idnits 2.17.1 draft-ietf-v6ops-nat64-deployment-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 : ---------------------------------------------------------------------------- == There are 2 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 1 instance of lines with non-RFC3849-compliant IPv6 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (April 19, 2019) is 1824 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-07) exists of draft-huitema-quic-dnsoquic-06 == Outdated reference: A later version (-09) exists of draft-ietf-6man-ra-pref64-00 == Outdated reference: A later version (-02) exists of draft-li-intarea-nat64-prefix-dhcp-option-01 == Outdated reference: A later version (-06) exists of draft-lmhp-v6ops-transition-comparison-02 == Outdated reference: A later version (-04) exists of draft-palet-v6ops-464xlat-opt-cdn-caches-01 Summary: 0 errors (**), 0 flaws (~~), 8 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 v6ops J. Palet Martinez 3 Internet-Draft The IPv6 Company 4 Intended status: Informational April 19, 2019 5 Expires: October 21, 2019 7 Additional NAT64/464XLAT Deployment Guidelines in Operator and 8 Enterprise Networks 9 draft-ietf-v6ops-nat64-deployment-05 11 Abstract 13 This document describes how NAT64 (including 464XLAT) can be deployed 14 in an IPv6 network, whether cellular ISP, broadband ISP, or 15 enterprise, and possible optimizations. The document also discusses 16 issues to be considered when having IPv6-only connectivity, 17 regarding: a) DNS64, b) applications or devices that use literal IPv4 18 addresses or non-IPv6 compliant APIs, and c) IPv4-only hosts or 19 applications. 21 Status of This Memo 23 This Internet-Draft is submitted in full conformance with the 24 provisions of BCP 78 and BCP 79. 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF). Note that other groups may also distribute 28 working documents as Internet-Drafts. The list of current Internet- 29 Drafts is at https://datatracker.ietf.org/drafts/current/. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 This Internet-Draft will expire on October 21, 2019. 38 Copyright Notice 40 Copyright (c) 2019 IETF Trust and the persons identified as the 41 document authors. All rights reserved. 43 This document is subject to BCP 78 and the IETF Trust's Legal 44 Provisions Relating to IETF Documents 45 (https://trustee.ietf.org/license-info) in effect on the date of 46 publication of this document. Please review these documents 47 carefully, as they describe your rights and restrictions with respect 48 to this document. Code Components extracted from this document must 49 include Simplified BSD License text as described in Section 4.e of 50 the Trust Legal Provisions and are provided without warranty as 51 described in the Simplified BSD License. 53 Table of Contents 55 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 56 2. Requirements Language . . . . . . . . . . . . . . . . . . . . 5 57 3. NAT64 Deployment Scenarios . . . . . . . . . . . . . . . . . 5 58 3.1. Known to Work . . . . . . . . . . . . . . . . . . . . . . 6 59 3.1.1. Service Provider NAT64 with DNS64 . . . . . . . . . . 6 60 3.1.2. Service Provider Offering 464XLAT, with DNS64 . . . . 8 61 3.1.3. Service Provider Offering 464XLAT, without DNS64 . . 11 62 3.2. Known to Work Under Special Conditions . . . . . . . . . 14 63 3.2.1. Service Provider NAT64 without DNS64 . . . . . . . . 14 64 3.2.2. Service Provider NAT64; DNS64 in the IPv6 hosts . . . 15 65 3.2.3. Service Provider NAT64; DNS64 in the IPv4-only 66 remote network . . . . . . . . . . . . . . . . . . . 16 67 3.3. Comparing the Scenarios . . . . . . . . . . . . . . . . . 16 68 4. Issues to be Considered . . . . . . . . . . . . . . . . . . . 18 69 4.1. DNSSEC Considerations and Possible Approaches . . . . . . 18 70 4.1.1. Not using DNS64 . . . . . . . . . . . . . . . . . . . 20 71 4.1.2. DNSSEC validator aware of DNS64 . . . . . . . . . . . 21 72 4.1.3. Stub validator . . . . . . . . . . . . . . . . . . . 21 73 4.1.4. CLAT with DNS proxy and validator . . . . . . . . . . 21 74 4.1.5. ACL of clients . . . . . . . . . . . . . . . . . . . 22 75 4.1.6. Mapping-out IPv4 addresses . . . . . . . . . . . . . 22 76 4.2. DNS64 and Reverse Mapping . . . . . . . . . . . . . . . . 22 77 4.3. Using 464XLAT with/without DNS64 . . . . . . . . . . . . 22 78 4.4. Foreign DNS . . . . . . . . . . . . . . . . . . . . . . . 23 79 4.4.1. Manual Configuration of Foreign DNS . . . . . . . . . 24 80 4.4.2. DNS Privacy . . . . . . . . . . . . . . . . . . . . . 24 81 4.4.3. Split DNS . . . . . . . . . . . . . . . . . . . . . . 25 82 4.5. Well-Known Prefix (WKP) vs Network-Specific Prefix (NSP) 25 83 4.6. IPv4 literals and old APIs . . . . . . . . . . . . . . . 25 84 4.7. IPv4-only Hosts or Applications . . . . . . . . . . . . . 26 85 4.8. CLAT Translation Considerations . . . . . . . . . . . . . 26 86 4.9. EAMT Considerations . . . . . . . . . . . . . . . . . . . 27 87 5. Summary of Deployment Recommendations for NAT64/464XLAT . . . 27 88 6. Deployment of NAT64 in Enterprise Networks . . . . . . . . . 30 89 7. Security Considerations . . . . . . . . . . . . . . . . . . . 31 90 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 32 91 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 32 92 10. ANNEX A: Example of Broadband Deployment with 464XLAT . . . . 32 93 11. ANNEX B: CLAT Implementation . . . . . . . . . . . . . . . . 36 94 12. ANNEX C: Benchmarking . . . . . . . . . . . . . . . . . . . . 36 95 13. ANNEX D: Changes from -00 to -01/-02 . . . . . . . . . . . . 36 96 14. ANNEX E: Changes from -02 to -03 . . . . . . . . . . . . . . 37 97 15. ANNEX F: Changes from -03 to -04 . . . . . . . . . . . . . . 37 98 16. ANNEX G: Changes from -04 to -05 . . . . . . . . . . . . . . 37 99 17. References . . . . . . . . . . . . . . . . . . . . . . . . . 37 100 17.1. Normative References . . . . . . . . . . . . . . . . . . 37 101 17.2. Informative References . . . . . . . . . . . . . . . . . 39 102 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 42 104 1. Introduction 106 Stateful NAT64 ([RFC6146]) describes a stateful IPv6 to IPv4 107 translation mechanism, which allows IPv6-only hosts to communicate 108 with IPv4-only servers using unicast UDP, TCP, or ICMP, by means of 109 IPv4 public addresses sharing, among multiple IPv6-only hosts. 110 Unless otherwise stated, references in the rest of this document to 111 NAT64 (function) should be interpreted as to Stateful NAT64. 113 The translation of the packet headers is done using the IP/ICMP 114 translation algorithm defined in [RFC7915] and algorithmically 115 translating the IPv4 addresses to IPv6 addresses and vice versa, 116 following [RFC6052]. 118 DNS64 ([RFC6147]) is in charge of the synthesis of AAAA records from 119 the A records, so only works for applications making use of DNS. It 120 was designed to avoid changes in both, the IPv6-only hosts and the 121 IPv4-only server, so they can use a NAT64 function. As discussed in 122 Section 5.5 of [RFC6147], a security-aware and validating host has to 123 perform the DNS64 function locally. 125 However, the use of NAT64 and/or DNS64 present three drawbacks: 127 a. Because DNS64 ([RFC6147]) modifies DNS answers, and DNSSEC is 128 designed to detect such modifications, DNS64 ([RFC6147]) may 129 potentially break DNSSEC, depending on a number of factors, such 130 as the location of the DNS64 function (at a DNS server or 131 validator, at the end host, ...), how it has been configured, if 132 the end-hosts is validating, etc. 134 b. Because the need of using DNS64 ([RFC6147]) or an alternative 135 "host/application built-in" mechanism for address synthesis, 136 there may be an issue for NAT64 ([RFC6146]), as it doesn't work 137 when IPv4 literal addresses or non-IPv6 compliant APIs are being 138 used. 140 c. NAT64 alone, was not designed to provide a solution for IPv4-only 141 hosts or applications located within a network which are 142 connected to a service provider IPv6-only access, as it was 143 designed for a very specific scenario ([RFC6144], Section 2.1). 145 Above drawbacks may be true if part of, an enterprise network, is 146 connected to other parts of the same network or third-party networks 147 by means of IPv6-only connectivity. This is just an example which 148 may apply to many other similar cases. All them are deployment 149 specific. 151 According to that, across this document, the use of "operator", 152 "operator network", "service provider", and similar ones, are 153 interchangeable with equivalent cases of enterprise networks (and 154 similar ones). This may be also the case for "managed end-user 155 networks". 157 An analysis of stateful IPv4/IPv6 mechanisms is provided in 158 [RFC6889]. 160 This document looks into different possible NAT64 ([RFC6146]) 161 deployment scenarios, including IPv4-IPv6-IPv4 (464 for short) and 162 similar ones, which were not documented in [RFC6144], such as 464XLAT 163 ([RFC6877]), in operator (broadband and cellular) and enterprise 164 networks, and provides guidelines to avoid operational issues. 166 Towards that, this document first looks into the possible NAT64 167 deployment scenarios (split in "known to work" and "known to work 168 under special conditions"), providing a quick and generic comparison 169 table among them. Then the document describes the issues that an 170 operator need to understand on different matters that will allow to 171 define what is the best approach/scenario for each specific network 172 case. A summary provides some recommendations and decision points. 173 A section with clarifications on the usage of this document for 174 enterprise networks, is also provided. Finally, an annex provides an 175 example of a broadband deployment using 464XLAT and another annex 176 provides hints for a CLAT implementation. 