idnits 2.17.1 draft-ietf-rmcat-eval-criteria-13.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 9, 2020) is 1502 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-11) exists of draft-ietf-rmcat-wireless-tests-09 Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 RMCAT WG V. Singh 3 Internet-Draft callstats.io 4 Intended status: Informational J. Ott 5 Expires: September 10, 2020 Technical University of Munich 6 S. Holmer 7 Google 8 March 9, 2020 10 Evaluating Congestion Control for Interactive Real-time Media 11 draft-ietf-rmcat-eval-criteria-13 13 Abstract 15 The Real-time Transport Protocol (RTP) is used to transmit media in 16 telephony and video conferencing applications. This document 17 describes the guidelines to evaluate new congestion control 18 algorithms for interactive point-to-point real-time media. 20 Status of This Memo 22 This Internet-Draft is submitted in full conformance with the 23 provisions of BCP 78 and BCP 79. 25 Internet-Drafts are working documents of the Internet Engineering 26 Task Force (IETF). Note that other groups may also distribute 27 working documents as Internet-Drafts. The list of current Internet- 28 Drafts is at https://datatracker.ietf.org/drafts/current/. 30 Internet-Drafts are draft documents valid for a maximum of six months 31 and may be updated, replaced, or obsoleted by other documents at any 32 time. It is inappropriate to use Internet-Drafts as reference 33 material or to cite them other than as "work in progress." 35 This Internet-Draft will expire on September 10, 2020. 37 Copyright Notice 39 Copyright (c) 2020 IETF Trust and the persons identified as the 40 document authors. All rights reserved. 42 This document is subject to BCP 78 and the IETF Trust's Legal 43 Provisions Relating to IETF Documents 44 (https://trustee.ietf.org/license-info) in effect on the date of 45 publication of this document. Please review these documents 46 carefully, as they describe your rights and restrictions with respect 47 to this document. Code Components extracted from this document must 48 include Simplified BSD License text as described in Section 4.e of 49 the Trust Legal Provisions and are provided without warranty as 50 described in the Simplified BSD License. 52 Table of Contents 54 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 55 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 56 3. Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 57 3.1. RTP Log Format . . . . . . . . . . . . . . . . . . . . . 5 58 4. List of Network Parameters . . . . . . . . . . . . . . . . . 6 59 4.1. One-way Propagation Delay . . . . . . . . . . . . . . . . 6 60 4.2. End-to-end Loss . . . . . . . . . . . . . . . . . . . . . 6 61 4.3. Drop Tail Router Queue Length . . . . . . . . . . . . . . 7 62 4.4. Loss generation model . . . . . . . . . . . . . . . . . . 7 63 4.5. Jitter models . . . . . . . . . . . . . . . . . . . . . . 7 64 4.5.1. Random Bounded PDV (RBPDV) . . . . . . . . . . . . . 8 65 4.5.2. Approximately Random Subject to No-Reordering Bounded 66 PDV (NR-RPVD) . . . . . . . . . . . . . . . . 9 67 4.5.3. Recommended distribution . . . . . . . . . . . . . . 10 68 5. Traffic Models . . . . . . . . . . . . . . . . . . . . . . . 10 69 5.1. TCP traffic model . . . . . . . . . . . . . . . . . . . . 10 70 5.2. RTP Video model . . . . . . . . . . . . . . . . . . . . . 11 71 5.3. Background UDP . . . . . . . . . . . . . . . . . . . . . 11 72 6. Security Considerations . . . . . . . . . . . . . . . . . . . 11 73 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 74 8. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 12 75 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 12 76 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 77 10.1. Normative References . . . . . . . . . . . . . . . . . . 13 78 10.2. Informative References . . . . . . . . . . . . . . . . . 13 79 Appendix A. Change Log . . . . . . . . . . . . . . . . . . . . . 14 80 A.1. Changes in draft-ietf-rmcat-eval-criteria-07 . . . . . . 15 81 A.2. Changes in draft-ietf-rmcat-eval-criteria-06 . . . . . . 15 82 A.3. Changes in draft-ietf-rmcat-eval-criteria-05 . . . . . . 