idnits 2.17.1 draft-ietf-avt-rapid-rtp-sync-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.i or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) 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 : ---------------------------------------------------------------------------- == The 'Updates: ' line in the draft header should list only the _numbers_ of the RFCs which will be updated by this document (if approved); it should not include the word 'RFC' in the list. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 10, 2009) is 5433 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-19) exists of draft-ietf-avt-rtcpssm-18 ** Obsolete normative reference: RFC 5285 (ref. '6') (Obsoleted by RFC 8285) ** Obsolete normative reference: RFC 1305 (ref. '7') (Obsoleted by RFC 5905) == Outdated reference: A later version (-27) exists of draft-ietf-avt-rtp-svc-18 == Outdated reference: A later version (-07) exists of draft-ietf-avt-dtls-srtp-05 == Outdated reference: A later version (-22) exists of draft-zimmermann-avt-zrtp-13 -- Obsolete informational reference (is this intentional?): RFC 5117 (ref. '15') (Obsoleted by RFC 7667) Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group C. Perkins 3 Internet-Draft University of Glasgow 4 Updates: RFC3550 T. Schierl 5 (if approved) Fraunhofer HHI 6 Intended status: Standards Track June 10, 2009 7 Expires: December 12, 2009 9 Rapid Synchronisation of RTP Flows 10 draft-ietf-avt-rapid-rtp-sync-02.txt 12 Status of this Memo 14 This Internet-Draft is submitted to IETF in full conformance with the 15 provisions of BCP 78 and BCP 79. 17 Internet-Drafts are working documents of the Internet Engineering 18 Task Force (IETF), its areas, and its working groups. Note that 19 other groups may also distribute working documents as Internet- 20 Drafts. 22 Internet-Drafts are draft documents valid for a maximum of six months 23 and may be updated, replaced, or obsoleted by other documents at any 24 time. It is inappropriate to use Internet-Drafts as reference 25 material or to cite them other than as "work in progress." 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/ietf/1id-abstracts.txt. 30 The list of Internet-Draft Shadow Directories can be accessed at 31 http://www.ietf.org/shadow.html. 33 This Internet-Draft will expire on December 12, 2009. 35 Copyright Notice 37 Copyright (c) 2009 IETF Trust and the persons identified as the 38 document authors. All rights reserved. 40 This document is subject to BCP 78 and the IETF Trust's Legal 41 Provisions Relating to IETF Documents in effect on the date of 42 publication of this document (http://trustee.ietf.org/license-info). 43 Please review these documents carefully, as they describe your rights 44 and restrictions with respect to this document. 46 Abstract 48 This memo outlines how RTP sessions are synchronised, and discusses 49 how rapidly such synchronisation can occur. We show that most RTP 50 sessions can be synchronised immediately, but that the use of video 51 switching multipoint conference units (MCUs) or large source specific 52 multicast (SSM) groups can greatly increase the synchronisation 53 delay. This increase in delay can be unacceptable to some 54 applications that use layered and/or multi-description codecs. 56 This memo introduces three mechanisms to reduce the synchronisation 57 delay for such sessions. First, it updates the RTP Control Protocol 58 (RTCP) timing rules to reduce the initial synchronisation delay for 59 SSM sessions. Second, a new feedback packet is defined for use with 60 the Extended RTP Profile for RTCP-based Feedback (RTP/AVPF), allowing 61 video switching MCUs to rapidly request resynchronisation. Finally, 62 new RTP header extensions are defined to allow rapid synchronisation 63 of late joiners, and guarantee correct timestamp based decoding order 64 recovery for layered codecs in the presence of clock skew. 66 Table of Contents 68 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 69 2. Synchronisation of RTP Flows . . . . . . . . . . . . . . . . . 5 70 2.1. Initial Synchronisation Delay . . . . . . . . . . . . . . 5 71 2.1.1. Unicast Sessions . . . . . . . . . . . . . . . . . . . 6 72 2.1.2. Source Specific Multicast (SSM) Sessions . . . . . . . 6 73 2.1.3. Any Source Multicast (ASM) Sessions . . . . . . . . . 7 74 2.1.4. Discussion . . . . . . . . . . . . . . . . . . . . . . 9 75 2.2. Synchronisation for Late Joiners . . . . . . . . . . . . . 9 76 3. Reducing RTP Synchronisation Delays . . . . . . . . . . . . . 10 77 3.1. Reduced Initial RTCP Interval for SSM Senders . . . . . . 10 78 3.2. Rapid Resynchronisation Request . . . . . . . . . . . . . 10 79 3.3. In-band Delivery of Synchronisation Metadata . . . . . . . 11 80 4. Application to Decoding Order Recovery in Layered Codecs . . . 13 81 4.1. Problem description . . . . . . . . . . . . . . . . . . . 13 82 4.2. In-band Synchronisation for Decoding Order Recovery . . . 14 83 4.3. Timestamp based decoding order recovery . . . . . . . . . 15 84 5. Security Considerations . . . . . . . . . . . . . . . . . . . 19 85 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 19 86 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 20 87 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 20 88 8.1. Normative References . . . . . . . . . . . . . . . . . . . 20 89 8.2. Informative References . . . . . . . . . . . . . . . . . . 21 90 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 22 92 1. Introduction 94 When using RTP to deliver multimedia content it's often necessary to 95 synchronise playout of audio and video components of a presentation. 96 This is achieved using information contained in RTP Control Protocol 97 (RTCP) Sender Report (SR) packets [1]. These are sent periodically, 98 and the components of a multimedia session cannot be synchronised 99 until sufficient RTCP SR packets have been received for each RTP flow 100 to allow the receiver to establish mappings between the media clock 101 used for each RTP flow, and the common (NTP-format) reference clock 102 used to establish synchronisation. 