idnits 2.17.1 draft-ietf-v6ops-rfc6555bis-05.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 (September 10, 2017) is 2391 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 120, but not defined == Missing Reference: 'Section 4' is mentioned on line 164, but not defined == Missing Reference: 'Section 5' is mentioned on line 202, but not defined == Missing Reference: 'Section 6' is mentioned on line 169, 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 September 10, 2017 6 Expires: March 14, 2018 8 Happy Eyeballs Version 2: Better Connectivity Using Concurrency 9 draft-ietf-v6ops-rfc6555bis-05 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 March 14, 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 . . . . . 8 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 . . . . . . . . . . . . . . . . . . . . . . . 12 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 expands on "Happy Eyeballs" [RFC6555], a technique of 96 reducing user-visible delays on dual-stack hosts. Now that this 97 approach has been deployed at scale and measured for several years, 98 the algorithm specification can be refined to improve its reliability 99 and generalization. This document recommends an algorithm of racing 100 resolved addresses that has several stages of ordering and racing to 101 avoid delays to the user whenever possible, while preferring the use 102 of IPv6. Specifically, it discusses how to handle DNS queries when 103 starting a connection on a dual-stack client, how to create an 104 ordered list of destination addresses to which to attempt 105 connections, and how to race the connection attempts. 107 1.1. Requirements Language 109 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 110 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 111 "OPTIONAL" in this document are to be interpreted as described in 112 "Key words for use in RFCs to Indicate Requirement Levels" [RFC2119]. 114 2. Overview 116 This document defines a method of connection establishment, named 117 "Happy Eyeballs Connection Setup". This approach has several 118 distinct phases: 120 1. Initiation of asynchronous DNS queries [Section 3] 122 2. Sorting of resolved destination addresses [Section 4] 124 3. Initiation of asynchronous connection attempts [Section 5] 126 4. Establishment of one connection, which cancels all other attempts 127 Note that this document assumes that the host destination address 128 preference policy favors IPv6 over IPv4. If the host is configured 129 differently, the recommendations in this document can be easily 130 adapted. 132 3. Hostname Resolution Query Handling 134 When a client has both IPv4 and IPv6 connectivity, and is trying to 135 establish a connection with a named host, it needs to send out both 136 AAAA and A DNS queries. Both queries SHOULD be made as soon after 137 one another as possible, with the AAAA query made first, immediately 138 followed by the A query. 140 Implementations SHOULD NOT wait for both families of answers to 141 return before attempting connection establishment. If one query 142 fails to return, or takes significantly longer to return, waiting for 143 the second address family can significantly delay the connection 144 establishment of the first one. Therefore, the client SHOULD treat 145 DNS resolution as asynchronous. Note that if the platform does not 146 offer an asynchronous DNS API, this behavior can be simulated by 147 making two separate synchronous queries on different threads, one per 148 address family. 150 The RECOMMENDED algorithm proceeds as follows: if a positive AAAA 151 response (a response with at least one valid AAAA record) is received 152 first, the first IPv6 connection attempt is immediately started. If 153 a positive A response is received first due to reordering, the client 154 SHOULD wait for a short time for the AAAA response to ensure 155 preference is given to IPv6 (it is common for the AAAA response to 156 follow the A response by a few milliseconds). This delay will be 157 referred to as the "Resolution Delay". The RECOMMENDED value for the 158 Resolution Delay is 50 milliseconds. If a positive AAAA response is 159 received within the Resolution Delay period, the client immediately 160 starts the IPv6 connection attempt. If a negative AAAA response (no 161 error, no data) is received within the Resolution Delay period or the 162 AAAA response has not been received by the end of the Resolution 163 Delay period, the client SHOULD proceed to Sorting Addresses 164 [Section 4] and staggered connection attempts [Section 5] using any 165 IPv4 addresses returned so far. If the AAAA response arrives while 166 these connection attempts are in progress, but before any connection 167 has been established, then the newly received IPv6 addresses are 168 incorporated into the list of available candidate addresses 169 [Section 6] and the process of connection attempts will continue with 170 the IPv6 addresses added, until one connection is established. 