idnits 2.17.1 draft-ietf-rmcat-wireless-tests-08.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 302 has weird spacing: '...nternet fix...' -- The document date (July 5, 2019) is 1728 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 -- Looks like a reference, but probably isn't: '40' on line 335 -- Looks like a reference, but probably isn't: '150' on line 335 == 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 813 -- Looks like a reference, but probably isn't: '8' on line 813 -- Looks like a reference, but probably isn't: '12' on line 813 -- Looks like a reference, but probably isn't: '16' on line 813 -- Looks like a reference, but probably isn't: '20' on line 813 == Unused Reference: 'I-D.ietf-rmcat-cc-requirements' is defined on line 985, but no explicit reference was found in the text == Outdated reference: A later version (-14) exists of draft-ietf-rmcat-eval-criteria-08 Summary: 0 errors (**), 0 flaws (~~), 8 warnings (==), 8 comments (--).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: January 6, 2020 X. Zhu 6 J. Fu 7 W. Tan 8 M. Ramalho 9 Cisco Systems 10 July 5, 2019 12 Evaluation Test Cases for Interactive Real-Time Media over Wireless 13 Networks 14 draft-ietf-rmcat-wireless-tests-08 16 Abstract 18 The Real-time Transport Protocol (RTP) is a common transport choice 19 for interactive multimedia communication applications. The 20 performance of such applications typically depends on a well- 21 functioning congestion control algorithm. To ensure seamless and 22 robust user experience, a well-designed RTP-based congestion control 23 algorithm should work well across all access network types. This 24 document describes test cases for evaluating performances of such 25 congestion control algorithms over LTE and Wi-Fi networks. 27 Status of This Memo 29 This Internet-Draft is submitted in full conformance with the 30 provisions of BCP 78 and BCP 79. 32 Internet-Drafts are working documents of the Internet Engineering 33 Task Force (IETF). Note that other groups may also distribute 34 working documents as Internet-Drafts. The list of current Internet- 35 Drafts is at https://datatracker.ietf.org/drafts/current/. 37 Internet-Drafts are draft documents valid for a maximum of six months 38 and may be updated, replaced, or obsoleted by other documents at any 39 time. It is inappropriate to use Internet-Drafts as reference 40 material or to cite them other than as "work in progress." 42 This Internet-Draft will expire on January 6, 2020. 44 Copyright Notice 46 Copyright (c) 2019 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents 51 (https://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the Simplified BSD License. 59 Table of Contents 61 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 62 2. Terminologies . . . . . . . . . . . . . . . . . . . . . . . . 3 63 3. Cellular Network Specific Test Cases . . . . . . . . . . . . 3 64 3.1. Varying Network Load . . . . . . . . . . . . . . . . . . 6 65 3.1.1. Network Connection . . . . . . . . . . . . . . . . . 6 66 3.1.2. Simulation Setup . . . . . . . . . . . . . . . . . . 7 67 3.2. Bad Radio Coverage . . . . . . . . . . . . . . . . . . . 9 68 3.2.1. Network connection . . . . . . . . . . . . . . . . . 9 69 3.2.2. Simulation Setup . . . . . . . . . . . . . . . . . . 9 70 3.3. Desired Evaluation Metrics for cellular test cases . . . 10 71 4. Wi-Fi Networks Specific Test Cases . . . . . . . . . . . . . 10 72 4.1. Bottleneck in Wired Network . . . . . . . . . . . . . . . 12 73 4.1.1. Network topology . . . . . . . . . . . . . . . . . . 12 74 4.1.2. Test setup . . . . . . . . . . . . . . . . . . . . . 13 75 4.1.3. Typical test scenarios . . . . . . . . . . . . . . . 14 76 4.1.4. Expected behavior . . . . . . . . . . . . . . . . . . 15 77 4.2. Bottleneck in Wi-Fi Network . . . . . . . . . . . . . . . 15 78 4.2.1. Network topology . . . . . . . . . . . . . . . . . . 15 79 4.2.2. Test setup . . . . . . . . . . . . . . . . . . . . . 15 80 4.2.3. Typical test scenarios . . . . . . . . . . . . . . . 17 81 4.2.4. Expected behavior . . . . . . . . . . . . . . . . . . 18 82 4.3. Other Potential Test Cases . . . . . . . . . . . . . . . 19 83 4.3.1. EDCA/WMM usage . . . . . . . . . . . . . . . . . . . 19 84 4.3.2. Effects of Legacy 802.11b Devices . . . . . . . . . . 19 85 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 19 86 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20 87 7. Security Considerations . . . . . . . . . . . . . . . . . . . 20 88 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 20 89 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 20 90 9.1. Normative References . . . . . . . . . . . . . . . . . . 20 91 9.2. Informative References . . . . . . . . . . . . . . . . . 