104 Recently, concern has been expressed that this synchronisation delay 105 is problematic for some applications, for example those using layered 106 or multi-description video coding. This memo reviews the operations 107 of RTP synchronisation, and describes the synchronisation delay that 108 can be expected. Three backwards compatible extensions to the basic 109 RTP synchronisation mechanism are proposed: 111 o The RTCP transmission timing rules are relaxed for SSM senders, to 112 reduce the initial synchronisation latency for large SSM groups. 113 See Section 3.1. 115 o An enhancement to the Extended RTP Profile for RTCP-based Feedback 116 (RTP/AVPF) [2] is defined to allow receivers to request additional 117 RTCP SR packets, providing the metadata needed to synchronise RTP 118 flows. This can reduce the synchronisation delay when joining 119 sessions with large RTCP reporting intervals, in the presence of 120 packet loss, or when video switching MCUs are employed. See 121 Section 3.2. 123 o Two RTP header extensions are defined, to deliver synchronisation 124 metadata in-band with RTP data packets. These extensions provide 125 synchronisation metadata that is aligned with RTP data packets, 126 and so eliminate the need to estimate clock-skew between flows 127 before synchronisation. They can also reduce the need to receive 128 RTCP SR packets before flows can be synchronising, although it 129 does not eliminate the need for RTCP. See Section 3.3. 131 The immediate use-case for these extensions is to reduce the delay 132 due to synchronisation when joining a layered video session (e.g. an 133 H.264/SVC session in NI-T mode [9]). The extensions are not specific 134 to layered coding, however, and can be used in any environment when 135 synchronisation latency is an issue. 137 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 138 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 139 document are to be interpreted as described in RFC 2119 [3]. 141 2. Synchronisation of RTP Flows 143 RTP flows are synchronised by receivers based on information that is 144 contained in RTCP SR packets generated by senders (specifically, the 145 NTP and RTP timestamps). For multimedia sessions, each type of media 146 (e.g. audio or video) is sent in a separate RTP session, and the 147 receiver associates RTP flows to be synchronised by means of the 148 canonical end-point identifier (CNAME) item included in the RTCP 149 Source Description (SDES) packets generated by the sender or 150 signalled out of band [10]. For layered media, different layers can 151 be sent in different RTP sessions, or using different SSRC values 152 within a single RTP session; in both cases, the CNAME is used to 153 identify flows to be synchronised. To ensure synchronisation, an RTP 154 sender MUST therefore send periodic compound RTCP packets following 155 Section 6 of RFC 3550 [1]. 157 The timing of these periodic compound RTCP packets will depend on the 158 number of members in each RTP session, the fraction of those that are 159 sending data, the session bandwidth, the configured RTCP bandwidth 160 fraction, and whether the session is multicast or unicast (see RFC 161 3550 Section 6.2 for details). In summary, RTCP control traffic is 162 allocated a small fraction, generally 5%, of the session bandwidth, 163 and of that fraction, one quarter is allocated to active RTP senders, 164 while receivers use the remaining three quarters (these fractions can 165 be configured via SDP [11]). Each member of an RTP session derives 166 an RTCP reporting interval based on these fractions, whether the 167 session is multicast or unicast, the number of members it has 168 observed, and whether it is actively sending data or not. It then 169 sends a compound RTCP packet on average once per reporting interval 170 (the actual packet transmission time is randomised in the range [0.5 171 ... 1.5] times the reporting interval to avoid synchronisation of 172 reports). 174 A minimum reporting interval of 5 seconds is RECOMMENDED, except that 175 the delay before sending the initial report "MAY be set to half the 176 minimum interval to allow quicker notification that the new 177 participant is present" [1]. Also, for unicast sessions, "the delay 178 before sending the initial compound RTCP packet MAY be zero" [1]. In 179 addition, for unicast sessions, and for active senders in a multicast 180 session, the fixed minimum reporting interval MAY be scaled to "360 181 divided by the session bandwidth in kilobits/second. This minimum is 182 smaller than 5 seconds for bandwidths greater than 72 kb/s." [1] 184 2.1. Initial Synchronisation Delay 186 A multimedia session comprises a set of concurrent RTP sessions among 187 a common group of participants, using one RTP session for each media 188 type. For example, a videoconference (which is a multimedia session) 189 might contain an audio RTP session and a video RTP session. To allow 190 a receiver to synchronise the components of a multimedia session, a 191 compound RTCP packet containing an RTCP SR packet and an RTCP SDES 192 packet with a CNAME item MUST be sent to each of the RTP sessions in 193 the multimedia session. A receiver cannot synchronise playout across 194 the multimedia session until such RTCP packets have been received on 195 all of the component RTP sessions. If there is no packet loss, this 196 gives an expected initial synchronisation delay equal to the average 197 time taken to receive the first RTCP packet in the RTP session with 198 the longest RTCP reporting interval. This will vary between unicast 199 and multicast RTP sessions. 201 The initial synchronisation delay for layered sessions is similar to 202 that for multimedia sessions. The layers cannot be synchronised 203 until the RTCP SR and CNAME information has been received for each 204 layer in the session. 