idnits 2.17.1 draft-ietf-rmcat-wireless-tests-09.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 == Line 301 has weird spacing: '...nternet fix...' -- The document date (February 27, 2020) is 1518 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: '2D' is mentioned on line 318, but not defined == Missing Reference: '3D' is mentioned on line 318, but not defined == Missing Reference: '40ms' is mentioned on line 335, but not defined == Missing Reference: '150ms' is mentioned on line 335, but not defined == Missing Reference: 'Uplink' is mentioned on line 355, but not defined == Missing Reference: 'Downlink' is mentioned on line 355, but not defined -- Looks like a reference, but probably isn't: '4' on line 822 -- Looks like a reference, but probably isn't: '8' on line 822 -- Looks like a reference, but probably isn't: '12' on line 822 -- Looks like a reference, but probably isn't: '16' on line 822 -- Looks like a reference, but probably isn't: '20' on line 822 == Unused Reference: 'I-D.ietf-rmcat-cc-requirements' is defined on line 998, but no explicit reference was found in the text == Outdated reference: A later version (-14) exists of draft-ietf-rmcat-eval-criteria-11 Summary: 0 errors (**), 0 flaws (~~), 10 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Z. Sarker 3 Internet-Draft I. Johansson 4 Intended status: Informational Ericsson AB 5 Expires: August 30, 2020 X. Zhu 6 J. Fu 7 W. Tan 8 Cisco Systems 9 M. Ramalho 10 AcousticComms 11 February 27, 2020 13 Evaluation Test Cases for Interactive Real-Time Media over Wireless 14 Networks 15 draft-ietf-rmcat-wireless-tests-09 17 Abstract 19 The Real-time Transport Protocol (RTP) is a common transport choice 20 for interactive multimedia communication applications. The 21 performance of these applications typically depends on a well- 22 functioning congestion control algorithm. To ensure a seamless and 23 robust user experience, a well-designed RTP-based congestion control 24 algorithm should work well across all access network types. This 25 document describes test cases for evaluating performances of 26 candidate congestion control algorithms over cellular and Wi-Fi 27 networks. 29 Status of This Memo 31 This Internet-Draft is submitted in full conformance with the 32 provisions of BCP 78 and BCP 79. 34 Internet-Drafts are working documents of the Internet Engineering 35 Task Force (IETF). Note that other groups may also distribute 36 working documents as Internet-Drafts. The list of current Internet- 37 Drafts is at https://datatracker.ietf.org/drafts/current/. 39 Internet-Drafts are draft documents valid for a maximum of six months 40 and may be updated, replaced, or obsoleted by other documents at any 41 time. It is inappropriate to use Internet-Drafts as reference 42 material or to cite them other than as "work in progress." 44 This Internet-Draft will expire on August 30, 2020. 46 Copyright Notice 48 Copyright (c) 2020 IETF Trust and the persons identified as the 49 document authors. All rights reserved. 51 This document is subject to BCP 78 and the IETF Trust's Legal 52 Provisions Relating to IETF Documents 53 (https://trustee.ietf.org/license-info) in effect on the date of 54 publication of this document. Please review these documents 55 carefully, as they describe your rights and restrictions with respect 56 to this document. Code Components extracted from this document must 57 include Simplified BSD License text as described in Section 4.e of 58 the Trust Legal Provisions and are provided without warranty as 59 described in the Simplified BSD License. 61 Table of Contents 63 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 64 2. Terminologies . . . . . . . . . . . . . . . . . . . . . . . . 3 65 3. Cellular Network Specific Test Cases . . . . . . . . . . . . 3 66 3.1. Varying Network Load . . . . . . . . . . . . . . . . . . 6 67 3.1.1. Network Connection . . . . . . . . . . . . . . . . . 6 68 3.1.2. Simulation Setup . . . . . . . . . . . . . . . . . . 7 69 3.2. Bad Radio Coverage . . . . . . . . . . . . . . . . . . . 9 70 3.2.1. Network connection . . . . . . . . . . . . . . . . . 9 71 3.2.2. Simulation Setup . . . . . . . . . . . . . . . . . . 9 72 3.3. Desired Evaluation Metrics for cellular test cases . . . 10 73 4. Wi-Fi Networks Specific Test Cases . . . . . . . . . . . . . 10 74 4.1. Bottleneck in Wired Network . . . . . . . . . . . . . . . 12 75 4.1.1. Network topology . . . . . . . . . . . . . . . . . . 12 76 4.1.2. Test setup . . . . . . . . . . . . . . . . . . . . . 13 77 4.1.3. Typical test scenarios . . . . . . . . . . . . . . . 14 78 4.1.4. Expected behavior . . . . . . . . . . . . . . . . . . 15 79 4.2. Bottleneck in Wi-Fi Network . . . . . . . . . . . . . . . 15 80 4.2.1. Network topology . . . . . . . . . . . . . . . . . . 15 81 4.2.2. Test setup . . . . . . . . . . . . . . . . . . . . . 15 82 4.2.3. Typical test scenarios . . . . . . . . . . . . . . . 17 83 4.2.4. Expected behavior . . . . . . . . . . . . . . . . . . 18 84 4.3. Other Potential Test Cases . . . . . . . . . . . . . . . 19 85 4.3.1. EDCA/WMM usage . . . . . . . . . . . . . . . . . . . 19 86 4.3.2. Effect of heterogeneous link rates . . . . . . . . . 19 87 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20 88 6. Security Considerations . . . . . . . . . . . . . . . . . . . 20 89 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 20 90 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 20 91 8.1. Normative References . . . . . . . . . . . . . . . . . . 