21 92 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 22 94 1. Introduction 96 Wireless networks (both cellular and Wi-Fi [IEEE802.11]) are an 97 integral part of the Internet. Mobile devices connected to the 98 wireless networks account for an increasingly more significant 99 portion of the media traffic over the Internet. Application 100 scenarios range from video conferencing calls in a bus or train to 101 media consumption by someone on a living room couch. 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. Even 105 though basic test cases for evaluating RTP-based congestion control 106 schemes as defined in [I-D.ietf-rmcat-eval-test] have covered many 107 effects of the impairments common to both wired and wireless 108 networks, there remain characteristics and dynamics unique to a given 109 wireless environment. For example, in LTE networks, the base station 110 maintains individual queues per radio bearer per user hence it leads 111 to a different nature of interactions between traffic flows of 112 different users. This contrasts with wired networks, 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 RMCAT evaluation criteria document [I-D.ietf-rmcat-eval-criteria] 120 provides the guideline for evaluating candidate algorithms and 121 recognizes the importance of testing over wireless access networks. 122 However, it does not describe any specific test cases for performance 123 evaluation of candidate algorithms. This document describes test 124 cases specifically targeting cellular networks such as LTE networks 125 and Wi-Fi networks. 127 2. Terminologies 129 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 130 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 131 "OPTIONAL" in this document are to be interpreted as described in BCP 132 14 [RFC2119] [RFC8174] when, and only when, they appear in all 133 capitals, as shown here. 135 3. Cellular Network Specific Test Cases 137 A cellular environment is more complicated than its wireline 138 counterpart since it seeks to provide services in the context of 139 variable available bandwidth, location dependencies and user 140 mobilities at different speeds. In a cellular network, the user may 141 reach the cell edge which may lead to a significant amount of 142 retransmissions to deliver the data from the base station to the 143 destination and vice versa. These network links or radio links will 144 often act as a bottleneck for the rest of the network and will 145 eventually lead to excessive delays or packet drops. An efficient 146 retransmission or link adaptation mechanism can reduce the packet 147 loss probability but there will still be some packet losses and delay 148 variations. Moreover, with increased cell load or handover to a 149 congested cell, congestion in the transport network will become even 150 worse. Besides, there are certain characteristics which make the 151 cellular network different from and more challenging than other types 152 of access networks such as Wi-Fi and wired network. In a cellular 153 network -- 155 o The bottleneck is often a shared link with relatively few users. 157 * The cost per bit over the shared link varies over time and is 158 different for different users. 160 * Leftover/unused resource can be consumed by other greedy users. 162 o Queues are always per radio bearer hence each user can have many 163 of such queues. 165 o Users can experience both Inter and Intra Radio Access Technology 166 (RAT) handovers (see [HO-def-3GPP] for the definition of 167 "handover"). 169 o Handover between cells or change of serving cells (as described in 170 [HO-LTE-3GPP] and [HO-UMTS-3GPP]) might cause user plane 171 interruptions which can lead to bursts of packet losses, delay 172 and/or jitter. The exact behavior depends on the type of radio 173 bearer. Typically, the default best-effort bearers do not 174 generate packet loss, instead, packets are queued up and 175 transmitted once the handover is completed. 177 o The network part decides how much the user can transmit. 179 o The cellular network has variable link capacity per user 181 * Can vary as fast as a period of milliseconds. 183 * Depends on many factors (such as distance, speed, interference, 184 different flows). 186 * Uses complex and smart link adaptation which makes the link 187 behavior ever more dynamic. 189 * The scheduling priority depends on the estimated throughput. 191 o Both Quality of Service (QoS) and non-QoS radio bearers can be 192 used. 194 Hence, a real-time communication application operating in such a 195 cellular network needs to cope with a shared bottleneck link and 196 variable link capacity, events like handover, non-congestion related 197 loss, abrupt changes in bandwidth (both short term and long term) due 198 to handover, network load and bad radio coverage. Even though 3GPP 199 define QoS bearers [QoS-3GPP] to ensure high-quality user experience, 200 adaptive real-time applications are desired. 