idnits 2.17.1 draft-ietf-v6ops-nd-cache-init-04.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 (August 20, 2020) is 1339 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-07) exists of draft-ietf-6man-grand-01 -- Obsolete informational reference (is this intentional?): RFC 4941 (Obsoleted by RFC 8981) Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 v6ops J. Linkova 3 Internet-Draft Google 4 Intended status: Informational August 20, 2020 5 Expires: February 21, 2021 7 Neighbor Cache Entries on First-Hop Routers: Operational Considerations 8 draft-ietf-v6ops-nd-cache-init-04 10 Abstract 12 Neighbor Discovery (RFC4861) is used by IPv6 nodes to determine the 13 link-layer addresses of neighboring nodes as well as to discover and 14 maintain reachability information. This document discusses how the 15 neighbor discovery state machine on a first-hop router is causing 16 user-visible connectivity issues when a new (not being seen on the 17 network before) IPv6 address is being used. 19 Status of This Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at https://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on February 21, 2021. 36 Copyright Notice 38 Copyright (c) 2020 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (https://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 54 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 4 55 1.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 56 2. Proposed Solution . . . . . . . . . . . . . . . . . . . . . . 5 57 2.1. Solution Requirements . . . . . . . . . . . . . . . . . . 5 58 2.2. Solution Overview . . . . . . . . . . . . . . . . . . . . 5 59 3. Solutions Considered but Discarded . . . . . . . . . . . . . 6 60 3.1. Do Nothing . . . . . . . . . . . . . . . . . . . . . . . 7 61 3.2. Change to the Registration-Based Neighbor Discovery . . . 7 62 3.3. Host Sending NS to the Router Address from Its GUA . . . 7 63 3.4. Host Sending Router Solicitation from its GUA . . . . . . 8 64 3.5. Routers Populating Their Caches by Gleaning From Neighbor 65 Discovery Packets . . . . . . . . . . . . . . . . . . . . 9 66 3.6. Initiating Hosts-to-Routers Communication . . . . . . . . 9 67 3.7. Transit Dataplane Traffic From a New Address Triggering 68 Address Resolution . . . . . . . . . . . . . . . . . . . 10 69 4. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 10 70 5. Security Considerations . . . . . . . . . . . . . . . . . . . 10 71 6. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 10 72 7. References . . . . . . . . . . . . . . . . . . . . . . . . . 11 73 7.1. Normative References . . . . . . . . . . . . . . . . . . 11 74 7.2. Informative References . . . . . . . . . . . . . . . . . 12 75 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 12 77 1. Introduction 79 The section 7.2.5 of [RFC4861] states: "When a valid Neighbor 80 Advertisement is received (either solicited or unsolicited), the 81 Neighbor Cache is searched for the target's entry. If no entry 82 exists, the advertisement SHOULD be silently discarded. There is no 83 need to create an entry if none exists, since the recipient has 84 apparently not initiated any communication with the target." 86 This approach is perfectly suitable for host-to-host communications, 87 which are in most cases bi-directional, and it could be expected that 88 if a host A has an ND cache entry for the host B IPv6 address, host B 89 also has the corresponding ND entry for the host A address in its 90 cache. However when a host communicates to off-link destinations via 91 its first-hop router, that logic does not apply. The most typical 92 scenario when the problem may arise is a host joining the network, 93 forming a new address and using that address for accessing the 94 Internet: 96 1. A host joins the network and receives a Router Advertisement (RA) 97 packet from the first-hop router (either a periodic unsolicited 98 RA or a response to a Router Solicitation sent by the host). The 99 RA contains information the host needs to perform SLAAC and to 100 configure its network stack. As in most cases the RA also 101 contains the Source link-layer address of the router, the host 102 can populate its Neighbor Cache with the router's link-local and 103 link-layer addresses. 