20 92 8.2. Informative References . . . . . . . . . . . . . . . . . 21 93 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 22 95 1. Introduction 97 Wireless networks (both cellular and Wi-Fi [IEEE802.11]) are an 98 integral and increasingly more significant part of the Internet. 99 Typical application scenarios for interactive multimedia 100 communication over wireless include from video conferencing calls in 101 a bus or train as well as live media streaming at home. It is well 102 known that the characteristics and technical challenges for 103 supporting multimedia services over wireless are very different from 104 those of providing the same service over a wired network. Although 105 the basic test cases as defined in [I-D.ietf-rmcat-eval-test] have 106 covered many common effects of network impairments for evaluating 107 RTP-based congestion control schemes, they remain to be tested over 108 characteristics and dynamics unique to a given wireless environment. 109 For example, in cellular networks, the base station maintains 110 individual queues per radio bearer per user hence it leads to a 111 different nature of interactions between traffic flows of different 112 users. This contrasts with the wired network setting where traffic 113 flows from all users share the same queue. Furthermore, user 114 mobility patterns in a cellular network differ from those in a Wi-Fi 115 network. Therefore, it is important to evaluate the performance of 116 proposed candidate RTP-based congestion control solutions over 117 cellular mobile networks and over Wi-Fi networks respectively. 119 The draft [I-D.ietf-rmcat-eval-criteria] provides the guideline for 120 evaluating candidate algorithms and recognizes the importance of 121 testing over wireless access networks. However, it does not describe 122 any specific test cases for performance evaluation of candidate 123 algorithms. This document describes test cases specifically 124 targeting cellular and Wi-Fi networks. 126 2. Terminologies 128 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 129 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 130 "OPTIONAL" in this document are to be interpreted as described in BCP 131 14 [RFC2119] [RFC8174] when, and only when, they appear in all 132 capitals, as shown here. 134 3. Cellular Network Specific Test Cases 136 A cellular environment is more complicated than its wireline 137 counterpart since it seeks to provide services in the context of 138 variable available bandwidth, location dependencies and user 139 mobilities at different speeds. In a cellular network, the user may 140 reach the cell edge which may lead to a significant amount of 141 retransmissions to deliver the data from the base station to the 142 destination and vice versa. These radio links will often act as a 143 bottleneck for the rest of the network and will eventually lead to 144 excessive delays or packet drops. An efficient retransmission or 145 link adaptation mechanism can reduce the packet loss probability but 146 there will remain some packet losses and delay variations. Moreover, 147 with increased cell load or handover to a congested cell, congestion 148 in the transport network will become even worse. Besides, there 149 exist certain characteristics that distinguish the cellular network 150 from other wireless access networks such as Wi-Fi. In a cellular 151 network -- 153 o The bottleneck is often a shared link with relatively few users. 155 * The cost per bit over the shared link varies over time and is 156 different for different users. 158 * Leftover/unused resources can be consumed by other greedy 159 users. 161 o Queues are always per radio bearer hence each user can have many 162 such queues. 164 o Users can experience both Inter and Intra Radio Access Technology 165 (RAT) handovers (see [HO-def-3GPP] for the definition of 166 "handover"). 168 o Handover between cells or change of serving cells (as described in 169 [HO-LTE-3GPP] and [HO-UMTS-3GPP]) might cause user plane 170 interruptions which can lead to bursts of packet losses, delay 171 and/or jitter. The exact behavior depends on the type of radio 172 bearer. Typically, the default best-effort bearers do not 173 generate packet loss, instead, packets are queued up and 174 transmitted once the handover is completed. 176 o The network part decides how much the user can transmit. 178 o The cellular network has variable link capacity per user. 180 * It can vary as fast as a period of milliseconds. 182 * It depends on many factors (such as distance, speed, 183 interference, different flows). 185 * It uses complex and smart link adaptation which makes the link 186 behavior ever more dynamic. 188 * The scheduling priority depends on the estimated throughput. 190 o Both Quality of Service (QoS) and non-QoS radio bearers can be 191 used. 193 Hence, a real-time communication application operating over a 194 cellular network needs to cope with a shared bottleneck link and 195 variable link capacity, events like handover, non-congestion related 196 loss, abrupt changes in bandwidth (both short term and long term) due 197 to handover, network load and bad radio coverage. Even though 3GPP 198 has defined QoS bearers [QoS-3GPP] to ensure high-quality user 199 experience, it is still preferable for real-time applications to 200 behave in an adaptive manner. 