202 Different mobile operators deploy their own cellular network 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 3G and 4G radio technologies can 207 support the high bandwidth requirements from real-time interactive 208 video applications. The future real-time interactive application 209 will impose even greater demand on cellular network performance which 210 makes 4G (and beyond radio technologies) more suitable access 211 technology for such genre of application. 213 The key factors to define test cases for cellular networks are 215 o Shared and varying link capacity 217 o Mobility 219 o Handover 221 However, for cellular networks, it is very hard to separate such 222 events from one another as these events are heavily related. Hence 223 instead of devising separate test cases for all those important 224 events, we have divided the test case into two categories. It should 225 be noted that the goal of the following test cases is to evaluate the 226 performance of candidate algorithms over the radio interface of the 227 cellular network. Hence it is assumed that the radio interface is 228 the bottleneck link between the communicating peers and that the core 229 network does not add any extra congestion in the path. Also, the 230 combination of multiple access technologies such as one user has LTE 231 connection and another has Wi-Fi connection is kept out of the scope 232 of this document. However, later those additional scenarios can also 233 be added in this list of test cases. While defining the test cases 234 we assumed a typical real-time telephony scenario over cellular 235 networks where one real-time session consists of one voice stream and 236 one video stream. 238 Even though it is possible to carry out tests over operational LTE 239 (and 5G) networks, and actually such tests are already available 240 today, these tests cannot in the general case be carried out in a 241 deterministic fashion or to ensure repeatability. The main reason is 242 that these networks are in the control of cellular operators and 243 there exist various amounts of competing traffic in the same cell(s). 244 In practice, it is only in underground mines that one can carry out 245 near deterministic testing. Even there, it is not guaranteed either 246 as workers in the mines may carry with them their personal mobile 247 phones. Furthermore, the underground mining setting may not reflect 248 typical usage patterns in an urban setting. We, therefore, recommend 249 that an LTE network simulator is used for the test cases defined in 250 this document, for example --- NS-3 LTE simulator [LTE-simulator]. 252 3.1. Varying Network Load 254 The goal of this test is to evaluate the performance of the candidate 255 congestion control algorithm under varying network load. The network 256 load variation is created by adding and removing network users a.k.a. 257 User Equipments (UEs) during the simulation. In this test case, each 258 of the user/UE in the media session is an RMCAT compliant endpoint. 259 The arrival of users follows a Poisson distribution proportional to 260 the length of the call so as to keep the number of users per cell 261 fairly constant during the evaluation period. At the beginning of 262 the simulation, there should be enough time to warm-up the network. 263 This is to avoid running the evaluation in an empty network where 264 network nodes are having empty buffers, low interference at the 265 beginning of the simulation. This network initialization period is 266 therefore excluded from the evaluation period. 268 This test case also includes user mobility and some competing 269 traffic. The latter includes both same kind of flows (with same 270 adaptation algorithms) and different kind of flows (with different 271 services and congestion control schemes). The investigated 272 congestion control algorithms should show maximum possible network 273 utilization and stability in terms of rate variations, lowest 274 possible end to end frame latency, network latency and Packet Loss 275 Rate (PLR) at different cell load level. 277 3.1.1. Network Connection 279 Each mobile user is connected to a fixed user. The connection 280 between the mobile user and fixed user consists of an LTE radio 281 access, an Evolved Packet Core (EPC) and an Internet connection. The 282 mobile user is connected to the EPC using LTE radio access technology 283 which is further connected to the Internet. The fixed user is 284 connected to the Internet via wired connection with sufficiently high 285 bandwidth, for instance, 10 Gbps, so that the system is resource- 286 limited on the wireless interface. The wired connection to the 287 Internet in this setup does not introduce any network impairments to 288 the test; it only adds 10 ms of one-way propagation delay. 290 The path from the fixed user to the mobile users is defined as 291 "Downlink" and the path from the mobile users to the fixed user is 292 defined as "Uplink". We assume that only uplink or downlink is 293 congested for mobile users. Hence, we recommend that the uplink and 294 downlink simulations are run separately. 