206 2.1.1. Unicast Sessions 208 For unicast multimedia or layered sessions, senders SHOULD transmit 209 an initial compound RTCP packet (containing an RTCP SR packet and an 210 RTCP SDES packet with a CNAME item) immediately on joining each RTP 211 session in the multimedia session. The individual RTP sessions are 212 considered to be joined once any in-band signalling for NAT traversal 213 (e.g. [12]) and/or security keying (e.g. [13],[14]) has concluded, 214 and the media path is open. This implies that the initial RTCP 215 packet is sent in parallel with the first data packet following the 216 guidance in RFC 3550 that "the delay before sending the initial 217 compound RTCP packet MAY be zero" and, in the absence of any packet 218 loss, flows can be synchronised immediately. 220 Note that NAT pinholes, firewall holes, quality-of-service, and media 221 security keys should have been negotiated as part of the signalling, 222 whether in-band or out-of-band, before the first RTCP packet is sent. 223 This should ensure that any middleboxes are ready to accept traffic, 224 and reduce the likelihood that the initial RTCP packet will be lost. 226 2.1.2. Source Specific Multicast (SSM) Sessions 228 For multicast sessions, the delay before sending the initial RTCP 229 packet, and hence the synchronisation delay, varies with the session 230 bandwidth and the number of members in the session. For a multicast 231 multimedia or layered session, the average synchronisation delay will 232 depend on the slowest of the component RTP sessions; this will 233 generally be the session with the lowest bandwidth (assuming all the 234 RTP sessions have the same number of members). 236 When sending to a multicast group, the reduced minimum RTCP reporting 237 interval of 360 seconds divided by the session bandwidth in kilobits 238 per second [1] should be used when synchronisation latency is likely 239 to be an issue. Also, as usual, the reporting interval is halved for 240 the first RTCP packet. Depending on the session bandwidth and the 241 number of members, this gives the average synchronisation delays 242 shown in Figure 1. 244 Session| Number of receivers: 245 Bandwidth| 2 3 4 5 10 100 1000 10000 246 --+------------------------------------------------ 247 8 kbps| 2.73 4.10 5.47 5.47 5.47 5.47 5.47 5.47 248 16 kbps| 2.50 2.50 2.73 2.73 2.73 2.73 2.73 2.73 249 32 kbps| 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50 250 64 kbps| 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50 251 128 kbps| 1.41 1.41 1.41 1.41 1.41 1.41 1.41 1.41 252 256 kbps| 0.70 0.70 0.70 0.70 0.70 0.70 0.70 0.70 253 512 kbps| 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35 254 1 Mbps| 0.18 0.18 0.18 0.18 0.18 0.18 0.18 0.18 255 2 Mbps| 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 256 4 Mbps| 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 258 Figure 1: Average RTCP reporting interval in seconds for an RTP 259 Session with 1 sender. 261 These numbers assume a source specific multicast channel with a 262 single active sender, which the rules in RFC 3550 section 6.3 give a 263 fixed fraction of the RTCP bandwidth irrespective of the number of 264 receivers. It can be seen that they are sufficient for lip- 265 synchronisation without excessive delay, but might be viewed as 266 having too much latency for synchronising parts of a layered video 267 stream. 269 The RTCP interval is randomised in the usual manner, so the minimum 270 synchronisation delay will be half these intervals, and the maximum 271 delay will be 1.5 times these intervals. Note also that these RTCP 272 intervals are calculated assuming perfect knowledge of the number of 273 members in the session. 275 2.1.3. Any Source Multicast (ASM) Sessions 277 For ASM sessions, the fraction of members that are senders plays an 278 important role, and causes more variation in average RTCP reporting 279 interval. This is illustrated in Figure 2 and Figure 3, which show 280 the RTCP reporting interval for the same session bandwidths and 281 receiver populations as the SSM session described in Figure 1, but 282 for sessions with 2 and 10 senders respectively. It can be seen that 283 the initial synchronisation delay scales with the number of senders 284 (this is to ensure that the total RTCP traffic from all group members 285 does not grow without bound) and can be significantly larger than for 286 single source groups. Despite this, the initial synchronisation time 287 remains acceptable for lip-synchronisation in typical small-to-medium 288 sized group conferencing scenarios. 290 Session| Number of receivers: 291 Bandwidth| 2 3 4 5 10 100 1000 10000 292 --+------------------------------------------------ 293 8 kbps| 2.73 4.10 5.47 6.84 10.94 10.94 10.94 10.94 294 16 kbps| 2.50 2.50 2.73 3.42 5.47 5.47 5.47 5.47 295 32 kbps| 2.50 2.50 2.50 2.50 2.73 2.73 2.73 2.73 296 64 kbps| 2.50 2.50 2.50 2.50 2.50 2.50 2.50 2.50 297 128 kbps| 1.41 1.41 1.41 1.41 1.41 1.41 1.41 1.41 298 256 kbps| 0.70 0.70 0.70 0.70 0.70 0.70 0.70 0.70 299 512 kbps| 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35 300 1 Mbps| 0.18 0.18 0.18 0.18 0.18 0.18 0.18 0.18 301 2 Mbps| 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 302 4 Mbps| 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 304 Figure 2: Average RTCP reporting interval in seconds for an RTP 305 Session with 2 senders. 307 Session| Number of receivers: 308 Bandwidth| 2 3 4 5 10 100 1000 10000 309 --+------------------------------------------------ 310 8 kbps| 2.73 4.10 5.47 6.84 13.67 54.69 54.69 54.69 311 16 kbps| 2.50 2.50 2.73 3.42 6.84 27.34 27.34 27.34 312 32 kbps| 2.50 2.50 2.50 2.50 3.42 13.67 13.67 13.67 313 64 kbps| 2.50 2.50 2.50 2.50 2.50 6.84 6.84 6.84 314 128 kbps| 1.41 1.41 1.41 1.41 1.41 3.42 3.42 3.42 315 256 kbps| 0.70 0.70 0.70 0.70 0.70 1.71 1.71 1.71 316 512 kbps| 0.35 0.35 0.35 0.35 0.35 0.85 0.85 0.85 317 1 Mbps| 0.18 0.18 0.18 0.18 0.18 0.43 0.43 0.43 318 2 Mbps| 0.09 0.09 0.09 0.09 0.09 0.21 0.21 0.21 319 4 Mbps| 0.04 0.04 0.04 0.04 0.04 0.11 0.11 0.11 321 Figure 3: Average RTCP reporting interval in seconds for an RTP 322 Session with 10 senders. 324 Note that multi-sender groups implemented using multi-unicast with a 325 central RTP translator (Topo-Translator in the terminology of [15]) 326 or mixer (Topo-Mixer), or some forms of video switching MCU (Topo- 327 Video-switch-MCU) distribute RTCP packets to all members of the 328 group, and so scale in the same way as an ASM group with regards to 329 initial synchronisation latency. 