idnits 2.17.1 draft-ietf-v6ops-rfc6555bis-01.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 (June 5, 2017) is 2515 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 119, but not defined == Missing Reference: 'Section 4' is mentioned on line 156, but not defined == Missing Reference: 'Section 5' is mentioned on line 193, but not defined == Missing Reference: 'Section 6' is mentioned on line 161, 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 June 5, 2017 6 Expires: December 7, 2017 8 Happy Eyeballs Version 2: Better Connectivity Using Concurrency 9 draft-ietf-v6ops-rfc6555bis-01 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 http://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 December 7, 2017. 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 (http://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 . . . . . . . . . 4 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 . . . . . . . . . . . 8 68 7.3. Virtual Private Networks . . . . . . . . . . . . . . . . 9 69 8. Summary of Configurable Values . . . . . . . . . . . . . . . 10 70 9. Limitations . . . . . . . . . . . . . . . . . . . . . . . . . 10 71 9.1. Path Maximum Transmission Unit Discovery . . . . . . . . 10 72 9.2. Application Layer . . . . . . . . . . . . . . . . . . . . 11 73 9.3. Hiding Operational Issues . . . . . . . . . . . . . . . . 11 74 10. Security Considerations . . . . . . . . . . . . . . . . . . . 11 75 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11 76 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 11 77 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 11 78 13.1. Normative References . . . . . . . . . . . . . . . . . . 11 79 13.2. Informative References . . . . . . . . . . . . . . . . . 12 80 Appendix A. Differences from RFC6555 . . . . . . . . . . . . . . 13 81 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 13 83 1. Introduction 85 Many communication protocols operated over the modern Internet use 86 host names. These often correspond to multiple IP addresses, whose 87 performance can vary. Since specific addresses or address families 88 (IPv4 or IPv6) may be blocked, broken, or sub-optimal on a network, 89 clients that attempt multiple connections in parallel have a higher 90 chance of establishing a connection sooner. This document specifies 91 requirements for algorithms that reduce this user-visible delay and 92 provides an algorithm. 94 This documents expands on "Happy Eyeballs" [RFC6555], a technique of 95 reducing user-visible delays on dual-stack hosts. Now that this 96 approach has been deployed at scale and measured for several years, 97 the algorithm specification can be refined to improve its reliability 98 and generalization. This document recommends an algorithm of racing 99 resolved addresses that has several stages of ordering and racing to 100 avoid delays to the user whenever possible, while preferring the use 101 of IPv6. Specifically, it discusses how to handle DNS queries when 102 starting a connection on a dual-stack client, how to create an 103 ordered list of addresses to which to attempt connections, and how to 104 race the connection attempts. 106 1.1. Requirements Language 108 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 109 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 110 "OPTIONAL" in this document are to be interpreted as described in 111 "Key words for use in RFCs to Indicate Requirement Levels" [RFC2119]. 113 2. Overview 115 This document defines a method of connection establishment, named 116 "Happy Eyeballs Connection Setup". This approach has several 117 distinct phases: 119 1. Initiation of asynchronous DNS queries [Section 3] 121 2. Sorting of resolved addresses [Section 4] 123 3. Initiation of asynchronous connection attempts [Section 5] 125 4. Establishment of one connection, which cancels all other attempts 127 Note that this document assumes that the host address preference 128 policy favors IPv6 over IPv4. If the host is configured differently, 129 the recommendations in this document can be easily adapted. 131 3. Hostname Resolution Query Handling 133 When a client has both IPv4 and IPv6 connectivity, and is trying to 134 establish a connection with a named host, it needs to send out both 135 AAAA and A DNS queries. Both queries SHOULD be made as soon after 136 one another as possible, with the AAAA query made first, immediately 137 followed by the A query. 