296 uplink 297 ++))) +--------------------------> 298 ++-+ ((o)) 299 | | / \ +-------+ +------+ +---+ 300 +--+ / \----+ +-----+ +----+ | 301 / \ +-------+ +------+ +---+ 302 UE BS EPC Internet fixed 303 <--------------------------+ 304 downlink 306 Figure 1: Simulation Topology 308 3.1.2. Simulation Setup 310 The values enclosed within "[ ]" for the following simulation 311 attributes follow the same notion as in [I-D.ietf-rmcat-eval-test]. 312 The desired simulation setup is as follows -- 314 1. Radio environment: 316 A. Deployment and propagation model: 3GPP case 1 [Deployment] 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): [40, 150] 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 of the user/UE in the media session is an RMCAT compliant 394 endpoint. The arrival of users follows a Poisson distribution 395 proportional to the length of the call, so as to keep the number of 396 users per cell fairly constant during the evaluation period. At the 397 beginning of the simulation, there should be enough amount of time to 398 warm-up the network. This is to avoid running the evaluation in an 399 empty network where network nodes are having empty buffers, low 400 interference at the beginning of the simulation. This network 401 initialization period is therefore excluded from the evaluation 402 period. 404 This test case also includes user mobility and some competing 405 traffic. The latter includes the same kind of flows (with same 406 adaptation algorithms). The investigated congestion control 407 algorithms should result in maximum possible network utilization and 408 stability in terms of rate variations, lowest possible end to end 409 frame latency, network latency and Packet Loss Rate (PLR) at 410 different cell load levels. 412 3.2.1. Network connection 414 Same as defined in Section 3.1.1 416 3.2.2. Simulation Setup 418 The desired simulation setup is the same as the Varying Network Load 419 test case defined in Section 3.1 except the following changes: 421 1. Radio environment: Same as defined in Section 3.1.2 except the 422 following: 424 A. Deployment and propagation model: 3GPP case 3 [Deployment] 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 RMCAT evaluation criteria document [I-D.ietf-rmcat-eval-criteria] 446 defines the metrics to be used to evaluate candidate algorithms. 447 However, looking at the nature and distinction of cellular networks 448 we recommend that at least the following metrics be used to evaluate 449 the performance of the candidate algorithms for the test cases 450 defined in this document. 452 The desired metrics are -- 454 o Average cell throughput (for all cells), shows cell utilizations. 456 o Application sending and receiving bitrate, goodput. 458 o Packet Loss Rate (PLR). 460 o End to end Media frame delay. For video, this means the delay 461 from capture to display. 463 o Transport delay. 465 o Algorithm stability in terms of rate variation. 467 4. Wi-Fi Networks Specific Test Cases 469 Given the prevalence of Internet access links over Wi-Fi, it is 470 important to evaluate candidate RMCAT congestion control solutions 471 over test cases that include Wi-Fi access lines. Such evaluations 472 should also highlight the inherently different characteristics of Wi- 473 Fi networks in contrast to wired networks: 475 o The wireless radio channel is subject to interference from nearby 476 transmitters, multipath fading, and shadowing, causing 477 fluctuations in link throughput and sometimes an error-prone 478 communication environment 480 o Available network bandwidth is not only shared over the air 481 between concurrent users but also between uplink and downlink 482 traffic due to the half-duplex nature of wireless transmission 483 medium. 485 o Packet transmissions over Wi-Fi are susceptible to contentions and 486 collisions over the air. Consequently, traffic load beyond a 487 certain utilization level over a Wi-Fi network can introduce 488 frequent collisions over the air and significant network overhead, 489 as well as packet drops due to buffer overflow at the 490 transmitters. This, in turn, leads to excessive delay, 491 retransmissions, packet losses and lower effective bandwidth for 492 applications. Note, however, that the consequent delay and loss 493 patterns caused by collisions are qualitatively different from 494 those induced by congestion over a wired connection. 496 o The IEEE 802.11 standard (i.e., Wi-Fi) supports multi-rate 497 transmission capabilities by dynamically choosing the most 498 appropriate modulation scheme for the given received signal 499 strength. A different choice of physical-layer rate leads to 500 different application-layer throughput. 