331 2.1.4. Discussion 333 For unicast sessions, the existing RTCP SR-based mechanism allows for 334 immediate synchronisation, provided the initial RTCP packet is not 335 lost. 337 For SSM sessions, the initial synchronisation delay is sufficient for 338 lip-synchronisation, but may be larger than desired for some layered 339 codecs. The rationale for not sending immediate RTCP packets for 340 multicast groups is to avoid implosion of requests when large numbers 341 of members simultaneously join the group ("flash crowd"). This is 342 not an issue for SSM senders, since there can be at most one sender, 343 so it is desirable to allow SSM senders to send an immediate RTCP SR 344 on joining a session (as is currently allowed for unicast sessions, 345 which also don't suffer from the implosion problem). SSM receivers 346 using unicast feedback would not be allowed to send immediate RTCP. 347 For ASM sessions, implosion of responses is a concern, so no change 348 is proposed to the RTCP timing rules. 350 In all cases, it is possible that the initial RTCP SR packet is lost. 351 In this case, the receiver will not be able to synchronise the media 352 until the reporting interval has passed, and the next RTCP SR packet 353 is sent. This is undesirable. Section 3.2 defines a new RTP/AVPF 354 transport layer feedback message to request an RTCP SR be generated, 355 allowing rapid resynchronisation in the case of packet loss. 357 2.2. Synchronisation for Late Joiners 359 Synchronisation between RTP sessions is potentially slower for late 360 joiners than for participants present at the start of the session. 361 The reasons for this are two-fold: 363 1. Many of the optimisations that allow rapid transmission of RTCP 364 SR packets apply only at the start of a session. This implies 365 that a new participant may have to wait a complete RTCP reporting 366 interval for each session before receiving the necessary data to 367 synchronise media streams. This might potentially take several 368 seconds, depending on the configured session bandwidth and the 369 number of participants. 371 2. Additional synchronisation delay comes from the nature of the 372 RTCP timing rules. Packets are generated on average once per 373 reporting interval, but with the exact transmission times being 374 randomised +/- 50% to avoid synchronisation of reports. This is 375 important to avoid network congestion in multicast sessions, but 376 does mean that the timing of RTCP SR reports for different RTP 377 sessions isn't synchronised. Accordingly, a receiver must 378 estimate the skew on the NTP-format clock in order to align RTP 379 timestamps across sessions. This estimation is an essential part 380 of an RTP synchronisation implementation, and can be done with 381 high accuracy given sufficient reports. Collecting sufficient 382 RTCP SR data to perform this estimation, however, may require 383 reception of several RTCP reports, further increasing the 384 synchronisation delay. 386 These delays are likely an issue for tuning in to an ongoing 387 multicast RTP session, or for video switching MCUs. 389 3. Reducing RTP Synchronisation Delays 391 Three backwards compatible RTP extensions are defined to reduce the 392 possible synchronisation delay: a reduced initial RTCP interval for 393 SSM senders, a rapid resynchronisation request message, and RTP 394 header extensions that can convey synchronisation metadata in-band. 396 3.1. Reduced Initial RTCP Interval for SSM Senders 398 In SSM sessions where the initial synchronisation delay is important, 399 the RTP sender MAY set the delay before sending the initial compound 400 RTCP packet to zero, and send its first RTCP packet immediately upon 401 joining the SSM session. RTP receivers in an SSM session, sending 402 unicast RTCP feedback, MUST NOT send RTCP packets with zero initial 403 delay; the timing rules defined in [4] apply unchanged to receivers. 405 3.2. Rapid Resynchronisation Request 407 The general format of an RTP/AVPF transport layer feedback message is 408 shown in Figure 4 (see [2] for details). 410 0 1 2 3 411 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 412 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 413 |V=2|P| FMT | PT=RTPFB=205 | length | 414 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 415 | SSRC of packet sender | 416 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 417 | SSRC of media source | 418 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 419 : Feedback Control Information (FCI) : 420 : : 422 Figure 4: RTP/AVP Transport Layer Feedback Message 424 A new feedback message type, RTCP-SR-REQ, is defined with FMT = 5. 426 The Feedback Control Information (FCI) part of the feedback message 427 MUST be empty. The SSRC of packet sender indicates the member that 428 is unable to synchronise media streams, while the SSRC of media 429 source indicates the sender of the media it is unable to synchronise. 430 The length MUST equal 2. 432 This feedback message MAY be sent by a receiver to indicate that it's 433 unable to synchronise some media streams, and desires that the media 434 source transmit an RTCP SR packet as soon as possible (within the 435 constraints of the RTCP timing rules for early feedback). When it 436 receives such an indication, the media source SHOULD generate an RTCP 437 SR packet as soon as possible within the RTCP early feedback rules. 438 If the use of non-compound RTCP [5] was previously negotiated, both 439 the feedback request and the RTCP SR response may be sent as non- 440 compound RTCP packets. The RTCP-SR-REQ packet MAY be repeated once 441 per RTCP reporting interval if no RTCP SR packet is forthcoming. 