172 3.1. Handling Multiple DNS Server Addresses 174 If multiple DNS server addresses are configured for the current 175 network, the client may have the option of sending its DNS queries 176 over IPv4 or IPv6. In keeping with the Happy Eyeballs approach, 177 queries SHOULD be sent over IPv6 first (note that this is not 178 referring to the sending of AAAA or A queries, but rather the address 179 of the DNS server itself and IP version used to transport DNS 180 messages). If DNS queries sent to the IPv6 address do not receive 181 responses, that address may be marked as penalized, and queries can 182 be sent to other DNS server addresses. 184 As native IPv6 deployments become more prevalent, and IPv4 addresses 185 are exhausted, it is expected that IPv6 connectivity will have 186 preferential treatment within networks. If a DNS server is 187 configured to be accessible over IPv6, IPv6 should be assumed to be 188 the preferred address family. 190 Client systems SHOULD NOT have an explicit limit to the number of DNS 191 servers that can be configured, either manually or by the network. 192 If such a limit is required by hardware limitations, it is 193 RECOMMENDED to use at least one address from each address family from 194 the available list. 196 4. Sorting Addresses 198 Before attempting to connect to any of the resolved destination 199 addresses, the client should define the order in which to start the 200 attempts. Once the order has been defined, the client can use a 201 simple algorithm for racing each option after a short delay 202 [Section 5]. It is important that the ordered list involves all 203 addresses from both families that have been received by this point, 204 as this allows the client to get the racing effect of Happy Eyeballs 205 for the entire list, not just the first IPv4 and first IPv6 206 addresses. 208 First, the client MUST sort the addresses received up to this point 209 using Destination Address Selection ([RFC6724], Section 6). 211 If the client is stateful and has history of expected round-trip 212 times (RTT) for the routes to access each address, it SHOULD add a 213 Destination Address Selection rule between rules 8 and 9 that prefers 214 addresses with lower RTTs. If the client keeps track of which 215 addresses it has used in the past, it SHOULD add another destination 216 address selection rule between the RTT rule and rule 9, which prefers 217 used addresses over unused ones. This helps servers that use the 218 client's IP address during authentication, as is the case for TCP 219 Fast Open [RFC7413] and some HTTP cookies. This historical data MUST 220 NOT be used across networks, and SHOULD be flushed on network 221 changes. 223 Next, the client SHOULD modify the ordered list to interleave address 224 families. Whichever address family is first in the list should be 225 followed by an address of the other address family; that is, if the 226 first address in the sorted list is IPv6, then the first IPv4 address 227 should be moved up in the list to be second in the list. An 228 implementation MAY want to favor one address family more by allowing 229 multiple addresses of that family to be attempted before trying the 230 other family. The number of contiguous addresses of the first 231 address family will be referred to as the "First Address Family 232 Count", and can be a configurable value. This is performed to avoid 233 waiting through a long list of addresses from a given address family 234 if connectivity over that address family is impaired. 236 Note that the address selection described in this section only 237 applies to destination addresses; Source Address Selection 238 ([RFC6724], Section 5) is performed once per destination address and 239 is out of scope of this document. 241 5. Connection Attempts 243 Once the list of addresses received up to this point has been 244 constructed, the client will attempt to make connections. In order 245 to avoid unreasonable network load, connection attempts SHOULD NOT be 246 made simultaneously. Instead, one connection attempt to a single 247 address is started first, followed by the others in the list, one at 248 a time. Starting a new connection attempt does not affect previous 249 attempts, as multiple connection attempts may occur in parallel. 250 Once one of the connection attempts succeeds (generally when the TCP 251 handshake completes), all other connections attempts that have not 252 yet succeeded SHOULD be cancelled. Any address that was not yet 253 attempted as a connection SHOULD be ignored. At that time, the 254 asynchronous DNS query MAY be cancelled as new addresses will not be 255 used for this connection. However, the DNS client resolver SHOULD 256 still process DNS replies from the network for a short period of time 257 (RECOMMENDED at 1 second), as they will populate the DNS cache and 258 can be used for subsequent connections. 