idnits 2.17.1 draft-ietf-payload-flexible-fec-scheme-19.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 : ---------------------------------------------------------------------------- -- The document has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: The value for the repair window duration is related to the maximum L and D values that are expected during a FLEX FEC session and therefore cannot be chosen arbitrarily. Repair packets that include L and D values larger than the repair window MUST not be sent. The rate of the source streams should also be considered, as the repair window duration should ideally span several packetization intervals in order to leverage the error correction capabilities of the parity code. -- The document date (March 28, 2019) is 1856 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) == Missing Reference: '0-14' is mentioned on line 721, but not defined == Missing Reference: '15-45' is mentioned on line 723, but not defined == Missing Reference: '46-109' is mentioned on line 725, but not defined == Missing Reference: 'RFCXXXX' is mentioned on line 1059, but not defined ** Obsolete normative reference: RFC 4566 (Obsoleted by RFC 8866) -- Obsolete informational reference (is this intentional?): RFC 2326 (Obsoleted by RFC 7826) -- Obsolete informational reference (is this intentional?): RFC 2733 (Obsoleted by RFC 5109) -- Obsolete informational reference (is this intentional?): RFC 6347 (Obsoleted by RFC 9147) Summary: 1 error (**), 0 flaws (~~), 6 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 PAYLOAD M. Zanaty 3 Internet-Draft Cisco 4 Intended status: Standards Track V. Singh 5 Expires: September 29, 2019 callstats.io 6 A. Begen 7 Networked Media 8 G. Mandyam 9 Qualcomm Inc. 10 March 28, 2019 12 RTP Payload Format for Flexible Forward Error Correction (FEC) 13 draft-ietf-payload-flexible-fec-scheme-19 15 Abstract 17 This document defines new RTP payload formats for the Forward Error 18 Correction (FEC) packets that are generated by the non-interleaved 19 and interleaved parity codes from source media encapsulated in RTP. 20 These parity codes are systematic codes (Flexible FEC, or "FLEX 21 FEC"), where a number of FEC repair packets are generated from a set 22 of source packets from one or more source RTP streams. These FEC 23 repair packets are sent in a redundancy RTP stream separate from the 24 source RTP stream(s) that carries the source packets. RTP source 25 packets that were lost in transmission can be reconstructed using the 26 source and repair packets that were received. The non-interleaved 27 and interleaved parity codes which are defined in this specification 28 offer a good protection against random and bursty packet losses, 29 respectively, at a cost of complexity. The RTP payload formats that 30 are defined in this document address scalability issues experienced 31 with the earlier specifications, and offer several improvements. Due 32 to these changes, the new payload formats are not backward compatible 33 with earlier specifications, but endpoints that do not implement this 34 specification can still work by simply ignoring the FEC repair 35 packets. 37 Status of This Memo 39 This Internet-Draft is submitted in full conformance with the 40 provisions of BCP 78 and BCP 79. 42 Internet-Drafts are working documents of the Internet Engineering 43 Task Force (IETF). Note that other groups may also distribute 44 working documents as Internet-Drafts. The list of current Internet- 45 Drafts is at https://datatracker.ietf.org/drafts/current/. 47 Internet-Drafts are draft documents valid for a maximum of six months 48 and may be updated, replaced, or obsoleted by other documents at any 49 time. It is inappropriate to use Internet-Drafts as reference 50 material or to cite them other than as "work in progress." 52 This Internet-Draft will expire on September 29, 2019. 54 Copyright Notice 56 Copyright (c) 2019 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents 61 (https://trustee.ietf.org/license-info) in effect on the date of 62 publication of this document. Please review these documents 63 carefully, as they describe your rights and restrictions with respect 64 to this document. Code Components extracted from this document must 65 include Simplified BSD License text as described in Section 4.e of 66 the Trust Legal Provisions and are provided without warranty as 67 described in the Simplified BSD License. 69 Table of Contents 71 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 72 1.1. Parity Codes . . . . . . . . . . . . . . . . . . . . . . 4 73 1.1.1. One-Dimensional (1-D) Non-interleaved (Row) FEC 74 Protection . . . . . . . . . . . . . . . . . . . . . 5 75 1.1.2. 1-D Interleaved (Column) FEC Protection . . . . . . . 6 76 1.1.3. Use Cases for 1-D FEC Protection . . . . . . . . . . 7 77 1.1.4. Two-Dimensional (2-D) (Row and Column) FEC Protection 8 78 1.1.5. FEC Protection with Flexible Mask . . . . . . . . . . 10 79 1.1.6. FEC Overhead Considerations . . . . . . . . . . . . . 10 80 1.1.7. FEC Protection with Retransmission . . . . . . . . . 10 81 1.1.8. Repair Window Considerations . . . . . . . . . . . . 10 82 2. Requirements Notation . . . . . . . . . . . . . . . . . . . . 11 83 3. Definitions and Notations . . . . . . . . . . . . . . . . . . 11 84 3.1. Definitions . . . . . . . . . . . . . . . . . . . . . . . 11 85 3.2. Notations . . . . . . . . . . . . . . . . . . . . . . . . 12 86 4. Packet Formats . . . . . . . . . . . . . . . . . . . . . . . 12 87 4.1. Source Packets . . . . . . . . . . . . . . . . . . . . . 12 88 4.2. FEC Repair Packets . . . . . . . . . . . . . . . . . . . 13 89 4.2.1. RTP Header of FEC Repair Packets . . . . . . . . . . 13 90 4.2.2. FEC Header of FEC Repair Packets . . . . . . . . . . 15 91 5. Payload Format Parameters . . . . . . . . . . . . . . . . . . 20 92 5.1. Media Type Registration - Parity Codes . . . . . . . . . 20 93 5.1.1. Registration of audio/flexfec . . . . . . . . . . . . 21 94 5.1.2. Registration of video/flexfec . . . . . . . . . . . . 22 95 5.1.3. Registration of text/flexfec . . . . . . . . . . . . 23 96 5.1.4. Registration of application/flexfec . . . . . . . . . 24 98 5.2. Mapping to SDP Parameters . . . . . . . . . . . . . . . . 25 99 5.2.1. Offer-Answer Model Considerations . . . . . . . . . . 25 100 5.2.2. Declarative Considerations . . . . . . . . . . . . . 26 101 6. Protection and Recovery Procedures - Parity Codes . . . . . . 26 102 6.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 26 103 6.2. Repair Packet Construction . . . . . . . . . . . . . . . 26 104 6.3. Source Packet Reconstruction . . . . . . . . . . . . . . 28 105 6.3.1. Associating the Source and Repair Packets . . . . . . 28 106 6.3.2. Recovering the RTP Header . . . . . . . . . . . . . . 30 107 6.3.3. Recovering the RTP Payload . . . . . . . . . . . . . 31 108 6.3.4. Iterative Decoding Algorithm for the 2-D Parity FEC 109 Protection . . . . . . . . . . . . . . . . . . . . . 31 110 7. Signaling Requirements . . . . . . . . . . . . . . . . . . . 34 111 7.1. SDP Examples . . . . . . . . . . . . . . . . . . . . . . 35 112 7.1.1. Example SDP for Flexible FEC Protection with in-band 113 SSRC mapping . . . . . . . . . . . . . . . . . . . . 35 114 7.1.2. Example SDP for Flexible FEC Protection with explicit 115 signalling in the SDP . . . . . . . . . . . . . . . . 35 116 7.2. On the Use of the RTP Stream Identifier Source 117 Description . . . . . . . . . . . . . . . . . . . . . . . 36 118 8. Congestion Control Considerations . . . . . . . . . . . . . . 36 119 9. Security Considerations . . . . . . . . . . . . . . . . . . . 37 120 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 37 121 11. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 38 122 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 38 123 12.1. Normative References . . . . . . . . . . . . . . . . . . 38 124 12.2. Informative References . . . . . . . . . . . . . . . . . 39 125 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 41 127 1. Introduction 129 This document defines new RTP payload formats for the Forward Error 130 Correction (FEC) that is generated by the non-interleaved and 131 interleaved parity codes from a source media encapsulated in RTP 132 [RFC3550]. The type of the source media protected by these parity 133 codes can be audio, video, text or application. The FEC data are 134 generated according to the media type parameters, which are 135 communicated out-of-band (e.g., in SDP). Furthermore, the 136 associations or relationships between the source and repair RTP 137 streams may be communicated in-band or out-of-band. The in-band 138 mechanism is advantageous when the endpoint is adapting the FEC 139 parameters. The out-of-band mechanism may be preferable when the FEC 140 parameters are fixed. While this document fully defines the use of 141 FEC to protect RTP streams, it also leverages several definitions 142 along with the basic source/repair header description from [RFC6363] 143 in their application to the parity codes defined here. 145 The Redundancy RTP Stream [RFC7656] repair packets proposed in this 146 document protect the Source RTP Stream packets that belong to the 147 same RTP session. 149 The RTP payload formats that are defined in this document address the 150 scalability issues experienced with the formats defined in earlier 151 specifications including [RFC2733], [RFC5109] and [SMPTE2022-1]. 153 1.1. Parity Codes 155 Both the non-interleaved and interleaved parity codes use the 156 eXclusive OR (XOR) operation to generate the repair packets. The 157 following steps take place: 159 1. The sender determines a set of source packets to be protected by 160 FEC based on the media type parameters. 162 2. The sender applies the XOR operation on the source packets to 163 generate the required number of repair packets. 