idnits 2.17.1 draft-ietf-v6ops-rfc6555bis-06.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 : ---------------------------------------------------------------------------- -- The document has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? -- The draft header indicates that this document obsoletes RFC6555, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (October 20, 2017) is 2380 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'Section 3' is mentioned on line 123, but not defined == Missing Reference: 'Section 4' is mentioned on line 168, but not defined == Missing Reference: 'Section 5' is mentioned on line 206, but not defined == Missing Reference: 'Section 6' is mentioned on line 173, but not defined ** Obsolete normative reference: RFC 6555 (Obsoleted by RFC 8305) Summary: 1 error (**), 0 flaws (~~), 5 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network D. Schinazi 3 Internet-Draft T. Pauly 4 Obsoletes: 6555 (if approved) Apple Inc. 5 Intended status: Standards Track October 20, 2017 6 Expires: April 23, 2018 8 Happy Eyeballs Version 2: Better Connectivity Using Concurrency 9 draft-ietf-v6ops-rfc6555bis-06 11 Abstract 13 Many communication protocols operated over the modern Internet use 14 host names. These often resolve to multiple IP addresses, each of 15 which may have different performance and connectivity 16 characteristics. Since specific addresses or address families (IPv4 17 or IPv6) may be blocked, broken, or sub-optimal on a network, clients 18 that attempt multiple connections in parallel have a higher chance of 19 establishing a connection sooner. This document specifies 20 requirements for algorithms that reduce this user-visible delay and 21 provides an example algorithm. 23 Status of This Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at https://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six months 34 and may be updated, replaced, or obsoleted by other documents at any 35 time. It is inappropriate to use Internet-Drafts as reference 36 material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on April 23, 2018. 40 Copyright Notice 42 Copyright (c) 2017 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents 47 (https://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with respect 50 to this document. Code Components extracted from this document must 51 include Simplified BSD License text as described in Section 4.e of 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 58 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 59 2. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 3 60 3. Hostname Resolution Query Handling . . . . . . . . . . . . . 4 61 3.1. Handling Multiple DNS Server Addresses . . . . . . . . . 5 62 4. Sorting Addresses . . . . . . . . . . . . . . . . . . . . . . 5 63 5. Connection Attempts . . . . . . . . . . . . . . . . . . . . . 6 64 6. DNS Answer Changes during Happy Eyeballs Connection Setup . . 7 65 7. Supporting IPv6-only Networks with NAT64 and DNS64 . . . . . 7 66 7.1. IPv4 Address Literals . . . . . . . . . . . . . . . . . . 8 67 7.2. Host Names with Broken AAAA Records . . . . . . . . . . . 9 68 7.3. Virtual Private Networks . . . . . . . . . . . . . . . . 10 69 8. Summary of Configurable Values . . . . . . . . . . . . . . . 11 70 9. Limitations . . . . . . . . . . . . . . . . . . . . . . . . . 11 71 9.1. Path Maximum Transmission Unit Discovery . . . . . . . . 12 72 9.2. Application Layer . . . . . . . . . . . . . . . . . . . . 12 73 9.3. Hiding Operational Issues . . . . . . . . . . . . . . . . 12 74 10. Security Considerations . . . . . . . . . . . . . . . . . . . 12 75 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 76 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 13 77 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 13 78 13.1. Normative References . . . . . . . . . . . . . . . . . . 13 79 13.2. Informative References . . . . . . . . . . . . . . . . . 14 80 Appendix A. Differences from RFC6555 . . . . . . . . . . . . . . 14 81 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 15 83 1. Introduction 85 Many communication protocols operated over the modern Internet use 86 host names. These often resolve to multiple IP addresses, each of 87 which may have different performance and connectivity 88 characteristics. Since specific addresses or address families (IPv4 89 or IPv6) may be blocked, broken, or sub-optimal on a network, clients 90 that attempt multiple connections in parallel have a higher chance of 91 establishing a connection sooner. This document specifies 92 requirements for algorithms that reduce this user-visible delay and 93 provides an example algorithm. 