139 Implementations MUST NOT wait for both families of answers to return 140 before attempting connection establishment. If one query fails to 141 return, or takes significantly longer to return, waiting for the 142 second address family can significantly delay the connection 143 establishment of the first one. Therefore, the client MUST treat DNS 144 resolution as asynchronous. Note that if the platform does not offer 145 an asynchronous DNS API, this behavior can be simulated by making two 146 separate synchronous queries on different threads, one per address 147 family. If the AAAA query returns first, the first IPv6 connection 148 attempt MUST be immediately started. If the A query returns first, 149 the client SHOULD wait for a short time for the AAAA response. This 150 delay will be referred to as the "Resolution Delay". The RECOMMENDED 151 value for the Resolution Delay is 50 milliseconds. If the AAAA 152 response is received within the Resolution Delay period, the client 153 MUST immediately start the IPv6 connection attempt. If, at the end 154 of the Resolution Delay period, the AAAA response has not been 155 received but the A response has been received, the client SHOULD 156 proceed to Sorting Addresses [Section 4] and staggered connection 157 attempts [Section 5] using only the IPv4 addresses returned so far. 158 If the AAAA response arrives while these connection attempts are in 159 progress, but before any connection has been established, then the 160 newly received IPv6 addresses are incorporated into the list of 161 available candidate addresses [Section 6] and the process of 162 connection attempts will continue with the IPv6 addresses added, 163 until one connection is established. 165 3.1. Handling Multiple DNS Server Addresses 167 If multiple DNS server addresses are configured for the current 168 network, the client may have the option of sending its DNS queries 169 over IPv4 or IPv6. In keeping with the Happy Eyeballs approach, 170 queries SHOULD be sent over IPv6 first (note that this is not 171 referring to the sending of AAAA or A queries, but rather the address 172 of the DNS server itself). If DNS queries sent to the IPv6 address 173 do not receive responses, that address may be marked as penalized, 174 and queries can be sent to other DNS server addresses. 176 As native IPv6 deployments become more prevalent, and IPv4 addresses 177 are exhausted, it is expected that IPv6 connectivity will have 178 preferential treatment within networks. If a DNS server is 179 configured to be accessible over IPv6, IPv6 should be assumed to be 180 the preferred address family. 182 Client systems SHOULD NOT have an explicit limit to the number of DNS 183 servers that can be configured, either manually or by the network. 184 If such a limit is required by hardware limitations, it is 185 RECOMMENDED to use at least one address from each address family from 186 the available list. 188 4. Sorting Addresses 190 Before attempting to connect to any of the resolved addresses, the 191 client should define the order in which to start the attempts. Once 192 the order has been defined, the client can use a simple algorithm for 193 racing each option after a short delay [Section 5]. It is important 194 that the ordered list involves all addresses from both families, as 195 this allows the client to get the racing effect of Happy Eyeballs for 196 the entire list, not just the first IPv4 and first IPv6 addresses. 198 First, the client MUST sort the addresses using Destination Address 199 Selection ([RFC6724], Section 6). 201 If the client is stateful and has history of expected round-trip 202 times (RTT) for the routes to access each address, it SHOULD add a 203 Destination Address Selection rule between rules 8 and 9 that prefers 204 addresses with lower RTTs. If the client keeps track of which 205 addresses it has used in the past, it SHOULD add another destination 206 address selection rule between the RTT rule and rule 9, which prefers 207 used addresses over unused ones. This helps servers that use the 208 client's IP address during authentication, as is the case for TCP 209 Fast Open ([RFC7413]) and some HTTP cookies. This historical data 210 MUST NOT be used across networks, and SHOULD be flushed on network 211 changes. 213 Next, the client SHOULD modify the ordered list to interleave address 214 families. Whichever address family is first in the list should be 215 followed by an address of the other address family; that is, if the 216 first address in the sorted list is IPv6, then the first IPv4 address 217 should be moved up in the list to be second in the list. An 218 implementation MAY want to favor one address family more by allowing 219 multiple addresses of that family to be attempted before trying the 220 other family. The number of contiguous addresses of the first 221 address family will be referred to as the "First Address Family 222 Count", and can be a configurable value. This is performed to avoid 223 waiting through a long list of addresses from a given address family 224 if connectivity over that address family is impaired. 226 5. Connection Attempts 228 Once the list of addresses has been constructed, the client will 229 attempt to make connections. In order to avoid unreasonable network 230 load, connection attempts SHOULD NOT be made simultaneously. 