178 [RFC7269] already provides information about NAT64 deployment options 179 and experiences. Both, this document and [RFC7269] are 180 complementary, as they are looking into different deployment 181 considerations and furthermore, this document is considering the 182 updated deployment experience and newer standards. 184 The target deployment scenarios in this document may be covered as 185 well by other IPv4-as-a-Service (IPv4aaS) transition mechanisms. 186 Note that this is true only for the case of broadband networks, as in 187 the case of cellular networks the only supported solution is the use 188 of NAT64/464XLAT. So, it is out of scope of this document to provide 189 a comparison among the different IPv4aaS transition mechanisms, which 190 is being analyzed already in [I-D.lmhp-v6ops-transition-comparison]. 192 Consequently, this document should not be understood as a guide for 193 an operator or enterprise to decide which IPv4aaS is the best one for 194 its own network. Instead it should be used as a tool for 195 understanding all the implications, including relevant documents (or 196 even specific parts of them), for the deployment of NAT64/464XLAT and 197 facilitate the decision process regarding specific deployment 198 details. 200 2. Requirements Language 202 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 203 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 204 "OPTIONAL" in this document are to be interpreted as described in BCP 205 14 [RFC2119] [RFC8174] when, and only when, they appear in all 206 capitals, as shown here. 208 3. NAT64 Deployment Scenarios 210 Section 7 of DNS64 ([RFC6147]), provides three scenarios, depending 211 on the location of the DNS64 function. However, since the 212 publication of that document, other deployment scenarios and NAT64 213 use cases need to be considered in actual networks, despite some of 214 them were specifically ruled out by the original NAT64/DNS64 work. 216 Consequently, the perspective in this document is to broaden those 217 scenarios, including a few new ones. However, in order to be able to 218 reduce the number of possible cases, we work under the assumption 219 that typically, the service provider wants to make sure that all the 220 customers have a service without failures. This means considering 221 the following assumptions for the worst possible case: 223 a. There are hosts that will be validating DNSSEC. 225 b. IPv4 literal addresses and non-IPv6 compliant APIs are being 226 used. 228 c. There are IPv4-only hosts or applications beyond the IPv6-only 229 link (e.g., tethering in cellular networks). 231 The document uses a common set of possible "participant entities": 233 1. An IPv6-only access network (IPv6). 235 2. An IPv4-only remote network/server/service (IPv4). 237 3. A NAT64 function (NAT64) in the service provider. 239 4. A DNS64 function (DNS64) in the service provider. 241 5. An external service provider offering the NAT64 function and/or 242 the DNS64 function (extNAT64/extDNS64). 244 6. 464XLAT customer side translator (CLAT). 246 Note that the nomenclature used in parenthesis is the one that, for 247 short, will be used in the figures. 249 The possible scenarios are split in two general categories: 251 1. Known to work. 253 2. Known to work under special conditions. 255 3.1. Known to Work 257 The scenarios in this category are known to work. Each one may have 258 different pros and cons, and in some cases the trade-offs, maybe 259 acceptable for some operators. 261 3.1.1. Service Provider NAT64 with DNS64 263 In this scenario, the service provider offers both, the NAT64 and the 264 DNS64 functions. 266 This is the most common scenario as originally considered by the 267 designers of NAT64 ([RFC6146]) and DNS64 ([RFC6147]), however also 268 may have the implications related the DNSSEC. 270 This scenario also may fail to solve the issue of IPv4 literal 271 addresses or non-IPv6 compliant APIs, as well as the issue of 272 IPv4-only hosts or applications behind the IPv6-only access network. 274 +----------+ +----------+ +----------+ 275 | | | NAT64 | | | 276 | IPv6 +--------+ + +--------+ IPv4 | 277 | | | DNS64 | | | 278 +----------+ +----------+ +----------+ 280 Figure 1: NAT64 with DNS64 282 A similar scenario will be if the service provider offers only the 283 DNS64 function, and the NAT64 function is provided by an outsourcing 284 agreement with an external provider. All the considerations in the 285 previous paragraphs of this section are the same for this sub-case. 287 +----------+ +----------+ 288 | | | | 289 | extNAT64 +--------+ IPv4 | 290 | | | | 291 +----+-----+ +----------+ 292 | 293 | 294 +----------+ +----+-----+ 295 | | | | 296 | IPv6 +--------+ DNS64 + 297 | | | | 298 +----------+ +----------+ 300 Figure 2: NAT64 in external service provider 302 As well, is equivalent to the scenario where the outsourcing 303 agreement with the external provider is to provide both the NAT64 and 304 DNS64 functions. Once more, all the considerations in the previous 305 paragraphs of this section are the same for this sub-case. 307 +----------+ +----------+ 308 | extNAT64 | | | 309 | + +-------+ IPv4 | 310 | extDNS64 | | | 311 +----+-----+ +----------+ 312 | 313 +----------+ | 314 | | | 315 | IPv6 +-------------+ 316 | | 317 +----------+ 319 Figure 3: NAT64 and DNS64 in external provider 321 One more equivalent scenario will be if the service provider offers 322 the NAT64 function only, and the DNS64 function is from an external 323 provider with or without a specific agreement among them. This is a 324 scenario already common today, as several "global" service providers 325 provide free DNS/DNS64 services and users often configure manually 326 their DNS. This will only work if both the NAT64 and the DNS64 327 functions are using the WKP (Well-Known Prefix) or the same NSP 328 (Network-Specific Prefix). All the considerations in the previous 329 paragraphs of this section are the same for this sub-case. 331 Of course, if the external DNS64 function is agreed with the service 332 provider, then we are in the same case as in the previous ones 333 already depicted in this scenario. 335 +----------+ 336 | | 337 | extDNS64 | 338 | | 339 +----+-----+ 340 | 341 | 342 +----------+ +----+-----+ +----------+ 343 | | | | | | 344 | IPv6 +--------+ NAT64 +--------+ IPv4 | 345 | | | | | | 346 +----------+ +----------+ +----------+ 348 Figure 4: NAT64; DNS64 by external provider 350 3.1.2. Service Provider Offering 464XLAT, with DNS64 352 464XLAT ([RFC6877]) describes an architecture that provides IPv4 353 connectivity across a network, or part of it, when it is only 354 natively transporting IPv6. [RFC7849] already suggest the need to 355 support the CLAT function in order to ensure the IPv4 service 356 continuity in IPv6-only cellular deployments. 358 In order to do that, 464XLAT ([RFC6877]) relies on the combination of 359 existing protocols: 361 1. The customer-side translator (CLAT) is a stateless IPv4 to IPv6 362 translator (NAT46) ([RFC7915]) implemented in the end-user device 363 or CE (Customer Edge Router), located at the "customer edge" of 364 the network. 366 2. The provider-side translator (PLAT) is a stateful NAT64 367 ([RFC6146]), implemented typically at in the operator network. 369 3. Optionally, DNS64 ([RFC6147]), may allow an optimization: a 370 single translation at the NAT64, instead of two translations 371 (NAT46+NAT64), when the application at the end-user device 372 supports IPv6 DNS (uses AAAA Resource Records). 374 Note that even if in the 464XLAT ([RFC6877]) terminology, the 375 provider-side translator is referred as PLAT, for simplicity and 376 uniformity, across this document is always referred as NAT64 377 (function). 379 In this scenario the service provider deploys 464XLAT with a DNS64 380 function. 382 As a consequence, the DNSSEC issues remain, unless the host is doing 383 the address synthesis. 385 464XLAT ([RFC6877]) is a very simple approach to cope with the major 386 NAT64+DNS64 drawback: Not working with applications or devices that 387 use literal IPv4 addresses or non-IPv6 compliant APIs. 389 464XLAT ([RFC6877]) has been used initially mainly in IPv6-only 390 cellular networks. By supporting a CLAT function, the end-user 391 device applications can access IPv4-only end-networks/applications, 392 despite those applications or devices use literal IPv4 addresses or 393 non-IPv6 compliant APIs. 395 In addition to that, in the same example of the cellular network 396 above, if the User Equipment (UE) provides tethering, other devices 397 behind it will be presented with a traditional NAT44, in addition to 398 the native IPv6 support, so clearly it allows IPv4-only hosts behind 399 the IPv6-only access network. 401 Furthermore, as discussed in [RFC6877], 464XLAT can be used in 402 broadband IPv6 network architectures, by implementing the CLAT 403 function at the CE. 405 The support of this scenario offers two additional advantages: 407 o DNS load optimization: A CLAT should implement a DNS proxy (as per 408 [RFC5625]), so that only IPv6 native queries and only for AAAA 409 records are sent to the DNS64 server. Otherwise doubling the 410 number of queries may impact the DNS infrastructure. 412 o Connection establishment delay optimization: If the UE/CE 413 implementation is detecting the presence of a DNS64 function, it 414 may issue only the AAAA query, instead of both the AAAA and A 415 queries. 417 In order to understand all the communication possibilities, let's 418 assume the following representation of two dual-stack peers: 420 +-------+ .-----. .-----. 421 | | / \ / \ 422 .-----. | Res./ | / IPv6- \ .-----. / IPv4- \ 423 / Local \ | SOHO +--( only )---( NAT64 )---( only ) 424 / \ | | \ flow /\ `-----' \ flow / 425 ( Dual- )--+ IPv6 | \ / \ / \ / 426 \ Stack / | CE | `--+--' \ .-----. / `--+--' 427 \ Peer / | with | | \ / Remote\/ | 428 `-----' | CLAT | +---+----+ / \ +---+----+ 429 | | |DNS/IPv6| ( Dual- ) |DNS/IPv4| 430 +-------+ | with | \ Stack / +--------+ 431 | DNS64 | \ Peer / 432 +--------+ `-----' 434 Figure A: Representation of 464XLAT among two peers with DNS64 436 The possible communication paths, among the IPv4/IPv6 stacks of both 437 peers, in this case, are: 439 a. Local-IPv6 to Remote-IPv6: Regular DNS and native IPv6 among 440 peers. 