443 When using SSM sessions with unicast feedback, is possible that the 444 feedback target and media source are not co-located. If a feedback 445 target receives an RTCP-SR-REQ feedback message in such a case, the 446 request should be forwarded to the media source. The mechanism to be 447 used for forwarding such requests is not defined here. 449 3.3. In-band Delivery of Synchronisation Metadata 451 The RTP header extension mechanism defined in [6] can be adopted to 452 carry an OPTIONAL NTP format wall clock timestamp in RTP data 453 packets. If such a timestamp is included, it MUST correspond to the 454 same time instant as the RTP timestamp in the packet's header, and 455 MUST be derived from the same clock used to generate the NTP format 456 timestamps included in RTCP SR packets. Provided it has knowledge of 457 the SSRC to CNAME mapping, either from prior receipt of an RTCP CNAME 458 packet or via out-of-band signalling [10], the receiver can use the 459 information provided as input to the synchronisation algorithm, in 460 exactly the same way as if an additional RTCP SR packet was been 461 received for the flow. 463 Two variants are defined for this header extension. The first 464 variant extends the RTP header with a 64 bit NTP timestamp format 465 timestamp as defined in [7]. The second variant carries the lower 24 466 bit part of the Seconds of a NTP timestamp format timestamp and the 467 32 bit of the Fraction of a NTP timestamp format timestamp. The 468 formats of the two variants are shown below. 470 Variant A/64-bit NTP RTP header extension (length: 16 bytes): 472 0 1 2 3 473 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 474 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 475 |V=2|P|1| CC |M| PT | sequence number | 476 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+R 477 | timestamp |T 478 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+P 479 | synchronization source (SSRC) identifier | 480 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 481 | 0xBE | 0xDE | length=3 | 482 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+E 483 | ID-A | L=7 | NTP timestamp format - Seconds (bit 0-23) |x 484 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+t 485 |NTP Sec.(24-31)| NTP timestamp format - Fraction(bit 0-23) |n 486 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 487 |NTP Frc.(24-31)| 0 (pad) | 0 (pad) | 0 (pad) | 488 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 489 | payload data | 490 | .... | 491 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 493 Variant B/56-bit NTP RTP header extension (length: 12 bytes): 495 0 1 2 3 496 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 497 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 498 |V=2|P|1| CC |M| PT | sequence number | 499 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+R 500 | timestamp |T 501 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+P 502 | synchronization source (SSRC) identifier | 503 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 504 | 0xBE | 0xDE | length=2 | 505 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+E 506 | ID-B | L=6 | NTP timestamp format - Seconds (bit 8-31) |x 507 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+t 508 | NTP timestamp format - Fraction (bit 0-31) |n 509 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 510 | payload data | 511 | .... | 512 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 514 An NTP timestamp format timestamp MAY be included on any RTP packets 515 the sender chooses, but it is RECOMMENDED when performing timestamp 516 based decoding order recovery for layered codecs transported in 517 multiple RTP flows, as further specified in Section 4.2. This header 518 extension MAY be also sent on the RTP packets corresponding to a 519 video random access point, and on the associated audio packets, to 520 allow rapid synchronisation for late joiners in multimedia sessions, 521 and in video switching scenarios. 523 Note: The inclusion of an RTP header extension will reduce the 524 efficiency of RTP header compression, if it is used. Furthermore, 525 middle boxes which do not understand the header extensions may remove 526 them or may not update the content according to this memo. 528 In all cases, irrespective of whether in-band NTP timestamp format 529 timestamps are included or not, regular RTCP SR packets MUST be sent 530 to provide backwards compatibility with receivers that synchronize 531 RTP flows according to [1], and robustness in the face of middleboxes 532 (RTP translators) that might strip RTP header extensions. The sender 533 reports are also required to receive the upper 8 bit of the Seconds 534 of the NTP timestamp format timestamp not included in the Variant 535 B/56-bit NTP RTP header extension (although this may generally be 536 inferred from context). 538 When the SDP is used, the use of the RTP header extensions defined 539 above MUST be indicated as specified in [6]. Therefore the following 540 URIs MUST be used: 542 o The URI used for signaling the use of Variant A/64-bit NTP RTP 543 header extension in SDP is "urn:ietf:params:rtp-hdrext:ntp-64". 545 o The URI used for signaling the use of Variant B/56-bit NTP RTP 546 header extension in SDP is "urn:ietf:params:rtp-hdrext:ntp-56". 548 4. Application to Decoding Order Recovery in Layered Codecs 550 Based on the timestamp contained in each RTP data packet, and the 551 mapping to an NTP format wallclock time, a decoding order recovery 552 process may be applied if a media as result of a layered coding 553 process is transported in multiple RTP flows. This recovers the 554 decoding order of media frames or samples at the receiver. 555 Especially when transporting layered video, the decoding order 556 recovery process is not straight forward. In this section, we 557 provide guidance on how to use RTP/NTP timing information for 558 decoding order recovery. 560 4.1. Problem description 562 One option for decoding order recovery in layered codecs is to use 563 the NTP/sample presentation timestamps to reorder data of the same 564 layered media transported in multiple RTP flows. For a timestamp- 565 based decoding order recovery process, it is crucial to allow exact 566 alignment of media frames respectively samples using the NTP timing 567 information. 569 In the presence of clock skew in used clock for wallclock timestamp 570 generation, it may not be possible to derive exact matching NTP 571 timestamps using the NTP format wallclock in each RTP flow's RTCP 572 sender reports. This is due to the fact that RTCP sender reports are 573 not sent at the same point of time in the multiple RTP flows 574 transporting data of the same layered media, while having a skew 575 between those wallclock samples in the RTP flows RTCP sender reports. 