15 83 A.4. Changes in draft-ietf-rmcat-eval-criteria-04 . . . . . . 15 84 A.5. Changes in draft-ietf-rmcat-eval-criteria-03 . . . . . . 15 85 A.6. Changes in draft-ietf-rmcat-eval-criteria-02 . . . . . . 15 86 A.7. Changes in draft-ietf-rmcat-eval-criteria-01 . . . . . . 15 87 A.8. Changes in draft-ietf-rmcat-eval-criteria-00 . . . . . . 16 88 A.9. Changes in draft-singh-rmcat-cc-eval-04 . . . . . . . . . 16 89 A.10. Changes in draft-singh-rmcat-cc-eval-03 . . . . . . . . . 16 90 A.11. Changes in draft-singh-rmcat-cc-eval-02 . . . . . . . . . 16 91 A.12. Changes in draft-singh-rmcat-cc-eval-01 . . . . . . . . . 16 92 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 17 94 1. Introduction 96 This memo describes the guidelines to help with evaluating new 97 congestion control algorithms for interactive point-to-point real 98 time media. The requirements for the congestion control algorithm 99 are outlined in [I-D.ietf-rmcat-cc-requirements]). This document 100 builds upon previous work at the IETF: Specifying New Congestion 101 Control Algorithms [RFC5033] and Metrics for the Evaluation of 102 Congestion Control Algorithms [RFC5166]. 104 The guidelines proposed in the document are intended to help prevent 105 a congestion collapse, promote fair capacity usage and optimize the 106 media flow's throughput. Furthermore, the proposed algorithms are 107 expected to operate within the envelope of the circuit breakers 108 defined in RFC8083 [RFC8083]. 110 This document only provides the broad set of network parameters and 111 and traffic models for evaluating a new congestion control algorithm. 112 The minimal requirements for congestion control proposals is to 113 produce or present results for the test scenarios described in 114 [I-D.ietf-rmcat-eval-test] (Basic Test Cases), which also defines the 115 specifics for the test cases. Additionally, proponents may produce 116 evaluation results for the wireless test scenarios 117 [I-D.ietf-rmcat-wireless-tests]. 119 This document does not cover application-specific implications of 120 congestion control algorithms and how those could be evaluated. 121 Therefore, no quality metrics are defined for performance evaluation; 122 quality metrics and algorithms to infer those vary between media 123 types. Metrics and algorithms to assess, e.g., quality of experience 124 evolve continuously so that determining suitable choices is left for 125 future work. However, there is consensus that each congestion 126 control algorithm should be able to show that it is useful for 127 interactive video by performing analysis using a real codecs and 128 video sequences and state-of-the-art quality metrics. 130 Beyond optimizing individual metrics, real-time applications may have 131 further options to trade off performance, e.g., across multiple 132 media; refer to the RMCAT requirements 133 [I-D.ietf-rmcat-cc-requirements] document. Such trade-offs may be 134 defined in the future. 136 2. Terminology 138 The terminology defined in RTP [RFC3550], RTP Profile for Audio and 139 Video Conferences with Minimal Control [RFC3551], RTCP Extended 140 Report (XR) [RFC3611], Extended RTP Profile for RTCP-based Feedback 141 (RTP/AVPF) [RFC4585] and Support for Reduced-Size RTCP [RFC5506] 142 apply. 144 3. Metrics 146 This document specifies testing criteria for evaluating congestion 147 control algorithms for RTP media flows. Proposed algorithms are to 148 prove their performance by means of simulation and/or emulation 149 experiments for all the cases described. 151 Each experiment is expected to log every incoming and outgoing packet 152 (the RTP logging format is described in Section 3.1). The logging 153 can be done inside the application or at the endpoints using PCAP 154 (packet capture, e.g., tcpdump [tcpdump], wireshark [wireshark]). 