165 3. The sender sends the repair packet(s) along with the source 166 packets, in different RTP streams, to the receiver(s). The 167 repair packets may be sent proactively or on-demand based on RTCP 168 feedback messages such as NACK [RFC4585]. 170 At the receiver side, if all of the source packets are successfully 171 received, there is no need for FEC recovery and the repair packets 172 are discarded. However, if there are missing source packets, the 173 repair packets can be used to recover the missing information. 174 Figure 1 and Figure 2 describe example block diagrams for the 175 systematic parity FEC encoder and decoder, respectively. 177 +------------+ 178 +--+ +--+ +--+ +--+ --> | Systematic | --> +--+ +--+ +--+ +--+ 179 +--+ +--+ +--+ +--+ | Parity FEC | +--+ +--+ +--+ +--+ 180 | Encoder | 181 | (Sender) | --> +==+ +==+ 182 +------------+ +==+ +==+ 184 Source Packet: +--+ Repair Packet: +==+ 185 +--+ +==+ 187 Figure 1: Block diagram for systematic parity FEC encoder 188 +------------+ 189 +--+ X X +--+ --> | Systematic | --> +--+ +--+ +--+ +--+ 190 +--+ +--+ | Parity FEC | +--+ +--+ +--+ +--+ 191 | Decoder | 192 +==+ +==+ --> | (Receiver) | 193 +==+ +==+ +------------+ 195 Source Packet: +--+ Repair Packet: +==+ Lost Packet: X 196 +--+ +==+ 198 Figure 2: Block diagram for systematic parity FEC decoder 200 In Figure 2, it is clear that the FEC repair packets have to be 201 received by the endpoint within a certain amount of time for the FEC 202 recovery process to be useful. The repair window is defined as the 203 time that spans a FEC block, which consists of the source packets and 204 the corresponding repair packets. At the receiver side, the FEC 205 decoder SHOULD buffer source and repair packets at least for the 206 duration of the repair window, to allow all the repair packets to 207 arrive. The FEC decoder can start decoding the already received 208 packets sooner; however, it should not register a FEC decoding 209 failure until it waits at least for the duration of the repair 210 window. 212 1.1.1. One-Dimensional (1-D) Non-interleaved (Row) FEC Protection 214 Consider a group of D x L source packets that have sequence numbers 215 starting from 1 running to D x L, and a repair packet is generated by 216 applying the XOR operation to every L consecutive packets as sketched 217 in Figure 3. This process is referred to as 1-D non-interleaved FEC 218 protection. As a result of this process, D repair packets are 219 generated, which are referred to as non-interleaved (or row) FEC 220 repair packets. In general D and L represent values that describe 221 how packets are grouped together from a depth and length perspective 222 (respectively) when interleaving all D x L source packets. 224 +--------------------------------------------------+ --- +===+ 225 | S_1 S_2 S3 ... S_L | + |XOR| = |R_1| 226 +--------------------------------------------------+ --- +===+ 227 +--------------------------------------------------+ --- +===+ 228 | S_L+1 S_L+2 S_L+3 ... S_2xL | + |XOR| = |R_2| 229 +--------------------------------------------------+ --- +===+ 230 . . . . . . 231 . . . . . . 232 . . . . . . 233 +--------------------------------------------------+ --- +===+ 234 | S_(D-1)xL+1 S_(D-1)xL+2 S_(D-1)xL+3 ... S_DxL | + |XOR| = |R_D| 235 +--------------------------------------------------+ --- +===+ 237 Figure 3: Generating non-interleaved (row) FEC repair packets 239 1.1.2. 1-D Interleaved (Column) FEC Protection 241 If the XOR operation is applied to the group of the source packets 242 whose sequence numbers are L apart from each other, as sketched in 243 Figure 4. In this case the endpoint generates L repair packets. 244 This process is referred to as 1-D interleaved FEC protection, and 245 the resulting L repair packets are referred to as interleaved (or 246 column) FEC repair packets. 248 +-------------+ +-------------+ +-------------+ +-------+ 249 | S_1 | | S_2 | | S3 | ... | S_L | 250 | S_L+1 | | S_L+2 | | S_L+3 | ... | S_2xL | 251 | . | | . | | | | | 252 | . | | . | | | | | 253 | . | | . | | | | | 254 | S_(D-1)xL+1 | | S_(D-1)xL+2 | | S_(D-1)xL+3 | ... | S_DxL | 255 +-------------+ +-------------+ +-------------+ +-------+ 256 + + + + 257 ------------- ------------- ------------- ------- 258 | XOR | | XOR | | XOR | ... | XOR | 259 ------------- ------------- ------------- ------- 260 = = = = 261 +===+ +===+ +===+ +===+ 262 |C_1| |C_2| |C_3| ... |C_L| 263 +===+ +===+ +===+ +===+ 265 Figure 4: Generating interleaved (column) FEC repair packets 267 1.1.3. Use Cases for 1-D FEC Protection 269 A sender may generate one non-interleaved repair packet out of L 270 consecutive source packets or one interleaved repair packet out of D 271 non-consecutive source packets. Regardless of whether the repair 272 packet is a non-interleaved or an interleaved one, it can provide a 273 full recovery of the missing information if there is only one packet 274 missing among the corresponding source packets. This implies that 275 1-D non-interleaved FEC protection performs better when the source 276 packets are randomly lost. However, if the packet losses occur in 277 bursts, 1-D interleaved FEC protection performs better provided that 278 L is chosen large enough, i.e., L-packet duration is not shorter than 279 the observed burst duration. If the sender generates non-interleaved 280 FEC repair packets and a burst loss hits the source packets, the 281 repair operation fails. This is illustrated in Figure 5. 283 +---+ +---+ +===+ 284 | 1 | X X | 4 | |R_1| 285 +---+ +---+ +===+ 287 +---+ +---+ +---+ +---+ +===+ 288 | 5 | | 6 | | 7 | | 8 | |R_2| 289 +---+ +---+ +---+ +---+ +===+ 291 +---+ +---+ +---+ +---+ +===+ 292 | 9 | | 10| | 11| | 12| |R_3| 293 +---+ +---+ +---+ +---+ +===+ 295 Figure 5: Example scenario where 1-D non-interleaved FEC protection 296 fails error recovery (Burst Loss) 298 The sender may generate interleaved FEC repair packets to combat with 299 the bursty packet losses. However, two or more random packet losses 300 may hit the source and repair packets in the same column. In that 301 case, the repair operation fails as well. This is illustrated in 302 Figure 6. Note that it is possible that two burst losses may occur 303 back-to-back, in which case interleaved FEC repair packets may still 304 fail to recover the lost data. 306 +---+ +---+ +---+ 307 | 1 | X | 3 | | 4 | 308 +---+ +---+ +---+ 310 +---+ +---+ +---+ 311 | 5 | X | 7 | | 8 | 312 +---+ +---+ +---+ 314 +---+ +---+ +---+ +---+ 315 | 9 | | 10| | 11| | 12| 316 +---+ +---+ +---+ +---+ 318 +===+ +===+ +===+ +===+ 319 |C_1| |C_2| |C_3| |C_4| 320 +===+ +===+ +===+ +===+ 322 Figure 6: Example scenario where 1-D interleaved FEC protection fails 323 error recovery (Periodic Loss) 325 1.1.4. Two-Dimensional (2-D) (Row and Column) FEC Protection 327 In networks where the source packets are lost both randomly and in 328 bursts, the sender ought to generate both non-interleaved and 329 interleaved FEC repair packets. This type of FEC protection is known 330 as 2-D parity FEC protection. At the expense of generating more FEC 331 repair packets, thus increasing the FEC overhead, 2-D FEC provides 332 superior protection against mixed loss patterns. However, it is 333 still possible for 2-D parity FEC protection to fail to recover all 334 of the lost source packets if a particular loss pattern occurs. An 335 example scenario is illustrated in Figure 7. 337 +---+ +---+ +===+ 338 | 1 | X X | 4 | |R_1| 339 +---+ +---+ +===+ 341 +---+ +---+ +---+ +---+ +===+ 342 | 5 | | 6 | | 7 | | 8 | |R_2| 343 +---+ +---+ +---+ +---+ +===+ 345 +---+ +---+ +===+ 346 | 9 | X X | 12| |R_3| 347 +---+ +---+ +===+ 349 +===+ +===+ +===+ +===+ 350 |C_1| |C_2| |C_3| |C_4| 351 +===+ +===+ +===+ +===+ 353 Figure 7: Example scenario #1 where 2-D parity FEC protection fails 354 error recovery 356 2-D parity FEC protection also fails when at least two rows are 357 missing a source and the FEC packet and the missing source packets 358 (in at least two rows) are aligned in the same column. An example 359 loss pattern is sketched in Figure 8. Similarly, 2-D parity FEC 360 protection cannot repair all missing source packets when at least two 361 columns are missing a source and the FEC packet and the missing 362 source packets (in at least two columns) are aligned in the same row. 364 +---+ +---+ +---+ 365 | 1 | | 2 | X | 4 | X 366 +---+ +---+ +---+ 368 +---+ +---+ +---+ +---+ +===+ 369 | 5 | | 6 | | 7 | | 8 | |R_2| 370 +---+ +---+ +---+ +---+ +===+ 372 +---+ +---+ +---+ 373 | 9 | | 10| X | 12| X 374 +---+ +---+ +---+ 376 +===+ +===+ +===+ +===+ 377 |C_1| |C_2| |C_3| |C_4| 378 +===+ +===+ +===+ +===+ 380 Figure 8: Example scenario #2 where 2-D parity FEC protection fails 381 error recovery 383 1.1.5. FEC Protection with Flexible Mask 385 It is possible to define FEC protection for selected packets in the 386 source stream. This would enable differential protection, i.e. 387 application of FEC selectively to packets that require a higher level 388 of reliability then the other packets in the source stream. The 389 sender will be required to send a bitmap indicating the packets to be 390 protected, i.e. a "mask", to the receiver. Since the mask can be 391 modified during an RTP session ("flexible mask"), this kind of FEC 392 protection can also be used to implement FEC dynamically (e.g. for 393 adaptation to different types of traffic during the RTP session). 395 1.1.6. FEC Overhead Considerations 397 The overhead is defined as the ratio of the number of bytes belonging 398 to the repair packets to the number of bytes belonging to the 399 protected source packets. 401 Generally, repair packets are larger in size compared to the source 402 packets. Also, not all the source packets are necessarily equal in 403 size. However, assuming that each repair packet carries an equal 404 number of bytes as carried by a source packet, the overhead for 405 different FEC protection methods can be computed as follows: 407 o 1-D Non-interleaved FEC Protection: Overhead = 1/L 409 o 1-D Interleaved FEC Protection: Overhead = 1/D 411 o 2-D Parity FEC Protection: Overhead = 1/L + 1/D 413 where L and D are the number of columns and rows in the source block, 414 respectively. 