202 Different mobile operators deploy their own cellular networks with 203 their own set of network functionalities and policies. Usually, a 204 mobile operator network includes 2G, EDGE, 3G and 4G radio access 205 technologies. Looking at the specifications of such radio 206 technologies it is evident that only the more recent radio 207 technologies can support the high bandwidth requirements from real- 208 time interactive video applications. The future real-time 209 interactive application will impose even greater demand on cellular 210 network performance which makes 4G (and beyond) radio technologies 211 more suitable for such genre of application. 213 The key factors in defining test cases for cellular networks are: 215 o Shared and varying link capacity 217 o Mobility 219 o Handover 221 However, these factors are typically highly correlated in a cellular 222 network. Therefore, instead of devising separate test cases for 223 individual important events, we have divided the test case into two 224 categories. It should be noted that the goal of the following test 225 cases is to evaluate the performance of candidate algorithms over the 226 radio interface of the cellular network. Hence it is assumed that 227 the radio interface is the bottleneck link between the communicating 228 peers and that the core network does not introduce any extra 229 congestion along the path. Consequently, this draft has kept as out 230 of scope the combination of multiple access technologies involving 231 both cellular and Wi-Fi users. In this latter case the shared 232 bottleneck is likely at the wired backhaul link. These test cases 233 further assume a typical real-time telephony scenario where one real- 234 time session consists of one voice stream and one video stream. 236 Even though it is possible to carry out tests over operational 237 cellular networks (e.g., LTE/5G), and actually such tests are already 238 available today, these tests cannot in general be carried out in a 239 deterministic fashion to ensure repeatability. The main reason is 240 that these networks are controlled by cellular operators and there 241 exist various amounts of competing traffic in the same cell(s). In 242 practice, it is only in underground mines that one can carry out near 243 deterministic testing. Even there, it is not guaranteed either as 244 workers in the mines may carry with them their personal mobile 245 phones. Furthermore, the underground mining setting may not reflect 246 typical usage patterns in an urban setting. We, therefore, recommend 247 that a cellular network simulator is used for the test cases defined 248 in this document, for example -- the LTE simulator in [NS-3]. 250 3.1. Varying Network Load 252 The goal of this test is to evaluate the performance of the candidate 253 congestion control algorithm under varying network load. The network 254 load variation is created by adding and removing network users a.k.a. 255 User Equipments (UEs) during the simulation. In this test case, each 256 user/UE in the media session is an endpoint following RTP-based 257 congestion control. User arrivals follow a Poisson distribution 258 proportional to the length of the call, to keep the number of users 259 per cell fairly constant during the evaluation period. At the 260 beginning of the simulation, there should be enough time to warm-up 261 the network. This is to avoid running the evaluation in an empty 262 network where network nodes are having empty buffers, low 263 interference at the beginning of the simulation. This network 264 initialization period should be excluded from the evaluation period. 266 This test case also includes user mobility and some competing 267 traffic. The latter includes both the same types of flows (with same 268 adaptation algorithms) and different types of flows (with different 269 services and congestion control schemes). The investigated 270 congestion control algorithms should show maximum possible network 271 utilization and stability in terms of rate variations, lowest 272 possible end to end frame latency, network latency and Packet Loss 273 Rate (PLR) at different cell load level. 275 3.1.1. Network Connection 277 Each mobile user is connected to a fixed user. The connection 278 between the mobile user and fixed user consists of a cellular radio 279 access, an Evolved Packet Core (EPC) and an Internet connection. The 280 mobile user is connected to the EPC using cellular radio access 281 technology which is further connected to the Internet. At the other 282 end, the fixed user is connected to the Internet via wired connection 283 with sufficiently high bandwidth, for instance, 10 Gbps, so that the 284 system bottleneck is on the cellular radio access interface. The 285 wired connection to in this setup does not introduce any network 286 impairments to the test; it only adds 10 ms of one-way propagation 287 delay. 289 The path from the fixed user to the mobile users is defined as 290 "Downlink" and the path from the mobile users to the fixed user is 291 defined as "Uplink". We assume that only uplink or downlink is 292 congested for mobile users. Hence, we recommend that the uplink and 293 downlink simulations are run separately. 295 uplink 296 ++))) +--------------------------> 297 ++-+ ((o)) 298 | | / \ +-------+ +------+ +---+ 299 +--+ / \----+ +-----+ +----+ | 300 / \ +-------+ +------+ +---+ 301 UE BS EPC Internet fixed 302 <--------------------------+ 303 downlink 305 Figure 1: Simulation Topology 307 3.1.2. Simulation Setup 309 The values enclosed within "[ ]" for the following simulation 310 attributes follow the same notion as in [I-D.ietf-rmcat-eval-test]. 