95 This document defines the algorithm for "Happy Eyeballs", a technique 96 of reducing user-visible delays on dual-stack hosts. This definition 97 obsoletes the original description in [RFC6555]. Now that this 98 approach has been deployed at scale and measured for several years, 99 the algorithm specification can be refined to improve its reliability 100 and generalization. 102 The Happy Eyeballs algorithm of racing resolved addresses has several 103 stages of ordering and racing to avoid delays to the user whenever 104 possible, while preferring the use of IPv6. This document discusses 105 how to handle DNS queries when starting a connection on a dual-stack 106 client, how to create an ordered list of destination addresses to 107 which to attempt connections, and how to race the connection 108 attempts. 110 1.1. Requirements Language 112 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 113 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 114 "OPTIONAL" in this document are to be interpreted as described in 115 "Key words for use in RFCs to Indicate Requirement Levels" [RFC2119]. 117 2. Overview 119 This document defines a method of connection establishment, named 120 "Happy Eyeballs Connection Setup". This approach has several 121 distinct phases: 123 1. Initiation of asynchronous DNS queries [Section 3] 125 2. Sorting of resolved destination addresses [Section 4] 127 3. Initiation of asynchronous connection attempts [Section 5] 129 4. Establishment of one connection, which cancels all other attempts 130 Note that this document assumes that the host destination address 131 preference policy favors IPv6 over IPv4. IPv6 has many desirable 132 properties designed to be improvements over IPv4 [RFC8200]. If the 133 host is configured to have a different preference, the 134 recommendations in this document can be easily adapted. 136 3. Hostname Resolution Query Handling 138 When a client has both IPv4 and IPv6 connectivity, and is trying to 139 establish a connection with a named host, it needs to send out both 140 AAAA and A DNS queries. Both queries SHOULD be made as soon after 141 one another as possible, with the AAAA query made first, immediately 142 followed by the A query. 144 Implementations SHOULD NOT wait for both families of answers to 145 return before attempting connection establishment. If one query 146 fails to return, or takes significantly longer to return, waiting for 147 the second address family can significantly delay the connection 148 establishment of the first one. Therefore, the client SHOULD treat 149 DNS resolution as asynchronous. Note that if the platform does not 150 offer an asynchronous DNS API, this behavior can be simulated by 151 making two separate synchronous queries on different threads, one per 152 address family. 154 The RECOMMENDED algorithm proceeds as follows: if a positive AAAA 155 response (a response with at least one valid AAAA record) is received 156 first, the first IPv6 connection attempt is immediately started. If 157 a positive A response is received first due to reordering, the client 158 SHOULD wait for a short time for the AAAA response to ensure 159 preference is given to IPv6 (it is common for the AAAA response to 160 follow the A response by a few milliseconds). This delay will be 161 referred to as the "Resolution Delay". The RECOMMENDED value for the 162 Resolution Delay is 50 milliseconds. If a positive AAAA response is 163 received within the Resolution Delay period, the client immediately 164 starts the IPv6 connection attempt. If a negative AAAA response (no 165 error, no data) is received within the Resolution Delay period or the 166 AAAA response has not been received by the end of the Resolution 167 Delay period, the client SHOULD proceed to Sorting Addresses 168 [Section 4] and staggered connection attempts [Section 5] using any 169 IPv4 addresses returned so far. If the AAAA response arrives while 170 these connection attempts are in progress, but before any connection 171 has been established, then the newly received IPv6 addresses are 172 incorporated into the list of available candidate addresses 173 [Section 6] and the process of connection attempts will continue with 174 the IPv6 addresses added, until one connection is established. 176 3.1. Handling Multiple DNS Server Addresses 178 If multiple DNS server addresses are configured for the current 179 network, the client may have the option of sending its DNS queries 180 over IPv4 or IPv6. In keeping with the Happy Eyeballs approach, 181 queries SHOULD be sent over IPv6 first (note that this is not 182 referring to the sending of AAAA or A queries, but rather the address 183 of the DNS server itself and IP version used to transport DNS 184 messages). If DNS queries sent to the IPv6 address do not receive 185 responses, that address may be marked as penalized, and queries can 186 be sent to other DNS server addresses. 