416 1.1.7. FEC Protection with Retransmission 418 This specification supports both forward error correction, i.e. 419 before any loss is reported, as well as retransmission of source 420 packets after loss is reported. The retransmission includes the RTP 421 header of the source packet in addition to the payload. Therefore, 422 endpoints supporting other RTP retransmission methods (see [RFC4588]) 423 in addition to FLEX FEC MUST only use the FLEX FEC retransmission 424 method. 426 1.1.8. Repair Window Considerations 428 The value for the repair window duration is related to the maximum L 429 and D values that are expected during a FLEX FEC session and 430 therefore cannot be chosen arbitrarily. Repair packets that include 431 L and D values larger than the repair window MUST not be sent. The 432 rate of the source streams should also be considered, as the repair 433 window duration should ideally span several packetization intervals 434 in order to leverage the error correction capabilities of the parity 435 code. 437 Since the FEC configuration can change with each repair packet (see 438 Section 4.2.2), for any given repair packet the FLEX FEC receiver 439 MUST support all possible L and D combinations (both 1-D and 2-D 440 interleaved over all source flows) and all flexible mask 441 configurations (over all source flows) within the repair window to 442 which it has agreed (e.g. through SDP or out-of-band signaling) for a 443 FLEX FEC RTP session. In addition, the FLEX FEC receiver MUST 444 support receipt of a retransmission of any source flow packet within 445 the repair window to which it has agreed. 447 2. Requirements Notation 449 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 450 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 451 "OPTIONAL" in this document are to be interpreted as described in BCP 452 14 [RFC2119] [RFC8174] when, and only when, they appear in all 453 capitals, as shown here. 455 3. Definitions and Notations 457 3.1. Definitions 459 This document uses a number of definitions from [RFC6363]. 461 1-D Non-interleaved Row FEC: A protection scheme that operates on 462 consecutive source packets in the source block, able to recover a 463 single lost source packet per row of the source block. 465 1-D Interleaved Column FEC: A protection scheme that operates on 466 interleaved source packets in the source block, able to recover a 467 single lost source packet per column of the source block. 469 2-D FEC: A protection scheme that combines row and column FEC. 471 Source Block: A set of source packets that are protected by a set 472 of 1-D or 2-D FEC repair packets. 474 FEC Block: A source block and its corresponding FEC repair 475 packets. 477 Repair Window: The time that spans a FEC block, which consists of 478 the source packets and the corresponding FEC repair packets. 480 XOR Parity Codes: A FEC code which uses the eXclusive OR (XOR) 481 parity operation to encode a set of source packets to form a FEC 482 repair packet. 484 3.2. Notations 486 L: Number of columns of the source block (length of each row). 488 D: Number of rows of the source block (depth of each column). 490 bitmask: A 15-bit, 46-bit, or 110-bit mask indicating which source 491 packets are protected by a FEC repair packet. If the bit i in the 492 mask is set to 1, the source packet number N + i is protected by 493 this FEC repair packet, where N is the sequence number base 494 indicated in the FEC repair packet. The most significant bit of 495 the mask corresponds to i=0. The least significant bit of the 496 mask corresponds to i=14 in the 15-bit mask, i=45 in the 46-bit 497 mask, or i=109 in the 110-bit mask. 499 4. Packet Formats 501 This section describes the formats of the source packets and defines 502 the formats of the FEC repair packets. 504 4.1. Source Packets 506 The source packets contain the information that identifies the source 507 block and the position within the source block occupied by the 508 packet. Since the source packets that are carried within an RTP 509 stream already contain unique sequence numbers in their RTP headers 510 [RFC3550], the source packets can be identified in a straightforward 511 manner and there is no need to append additional field(s). The 512 primary advantage of not modifying the source packets in any way is 513 that it provides backward compatibility for the receivers that do not 514 support FEC at all. In multicast scenarios, this backward 515 compatibility becomes quite useful as it allows the non-FEC-capable 516 and FEC-capable receivers to receive and interpret the same source 517 packets sent in the same multicast session. 519 The source packets are transmitted as usual without altering them. 520 They are used along with the FEC repair packets to recover any 521 missing source packets, making this scheme a systematic code. 523 The source packets are full RTP packets with optional CSRC list, RTP 524 header extension, and padding. If any of these optional elements are 525 present in the source RTP packet, and that source packet is lost, 526 they are recovered by the FEC repair operation, which recovers the 527 full source RTP packet including these optional elements. 529 4.2. FEC Repair Packets 531 The FEC repair packets will contain information that identifies the 532 source block they pertain to and the relationship between the 533 contained repair packets and the original source block. For this 534 purpose, the RTP header of the repair packets is used, as well as 535 another header within the RTP payload, called the FEC header, as 536 shown in Figure 9. 538 Note that all the source stream packets that are protected by a 539 particular FEC packet need to be in the same RTP session. 541 +------------------------------+ 542 | IP Header | 543 +------------------------------+ 544 | Transport Header | 545 +------------------------------+ 546 | RTP Header | 547 +------------------------------+ ---+ 548 | FEC Header | | 549 +------------------------------+ | RTP Payload 550 | Repair "Payload" | | 551 +------------------------------+ ---+ 553 Figure 9: Format of FEC repair packets 555 The Repair "Payload", which follows the FEC Header, includes repair 556 of everything following the fixed 12-byte RTP header of each source 557 packet, including any CSRC identifier list and header extensions if 558 present. 560 4.2.1. RTP Header of FEC Repair Packets 562 The RTP header is formatted according to [RFC3550] with some further 563 clarifications listed below: 565 Version (V) 2 bits: This MUST be set to 2 (binary 10), as this 566 specification requires all source RTP packets and all FEC repair 567 packets to use RTP version 2. 569 Padding (P) bit: Source packets can have optional RTP padding, 570 which can be recovered. FEC repair packets can have optional RTP 571 padding, which is independent of the RTP padding of the source 572 packets. 574 Extension (X) bit: Source packets can have optional RTP header 575 extensions, which can be recovered. FEC repair packets can have 576 optional RTP header extensions, which are independent of the RTP 577 header extensions of the source packets. 579 CSRC Count (CC) 4 bits, and CSRC List (CSRC_i) 32 bits each: 580 Source packets can have an optional CSRC list and count, which can 581 be recovered. FEC repair packets MUST use the CSRC list and count 582 to specify the SSRC(s) of the source RTP stream(s) protected by 583 this FEC repair packet. 585 Marker (M) bit: This bit is not used for this payload type, and 586 SHALL be set to 0 by senders, and SHALL be ignored by receivers. 588 Payload Type: The (dynamic) payload type for the FEC repair 589 packets is determined through out-of-band means (e.g. SDP). Note 590 that this document registers new payload formats for the repair 591 packets (Refer to Section 5 for details). According to [RFC3550], 592 an RTP receiver that cannot recognize a payload type must discard 593 it. This provides backward compatibility. If a non-FEC-capable 594 receiver receives a repair packet, it will not recognize the 595 payload type, and hence, will discard the repair packet. 597 Sequence Number (SN): The sequence number follows the standard 598 definition provided in [RFC3550]. Therefore it must be one higher 599 than the sequence number in the previously transmitted repair 600 packet, and the initial value of the sequence number should be 601 random (i.e. unpredictable). 603 Timestamp (TS): The timestamp SHALL be set to a time corresponding 604 to the repair packet's transmission time. Note that the timestamp 605 value has no use in the actual FEC protection process and is 606 usually useful for jitter calculations. 608 Synchronization Source (SSRC): The SSRC value for each repair 609 stream SHALL be randomly assigned as per the guidelines provided 610 in Section 8 of [RFC3550]. This allows the sender to multiplex 611 the source and repair RTP streams in the same RTP session, or 612 multiplex multiple repair streams in an RTP session. The repair 613 streams' SSRC's CNAME SHOULD be identical to the CNAME of the 614 source RTP stream(s) that this repair stream protects. An FEC 615 stream that protects multiple source RTP streams with different 616 CNAME's uses the CNAME associated with the entity generating the 617 FEC stream or the CNAME of the entity on whose behalf it performs 618 the protection operation. In cases when the repair stream covers 619 packets from multiple source RTP streams with different CNAME 620 values and none of these CNAME values can be associated with the 621 entity generating the FEC stream, any of these CNAME values MAY be 622 used. 624 In some networks, the RTP Source, which produces the source 625 packets and the FEC Source, which generates the repair packets 626 from the source packets may not be the same host. In such 627 scenarios, using the same CNAME for the source and repair RTP 628 streams means that the RTP Source and the FEC Source will share 629 the same CNAME (for this specific source-repair stream 630 association). A common CNAME may be produced based on an 631 algorithm that is known both to the RTP and FEC Source [RFC7022]. 