105 2. The host starts opening connections to off-link destinations. A 106 very common use case is a mobile device sending probes to detect 107 the Internet connectivity and/or the presence of a captive portal 108 on the network. To speed up that process many implementations 109 use Optimistic Duplicate Address Detection [RFC4429] which allows 110 them to send probes from their GUA before the DAD process is 111 completed. At that moment the device ND cache contains all 112 information required to send those probes (such as the default 113 router link-local the link-layer addresses). The router ND 114 cache, however, might contain an entry for the device link-local 115 address (if the device has been performing the address resolution 116 for the router LLA), but there are no entries for the device GUA. 118 3. Return traffic is received by the first-hop router. As the 119 router does not have any ND cache entry for the host GUA yet, the 120 router starts the neighbor discovery process by creating an 121 INCOMPLETE cache entry and then sending an NS to the Solicited 122 Node Multicast Address. Most router implementations buffer only 123 one data packet while resolving the packet destination address, 124 so it would drop all subsequent packets for the host GUA, until 125 the address resolution process is completed. 127 4. If the host sends multiple probes in parallel, it would consider 128 all but one of them failed. It leads to user-visible delay in 129 connecting to the network, especially if the host implements some 130 form of backoff mechanism and does not retransmit the probes as 131 soon as possible. 133 This scenario illustrates the problem occurring when the device 134 connects to the network for the first time or after a timeout long 135 enough for the device address to be removed from the router's 136 neighbor cache. However, the same sequence of events happen when the 137 host starts using a new GUA previously unseen by the router, such as 138 a new privacy address [RFC4941] or if the router's Neighbor Cache has 139 been flushed. 141 While in dual-stack networks this problem might be hidden by Happy 142 Eyeballs [RFC8305] it manifests quite clearly in IPv6-only 143 environments, especially wireless ones, leading to poor user 144 experience and contributing to a negative perception of IPv6-only 145 solutions as unstable and non-deployable. 147 This document discusses the operational implications of not 148 proactively creating Neighbor Cache entries on first-hop routers and 149 summarizes various approaches to mitigate the problem. 151 1.1. Requirements Language 153 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 154 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 155 "OPTIONAL" in this document are to be interpreted as described in BCP 156 14 [RFC2119] [RFC8174] when, and only when, they appear in all 157 capitals, as shown here. 159 1.2. Terminology 161 ND: Neighbor Discovery, [RFC4861]. 163 SLAAC: IPv6 Stateless Address Autoconfiguration, [RFC4862]. 165 NS: Neighbor Solicitation, [RFC4861]. 167 NA: Neighbor Advertisement, [RFC4861]. 169 RS: Router Solicitation, [RFC4861]. 171 RA: Router Advertisement, [RFC4861]. 173 SLLA: Source link-layer Address, an option in the ND packets 174 containing the link-layer address of the sender of the packet, 175 [RFC4861]. 177 TLLA: Target link-layer Address, an option in the ND packets 178 containing the link-layer address of the target, [RFC4861]. 180 GUA: Global Unicast Address, [RFC4291]. 182 DAD: Duplicate Address Detection, [RFC4862]. 184 Optimistic DAD: a modification of DAD, [RFC4429]. 186 FCFS SAVI: First-Come, First-Served Source Address Validation, 187 [RFC6620]. 189 2. Proposed Solution 191 2.1. Solution Requirements 193 It would be highly desirable to improve the Neighbor Discovery 194 mechanics so routers have a usable cache entry for a host address by 195 the time the router receives the first packet for that address. In 196 particular: 198 o If the router does not have a Neighbor Cache entry for the 199 address, a STALE entry needs to be created. 201 o The solution needs to work for Optimistic addresses as well. 202 Devices implementing the Optimistic DAD usually attempt to 203 minimize the delay in connecting to the network and therefore are 204 more likely to be affected by the problem described in this 205 document. 207 o In case of duplicate addresses present in the network, the 208 proposed solution MUST NOT override the existing entry. 210 o In topologies with multiple first-hop routers the cache needs to 211 be updated on all of them, as traffic might be asymmetric: 212 outgoing flows leaving the network via one router while the return 213 traffic enters the segment via another one. 215 In addition the solution MUST NOT exacerbate issues described in 216 [RFC6583] and MUST be compatible with the recommendations provided in 217 [RFC6583]. 