576 If the RTCP SR packets are not send synchronously in the multiple RTP 577 flows, they therefore do not contain the same NTP wallclock 578 timestamp. If there is a skew present in the clock used for NTP 579 wallclock timestamp generation, using different wallclock timestamps 580 for the same sampling instance in the RTP flow inevitably leads to 581 non-matching NTP timestamps generated from RTP timestamps and 582 wallclock timestamp in the multiple RTP flows. In order to allow a 583 common and straight forward timestamp-based decoding order recovery 584 process, it is important to guarantee exact matching of NTP 585 timestamps. Thus in the presence of non-perfect clocks, which should 586 be the normal case, an additional mechanism SHALL be used. An exact 587 inter-flow alignment of NTP timestamps can be guaranteed, if an RTP 588 header extension containing an NTP timestamp is always inserted at 589 the same timing position in all the RTP flows in question, and if 590 those NTP header extensions are used to update the NTP-RTP relation 591 in all RTP flows at the same point of time. This is called 592 synchronous insertion of RTP header extensions in the following. 594 4.2. In-band Synchronisation for Decoding Order Recovery 596 The RTP header extension to convey an NTP timestamp SHOULD be used 597 with a layered, multi-description, or multi-view codec, to provide 598 exact matching of NTP timestamps between layers, descriptions, or 599 views transported in different RTP flows to allow timestamp-based 600 decoding order recovery. If this header extension is inserted for 601 RTP flows transporting samples or parts of samples of the same 602 layered media, it SHALL be included at least once in each of the RTP 603 flows of the same media for the sampling time instance of an 604 insertion of an RTP header extension. Such synchronously inserted 605 RTP header extensions SHALL contain the same NTP format wallclock 606 timestamp. The frequency of inserting the header extensions in the 607 RTP flows is up to the sender, but it should be noticed that higher 608 insertion frequencies obviously lead to higher synchronization 609 frequencies. For use cases where the same clock source has been used 610 to generate the RTP timestamps in the multiple RTP flows, an 611 application MAY rely on the RTP timestamps only for decoding order 612 recovery starting from the point of synchronous insertion of the RTP 613 header extensions containing NTP timestamps. 615 Note: If the decoding order of RTP flows is given by any means (as 616 e.g., for RTP session by mechanism defined in [8]), the NTP timestamp 617 provided by the header extension allows to collect data of the same 618 sample from the RTP flows, forming the sample decoding order. There 619 may be future mechanism to allow indication of dependencies of RTP 620 flows transported as RTP streams using SSRC multiplexing 622 It is RECOMMENDED that the receiver uses for timestamp-based decoding 623 order recovery the NTP timestamps provided in the RTP header 624 extensions only, if such extensions are present for the RTP flows. 625 Section 4.3 gives further details about the timestamp-based decoding 626 order recovery. 628 Note: Using the RTP header extensions described above allows the 629 receiver to find the corresponding sample of the layered media, or 630 parts thereof, in all RTP flows at the instant the RTP header 631 extension is inserted into the flows. This guarantees that any clock 632 skew present in the NTP timestamp generation process based on RTCP 633 sender reports is avoided, and so allows direct comparison of NTP 634 timestamps across multiple RTP flows. Furthermore, this approach 635 solves the possible problem of clock skews identified for the NI-T 636 mode as defined in [9]. To ensure the absence of clock skew, a 637 header extension containing the NTP timestamp MUST be inserted into 638 the RTP flows comprising a layered media stream at the same instant 639 in each RTP flow. This may require the insertion of extra packets in 640 some of the RTP flows, since in layered video codecs not all sampling 641 instances may be present in all the flows. If such a header 642 extension is included in all flows at a sampling time instance, the 643 NTP timestamps for samples following in decoding order the RTP header 644 insertion point can be constructed using the RTP timestamps and 645 identical reference NTP timestamps in the NTP header extension in all 646 RTP flows. It should be noted that the frequency of inserting the 647 RTP header extension containing the NTP timestamp is crucial in 648 presence of clock skew, since the points of insertion may be the only 649 points for a receiver to start the decoding order recovery. 651 4.3. Timestamp based decoding order recovery 653 If parts or complete samples as result of a layered coding process 654 are transported as different RTP flows, i.e. as different RTP 655 streams, and/or as different RTP sessions, a decoding order recovery 656 process is required to reorder the samples or parts of samples 657 received. Such mechanism may be based on the NTP presentation 658 timestamp which can be derived from the RTP timestamp using the NTP 659 wallclock provided in the RTCP sender report packets. 661 In order to guarantee the exact alignment of those derived NTP 662 presentation timestamps, the RTP header extension as defined in this 663 memo in Section 3.3 allows the receiver to start the decoding order 664 recovery before the reception of a RTCP sender report if the RTP 665 header extension is earlier provided in the RTP flow. Using the RTP 666 header extensions may be the only way to allow correct decoding order 667 recovery based on exact matching of NTP timestamps in the presence of 668 clock skew in the clock used for generating the NTP wallclock. 