260 A simple implementation can have a fixed delay for how long to wait 261 before starting the next connection attempt. This delay is referred 262 to as the "Connection Attempt Delay". One recommended value for this 263 delay is 250 milliseconds. If the client has historical RTT data, it 264 can also use the expected RTT to choose a more nuanced delay value. 265 The recommended formula for calculating the delay after starting a 266 connection attempt is: MAX( 1.25 * RTT_MEAN + 4 * RTT_VARIANCE, 2 * 267 RTT_MEAN ), where the RTT values are based on the statistics for 268 previous address used. If the TCP implementation leverages 269 historical RTT data to compute SYN timeout, these algorithms should 270 match so that a new attempt will be started at the same time as the 271 previous is sending its second TCP SYN. While TCP implementations 272 often leverage an exponential backoff when they detect packet loss, 273 the "Connection Attempt Delay" SHOULD NOT perform such an aggressive 274 backoff, as it would harm user experience. 276 The Connection Attempt Delay MUST have a lower bound, especially if 277 it is computed using historical data. More specifically, a 278 subsequent connection MUST NOT be started within 10 milliseconds of 279 the previous attempt. The recommended minimum value is 100 280 milliseconds, which is referred to as the "Minimum Connection Attempt 281 Delay". This minimum value is required to avoid congestion collapse 282 in the presence of high packet loss rates. The Connection Attempt 283 Delay SHOULD have an upper bound, referred to as the "Maximum 284 Connection Attempt Delay". The current recommended value is 2 285 seconds. 287 6. DNS Answer Changes during Happy Eyeballs Connection Setup 289 If, during the course of connection establishment, the DNS answers 290 change either by adding resolved addresses (for example, due to DNS 291 push notifications [DNS-PUSH]), or removing previously resolved 292 addresses (for example, due to expiry of the TTL on that DNS record), 293 the client should react based on its current progress. 295 If an address is removed from the list that already had a connection 296 attempt started, the connection attempt SHOULD NOT be cancelled, but 297 rather be allowed to continue. If the removed address had not yet 298 had a connection attempt started, it SHOULD be removed from the list 299 of addresses to try. 301 If an address is added to the list, it should be sorted into the list 302 of addresses not yet attempted according to the rules above 303 (Section 4). 305 7. Supporting IPv6-only Networks with NAT64 and DNS64 307 While many IPv6 transition protocols have been standardized and 308 deployed, most are transparent to client devices. The combined use 309 of NAT64 [RFC6146] and DNS64 [RFC6147] is a popular solution that is 310 being deployed and requires changes in client devices. One possible 311 way to handle these networks is for the client device networking 312 stack to implement 464XLAT [RFC6877]. 464XLAT has the advantage of 313 not requiring changes to user space software, however it requires 314 per-packet translation if the application is using IPv4 literals and 315 does not encourage client application software to support native 316 IPv6. On platforms that do not support 464XLAT, the Happy Eyeballs 317 engine SHOULD follow the recommendations in this section to properly 318 support IPv6-only networks with NAT64 and DNS64. 320 The features described in this section SHOULD only be enabled when 321 the host detects one of these networks. A simple heuristic to 322 achieve that is to check if the network offers routable IPv6 323 addressing, does not offer routable IPv4 addressing, and offers a DNS 324 resolver address. 326 7.1. IPv4 Address Literals 328 If client applications or users wish to connect to IPv4 address 329 literals, the Happy Eyeballs engine will need to perform NAT64 330 address synthesis for them. The solution is similar to "Bump-in-the- 331 Host" [RFC6535] but is implemented inside the Happy Eyeballs library. 333 When an IPv4 address is passed in to the library instead of a host 334 name, the device queries the network for the NAT64 prefix using 335 "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" 336 [RFC7050] then synthesizes an appropriate IPv6 address (or several) 337 using the encoding described in "IPv6 Addressing of IPv4/IPv6 338 Translators" [RFC6052]. The synthesized addresses are then inserted 339 into the list of addresses as if they were results from DNS queries; 340 connection attempts follow the algorithm described above (Section 5). 