311 The desired simulation setup is as follows -- 313 1. Radio environment: 315 A. Deployment and propagation model: 3GPP case 1 (see 316 [HO-deploy-3GPP]) 318 B. Antenna: Multiple-Input and Multiple-Output (MIMO), [2D, 3D] 320 C. Mobility: [3km/h, 30km/h] 322 D. Transmission bandwidth: 10Mhz 324 E. Number of cells: multi-cell deployment (3 Cells per Base 325 Station (BS) * 7 BS) = 21 cells 327 F. Cell radius: 166.666 Meters 329 G. Scheduler: Proportional fair with no priority 331 H. Bearer: Default bearer for all traffic. 333 I. Active Queue Management (AQM) settings: AQM [on,off] 335 2. End-to-end Round Trip Time (RTT): [40ms, 150ms] 337 3. User arrival model: Poisson arrival model 339 4. User intensity: 341 * Downlink user intensity: {0.7, 1.4, 2.1, 2.8, 3.5, 4.2, 4.9, 342 5.6, 6.3, 7.0, 7.7, 8.4, 9,1, 9.8, 10.5} 344 * Uplink user intensity : {0.7, 1.4, 2.1, 2.8, 3.5, 4.2, 4.9, 345 5.6, 6.3, 7.0} 347 5. Simulation duration: 91s 349 6. Evaluation period: 30s-60s 351 7. Media traffic: 353 1. Media type: Video 355 a. Media direction: [Uplink, Downlink] 357 b. Number of Media source per user: One (1) 359 c. Media duration per user: 30s 361 d. Media source: same as defined in Section 4.3 of 362 [I-D.ietf-rmcat-eval-test] 364 2. Media Type: Audio 366 a. Media direction: Uplink and Downlink 368 b. Number of Media source per user: One (1) 370 c. Media duration per user: 30s 372 d. Media codec: Constant Bit Rate (CBR) 374 e. Media bitrate: 20 Kbps 376 f. Adaptation: off 378 8. Other traffic models: 380 * Downlink simulation: Maximum of 4Mbps/cell (web browsing or 381 FTP traffic following default TCP congestion control 382 [RFC5681]) 384 * Unlink simulation: Maximum of 2Mbps/cell (web browsing or FTP 385 traffic following default TCP congestion control [RFC5681]) 387 3.2. Bad Radio Coverage 389 The goal of this test is to evaluate the performance of candidate 390 congestion control algorithm when users visit part of the network 391 with bad radio coverage. The scenario is created by using a larger 392 cell radius than that in the previous test case. In this test case, 393 each user/UE in the media session is an RMCAT compliant endpoint. 394 User arrivals follow a Poisson distribution proportional to the 395 length of the call, to keep the number of users per cell fairly 396 constant during the evaluation period. At the beginning of the 397 simulation, there should be enough amount of time to warm-up the 398 network. This is to avoid running the evaluation in an empty network 399 where network nodes are having empty buffers, low interference at the 400 beginning of the simulation. This network initialization period 401 should be excluded from the evaluation period. 403 This test case also includes user mobility and some competing 404 traffic. The latter includes the same kind of flows (with same 405 adaptation algorithms). The investigated congestion control 406 algorithms should result in maximum possible network utilization and 407 stability in terms of rate variations, lowest possible end to end 408 frame latency, network latency and Packet Loss Rate (PLR) at 409 different cell load levels. 411 3.2.1. Network connection 413 Same as defined in Section 3.1.1 415 3.2.2. Simulation Setup 417 The desired simulation setup is the same as the Varying Network Load 418 test case defined in Section 3.1 except the following changes: 420 1. Radio environment: Same as defined in Section 3.1.2 except the 421 following: 423 A. Deployment and propagation model: 3GPP case 3 (see 424 [HO-deploy-3GPP]) 426 B. Cell radius: 577.3333 Meters 427 C. Mobility: 3km/h 429 2. User intensity = {0.7, 1.4, 2.1, 2.8, 3.5, 4.2, 4.9, 5.6, 6.3, 430 7.0} 432 3. Media traffic model: Same as defined in Section 3.1.2 434 4. Other traffic models: 436 * Downlink simulation: Maximum of 2Mbps/cell (web browsing or 437 FTP traffic following default TCP congestion control 438 [RFC5681]) 440 * Unlink simulation: Maximum of 1Mbps/cell (web browsing or FTP 441 traffic following default TCP congestion control [RFC5681]) 443 3.3. Desired Evaluation Metrics for cellular test cases 445 The evaluation criteria document [I-D.ietf-rmcat-eval-criteria] 446 defines the metrics to be used to evaluate candidate algorithms. 447 Considering the nature and distinction of cellular networks we 448 recommend that at least the following metrics be used to evaluate the 449 performance of the candidate algorithms: 451 o Average cell throughput (for all cells), shows cell utilizations. 453 o Application sending and receiving bitrate, goodput. 455 o Packet Loss Rate (PLR). 457 o End-to-end Media frame delay. For video, this means the delay 458 from capture to display. 460 o Transport delay. 462 o Algorithm stability in terms of rate variation. 464 4. Wi-Fi Networks Specific Test Cases 466 Given the prevalence of Internet access links over Wi-Fi, it is 467 important to evaluate candidate RTP-based congestion control 468 solutions over test cases that include Wi-Fi access links. Such 469 evaluations should highlight the inherently different characteristics 470 of Wi-Fi networks in contrast to their wired counterparts: 472 o The wireless radio channel is subject to interference from nearby 473 transmitters, multipath fading, and shadowing. These effects lead 474 to fluctuations in the link throughput and sometimes an error- 475 prone communication environment. 477 o Available network bandwidth is not only shared over the air 478 between concurrent users but also between uplink and downlink 479 traffic due to the half-duplex nature of the wireless transmission 480 medium. 