632 This usage is compliant with [RFC3550]. 634 Note that due to the randomness of the SSRC assignments, there is 635 a possibility of SSRC collision. In such cases, the collisions 636 must be resolved as described in [RFC3550]. 638 4.2.2. FEC Header of FEC Repair Packets 640 The format of the FEC header has 3 variants, depending on the values 641 in the first 2 bits (R and F bits) as shown in Figure 10. Note that 642 R and F stand for "retransmit" and "fixed block", respectively. Two 643 of these variants are meant to describe different methods for 644 deriving the source data from a source packet for a repair packet. 645 This allows for customizing the FEC method to allow for robustness 646 against different levels of burst errors and random packet losses. 647 The third variant is for a straight retransmission of the source 648 packet. 650 0 1 2 3 651 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 1 652 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 653 |R|F|P|X| CC |M| PT recovery | ...varies depending on R/F... | 654 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 655 | | 656 | ...varies depending on R/F... | 657 | | 658 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 659 : Repair "Payload" follows FEC Header : 660 : : 662 Figure 10: FEC Header 664 The Repair "Payload", which follows the FEC Header, includes repair 665 of everything following the fixed 12-byte RTP header of each source 666 packet, including any CSRC identifier list and header extensions if 667 present. An overview on how the repair payload can be used to 668 recover source packets is provided Section 6. 670 +---+---+-----------------------------------------------------+ 671 | R | F | FEC Header variant | 672 +---+---+-----------------------------------------------------+ 673 | 0 | 0 | Flexible FEC Mask fields indicate source packets | 674 | 0 | 1 | Fixed FEC L/D (cols/rows) indicate source packets | 675 | 1 | 0 | Retransmission of a single source packet | 676 | 1 | 1 | Reserved for future use, MUST NOT send, MUST ignore | 677 +---+---+-----------------------------------------------------+ 679 Figure 11: R and F bit values for FEC Header variants 681 The first variant, when R=0 and F=0, has a mask to signal protected 682 source packets, as shown in Figure 12. 684 The second variant, when R=0 and F=1, has a number of columns (L) and 685 rows (D) to signal protected source packets, as shown in Figure 13. 687 The final variant, when R=1 and F=0, is a retransmission format as 688 shown in Figure 15. 690 No variant presently uses R=1 and F=1, which is reserved for future 691 use. Current FLEX FEC implementations MUST NOT send packets with 692 this variant, and receivers MUST ignore these packets. Future FLEX 693 FEC implementations may use this by updating the media type 694 registration. 696 The FEC header for all variants consists of the following common 697 fields: 699 o The R bit MUST be set to 1 to indicate a retransmission packet, 700 and MUST be set to 0 for FEC repair packets. 702 o The F bit indicates the type of FEC repair packets, as shown in 703 Figure 11, when the R bit is 0. The F bit MUST be set to 0 when 704 the R bit is 1 for retransmission packets. 706 o The P, X, CC, M and PT recovery fields are used to determine the 707 corresponding fields of the recovered packets (see also 708 Section 6.3.2). 710 4.2.2.1. FEC Header with Flexible Mask 712 When R=0 and F=0, the FEC Header includes flexible mask fields. 714 0 1 2 3 715 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 1 716 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 717 |0|0|P|X| CC |M| PT recovery | length recovery | 718 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 719 | TS recovery | 720 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 721 | SN base_i |k| Mask [0-14] | 722 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 723 |k| Mask [15-45] (optional) | 724 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 725 | Mask [46-109] (optional) | 726 | | 727 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 728 | ... next SN base and Mask for CSRC_i in CSRC list ... | 729 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 730 : Repair "Payload" follows FEC Header : 731 : : 733 Figure 12: FEC Header for F=0 735 o The Length recovery (16 bits) field is used to determine the 736 length of the recovered packets. This length includes all octets 737 following the fixed 12-byte RTP header of source packets, 738 including CSRC list and optional header extension(s) if present. 739 It excludes the fixed 12-byte RTP header of source packets. 741 o The TS recovery (32 bits) field is used to determine the timestamp 742 of the recovered packets. 744 o The CSRC_i (32 bits) field in the RTP Header (not FEC Header) 745 describes the SSRC of the source packets protected by this 746 particular FEC packet. If a FEC packet protects multiple SSRCs 747 (indicated by the CSRC Count > 1 in the RTP Header), there will be 748 multiple blocks of data containing the SN base and Mask fields. 750 o The SN base_i (16 bits) field indicates the lowest sequence 751 number, taking wrap around into account, of the source packets for 752 a particular SSRC (indicated in CSRC_i) protected by this repair 753 packet. 755 o The Mask fields indicate a bitmask of which source packets are 756 protected by this FEC repair packet, where bit j of the mask set 757 to 1 indicates that the source packet with sequence number (SN 758 base_i + j) is protected by this FEC repair packet, where j=0 is 759 the most significant bit in the mask. 761 o The k-bit in the bitmasks indicates if the mask is 15, 46, or 110 762 bits. k=1 denotes that another mask follows, and k=0 denotes that 763 it is the last block of mask. 765 o The Repair "Payload", which follows the FEC Header, includes 766 repair of everything following the fixed 12-byte RTP header of 767 each source packet, including any CSRC identifier list and header 768 extensions if present. 770 4.2.2.2. FEC Header with Fixed L Columns and D Rows 772 When R=0 and F=1, the FEC Header includes L and D fields for fixed 773 columns and rows. The other fields are the same as the prior 774 section. As in the previous section, the CSRC_i (32 bits) field in 775 the RTP Header (not FEC Header) describes the SSRC of the source 776 packets protected by this particular FEC packet. If there are 777 multiple SSRC's protected by the FEC packet, then there will be 778 multiple blocks of data containing an SN base along with L and D 779 fields. 781 0 1 2 3 782 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 1 783 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 784 |0|1|P|X| CC |M| PT recovery | length recovery | 785 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 786 | TS recovery | 787 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 788 | SN base_i | L (columns) | D (rows) | 789 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 790 | ... next SN base and L/D for CSRC_i in CSRC list ... | 791 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 792 : Repair "Payload" follows FEC Header : 793 : : 795 Figure 13: FEC Header for F=1 797 Consequently, the following conditions occur for L and D values: 799 If L=0, D=0, reserved for future use, 800 MUST NOT send, MUST ignore if received. 802 If L>0, D=0, indicates row FEC, and no column FEC will follow (1D). 803 Source packets for each row: SN, SN+1, ..., SN+(L-1) 805 If L>0, D=1, indicates row FEC, and column FEC will follow (2D). 806 Source packets for each row: SN, SN+1, ..., SN+(L-1) 807 Source packets for each col: SN, SN+L, ..., SN+(D-1)*L 808 After all row FEC packets have been sent, 809 then the column FEC packets will be sent. 811 If L>0, D>1, indicates column FEC of every L packet, D times. 812 Source packets for each col: SN, SN+L, ..., SN+(D-1)*L 814 Figure 14: Interpreting the L and D field values 816 Given the 8-bit limit on L and D (as depicted in Figure 13), the 817 maximum value of either parameter is 255. If L=0 and D=0 are in a 818 packet, then the repair packet MUST be ignored by the receiver. In 819 addition when L=1 and D=0, the repair packet becomes a retransmission 820 of a corresponding source packet. 822 The values of L and D for a given block of recovery data will 823 correspond to the type of recovery in use for that block of data. In 824 particular, for 2-D repair, the (L,D) values may not be constant 825 across all packets for a given SSRC being repaired. Similarly, the L 826 and D values can differ across different blocks of repair data 827 (repairing different SSRCs) in a single packet. If the values of L 828 and D result in a repair packet that exceed the repair window of the 829 FLEX FEC session, then the repair packet MUST be ignored. 831 It should be noted that the flexible mask-based approach may be 832 inefficient for protecting a large number of source packets, or 833 impossible to signal if larger than the largest mask size. In such 834 cases, the fixed columns and rows variant may be more useful. 836 4.2.2.3. FEC Header for Retransmissions 838 When R=1 and F=0, the FEC packet is a retransmission of a single 839 source packet. Note that the layout of this retransmission packet is 840 different from other FEC repair packets. The sequence number (SN 841 base_i) replaces the length recovery in the FEC header, since the 842 length is already known for a single packet. There are no L, D or 843 Mask fields, since only a single packet is retransmitted, identified 844 by the sequence number in the FEC header. The source packet SSRC is 845 included in the FEC header for retransmissions, not in the RTP header 846 CSRC list as in the FEC header variants with R=0. When performing 847 retransmissions, a single repair packet stream (SSRC) MAY be used for 848 retransmitting packets from multiple source packet streams (SSRCs), 849 as well as transmitting FEC repair packets that protect multiple 850 source packet streams (SSRCs). 