502 o Presence of legacy 802.11b networks can significantly slow down 503 the rest of a modern Wi-Fi network. As discussed in [Heusse2003], 504 the main reason for such abnomaly is that it takes longer to 505 transmit the same packet over a slower link than over a faster 506 link. 508 o Handover from one Wi-Fi Access Point (AP) to another may lead to 509 packet delay and losses during the process. 511 o IEEE 802.11e defined EDCA/WMM (Enhanced DCF Channel Access/Wi-Fi 512 Multi-Media) to give voice and video streams higher priority over 513 pure data applications (e.g., file transfers). 515 In summary, the presence of Wi-Fi access links in different network 516 topologies can exert different impact on the network performance in 517 terms of application-layer effective throughput, packet loss rate, 518 and packet delivery delay. These, in turn, influence the behavior of 519 end-to-end real-time multimedia congestion control. 521 Unless otherwise mentioned, test cases in this section are described 522 using the underlying PHY- and MAC-layer parameters based on the IEEE 523 802.11n Standard. Statistics collected from enterprise Wi-Fi 524 networks show that the two dominant physical modes are 802.11n and 525 802.11ac, accounting for 41% and 58% of connected devices. As Wi-Fi 526 standards evolve over time -- for instance, with the introduction of 527 the emerging Wi-Fi 6 (802.11ax) products -- the PHY- and MAC-layer 528 test case specifications need to be updated accordingly to reflect 529 such changes. 531 Typically, a Wi-Fi access network connects to a wired infrastructure. 532 Either the wired or the Wi-Fi segment of the network could be the 533 bottleneck. In the following sections, we describe basic test cases 534 for both scenarios separately. The same set of performance metrics 535 as in [I-D.ietf-rmcat-eval-test]) should be collected for each test 536 case. 538 All test cases described below can be carried out using simulations, 539 e.g. based on [ns-2] or [ns-3]. When feasible, it is also encouraged 540 to perform testbed-based evaluations using Wi-Fi access points and 541 endpoints running up-to-date IEEE 802.11 protocols, such as 802.11ac 542 and the emerging Wi-Fi 6, to verify the viability of the candidate 543 schemes. 545 4.1. Bottleneck in Wired Network 547 The test scenarios below are intended to mimic the setup of video 548 conferencing over Wi-Fi connections from the home. Typically, the 549 Wi-Fi home network is not congested and the bottleneck is present 550 over the wired home access link. Although it is expected that test 551 evaluation results from this section are similar to those from test 552 cases defined for wired networks (see [I-D.ietf-rmcat-eval-test]), it 553 is still worthwhile to run through these tests as sanity checks. 555 4.1.1. Network topology 557 Figure 2 shows the network topology of Wi-Fi test cases. The test 558 contains multiple mobile nodes (MNs) connected to a common Wi-Fi 559 access point (AP) and their corresponding wired clients on fixed 560 nodes (FNs). Each connection carries either a RMCAT or a TCP traffic 561 flow. Directions of the flows can be uplink, downlink, or bi- 562 directional. 564 Uplink 565 +----------------->+ 566 +------+ +------+ 567 | MN_1 |)))) /=====| FN_1 | 568 +------+ )) // +------+ 569 . )) // . 570 . )) // . 571 . )) // . 572 +------+ +----+ +-----+ +------+ 573 | MN_N | ))))))) | | | |========| FN_N | 574 +------+ | | | | +------+ 575 | AP |=========| FN0 | 576 +----------+ | | | | +----------+ 577 | MN_tcp_1 | )))) | | | |======| MN_tcp_1 | 578 +----------+ +----+ +-----+ +----------+ 579 . )) \\ . 580 . )) \\ . 581 . )) \\ . 582 +----------+ )) \\ +----------+ 583 | MN_tcp_M |))) \=====| MN_tcp_M | 584 +----------+ +----------+ 585 +<-----------------+ 586 Downlink 588 Figure 2: Network topology for Wi-Fi test cases 590 4.1.2. Test setup 592 o Test duration: 120s 594 o Wi-Fi network characteristics: 596 * Radio propagation model: Log-distance path loss propagation 597 model [NS3WiFi] 599 * PHY- and MAC-layer configuration: IEEE 802.11n 601 * MCS Index at 11: 16-QAM 1/2, Raw Data Rate@52Mbps 603 o Wired path characteristics: 605 * Path capacity: 1Mbps 607 * One-Way propagation delay: 50ms. 609 * Maximum end-to-end jitter: 30ms 611 * Bottleneck queue type: Drop tail. 613 * Bottleneck queue size: 300ms. 615 * Path loss ratio: 0%. 617 o Application characteristics: 619 * Media Traffic: 621 + Media type: Video 623 + Media direction: See Section 4.1.3 625 + Number of media sources (N): See Section 4.1.3 627 + Media timeline: 629 - Start time: 0s. 631 - End time: 119s. 633 * Competing traffic: 635 + Type of sources: long-lived TCP or CBR over UDP 637 + Traffic direction: See Section 4.1.3 639 + Number of sources (M): See Section 4.1.3 641 + Congestion control: Default TCP congestion control [RFC5681] 642 or constant-bit-rate (CBR) traffic over UDP. 644 + Traffic timeline: See Section 4.1.3 646 4.1.3. Typical test scenarios 648 o Single uplink RMCAT flow: N=1 with uplink direction and M=0. 650 o One pair of bi-directional RMCAT flows: N=2 (with one uplink flow 651 and one downlink flow); M=0. 