188 As native IPv6 deployments become more prevalent, and IPv4 addresses 189 are exhausted, it is expected that IPv6 connectivity will have 190 preferential treatment within networks. If a DNS server is 191 configured to be accessible over IPv6, IPv6 should be assumed to be 192 the preferred address family. 194 Client systems SHOULD NOT have an explicit limit to the number of DNS 195 servers that can be configured, either manually or by the network. 196 If such a limit is required by hardware limitations, it is 197 RECOMMENDED to use at least one address from each address family from 198 the available list. 200 4. Sorting Addresses 202 Before attempting to connect to any of the resolved destination 203 addresses, the client should define the order in which to start the 204 attempts. Once the order has been defined, the client can use a 205 simple algorithm for racing each option after a short delay 206 [Section 5]. It is important that the ordered list involves all 207 addresses from both families that have been received by this point, 208 as this allows the client to get the racing effect of Happy Eyeballs 209 for the entire list, not just the first IPv4 and first IPv6 210 addresses. 212 First, the client MUST sort the addresses received up to this point 213 using Destination Address Selection ([RFC6724], Section 6). 215 If the client is stateful and has history of expected round-trip 216 times (RTT) for the routes to access each address, it SHOULD add a 217 Destination Address Selection rule between rules 8 and 9 that prefers 218 addresses with lower RTTs. If the client keeps track of which 219 addresses it has used in the past, it SHOULD add another destination 220 address selection rule between the RTT rule and rule 9, which prefers 221 used addresses over unused ones. This helps servers that use the 222 client's IP address during authentication, as is the case for TCP 223 Fast Open [RFC7413] and some HTTP cookies. This historical data MUST 224 NOT be used across networks, and SHOULD be flushed on network 225 changes. 227 Next, the client SHOULD modify the ordered list to interleave address 228 families. Whichever address family is first in the list should be 229 followed by an address of the other address family; that is, if the 230 first address in the sorted list is IPv6, then the first IPv4 address 231 should be moved up in the list to be second in the list. An 232 implementation MAY want to favor one address family more by allowing 233 multiple addresses of that family to be attempted before trying the 234 other family. The number of contiguous addresses of the first 235 address family will be referred to as the "First Address Family 236 Count", and can be a configurable value. This is performed to avoid 237 waiting through a long list of addresses from a given address family 238 if connectivity over that address family is impaired. 240 Note that the address selection described in this section only 241 applies to destination addresses; Source Address Selection 242 ([RFC6724], Section 5) is performed once per destination address and 243 is out of scope of this document. 245 5. Connection Attempts 247 Once the list of addresses received up to this point has been 248 constructed, the client will attempt to make connections. In order 249 to avoid unreasonable network load, connection attempts SHOULD NOT be 250 made simultaneously. Instead, one connection attempt to a single 251 address is started first, followed by the others in the list, one at 252 a time. Starting a new connection attempt does not affect previous 253 attempts, as multiple connection attempts may occur in parallel. 254 Once one of the connection attempts succeeds (generally when the TCP 255 handshake completes), all other connections attempts that have not 256 yet succeeded SHOULD be cancelled. Any address that was not yet 257 attempted as a connection SHOULD be ignored. At that time, the 258 asynchronous DNS query MAY be cancelled as new addresses will not be 259 used for this connection. However, the DNS client resolver SHOULD 260 still process DNS replies from the network for a short period of time 261 (RECOMMENDED at 1 second), as they will populate the DNS cache and 262 can be used for subsequent connections. 