852 This FEC header layout is identical to the source RTP (version 2) 853 packet, starting with its RTP header, where the retransmission 854 "payload" is everything following the fixed 12-byte RTP header of the 855 source packet, including CSRC list and extensions if present. 856 Therefore, the only operation needed for sending retransmissions is 857 to prepend a new RTP header to the source packet. 859 0 1 2 3 860 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 1 861 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 862 |1|0|P|X| CC |M| Payload Type| Sequence Number | 863 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 864 | Timestamp | 865 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 866 | SSRC | 867 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 868 : Retransmission "Payload" follows FEC Header : 869 : : 871 Figure 15: FEC Header for Retransmission 873 5. Payload Format Parameters 875 This section provides the media subtype registration for the non- 876 interleaved and interleaved parity FEC. The parameters that are 877 required to configure the FEC encoding and decoding operations are 878 also defined in this section. If no specific FEC code is specified 879 in the subtype, then the FEC code defaults to the parity code defined 880 in this specification. 882 5.1. Media Type Registration - Parity Codes 884 This registration is done using the template defined in [RFC6838] and 885 following the guidance provided in [RFC4855] along with [RFC4856]. 887 Note to the RFC Editor: In the following sections, please replace 888 "XXXX" with the number of this document prior to publication as an 889 RFC. 891 5.1.1. Registration of audio/flexfec 893 Type name: audio 895 Subtype name: flexfec 897 Required parameters: 899 o rate: The RTP timestamp (clock) rate. The rate SHALL be larger 900 than 1000 Hz to provide sufficient resolution to RTCP operations. 901 However, it is RECOMMENDED to select the rate that matches the 902 rate of the protected source RTP stream. 904 o repair-window: The time that spans the source packets and the 905 corresponding repair packets. The size of the repair window is 906 specified in microseconds. 908 Encoding considerations: This media type is framed (See Section 4.8 909 in the template document [RFC6838]) and contains binary data. 911 Security considerations: See Section 9 of [RFCXXXX]. 913 Interoperability considerations: None. 915 Published specification: [RFCXXXX]. 917 Applications that use this media type: Multimedia applications that 918 want to improve resiliency against packet loss by sending redundant 919 data in addition to the source media. 921 Fragment identifier considerations: None. 923 Additional information: None. 925 Person & email address to contact for further information: IESG 926 and IETF Audio/Video Transport Payloads Working Group 927 (or it's successor as delegated by the IESG). 929 Intended usage: COMMON. 931 Restriction on usage: This media type depends on RTP framing, and 932 hence, is only defined for transport via RTP [RFC3550]. 934 Author: Varun Singh . 936 Change controller: IETF Audio/Video Transport Payloads Working Group 937 delegated from the IESG (or it's successor as delegated by the IESG). 939 5.1.2. Registration of video/flexfec 941 Type name: video 943 Subtype name: flexfec 945 Required parameters: 947 o rate: The RTP timestamp (clock) rate. The rate SHALL be larger 948 than 1000 Hz to provide sufficient resolution to RTCP operations. 949 However, it is RECOMMENDED to select the rate that matches the 950 rate of the protected source RTP stream. 952 o repair-window: The time that spans the source packets and the 953 corresponding repair packets. The size of the repair window is 954 specified in microseconds. 956 Encoding considerations: This media type is framed (See Section 4.8 957 in the template document [RFC6838]) and contains binary data. 959 Security considerations: See Section 9 of [RFCXXXX]. 961 Interoperability considerations: None. 963 Published specification: [RFCXXXX]. 965 Applications that use this media type: Multimedia applications that 966 want to improve resiliency against packet loss by sending redundant 967 data in addition to the source media. 969 Fragment identifier considerations: None. 971 Additional information: None. 973 Person & email address to contact for further information: IESG 974 and IETF Audio/Video Transport Payloads Working Group 975 (or it's successor as delegated by the IESG). 977 Intended usage: COMMON. 979 Restriction on usage: This media type depends on RTP framing, and 980 hence, is only defined for transport via RTP [RFC3550]. 982 Author: Varun Singh . 984 Change controller: IETF Audio/Video Transport Payloads Working Group 985 delegated from the IESG (or it's successor as delegated by the IESG). 987 5.1.3. Registration of text/flexfec 989 Type name: text 991 Subtype name: flexfec 993 Required parameters: 995 o rate: The RTP timestamp (clock) rate. The rate SHALL be larger 996 than 1000 Hz to provide sufficient resolution to RTCP operations. 997 However, it is RECOMMENDED to select the rate that matches the 998 rate of the protected source RTP stream. 1000 o repair-window: The time that spans the source packets and the 1001 corresponding repair packets. The size of the repair window is 1002 specified in microseconds. 1004 Encoding considerations: This media type is framed (See Section 4.8 1005 in the template document [RFC6838]) and contains binary data. 1007 Security considerations: See Section 9 of [RFCXXXX]. 1009 Interoperability considerations: None. 1011 Published specification: [RFCXXXX]. 1013 Applications that use this media type: Multimedia applications that 1014 want to improve resiliency against packet loss by sending redundant 1015 data in addition to the source media. 1017 Fragment identifier considerations: None. 1019 Additional information: None. 1021 Person & email address to contact for further information: IESG 1022 and IETF Audio/Video Transport Payloads Working Group 1023 (or it's successor as delegated by the IESG). 1025 Intended usage: COMMON. 1027 Restriction on usage: This media type depends on RTP framing, and 1028 hence, is only defined for transport via RTP [RFC3550]. 1030 Author: Varun Singh . 1032 Change controller: IETF Audio/Video Transport Payloads Working Group 1033 delegated from the IESG (or it's successor as delegated by the IESG). 1035 5.1.4. Registration of application/flexfec 1037 Type name: application 1039 Subtype name: flexfec 1041 Required parameters: 1043 o rate: The RTP timestamp (clock) rate. The rate SHALL be larger 1044 than 1000 Hz to provide sufficient resolution to RTCP operations. 1045 However, it is RECOMMENDED to select the rate that matches the 1046 rate of the protected source RTP stream. 1048 o repair-window: The time that spans the source packets and the 1049 corresponding repair packets. The size of the repair window is 1050 specified in microseconds. 1052 Encoding considerations: This media type is framed (See Section 4.8 1053 in the template document [RFC6838]) and contains binary data. 1055 Security considerations: See Section 9 of [RFCXXXX]. 1057 Interoperability considerations: None. 1059 Published specification: [RFCXXXX]. 1061 Applications that use this media type: Multimedia applications that 1062 want to improve resiliency against packet loss by sending redundant 1063 data in addition to the source media. 1065 Fragment identifier considerations: None. 1067 Additional information: None. 1069 Person & email address to contact for further information: IESG 1070 and IETF Audio/Video Transport Payloads Working Group 1071 (or it's successor as delegated by the IESG). 1073 Intended usage: COMMON. 1075 Restriction on usage: This media type depends on RTP framing, and 1076 hence, is only defined for transport via RTP [RFC3550]. 1078 Author: Varun Singh . 1080 Change controller: IETF Audio/Video Transport Payloads Working Group 1081 delegated from the IESG (or it's successor as delegated by the IESG). 1083 5.2. Mapping to SDP Parameters 1085 Applications that use the RTP transport commonly use Session 1086 Description Protocol (SDP) [RFC4566] to describe their RTP sessions. 1087 The information that is used to specify the media types in an RTP 1088 session has specific mappings to the fields in an SDP description. 1089 This section provides these mappings for the media subtypes 1090 registered by this document. Note that if an application does not 1091 use SDP to describe the RTP sessions, an appropriate mapping must be 1092 defined and used to specify the media types and their parameters for 1093 the control/description protocol employed by the application. 1095 The mapping of the media type specification for "flexfec" and its 1096 associated parameters in SDP is as follows: 1098 o The media type (e.g., "application") goes into the "m=" line as 1099 the media name. 1101 o The media subtype goes into the "a=rtpmap" line as the encoding 1102 name. The RTP clock rate parameter ("rate") also goes into the 1103 "a=rtpmap" line as the clock rate. 1105 o The remaining required payload-format-specific parameters go into 1106 the "a=fmtp" line by copying them directly from the media type 1107 string as a semicolon-separated list of parameter=value pairs. 1109 SDP examples are provided in Section 7.1. 