442 b. Local-IPv6 to Remote-IPv4: DNS64 and NAT64 translation. 444 c. Local-IPv4 to Remote-IPv6: Not possible unless the CLAT 445 implements EAMT as indicated by Section 4.9. In principle, it is 446 not expected that services are deployed in Internet using 447 IPv6-only, unless there is certainty that peers will also be 448 IPv6-capable. 450 d. Local-IPv4 to Remote-IPv4: DNS64, CLAT and NAT64 translations. 452 e. Local-IPv4 to Remote-dual-stack using EAMT optimization: If the 453 CLAT implements EAMT as indicated by Section 4.9, instead of 454 using the path d. above, NAT64 translation is avoided and the 455 flow will use IPv6 from the CLAT to the destination. 457 The rest of the figures in this section show different choices for 458 placing the different elements. 460 +----------+ +----------+ +----------+ 461 | IPv6 | | NAT64 | | | 462 | + +--------+ + +--------+ IPv4 | 463 | CLAT | | DNS64 | | | 464 +----------+ +----------+ +----------+ 466 Figure 5: 464XLAT with DNS64 468 A similar scenario will be if the service provider offers only the 469 DNS64 function, and the NAT64 function is provided by an outsourcing 470 agreement with an external provider. All the considerations in the 471 previous paragraphs of this section are the same for this sub-case. 473 +----------+ +----------+ 474 | | | | 475 | extNAT64 +--------+ IPv4 | 476 | | | | 477 +----+-----+ +----------+ 478 | 479 | 480 +----------+ +----+-----+ 481 | IPv6 | | | 482 | + +--------+ DNS64 + 483 | CLAT | | | 484 +----------+ +----------+ 486 Figure 6: 464XLAT with DNS64; NAT64 in external provider 488 As well, is equivalent to the scenario where the outsourcing 489 agreement with the external provider is to provide both the NAT64 and 490 DNS64 functions. Once more, all the considerations in the previous 491 paragraphs of this section are the same for this sub-case. 493 +----------+ +----------+ 494 | extNAT64 | | | 495 | + +--------+ IPv4 | 496 | extDNS64 | | | 497 +----+-----+ +----------+ 498 | 499 +----------+ | 500 | IPv6 | | 501 | + +-------------+ 502 | CLAT | 503 +----------+ 505 Figure 7: 464XLAT with DNS64; NAT64 and DNS64 in external provider 507 3.1.3. Service Provider Offering 464XLAT, without DNS64 509 The major advantage of this scenario, using 464XLAT without DNS64, is 510 that the service provider ensures that DNSSEC is never broken, even 511 in case the user modifies the DNS configuration. Nevertheless, some 512 CLAT implementations or applications may expose an extra delay, which 513 is inducted by the dual A/AAAA queries (and wait for both responses), 514 unless Happy Eyeballs v2 (HEv2, [RFC8305]) is also present. 516 A possible variation of this scenario is the case when DNS64 is used 517 only for the discovery of the NAT64 prefix. The rest of the document 518 is not considering it as a different scenario, because once the 519 prefix has been discovered, the DNS64 function is not used, so it 520 behaves as if the DNS64 synthesis function is not present. 522 In this scenario, as in the previous one, there are no issues related 523 to IPv4-only hosts (or IPv4-only applications) behind the IPv6-only 524 access network, neither related to the usage of IPv4 literals or non- 525 IPv6 compliant APIs. 527 The support of this scenario offers one advantage: 529 o DNS load optimization: A CLAT should implement a DNS proxy (as per 530 [RFC5625]), so that only IPv6 native queries are sent to the DNS64 531 server. Otherwise doubling the number of queries may impact the 532 DNS infrastructure. 534 As indicated earlier, the connection establishment delay optimization 535 is achieved only in the case of devices, Operating Systems, or 536 applications that use HEv2 ([RFC8305]), which is very common. 538 Let's assume the representation of two dual-stack peers as in the 539 previous case: 541 +-------+ .-----. .-----. 542 | | / \ / \ 543 .-----. | Res./ | / IPv6- \ .-----. / IPv4- \ 544 / Local \ | SOHO +--( only )---( NAT64 )---( only ) 545 / \ | | \ flow /\ `-----' \ flow / 546 ( Dual- )--+ IPv6 | \ / \ / \ / 547 \ Stack / | CE | `--+--' \ .-----. / `--+--' 548 \ Peer / | with | | \ / Remote\/ | 549 `-----' | CLAT | +---+----+ / \ +---+----+ 550 | | |DNS/IPv6| ( Dual- ) |DNS/IPv4| 551 +-------+ +--------+ \ Stack / +--------+ 552 \ Peer / 553 `-----' 555 Figure B: Representation of 464XLAT among two peers without DNS64 557 The possible communication paths, among the IPv4/IPv6 stacks of both 558 peers, in this case, are: 560 a. Local-IPv6 to Remote-IPv6: Regular DNS and native IPv6 among 561 peers. 563 b. Local-IPv6 to Remote-IPv4: Regular DNS, CLAT and NAT64 564 translations. 566 c. Local-IPv4 to Remote-IPv6: Not possible unless the CLAT 567 implements EAMT as indicated by Section 4.9. In principle, it is 568 not expected that services are deployed in Internet using 569 IPv6-only, unless there is certainty that peers will also be 570 IPv6-capable. 572 d. Local-IPv4 to Remote-IPv4: Regular DNS, CLAT and NAT64 573 translations. 575 e. Local-IPv4 to Remote-dual-stack using EAMT optimization: If the 576 CLAT implements EAMT as indicated by Section 4.9, instead of 577 using the path d. above, NAT64 translation is avoided and the 578 flow will use IPv6 from the CLAT to the destination. 580 It needs to be noticed that this scenario works while the local 581 hosts/applications are dual-stack (which is the current situation), 582 because the connectivity from a local-IPv6 to a remote-IPv4 is not 583 possible without an AAAA synthesis. This aspect is important only 584 when in the LANs behind the CLAT there are IPv6-only hosts and they 585 need to communicate with remote IPv4-only hosts. However, it doesn't 586 look a sensible approach from an Operating System or application 587 vendor perspective, to provide IPv6-only support unless, similarly to 588 case c above, there is certainty of peers supporting IPv6 as well. A 589 solution approach to this is also presented in 590 [I-D.palet-v6ops-464xlat-opt-cdn-caches]. 592 The following figures show different choices for placing the 593 different elements. 595 +----------+ +----------+ +----------+ 596 | IPv6 | | | | | 597 | + +--------+ NAT64 +--------+ IPv4 | 598 | CLAT | | | | | 599 +----------+ +----------+ +----------+ 601 Figure 8: 464XLAT without DNS64 603 This is equivalent to the scenario where there is an outsourcing 604 agreement with an external provider for the NAT64 function. All the 605 considerations in the previous paragraphs of this section are the 606 same for this sub-case. 608 +----------+ +----------+ 609 | | | | 610 | extNAT64 +--------+ IPv4 | 611 | | | | 612 +----+-----+ +----------+ 613 | 614 +----------+ | 615 | IPv6 | | 616 | + +-------------+ 617 | CLAT | 618 +----------+ 620 Figure 9: 464XLAT without DNS64; NAT64 in external provider 622 3.2. Known to Work Under Special Conditions 624 The scenarios in this category are known to not work unless 625 significant effort is devoted to solve the issues, or are intended to 626 solve problems across "closed" networks, instead of as a general 627 Internet access usage. In addition to the different pros, cons and 628 trade-offs, which may be acceptable for some operators, they have 629 implementation difficulties, as they are beyond the original 630 expectations of the NAT64/DNS64 original intent. 632 3.2.1. Service Provider NAT64 without DNS64 634 In this scenario, the service provider offers a NAT64 function, 635 however there is no DNS64 function support at all. 637 As a consequence, an IPv6 host in the IPv6-only access network, will 638 not be able to detect the presence of DNS64 by means of [RFC7050], 639 neither to learn the IPv6 prefix to be used for the NAT64 function. 641 This can be sorted out as indicated in Section 4.1.1. 643 However, despite that, because the lack of the DNS64 function, the 644 IPv6 host will not be able to obtain AAAA synthesized records, so the 645 NAT64 function becomes useless. 647 An exception to this "useless" scenario will be manually configure 648 mappings between the A records of each of the IPv4-only remote hosts 649 and the corresponding AAAA records, with the WKP (Well-Known Prefix) 650 or NSP (Network-Specific Prefix) used by the service provider NAT64 651 function, as if they were synthesized by a DNS64 function. 653 This mapping could be done by several means, typically at the 654 authoritative DNS server, or at the service provider resolvers by 655 means of DNS RPZ (Response Policy Zones, [I-D.vixie-dns-rpz]) or 656 equivalent functionality. DNS RPZ, may have implications in DNSSEC, 657 if the zone is signed. Also, if the service provider is using an 658 NSP, having the mapping at the authoritative server, may create 659 troubles to other parties trying to use different NSP or the WKP, 660 unless multiple DNS "views" (split-DNS) is also being used at the 661 authoritative servers. 663 Generally, the mappings alternative, will only make sense if a few 664 set of IPv4-only remote hosts need to be accessed by a single network 665 (or a small number of them), which support IPv6-only in the access. 666 This will require some kind of mutual agreement for using this 667 procedure, so it doesn't care if they become a trouble for other 668 parties across Internet ("closed services"). 670 In any case, this scenario doesn't solve the issue of IPv4 literal 671 addresses or non-IPv6 compliant APIs, neither it solves the problem 672 of IPv4-only hosts within that IPv6-only access network. 674 +----------+ +----------+ +----------+ 675 | | | | | | 676 | IPv6 +--------+ NAT64 +--------+ IPv4 | 677 | | | | | | 678 +----------+ +----------+ +----------+ 680 Figure 10: NAT64 without DNS64 682 3.2.2. Service Provider NAT64; DNS64 in the IPv6 hosts 684 In this scenario, the service provider offers the NAT64 function, but 685 not the DNS64 function. However, the IPv6 hosts have a built-in 686 DNS64 function. 