155 The following metrics are calculated based on the information in the 156 packet logs: 158 1. Sending rate, Receiver rate, Goodput (measured at 200ms 159 intervals) 161 2. Packets sent, Packets received 163 3. Bytes sent, bytes received 165 4. Packet delay 167 5. Packets lost, Packets discarded (from the playout or de-jitter 168 buffer) 170 6. If using, retransmission or FEC: post-repair loss 172 7. Self-Fairness and Fairness with respect to cross traffic: 173 Experiments testing a given congestion control proposal must 174 report on relative ratios of the average throughput (measured at 175 coarser time intervals) obtained by each RTP media stream. In 176 the presence of background cross-traffic such as TCP, the report 177 must also include the relative ratio between average throughput 178 of RTP media streams and cross-traffic streams. 179 During static periods of a test (i.e., when bottleneck bandwidth 180 is constant and no arrival/departure of streams), these report 181 on relative ratios serve as an indicator of how fair the RTP 182 streams share bandwidth amongst themselves and against cross- 183 traffic streams. The throughput measurement interval should be 184 set at a few values (for example, at 1s, 5s, and 20s) in order 185 to measure fairness across different time scales. 186 As a general guideline, the relative ratio between congestion 187 controlled RTP flows with the same priority level and similar 188 path RTT should be bounded between (0.333 and 3.) For example, 189 see the test scenarios described in [I-D.ietf-rmcat-eval-test]. 191 8. Convergence time: The time taken to reach a stable rate at 192 startup, after the available link capacity changes, or when new 193 flows get added to the bottleneck link. 195 9. Instability or oscillation in the sending rate: The frequency or 196 number of instances when the sending rate oscillates between an 197 high watermark level and a low watermark level, or vice-versa in 198 a defined time window. For example, the watermarks can be set 199 at 4x interval: 500 Kbps, 2 Mbps, and a time window of 500ms. 201 10. Bandwidth Utilization, defined as ratio of the instantaneous 202 sending rate to the instantaneous bottleneck capacity. This 203 metric is useful only when a congestion controlled RTP flow is 204 by itself or competing with similar cross-traffic. 206 Note that the above metrics are all objective application-independent 207 metrics. Refer to Section 3, in [I-D.ietf-netvc-testing] for 208 objective metrics for evaluating codecs. 210 From the logs the statistical measures (min, max, mean, standard 211 deviation and variance) for the whole duration or any specific part 212 of the session can be calculated. Also the metrics (sending rate, 213 receiver rate, goodput, latency) can be visualized in graphs as 214 variation over time, the measurements in the plot are at 1 second 215 intervals. Additionally, from the logs it is possible to plot the 216 histogram or CDF of packet delay. 218 3.1. RTP Log Format 220 Having a common log format simplifies running analyses across and 221 comparing different measurements. The log file should be tab or 222 comma separated containing the following details: 224 Send or receive timestamp (unix) 225 RTP payload type 226 SSRC 227 RTP sequence no 228 RTP timestamp 229 marker bit 230 payload size 232 If the congestion control implements, retransmissions or FEC, the 233 evaluation should report both packet loss (before applying error- 234 resilience) and residual packet loss (after applying error- 235 resilience). 237 These data should suffice to compute the media-encoding independent 238 metrics described above. Use of a common log will allow simplified 239 post-processing and analysis across different implementations. 241 4. List of Network Parameters 243 The implementors initially are encouraged to choose evaluation 244 settings from the following values: 246 4.1. One-way Propagation Delay 248 Experiments are expected to verify that the congestion control is 249 able to work across a broad range of path characteristics, also 250 including challenging situations, for example over trans-continental 251 and/or satellite links. Tests thus account for the following 252 different latencies: 254 1. Very low latency: 0-1ms 256 2. Low latency: 50ms 258 3. High latency: 150ms 260 4. Extreme latency: 300ms 262 4.2. End-to-end Loss 264 Many paths in the Internet today are largely lossless but, with 265 wireless networks and interference, towards remote regions, or in 266 scenarios featuring high/fast mobility, media flows may exhibit 267 substantial packet loss. This variety needs to be reflected 268 appropriately by the tests. 