653 o One pair of bi-directional RMCAT flows, one on-off CBR over UDP 654 flow on uplink: N=2 (with one uplink flow and one downlink flow); 655 M=1 (uplink). CBR flow ON time at 0s-60s, OFF time at 60s-119s. 657 o One pair of bi-directional RMCAT flows, one off-on CBR over UDP 658 flow on uplink: N=2 (with one uplink flow and one downlink flow); 659 M=1 (uplink). OFF time for UDP flow: 0s-60s; ON time: 60s-119s. 661 o One RMCAT flow competing against one long-live TCP flow over 662 uplink: N=1 (uplink) and M = 1(uplink), TCP start time at 0s and 663 end time at 119s. 665 4.1.4. Expected behavior 667 o Single uplink RMCAT flow: the candidate algorithm is expected to 668 detect the path capacity constraint, to converge to bottleneck 669 link capacity and to adapt the flow to avoid unwanted oscillation 670 when the sending bit rate is approaching the bottleneck link 671 capacity. No excessive oscillations in the media rate should be 672 present. 674 o Bi-directional RMCAT flows: It is expected that the candidate 675 algorithm is able to converge to the bottleneck capacity of the 676 wired path on both directions despite the presence of measurement 677 noise over the Wi-Fi connection. In the presence of background 678 TCP or CBR over UDP traffic, the rate of RMCAT flows should adapt 679 in a timely manner to changes in the available bottleneck 680 bandwidth. 682 o One RMCAT flow competing with long-live TCP flow over uplink: the 683 candidate algorithm should be able to avoid congestion collapse, 684 and to stabilize at a fair share of the bottleneck link capacity. 686 4.2. Bottleneck in Wi-Fi Network 688 These test cases assume that the wired portion along the media path 689 is well-provisioned whereas the bottleneck exists over the Wi-Fi 690 access network. This is to mimic the application scenarios typically 691 encountered by users in an enterprise environment or at a coffee 692 house. 694 4.2.1. Network topology 696 Same as defined in Section 4.1.1 698 4.2.2. Test setup 700 o Test duration: 120s 702 o Wi-Fi network characteristics: 704 * Radio propagation model: Log-distance path loss propagation 705 model [NS3WiFi] 707 * PHY- and MAC-layer configuration: IEEE 802.11n 708 * MCS Index at 11: 16-QAM 1/2, Raw Data Rate at 52Mbps 710 o Wired path characteristics: 712 * Path capacity: 100Mbps. 714 * One-Way propagation delay: 50ms. 716 * Maximum end-to-end jitter: 30ms. 718 * Bottleneck queue type: Drop tail. 720 * Bottleneck queue size: 300ms. 722 * Path loss ratio: 0%. 724 o Application characteristics: 726 * Media Traffic: 728 + Media type: Video 730 + Media direction: See Section 4.2.3. 732 + Number of media sources (N): See Section 4.2.3. 734 + Media timeline: 736 - Start time: 0s. 738 - End time: 119s. 740 * Competing traffic: 742 + Type of sources: long-lived TCP or CBR over UDP. 744 + Number of sources (M): See Section 4.2.3. 746 + Traffic direction: See Section 4.2.3. 748 + Congestion control: Default TCP congestion control [RFC5681] 749 or constant-bit-rate (CBR) traffic over UDP. 751 + Traffic timeline: See Section 4.2.3. 753 4.2.3. Typical test scenarios 755 This section describes a few test scenarios that are deemed as 756 important for understanding the behavior of a candidate RMCAT 757 solution over a Wi-Fi network. 759 a. Multiple RMCAT Flows Sharing the Wireless Downlink: N=16 (all 760 downlink); M = 0. This test case is for studying the impact of 761 contention on the multiple concurrent RMCAT flows. For an 762 802.11n network, given the MCS Index of 11 and the corresponding 763 raw data rate of 52Mbps, the total application-layer throughput 764 (assuming reasonable distance, low interference and infrequent 765 contentions caused by competing streams) is around 20Mbps. 766 Consequently, a total of N=16 RMCAT flows are needed to saturate 767 the wireless interface in this experiment. Evaluation of a given 768 candidate solution should focus on whether downlink RMCAT flows 769 can stabilize at a fair share of total application-layer 770 throughput. 772 b. Multiple RMCAT Flows Sharing the Wireless Uplink: N = 16 (all 773 downlink); M = 0. When multiple clients attempt to transmit 774 video packets uplink over the wireless interface, they introduce 775 more frequent contentions and potential collisions. Per-flow 776 throughput is expected to be lower than that in the previous 777 downlink-only scenario. Evaluation of a given candidate solution 778 should focus on whether uplink flows can stabilize at a fair 779 share of application-layer throughput. 781 c. Multiple Bi-directional RMCAT Flows: N = 16 (8 uplink and 8 782 downlink); M = 0. The goal of this test is to evaluate the 783 performance of the candidate solution in terms of bandwidth 784 fairness between uplink and downlink flows. 