688 This may become common if the DNS64 function is implemented in all 689 the IPv6 hosts/stacks, which is not the actual situation, but it may 690 happen in the medium-term. At this way, the DNSSEC validation is 691 performed on the A record, and then the host can use the DNS64 692 function so to be able to use the NAT64 function, without any DNSSEC 693 issues. 695 This scenario fails to solve the issue of IPv4 literal addresses or 696 non-IPv6 compliant APIs, unless the IPv6 hosts also supports HEv2 697 ([RFC8305], Section 7.1), which may solve that issue. 699 However, this scenario still fails to solve the problem of IPv4-only 700 hosts or applications behind the IPv6-only access network. 702 +----------+ +----------+ +----------+ 703 | IPv6 | | | | | 704 | + +--------+ NAT64 +--------+ IPv4 | 705 | DNS64 | | | | | 706 +----------+ +----------+ +----------+ 708 Figure 11: NAT64; DNS64 in IPv6 hosts 710 3.2.3. Service Provider NAT64; DNS64 in the IPv4-only remote network 712 In this scenario, the service provider offers the NAT64 function 713 only. The remote IPv4-only network offers the DNS64 function. 715 This is not common, and looks like doesn't make too much sense that a 716 remote network, not deploying IPv6, is providing a DNS64 function. 717 As in the case of the scenario depicted in Section 3.2.1, it will 718 only work if both sides are using the WKP or the same NSP, so the 719 same considerations apply. It can be also tuned to behave as in 720 Section 3.1.1 722 This scenario still fails to solve the issue of IPv4 literal 723 addresses or non-IPv6 compliant APIs. 725 This scenario also fails to solve the problem of IPv4-only hosts or 726 applications behind the IPv6-only access network. 728 +----------+ +----------+ +----------+ 729 | | | | | IPv4 | 730 | IPv6 +--------+ NAT64 +--------+ + | 731 | | | | | DNS64 | 732 +----------+ +----------+ +----------+ 734 Figure 12: NAT64; DNS64 in the IPv4-only 736 3.3. Comparing the Scenarios 738 This section compares the different scenarios, including the possible 739 variations (each one represented in the precedent sections by a 740 different figure), looking at the following criteria: 742 a. DNSSEC: Are there hosts validating DNSSEC? 744 b. Literal/APIs: Are there applications using IPv4 literals or non- 745 IPv6 compliant APIs? 747 c. IPv4-only: Are there hosts or applications using IPv4-only? 749 d. Foreign DNS: Is the scenario surviving if the user, Operating 750 System, applications or devices change the DNS? 752 e. DNS load opt. (DNS load optimization): Are there extra queries 753 that may impact DNS infrastructure?. 755 f. Connect. opt. (Connection establishment delay optimization): Is 756 the UE/CE issuing only the AAAA query or also an A query and 757 waiting for both responses? 759 In the next table, the columns represent each of the scenarios from 760 the previous sections, by the figure number. The possible values 761 are: 763 o "-" Scenario "bad" for that criteria. 765 o "+" Scenario "good" for that criteria. 767 o "*" Scenario "bad" for that criteria, however it is typically 768 resolved, with the support of HEv2 ([RFC8305]). 770 In some cases "countermeasures", alternative or special 771 configurations, may be available for the criteria designated as 772 "bad". So this comparison is considering a generic case, as a quick 773 comparison guide. In some cases, a "bad" criteria is not necessarily 774 a negative aspect, all it depends on the specific needs/ 775 characteristics of the network where the deployment will take place. 776 For instance, in a network which has only IPv6-only hosts and apps 777 using only DNS and IPv6-compliant APIs, there is no impact using only 778 NAT64 and DNS64, but if the hosts may validate DNSSEC, that item is 779 still relevant. 781 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 782 | Item / Figure | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 783 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 784 | DNSSEC | - | - | - | - | - | - | - | + | + | + | + | + | 785 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 786 | Literal/APIs | - | - | - | - | + | + | + | + | + | - | - | - | 787 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 788 | IPv4-only | - | - | - | - | + | + | + | + | + | - | - | - | 789 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 790 | Foreign DNS | - | - | - | - | + | + | + | + | + | - | + | - | 791 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 792 | DNS load opt. | + | + | + | + | + | + | + | + | + | + | + | + | 793 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 794 | Connect. opt. | + | + | + | + | + | + | + | * | * | + | + | + | 795 +----------------+---+---+---+---+---+---+---+---+---+----+----+----+ 797 Figure 13: Scenario Comparison 799 As a general conclusion, we should note that, if the network must 800 support applications using any of the following: 802 o IPv4 literals 804 o non-IPv6-compliant APIs 806 o IPv4-only hosts or applications 808 Then, only the scenarios with 464XLAT, a CLAT function, or equivalent 809 built-in local address synthesis features, will provide a valid 810 solution. Further to that, those scenarios will also keep working if 811 the DNS configuration is modified. Clearly also, depending on if 812 DNS64 is used or not, DNSSEC may be broken for those hosts doing 813 DNSSEC validation. 815 All the scenarios are good in terms of DNS load optimization, and in 816 the case of 464XLAT it may provide an extra degree of optimization. 817 Finally, all them are also good in terms of connection establishment 818 delay optimization. However, in the case of 464XLAT without DNS64, 819 it requires the usage of HEv2. This is not an issue, as commonly it 820 is available in actual Operating Systems. 822 4. Issues to be Considered 824 This section reviews the different issues that an operator needs to 825 consider towards a NAT64/464XLAT deployment, as they may bring to 826 specific decision points about how to approach that deployment. 828 4.1. DNSSEC Considerations and Possible Approaches 830 As indicated in Section 8 of [RFC6147] (DNS64, Security 831 Considerations), because DNS64 modifies DNS answers and DNSSEC is 832 designed to detect such modifications, DNS64 may break DNSSEC. 834 If a device connected to an IPv6-only WAN, queries for a domain name 835 in a signed zone, by means of a recursive name server that supports 836 DNS64, and the result is a synthesized AAAA record, and the recursive 837 name server is configured to perform DNSSEC validation and has a 838 valid chain of trust to the zone in question, it will 839 cryptographically validate the negative response from the 840 authoritative name server. This is the expected DNS64 behavior: The 841 recursive name server actually "lies" to the client device. However, 842 in most of the cases, the client will not notice it, because 843 generally, they don't perform validation themselves and instead, rely 844 on the recursive name servers. 846 A validating DNS64 resolver in fact, increase the confidence on the 847 synthetic AAAA, as it has validated that a non-synthetic AAAA for 848 sure, doesn't exists. However, if the client device is 849 NAT64-oblivious (most common case) and performs DNSSEC validation on 850 the AAAA record, it will fail as it is a synthesized record. 852 The best possible scenario from DNSSEC point of view, is when the 853 client requests the DNS64 server to perform the DNSSEC validation (by 854 setting the DO bit to 1 and the CD bit to 0). In this case, the 855 DNS64 server validates the data, thus tampering may only happen 856 inside the DNS64 server (which is considered as a trusted part, thus 857 its likelihood is low) or between the DNS64 server and the client. 858 All other parts of the system (including transmission and caching) 859 are protected by DNSSEC ([Threat-DNS64]). 861 Similarly, if the client querying the recursive name server is 862 another name server configured to use it as a forwarder, and is 863 performing DNSSEC validation, it will also fail on any synthesized 864 AAAA record. 866 All those considerations are extensively covered in Sections 3, 5.5 867 and 6.2 of [RFC6147]. 869 A solution to avoid DNSSEC issues, will be that all the signed zones 870 also provide IPv6 connectivity, together with the corresponding AAAA 871 records. However, this is out of the control of the operator needing 872 to deploy a NAT64 function. This has been proposed already in 873 [I-D.bp-v6ops-ipv6-ready-dns-dnssec]. 875 An alternative solution, which was the one considered while 876 developing [RFC6147], is that validators will be DNS64-aware, so 877 could perform the necessary discovery and do their own synthesis. 878 That was done under the expectation that it was sufficiently early in 879 the validator-deployment curve that it would be ok to break certain 880 DNSSEC assumptions for networks who were really stuck in a NAT64/ 881 DNS64-needing world. 883 As already indicated, the scenarios in the previous section, are in 884 fact somehow simplified, looking at the worst possible case. Saying 885 it in a different way: "trying to look for the most perfect 886 approach". DNSSEC breach will not happen if the end-host is not 887 doing validation. 889 Existing previous studies seems to indicate that the figures of 890 DNSSEC actually broken by using DNS64 will be around 1.7% 891 ([About-DNS64]) of the cases. However we can not negate that this 892 may increase, as DNSSEC deployment grows. Consequently, a decision 893 point for the operator must depend on "do I really care for that 894 percentage of cases and the impact in my helpdesk or can I provide 895 alternative solutions for them?". Some possible solutions may be 896 taken, as depicted in the next sections. 898 4.1.1. Not using DNS64 900 A solution will be to avoid using DNS64, but as already indicated 901 this is not possible in all the scenarios. 