270 To model a wide range of lossy links, the experiments can choose one 271 of the following loss rates, the fractional loss is the ratio of 272 packets lost and packets sent. 274 1. no loss: 0% 276 2. 1% 278 3. 5% 280 4. 10% 282 5. 20% 284 4.3. Drop Tail Router Queue Length 286 Routers should be configured to use Drop Trail queues in the 287 experiments due to their (still) prevalent nature. Experimentation 288 with AQM schemes is encouraged but not mandatory. 290 The router queue length is measured as the time taken to drain the 291 FIFO queue. It has been noted in various discussions that the queue 292 length in the current deployed Internet varies significantly. While 293 the core backbone network has very short queue length, the home 294 gateways usually have larger queue length. Those various queue 295 lengths can be categorized in the following way: 297 1. QoS-aware (or short): 70ms 299 2. Nominal: 300-500ms 301 3. Buffer-bloated: 1000-2000ms 303 Here the size of the queue is measured in bytes or packets and to 304 convert the queue length measured in seconds to queue length in 305 bytes: 307 QueueSize (in bytes) = QueueSize (in sec) x Throughput (in bps)/8 309 4.4. Loss generation model 311 Many models for generating packet loss are available, some yield 312 correlated, others independent losses; losses can also be extracted 313 from packet traces. As a (simple) minimum loss model with minimal 314 parameterization (i.e., the loss rate), independent random losses 315 must be used in the evaluation. 317 It is known that independent loss models may reflect reality poorly 318 and hence more sophisticated loss models could be considered. 319 Suitable models for correlated losses includes the Gilbert-Elliot 320 model [gilbert-elliott] and losses generated by modeling a queue 321 including its (different) drop behaviors. 323 4.5. Jitter models 325 This section defines jitter models for the purposes of this document. 326 When jitter is to be applied to both the congestion controlled RTP 327 flow and any competing flow (such as a TCP competing flow), the 328 competing flow will use the jitter definition below that does not 329 allow for re-ordering of packets on the competing flow (see NR-RBPDV 330 definition below). 332 Jitter is an overloaded term in communications. It is typically used 333 to refer to the variation of a metric (e.g., delay) with respect to 334 some reference metric (e.g., average delay or minimum delay). For 335 example, RFC 3550 jitter is computed as the smoothed difference in 336 packet arrival times relative to their respective expected arrival 337 times, which is particularly meaningful if the underlying packet 338 delay variation was caused by a Gaussian random process. 340 Because jitter is an overloaded term, we use the term Packet Delay 341 Variation (PDV) instead to describe the variation of delay of 342 individual packets in the same sense as the IETF IPPM WG has defined 343 PDV in their documents (e.g., RFC 3393) and as the ITU-T SG16 has 344 defined IP Packet Delay Variation (IPDV) in their documents (e.g., 345 Y.1540). 347 Most PDV distributions in packet network systems are one-sided 348 distributions, the measurement of which with a finite number of 349 measurement samples results in one-sided histograms. In the usual 350 packet network transport case, there is typically one packet that 351 transited the network with the minimum delay; a (large) number of 352 packets transit the network within some (smaller) positive variation 353 from this minimum delay, and a (small) number of the packets transit 354 the network with delays higher than the median or average transit 355 time (these are outliers). Although infrequent, outliers can cause 356 significant deleterious operation in adaptive systems and should be 357 considered in rate adaptation designs for RTP congestion control. 