342 7.2. Host Names with Broken AAAA Records 344 At the time of writing, there exist a small but non negligible number 345 of host names that resolve to valid A records and broken AAAA 346 records, which we define as AAAA records that contain seemingly valid 347 IPv6 addresses but those addresses never reply when contacted on the 348 usual ports. These can be for example caused by: 350 o Mistyping of the IPv6 address in the DNS zone configuration 352 o Routing black holes 354 o Service outages 356 While an algorithm complying with the other sections of this document 357 would correctly handle such host names on a dual-stack network, they 358 will not necessarily function correctly on IPv6-only networks with 359 NAT64 and DNS64. Since DNS64 recursive resolvers rely on the 360 authoritative name servers sending negative ("no error no answer") 361 responses for AAAA records in order to synthesize, they will not 362 synthesize records for these particular host names, and will instead 363 pass through the broken AAAA record. 365 In order to support these scenarios, the client device needs to query 366 the DNS for the A record then perform local synthesis. Since these 367 types of host names are rare, and in order to minimize load on DNS 368 servers, this A query should only be performed when the client has 369 given up on the AAAA records it initially received. This can be 370 achieved by using a longer timeout, referred to as the "Last Resort 371 Local Synthesis Delay" and RECOMMENDED at 2 seconds. The timer is 372 started when the last connection attempt is fired. If no connection 373 attempt has succeeded when this timer fires, the device queries the 374 DNS for the IPv4 address and on reception of a valid A record, treats 375 it as if it were provided by the application (Section 7.1). 377 7.3. Virtual Private Networks 379 Some Virtual Private Networks (VPN) may be configured to handle DNS 380 queries from the device. The configuration could encompass all 381 queries, or a subset such as "*.internal.example.com". These VPNs 382 can also be configured to only route part of the IPv4 address space, 383 such as 192.0.2.0/24. However, if an internal hostname resolves to 384 an external IPv4 address, these can cause issues if the underlying 385 network is IPv6-only. As an example, let's assume that 386 "www.internal.example.com" has exactly one A record, 198.51.100.42, 387 and no AAAA records. The client will send the DNS query to the 388 company's recursive resolver and that resolver will reply with these 389 records. The device now only has an IPv4 address to connect to, and 390 no route to that address. Since the company's resolver does not know 391 the NAT64 prefix of the underlying network, it cannot synthesize the 392 address. Similarly, the underlying network's DNS64 recursive 393 resolver does not know the company's internal addresses, so it cannot 394 resolve the hostname. Because of this, the client device needs to 395 resolve the A record using the company's resolver then locally 396 synthesize an IPv6 address, as if the resolved IPv4 address were 397 provided by the application (Section 7.1). 399 8. Summary of Configurable Values 401 The values that may be configured as defaults on a client for use in 402 Happy Eyeballs are as follows: 404 o Resolution Delay (Section 3): The time to wait for a AAAA response 405 after receiving an A response. RECOMMENDED at 50 milliseconds. 407 o First Address Family Count (Section 4): The number of addresses 408 belonging to the first address family (such as IPv6) that should 409 be attempted before attempting another address family. 410 RECOMMENDED as 1, or 2 to more aggressively favor one address 411 family. 413 o Connection Attempt Delay (Section 5): The time to wait between 414 connection attempts in the absence of RTT data. RECOMMENDED at 415 250 milliseconds. 417 o Minimum Connection Attempt Delay (Section 5): The minimum time to 418 wait between connection attempts. RECOMMENDED at 100 419 milliseconds. MUST NOT be less than 10 milliseconds. 421 o Maximum Connection Attempt Delay (Section 5): The maximum time to 422 wait between connection attempts. RECOMMENDED at 2 seconds. 424 o Last Resort Local Synthesis Delay (Section 7.2): The time to wait 425 after starting the last IPv6 attempt and before sending the A 426 query. RECOMMENDED at 2 seconds. 428 The delay values described in this section were determined 429 empirically by measuring the timing of connections on a very wide set 430 of production devices. They were picked to reduce wait times noticed 431 by users while minimizing load on the network. As time passes, it is 432 expected that the properties of networks will evolve. For that 433 reason, it is expected that these values will change over time. 434 Implementors should feel welcome to use different values without 435 changing this specification. Since IPv6 issues are expected to be 436 less common, the delays SHOULD be increased with time as client 437 software is updated. 439 9. Limitations 441 Happy Eyeballs will handle initial connection failures at the TCP/IP 442 layer, however other failures or performance issues may still affect 443 the chosen connection. 