idnits 2.17.1 draft-perkins-intarea-multicast-ieee802-02.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 13, 2017) is 2593 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '2' on line 338 == Missing Reference: 'ICMPv6' is mentioned on line 342, but not defined Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Area WG C. Perkins 3 Internet-Draft Futurewei 4 Intended status: Informational D. Stanley 5 Expires: September 14, 2017 HPE 6 W. Kumari 7 Google 8 JC. Zuniga 9 SIGFOX 10 March 13, 2017 12 Multicast Considerations over IEEE 802 Wireless Media 13 draft-perkins-intarea-multicast-ieee802-02 15 Abstract 17 Some performance issues have been observed when multicast packet 18 transmissions of IETF protocols are used over IEEE 802 wireless 19 media. Even though enhamcements for multicast transmissions have 20 been designed at both IETF and IEEE 802, there seems to exist a 21 disconnect between specifications, implementations and configuration 22 choices. 24 This draft describes the different issues that have been observed, 25 the multicast enhancement features that have been specified at IETF 26 and IEEE 802 for wireless media, as well as the operational chioces 27 that can be taken to improve the performace of the network. Finally, 28 it provides some recommendations about the usage and combination of 29 these features and operational choices. 31 Status of This Memo 33 This Internet-Draft is submitted in full conformance with the 34 provisions of BCP 78 and BCP 79. 36 Internet-Drafts are working documents of the Internet Engineering 37 Task Force (IETF). Note that other groups may also distribute 38 working documents as Internet-Drafts. The list of current Internet- 39 Drafts is at http://datatracker.ietf.org/drafts/current/. 41 Internet-Drafts are draft documents valid for a maximum of six months 42 and may be updated, replaced, or obsoleted by other documents at any 43 time. It is inappropriate to use Internet-Drafts as reference 44 material or to cite them other than as "work in progress." 46 This Internet-Draft will expire on September 14, 2017. 48 Copyright Notice 50 Copyright (c) 2017 IETF Trust and the persons identified as the 51 document authors. All rights reserved. 53 This document is subject to BCP 78 and the IETF Trust's Legal 54 Provisions Relating to IETF Documents 55 (http://trustee.ietf.org/license-info) in effect on the date of 56 publication of this document. Please review these documents 57 carefully, as they describe your rights and restrictions with respect 58 to this document. Code Components extracted from this document must 59 include Simplified BSD License text as described in Section 4.e of 60 the Trust Legal Provisions and are provided without warranty as 61 described in the Simplified BSD License. 63 Table of Contents 65 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 66 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 67 3. Identified mulitcast issues . . . . . . . . . . . . . . . . . 4 68 3.1. Issues at Layer 2 and below . . . . . . . . . . . . . . . 4 69 3.1.1. Multicast reliability . . . . . . . . . . . . . . . . 4 70 3.1.2. Lower data rate . . . . . . . . . . . . . . . . . . . 4 71 3.1.3. Power-save effects on multicast . . . . . . . . . . . 5 72 3.2. Issues at Layer 3 and above . . . . . . . . . . . . . . . 5 73 3.2.1. IPv4 issues . . . . . . . . . . . . . . . . . . . . . 5 74 3.2.2. IPv6 issues . . . . . . . . . . . . . . . . . . . . . 5 75 3.2.3. MLD issues . . . . . . . . . . . . . . . . . . . . . 6 76 3.2.4. Spurious Neighbor Discovery . . . . . . . . . . . . . 6 77 4. Multicast protocol optimizations . . . . . . . . . . . . . . 7 78 4.1. Proxy ARP in 802.11-2012 . . . . . . . . . . . . . . . . 7 79 4.2. Buffering to improve Power-Save . . . . . . . . . . . . . 8 80 4.3. IPv6 support in 802.11-2012 . . . . . . . . . . . . . . . 8 81 4.4. Conversion of multicast to unicast . . . . . . . . . . . 8 82 4.5. Directed Multicast Service (DMS) . . . . . . . . . . . . 8 83 4.6. GroupCast with Retries (GCR) . . . . . . . . . . . . . . 9 84 5. Operational optimizations . . . . . . . . . . . . . . . . . . 10 85 5.1. Mitigating Problems from Spurious Neighbor Discovery . . 10 86 6. Multicast Considerations for Other Wireless Media . . . . . . 12 87 7. Recommendations . . . . . . . . . . . . . . . . . . . . . . . 12 88 8. Security Considerations . . . . . . . . . . . . . . . . . . . 