359 In this section we define two different bounded PDV characteristics, 360 1) Random Bounded PDV and 2) Approximately Random Subject to No- 361 Reordering Bounded PDV. 363 The former, 1) Random Bounded PDV is presented for information only, 364 while the latter, 2) Approximately Random Subject to No-Reordering 365 Bounded PDV, must be used in the evaluation. 367 4.5.1. Random Bounded PDV (RBPDV) 369 The RBPDV probability distribution function (PDF) is specified to be 370 of some mathematically describable function which includes some 371 practical minimum and maximum discrete values suitable for testing. 372 For example, the minimum value, x_min, might be specified as the 373 minimum transit time packet and the maximum value, x_max, might be 374 defined to be two standard deviations higher than the mean. 376 Since we are typically interested in the distribution relative to the 377 mean delay packet, we define the zero mean PDV sample, z(n), to be 378 z(n) = x(n) - x_mean, where x(n) is a sample of the RBPDV random 379 variable x and x_mean is the mean of x. 381 We assume here that s(n) is the original source time of packet n and 382 the post-jitter induced emission time, j(n), for packet n is: 384 j(n) = {[z(n) + x_mean] + s(n)}. 386 It follows that the separation in the post-jitter time of packets n 387 and n+1 is {[s(n+1)-s(n)] - [z(n)-z(n+1)]}. Since the first term is 388 always a positive quantity, we note that packet reordering at the 389 receiver is possible whenever the second term is greater than the 390 first. Said another way, whenever the difference in possible zero 391 mean PDV sample delays (i.e., [x_max-x_min]) exceeds the inter- 392 departure time of any two sent packets, we have the possibility of 393 packet re-ordering. 395 There are important use cases in real networks where packets can 396 become re-ordered such as in load balancing topologies and during 397 route changes. However, for the vast majority of cases there is no 398 packet re-ordering because most of the time packets follow the same 399 path. Due to this, if a packet becomes overly delayed, the packets 400 after it on that flow are also delayed. This is especially true for 401 mobile wireless links where there are per-flow queues prior to base 402 station scheduling. Owing to this important use case, we define 403 another PDV profile similar to the above, but one that does not allow 404 for re-ordering within a flow. 406 4.5.2. Approximately Random Subject to No-Reordering Bounded PDV (NR- 407 RPVD) 409 No Reordering RPDV, NR-RPVD, is defined similarly to the above with 410 one important exception. Let serial(n) be defined as the 411 serialization delay of packet n at the lowest bottleneck link rate 412 (or other appropriate rate) in a given test. Then we produce all the 413 post-jitter values for j(n) for n = 1, 2, ... N, where N is the 414 length of the source sequence s to be offset-ed. The exception can 415 be stated as follows: We revisit all j(n) beginning from index n=2, 416 and if j(n) is determined to be less than [j(n-1)+serial(n-1)], we 417 redefine j(n) to be equal to [j(n-1)+serial(n-1)] and continue for 418 all remaining n (i.e., n = 3, 4, .. N). This models the case where 419 the packet n is sent immediately after packet (n-1) at the bottleneck 420 link rate. Although this is generally the theoretical minimum in 421 that it assumes that no other packets from other flows are in-between 422 packet n and n+1 at the bottleneck link, it is a reasonable 423 assumption for per flow queuing. 425 We note that this assumption holds for some important exception 426 cases, such as packets immediately following outliers. There are a 427 multitude of software controlled elements common on end-to-end 428 Internet paths (such as firewalls, ALGs and other middleboxes) which 429 stop processing packets while servicing other functions (e.g., 430 garbage collection). Often these devices do not drop packets, but 431 rather queue them for later processing and cause many of the 432 outliers. Thus NR-RPVD models this particular use case (assuming 433 serial(n+1) is defined appropriately for the device causing the 434 outlier) and thus is believed to be important for adaptation 435 development for congestion controlled RTP streams. 