903 The use of DNS64 is a key component for some networks, in order to 904 comply with traffic performance metrics, monitored by some 905 governmental bodies and other institutions. 907 One drawback of not having a DNS64 at the network side, is that is 908 not possible to heuristically discover the NAT64 ([RFC7050]). 909 Consequently, an IPv6 host behind the IPv6-only access network, will 910 not be able to detect the presence of the NAT64 function, neither to 911 learn the IPv6 prefix to be used for it, unless it is configured by 912 alternative means. 914 The discovery of the IPv6 prefix could be solved by means of adding 915 the relevant AAAA records to the ipv4only.arpa. zone of the service 916 provider recursive servers, i.e., if using the WKP (64:ff9b::/96): 918 ipv4only.arpa. SOA . . 0 0 0 0 0 919 ipv4only.arpa. NS . 920 ipv4only.arpa. AAAA 64:ff9b::192.0.0.170 921 ipv4only.arpa. AAAA 64:ff9b::192.0.0.171 922 ipv4only.arpa. A 192.0.0.170 923 ipv4only.arpa. A 192.0.0.171 925 An alternative option to the above, is the use of DNS RPZ 926 ([I-D.vixie-dns-rpz]) or equivalent functionalities. Note that this 927 may impact DNSSEC if the zone is signed. 929 One more alternative, only valid in environments with PCP support 930 (for both the hosts or CEs and for the service provider network), is 931 to follow [RFC7225] (Discovering NAT64 IPv6 Prefixes using PCP). 933 Other alternatives may be available in the future. All them are 934 extensively discussed in [RFC7051], however the deployment evolution 935 has evolved many considerations from that document. New options are 936 being documented, such using Router Advertising 937 ([I-D.ietf-6man-ra-pref64]) or DHCPv6 options 938 ([I-D.li-intarea-nat64-prefix-dhcp-option]). 940 It may be convenient the simultaneous support of several of the 941 possible approaches, in order to ensure that clients with different 942 ways to configure the NAT64 prefix, successfully obtain it. This is 943 also convenient even if DNS64 is being used. 945 4.1.2. DNSSEC validator aware of DNS64 947 In general, by default, DNS servers with DNS64 function, will not 948 synthesize AAAA responses if the DNSSEC OK (DO) flag was set in the 949 query. In this case, as only an A record is available, it means that 950 the CLAT will take the responsibility, as in the case of literal IPv4 951 addresses, to keep that traffic flow end-to-end as IPv4, so DNSSEC is 952 not broken. However, this will not work if a CLAT function is not 953 present as the hosts will not be able to use IPv4 (scenarios without 954 464XLAT). 956 4.1.3. Stub validator 958 If the DO flag is set and the client device performs DNSSEC 959 validation, and the Checking Disabled (CD) flag is set for a query, 960 the DNS64 recursive server will not synthesize AAAA responses. In 961 this case, the client could perform the DNSSEC validation with the A 962 record and then synthesize the AAAA ([RFC6052]). For that to be 963 possible, the client must have learned beforehand the NAT64 prefix 964 using any of the available methods ([RFC7050], [RFC7225], 965 [I-D.ietf-6man-ra-pref64], 966 [I-D.li-intarea-nat64-prefix-dhcp-option]). This allows the client 967 device to avoid using the DNS64 function and still use NAT64 even 968 with DNSSEC. 970 If the end-host is IPv4-only, this will not work if a CLAT function 971 is not present (scenarios without 464XLAT). 973 Some devices or Operating Systems may implement, instead of a CLAT, 974 an equivalent function by using Bump-in-the-Host ([RFC6535]), 975 implemented as part of HEv2 (Section 7.1 of [RFC8305]). In this 976 case, the considerations in the above paragraphs are also applicable. 978 4.1.4. CLAT with DNS proxy and validator 980 If a CE includes CLAT support and also a DNS proxy, as indicated in 981 Section 6.4 of [RFC6877], the CE could behave as a stub validator on 982 behalf of the client devices. Then, following the same approach 983 described in the Section 4.1.3, the DNS proxy actually will "lie" to 984 the client devices, which in most of the cases will not notice it, 985 unless they perform validation by themselves. Again, this allow the 986 client devices to avoid using the DNS64 function and still use NAT64 987 with DNSSEC. 989 Once more, this will not work without a CLAT function (scenarios 990 without 464XLAT). 992 4.1.5. ACL of clients 994 In cases of dual-stack clients, the AAAA queries typically take 995 preference over A queries. If DNS64 is enabled for those clients, 996 will never get A records, even for IPv4-only servers. As a 997 consequence, if the IPv4-only servers are in the path before the 998 NAT64 function, the clients will never reach them. If DNSSEC is 999 being used for all those flows, specific addresses or prefixes can be 1000 left-out of the DNS64 synthesis by means of ACLs. 1002 Once more, this will not work without a CLAT function (scenarios 1003 without 464XLAT). 1005 4.1.6. Mapping-out IPv4 addresses 1007 If there are well-known specific IPv4 addresses or prefixes using 1008 DNSSEC, they can be mapped-out of the DNS64 synthesis. 1010 Even if this is not related to DNSSEC, this "mapping-out" feature is 1011 actually, quite commonly used to ensure that [RFC1918] addresses (for 1012 example used by LAN servers) are not synthesized to AAAA. 1014 Once more, this will not work without a CLAT function (scenarios 1015 without 464XLAT). 1017 4.2. DNS64 and Reverse Mapping 1019 When a client device, using DNS64 tries to reverse-map a synthesized 1020 IPv6 address, the name server responds with a CNAME record pointing 1021 the domain name used to reverse-map the synthesized IPv6 address (the 1022 one under ip6.arpa), to the domain name corresponding to the embedded 1023 IPv4 address (under in-addr.arpa). 1025 This is the expected behavior, so no issues need to be considered 1026 regarding DNS reverse mapping. 1028 4.3. Using 464XLAT with/without DNS64 1030 In the case the client device is IPv6-only (either because the stack 1031 or application is IPv6-only, or because it is connected via an 1032 IPv6-only LAN) and the remote server is IPv4-only (either because the 1033 stack is IPv4-only, or because it is connected via an IPv4-only LAN), 1034 only NAT64 combined with DNS64 will be able to provide access among 1035 both. Because DNS64 is then required, DNSSEC validation will be only 1036 possible if the recursive name server is validating the negative 1037 response from the authoritative name server and the client is not 1038 performing validation. 1040 Note that is not expected at this stage of the transition, that 1041 applications, devices or Operating Systems are IPv6-only. It will 1042 not be a sensible decision for a developer to work on that direction, 1043 unless it is clear that the deployment scenario fully supports it. 1045 On the other hand, an end-user or enterprise network may decide to 1046 run IPv6-only in the LANs. In case there is any chance for 1047 applications to be IPv6-only, the Operating System may be responsible 1048 either for doing a local address synthesis, or alternatively, setting 1049 up some kind of on-demand VPN (IPv4-in-IPv6), which need to be 1050 supported by that network. This may become very common in enterprise 1051 networks, where "Unique IPv6 Prefix per Host" [RFC8273] is supported. 1053 However, when the client device is dual-stack and/or connected in a 1054 dual-stack LAN by means of a CLAT function (or has a built-in CLAT 1055 function), DNS64 is an option. 1057 1. With DNS64: If DNS64 is used, most of the IPv4 traffic (except if 1058 using literal IPv4 addresses or non-IPv6 compliant APIs) will not 1059 use the CLAT, so will use the IPv6 path and only one translation 1060 will be done at the NAT64. This may break DNSSEC, unless 1061 measures as described in the precedent sections are taken. 1063 2. Without DNS64: If DNS64 is not used, all the IPv4 traffic will 1064 make use of the CLAT, so two translations are required (NAT46 at 1065 the CLAT and NAT64 at the PLAT), which adds some overhead in 1066 terms of the extra NAT46 translation. However, this avoids the 1067 AAAA synthesis and consequently will never break DNSSEC. 1069 Note that the extra translation, when DNS64 is not used, takes place 1070 at the CLAT, which means no extra overhead for the operator. 1071 However, adds potential extra delays to establish the connections, 1072 and no perceptible impact for a CE in a broadband network, while it 1073 may have some impact in a battery powered device. This cost for a 1074 battery powered device, is possibly comparable to the cost when the 1075 device is doing a local address synthesis (see Section 7.1 of 1076 [RFC8305]). 1078 4.4. Foreign DNS 1080 Clients, devices or applications in a service provider network, may 1081 use DNS servers from other networks. This may be the case either if 1082 individual applications use their own DNS server, the Operating 1083 System itself or even the CE, or combinations of the above. 1085 Those "foreign" DNS servers may not support DNS64, which will mean 1086 that those scenarios that require a DNS64 may not work. However, if 1087 a CLAT function is available, the considerations in Section 4.3 will 1088 apply. 1090 In the case that the foreign DNS supports the DNS64 function, we may 1091 be in the situation of providing incorrect configurations parameters, 1092 for example un-matching WKP or NSP, or a case such the one described 1093 in Section 3.2.3. 1095 Having a CLAT function, even if using foreign DNS without a DNS64 1096 function, ensures that everything will work, so the CLAT must be 1097 considered as an advantage even against user configuration errors. 1098 The cost of this, is that all the traffic will use a double 1099 translation (NAT46 at the CLAT and NAT64 at the operator network), 1100 unless there is support for EAMT (Section 4.9). 1102 An exception to that is the case when there is a CLAT function at the 1103 CE, which is not able to obtain the correct configuration parameters 1104 (again, un-matching WKP or NSP). 