1111 5.2.1. Offer-Answer Model Considerations 1113 When offering parity FEC over RTP using SDP in an Offer/Answer model 1114 [RFC3264], the following considerations apply: 1116 o A sender application will indicate a repair window consistent with 1117 the desired amount of protection. Note that since the sender can 1118 change the FEC configuration on a packet-by-packet basis, the 1119 receiver must support any valid FLEX FEC configuration within the 1120 repair window associated with the offer (see Section 4.2.2). If 1121 the receiver cannot support the offered repair window it MUST 1122 reject the offer. 1124 o The size of the repair-window is related to the maximum delay 1125 between the transmission of a source packet and the associated 1126 repair packet. This directly impacts the buffering requirement on 1127 the receiver side and the receiver must consider this when 1128 choosing an offer. 1130 o Any unknown option in the offer must be ignored and deleted from 1131 the answer (see Section 6 of [RFC3264]). If FEC is not desired by 1132 the receiver, it can be deleted from the answer. 1134 5.2.2. Declarative Considerations 1136 In declarative usage, like SDP in the Real-time Streaming Protocol 1137 (RTSP, for RTSP 1.0 see [RFC2326] and for RTSP 2.0 see [RFC7826]) or 1138 the Session Announcement Protocol (SAP) [RFC2974], the following 1139 considerations apply: 1141 o The payload format configuration parameters are all declarative 1142 and a participant MUST use the configuration that is provided for 1143 the session. 1145 o More than one configuration may be provided (if desired) by 1146 declaring multiple RTP payload types. In that case, the receivers 1147 should choose the repair stream that is best for them. 1149 6. Protection and Recovery Procedures - Parity Codes 1151 This section provides a complete specification of the 1-D and 2-D 1152 parity codes and their RTP payload formats. It does not apply to the 1153 single packet retransmission format (R=1 in the FEC Header). 1155 6.1. Overview 1157 The following sections specify the steps involved in generating the 1158 repair packets and reconstructing the missing source packets from the 1159 repair packets. 1161 6.2. Repair Packet Construction 1163 The RTP Header of a repair packet is formed based on the guidelines 1164 given in Section 4.2. 1166 The FEC Header and Repair "Payload" of repair packets are formed by 1167 applying the XOR operation on the bit strings that are generated from 1168 the individual source packets protected by this particular repair 1169 packet. The set of the source packets that are associated with a 1170 given repair packet can be computed by the formula given in 1171 Section 6.3.1. 1173 The bit string is formed for each source packet by concatenating the 1174 following fields together in the order specified: 1176 o The first 16 bits of the RTP header (16 bits), though the first 1177 two (version) bits will be ignored by the recovery procedure. 1179 o Unsigned network-ordered 16-bit representation of the source 1180 packet length in bytes minus 12 (for the fixed RTP header), i.e., 1181 the sum of the lengths of all the following if present: the CSRC 1182 list, extension header, RTP payload and RTP padding (16 bits). 1184 o The timestamp of the RTP header (32 bits). 1186 o All octets after the fixed 12-byte RTP header. (Note the SSRC 1187 field is skipped.) 1189 The FEC bit string is generated by applying the parity operation on 1190 the bit strings produced from the source packets. The FEC header is 1191 generated from the FEC bit string as follows: 1193 o The first (most significant) 2 bits in the FEC bit string, which 1194 contain the RTP version field, are skipped. The R and F bits in 1195 the FEC header are set to the appropriate value, i.e., it depends 1196 on the chosen format variant. As a consequence of overwriting the 1197 RTP version field with the R and F bits, this payload format only 1198 supports RTP version 2. 1200 o The next bit in the FEC bit string is written into the P recovery 1201 bit in the FEC header. 1203 o The next bit in the FEC bit string is written into the X recovery 1204 bit in the FEC header. 1206 o The next 4 bits of the FEC bit string are written into the CC 1207 recovery field in the FEC header. 1209 o The next bit is written into the M recovery bit in the FEC header. 1211 o The next 7 bits of the FEC bit string are written into the PT 1212 recovery field in the FEC header. 1214 o The next 16 bits are written into the length recovery field in the 1215 FEC header. 1217 o The next 32 bits of the FEC bit string are written into the TS 1218 recovery field in the FEC header. 1220 o The lowest Sequence Number of the source packets protected by this 1221 repair packet is written into the Sequence Number Base field in 1222 the FEC header. This needs to be repeated for each SSRC that has 1223 packets included in the source block. 1225 o Depending on the chosen FEC header variant, the mask(s) are set 1226 when F=0, or the L and D values are set when F=1. This needs to 1227 be repeated for each SSRC that has packets included in the source 1228 block. 1230 o The rest of the FEC bit string, which contains everything after 1231 the fixed 12-byte RTP header of the source packet, is written into 1232 the Repair "Payload" following the FEC header, where "Payload" 1233 refers to everything after the fixed 12-byte RTP header, including 1234 extensions, CSRC list, true payloads, and padding. 1236 If the lengths of the source packets are not equal, each shorter 1237 packet MUST be padded to the length of the longest packet by adding 1238 octet 0's at the end. 1240 Due to this possible padding and mandatory FEC header, a repair 1241 packet has a larger size than the source packets it protects. This 1242 may cause problems if the resulting repair packet size exceeds the 1243 Maximum Transmission Unit (MTU) size of the path over which the 1244 repair stream is sent. 1246 6.3. Source Packet Reconstruction 1248 This section describes the recovery procedures that are required to 1249 reconstruct the missing source packets. The recovery process has two 1250 steps. In the first step, the FEC decoder determines which source 1251 and repair packets should be used in order to recover a missing 1252 packet. In the second step, the decoder recovers the missing packet, 1253 which consists of an RTP header and RTP payload. 1255 The following describes the RECOMMENDED algorithms for the first and 1256 second steps. Based on the implementation, different algorithms MAY 1257 be adopted. However, the end result MUST be identical to the one 1258 produced by the algorithms described below. 1260 Note that the same algorithms are used by the 1-D parity codes, 1261 regardless of whether the FEC protection is applied over a column or 1262 a row. The 2-D parity codes, on the other hand, usually require 1263 multiple iterations of the procedures described here. This iterative 1264 decoding algorithm is further explained in Section 6.3.4. 1266 6.3.1. Associating the Source and Repair Packets 1268 Before associating source and repair packets, the receiver must know 1269 in which RTP sessions the source and repair respectively are being 1270 sent. After this is established by the receiver the first step is 1271 associating the source and repair packets. This association can be 1272 via flexible bitmasks, or fixed L and D offsets which can be in the 1273 FEC header or signaled in SDP in optional payload format parameters 1274 when L=D=0 in the FEC header. 1276 6.3.1.1. Using Bitmasks 1278 To use flexible bitmasks, the first two FEC header bits MUST have R=0 1279 and F=0. A 15-bit, 46-bit, or 110-bit mask indicates which source 1280 packets are protected by a FEC repair packet. If the bit i in the 1281 mask is set to 1, the source packet number N + i is protected by this 1282 FEC repair packet, where N is the sequence number base indicated in 1283 the FEC header. The most significant bit of the mask corresponds to 1284 i=0. The least significant bit of the mask corresponds to i=14 in 1285 the 15-bit mask, i=45 in the 46-bit mask, or i=109 in the 110-bit 1286 mask. 1288 The bitmasks are able to represent arbitrary protection patterns, for 1289 example, 1-D interleaved, 1-D non-interleaved, 2-D. 1291 6.3.1.2. Using L and D Offsets 1293 Denote the set of the source packets associated with repair packet p* 1294 by set T(p*). Note that in a source block whose size is L columns by 1295 D rows, set T includes D source packets plus one repair packet for 1296 the FEC protection applied over a column, and L source packets plus 1297 one repair packet for the FEC protection applied over a row. Recall 1298 that 1-D interleaved and non-interleaved FEC protection can fully 1299 recover the missing information if there is only one source packet 1300 missing per column or row in set T. If there are more than one 1301 source packets missing per column or row in set T, 1-D FEC protection 1302 may fail to recover all the missing information. 1304 When value of L is non-zero, the 8-bit fields indicate the offset of 1305 packets protected by an interleaved (D>0) or non-interleaved (D=0) 1306 FEC packet. Using a combination of interleaved and non-interleaved 1307 FEC repair packets can form 2-D protection patterns. 1309 Mathematically, for any received repair packet, p*, the sequence 1310 numbers of the source packets that are protected by this repair 1311 packet are determined as follows, where SN is the sequence number 1312 base in the FEC header: 1314 For each SSRC (in CSRC list): 1315 When D <= 1: Source packets for each row: SN, SN+1, ..., SN+(L-1) 1316 When D > 1: Source packets for each col: SN, SN+L, ..., SN+(D-1)*L 1318 6.3.2. Recovering the RTP Header 1320 For a given set T, the procedure for the recovery of the RTP header 1321 of the missing packet, whose sequence number is denoted by SEQNUM, is 1322 as follows: 1324 1. For each of the source packets that are successfully received in 1325 T, compute the 80-bit string by concatenating the first 64 bits 1326 of their RTP header and the unsigned network-ordered 16-bit 1327 representation of their length in bytes minus 12. 