482 o Packet transmissions over Wi-Fi are susceptible to contentions and 483 collisions over the air. Consequently, traffic load beyond a 484 certain utilization level over a Wi-Fi network can introduce 485 frequent collisions over the air and significant network overhead, 486 as well as packet drops due to buffer overflow at the 487 transmitters. This, in turn, leads to excessive delay, 488 retransmissions, packet losses and lower effective bandwidth for 489 applications. Note further that the collision-induced delay and 490 loss patterns are qualitatively different from those caused by 491 congestion over a wired connection. 493 o The IEEE 802.11 standard (i.e., Wi-Fi) supports multi-rate 494 transmission capabilities by dynamically choosing the most 495 appropriate modulation and coding scheme (MCS) for the given 496 received signal strength. A different choice in the MCS Index 497 leads to different physical-layer (PHY-layer) link rates and 498 consequently different application-layer throughput. 500 o The presence of legacy devices (e.g., ones operating only in IEEE 501 802.11b) at a much lower PHY-layer link rate can significantly 502 slow down the rest of a modern Wi-Fi network. As discussed in 503 [Heusse2003], the main reason for such anomaly is that it takes 504 much longer to transmit the same packet over a slower link than 505 over a faster link, thereby consuming a substantial portion of air 506 time. 508 o Handover from one Wi-Fi Access Point (AP) to another may lead to 509 excessive packet delays and losses during the process. 511 o IEEE 802.11e has introduced the Enhanced Distributed Channel 512 Access (EDCA) mechanism to allow different traffic categories to 513 contend for channel access using different random back-off 514 parameters. This mechanism is a mandatory requirement for the Wi- 515 Fi Multimedia (WMM) certification in Wi-Fi Alliance. It allows 516 for prioritization of real-time application traffic such as voice 517 and video over non-urgent data transmissions (e.g., file 518 transfer). 520 In summary, the presence of Wi-Fi access links in different network 521 topologies can exert different impact on the network performance in 522 terms of application-layer effective throughput, packet loss rate, 523 and packet delivery delay. These, in turn, will influence the 524 behavior of end-to-end real-time multimedia congestion control. 526 Unless otherwise mentioned, the test cases in this section choose the 527 PHY- and MAC-layer parameters based on the IEEE 802.11n Standard. 528 Statistics collected from enterprise Wi-Fi networks show that the two 529 dominant physical modes are 802.11n and 802.11ac, accounting for 41% 530 and 58% of connected devices. As Wi-Fi standards evolve over time -- 531 for instance, with the introduction of the emerging Wi-Fi 6 (based on 532 IEEE 802.11ax) products -- the PHY- and MAC-layer test case 533 specifications need to be updated accordingly to reflect such 534 changes. 536 Typically, a Wi-Fi access network connects to a wired infrastructure. 537 Either the wired or the Wi-Fi segment of the network can be the 538 bottleneck. The following sections describe basic test cases for 539 both scenarios separately. The same set of performance metrics as in 540 [I-D.ietf-rmcat-eval-test]) should be collected for each test case. 542 We recommend to carry out the test cases as defined in this document 543 using a simulator, such as [NS-2] or [NS-3]. When feasible, it is 544 encouraged to perform testbed-based evaluations using Wi-Fi access 545 points and endpoints running up-to-date IEEE 802.11 protocols, such 546 as 802.11ac and the emerging Wi-Fi 6, so as to verify the viability 547 of the candidate schemes. 549 4.1. Bottleneck in Wired Network 551 The test scenarios below are intended to mimic the setup of video 552 conferencing over Wi-Fi connections from the home. Typically, the 553 Wi-Fi home network is not congested and the bottleneck is present 554 over the wired home access link. Although it is expected that test 555 evaluation results from this section are similar to those as in 556 [I-D.ietf-rmcat-eval-test], it is still worthwhile to run through 557 these tests as sanity checks. 559 4.1.1. Network topology 561 Figure 2 shows the network topology of Wi-Fi test cases. The test 562 contains multiple mobile nodes (MNs) connected to a common Wi-Fi 563 access point (AP) and their corresponding wired clients on fixed 564 nodes (FNs). Each connection carries either a RTP-based media flow 565 or a TCP traffic flow. Directions of the flows can be uplink (i.e., 566 from mobile nodes to fixed nodes), downlink (i.e., from fixed nodes 567 to mobile nodes), or bi-directional. The total number of 568 uplink/downlink/bi-directional flows for RTP-based media traffic and 569 TCP traffic are denoted as N and M, respectively. 571 Uplink 572 +----------------->+ 573 +------+ +------+ 574 | MN_1 |)))) /=====| FN_1 | 575 +------+ )) // +------+ 576 . )) // . 577 . )) // . 578 . )) // . 579 +------+ +----+ +-----+ +------+ 580 | MN_N | ))))))) | | | |========| FN_N | 581 +------+ | | | | +------+ 582 | AP |=========| FN0 | 583 +----------+ | | | | +----------+ 584 | MN_tcp_1 | )))) | | | |======| FN_tcp_1 | 585 +----------+ +----+ +-----+ +----------+ 586 . )) \\ . 587 . )) \\ . 588 . )) \\ . 