437 4.5.3. Recommended distribution 439 Whether Random Bounded PDV or Approximately Random Subject to No- 440 Reordering Bounded PDV, it is recommended that z(n) is distributed 441 according to a truncated Gaussian for the above jitter models: 443 z(n) ~ |max(min(N(0, std^2), N_STD * std), -N_STD * std)| 445 where N(0, std^2) is the Gaussian distribution with zero mean and 446 standard deviation std. Recommended values: 448 o std = 5 ms 450 o N_STD = 3 452 5. Traffic Models 454 5.1. TCP traffic model 456 Long-lived TCP flows will download data throughout the session and 457 are expected to have infinite amount of data to send or receive. 458 This roughly applies, for example, when downloading software 459 distributions. 461 Each short TCP flow is modeled as a sequence of file downloads 462 interleaved with idle periods. Not all short TCP flows start at the 463 same time, i.e., some start in the ON state while others start in the 464 OFF state. 466 The short TCP flows can be modeled as follows: 30 connections start 467 simultaneously fetching small (30-50 KB) amounts of data, evenly 468 distributed. This covers the case where the short TCP flows are 469 fetching web page resources rather than video files. 471 The idle period between bursts of starting a group of TCP flows is 472 typically derived from an exponential distribution with the mean 473 value of 10 seconds. 475 [These values were picked based on the data available at 476 http://httparchive.org/interesting.php as of October 2015]. 478 Many different TCP congestion control schemes are deployed today. 479 Therefore, experimentation with a range of different schemes, 480 especially including CUBIC, is encouraged. Experiments must document 481 in detail which congestion control schemes they tested against and 482 which parameters were used. 484 5.2. RTP Video model 486 [RFC8593] describes two types of video traffic models for evaluating 487 candidate algorithms for RTP congestion control. The first model 488 statistically characterizes the behavior of a video encoder, whereas 489 the second model uses video traces. 491 Sample video test sequences are available at [xiph-seq]. The 492 following two video streams are the recommended minimum for testing: 493 Foreman (CIF sequence) and FourPeople (720p); both come as raw video 494 data to be encoded dynamically. As these video sequences are short 495 (300 and 600 frames, respectively, they shall be stitched together 496 repeatedly until the desired length is reached. 498 5.3. Background UDP 500 Background UDP flow is modeled as a constant bit rate (CBR) flow. It 501 will download data at a particular CBR rate for the complete session, 502 or will change to particular CBR rate at predefined intervals. The 503 inter packet interval is calculated based on the CBR and the packet 504 size (is typically set to the path MTU size, the default value can be 505 1500 bytes). 507 Note that new transport protocols such as QUIC may use UDP but, due 508 to their congestion control algorithms, will exhibit behavior 509 conceptually similar in nature to TCP flows above and can thus be 510 subsumed by the above, including the division into short- and long- 511 lived flows. As QUIC evolves independently of TCP congestion control 512 algorithms, its future congestion control should be considered as 513 competing traffic as appropriate. 515 6. Security Considerations 517 This document specifies evaluation criteria and parameters for 518 assessing and comparing the performance of congestion control 519 protocols and algorithms for real-time communication. This memo 520 itself is thus not subject to security considerations but the 521 protocols and algorithms evaluated may be. In particular, successful 522 operation under all tests defined in this document may suffice for a 523 comparative evaluation but must not be interpreted that the protocol 524 is free of risks when deployed on the Internet as briefly described 525 in the following by example. 