12 89 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 90 10. Informative References . . . . . . . . . . . . . . . . . . . 12 91 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 13 93 1. Introduction 95 Many IETF protocols depend on multicast/broadcast for delivery of 96 control messages to multiple receivers. Multicast is used for 97 various purposes such as neighborhood discovery, network flooding, 98 address resolution, as well as for reduction in media access for the 99 transmission of data that is intended for multiple receivers. 101 IETF protocols typically rely on network protocol layering in order 102 to reduce or eliminate any dependence of higher level protocols on 103 the specific nature of the MAC layer protocols or the physical media. 104 In the case of multicast transmissions, higher level protocols have 105 traditionally been designed as if transmitting a packet to an IP 106 address had the same cost in interference and network media access, 107 regardless of whether the destination IP address is a unicast address 108 or a multicast or broadcast address. This model was reasonable for 109 networks where the physical medium was usually wired, like Ethernet. 110 Unfortunately, for many wireless media, the costs to access the 111 medium can be quite different. Some enhancements have been designed 112 in IETF protocols that are assumed to work primarily over wilress 113 media. However, these enhancements are usually implemented in 114 limited deployments and not widely spread on most wireless networks. 116 IEEE 802 wireless protocols have been designed with certain fetures 117 to support multicat traffic. For instance, lower modulations are 118 used to transmit multicast frames, so that these can be received by 119 all stations in the cell, regardless of the distance or path 120 attenuation from the base station or access point. However, these 121 lower modulation transmissions take longer on the medium and 122 therefore they reduce the capabilities to transmit more high 123 efficiency traffic with higher order modulations to stations that may 124 be in closer vicinity. Due to these and other reasons, some IEEE 802 125 working groups like 802.11 have designed several features to improve 126 the performance of multicast transmissions at Layer 2 [REF 127 11-15-1261-03]. Besides protocol design features, some operational 128 and configuration enhancements can also be applied to overcome the 129 network performance issues created by multicast traffic. 131 This Internet Draft identifies the problems created by the usage of 132 multicast traffic over wireless networks. It also highlights the 133 different enhancements that have been designed at IETF and IEEE 802, 134 as well as the operational choices that can be taken, to ameliorate 135 the effects of multicast traffic. Some recommendations about the 136 usage and combinations of these enhancements are also provided. 138 2. Terminology 140 This document uses the following definitions: 142 AP 143 IEEE 802.11 Access Point. 145 STA 146 IEEE 802.11 station. 148 basic rate 149 The "lowest common denominator" data rate at which multicast and 150 broadcast traffic is generally transmitted. 152 MCS 153 Modulation and Coding Scheme. 155 3. Identified mulitcast issues 157 3.1. Issues at Layer 2 and below 159 In this section we list some of the issues related to the use of 160 multicast transmissions over IEEE 802 wireless technologies. 162 3.1.1. Multicast reliability 164 Multicast traffic is typically much less reliable than unicast 165 traffic. Since multicast makes point-to-multipoint communications, 166 multiple acknowledgements would be needed to guarantee the reception 167 on all recepients. 169 3.1.2. Lower data rate 171 Because lower MCS have longer range but also lower data rate, 172 multicast / broadcast traffic is generally transmitted at the lowest 173 common denominator rate, also known as a basic rate. On IEEE 802.11 174 networks (aka Wi-Fi), this rate might be as low as 6 Mbps, when some 175 unicast links in the same cell can be operating at rates up to 600 176 Mbps. Transmissions at a lower rate require more occupancy of the 177 wireless medium and thus restrict the airtime for all other medium 178 communications and degrade the overall capacity. 