589 +----------+ )) \\ +----------+ 590 | MN_tcp_M |))) \=====| FN_tcp_M | 591 +----------+ +----------+ 592 +<-----------------+ 593 Downlink 595 Figure 2: Network topology for Wi-Fi test cases 597 4.1.2. Test setup 599 o Test duration: 120s 601 o Wi-Fi network characteristics: 603 * Radio propagation model: Log-distance path loss propagation 604 model (see [NS3WiFi]) 606 * PHY- and MAC-layer configuration: IEEE 802.11n 608 * MCS Index at 11: 16-QAM 1/2, Raw Data Rate at 52Mbps 610 o Wired path characteristics: 612 * Path capacity: 1Mbps 614 * One-Way propagation delay: 50ms. 616 * Maximum end-to-end jitter: 30ms 618 * Bottleneck queue type: Drop tail. 620 * Bottleneck queue size: 300ms. 622 * Path loss ratio: 0%. 624 o Application characteristics: 626 * Media Traffic: 628 + Media type: Video 630 + Media direction: See Section 4.1.3 632 + Number of media sources (N): See Section 4.1.3 634 + Media timeline: 636 - Start time: 0s. 638 - End time: 119s. 640 * Competing traffic: 642 + Type of sources: long-lived TCP or CBR over UDP 644 + Traffic direction: See Section 4.1.3 646 + Number of sources (M): See Section 4.1.3 648 + Congestion control: Default TCP congestion control [RFC5681] 649 or constant-bit-rate (CBR) traffic over UDP. 651 + Traffic timeline: See Section 4.1.3 653 4.1.3. Typical test scenarios 655 o Single uplink RTP-based media flow: N=1 with uplink direction and 656 M=0. 658 o One pair of bi-directional RTP-based media flows: N=2 (i.e., one 659 uplink flow and one downlink flow); M=0. 661 o One pair of bi-directional RTP-based media flows: N=2; one uplink 662 on-off CBR flow over UDP: M=1 (uplink). The CBR flow has ON time 663 at t=0s-60s and OFF time at t=60s-119s. 665 o One pair of bi-directional RTP-based media flows: N=2; one uplink 666 off-on CBR flow over UDP: M=1 (uplink). The CBR flow has OFF time 667 at t=0s-60s and ON time at t=60s-119s. 669 o One RTP-based media flow competing against one long-live TCP flow 670 in the uplink direction: N=1 (uplink) and M = 1(uplink). The TCP 671 flow has start time at t=0s and end time at t=119s. 673 4.1.4. Expected behavior 675 o Single uplink RTP-based media flow: the candidate algorithm is 676 expected to detect the path capacity constraint, to converge to 677 the bottleneck link capacity, and to adapt the flow to avoid 678 unwanted oscillations when the sending bit rate is approaching the 679 bottleneck link capacity. No excessive oscillations in the media 680 rate should be present. 682 o Bi-directional RTP-based media flows: the candidate algorithm is 683 expected to converge to the bottleneck capacity of the wired path 684 in both directions despite the presence of measurement noise over 685 the Wi-Fi connection. In the presence of background TCP or CBR 686 over UDP traffic, the rate of RTP-based media flows should adapt 687 promptly to the arrival and departure of background traffic flows. 689 o One RTP-based media flow competing with long-live TCP flow in the 690 uplink direction: the candidate algorithm is expected to avoid 691 congestion collapse and to stabilize at a fair share of the 692 bottleneck link capacity. 694 4.2. Bottleneck in Wi-Fi Network 696 The test cases in this section assume that the wired segment along 697 the media path is well-provisioned whereas the bottleneck exists over 698 the Wi-Fi access network. This is to mimic the application scenarios 699 typically encountered by users in an enterprise environment or at a 700 coffee house. 702 4.2.1. Network topology 704 Same as defined in Section 4.1.1 706 4.2.2. Test setup 708 o Test duration: 120s 710 o Wi-Fi network characteristics: 712 * Radio propagation model: Log-distance path loss propagation 713 model (see [NS3WiFi]) 715 * PHY- and MAC-layer configuration: IEEE 802.11n 716 * MCS Index at 11: 16-QAM 1/2, Raw Data Rate at 52Mbps 718 o Wired path characteristics: 720 * Path capacity: 100Mbps. 722 * One-Way propagation delay: 50ms. 724 * Maximum end-to-end jitter: 30ms. 726 * Bottleneck queue type: Drop tail. 728 * Bottleneck queue size: 300ms. 730 * Path loss ratio: 0%. 732 o Application characteristics: 734 * Media Traffic: 736 + Media type: Video 738 + Media direction: See Section 4.2.3. 740 + Number of media sources (N): See Section 4.2.3. 742 + Media timeline: 744 - Start time: 0s. 746 - End time: 119s. 748 * Competing traffic: 750 + Type of sources: long-lived TCP or CBR over UDP. 752 + Number of sources (M): See Section 4.2.3. 754 + Traffic direction: See Section 4.2.3. 756 + Congestion control: Default TCP congestion control [RFC5681] 757 or constant-bit-rate (CBR) traffic over UDP. 759 + Traffic timeline: See Section 4.2.3. 761 4.2.3. Typical test scenarios 763 This section describes a few test scenarios that are deemed as 764 important for understanding the behavior of a candidate RTP-based 765 congestion control scheme over a Wi-Fi network. 767 a. Multiple RTP-based media flows sharing the wireless downlink: 768 N=16 (all downlink); M = 0. This test case is for studying the 769 impact of contention on the multiple concurrent media flows. For 770 an 802.11n network, given the MCS Index of 11 and the 771 corresponding link rate of 52Mbps, the total application-layer 772 throughput (assuming reasonable distance, low interference and 773 infrequent contentions caused by competing streams) is around 774 20Mbps. A total of N=16 RTP-based media flows (with a maximum 775 rate of 1.5Mbps each) are expected to saturate the wireless 776 interface in this experiment. Evaluation of a given candidate 777 scheme should focus on whether the downlink media flows can 778 stabilize at a fair share of the total application-layer 779 throughput. 