1329 2. For the repair packet in T, extract the FEC bit string as the 1330 first 80 bits of the FEC header. 1332 3. Calculate the recovered bit string as the XOR of the bit strings 1333 generated from all source packets in T and the FEC bit string 1334 generated from the repair packet in T. 1336 4. Create a new packet with the standard 12-byte RTP header and no 1337 payload. 1339 5. Set the version of the new packet to 2. Skip the first 2 bits 1340 in the recovered bit string. 1342 6. Set the Padding bit in the new packet to the next bit in the 1343 recovered bit string. 1345 7. Set the Extension bit in the new packet to the next bit in the 1346 recovered bit string. 1348 8. Set the CC field to the next 4 bits in the recovered bit string. 1350 9. Set the Marker bit in the new packet to the next bit in the 1351 recovered bit string. 1353 10. Set the Payload type in the new packet to the next 7 bits in the 1354 recovered bit string. 1356 11. Set the SN field in the new packet to SEQNUM. 1358 12. Take the next 16 bits of the recovered bit string and set the 1359 new variable Y to whatever unsigned integer this represents 1360 (assuming network order). Convert Y to host order. Y 1361 represents the length of the new packet in bytes minus 12 (for 1362 the fixed RTP header), i.e., the sum of the lengths of all the 1363 following if present: the CSRC list, header extension, RTP 1364 payload and RTP padding. 1366 13. Set the TS field in the new packet to the next 32 bits in the 1367 recovered bit string. 1369 14. Set the SSRC of the new packet to the SSRC of the missing source 1370 RTP stream. 1372 This procedure recovers the header of an RTP packet up to (and 1373 including) the SSRC field. 1375 6.3.3. Recovering the RTP Payload 1377 Following the recovery of the RTP header, the procedure for the 1378 recovery of the RTP "payload" is as follows, where "payload" refers 1379 to everything following the fixed 12-byte RTP header, including 1380 extensions, CSRC list, true payload and padding. 1382 1. Allocate Y additional bytes for the new packet generated in 1383 Section 6.3.2. 1385 2. For each of the source packets that are successfully received in 1386 T, compute the bit string from the Y octets of data starting with 1387 the 13th octet of the packet. If any of the bit strings 1388 generated from the source packets has a length shorter than Y, 1389 pad them to that length. The zero-padding octets MUST be added 1390 at the end of the bit string. Note that the information of the 1391 first 8 octets are protected by the FEC header. 1393 3. For the repair packet in T, compute the FEC bit string from the 1394 repair packet payload, i.e., the Y octets of data following the 1395 FEC header. Note that the FEC header may be different sizes 1396 depending on the variant and bitmask size. 1398 4. Calculate the recovered bit string as the XOR of the bit strings 1399 generated from all source packets in T and the FEC bit string 1400 generated from the repair packet in T. 1402 5. Set the last Y octets in the new packet to the recovered bit 1403 string. 1405 6.3.4. Iterative Decoding Algorithm for the 2-D Parity FEC Protection 1407 In 2-D parity FEC protection, the sender generates both non- 1408 interleaved and interleaved FEC repair packets to combat with the 1409 mixed loss patterns (random and bursty). At the receiver side, these 1410 FEC packets are used iteratively to overcome the shortcomings of the 1411 1-D non-interleaved/interleaved FEC protection and improve the 1412 chances of full error recovery. 1414 The iterative decoding algorithm runs as follows: 1416 1. Set num_recovered_until_this_iteration to zero 1418 2. Set num_recovered_so_far to zero 1420 3. Recover as many source packets as possible by using the non- 1421 interleaved FEC repair packets as outlined in Section 6.3.2 and 1422 Section 6.3.3, and increase the value of num_recovered_so_far by 1423 the number of recovered source packets. 1425 4. Recover as many source packets as possible by using the 1426 interleaved FEC repair packets as outlined in Section 6.3.2 and 1427 Section 6.3.3, and increase the value of num_recovered_so_far by 1428 the number of recovered source packets. 1430 5. If num_recovered_so_far > num_recovered_until_this_iteration 1431 ---num_recovered_until_this_iteration = num_recovered_so_far 1432 ---Go to step 3 1433 Else 1434 ---Terminate 1436 The algorithm terminates either when all missing source packets are 1437 fully recovered or when there are still remaining missing source 1438 packets but the FEC repair packets are not able to recover any more 1439 source packets. For the example scenarios when the 2-D parity FEC 1440 protection fails full recovery, refer to Section 1.1.4. Upon 1441 termination, variable num_recovered_so_far has a value equal to the 1442 total number of recovered source packets. 1444 Example: 1446 Suppose that the receiver experienced the loss pattern sketched in 1447 Figure 16. 1449 +---+ +---+ +===+ 1450 X X | 3 | | 4 | |R_1| 1451 +---+ +---+ +===+ 1453 +---+ +---+ +---+ +---+ +===+ 1454 | 5 | | 6 | | 7 | | 8 | |R_2| 1455 +---+ +---+ +---+ +---+ +===+ 1457 +---+ +---+ +===+ 1458 | 9 | X X | 12| |R_3| 1459 +---+ +---+ +===+ 1461 +===+ +===+ +===+ +===+ 1462 |C_1| |C_2| |C_3| |C_4| 1463 +===+ +===+ +===+ +===+ 1465 Figure 16: Example loss pattern for the iterative decoding algorithm 1467 The receiver executes the iterative decoding algorithm and recovers 1468 source packets #1 and #11 in the first iteration. The resulting 1469 pattern is sketched in Figure 17. 1471 +---+ +---+ +---+ +===+ 1472 | 1 | X | 3 | | 4 | |R_1| 1473 +---+ +---+ +---+ +===+ 1475 +---+ +---+ +---+ +---+ +===+ 1476 | 5 | | 6 | | 7 | | 8 | |R_2| 1477 +---+ +---+ +---+ +---+ +===+ 1479 +---+ +---+ +---+ +===+ 1480 | 9 | X | 11| | 12| |R_3| 1481 +---+ +---+ +---+ +===+ 1483 +===+ +===+ +===+ +===+ 1484 |C_1| |C_2| |C_3| |C_4| 1485 +===+ +===+ +===+ +===+ 1487 Figure 17: The resulting pattern after the first iteration 1489 Since the if condition holds true, the receiver runs a new iteration. 1490 In the second iteration, source packets #2 and #10 are recovered, 1491 resulting in a full recovery as sketched in Figure 18. 1493 +---+ +---+ +---+ +---+ +===+ 1494 | 1 | | 2 | | 3 | | 4 | |R_1| 1495 +---+ +---+ +---+ +---+ +===+ 1497 +---+ +---+ +---+ +---+ +===+ 1498 | 5 | | 6 | | 7 | | 8 | |R_2| 1499 +---+ +---+ +---+ +---+ +===+ 1501 +---+ +---+ +---+ +---+ +===+ 1502 | 9 | | 10| | 11| | 12| |R_3| 1503 +---+ +---+ +---+ +---+ +===+ 1505 +===+ +===+ +===+ +===+ 1506 |C_1| |C_2| |C_3| |C_4| 1507 +===+ +===+ +===+ +===+ 1509 Figure 18: The resulting pattern after the second iteration 1511 7. Signaling Requirements 1513 Out-of-band signaling should be designed to enable the receiver to 1514 identify the RTP streams associated with source packets and repair 1515 packets, respectively. At a minimum, the signaling must be designed 1516 to allow the receiver to 1518 o Determine whether one or more source RTP streams will be sent. 1520 o Determine whether one or more repair RTP streams will be sent. 1522 o Associate the appropriate SSRC's to both source and repair 1523 streams. 1525 o Clearly identify which SSRC's are associated with each source 1526 block. 1528 o Clearly identify which repair packets correspond to which source 1529 blocks. 1531 o Make use of repair packets to recover source data associated with 1532 specific SSRC's. 1534 This section provides several Session Description Protocol (SDP) 1535 examples to demonstrate how these requirements can be met. 1537 7.1. SDP Examples 1539 This section provides two SDP [RFC4566] examples. The examples use 1540 the FEC grouping semantics defined in [RFC5956]. 1542 7.1.1. Example SDP for Flexible FEC Protection with in-band SSRC 1543 mapping 1545 In this example, we have one source video stream and one FEC repair 1546 stream. The source and repair streams are multiplexed on different 1547 SSRCs. The repair window is set to 200 ms. 1549 v=0 1550 o=mo 1122334455 1122334466 IN IP4 fec.example.com 1551 s=FlexFEC minimal SDP signalling Example 1552 t=0 0 1553 m=video 30000 RTP/AVP 96 98 1554 c=IN IP4 233.252.0.1/127 1555 a=rtpmap:96 VP8/90000 1556 a=rtpmap:98 flexfec/90000 1557 a=fmtp:98; repair-window=200000 1559 7.1.2. Example SDP for Flexible FEC Protection with explicit signalling 1560 in the SDP 1562 This example shows one source video stream (ssrc:1234) and one FEC 1563 repair streams (ssrc:2345). One FEC group is formed with the 1564 "a=ssrc-group:FEC-FR 1234 2345" line. The source and repair streams 1565 are multiplexed on different SSRCs. The repair window is set to 200 1566 ms. 1568 v=0 1569 o=ali 1122334455 1122334466 IN IP4 fec.example.com 1570 s=2-D Parity FEC with no in band signalling Example 1571 t=0 0 1572 m=video 30000 RTP/AVP 100 110 1573 c=IN IP4 192.0.2.0/24 1574 a=rtpmap:100 MP2T/90000 1575 a=rtpmap:110 flexfec/90000 1576 a=fmtp:110; repair-window:200000 1577 a=ssrc:1234 1578 a=ssrc:2345 1579 a=ssrc-group:FEC-FR 1234 2345 1581 7.2. On the Use of the RTP Stream Identifier Source Description 1583 The RTP Stream Identifier Source Description [I-D.ietf-avtext-rid] is 1584 a format that can be used to identify a single RTP source stream 1585 along with an associated repair stream. However, this specification 1586 already defines a method of source and repair stream identification 1587 that can enable protection of multiple source streams with a single 1588 repair stream. Therefore the RTP Stream Idenfifer Source Description 1589 SHOULD NOT be used for the Flexible FEC payload format 1591 8. Congestion Control Considerations 1593 FEC is an effective approach to provide applications resiliency 1594 against packet losses. However, in networks where the congestion is 1595 a major contributor to the packet loss, the potential impacts of 1596 using FEC should be considered carefully before injecting the repair 1597 streams into the network. In particular, in bandwidth-limited 1598 networks, FEC repair streams may consume a significant part of the 1599 available bandwidth and consequently may congest the network. In 1600 such cases, the applications MUST NOT arbitrarily increase the amount 1601 of FEC protection since doing so may lead to a congestion collapse. 