180 Wired multicast affects wireless LANs because the AP extends the 181 wired segment and multicast / broadcast frames on the wired LAN side 182 are copied to WLAN. Since broadcast messages are transmitted at the 183 most robust MCS, this implies that large frames sent at slow rate 184 over the air. 186 3.1.3. Power-save effects on multicast 188 Multicast can work poorly with the power-save mechanisms defined in 189 IEEE 802.11. 191 o Both unicast and multicast traffic can be delayed by power-saving 192 mechanisms. 193 o Unicast is delayed until a STA wakes up and asks for it. 194 Additionally, unicast traffic may be delayed to improve power 195 save, efficiency and increase probability of aggregation. 196 o Multicast traffic is delayed in a wireless network if any of the 197 STAs in that network are power savers. All STAs have to be awake 198 at a known time to receive multicast traffic. 199 o Packets can also be discarded due to buffer limitations in the AP 200 and non-AP STA. 202 3.2. Issues at Layer 3 and above 204 In this section we mention a few representative IETF protocols, and 205 describe some possible negative effects due to performance 206 degradation when using multicast transmissions for control messages. 207 Common uses of multicast include: 209 o Control plane for IPv4 and IPv6 210 o ARP and Neighbor Discovery 211 o Service discovery 212 o Applications (video delivery, stock data etc) 213 o Other L3 protocols (non-IP) 215 3.2.1. IPv4 issues 217 The following list contains a few representative IPv4 protocols using 218 multicast. 220 o ARP 221 o DHCP 222 o mDNS 224 After initial configuration, ARP and DHCP occur much less commonly. 226 3.2.2. IPv6 issues 228 The following list contains a few representative IPv6 protocols using 229 multicast. IPv6 makes much more extensive use of multicast. 231 o DHCPv6 232 o Liveness detection (NUD) 233 o Some control plane protocols are not very tolerant of packet loss, 234 especially neighbor discovery. 235 o Services may be considered lost if several consecutive packets 236 fail. 238 Address Resolution 240 Service Discovery 242 Route Discovery 244 Decentralized Address Assignment 246 Geographic routing 248 3.2.3. MLD issues 250 Multicast Listener Discovery(MLD) [RFC4541] is often used to identify 251 members of a multicast group that are connected to the ports of a 252 switch. Forwarding multicast frames into a WiFi-enabled area can use 253 such switch support for hardware forwarding state information. 254 However, since IPv6 makes heavy use of multicast, each STA with an 255 IPv6 address will require state on the switch for several and 256 possibly many multicast solicited-node addresses. Multicast 257 addresses that do not have forwarding state installed (perhaps due to 258 hardware memory limitations on the switch) cause frames to be flooded 259 on all ports of the switch. 261 3.2.4. Spurious Neighbor Discovery 263 On the Internet there is a "background radiation" of scanning traffic 264 (people scanning for vulnerable machines) and backscatter (responses 265 from spoofed traffic, etc). This means that the router is constantly 266 getting packets destined for machines whose IP addresses may or may 267 not be in use. In the cases where the IP is assigned to a machine, 268 the router broadcasts an ARP request, gets back an ARP reply, caches 269 this and then can deliver traffic to the host. In the cases where 270 the IP address is not in use, the router broadcasts one (or more) ARP 271 requests, and never gets a reply. This means that it does not 272 populate the ARP cache, and the next time there is traffic for that 273 IP address it will broadcast ARP requests again. 275 The rate of these ARP requests is proportional to the size of the 276 subnets, the rate of scanning and backscatter, and how long the 277 router keeps state on non-responding ARPs. As it turns out, this 278 rate is inversely proportional to how occupied the subnet is (valid 279 ARPs end up in a cache, stopping the broadcasting; unused IPs never 280 respond, and so cause more broadcasts). Depending on the address 281 space in use, the time of day, how occupied the subnet is, and other 282 unknown factors, on the order of 2000 broadcasts per second have been 283 observed at the IETF NOCs. 285 On a wired network, there is not a huge difference amongst unicast, 286 multicast and broadcast traffic; but this is not true in the wireless 287 realm. Wireless equipment often is unable to send this amount of 288 broadcast and multicast traffic. Consequently, on the wireless 289 networks, we observe a significant amount of dropped broadcast and 290 multicast packets. This, in turn, means that when a host connects it 291 is often not able to complete DHCP, and IPv6 RAs get dropped, leading 292 to users being unable to use the network. 