264 A simple implementation can have a fixed delay for how long to wait 265 before starting the next connection attempt. This delay is referred 266 to as the "Connection Attempt Delay". One recommended value for a 267 default delay is 250 milliseconds. A more nuanced implementation's 268 delay should correspond to the time when the previous attempt is 269 sending its second TCP SYN, based on TCP's retransmission timer 270 [RFC6298]. If the client has historical RTT data gathered from other 271 connections to the same host or prefix, it can use this information 272 to influence its delay. Note that this algorithm should only try to 273 approximate the time of the first SYN retransmission, and not any 274 further retransmissions which may be influenced by exponential timer 275 back off. 277 The Connection Attempt Delay MUST have a lower bound, especially if 278 it is computed using historical data. More specifically, a 279 subsequent connection MUST NOT be started within 10 milliseconds of 280 the previous attempt. The recommended minimum value is 100 281 milliseconds, which is referred to as the "Minimum Connection Attempt 282 Delay". This minimum value is required to avoid congestion collapse 283 in the presence of high packet loss rates. The Connection Attempt 284 Delay SHOULD have an upper bound, referred to as the "Maximum 285 Connection Attempt Delay". The current recommended value is 2 286 seconds. 288 6. DNS Answer Changes during Happy Eyeballs Connection Setup 290 If, during the course of connection establishment, the DNS answers 291 change either by adding resolved addresses (for example, due to DNS 292 push notifications [DNS-PUSH]), or removing previously resolved 293 addresses (for example, due to expiry of the TTL on that DNS record), 294 the client should react based on its current progress. 296 If an address is removed from the list that already had a connection 297 attempt started, the connection attempt SHOULD NOT be cancelled, but 298 rather be allowed to continue. If the removed address had not yet 299 had a connection attempt started, it SHOULD be removed from the list 300 of addresses to try. 302 If an address is added to the list, it should be sorted into the list 303 of addresses not yet attempted according to the rules above 304 (Section 4). 306 7. Supporting IPv6-only Networks with NAT64 and DNS64 308 While many IPv6 transition protocols have been standardized and 309 deployed, most are transparent to client devices. The combined use 310 of NAT64 [RFC6146] and DNS64 [RFC6147] is a popular solution that is 311 being deployed and requires changes in client devices. One possible 312 way to handle these networks is for the client device networking 313 stack to implement 464XLAT [RFC6877]. 464XLAT has the advantage of 314 not requiring changes to user space software, however it requires 315 per-packet translation if the application is using IPv4 literals and 316 does not encourage client application software to support native 317 IPv6. On platforms that do not support 464XLAT, the Happy Eyeballs 318 engine SHOULD follow the recommendations in this section to properly 319 support IPv6-only networks with NAT64 and DNS64. 321 The features described in this section SHOULD only be enabled when 322 the host detects one of these networks. A simple heuristic to 323 achieve that is to check if the network offers routable IPv6 324 addressing, does not offer routable IPv4 addressing, and offers a DNS 325 resolver address. 327 7.1. IPv4 Address Literals 329 If client applications or users wish to connect to IPv4 address 330 literals, the Happy Eyeballs engine will need to perform NAT64 331 address synthesis for them. The solution is similar to "Bump-in-the- 332 Host" [RFC6535] but is implemented inside the Happy Eyeballs library. 334 When an IPv4 address is passed in to the library instead of a host 335 name, the device queries the network for the NAT64 prefix using 336 "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" 337 [RFC7050] then synthesizes an appropriate IPv6 address (or several) 338 using the encoding described in "IPv6 Addressing of IPv4/IPv6 339 Translators" [RFC6052]. The synthesized addresses are then inserted 340 into the list of addresses as if they were results from DNS queries; 341 connection attempts follow the algorithm described above (Section 5). 343 7.2. Host Names with Broken AAAA Records 345 At the time of writing, there exist a small but non negligible number 346 of host names that resolve to valid A records and broken AAAA 347 records, which we define as AAAA records that contain seemingly valid 348 IPv6 addresses but those addresses never reply when contacted on the 349 usual ports. These can be for example caused by: 351 o Mistyping of the IPv6 address in the DNS zone configuration 353 o Routing black holes 355 o Service outages 357 While an algorithm complying with the other sections of this document 358 would correctly handle such host names on a dual-stack network, they 359 will not necessarily function correctly on IPv6-only networks with 360 NAT64 and DNS64. Since DNS64 recursive resolvers rely on the 361 authoritative name servers sending negative ("no error no answer") 362 responses for AAAA records in order to synthesize, they will not 363 synthesize records for these particular host names, and will instead 364 pass through the broken AAAA record. 