, C., Alvestrand, H., and C. Jennings, 2065 "Negotiating Media Multiplexing Using the Session 2066 Description Protocol (SDP)", draft-ietf-mmusic-sdp-bundle- 2067 negotiation-22 (work in progress), June 2015. 2069 [I-D.westerlund-avtext-sdes-hdr-ext] 2070 Westerlund, M., Even, R., and M. Zanaty, "RTP Header 2071 Extension for RTCP Source Description Items", draft- 2072 westerlund-avtext-sdes-hdr-ext-03 (work in progress), 2073 November 2014. 2075 [RFC1112] Deering, S., "Host extensions for IP multicasting", STD 5, 2076 RFC 1112, August 1989. 2078 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 2079 Address Translator (Traditional NAT)", RFC 3022, January 2080 2001. 2082 [RFC3569] Bhattacharyya, S., "An Overview of Source-Specific 2083 Multicast (SSM)", RFC 3569, July 2003. 2085 [RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K. 2086 Norrman, "The Secure Real-time Transport Protocol (SRTP)", 2087 RFC 3711, March 2004. 2089 [RFC4575] Rosenberg, J., Schulzrinne, H., and O. Levin, "A Session 2090 Initiation Protocol (SIP) Event Package for Conference 2091 State", RFC 4575, August 2006. 2093 [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for 2094 IP", RFC 4607, August 2006. 2096 [RFC5104] Wenger, S., Chandra, U., Westerlund, M., and B. Burman, 2097 "Codec Control Messages in the RTP Audio-Visual Profile 2098 with Feedback (AVPF)", RFC 5104, February 2008. 2100 [RFC5117] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 5117, 2101 January 2008. 2103 [RFC5285] Singer, D. and H. Desineni, "A General Mechanism for RTP 2104 Header Extensions", RFC 5285, July 2008. 2106 [RFC5760] Ott, J., Chesterfield, J., and E. Schooler, "RTP Control 2107 Protocol (RTCP) Extensions for Single-Source Multicast 2108 Sessions with Unicast Feedback", RFC 5760, February 2010. 2110 [RFC5766] Mahy, R., Matthews, P., and J. Rosenberg, "Traversal Using 2111 Relays around NAT (TURN): Relay Extensions to Session 2112 Traversal Utilities for NAT (STUN)", RFC 5766, April 2010. 2114 [RFC6285] Ver Steeg, B., Begen, A., Van Caenegem, T., and Z. Vax, 2115 "Unicast-Based Rapid Acquisition of Multicast RTP 2116 Sessions", RFC 6285, June 2011. 2118 [RFC6465] Ivov, E., Marocco, E., and J. Lennox, "A Real-time 2119 Transport Protocol (RTP) Header Extension for Mixer-to- 2120 Client Audio Level Indication", RFC 6465, December 2011. 2122 [RFC7201] Westerlund, M. and C. Perkins, "Options for Securing RTP 2123 Sessions", RFC 7201, April 2014. 2125 Authors' Addresses 2127 Magnus Westerlund 2128 Ericsson 2129 Farogatan 6 2130 SE-164 80 Kista 2131 Sweden 2133 Phone: +46 10 714 82 87 2134 Email: magnus.westerlund@ericsson.com 2136 Stephan Wenger 2137 Vidyo 2138 433 Hackensack Ave 2139 Hackensack, NJ 07601 2140 USA 2142 Email: stewe@stewe.org