231 Instead, one connection attempt to a single address is started first, 232 followed by the others in the list, one at a time. Starting a new 233 connection attempt does not affect previous attempts, as multiple 234 connection attempts may occur in parallel. Once one of the 235 connection attempts succeeds (generally when the TCP handshake 236 completes), all other connections attempts that have not yet 237 succeeded SHOULD be cancelled. Any address that was not yet 238 attempted as a connection SHOULD be ignored. 240 A simple implementation can have a fixed delay for how long to wait 241 before starting the next connection attempt. This delay is referred 242 to as the "Connection Attempt Delay". One recommended value for this 243 delay is 250 milliseconds. If the client has historical RTT data, it 244 can also use the expected RTT to choose a more nuanced delay value. 245 The recommended formula for calculating the delay after starting a 246 connection attempt is: MAX( 1.25 * RTT_MEAN + 4 * RTT_VARIANCE, 2 * 247 RTT_MEAN ), where the RTT values are based on the statistics for 248 previous address used. If the TCP implementation leverages 249 historical RTT data to compute SYN timeout, these algorithms should 250 match so that a new attempt will be started at the same time as the 251 previous is sending its second TCP SYN. While TCP implementations 252 often leverage an exponential backoff when they detect packet loss, 253 the "Connection Attempt Delay" SHOULD NOT such an aggressive backoff, 254 as it would harm user experience. 256 The Connection Attempt Delay MUST have a lower bound, especially if 257 it is computed using historical data. More specifically, a 258 subsequent connection MUST NOT be started within 10 milliseconds of 259 the previous attempt. The recommended minimum value is 100 260 milliseconds, which is referred to as the "Minimum Connection Attempt 261 Delay". This minimum value is required to avoid congestion collapse 262 in the presence of high packet loss rates. The Connection Attempt 263 Delay SHOULD have an upper bound, referred to as the "Maximum 264 Connection Attempt Delay". The current recommended value is 2 265 seconds. 267 6. DNS Answer Changes during Happy Eyeballs Connection Setup 269 If, during the course of connection establishment, the DNS answers 270 change either by adding resolved addresses (for example, due to DNS 271 push notifications [DNS-PUSH]), or removing previously resolved 272 addresses (for example, due to expiry of the TTL on that DNS record), 273 the client should react based on its current progress. 275 If an address is removed from the list that already had a connection 276 attempt started, the connection attempt SHOULD NOT be cancelled, but 277 rather be allowed to continue. If the removed address had not yet 278 had a connection attempt started, it SHOULD be removed from the list 279 of addresses to try. 281 If an address is added to the list, it should be sorted into the list 282 of addresses not yet attempted according to the rules above 283 (Section 4). 285 7. Supporting IPv6-only Networks with NAT64 and DNS64 287 While many IPv6 transition protocols have been standardized and 288 deployed, most are transparent to client devices. The combined use 289 of NAT64 [RFC6146] and DNS64 [RFC6147] is a popular solution that is 290 being deployed and requires changes in client devices. One possible 291 way to handle these networks is for the client device networking 292 stack to implement 464XLAT [RFC6877]. 464XLAT has the advantage of 293 not requiring changes to user space software, however it requires 294 per-packet translation and does not encourage client application 295 software to support native IPv6. On platforms that do not support 296 464XLAT, the Happy Eyeballs engine SHOULD follow the recommendations 297 in this section to properly support IPv6-only networks with NAT64 and 298 DNS64. 300 The features described in this section SHOULD only be enabled when 301 the host detects one of these networks. A simple heuristic to 302 achieve that is to check if the networks offers routable IPv6 303 addressing, does not offer routable IPv4 addressing, and offers a DNS 304 resolver address. 306 7.1. IPv4 Address Literals 308 If client applications or users wish to connect to IPv4 address 309 literals, the Happy Eyeballs engine will need to perform NAT64 310 address synthesis for them. The solution is similar to "Bump-in-the- 311 Host" [RFC6535] but is implemented inside the Happy Eyeballs library. 313 When an IPv4 address is passed in to the library instead of a host 314 name, the device queries the network for the NAT64 prefix using 315 "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" 316 [RFC7050] then synthesizes an appropriate IPv6 address (or several) 317 using the encoding described in "IPv6 Addressing of IPv4/IPv6 318 Translators" [RFC6052]. The synthesized addresses are then inserted 319 into the list of addresses as if they were results from DNS queries; 320 connection attempts follow the algorithm described above (Section 5). 