219 2.2. Solution Overview 221 The Neighbor Discovery is designed to allow IPv6 nodes to discover 222 neighboring nodes reachability and learn IPv6 to link-layer addresses 223 mapping. Therefore ND seems to be the most appropriate tool to 224 inform the first-hop routers about addresses the host is going to 225 use. 227 Section 4.4 of [RFC4861] says: 229 "A node sends Neighbor Advertisements in response to Neighbor 230 Solicitations and sends unsolicited Neighbor Advertisements in order 231 to (unreliably) propagate new information quickly." 233 Propagating information about new GUA as quickly as possible is 234 exactly what is required to solve the problem outlined in this 235 document. Therefore the host might send an unsolicited NA with the 236 target link-layer address option to advertise its GUA as soon as the 237 said address enters Optimistic or Preferred state. 239 The proposed solution is discussed in [I-D.ietf-6man-grand]. In 240 summary, the following changes to [RFC4861] are suggested: 242 o A node SHOULD send up to MAX_NEIGHBOR_ADVERTISEMENT unsolicited NA 243 packet with the Override flag cleared to all-routers multicast 244 address (ff02::2) as soon as one of the following events happens: 246 * (if Optimistic DAD is used): a new Optimistic address is 247 assigned to the node interface. 249 * (if Optimistic DAD is not used): an address changes the state 250 from tentative to preferred. 252 o Routers SHOULD create a new STALE ND cache entry upon receiving 253 unsolicited NAs. 255 It should be noted that some routing and switching platforms have 256 implemented such behaviour already. Administrators could enable the 257 creation of neighbor discovery cache entries based on unsolicited NA 258 packets sent from the previously unknown neighbors on that interface. 260 Network devices implementing FCFS SAVI might drop Neighbor 261 Advertisements received through a Validating Port which is in the 262 TENTATIVE state (see Section 2.3.2 of[RFC6620]). Therefore hosts 263 using Optimistic DAD might not benefit from the proposed solution if 264 FCFS SAVI is implemented on the network infrastructure. 265 [I-D.ietf-6man-grand] discusses in more details how the proposed 266 solution interacts with SAVI. 268 3. Solutions Considered but Discarded 270 The problem could be addressed from different angles. Possible 271 approaches are: 273 o Just do nothing. 275 o Migrate from the "reactive" Neighbor Discovery ([RFC4861]) to the 276 registration-based mechanisms ([RFC8505]). 278 o The router creates new entries in its Neighbor Cache by gleaning 279 from Neighbor Discovery DAD messages. 281 o The host initiates bidirectional communication to the router using 282 the host GUA. 284 o Making the probing logic on hosts more robust. 286 o Increasing the buffer size on routers. 288 o Transit dataplane traffic from an unknown address (an address w/o 289 the corresponding neighbor cache entry) triggers an address 290 resolution process on the router. 292 It should be noted that some of those options are already implemented 293 by some vendors. The following sections discuss those approaches and 294 the reasons they were discarded. 296 3.1. Do Nothing 298 One of the possible approaches might be to declare that everything is 299 working as intended and let the upper-layer protocols to deal with 300 packet loss. The obvious drawbacks include: 302 o Unhappy users. 304 o Many support tickets. 306 o More resistance to deploy IPv6 and IPv6-Only networks. 308 3.2. Change to the Registration-Based Neighbor Discovery 310 The most radical approach would be to move away from the reactive ND 311 as defined in [RFC4861] and expand the registration-based ND 312 ([RFC6775], [RFC8505]) used in Low-Power Wireless Personal Area 313 Networks (6LoWPANs) to the rest of IPv6 deployments. This option 314 requires some investigation and discussions and seems to be an 315 overkill for the problem described in this document. 317 3.3. Host Sending NS to the Router Address from Its GUA 319 The host could force creating a STALE entry for its GUA in the router 320 ND cache by sending the following Neighbor Solicitation message: 322 o The NS source address is the host GUA. 324 o The destination address is the default router IPv6 address. 326 o The Source Link-Layer Address option contains the host link-layer 327 address. 329 o The target address is the host default router address (the default 330 router address the host received in the RA). 332 The main disadvantages of this approach are: 334 o Would not work for Optimistic addresses as section 2.2 of 335 [RFC4429] explicitly prohibits sending Neighbor Solicitations from 336 an Optimistic Address. 