786 d. Multiple Bi-directional RMCAT Flows with on-off CBR traffic: N = 787 16 (8 uplink and 8 downlink); M = 5(uplink). The goal of this 788 test is to evaluate the adaptation behavior of the candidate 789 solution when its available bandwidth changes due to the 790 departure of background traffic. The background traffic consists 791 of several (e.g., M=5) CBR flows transported over UDP. These 792 background flows are ON at times t=0-60s and are OFF at times 793 t=61-120s. 795 e. Multiple Bi-directional RMCAT Flows with off-on CBR traffic: N = 796 16 (8 uplink and 8 downlink); M = 5(uplink). The goal of this 797 test is to evaluate the adaptation behavior of the candidate 798 solution when its available bandwidth changes due to the arrival 799 of background traffic. The background traffic consists of 800 several (e.g., M=5) parallel CBR flows transported over UDP. 802 These background flows are OFF at times t=0-60s and are ON at 803 times t=61-120s. 805 f. Multiple Bi-directional RMCAT flows in the presence of background 806 TCP traffic: N=16 (8 uplink and 8 downlink); M = 5 (uplink). The 807 goal of this test is to evaluate how RMCAT flows compete against 808 TCP over a congested Wi-Fi network for a given candidate 809 solution. TCP start time: 40s, end time: 80s. 811 g. Varying number of RMCAT flows: A series of tests can be carried 812 out for the above test cases with different values of N, e.g., N 813 = [4, 8, 12, 16, 20]. The goal of this test is to evaluate how a 814 candidate RMCAT solution responds to varying traffic load/demand 815 over a congested Wi-Fi network. The start time of these RMCAT 816 flows is randomly distributed within a window of t=0-10s, whereas 817 their end times are randomly distributed within a window of 818 t=110-120s. 820 4.2.4. Expected behavior 822 o Multiple downlink RMCAT flows: each RMCAT flow should get its fair 823 share of the total bottleneck link bandwidth. Overall bandwidth 824 usage should not be significantly lower than that experienced by 825 the same number of concurrent downlink TCP flows. In other words, 826 the performance of multiple concurrent TCP flows will be used as a 827 performance benchmark for this test scenario. The end-to-end 828 delay and packet loss ratio experienced by each flow should be 829 within an acceptable range for real-time multimedia applications. 831 o Multiple uplink RMCAT flows: overall bandwidth usage shared by all 832 RMCAT flows should not be significantly lower than that 833 experienced by the same number of concurrent uplink TCP flows. In 834 other words, the performance of multiple concurrent TCP flows will 835 be used as a performance benchmark for this test scenario. 837 o Multiple bi-directional RMCAT flows with dynamic background 838 traffic carrying CBR flows over UDP: RMCAT flows should adapt in a 839 timely fashion to the resulting changes in available bandwidth. 841 o Multiple bi-directional RMCAT flows with dynamic background 842 traffic over TCP: during the presence of TCP background flows, the 843 overall bandwidth usage shared by all RMCAT flows should not be 844 significantly lower than those achieved by the same number of bi- 845 directional TCP flows. In other words, the performance of 846 multiple concurrent TCP flows will be used as a performance 847 benchmark for this test scenario. All downlink RMCAT flows are 848 expected to obtain similar bandwidth with respect to each other. 849 The throughput of RMCAT flows should decrease upon the arrival of 850 TCP background traffic and increase upon their departure, both 851 reactions should occur in a timely fashion (e.g., within 10s of 852 seconds). 854 o Varying number of bi-directional RMCAT flows: the test results for 855 varying values of N -- while keeping all other parameters constant 856 -- is expected to show steady and stable per-flow throughput for 857 each value of N. The average throughput of all RMCAT flows is 858 expected to stay constant around the maximum rate when N is small, 859 then gradually decrease with increasing number of RMCAT flows till 860 it reaches the minimum allowed rate, beyond which the offered load 861 to the Wi-Fi network (with a large value of N) is exceeding its 862 capacity. 864 4.3. Other Potential Test Cases 866 4.3.1. EDCA/WMM usage 868 EDCA/WMM is prioritized QoS with four traffic classes (or Access 869 Categories) with differing priorities. RMCAT flows should achieve 870 better performance (i.e., lower delay, fewer packet losses) with 871 EDCA/WMM enabled when competing against non-interactive background 872 traffic (e.g., file transfers). When most of the traffic over Wi-Fi 873 is dominated by media, however, turning on WMM may actually degrade 874 performance since all media flows now attempt to access the wireless 875 transmission medium more aggressively, thereby causing more frequent 876 collisions and collision-induced losses. This is a topic worthy of 877 further investigation. 879 4.3.2. Effects of Legacy 802.11b Devices 881 When there exist 802.11b devices connected to a modern 802.11 882 network, they may affect the performance of the whole network. 