322 7.2. Host Names with Broken AAAA Records 324 At the time of writing, there exist a small but non negligible number 325 of host names that resolve to valid A records and broken AAAA 326 records, which we define as AAAA records that contain seemingly valid 327 IPv6 addresses but those addresses never reply when contacted on the 328 usual ports. These can be for example caused by: 330 o Mistyping of the IPv6 address in the DNS zone configuration 332 o Routing black holes 334 o Service outages 336 While an algorithm complying with the other sections of this document 337 would correctly handle such host names on a dual-stack network, they 338 will not necessarily function correctly on IPv6-only networks with 339 NAT64 and DNS64. Since DNS64 recursive resolvers rely on the 340 authoritative name servers sending negative ("no error no answer") 341 responses for AAAA records in order to synthesize, they will not 342 synthesize records for these particular host names, and will instead 343 pass through the broken AAAA record. 345 In order to support these scenarios, the client device needs to query 346 the DNS for the A record then perform local synthesis. Since these 347 types of host names are rare, and in order to minimize load on DNS 348 servers, this A query should only be performed when the client has 349 given up on the AAAA records it initially received. This can be 350 achieved by using a longer timeout, referred to as the "Last Resort 351 Local Synthesis Delay" and RECOMMENDED at 2 seconds. The timer is 352 started when the last connection attempt is fired. If no connection 353 attempt has succeeded when this timer fires, the device queries the 354 DNS for the IPv4 address and on reception of a valid A record, treats 355 it as if it were provided by the application (Section 7.1). 357 7.3. Virtual Private Networks 359 Some Virtual Private Networks (VPN) may be configured to handle DNS 360 queries from the device. The configuration could encompass all 361 queries, or a subset such as "*.internal.example.com". These VPNs 362 can also be configured to only route part of the IPv4 address space, 363 such as 192.0.2.0/24. However, if an internal hostname resolves to 364 an external IPv4 address, these can cause issues if the underlying 365 network is IPv6-only. As an example, let's assume that 366 "www.internal.example.com" has exactly one A record, 198.51.100.42, 367 and no AAAA records. The client will send the DNS query to the 368 company's recursive resolver and that resolver will reply with these 369 records. The device now only has an IPv4 address to connect to, and 370 no route to that address. Since the company's resolver does not know 371 the NAT64 prefix of the underlying network, it cannot synthesize the 372 address. Similarly, the underlying network's DNS64 recursive 373 resolver does not know the company's internal addresses, so it cannot 374 resolve the hostname. Because of this, the client device needs to 375 resolve the A record using the company's resolver then locally 376 synthesize an IPv6 address, as if the resolved IPv4 address were 377 provided by the application (Section 7.1). 379 8. Summary of Configurable Values 381 The values that may be configured as defaults on a client for use in 382 Happy Eyeballs are as follows: 384 o Resolution Delay (Section 3): The time to wait for a AAAA response 385 after receiving an A response. RECOMMENDED at 50 milliseconds. 387 o First Address Family Count (Section 4): The number of addresses 388 belonging to the first address family (such as IPv6) that should 389 be attempted before attempting another address family. 390 RECOMMENDED as 1, or 2 to more aggressively favor one address 391 family. 393 o Connection Attempt Delay (Section 5): The time to wait between 394 connection attempts in the absence of RTT data. RECOMMENDED at 395 250 milliseconds. 397 o Minimum Connection Attempt Delay (Section 5): The minimum time to 398 wait between connection attempts. RECOMMENDED at 100 399 milliseconds. MUST NOT be less than 10 milliseconds. 401 o Maximum Connection Attempt Delay (Section 5): The maximum time to 402 wait between connection attempts. RECOMMENDED at 2 seconds. 404 o Last Resort Local Synthesis Delay (Section 7.2): The time to wait 405 after starting the last IPv6 attempt and before sending the A 406 query. RECOMMENDED at 2 seconds. 408 As time advances, it is expected that the properties of networks will 409 evolve. For that reason, it is expected that these values will 410 change over time. Implementors should feel welcome to use different 411 values without changing this specification. In particular, IPv6 412 issues are expected to be less common, therefore the Resolution Delay 413 SHOULD be increased with time as client software is updated. 