445 9.1. Path Maximum Transmission Unit Discovery 447 Since Happy Eyeballs is only active during the initial handshake and 448 TCP does not pass the initial handshake, issues related to MTU can be 449 masked and go unnoticed during Happy Eyeballs. Solving this issue is 450 out of scope of this document. One solution is to use Packetization 451 Layer Path MTU Discovery [RFC4821]. 453 9.2. Application Layer 455 If the DNS returns multiple addresses for different application 456 servers, the application itself may not be operational and functional 457 on all of them. Common examples include Transport Layer Security 458 (TLS) and the Hypertext Transport Protocol (HTTP). 460 9.3. Hiding Operational Issues 462 It has been observed in practice that Happy Eyeballs can hide issues 463 in networks. For example, if a misconfiguration causes IPv6 to 464 consistently fail on a given network while IPv4 is still functional, 465 Happy Eyeballs may impair the operator's ability to notice the issue. 466 It is recommended that network operators deploy external means of 467 monitoring to ensure functionality of all address families. 469 10. Security Considerations 471 This memo has no direct security considerations. 473 11. IANA Considerations 475 This memo includes no request to IANA. 477 12. Acknowledgments 479 The authors thank Dan Wing, Andrew Yourtchenko, and everyone else who 480 worked on the original Happy Eyeballs design [RFC6555], Josh 481 Graessley, Stuart Cheshire, and the rest of team at Apple that helped 482 implement and instrument this algorithm, and Jason Fesler and Paul 483 Saab who helped measure and refine this algorithm. The authors would 484 also like to thank Fred Baker, Nick Chettle, Lorenzo Colitti, Igor 485 Gashinsky, Geoff Huston, Jen Linkova, Paul Hoffman, Philip Homburg, 486 Warren Kumari, Erik Nygren, Jordi Palet Martinez, Rui Paulo, Stephen 487 Strowes, Jinmei Tatuya, Dave Thaler, Joe Touch and James Woodyatt for 488 their input and contributions. 490 13. References 492 13.1. Normative References 494 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 495 Requirement Levels", BCP 14, RFC 2119, 496 DOI 10.17487/RFC2119, March 1997, 497 . 499 [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU 500 Discovery", RFC 4821, DOI 10.17487/RFC4821, March 2007, 501 . 503 [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. 504 Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, 505 DOI 10.17487/RFC6052, October 2010, 506 . 508 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 509 NAT64: Network Address and Protocol Translation from IPv6 510 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 511 April 2011, . 513 [RFC6147] Bagnulo, M., Sullivan, A., Matthews, P., and I. van 514 Beijnum, "DNS64: DNS Extensions for Network Address 515 Translation from IPv6 Clients to IPv4 Servers", RFC 6147, 516 DOI 10.17487/RFC6147, April 2011, 517 . 519 [RFC6535] Huang, B., Deng, H., and T. Savolainen, "Dual-Stack Hosts 520 Using "Bump-in-the-Host" (BIH)", RFC 6535, 521 DOI 10.17487/RFC6535, February 2012, 522 . 524 [RFC6555] Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with 525 Dual-Stack Hosts", RFC 6555, DOI 10.17487/RFC6555, April 526 2012, . 528 [RFC6724] Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown, 529 "Default Address Selection for Internet Protocol Version 6 530 (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012, 531 . 533 [RFC7050] Savolainen, T., Korhonen, J., and D. Wing, "Discovery of 534 the IPv6 Prefix Used for IPv6 Address Synthesis", 535 RFC 7050, DOI 10.17487/RFC7050, November 2013, 536 . 538 13.2. Informative References 540 [DNS-PUSH] 541 Pusateri, T. and S. Cheshire, "DNS Push Notifications", 542 Work in Progress, draft-ietf-dnssd-push, March 2017. 544 [RFC6877] Mawatari, M., Kawashima, M., and C. Byrne, "464XLAT: 545 Combination of Stateful and Stateless Translation", 546 RFC 6877, DOI 10.17487/RFC6877, April 2013, 547 . 549 [RFC7413] Cheng, Y., Chu, J., Radhakrishnan, S., and A. Jain, "TCP 550 Fast Open", RFC 7413, DOI 10.17487/RFC7413, December 2014, 551 . 553 Appendix A. Differences from RFC6555 555 "Happy Eyeballs: Success with Dual-Stack Hosts" [RFC6555] mostly 556 concentrates on how to stagger connections to a hostname that has an 557 AAAA and an A record. This document additionally discusses: 559 o how to perform DNS queries to obtain these addresses 561 o how to handle multiple addresses from each address family 563 o how to handle DNS updates while connections are being raced 565 o how to leverage historical information 567 o how to support IPv6-only networks with NAT64 and DNS64 569 Authors' Addresses 571 David Schinazi 572 Apple Inc. 573 1 Infinite Loop 574 Cupertino, California 95014 575 US 577 Email: dschinazi@apple.com 579 Tommy Pauly 580 Apple Inc. 581 1 Infinite Loop 582 Cupertino, California 95014 583 US 585 Email: tpauly@apple.com