366 In order to support these scenarios, the client device needs to query 367 the DNS for the A record then perform local synthesis. Since these 368 types of host names are rare, and in order to minimize load on DNS 369 servers, this A query should only be performed when the client has 370 given up on the AAAA records it initially received. This can be 371 achieved by using a longer timeout, referred to as the "Last Resort 372 Local Synthesis Delay" and RECOMMENDED at 2 seconds. The timer is 373 started when the last connection attempt is fired. If no connection 374 attempt has succeeded when this timer fires, the device queries the 375 DNS for the IPv4 address and on reception of a valid A record, treats 376 it as if it were provided by the application (Section 7.1). 378 7.3. Virtual Private Networks 380 Some Virtual Private Networks (VPN) may be configured to handle DNS 381 queries from the device. The configuration could encompass all 382 queries, or a subset such as "*.internal.example.com". These VPNs 383 can also be configured to only route part of the IPv4 address space, 384 such as 192.0.2.0/24. However, if an internal hostname resolves to 385 an external IPv4 address, these can cause issues if the underlying 386 network is IPv6-only. As an example, let's assume that 387 "www.internal.example.com" has exactly one A record, 198.51.100.42, 388 and no AAAA records. The client will send the DNS query to the 389 company's recursive resolver and that resolver will reply with these 390 records. The device now only has an IPv4 address to connect to, and 391 no route to that address. Since the company's resolver does not know 392 the NAT64 prefix of the underlying network, it cannot synthesize the 393 address. Similarly, the underlying network's DNS64 recursive 394 resolver does not know the company's internal addresses, so it cannot 395 resolve the hostname. Because of this, the client device needs to 396 resolve the A record using the company's resolver then locally 397 synthesize an IPv6 address, as if the resolved IPv4 address were 398 provided by the application (Section 7.1). 400 8. Summary of Configurable Values 402 The values that may be configured as defaults on a client for use in 403 Happy Eyeballs are as follows: 405 o Resolution Delay (Section 3): The time to wait for a AAAA response 406 after receiving an A response. RECOMMENDED at 50 milliseconds. 408 o First Address Family Count (Section 4): The number of addresses 409 belonging to the first address family (such as IPv6) that should 410 be attempted before attempting another address family. 411 RECOMMENDED as 1, or 2 to more aggressively favor one address 412 family. 414 o Connection Attempt Delay (Section 5): The time to wait between 415 connection attempts in the absence of RTT data. RECOMMENDED at 416 250 milliseconds. 418 o Minimum Connection Attempt Delay (Section 5): The minimum time to 419 wait between connection attempts. RECOMMENDED at 100 420 milliseconds. MUST NOT be less than 10 milliseconds. 422 o Maximum Connection Attempt Delay (Section 5): The maximum time to 423 wait between connection attempts. RECOMMENDED at 2 seconds. 425 o Last Resort Local Synthesis Delay (Section 7.2): The time to wait 426 after starting the last IPv6 attempt and before sending the A 427 query. RECOMMENDED at 2 seconds. 429 The delay values described in this section were determined 430 empirically by measuring the timing of connections on a very wide set 431 of production devices. They were picked to reduce wait times noticed 432 by users while minimizing load on the network. As time passes, it is 433 expected that the properties of networks will evolve. For that 434 reason, it is expected that these values will change over time. 435 Implementors should feel welcome to use different values without 436 changing this specification. Since IPv6 issues are expected to be 437 less common, the delays SHOULD be increased with time as client 438 software is updated. 440 9. Limitations 442 Happy Eyeballs will handle initial connection failures at the TCP/IP 443 layer, however other failures or performance issues may still affect 444 the chosen connection. 446 9.1. Path Maximum Transmission Unit Discovery 448 Since Happy Eyeballs is only active during the initial handshake and 449 TCP does not pass the initial handshake, issues related to MTU can be 450 masked and go unnoticed during Happy Eyeballs. Solving this issue is 451 out of scope of this document. One solution is to use Packetization 452 Layer Path MTU Discovery [RFC4821]. 454 9.2. Application Layer 456 If the DNS returns multiple addresses for different application 457 servers, the application itself may not be operational and functional 458 on all of them. Common examples include Transport Layer Security 459 (TLS) and the Hypertext Transport Protocol (HTTP). 