294 4. Multicast protocol optimizations 296 This section lists some optimizations that have been specified in 297 IEEE 802 and IETF that are aimed at reducing or eliminating the 298 issues discussed in Section 3. 300 4.1. Proxy ARP in 802.11-2012 302 The AP knows all associated STAs MAC address and IP address; in other 303 words, the AP acts as the central "manager" for all the 802.11 STAs 304 in its BSS. Proxy ARP is easy to implement at the AP, and offers the 305 following advantages: 307 o Reduced broadcast traffic (transmitted at low MCS) on the wireless 308 medium 309 o STA benefits from extended power save in sleep mode, as ARP 310 requests are replied to by AP. 311 o Keeps ARP frames off the wireless medium. 312 o Changes are not needed to STA implementation. 314 Here is the specification language from clause 10.23.13 in [2] as 315 described in [dot11-proxyarp]: 317 When the AP supports Proxy ARP "[...] the AP shall maintain a 318 Hardware Address to Internet Address mapping for each associated 319 station, and shall update the mapping when the Internet Address of 320 the associated station changes. When the IPv4 address being 321 resolved in the ARP request packet is used by a non-AP STA 322 currently associated to the BSS, the proxy ARP service shall 323 respond on behalf of the non-AP STA" 325 4.2. Buffering to improve Power-Save 327 The AP acts on behalf of STAs in various ways. In order to improve 328 the power-saving feature for STAs in its BSS, the AP buffers frames 329 for delivery to the STA at the time when the STA is scheduled for 330 reception. 332 4.3. IPv6 support in 802.11-2012 334 IPv6 uses Neighbor Discovery Protocol (NDP) instead Every IPv6 node 335 subscribes to special multicast address Neighbor-Solicitation message 336 replaces ARP 338 Here is the specification language from-10.23.13 in [2]: 340 "When an IPv6 address is being resolved, the Proxy Neighbor 341 Discovery service shall respond with a Neighbor Advertisement 342 message [...] on behalf of an associated STA to an [ICMPv6] 343 Neighbor Solicitation message [...]. When MAC address mappings 344 change, the AP may send unsolicited Neighbor Advertisement 345 Messages on behalf of a STA." 347 NDP may be used to request additional information 349 o Maximum Transmission Unit 350 o Router Solicitation 351 o Router Advertisement, etc. 353 NDP messages are sent as group addressed (broadcast) frames in 354 802.11. Using the proxy operation helps to keep NDP messages off the 355 wireless medium. 357 4.4. Conversion of multicast to unicast 359 It is often possible to transmit control and data messages by using 360 unicast transmissions to each station individually. 362 4.5. Directed Multicast Service (DMS) 364 There are situations where more is needed than simply converting 365 multicast to unicast [Editor's note: citation needed]. For these 366 purposes, DMS enables a client to request that the AP transmit 367 multicast group addressed frames destined to the requesting clients 368 as individually addressed frames [i.e., convert multicast to 369 unicast]. 371 o DMS Requires 802.11n A-MSDUs 372 o Individually addressed frames are acknowledged and are buffered 373 for power save clients 374 o Requesting STA may specify traffic characteristics for DMS traffic 375 o DMS was defined in IEEE Std 802.11v-2011 377 DMS is not currently implemented in products. DMS does require 378 changes to both AP and STA implementation. 380 4.6. GroupCast with Retries (GCR) 382 GCR (defined in [dot11aa]) provides greater reliability by using 383 either unsolicited retries or a block acknowledgement mechanism. GCR 384 increases probability of broadcast frame reception success, but still 385 does not guarantee success. 387 For the block acknowledgement mechanism, the AP transmits each group 388 addressed frame as conventional group addressed transmission. 389 Retransmissions are group addressed, but hidden from non-11aa 390 clients. A directed block acknowledgement scheme is used to harvest 391 reception status from receivers; retransmissions are based upon these 392 responses. 394 GCR is suitable for all group sizes including medium to large groups. 