1106 However, it needs to be reinforced, that if there is not a CLAT 1107 function (scenarios without 464XLAT), an external DNS without DNS64 1108 support, will disallow any access to IPv4-only destination networks, 1109 and will not guarantee DNSSEC, so will behave as in the 1110 Section 3.2.1. 1112 The causes of "foreign DNS" could be classified in three main 1113 categories, as depicted in the following sub-sections. 1115 4.4.1. Manual Configuration of Foreign DNS 1117 It is becoming increasingly common that end-users or even devices or 1118 applications configure alternative DNS in their Operating Systems, 1119 and sometimes in CEs. 1121 4.4.2. DNS Privacy 1123 A new trend is for clients or applications to use mechanisms for DNS 1124 privacy/encryption, such as DNS over TLS ([RFC7858]), DNS over DTLS 1125 ([RFC8094]), DNS queries over HTTPS ([RFC8484]) or DNS over QUIC 1126 ([I-D.huitema-quic-dnsoquic]). Those are commonly cited as DoT, DoH 1127 and DoQ. 1129 Those DNS privacy/encryption options, currently are typically 1130 provided by the applications, not the Operating System vendors. At 1131 the time of writing this document, at least DoT and DoH standards 1132 have declared DNS64 (and consequently NAT64) out of their scope, so 1133 an application using them may break NAT64, unless a correctly 1134 configured CLAT function is used. 1136 4.4.3. Split DNS 1138 When networks or hosts use "split-DNS" (also called Split Horizon, 1139 DNS views or private DNS), the successful use of the DNS64 is not 1140 guaranteed. Section 4 of [RFC6950], analyses this case. 1142 A similar situation may happen in case of VPNs that force all the DNS 1143 queries through the VPN, ignoring the operator DNS64 function. 1145 4.5. Well-Known Prefix (WKP) vs Network-Specific Prefix (NSP) 1147 [RFC6052] (IPv6 Addressing of IPv4/IPv6 Translators), Section 3, 1148 discusses some considerations which are useful to decide if an 1149 operator should use the WKP or an NSP. 1151 Taking in consideration that discussion and other issues, we can 1152 summarize the possible decision points as: 1154 a. The WKP MUST NOT be used to represent non-global IPv4 addresses. 1155 If this is required because the network to be translated use non- 1156 global addresses, then an NSP is required. 1158 b. The WKP MAY appear in inter-domain routing tables, if the 1159 operator provides a NAT64 function to peers. However, in this 1160 case, special considerations related to BGP filtering are 1161 required and IPv4-embedded IPv6 prefixes longer than the WKP MUST 1162 NOT be advertised (or accepted) in BGP. An NSP may be a more 1163 appropriate option in those cases. 1165 c. If several NAT64 use the same prefix, packets from the same flow 1166 may be routed to different NAT64 in case of routing changes. 1167 This can be avoided either by using different prefixes for each 1168 NAT64 function, or by ensuring that all the NAT64 coordinate 1169 their state. Using an NSP could simplify that. 1171 d. If DNS64 is required and users, devices, Operating Systems or 1172 applications may change their DNS configuration, and deliberately 1173 choose an alternative DNS64 function, most probably alternative 1174 DNS64 will use by default the WKP. In that case, if an NSP is 1175 used by the NAT64 function, clients will not be able to use the 1176 operator NAT64 function, which will break connectivity to 1177 IPv4-only destinations. 1179 4.6. IPv4 literals and old APIs 1181 A host or application using literal IPv4 addresses or older APIs, 1182 behind a network with IPv6-only access, will not work unless any of 1183 the following alternatives is provided: 1185 o CLAT (or equivalent function). 1187 o HEv2 (Section 7.1, [RFC8305]). 1189 o Bump-in-the-Host ([RFC6535]) with a DNS64 function. 1191 Those alternatives will solve the problem for an end-host. However, 1192 if that end-hosts is providing "tethering" or an equivalent service 1193 to other hosts, that needs to be considered as well. In other words, 1194 in a case of a cellular network, it resolves the issue for the UE 1195 itself, but may be not the case for hosts behind it. 1197 Otherwise, the support of 464XLAT is the only valid and complete 1198 approach to resolve this issue. 1200 4.7. IPv4-only Hosts or Applications 1202 An IPv4-only hosts or application behind a network with IPv6-only 1203 access, will not work unless a CLAT function is present. 1205 464XLAT is the only valid approach to resolve this issue. 1207 4.8. CLAT Translation Considerations 1209 As described in Section 6.3 of [RFC6877] (IPv6 Prefix Handling), if 1210 the CLAT function can be configured with a dedicated /64 prefix for 1211 the NAT46 translation, then it will be possible to do a more 1212 efficient stateless translation. 1214 Otherwise, if this dedicated prefix is not available, the CLAT 1215 function will need to do a stateful translation, for example 1216 performing stateful NAT44 for all the IPv4 LAN packets, so they 1217 appear as coming from a single IPv4 address, and then in turn, 1218 stateless translated to a single IPv6 address. 1220 A possible setup, in order to maximize the CLAT performance, is to 1221 configure the dedicated translation prefix. This can be easily 1222 achieved automatically, if the broadband CE or end-user device is 1223 able to obtain a shorter prefix by means of DHCPv6-PD ([RFC8415]), or 1224 other alternatives. The CE can then use a specific /64 for the 1225 translation. This is also possible when broadband is provided by a 1226 cellular access. 1228 The above recommendation is often not possible for cellular networks, 1229 when connecting smartphones (as UEs), as generally they don't use 1230 DHCPv6-PD ([RFC8415]). Instead, a single /64 is provided for each 1231 PDP context and prefix sharing ([RFC6877]) is used. So, in this 1232 case, the UEs typically have a build-in CLAT function which is 1233 performing a stateful NAT44 translation before the stateless NAT46. 1235 4.9. EAMT Considerations 1237 Explicit Address Mappings for Stateless IP/ICMP Translation 1238 ([RFC7757]) provide a way to configure explicit mappings between IPv4 1239 and IPv6 prefixes of any length. When this is used, for example in a 1240 CLAT function, it may provide a simple mechanism in order to avoid 1241 traffic flows between IPv4-only nodes or applications and dual-stack 1242 destinations to be translated twice (NAT46 and NAT64), by creating 1243 mapping entries with the GUA of the IPv6-reachable destination. This 1244 optimization of the NAT64 usage is very useful in many scenarios, 1245 including CDNs and caches, as described in 1246 [I-D.palet-v6ops-464xlat-opt-cdn-caches]. 1248 In addition to that, it may provide as well a way for IPv4-only nodes 1249 or applications to communicate with IPv6-only destinations. 1251 5. Summary of Deployment Recommendations for NAT64/464XLAT 1253 NAT64/464XLAT has demonstrated to be a valid choice in several 1254 scenarios (IPv6-IPv4 and IPv4-IPv6-IPv4), with hundreds of millions 1255 of users, offering different choices of deployment, depending on each 1256 network case, needs and requirements. Despite that, this document is 1257 not an explicit recommendation for using this choice versus other 1258 IPv4aaS transition mechanisms. Instead, this document is a guide 1259 that facilitates evaluating a possible implementation of 1260 NAT64/464XLAT and key decision points about specific design 1261 considerations for its deployment. 1263 Depending on the specific requirements of each deployment case, DNS64 1264 may be a required function, while in other cases the adverse effects 1265 may be counterproductive. Similarly, in some cases a NAT64 function, 1266 together with a DNS64 function, may be a valid solution, when there 1267 is a certainty that IPv4-only hosts or applications do not need to be 1268 supported (Section 4.6 and Section 4.7). However, in other cases 1269 (i.e. IPv4-only devices or applications need to be supported), the 1270 limitations of NAT64/DNS64, may suggest the operator to look into 1271 464XLAT as a more complete solution. 1273 In the case of broadband managed networks (where the CE is provided 1274 or suggested/supported by the operator), in order to fully support 1275 the actual user needs (IPv4-only devices and applications, usage of 1276 IPv4 literals and old APIs), the 464XLAT scenario should be 1277 considered. In that case, it must support a CLAT function. 1279 If the operator provides DNS services, in order to increase 1280 performance by reducing the double translation for all the IPv4 1281 traffic, they may support a DNS64 function and avoid, as much as 1282 possible, breaking DNSSEC. In this case, if the DNS service is 1283 offering DNSSEC validation, then it must be in such way that it is 1284 aware of the DNS64. This is considered the simpler and safer 1285 approach, and may be combined as well with other recommendations 1286 described in this document: 1288 o DNS infrastructure MUST be aware of DNS64 (Section 4.1.2). 1290 o Devices running CLAT SHOULD follow the indications in 1291 Section 4.1.3 (Stub Validator). However, this may be out of the 1292 control of the operator. 1294 o CEs SHOULD include a DNS proxy and validator (Section 4.1.4). 1296 o Section 4.1.5 (ACL of clients) and Section 4.1.6 (Mapping-out IPv4 1297 addresses) MAY be considered by operators, depending on their own 1298 infrastructure. 1300 This "increased performance" approach has the disadvantage of 1301 potentially breaking DNSSEC for a small percentage of validating end- 1302 hosts versus the small impact of a double translation taking place in 1303 the CE. If CE performance is not an issue, which is the most 1304 frequent case, then a much safer approach is to not use DNS64 at all, 1305 and consequently, ensure that all the IPv4 traffic is translated at 1306 the CLAT (Section 4.3). 1308 If DNS64 is not used, at least one of the alternatives described in 1309 Section 4.1.1, must be followed in order to learn he NAT64 prefix. 