338 o If first-hop redundancy is deployed in the network, the NS would 339 reach the active router only, so all backup routers (or all active 340 routers ex. one) would not get their neighbor cache updated. 342 o Some wireless devices are known to alter ND packets and perform 343 various non-obvious forms of ND proxy actions. In some cases, 344 unsolicited NAs might not even reach the routers. 346 3.4. Host Sending Router Solicitation from its GUA 348 The host could send a router solicitation message to 'all routers' 349 multicast address, using its GUA as a source. If the host link-layer 350 address is included in the Source Link-Layer Address option, the 351 router would create a STALE entry for the host GUA as per the section 352 6.2.6 of [RFC4861]. However, this approach can not be used if the 353 GUA is in optimistic state: section 2.2 of [RFC4429] explicitly 354 prohibits using an Optimistic Address as the source address of a 355 Router Solicitation with a SLLAO as it might disrupt the rightful 356 owner of the address in the case of a collision. So for the 357 optimistic addresses the host can send an RS without SLLAO included. 358 In that case the router may respond with either a multicast or a 359 unicast RA (only the latter would create a cache entry). 361 This approach has the following drawbacks: 363 o If the address is in the Optimistic state the RS can not contain 364 SLLAO. As a result the router would only create a cache entry if 365 the solicited RAs is sent as as a unicast. Routers sending 366 solicited RAs as multicast would not create a new cache entry as 367 they do not need to send a unicast packet back to the host. 369 o There might be a random delay between receiving an RS and sending 370 a unicast RA back (and creating a cache entry) which might 371 undermine the idea of creating the cache entry proactively. 373 o Some wireless devices are known to fiddle with ND packets and 374 perform various non-obvious forms of ND proxy actions. In some 375 cases the RS might not even reach the routers. 377 3.5. Routers Populating Their Caches by Gleaning From Neighbor 378 Discovery Packets 380 Routers may be able to learn about new addresses by gleaning from the 381 DAD Neighbor Solicitation messages. The router could listen to all 382 solicited node multicast address groups and upon receiving a Neighbor 383 Solicitation from the unspecified address search its Neighbor Cache 384 for the solicitation's Target Address. If no entry exists, the 385 router may create an entry, set its reachability state to 386 'INCOMPLETE' and start the address resolution for that entry. 388 The same solution was proposed in 389 [I-D.halpern-6man-nd-pre-resolve-addr]. Some routing vendors support 390 such optimization already. However, this approach has a number of 391 drawbacks and therefore should not be used as the only solution: 393 o Routers need to receive all multicast Neighbor Discovery packets 394 which might negatively impact the routers CPU. 396 o If the router starts the address resolution as soon as it receives 397 the DAD Neighbor Solicitation the host might be still performing 398 DAD and the target address might be tentative. In that case, the 399 host SHOULD silently ignore the received Neighbor Solicitation 400 from the router as per the Section 5.4.3 of [RFC4862]. As a 401 result the router might not be able to complete the address 402 resolution before the return traffic arrives. 404 3.6. Initiating Hosts-to-Routers Communication 406 The host may force the router to start address resolution by sending 407 a data packet such as ping or traceroute to its default router link- 408 local address, using the GUA as a source address. As the RTT to the 409 default router is lower than RTT to any off-link destinations it's 410 quite likely that the router would start the neighbor discovery 411 process for the host GUA before the first packet of the returning 412 traffic arrives. 414 The downside of this approach includes: 416 o Data packets to the router LLA could be blocked by security policy 417 or control plane protection mechanism. 419 o Additional overhead for routers control plane (in addition to 420 processing ND packets, the data packet needs to be processed as 421 well). 423 o Unless the data packet is sent to 'all routers' ff02::2 multicast 424 address, if the network provides a first-hop redundancy then only 425 the active router would create a new cache entry. 