1602 If desired, stronger FEC protection MAY be applied only after the 1603 source rate has been reduced. 1605 In a network-friendly implementation, an application should avoid 1606 sending/receiving FEC repair streams if it knows that sending/ 1607 receiving those FEC repair streams would not help at all in 1608 recovering the missing packets. Examples of where FEC would not be 1609 beneficial are: (1) if the successful recovery rate as determined by 1610 RTCP feedback is low (see [RFC5725] and [RFC7509]), and (2) the 1611 application has a smaller latency requirement than the repair window 1612 adopted by the FEC configuration based on the expected burst loss 1613 duration and the target FEC overhead. It is RECOMMENDED that the 1614 amount and type (row, column, or both) of FEC protection is adjusted 1615 dynamically based on the packet loss rate and burst loss length 1616 observed by the applications. 1618 In multicast scenarios, it may be difficult to optimize the FEC 1619 protection per receiver. If there is a large variation among the 1620 levels of FEC protection needed by different receivers, it is 1621 RECOMMENDED that the sender offers multiple repair streams with 1622 different levels of FEC protection and the receivers join the 1623 corresponding multicast sessions to receive the repair stream(s) that 1624 is best for them. 1626 9. Security Considerations 1628 RTP packets using the payload format defined in this specification 1629 are subject to the security considerations discussed in the RTP 1630 specification [RFC3550] and in any applicable RTP profile. The main 1631 security considerations for the RTP packet carrying the RTP payload 1632 format defined within this memo are confidentiality, integrity and 1633 source authenticity. Confidentiality can be provided by encrypting 1634 the RTP payload. Integrity of the RTP packets is achieved through a 1635 suitable cryptographic integrity protection mechanism. Such a 1636 cryptographic system may also allow the authentication of the source 1637 of the payload. A suitable security mechanism for this RTP payload 1638 format should provide confidentiality, integrity protection, and at 1639 least source authentication capable of determining if an RTP packet 1640 is from a member of the RTP session. 1642 Note that the appropriate mechanism to provide security to RTP and 1643 payloads following this memo may vary. It is dependent on the 1644 application, transport and signaling protocol employed. Therefore, a 1645 single mechanism is not sufficient, although if suitable, using the 1646 Secure Real-time Transport Protocol (SRTP) [RFC3711] is recommended. 1647 Other mechanisms that may be used are IPsec [RFC4301] and Datagram 1648 Transport Layer Security (DTLS, see [RFC6347]) (RTP over UDP); other 1649 alternatives may exist. 1651 Given that FLEX FEC enables the protection of multiple source 1652 streams, there exists the possibility that multiple source buffers 1653 may be created that may not be used. An attacker could leverage 1654 unused source buffers to as a means of occupying memory in a FLEX FEC 1655 endpoint. In addition, an attack against the FEC parameters 1656 themselves (e.g. repair window, D or L values) can result in a 1657 receiver having to allocate source buffer space that may also lead to 1658 excessive consumption of resources. Similarly, a network attacker 1659 could modify the recovery fields corresponding to packet lengths 1660 (assuming there are no message integrity mechanisms) which in turn 1661 could force unnecessarily large memory allocations at the receiver. 1662 Moreover the application source data may not be perfectly matched 1663 with FLEX FEC source partitioning. If this is the case, there is a 1664 possibility for unprotected source data if, for instance, the FLEX 1665 FEC implementation discards data that does not fit perfectly into its 1666 source processing requirements. 1668 10. IANA Considerations 1670 New media subtypes are subject to IANA registration. For the 1671 registration of the payload formats and their parameters introduced 1672 in this document, refer to Section 5.1. 1674 11. Acknowledgments 1676 Some parts of this document are borrowed from [RFC5109]. Thus, the 1677 author would like to thank the editor of [RFC5109] and those who 1678 contributed to [RFC5109]. 1680 Thanks to Stephen Botzko , Bernard Aboba , Rasmus Brandt , Brian 1681 Baldino , Roni Even , Stefan Holmer , Jonathan Lennox , and Magnus 1682 Westerlund for providing valuable feedback on earlier versions of 1683 this draft. 1685 12. References 1687 12.1. Normative References 1689 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1690 Requirement Levels", BCP 14, RFC 2119, 1691 DOI 10.17487/RFC2119, March 1997, 1692 . 1694 [RFC3264] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model 1695 with Session Description Protocol (SDP)", RFC 3264, 1696 DOI 10.17487/RFC3264, June 2002, 1697 . 1699 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 1700 Jacobson, "RTP: A Transport Protocol for Real-Time 1701 Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, 1702 July 2003, . 1704 [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session 1705 Description Protocol", RFC 4566, DOI 10.17487/RFC4566, 1706 July 2006, . 1708 [RFC4855] Casner, S., "Media Type Registration of RTP Payload 1709 Formats", RFC 4855, DOI 10.17487/RFC4855, February 2007, 1710 . 1712 [RFC4856] Casner, S., "Media Type Registration of Payload Formats in 1713 the RTP Profile for Audio and Video Conferences", 1714 RFC 4856, DOI 10.17487/RFC4856, February 2007, 1715 . 1717 [RFC5956] Begen, A., "Forward Error Correction Grouping Semantics in 1718 the Session Description Protocol", RFC 5956, 1719 DOI 10.17487/RFC5956, September 2010, 1720 . 1722 [RFC6363] Watson, M., Begen, A., and V. Roca, "Forward Error 1723 Correction (FEC) Framework", RFC 6363, 1724 DOI 10.17487/RFC6363, October 2011, 1725 . 1727 [RFC6838] Freed, N., Klensin, J., and T. Hansen, "Media Type 1728 Specifications and Registration Procedures", BCP 13, 1729 RFC 6838, DOI 10.17487/RFC6838, January 2013, 1730 . 1732 [RFC7022] Begen, A., Perkins, C., Wing, D., and E. Rescorla, 1733 "Guidelines for Choosing RTP Control Protocol (RTCP) 1734 Canonical Names (CNAMEs)", RFC 7022, DOI 10.17487/RFC7022, 1735 September 2013, . 1737 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1738 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1739 May 2017, . 1741 12.2. Informative References 1743 [I-D.ietf-avtext-rid] 1744 Roach, A., Nandakumar, S., and P. Thatcher, "RTP Stream 1745 Identifier Source Description (SDES)", draft-ietf-avtext- 1746 rid-09 (work in progress), October 2016. 1748 [RFC2326] Schulzrinne, H., Rao, A., and R. Lanphier, "Real Time 1749 Streaming Protocol (RTSP)", RFC 2326, 1750 DOI 10.17487/RFC2326, April 1998, 1751 . 1753 [RFC2733] Rosenberg, J. and H. Schulzrinne, "An RTP Payload Format 1754 for Generic Forward Error Correction", RFC 2733, 1755 DOI 10.17487/RFC2733, December 1999, 1756 . 1758 [RFC2974] Handley, M., Perkins, C., and E. Whelan, "Session 1759 Announcement Protocol", RFC 2974, DOI 10.17487/RFC2974, 1760 October 2000, . 1762 [RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K. 1763 Norrman, "The Secure Real-time Transport Protocol (SRTP)", 1764 RFC 3711, DOI 10.17487/RFC3711, March 2004, 1765 . 1767 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1768 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 1769 December 2005, . 1771 [RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 1772 "Extended RTP Profile for Real-time Transport Control 1773 Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, 1774 DOI 10.17487/RFC4585, July 2006, 1775 . 1777 [RFC4588] Rey, J., Leon, D., Miyazaki, A., Varsa, V., and R. 1778 Hakenberg, "RTP Retransmission Payload Format", RFC 4588, 1779 DOI 10.17487/RFC4588, July 2006, 1780 . 1782 [RFC5109] Li, A., Ed., "RTP Payload Format for Generic Forward Error 1783 Correction", RFC 5109, DOI 10.17487/RFC5109, December 1784 2007, . 1786 [RFC5725] Begen, A., Hsu, D., and M. Lague, "Post-Repair Loss RLE 1787 Report Block Type for RTP Control Protocol (RTCP) Extended 1788 Reports (XRs)", RFC 5725, DOI 10.17487/RFC5725, February 1789 2010, . 1791 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 1792 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 1793 January 2012, . 1795 [RFC7509] Huang, R. and V. Singh, "RTP Control Protocol (RTCP) 1796 Extended Report (XR) for Post-Repair Loss Count Metrics", 1797 RFC 7509, DOI 10.17487/RFC7509, May 2015, 1798 . 1800 [RFC7656] Lennox, J., Gross, K., Nandakumar, S., Salgueiro, G., and 1801 B. Burman, Ed., "A Taxonomy of Semantics and Mechanisms 1802 for Real-Time Transport Protocol (RTP) Sources", RFC 7656, 1803 DOI 10.17487/RFC7656, November 2015, 1804 . 1806 [RFC7826] Schulzrinne, H., Rao, A., Lanphier, R., Westerlund, M., 1807 and M. Stiemerling, Ed., "Real-Time Streaming Protocol 1808 Version 2.0", RFC 7826, DOI 10.17487/RFC7826, December 1809 2016, . 1811 [SMPTE2022-1] 1812 "Forward Error Correction for Real-Time Video/Audio 1813 Transport over IP Networks", 2007. 1815 Authors' Addresses 1817 Mo Zanaty 1818 Cisco 1819 Raleigh, NC 1820 USA 1822 Email: mzanaty@cisco.com 1824 Varun Singh 1825 CALLSTATS I/O Oy 1826 Runeberginkatu 4c A 4 1827 Helsinki 00100 1828 Finland 1830 Email: varun.singh@iki.fi 1831 URI: http://www.callstats.io/ 1833 Ali Begen 1834 Networked Media 1835 Konya 1836 Turkey 1838 Email: ali.begen@networked.media 1840 Giridhar Mandyam 1841 Qualcomm Inc. 1842 5775 Morehouse Drive 1843 San Diego, CA 92121 1844 USA 1846 Phone: +1 858 651 7200 1847 Email: mandyam@qti.qualcomm.com