883 Additional test cases can be added to evaluate the impacts of legacy 884 devices on the performance of the candidate congestion control 885 algorithm. 887 5. Conclusion 889 This document defines a collection of test cases that are considered 890 important for cellular and Wi-Fi networks. Moreover, this document 891 also provides a framework for defining additional test cases over 892 wireless cellular/Wi-Fi networks. 894 6. IANA Considerations 896 This memo includes no request to IANA. 898 7. Security Considerations 900 The security considerations in [I-D.ietf-rmcat-eval-criteria] and the 901 relevant congestion control algorithms apply. The principles for 902 congestion control are described in [RFC2914], and in particular, any 903 new method MUST implement safeguards to avoid congestion collapse of 904 the Internet. 906 The evaluations of the test cases are intended to carry out in a 907 controlled lab environment. Hence, the applications, simulators and 908 network nodes ought to be well-behaved and should not impact the 909 desired results. It is important to take appropriate caution to 910 avoid leaking non-responsive traffic from unproven congestion 911 avoidance techniques onto the open Internet. 913 8. Acknowledgments 915 The authors would like to thank Tomas Frankkila, Magnus Westerlund, 916 Kristofer Sandlund, and Sergio Mena de la Cruz for their valuable 917 input and review comments regarding this draft. 919 9. References 921 9.1. Normative References 923 [Deployment] 924 TS 25.814, 3GPP., "Physical layer aspects for evolved 925 Universal Terrestrial Radio Access (UTRA)", October 2006, 926 . 929 [HO-def-3GPP] 930 TR 21.905, 3GPP., "Vocabulary for 3GPP Specifications", 931 December 2009, . 934 [HO-LTE-3GPP] 935 TS 36.331, 3GPP., "E-UTRA- Radio Resource Control (RRC); 936 Protocol specification", December 2011, 937 . 940 [HO-UMTS-3GPP] 941 TS 25.331, 3GPP., "Radio Resource Control (RRC); Protocol 942 specification", December 2011, 943 . 946 [I-D.ietf-rmcat-eval-criteria] 947 Singh, V., Ott, J., and S. Holmer, "Evaluating Congestion 948 Control for Interactive Real-time Media", draft-ietf- 949 rmcat-eval-criteria-08 (work in progress), November 2018. 951 [NS3WiFi] "Wi-Fi Channel Model in NS3 Simulator", 952 . 955 [QoS-3GPP] 956 TS 23.203, 3GPP., "Policy and charging control 957 architecture", June 2011, . 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 [RFC2914] Floyd, S., "Congestion Control Principles", BCP 41, 966 RFC 2914, DOI 10.17487/RFC2914, September 2000, 967 . 969 [RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion 970 Control", RFC 5681, DOI 10.17487/RFC5681, September 2009, 971 . 973 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 974 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 975 May 2017, . 977 9.2. Informative References 979 [Heusse2003] 980 Heusse, M., Rousseau, F., Berger-Sabbatel, G., and A. 981 Duda, "Performance anomaly of 802.11b", in Proc. 23th 982 Annual Joint Conference of the IEEE Computer and 983 Communications Societies, (INFOCOM'03), March 2003. 985 [I-D.ietf-rmcat-cc-requirements] 986 Jesup, R. and Z. Sarker, "Congestion Control Requirements 987 for Interactive Real-Time Media", draft-ietf-rmcat-cc- 988 requirements-09 (work in progress), December 2014. 990 [I-D.ietf-rmcat-eval-test] 991 Sarker, Z., Singh, V., Zhu, X., and M. Ramalho, "Test 992 Cases for Evaluating RMCAT Proposals", draft-ietf-rmcat- 993 eval-test-10 (work in progress), May 2019. 995 [IEEE802.11] 996 IEEE, "Standard for Information technology-- 997 Telecommunications and information exchange between 998 systems Local and metropolitan area networks--Specific 999 requirements Part 11: Wireless LAN Medium Access Control 1000 (MAC) and Physical Layer (PHY) Specifications", 2012. 1002 [LTE-simulator] 1003 "NS-3, A discrete-Event Network Simulator", 1004 . 1007 [ns-2] "The Network Simulator - ns-2", 1008 . 1010 [ns-3] "The Network Simulator - ns-3", . 1012 Authors' Addresses 1014 Zaheduzzaman Sarker 1015 Ericsson AB 1016 Laboratoriegraend 11 1017 Luleae 97753 1018 Sweden 1020 Phone: +46 107173743 1021 Email: zaheduzzaman.sarker@ericsson.com 1023 Ingemar Johansson 1024 Ericsson AB 1025 Laboratoriegraend 11 1026 Luleae 97753 1027 Sweden 1029 Phone: +46 10 7143042 1030 Email: ingemar.s.johansson@ericsson.com 1031 Xiaoqing Zhu 1032 Cisco Systems 1033 12515 Research Blvd., Building 4 1034 Austin, TX 78759 1035 USA 1037 Email: xiaoqzhu@cisco.com 1039 Jiantao Fu 1040 Cisco Systems 1041 707 Tasman Drive 1042 Milpitas, CA 95035 1043 USA 1045 Email: jianfu@cisco.com 1047 Wei-Tian Tan 1048 Cisco Systems 1049 725 Alder Drive 1050 Milpitas, CA 95035 1051 USA 1053 Email: dtan2@cisco.com 1055 Michael A. Ramalho 1056 Cisco Systems, Inc. 1057 8000 Hawkins Road 1058 Sarasota, FL 34241 1059 USA 1061 Phone: +1 919 476 2038 1062 Email: mramalho@cisco.com