461 9.3. Hiding Operational Issues 463 It has been observed in practice that Happy Eyeballs can hide issues 464 in networks. For example, if a misconfiguration causes IPv6 to 465 consistently fail on a given network while IPv4 is still functional, 466 Happy Eyeballs may impair the operator's ability to notice the issue. 467 It is recommended that network operators deploy external means of 468 monitoring to ensure functionality of all address families. 470 10. Security Considerations 472 This memo has no direct security considerations. 474 Note that applications should not rely upon a stable hostname-to- 475 address mapping to ensure any security properties, since DNS results 476 may change between queries. Happy Eyeballs may make it more likely 477 that subsequent connections to a single hostname use different IP 478 addresses. 480 11. IANA Considerations 482 This memo includes no request to IANA. 484 12. Acknowledgments 486 The authors thank Dan Wing, Andrew Yourtchenko, and everyone else who 487 worked on the original Happy Eyeballs design [RFC6555], Josh 488 Graessley, Stuart Cheshire, and the rest of team at Apple that helped 489 implement and instrument this algorithm, and Jason Fesler and Paul 490 Saab who helped measure and refine this algorithm. The authors would 491 also like to thank Fred Baker, Nick Chettle, Lorenzo Colitti, Igor 492 Gashinsky, Geoff Huston, Jen Linkova, Paul Hoffman, Philip Homburg, 493 Warren Kumari, Erik Nygren, Jordi Palet Martinez, Rui Paulo, Stephen 494 Strowes, Jinmei Tatuya, Dave Thaler, Joe Touch and James Woodyatt for 495 their input and contributions. 497 13. References 499 13.1. Normative References 501 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 502 Requirement Levels", BCP 14, RFC 2119, 503 DOI 10.17487/RFC2119, March 1997, 504 . 506 [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU 507 Discovery", RFC 4821, DOI 10.17487/RFC4821, March 2007, 508 . 510 [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. 511 Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, 512 DOI 10.17487/RFC6052, October 2010, 513 . 515 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 516 NAT64: Network Address and Protocol Translation from IPv6 517 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 518 April 2011, . 520 [RFC6147] Bagnulo, M., Sullivan, A., Matthews, P., and I. van 521 Beijnum, "DNS64: DNS Extensions for Network Address 522 Translation from IPv6 Clients to IPv4 Servers", RFC 6147, 523 DOI 10.17487/RFC6147, April 2011, 524 . 526 [RFC6298] Paxson, V., Allman, M., Chu, J., and M. Sargent, 527 "Computing TCP's Retransmission Timer", RFC 6298, 528 DOI 10.17487/RFC6298, June 2011, 529 . 531 [RFC6535] Huang, B., Deng, H., and T. Savolainen, "Dual-Stack Hosts 532 Using "Bump-in-the-Host" (BIH)", RFC 6535, 533 DOI 10.17487/RFC6535, February 2012, 534 . 536 [RFC6555] Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with 537 Dual-Stack Hosts", RFC 6555, DOI 10.17487/RFC6555, April 538 2012, . 540 [RFC6724] Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown, 541 "Default Address Selection for Internet Protocol Version 6 542 (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012, 543 . 545 [RFC7050] Savolainen, T., Korhonen, J., and D. Wing, "Discovery of 546 the IPv6 Prefix Used for IPv6 Address Synthesis", 547 RFC 7050, DOI 10.17487/RFC7050, November 2013, 548 . 550 13.2. Informative References 552 [DNS-PUSH] 553 Pusateri, T. and S. Cheshire, "DNS Push Notifications", 554 Work in Progress, draft-ietf-dnssd-push, March 2017. 556 [RFC6877] Mawatari, M., Kawashima, M., and C. Byrne, "464XLAT: 557 Combination of Stateful and Stateless Translation", 558 RFC 6877, DOI 10.17487/RFC6877, April 2013, 559 . 561 [RFC7413] Cheng, Y., Chu, J., Radhakrishnan, S., and A. Jain, "TCP 562 Fast Open", RFC 7413, DOI 10.17487/RFC7413, December 2014, 563 . 565 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 566 (IPv6) Specification", STD 86, RFC 8200, 567 DOI 10.17487/RFC8200, July 2017, 568 . 570 Appendix A. Differences from RFC6555 572 "Happy Eyeballs: Success with Dual-Stack Hosts" [RFC6555] mostly 573 concentrates on how to stagger connections to a hostname that has an 574 AAAA and an A record. This document additionally discusses: 576 o how to perform DNS queries to obtain these addresses 578 o how to handle multiple addresses from each address family 579 o how to handle DNS updates while connections are being raced 581 o how to leverage historical information 583 o how to support IPv6-only networks with NAT64 and DNS64 585 Authors' Addresses 587 David Schinazi 588 Apple Inc. 589 1 Infinite Loop 590 Cupertino, California 95014 591 US 593 Email: dschinazi@apple.com 595 Tommy Pauly 596 Apple Inc. 597 1 Infinite Loop 598 Cupertino, California 95014 599 US 601 Email: tpauly@apple.com