670 Furthermore, some use cases may allow to use synchronously inserted 671 RTP header extensions containing NTP timestamps to align the RTP 672 timestamps of the multiple RTP flows, i.e. use cases where the RTP 673 timestamps of the multiple RTP flows are generated from the same 674 clock source. In such use cases, starting from a synchronous 675 insertion of the RTP header extensions, the application may use the 676 detected difference of RTP random offset values in the multiple 677 sessions to align the media samples of parts of samples. 679 Since typically for layered video codecs as, e.g. SVC [9], the 680 decoding order is not equal to the presentation order of the media 681 samples, media samples or parts of media samples cannot be simply 682 ordered according to the presentation timestamp order. For this 683 reason, if transporting media samples or parts of media samples of a 684 layered, multi-view or multi description codec in different RTP 685 flows, the following rules SHOULD be kept for sending such flows: 687 Note: The following rules are typically kept for layered audio 688 codecs, which allows using the same algorithm for decoding order 689 recovery of audio samples. 691 Terminology: Following the decoding order of RTP flows as described 692 above, an RTP flow containing sample data which is required to be 693 accessed and/or decoded before decoding a second sample data of 694 another RTP flow is called a lower RTP flow with respect to the 695 second RTP flow. A second RTP flow, which requires for the decoding 696 process accessing and/or decoding the sample data of the lower RTP 697 flow is called the higher RTP flow. The lowest RTP flow is the flow, 698 which does not require the presence of any other data. 700 o The decoding order of media samples or part of the media samples 701 transported in different RTP flows SHOULD be derivable by any 702 means. This can be accomplished, e.g. by using the mechanisms 703 defined in [8] if the sample data or parts of the sample data are 704 transported in different RTP sessions or by any other means. 706 o For each two RTP flows the following rules SHOULD be true in order 707 to allow decoding order recovery based on matching NTP timestamps 708 present in the different RTP flows: 710 1. The order of the RTP samples within an RTP flow is equal to 711 the decoding order. 713 2. A higher RTP flow contains all the same sampling instances of 714 the lower RTP flow. A higher RTP flow may contain additional 715 sampling instances. 717 Note: In some cases, it may be required to add packets in higher RTP 718 flows in order to satisfy the second rule above. This may be 719 achieved by placing empty RTP packets (containing padding data only) 720 or by other payload means as, e.g. the Empty NAL unit packet as 721 defined in [9]. 723 If a packet must be inserted for satisfying the above rule, the NTP 724 timestamp of such an inserted packet MUST be set equal to the NTP 725 timestamp of a packet of the same sample present in any lower RTP 726 flow and the lowest RTP flow. This is easy to accomplish if the 727 packet can be inserted at the time of the RTP flow generation, since 728 the NTP timestamp must be the same for the inserted packet and the 729 packet of the corresponding sample. 731 The above rules allow the receiver to process the data of the RTP 732 flows as follows: 734 o Go through all received RTP flows starting with the highest RTP 735 flow and aggregate the sample data or parts of the sample data 736 with the same NTP timestamp in the order of RTP flows, starting 737 from the lowest RTP flow up to the highest RTP flow received, to 738 the sample with the NTP timestamp present in the highest RTP flow. 739 The NTP timestamps MAY be derived using RTCP sender reports or MAY 740 be directly taken from the NTP timestamp provided in an RTP header 741 extension. The order of RTP flows may e.g. be indicated by 742 mechanisms as defined in [8] or any other implicit or explicit 743 means. Repeat the aforementioned process for each different NTP 744 timestamp present in the highest RTP flow. 746 Informative example: The example shown in Figure 3 refers to three 747 RTP flows A, B and C containing a layered, a multi-view or a multi- 748 description media stream. In the example, the dependency signalling 749 as defined in [8] indicates that flow A is the lowest RTP flow, B is 750 the first higher RTP flow and depends on A, and C is the second 751 higher RTP flow corresponding to flow A and depends on A and B. A 752 media coding structure is used that results in samples present in 753 higher flows but not present in all lower flows. Flow A has the 754 lowest frame rate and Flow B and C have the same but higher frame 755 rate. The figure shows the full video samples with their 756 corresponding RTP timestamps "(x)". The video samples are already 757 re-ordered according to their RTP sequence number order. The figure 758 indicates for the received sample in decoding order within each RTP 759 flow, as well as the associated NTP media timestamps ("TS[..]"). 760 These timestamps may be derived using the NTP format wallclock 761 provided in the RTCP sender reports or as shown in the figure 762 directly from the NTP timestamp contained in the RTP header 763 extensions as indicate by the timestamp in "". Note that the 764 timestamps are not in increasing order since, in this example, the 765 decoding order is different from the output/presentation order. 767 The process first proceeds to the sample parts associated with the 768 first available synchronous insertion of NTP timestamp into RTP 769 header extensions at NTP media timestamp TS=[8] and starts in the 770 highest RTP flow C and removes/ignores all preceding sample parts (in 771 decoding order) to sample parts with TS=[8] in each of the de- 772 jittering buffers of RTP flows A, B, and C. Then, starting from flow 773 C, the first media timestamp available in decoding order (TS=[8]) is 774 selected and sample parts starting from RTP flow A, and flow B and C 775 are placed in order of the RTP flow dependency as indicated by 776 mechanisms defined in [8] (in the example for TS[8]: first flow B and 777 then flow C into the video sample VS(TS[8]) associated with NTP media 778 timestamp TS=[8]. Then the next media timestamp TS=[6] (RTP 779 timestamp=(4)) in order of appearance in the highest RTP flow C is 780 processed and the process described above is repeated. Note that 781 there may be video samples with no sample parts present, e.g., in the 782 lowest RTP flow A (see, e.g., TS=[5]). The decoding order recovery 783 process could be also started after receiving all RTP sender reports 784 "RTP"-"wallclock" mapping (indicated as timestamps "(x){y}") assuming 785 that there is no clock skew in the source used for the wallclock 786 generation. 