1311 The operator needs to consider that if the DNS configuration can be 1312 modified (Section 4.4, Section 4.4.2, Section 4.4.3), which most 1313 probably is impossible to avoid, there are chances that instead of 1314 configuring a DNS64 a foreign non-DNS64 is used. In a scenario with 1315 only a NAT64 function IPv4-only remote host will no longer be 1316 accesible. Instead, it will continue to work in the case of 464XLAT. 1318 Similar considerations need to be taken regarding the usage of a 1319 NAT64 WKP vs NSP (Section 4.5), as they must match with the 1320 configuration of the DNS64. In case of using foreign DNS, they may 1321 not match. If there is a CLAT and the configured foreign DNS is not 1322 a DNS64, the network will keep working only if other means of 1323 learning the NAT64 prefix are available. 1325 As described in Section 4.8, for broadband networks, the CEs 1326 supporting a CLAT function, SHOULD support DHCPv6-PD ([RFC8415]), or 1327 alternative means for configuring a shorter prefix. The CE SHOULD 1328 internally reserve one /64 for the stateless NAT46 translation. The 1329 operator must ensure that the customers get allocated prefixes 1330 shorter than /64 in order to support this optimization. One way or 1331 the other, this is not impacting the performance of the operator 1332 network. 1334 Operators may follow Section 7 of [RFC6877] (Deployment 1335 Considerations), for suggestions in order to take advantage of 1336 traffic engineering requirements. 1338 In the case of cellular networks, the considerations regarding DNSSEC 1339 may appear as out-of-scope, because UEs Operating Systems, commonly 1340 don't support DNSSEC. However, applications running on them may do, 1341 or it may be an Operating System "built-in" support in the future. 1342 Moreover, if those devices offer tethering, other client devices 1343 behind the UE, may be doing the validation, hence the relevance of a 1344 proper DNSSEC support by the operator network. 1346 Furthermore, cellular networks supporting 464XLAT ([RFC6877]) and 1347 "Discovery of the IPv6 Prefix Used for IPv6 Address Synthesis" 1348 ([RFC7050]), allow a progressive IPv6 deployment, with a single APN 1349 supporting all types of PDP context (IPv4, IPv6, IPv4v6). This 1350 approach allows the network to automatically serve every possible 1351 combinations of UEs. 1353 If the operator chooses to provide validation for the DNS64 prefix 1354 discovery, it must follow the advice from Section 3.1. of [RFC7050] 1355 (Validation of Discovered Pref64::/n). 1357 One last consideration, is that many networks may have a mix of 1358 different complex scenarios at the same time, for example, customers 1359 requiring 464XLAT, others not requiring it, customers requiring 1360 DNS64, others not, etc. In general, the different issues and the 1361 approaches described in this document can be implemented at the same 1362 time for different customers or parts of the network. That mix of 1363 approaches don't present any problem or incompatibility, as they work 1364 well together, being just a matter of appropriate and differentiated 1365 provisioning. In fact, the NAT64/464XLAT approach facilitates an 1366 operator offering both cellular and broadband services, to have a 1367 single IPv4aaS for both networks while differentiating the deployment 1368 key decisions to optimize each case. It even makes possible using 1369 hybrid CEs that have a main broadband access link and a backup via 1370 the cellular network. 1372 In an ideal world we could safely use DNS64, if the approach proposed 1373 in [I-D.bp-v6ops-ipv6-ready-dns-dnssec] is followed, avoiding the 1374 cases where DNSSEC may be broken. However, this will not solve the 1375 issues related to DNS Privacy and Split DNS. 1377 The only 100% safe solution, which also resolves all the issues, will 1378 be, in addition to having a CLAT function, not using a DNS64 but 1379 instead making sure that the hosts have a built-in address synthesis 1380 feature. Operators could manage to provide CEs with the CLAT 1381 function, however the built-in address synthesis feature is out of 1382 their control. If the synthesis is provided either by the Operating 1383 System (via its DNS resolver API) or by the application (via its own 1384 DNS resolver), in such way that the prefix used for the NAT64 1385 function is reachable for the host, the problem goes away. 1387 Whenever feasible, using EAMT ([RFC7757]) as indicated in 1388 Section 4.9, provides a very relevant optimization, avoiding double- 1389 translations. 1391 6. Deployment of NAT64 in Enterprise Networks 1393 The recommendations of this document can be used as well in 1394 enterprise networks, campus and other similar scenarios (including 1395 managed end-user networks). 1397 This include scenarios where the NAT64 function (and DNS64 function, 1398 if available) are under the control of that network (or can be 1399 configured manually according to that network specific requirements), 1400 and for whatever reasons, there is a need to provide "IPv6-only 1401 access" to any part of that network or it is IPv6-only connected to 1402 third party-networks. 1404 An example of that is the IETF meetings network itself, where both 1405 NAT64 and DNS64 functions are provided, presenting in this case the 1406 same issues as per Section 3.1.1. If there is a CLAT function in the 1407 IETF network, then there is no need to use DNS64 and it falls under 1408 the considerations of Section 3.1.3. Both scenarios have been tested 1409 and verified already in the IETF network itself. 1411 Next figures are only meant to represent a few of the possible 1412 scenarios, not pretending to be the only feasible ones. 1414 The following figure provides an example of an IPv6-only enterprise 1415 network connected with dual-stack to Internet and using local NAT64 1416 and DNS64 functions. 1418 +----------------------------------+ 1419 | Enterprise Network | 1420 | +----------+ +----------+ | +----------+ 1421 | | IPv6 | | NAT64 | | | IPv4 | 1422 | | only +--------+ + | +-------+ + | 1423 | | LANs | | DNS64 | | | IPv6 | 1424 | +----------+ +----------+ | +----------+ 1425 +----------------------------------+ 1427 Figure 14: IPv6-only enterprise with NAT64 and DNS64 1429 The following figure provides an example of a dual-stack (DS) 1430 enterprise network connected with dual-stack (DS) to Internet and 1431 using a CLAT function, without a DNS64 function. 1433 +----------------------------------+ 1434 | Enterprise Network | 1435 | +----------+ +----------+ | +----------+ 1436 | | IPv6 | | | | | IPv4 | 1437 | | + +--------+ NAT64 | +-------+ + | 1438 | | CLAT | | | | | IPv6 | 1439 | +----------+ +----------+ | +----------+ 1440 +----------------------------------+ 1442 Figure 15: DS enterprise with CLAT, DS Internet, without DNS64 1444 Finally, the following figure provides an example of an IPv6-only 1445 provider with a NAT64 function, and a dual-stack (DS) enterprise 1446 network by means of their own CLAT function, without a DNS64 1447 function. 1449 +----------------------------------+ 1450 | Enterprise Network | 1451 | +----------+ +----------+ | +----------+ 1452 | | IPv6 | | | | IPv6 | | 1453 | | + +--------+ CLAT | +--------+ NAT64 | 1454 | | IPv4 | | | | only | | 1455 | +----------+ +----------+ | +----------+ 1456 +----------------------------------+ 1458 Figure 16: DS enterprise with CLAT, IPv6-only Access, without DNS64 1460 7. Security Considerations 1462 This document does not have new specific security considerations 1463 beyond those already reported by each of the documents cited. 1465 8. IANA Considerations 1467 This document does not have any new specific IANA considerations. 1469 Note: This section is assuming that https://www.rfc- 1470 editor.org/errata/eid5152 is resolved, otherwise, this section may 1471 include the required text to resolve the issue. 1473 9. Acknowledgements 1475 The author would like to acknowledge the inputs of Gabor Lencse, 1476 Andrew Sullivan, Lee Howard, Barbara Stark, Fred Baker, Mohamed 1477 Boucadair, Alejandro D'Egidio, Dan Wing and Mikael Abrahamsson. 1479 Conversations with Marcelo Bagnulo, one of the co-authors of NAT64 1480 and DNS64, as well as several emails in mailing lists from Mark 1481 Andrews, have been very useful for this work. 1483 Christian Huitema inspired working in this document by suggesting 1484 that DNS64 should never be used, during a discussion regarding the 1485 deployment of CLAT in the IETF network. 1487 10. ANNEX A: Example of Broadband Deployment with 464XLAT 1489 This section summarizes how an operator may deploy an IPv6-only 1490 network for residential/SOHO customers, supporting IPv6 inbound 1491 connections, and IPv4-as-a-Service (IPv4aaS) by using 464XLAT. 1493 Note that an equivalent setup could also be provided for enterprise 1494 customers. In case they need to support IPv4 inbound connections, 1495 several mechanisms, depending on specific customer needs, allow that, 1496 for instance [RFC7757]. 1498 Conceptually, most part of the operator network could be IPv6-only 1499 (represented in the next pictures as "IPv6-only flow"), or even if 1500 this part of the network is actually dual-stack, only IPv6-access is 1501 available for some customers (i.e. residential customers). This part 1502 of the network connects the IPv6-only subscribers (by means of 1503 IPv6-only access links), to the IPv6 upstream providers, as well as 1504 to the IPv4-Internet by means of the NAT64 (PLAT in the 464XLAT 1505 terminology). 1507 The traffic flow from and back to the CE to services available in the 1508 IPv6 Internet (or even dual-stack remote services, when IPv6 is being 1509 used), is purely native IPv6 traffic, so there are no special 1510 considerations about it. 1512 Looking at the picture from the DNS perspective, there are remote 1513 networks with are IPv4-only, and typically will have only IPv4 DNS 1514 (DNS/IPv4), or at least will be seen as that from the CE perspective. 