427 3.7. Transit Dataplane Traffic From a New Address Triggering Address 428 Resolution 430 When a router receives a transit packet, it might check the presence 431 of the neighbor cache entry for the packet source address and if the 432 entry does not exist start address resolution process. This approach 433 does ensure that a Neighbor Cache entry is proactively created every 434 time a new, previously unseen GUA is used for sending offlink 435 traffic. However this approach has a number of limitations, in 436 particular: 438 o If traffic flows are asymmetrical the return traffic might not 439 transit the same router as the original traffic which triggered 440 the address resolution. So the neighbor cache entry is created on 441 the "wrong" router, not the one which actually needs the neighbor 442 cache entry for the host address. 444 o The functionality needs to be limited to explicitly configured 445 networks/interfaces, as the router needs to distinguish between 446 onlink addresses (ones the router needs to have Neighbor Cache 447 entries for) and the rest of the address space. 449 o Implementing such functionality is much more complicated than all 450 other solutions as it would involve complex data-control planes 451 interaction. 453 4. IANA Considerations 455 This memo asks the IANA for no new parameters. 457 5. Security Considerations 459 This memo documents the operational issue and does not introduce any 460 new security considerations. Security considerations of the proposed 461 solution are discussed in the corresponding section of 462 [I-D.ietf-6man-grand]. 464 6. Acknowledgements 466 Thanks to the following people (in alphabetical order) for their 467 review and feedback: Mikael Abrahamsson, Lorenzo Colitti, Owen 468 DeLong, Igor Gashinsky, Fernando Gont, Tatuya Jinmei, Erik Kline, 469 Warren Kumari, Jordi Palet Martinez, Michael Richardson, Dave Thaler, 470 Pascal Thubert, Loganaden Velvindron, Eric Vyncke. 472 7. References 474 7.1. Normative References 476 [I-D.ietf-6man-grand] 477 Linkova, J., "Gratuitous Neighbor Discovery: Creating 478 Neighbor Cache Entries on First-Hop Routers", draft-ietf- 479 6man-grand-01 (work in progress), July 2020. 481 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 482 Requirement Levels", BCP 14, RFC 2119, 483 DOI 10.17487/RFC2119, March 1997, 484 . 486 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 487 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 488 2006, . 490 [RFC4429] Moore, N., "Optimistic Duplicate Address Detection (DAD) 491 for IPv6", RFC 4429, DOI 10.17487/RFC4429, April 2006, 492 . 494 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 495 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 496 DOI 10.17487/RFC4861, September 2007, 497 . 499 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 500 Address Autoconfiguration", RFC 4862, 501 DOI 10.17487/RFC4862, September 2007, 502 . 504 [RFC6583] Gashinsky, I., Jaeggli, J., and W. Kumari, "Operational 505 Neighbor Discovery Problems", RFC 6583, 506 DOI 10.17487/RFC6583, March 2012, 507 . 509 [RFC6620] Nordmark, E., Bagnulo, M., and E. Levy-Abegnoli, "FCFS 510 SAVI: First-Come, First-Served Source Address Validation 511 Improvement for Locally Assigned IPv6 Addresses", 512 RFC 6620, DOI 10.17487/RFC6620, May 2012, 513 . 515 [RFC6775] Shelby, Z., Ed., Chakrabarti, S., Nordmark, E., and C. 516 Bormann, "Neighbor Discovery Optimization for IPv6 over 517 Low-Power Wireless Personal Area Networks (6LoWPANs)", 518 RFC 6775, DOI 10.17487/RFC6775, November 2012, 519 . 521 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 522 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 523 May 2017, . 525 [RFC8305] Schinazi, D. and T. Pauly, "Happy Eyeballs Version 2: 526 Better Connectivity Using Concurrency", RFC 8305, 527 DOI 10.17487/RFC8305, December 2017, 528 . 530 [RFC8505] Thubert, P., Ed., Nordmark, E., Chakrabarti, S., and C. 531 Perkins, "Registration Extensions for IPv6 over Low-Power 532 Wireless Personal Area Network (6LoWPAN) Neighbor 533 Discovery", RFC 8505, DOI 10.17487/RFC8505, November 2018, 534 . 536 7.2. Informative References 538 [I-D.halpern-6man-nd-pre-resolve-addr] 539 Chen, I. and J. Halpern, "Triggering ND Address Resolution 540 on Receiving DAD-NS", draft-halpern-6man-nd-pre-resolve- 541 addr-00 (work in progress), January 2014. 543 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 544 Extensions for Stateless Address Autoconfiguration in 545 IPv6", RFC 4941, DOI 10.17487/RFC4941, September 2007, 546 . 548 Author's Address 550 Jen Linkova 551 Google 552 1 Darling Island Rd 553 Pyrmont, NSW 2009 554 AU 556 Email: furry@google.com