788 C:-(0)----(2)----(7)<8>--(5)----(4)----(6)-----(11)----(9){10}- 789 | | | | | | | | 790 B:-(3)----(5)---(10)<8>--(8)----(7)----(9){7}--(14)----(12)---- 791 | | | | 792 A:---------------(3)<8>--(1)-------------------(7){12}-(5)----- 794 ---------------------------------------decoding/transmission order-> 795 TS:[1] [3] [8]=<8> [6] [5] [7] [12] [10] 797 Key: 798 A, B, C - RTP flows 799 Integer values in "()"- video sample with its RTP timestamp as 800 indicated in its RTP packet. 801 "|" - indicates corresponding samples / parts of 802 sample of the same video sample VS(TS[..]) 803 in the RTP flows. 804 Integer values in "[]"- NTP media timestamp TS, sampling time 805 as derived from the NTP timestamp associated 806 with the video sample AU(TS[..]), consisting 807 of sample parts in the flows above. 808 Integer values in "<>"- NTP media timestamp TS as directly 809 taken from the NTP RTP header extensions. 810 Integer values in "{}"- NTP media timestamp TS as provided in the 811 RTCP sender reports. 813 5. Security Considerations 815 The security considerations of the RTP specification [1], the 816 Extended RTP profile for RTCP-Based Feedback [2], and the General 817 Mechanism for RTP Header Extensions [6] apply. The extensions we 818 define in this memo are not believed to introduce any additional 819 security considerations. 821 6. IANA Considerations 823 NOTE TO RFC EDITOR: Please replace "RFC XXXX" in the following with 824 the RFC number assigned to this memo, and delete this note. 826 The IANA is requested to register one new value in the table of FMT 827 Values for RTPFB Payload Types [2] as follows: 829 Name: RTCP-SR-REQ 830 Long name: RTCP Rapid Resynchronisation Request 831 Value: 5 832 Reference: RFC XXXX 834 The IANA is also requested to register two new RTP Compact Header 835 Extensions [6], according to the following: 837 Extension URI: urn:ietf:params:rtp-hdrext:ntp-64 838 Description: Synchronisation metadata: 64-bit timestamp format 839 Contact: Thomas Schierl 840 IETF Audio/Video Transport Working Group 841 Reference: RFC XXXX 843 Extension URI: urn:ietf:params:rtp-hdrext:ntp-56 844 Description: Synchronisation metadata: 56-bit timestamp format 845 Contact: Thomas Schierl 846 IETF Audio/Video Transport Working Group 847 Reference: RFC XXXX 849 7. Acknowledgements 851 This memo has benefitted from discussions with numerous members of 852 the IETF AVT working group, including Jonathan Lennox, Magnus 853 Westerlund, Randell Jesup, Gerard Babonneau, Ingemar Johansson, Ali 854 Began, Ye-Kui Wang, and Roni Even. The header extension format of 855 Variant A in Section 3.3 was suggested by Dave Singer, matching a 856 similar mechanism specified by ISMA. 858 8. References 860 8.1. Normative References 862 [1] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson, 863 "RTP: A Transport Protocol for Real-Time Applications", STD 64, 864 RFC 3550, July 2003. 866 [2] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 867 "Extended RTP Profile for Real-time Transport Control Protocol 868 (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July 2006. 870 [3] Bradner, S., "Key words for use in RFCs to Indicate Requirement 871 Levels", BCP 14, RFC 2119, March 1997. 873 [4] Schooler, E., Ott, J., and J. Chesterfield, "RTCP Extensions 874 for Single-Source Multicast Sessions with Unicast Feedback", 875 draft-ietf-avt-rtcpssm-18 (work in progress), March 2009. 877 [5] Johansson, I. and M. Westerlund, "Support for Reduced-Size 878 Real-Time Transport Control Protocol (RTCP): Opportunities and 879 Consequences", RFC 5506, April 2009. 881 [6] Singer, D. and H. Desineni, "A General Mechanism for RTP Header 882 Extensions", RFC 5285, July 2008. 884 [7] Mills, D., "Network Time Protocol (Version 3) Specification, 885 Implementation", RFC 1305, March 1992. 887 [8] Schierl, T. and S. Wenger, "Signaling media decoding dependency 888 in Session Description Protocol (SDP)", 889 draft-ietf-mmusic-decoding-dependency-08 (work in progress), 890 April 2009. 892 8.2. Informative References 894 [9] Wenger, S., Wang, Y., Schierl, T., and A. Eleftheriadis, "RTP 895 Payload Format for SVC Video", draft-ietf-avt-rtp-svc-18 (work 896 in progress), March 2009. 898 [10] Lennox, J., Ott, J., and T. Schierl, "Source-Specific Media 899 Attributes in the Session Description Protocol (SDP)", 900 draft-ietf-mmusic-sdp-source-attributes-02 (work in progress), 901 October 2008. 903 [11] Casner, S., "Session Description Protocol (SDP) Bandwidth 904 Modifiers for RTP Control Protocol (RTCP) Bandwidth", RFC 3556, 905 July 2003. 907 [12] Rosenberg, J., "Interactive Connectivity Establishment (ICE): A 908 Protocol for Network Address Translator (NAT) Traversal for 909 Offer/Answer Protocols", draft-ietf-mmusic-ice-19 (work in 910 progress), October 2007. 912 [13] McGrew, D. and E. Rescorla, "Datagram Transport Layer Security 913 (DTLS) Extension to Establish Keys for Secure Real-time 914 Transport Protocol (SRTP)", draft-ietf-avt-dtls-srtp-05 (work 915 in progress), September 2008. 917 [14] Zimmermann, P., Johnston, A., and J. Callas, "ZRTP: Media Path 918 Key Agreement for Secure RTP", draft-zimmermann-avt-zrtp-13 919 (work in progress), January 2009. 921 [15] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 5117, 922 January 2008. 924 Authors' Addresses 926 Colin Perkins 927 University of Glasgow 928 Department of Computing Science 929 Sir Alwyn Williams Building 930 Lilybank Gardens 931 Glasgow G12 8QQ 932 UK 934 Email: csp@csperkins.org 936 Thomas Schierl 937 Fraunhofer HHI 938 Einsteinufer 37 939 D-10587 Berlin 940 Germany 942 Phone: +49-30-31002-227 943 Email: Thomas.Schierl@hhi.fraunhofer.de