415 9. Limitations 417 Happy Eyeballs will handle initial connection failures at the TCP/IP 418 layer, however other failures or performance issues may still affect 419 the chosen connection. 421 9.1. Path Maximum Transmission Unit Discovery 423 Since Happy Eyeballs is only active during the initial handshake and 424 TCP does not pass the initial handshake, issues related to MTU can be 425 masked and go unnoticed during Happy Eyeballs. Solving this issue is 426 out of scope of this document. 428 9.2. Application Layer 430 If the DNS returns multiple addresses for different application 431 servers, the application itself may not be operational and functional 432 on all of them. Common examples include Transport Layer Security 433 (TLS) and the Hypertext Transport Protocol (HTTP). 435 9.3. Hiding Operational Issues 437 It has been observed in practice that Happy Eyeballs can hide issues 438 in networks. For example, if a misconfiguration causes IPv6 to 439 consistently fail on a given network while IPv4 is still functional, 440 Happy Eyeballs may impair the operator's ability to notice the issue. 441 It is recommended that network operators deploy external means of 442 monitoring to ensure functionality of all address families. 444 10. Security Considerations 446 This memo has no direct security considerations. 448 11. IANA Considerations 450 This memo includes no request to IANA. 452 12. Acknowledgments 454 The authors thank Dan Wing, Andrew Yourtchenko, and everyone else who 455 worked on the original Happy Eyeballs design ([RFC6555]), Josh 456 Graessley, Stuart Cheshire, and the rest of team at Apple that helped 457 implement and instrument this algorithm, and Jason Fesler and Paul 458 Saab who helped measure and refine this algorithm. The authors would 459 also like to thank Fred Baker, Nick Chettle, Lorenzo Colitti, Igor 460 Gashinsky, Geoff Huston, Jen Linkova, Paul Hoffman, Philip Homburg, 461 Erik Nygren, Jordi Palet Martinez, Rui Paulo, Jinmei Tatuya, Dave 462 Thaler, Joe Touch and James Woodyatt for their input and 463 contributions. 465 13. References 467 13.1. Normative References 469 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 470 Requirement Levels", BCP 14, RFC 2119, 471 DOI 10.17487/RFC2119, March 1997, 472 . 474 [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. 475 Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, 476 DOI 10.17487/RFC6052, October 2010, 477 . 479 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 480 NAT64: Network Address and Protocol Translation from IPv6 481 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 482 April 2011, . 484 [RFC6147] Bagnulo, M., Sullivan, A., Matthews, P., and I. van 485 Beijnum, "DNS64: DNS Extensions for Network Address 486 Translation from IPv6 Clients to IPv4 Servers", RFC 6147, 487 DOI 10.17487/RFC6147, April 2011, 488 . 490 [RFC6535] Huang, B., Deng, H., and T. Savolainen, "Dual-Stack Hosts 491 Using "Bump-in-the-Host" (BIH)", RFC 6535, 492 DOI 10.17487/RFC6535, February 2012, 493 . 495 [RFC6555] Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with 496 Dual-Stack Hosts", RFC 6555, DOI 10.17487/RFC6555, April 497 2012, . 499 [RFC6724] Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown, 500 "Default Address Selection for Internet Protocol Version 6 501 (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012, 502 . 504 [RFC7050] Savolainen, T., Korhonen, J., and D. Wing, "Discovery of 505 the IPv6 Prefix Used for IPv6 Address Synthesis", 506 RFC 7050, DOI 10.17487/RFC7050, November 2013, 507 . 509 13.2. Informative References 511 [DNS-PUSH] 512 Pusateri, T. and S. Cheshire, "DNS Push Notifications", 513 Work in Progress, draft-ietf-dnssd-push, March 2017. 515 [RFC6877] Mawatari, M., Kawashima, M., and C. Byrne, "464XLAT: 516 Combination of Stateful and Stateless Translation", 517 RFC 6877, DOI 10.17487/RFC6877, April 2013, 518 . 520 [RFC7413] Cheng, Y., Chu, J., Radhakrishnan, S., and A. Jain, "TCP 521 Fast Open", RFC 7413, DOI 10.17487/RFC7413, December 2014, 522 . 524 Appendix A. Differences from RFC6555 526 "Happy Eyeballs: Success with Dual-Stack Hosts" [RFC6555] mostly 527 concentrates on how to stagger connections to a hostname that has an 528 AAAA and an A record. This document additionally discusses: 530 o how to perform DNS queries to obtain these addresses 532 o how to handle multiple addresses from each address family 534 o how to handle DNS updates while connections are being raced 536 o how to leverage historical information 538 o how to support IPv6-only networks with NAT64 and DNS64 540 Authors' Addresses 542 David Schinazi 543 Apple Inc. 544 1 Infinite Loop 545 Cupertino, California 95014 546 US 548 Email: dschinazi@apple.com 550 Tommy Pauly 551 Apple Inc. 552 1 Infinite Loop 553 Cupertino, California 95014 554 US 556 Email: tpauly@apple.com