395 As the number of devices in the group increases, GCR can send block 396 acknowledgement requests to only a small subset of the group. GCR 397 does require changes to both AP and STA implementation. 399 GCR may introduce unacceptable latency. After sending a group of 400 data frames to the group, the AP has do the following: 402 o unicast a Block Ack Request (BAR) to a subset of members. 403 o wait for the corresponding Block Ack (BA). 404 o retransmit any missed frames. 405 o resume other operations which may have been delayed. 407 This latency may not be acceptable for some traffic. 409 There are ongoing extensions in 802.11 to improve GCR performance. 411 o BAR is sent using downlink MU-MIMO (note that downlink MU-MIMO is 412 already specified in 802.11-REVmc 4.3). 413 o BA is sent using uplink MU-MIMO (which is a .11ax feature). 414 o Additional 802.11ax extensions are under consideration; see 415 [mc-ack-mux] 416 o Latency may also be reduced by simultaneously receiving BA 417 information from multiple clients. 419 5. Operational optimizations 421 This section lists some operational optimizations that can be 422 implemented when deploying wireless IEEE 802 networks to mitigate the 423 issues discussed in Section 3. 425 5.1. Mitigating Problems from Spurious Neighbor Discovery 427 ARP Sponges 429 An ARP Sponge sits on a network and learn which IPs addresses 430 are actually in use. It also listen for ARP requests, and, if 431 it sees an ARP for an IP address which it believes is not used, 432 it will reply with its own MAC address. This means that the 433 router now has an IP to MAC mapping, which it caches. If that 434 IP is later assigned to an machine (e.g using DHCP), the ARP 435 sponge will see this, and will stop replying for that address. 436 Gratuitous ARPs (or the machine ARPing for its gateway) will 437 replace the sponged address in the router ARP table. This 438 technique is quite effective; but, unfortunately, the ARP 439 sponge daemons were not really designed for this use (the 440 standard one [arpsponge], was designed to deal with the 441 disappearance of participants from an IXP) and so are not 442 optimized for this purpose. We have to run one daemon per 443 subnet, the tuning is tricky (the scanning rate versus the 444 population rate versus retires, etc.) and sometimes the daemons 445 just seem to stop, requiring a restart of the daemon and 446 causing disruption. 448 Router mitigations 450 Some routers (often those based on Linux) implement a "negative 451 ARP cache" daemon. Simply put, if the router does not see a 452 reply to an ARP it can be configured to cache this information 453 for some interval. Unfortunately, the core routers which we 454 are using do not support this. When a host connects to network 455 and gets an IP address, it will ARP for its default gateway 456 (the router). The router will update its cache with the IP to 457 host MAC mapping learnt from the request (passive ARP 458 learning). 460 Firewall unused space 462 The distribution of users on wireless networks / subnets 463 changes from meeting to meeting (e.g the "IETF-secure" SSID was 464 renamed to "IETF", fewer users use "IETF-legacy", etc). This 465 utilization is difficult to predict ahead of time, but we can 466 monitor the usage as attendees use the different networks. By 467 configuring multiple DHCP pools per subnet, and enabling them 468 sequentially, we can have a large subnet, but only assign 469 addresses from the lower portions of it. This means that we 470 can apply input IP access lists, which deny traffic to the 471 upper, unused portions. This means that the router does not 472 attempt to forward packets to the unused portions of the 473 subnets, and so does not ARP for it. This method has proven to 474 be very effective, but is somewhat of a blunt axe, is fairly 475 labor intensive, and requires coordination. 477 Disabling/filtering ARP requests 479 In general, the router does not need to ARP for hosts; when a 480 host connects, the router can learn the IP to MAC mapping from 481 the ARP request sent by that host. This means that we should 482 be able to disable and / or filter ARP requests from the 483 router. Unfortunately, ARP is a very low level / fundamental 484 part of the IP stack, and is often offloaded from the normal 485 control plane. While many routers can filter layer-2 traffic, 486 this is usually implemented as an input filter and / or has 487 limited ability to filter output broadcast traffic. This means 488 that the simple "just disable ARP or filter it outbound" seems 489 like a really simple (and obvious) solution, but 490 implementations / architectural issues make this difficult or 491 awkward in practice. 