781 b. Multiple RTP-based media flows sharing the wireless uplink:N = 16 782 (all downlink); M = 0. When multiple clients attempt to transmit 783 media packets uplink over the Wi-Fi network, they introduce more 784 frequent contentions and potential collisions. Per-flow 785 throughput is expected to be lower than that in the previous 786 downlink-only scenario. Evaluation of a given candidate scheme 787 should focus on whether the uplink flows can stabilize at a fair 788 share of the total application-layer throughput. 790 c. Multiple bi-directional RTP-based media flows: N = 16 (8 uplink 791 and 8 downlink); M = 0. The goal of this test is to evaluate the 792 performance of the candidate scheme in terms of bandwidth 793 fairness between uplink and downlink flows. 795 d. Multiple bi-directional RTP-based media flows with on-off CBR 796 traffic over UDP: N = 16 (8 uplink and 8 downlink); M = 5 797 (uplink). The goal of this test is to evaluate the adaptation 798 behavior of the candidate scheme when its available bandwidth 799 changes due to the departure of background traffic. The 800 background traffic consists of several (e.g., M=5) CBR flows 801 transported over UDP. These background flows are ON at time 802 t=0-60s and OFF at time t=61-120s. 804 e. Multiple bi-directional RTP-based media flows with off-on CBR 805 traffic over UDP: N = 16 (8 uplink and 8 downlink); M = 5 806 (uplink). The goal of this test is to evaluate the adaptation 807 behavior of the candidate scheme when its available bandwidth 808 changes due to the arrival of background traffic. The background 809 traffic consists of several (e.g., M=5) parallel CBR flows 810 transported over UDP. These background flows are OFF at time 811 t=0-60s and ON at times t=61-120s. 813 f. Multiple bi-directional RTP-based media flows in the presence of 814 background TCP traffic: N=16 (8 uplink and 8 downlink); M = 5 815 (uplink). The goal of this test is to evaluate how RTP-based 816 media flows compete against TCP over a congested Wi-Fi network 817 for a given candidate scheme. TCP flows have start time at t=40s 818 and end time at t=80s. 820 g. Varying number of RTP-based media flows: A series of tests can be 821 carried out for the above test cases with different values of N, 822 e.g., N = [4, 8, 12, 16, 20]. The goal of this test is to 823 evaluate how a candidate scheme responds to varying traffic load/ 824 demand over a congested Wi-Fi network. The start times of the 825 media flows are randomly distributes within a window of t=0-10s; 826 their end times are randomly distributed within a window of 827 t=110-120s. 829 4.2.4. Expected behavior 831 o Multiple downlink RTP-based media flows: each media flow is 832 expected to get its fair share of the total bottleneck link 833 bandwidth. Overall bandwidth usage should not be significantly 834 lower than that experienced by the same number of concurrent 835 downlink TCP flows. In other words, the behavior of multiple 836 concurrent TCP flows will be used as a performance benchmark for 837 this test scenario. The end-to-end delay and packet loss ratio 838 experienced by each flow should be within an acceptable range for 839 real-time multimedia applications. 841 o Multiple uplink RTP-based media flows: overall bandwidth usage by 842 all media flows should not be significantly lower than that 843 experienced by the same number of concurrent uplink TCP flows. In 844 other words, the behavior of multiple concurrent TCP flows will be 845 used as a performance benchmark for this test scenario. 847 o Multiple bi-directional RTP-based media flows with dynamic 848 background traffic carrying CBR flows over UDP: the media flows 849 are expected to adapt in a timely fashion to the changes in 850 available bandwidth introduced by the arrival/departure of 851 background traffic. 853 o Multiple bi-directional RTP-based media flows with dynamic 854 background traffic over TCP: during the presence of TCP background 855 flows, the overall bandwidth usage by all media flows should not 856 be significantly lower than those achieved by the same number of 857 bi-directional TCP flows. In other words, the behavior of 858 multiple concurrent TCP flows will be used as a performance 859 benchmark for this test scenario. All downlink media flows are 860 expected to obtain similar bandwidth as each other. The 861 throughput of each media flow is expected to decrease upon the 862 arrival of TCP background traffic and, conversely, increase upon 863 their departure. Both reactions should occur in a timely fashion, 864 for example, within 10s of seconds. 866 o Varying number of bi-directional RTP-based media flows: the test 867 results for varying values of N -- while keeping all other 868 parameters constant -- is expected to show steady and stable per- 869 flow throughput for each value of N. The average throughput of 870 all media flows is expected to stay constant around the maximum 871 rate when N is small, then gradually decrease with increasing 872 value of N till it reaches the minimum allowed rate, beyond which 873 the offered load to the Wi-Fi network exceeds its capacity (i.e., 874 with a very large value of N). 876 4.3. Other Potential Test Cases 878 4.3.1. EDCA/WMM usage 880 The EDCA/WMM mechanism defines prioritized QoS for four traffic 881 classes (or Access Categories). RTP-based real-time media flows 882 should achieve better performance in terms of lower delay and fewer 883 packet losses with EDCA/WMM enabled when competing against non- 884 interactive background traffic such as file transfers. When most of 885 the traffic over Wi-Fi is dominated by media, however, turning on WMM 886 may degrade performance since all media flows now attempt to access 887 the wireless transmission medium more aggressively, thereby causing 888 more frequent collisions and collision-induced losses. This is a 889 topic worthy of further investigation. 