527 Such evaluations are expected to be carried out in controlled 528 environments for limited numbers of parallel flows. As such, these 529 evaluations are by definition limited and will not be able to 530 systematically consider possible interactions or very large groups of 531 communicating nodes under all possible circumstances, so that careful 532 protocol design is advised to avoid incidentally contributing traffic 533 that could lead to unstable networks, e.g., (local) congestion 534 collapse. 536 This specification focuses on assessing the regular operation of the 537 protocols and algorithms under considerations. It does not suggest 538 checks against malicious use of the protocols -- by the sender, the 539 receiver, or intermediate parties, e.g., through faked, dropped, 540 replicated, or modified congestion signals. It is up to the protocol 541 specifications themselves to ensure that authenticity, integrity, 542 and/or plausibility of received signals are checked and the 543 appropriate actions (or non-actions) are taken. 545 7. IANA Considerations 547 There are no IANA impacts in this memo. 549 8. Contributors 551 The content and concepts within this document are a product of the 552 discussion carried out in the Design Team. 554 Michael Ramalho provided the text for the Jitter model. 556 9. Acknowledgments 558 Much of this document is derived from previous work on congestion 559 control at the IETF. 561 The authors would like to thank Harald Alvestrand, Anna Brunstrom, 562 Luca De Cicco, Wesley Eddy, Lars Eggert, Kevin Gross, Vinayak Hegde, 563 Randell Jesup, Mirja Kuehlewind, Karen Nielsen, Piers O'Hanlon, Colin 564 Perkins, Michael Ramalho, Zaheduzzaman Sarker, Timothy B. 565 Terriberry, Michael Welzl, Mo Zanaty, and Xiaoqing Zhu for providing 566 valuable feedback on earlier versions of this draft. Additionally, 567 also thank the participants of the design team for their comments and 568 discussion related to the evaluation criteria. 570 10. References 571 10.1. Normative References 573 [I-D.ietf-rmcat-cc-requirements] 574 Jesup, R. and Z. Sarker, "Congestion Control Requirements 575 for Interactive Real-Time Media", draft-ietf-rmcat-cc- 576 requirements-09 (work in progress), December 2014. 578 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 579 Jacobson, "RTP: A Transport Protocol for Real-Time 580 Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, 581 July 2003, . 583 [RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and 584 Video Conferences with Minimal Control", STD 65, RFC 3551, 585 DOI 10.17487/RFC3551, July 2003, 586 . 588 [RFC3611] Friedman, T., Ed., Caceres, R., Ed., and A. Clark, Ed., 589 "RTP Control Protocol Extended Reports (RTCP XR)", 590 RFC 3611, DOI 10.17487/RFC3611, November 2003, 591 . 593 [RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 594 "Extended RTP Profile for Real-time Transport Control 595 Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, 596 DOI 10.17487/RFC4585, July 2006, 597 . 599 [RFC5506] Johansson, I. and M. Westerlund, "Support for Reduced-Size 600 Real-Time Transport Control Protocol (RTCP): Opportunities 601 and Consequences", RFC 5506, DOI 10.17487/RFC5506, April 602 2009, . 604 [RFC8083] Perkins, C. and V. Singh, "Multimedia Congestion Control: 605 Circuit Breakers for Unicast RTP Sessions", RFC 8083, 606 DOI 10.17487/RFC8083, March 2017, 607 . 609 [RFC8593] Zhu, X., Mena, S., and Z. Sarker, "Video Traffic Models 610 for RTP Congestion Control Evaluations", RFC 8593, 611 DOI 10.17487/RFC8593, May 2019, 612 . 614 10.2. Informative References 616 [gilbert-elliott] 617 Hasslinger, G. and O. Hohlfeld, "The Gilbert-Elliott Model 618 for Packet Loss in Real Time Services on the Internet", 619 14th GI/ITG Conference - Measurement, Modelling and 620 Evalutation of Computer and Communication Systems , 3 621 2008. 623 [I-D.ietf-netvc-testing] 624 Daede, T., Norkin, A., and I. Brailovskiy, "Video Codec 625 Testing and Quality Measurement", draft-ietf-netvc- 626 testing-09 (work in progress), January 2020. 