493 NAT 495 The broadcasts are overwhelmingly being caused by outside 496 scanning / backscatter traffic. This means that, if we were to 497 NAT the entire (or a large portion) of the attendee networks, 498 there would be no NAT translation entries for unused addresses, 499 and so the router would never ARP for them. The IETF NOC has 500 discussed NATing the entire (or large portions) attendee 501 address space, but a: elegance and b: flaming torches and 502 pitchfork concerns means we have not attempted this yet. 504 Stateful firewalls 506 Another obvious solution would be to put a stateful firewall 507 between the wireless network and the Internet. This firewall 508 would block incoming traffic not associated with an outbound 509 request. The IETF philosophy has been to have the network as 510 open as possible / honor the end-to-end principle. An attendee 511 on the meeting network should be an Internet host, and should 512 be able to receive unsolicited requests. Unfortunately, 513 keeping the network working and stable is the first priority 514 and a stateful firewall may be required in order to achieve 515 this. 517 6. Multicast Considerations for Other Wireless Media 519 Many of the causes of performance degradation described in earlier 520 sections are also observable for wireless media other than 802.11. 522 For instance, problems with power save, excess media occupancy, and 523 poor reliability will also affect 802.15.3 and 802.15.4. However, 524 802.15 media specifications do not include similar mechanisms of the 525 type that have been developed for 802.11. In fact, the design 526 philosophy for 802.15 is more oriented towards minimality, with the 527 result that many such functions would more likely be relegated to 528 operation within higher layer protocols. This leads to a patchwork 529 of non-interoperable and vendor-specific solutions. See [uli] for 530 some additional discussion, and a proposal for a task group to 531 resolve similar issues, in which the multicast problems might be 532 considered for mitigation. 534 7. Recommendations 536 This section provides some recommendations about the usage and 537 combinations of the multicast enhancements described in Section 4 and 538 Section 5. 540 (FFS) 542 8. Security Considerations 544 This document does not introduce any security mechanisms, and does 545 not have any impact on existing security mechanisms. 547 9. IANA Considerations 549 This document does not specify any IANA actions. 551 10. Informative References 553 [arpsponge] 554 Arien Vijn, Steven Bakker, , "Arp Sponge", March 2015. 556 [dot11] P802.11, , "Part 11: Wireless LAN Medium Access Control 557 (MAC) and Physical Layer (PHY) Specifications", March 558 2012. 560 [dot11-proxyarp] 561 P802.11, , "Proxy ARP in 802.11ax", September 2015. 563 [dot11aa] P802.11, , "Part 11: Wireless LAN Medium Access Control 564 (MAC) and Physical Layer (PHY) Specifications Amendment 2: 565 MAC Enhancements for Robust Audio Video Streaming", March 566 2012. 568 [mc-ack-mux] 569 Yusuke Tanaka et al., , "Multiplexing of Acknowledgements 570 for Multicast Transmission", July 2015. 572 [mc-prob-stmt] 573 Mikael Abrahamsson and Adrian Stephens, , "Multicast on 574 802.11", March 2015. 576 [mc-props] 577 Adrian Stephens, , "IEEE 802.11 multicast properties", 578 March 2015. 580 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 581 "Considerations for Internet Group Management Protocol 582 (IGMP) and Multicast Listener Discovery (MLD) Snooping 583 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 584 . 586 [uli] Pat Kinney, , "LLC Proposal for 802.15.4", Nov 2015. 588 Authors' Addresses 590 Charles E. Perkins 591 Futurewei Inc. 592 2330 Central Expressway 593 Santa Clara, CA 95050 594 USA 596 Phone: +1-408-330-4586 597 Email: charliep@computer.org 599 Dorothy Stanley 600 Hewlett Packard Enterprise 601 2000 North Naperville Rd. 602 Naperville, IL 60566 603 USA 605 Phone: +1 630 979 1572 606 Email: dstanley@arubanetworks.com 607 Warren Kumari 608 Google 609 1600 Amphitheatre Parkway 610 Mountain View, CA 94043 611 USA 613 Email: warren@kumari.net 615 Juan Carlos Zuniga 616 SIGFOX 617 425 rue Jean Rostand 618 Labege 31670 619 France 621 Email: j.c.zuniga@ieee.org