891 4.3.2. Effect of heterogeneous link rates 893 As discussed in [Heusse2003], the presence of clients operating over 894 slow PHY-layer link rates (e.g., a legacy 802.11b device) connected 895 to a modern network may adversely impact the overall performance of 896 the network. Additional test cases can be devised to evaluate the 897 effect of clients with heterogeneous link rates on the performance of 898 the candidate congestion control algorithm. Such test cases, for 899 instance, can specify that the PHY-layer link rates for all clients 900 span over a wide range (e.g., 2Mbps to 54Mbps) for investigating its 901 effect on the congestion control behavior of the real-time 902 interactive applications. 904 5. IANA Considerations 906 This memo includes no request to IANA. 908 6. Security Considerations 910 The security considerations in [I-D.ietf-rmcat-eval-criteria] and the 911 relevant congestion control algorithms apply. The principles for 912 congestion control are described in [RFC2914], and in particular, any 913 new method MUST implement safeguards to avoid congestion collapse of 914 the Internet. 916 The evaluations of the test cases are intended to carry out in a 917 controlled lab environment. Hence, the applications, simulators and 918 network nodes ought to be well-behaved and should not impact the 919 desired results. It is important to take appropriate caution to 920 avoid leaking non-responsive traffic with unproven congestion 921 avoidance behavior onto the open Internet. 923 7. Acknowledgments 925 The authors would like to thank Tomas Frankkila, Magnus Westerlund, 926 Kristofer Sandlund, Sergio Mena de la Cruz, and Mirja Kuehlewind for 927 their valuable inputs and review comments regarding this draft. 929 8. References 931 8.1. Normative References 933 [HO-deploy-3GPP] 934 TS 25.814, 3GPP., "Physical layer aspects for evolved 935 Universal Terrestrial Radio Access (UTRA)", October 2006, 936 . 939 [I-D.ietf-rmcat-eval-criteria] 940 Singh, V., Ott, J., and S. Holmer, "Evaluating Congestion 941 Control for Interactive Real-time Media", draft-ietf- 942 rmcat-eval-criteria-11 (work in progress), February 2020. 944 [I-D.ietf-rmcat-eval-test] 945 Sarker, Z., Singh, V., Zhu, X., and M. Ramalho, "Test 946 Cases for Evaluating RMCAT Proposals", draft-ietf-rmcat- 947 eval-test-10 (work in progress), May 2019. 949 [IEEE802.11] 950 IEEE, "Standard for Information technology-- 951 Telecommunications and information exchange between 952 systems Local and metropolitan area networks--Specific 953 requirements Part 11: Wireless LAN Medium Access Control 954 (MAC) and Physical Layer (PHY) Specifications", 2012. 956 [NS3WiFi] "Wi-Fi Channel Model in ns-3 Simulator", 957 . 960 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 961 Requirement Levels", BCP 14, RFC 2119, 962 DOI 10.17487/RFC2119, March 1997, 963 . 965 [RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion 966 Control", RFC 5681, DOI 10.17487/RFC5681, September 2009, 967 . 969 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 970 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 971 May 2017, . 973 8.2. Informative References 975 [Heusse2003] 976 Heusse, M., Rousseau, F., Berger-Sabbatel, G., and A. 977 Duda, "Performance anomaly of 802.11b", in Proc. 23th 978 Annual Joint Conference of the IEEE Computer and 979 Communications Societies, (INFOCOM'03), March 2003. 981 [HO-def-3GPP] 982 TR 21.905, 3GPP., "Vocabulary for 3GPP Specifications", 983 December 2009, . 986 [HO-LTE-3GPP] 987 TS 36.331, 3GPP., "E-UTRA- Radio Resource Control (RRC); 988 Protocol specification", December 2011, 989 . 992 [HO-UMTS-3GPP] 993 TS 25.331, 3GPP., "Radio Resource Control (RRC); Protocol 994 specification", December 2011, 995 . 998 [I-D.ietf-rmcat-cc-requirements] 999 Jesup, R. and Z. Sarker, "Congestion Control Requirements 1000 for Interactive Real-Time Media", draft-ietf-rmcat-cc- 1001 requirements-09 (work in progress), December 2014. 1003 [NS-2] "ns-2", December 2014, 1004 . 1006 [NS-3] "ns-3 Network Simulator", . 1008 [QoS-3GPP] 1009 TS 23.203, 3GPP., "Policy and charging control 1010 architecture", June 2011, . 1013 [RFC2914] Floyd, S., "Congestion Control Principles", BCP 41, 1014 RFC 2914, DOI 10.17487/RFC2914, September 2000, 1015 . 1017 Authors' Addresses 1019 Zaheduzzaman Sarker 1020 Ericsson AB 1021 Laboratoriegraend 11 1022 Luleae 97753 1023 Sweden 1025 Phone: +46 107173743 1026 Email: zaheduzzaman.sarker@ericsson.com 1028 Ingemar Johansson 1029 Ericsson AB 1030 Laboratoriegraend 11 1031 Luleae 97753 1032 Sweden 1034 Phone: +46 10 7143042 1035 Email: ingemar.s.johansson@ericsson.com 1037 Xiaoqing Zhu 1038 Cisco Systems 1039 12515 Research Blvd., Building 4 1040 Austin, TX 78759 1041 USA 1043 Email: xiaoqzhu@cisco.com 1044 Jiantao Fu 1045 Cisco Systems 1046 771 Alder Drive 1047 Milpitas, CA 95035 1048 USA 1050 Email: jianfu@cisco.com 1052 Wei-Tian Tan 1053 Cisco Systems 1054 510 McCarthy Blvd 1055 Milpitas, CA 95035 1056 USA 1058 Email: dtan2@cisco.com 1060 Michael A. Ramalho 1061 AcousticComms Consulting 1062 6310 Watercrest Way Unit 203 1063 Lakewood Ranch, FL 34202-5211 1064 USA 1066 Phone: +1 732 832 9723 1067 Email: mar42@cornell.edu