628 [I-D.ietf-rmcat-eval-test] 629 Sarker, Z., Singh, V., Zhu, X., and M. Ramalho, "Test 630 Cases for Evaluating RMCAT Proposals", draft-ietf-rmcat- 631 eval-test-10 (work in progress), May 2019. 633 [I-D.ietf-rmcat-wireless-tests] 634 Sarker, Z., Johansson, I., Zhu, X., Fu, J., Tan, W., and 635 M. Ramalho, "Evaluation Test Cases for Interactive Real- 636 Time Media over Wireless Networks", draft-ietf-rmcat- 637 wireless-tests-09 (work in progress), February 2020. 639 [RFC5033] Floyd, S. and M. Allman, "Specifying New Congestion 640 Control Algorithms", BCP 133, RFC 5033, 641 DOI 10.17487/RFC5033, August 2007, 642 . 644 [RFC5166] Floyd, S., Ed., "Metrics for the Evaluation of Congestion 645 Control Mechanisms", RFC 5166, DOI 10.17487/RFC5166, March 646 2008, . 648 [tcpdump] "Homepage of tcpdump and libpcap", 649 https://www.tcpdump.org/index.html . 651 [wireshark] 652 "Homepage of Wireshark", https://www.wireshark.org . 654 [xiph-seq] 655 Daede, T., "Video Test Media Set", 656 https://media.xiph.org/video/derf/ . 658 Appendix A. Change Log 660 Note to the RFC-Editor: please remove this section prior to 661 publication as an RFC. 663 A.1. Changes in draft-ietf-rmcat-eval-criteria-07 665 Updated the draft according to the discussion at IETF-101. 667 o Updated the discussion on fairness. Thanks to Xiaoqing Zhu for 668 providing text. 670 o Fixed a simple loss model and provided pointers to more 671 sophisticated ones. 673 o Fixed the choice of the jitter model. 675 A.2. Changes in draft-ietf-rmcat-eval-criteria-06 677 o Updated Jitter. 679 A.3. Changes in draft-ietf-rmcat-eval-criteria-05 681 o Improved text surrounding wireless tests, video sequences, and 682 short-TCP model. 684 A.4. Changes in draft-ietf-rmcat-eval-criteria-04 686 o Removed the guidelines section, as most of the sections are now 687 covered: wireless tests, video model, etc. 689 o Improved Short TCP model based on the suggestion to use 690 httparchive.org. 692 A.5. Changes in draft-ietf-rmcat-eval-criteria-03 694 o Keep-alive version. 696 o Moved link parameters and traffic models from eval-test 698 A.6. Changes in draft-ietf-rmcat-eval-criteria-02 700 o Incorporated fairness test as a working test. 702 o Updated text on mimimum evaluation requirements. 704 A.7. Changes in draft-ietf-rmcat-eval-criteria-01 706 o Removed Appendix B. 708 o Removed Section on Evaluation Parameters. 710 A.8. Changes in draft-ietf-rmcat-eval-criteria-00 712 o Updated references. 714 o Resubmitted as WG draft. 716 A.9. Changes in draft-singh-rmcat-cc-eval-04 718 o Incorporate feedback from IETF 87, Berlin. 720 o Clarified metrics: convergence time, bandwidth utilization. 722 o Changed fairness criteria to fairness test. 724 o Added measuring pre- and post-repair loss. 726 o Added open issue of measuring video quality to appendix. 728 o clarified use of DropTail and AQM. 730 o Updated text in "Minimum Requirements for Evaluation" 732 A.10. Changes in draft-singh-rmcat-cc-eval-03 734 o Incorporate the discussion within the design team. 736 o Added a section on evaluation parameters, it describes the flow 737 and network characteristics. 739 o Added Appendix with self-fairness experiment. 741 o Changed bottleneck parameters from a proposal to an example set. 743 o 745 A.11. Changes in draft-singh-rmcat-cc-eval-02 747 o Added scenario descriptions. 749 A.12. Changes in draft-singh-rmcat-cc-eval-01 751 o Removed QoE metrics. 753 o Changed stability to steady-state. 755 o Added measuring impact against few and many flows. 757 o Added guideline for idle and data-limited periods. 759 o Added reference to TCP evaluation suite in example evaluation 760 scenarios. 762 Authors' Addresses 764 Varun Singh 765 CALLSTATS I/O Oy 766 Runeberginkatu 4c A 4 767 Helsinki 00100 768 Finland 770 Email: varun.singh@iki.fi 771 URI: https://www.callstats.io/about 773 Joerg Ott 774 Technical University of Munich 775 Faculty of Informatics 776 Boltzmannstrasse 3 777 Garching bei Muenchen, DE 85748 778 Germany 780 Email: ott@in.tum.de 782 Stefan Holmer 783 Google 784 Kungsbron 2 785 Stockholm 11122 786 Sweden 788 Email: holmer@google.com