1515 At the operator side, the DNS, as seen from the CE, is only IPv6 1516 (DNS/IPv6) and has also a DNS64 function. 1518 In the customer LANs side, there is actually one network, which of 1519 course could be split in different segments. The most common setup 1520 will be those segments being dual-stack, using global IPv6 addresses 1521 and [RFC1918] for IPv4, as usual in any regular residential/SOHO IPv4 1522 network. In the figure, it is represented as tree segments, just to 1523 show that the three possible setups are valid (IPv6-only, IPv4-only 1524 and dual-stack). 1526 .-----. +-------+ .-----. .-----. 1527 / IPv6- \ | | / \ / \ 1528 ( only )--+ Res./ | / IPv6- \ .-----. / IPv4- \ 1529 \ LANs / | SOHO +--( only )--( NAT64 )--( only ) 1530 `-----' | | \ flow / `-----' \ flow / 1531 .-----. | IPv6 | \ / \ / 1532 / IPv4- \ | CE | `--+--' `--+--' 1533 ( only )--+ with | | | 1534 \ LANs / | CLAT | +---+----+ +---+----+ 1535 `-----' | | |DNS/IPv6| |DNS/IPv4| 1536 .-----. +---+---+ | with | +--------+ 1537 / Dual- \ | | DNS64 | 1538 ( Stack )------| +--------+ 1539 \ LANs / 1540 `-----' 1542 Figure 17: CE setup with built-in CLAT with DNS64 1544 In addition to the regular CE setup, which will be typically access- 1545 technology dependent, the steps for the CLAT function configuration 1546 can be summarized as: 1548 1. Discovery of the PLAT (NAT64) prefix: It may be done using 1549 [RFC7050], or in those networks where PCP is supported, by means 1550 of [RFC7225], or other alternatives that may be available in the 1551 future, such as Router Advertising ([I-D.ietf-6man-ra-pref64]) or 1552 DHCPv6 options ([I-D.li-intarea-nat64-prefix-dhcp-option]). 1554 2. If the CLAT function allows stateless NAT46 translation, a /64 1555 from the pool typically provided to the CE by means of DHCPv6-PD 1556 [RFC8415], need to be set aside for that translation. Otherwise, 1557 the CLAT is forced to perform an intermediate stateful NAT44 1558 before the a stateless NAT46, as described in Section 4.8. 1560 A more detailed configuration approach is described in 1562 [I-D.ietf-v6ops-transition-ipv4aas]. 1564 The operator network needs to ensure that the correct responses are 1565 provided for the discovery of the PLAT prefix. It is highly 1566 recommended to follow [RIPE-690], in order to ensure that multiple 1567 /64s are available, including the one needed for the NAT46 stateless 1568 translation. 1570 The operator needs to understand other issues, described across this 1571 document, in order to take the relevant decisions. For example, if 1572 several NAT64 functions are needed in the context of scalability/ 1573 high-availability, an NSP should be considered (Section 4.5). 1575 More complex scenarios are possible, for example, if a network offers 1576 multiple NAT64 prefixes, destination-based NAT64 prefixes, etc. 1578 If the operator decides not to provide a DNS64 function, then this 1579 setup turns into the one in the following Figure. This will be also 1580 the setup that "will be seen" from the perspective of the CE, if a 1581 foreign DNS is used and consequently is not the operator-provided 1582 DNS64 function. 1584 .-----. +-------+ .-----. .-----. 1585 / IPv6- \ | | / \ / \ 1586 ( only )--+ Res./ | / IPv6- \ .-----. / IPv4- \ 1587 \ LANs / | SOHO +--( only )--( NAT64 )--( only ) 1588 `-----' | | \ flow / `-----' \ flow / 1589 .-----. | IPv6 | \ / \ / 1590 / IPv4- \ | CE | `--+--' `--+--' 1591 ( only )--+ with | | | 1592 \ LANs / | CLAT | +---+----+ +---+----+ 1593 `-----' | | |DNS/IPv6| |DNS/IPv4| 1594 .-----. +---+---+ +--------+ +--------+ 1595 / Dual- \ | 1596 ( Stack )------| 1597 \ LANs / 1598 `-----' 1600 Figure 18: CE setup with built-in CLAT without DNS64 1602 In this case, the discovery of the PLAT prefix needs to be arranged 1603 as indicated in Section 4.1.1. 1605 In this case, the CE doesn't have a built-in CLAT function, or the 1606 customer can choose to setup the IPv6 operator-managed CE in bridge 1607 mode (and optionally use an external router), or for example, there 1608 is an access technology that requires some kind of media converter 1609 (ONT for FTTH, CableModem for DOCSIS, etc.), the complete setup will 1610 look as in the next figure. Obviously, there will be some 1611 intermediate configuration steps for the bridge, depending on the 1612 specific access technology/protocols, which should not modify the 1613 steps already described in the previous cases for the CLAT function 1614 configuration. 1616 +-------+ .-----. .-----. 1617 | | / \ / \ 1618 | Res./ | / IPv6- \ .-----. / IPv4- \ 1619 | SOHO +--( only )--( NAT64 )--( only ) 1620 | | \ flow / `-----' \ flow / 1621 | IPv6 | \ / \ / 1622 | CE | `--+--' `--+--' 1623 | Bridge| | | 1624 | | +---+----+ +---+----+ 1625 | | |DNS/IPv6| |DNS/IPv4| 1626 +---+---+ +--------+ +--------+ 1627 | 1628 .-----. +---+---+ 1629 / IPv6- \ | | 1630 ( only )--+ IPv6 | 1631 \ LANs / | Router| 1632 `-----' | | 1633 .-----. | with | 1634 / IPv4- \ | CLAT | 1635 ( only )--+ | 1636 \ LANs / | | 1637 `-----' | | 1638 .-----. +---+---+ 1639 / Dual- \ | 1640 ( Stack )------| 1641 \ LANs / 1642 `-----' 1644 Figure 19: CE setup with bridged CLAT without DNS64 1646 It should be avoided that several routers (i.e., the operator 1647 provided CE and a downstream user provided router) enable 1648 simultaneously routing and/or CLAT, in order to avoid multiple NAT44 1649 and NAT46 levels, as well as ensuring the correct operation of 1650 multiple IPv6 subnets. In those cases, it is suggested the use of 1651 HNCP ([RFC8375]). 1653 Note that the procedure described here for the CE setup, can be 1654 simplified if the CE follows [I-D.ietf-v6ops-transition-ipv4aas]. 1656 11. ANNEX B: CLAT Implementation 1658 In addition to the regular set of features for a CE, a CLAT CE 1659 implementation requires support of: 1661 o [RFC7915] for the NAT46 function. 1663 o [RFC7050] for the PLAT prefix discovery. 1665 o [RFC7225] for the PLAT prefix discovery if PCP is supported. 1667 o [I-D.ietf-6man-ra-pref64] for the PLAT prefix discovery by means 1668 of Router Advertising. 1670 o If stateless NAT46 is supported, a mechanism to ensure that 1671 multiple /64 are available, such as DHCPv6-PD [RFC8415]. 1673 There are several OpenSource implementations of CLAT, such as: 1675 o Android: https://github.com/ddrown/android_external_android-clat. 1677 o Jool: https://www.jool.mx. 1679 o Linux: https://github.com/toreanderson/clatd. 1681 o OpenWRT: https://github.com/openwrt- 1682 routing/packages/blob/master/nat46/files/464xlat.sh. 1684 o VPP: https://git.fd.io/vpp/tree/src/plugins/nat. 1686 12. ANNEX C: Benchmarking 1688 [RFC8219] has defined a benchmarking methodology for IPv6 transition 1689 technologies. NAT64 and 464XLAT are addressed among the single and 1690 double translation technologies, respectively. DNS64 is addressed in 1691 Section 9, and the methodology is more elaborated in [DNS64-BM-Meth]. 1693 Several documents provide references to benchmarking results, for 1694 example in the case of DNS64, [DNS64-Benchm]. 1696 13. ANNEX D: Changes from -00 to -01/-02 1698 Section to be removed after WGLC. Significant updates are: 1700 1. Text changes across all the document. 1702 14. ANNEX E: Changes from -02 to -03 1704 Section to be removed after WGLC. Significant updates are: 1706 1. Added references to new cited documents. 1708 2. Reference to RFC8273 and on-demand IPv4-in-IPv6 VPN for IPv6-only 1709 LANs w/o DNS64. 1711 3. Overall review and editorial changes. 1713 15. ANNEX F: Changes from -03 to -04 1715 Section to be removed after WGLC. Significant updates are: 1717 1. Added text related to EAMT considerations. 1719 16. ANNEX G: Changes from -04 to -05 1721 Section to be removed after WGLC. Significant updates are: 1723 1. Added cross references to EAMT section. 1725 2. Reworded "foreing DNS section". 1727 3. Overall editorial review of text, pictures and nits correction. 1729 17. References 1731 17.1. Normative References 1733 [RFC1918] Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G., 1734 and E. Lear, "Address Allocation for Private Internets", 1735 BCP 5, RFC 1918, DOI 10.17487/RFC1918, February 1996, 1736 . 1738 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1739 Requirement Levels", BCP 14, RFC 2119, 1740 DOI 10.17487/RFC2119, March 1997, 1741 . 1743 [RFC5625] Bellis, R., "DNS Proxy Implementation Guidelines", 1744 BCP 152, RFC 5625, DOI 10.17487/RFC5625, August 2009, 1745 . 1747 [RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X. 1748 Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052, 1749 DOI 10.17487/RFC6052, October 2010, 1750 . 1752 [RFC6144] Baker, F., Li, X., Bao, C., and K. Yin, "Framework for 1753 IPv4/IPv6 Translation", RFC 6144, DOI 10.17487/RFC6144, 1754 April 2011, . 1756 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 1757 NAT64: Network Address and Protocol Translation from IPv6 1758 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 1759 April 2011, . 1761 [RFC6147] Bagnulo, M., Sullivan, A., Matthews, P., and I. van 1762 Beijnum, "DNS64: DNS Extensions for Network Address 1763 Translation from IPv6 Clients to IPv4 Servers", RFC 6147, 1764 DOI 10.17487/RFC6147, April 2011, 1765 . 1767 [RFC6535] Huang, B., Deng, H., and T. Savolainen, "Dual-Stack Hosts 1768 Using "Bump-in-the-Host" (BIH)", RFC 6535, 1769 DOI 10.17487/RFC6535, February 2012, 1770 . 1772 [RFC6877] Mawatari, M., Kawashima, M., and C. Byrne, "464XLAT: 1773 Combination of Stateful and Stateless Translation", 1774 RFC 6877, DOI 10.17487/RFC6877, April 2013, 1775 . 1777 [RFC7050] Savolainen, T., Korhonen, J., and D. Wing, "Discovery of 1778 the IPv6 Prefix Used for IPv6 Address Synthesis", 1779 RFC 7050, DOI 10.17487/RFC7050, November 2013, 1780 . 1782 [RFC7225] Boucadair, M., "Discovering NAT64 IPv6 Prefixes Using the 1783 Port Control Protocol (PCP)", RFC 7225, 1784 DOI 10.17487/RFC7225, May 2014, 1785 . 1787 [RFC7757] Anderson, T. and A. Leiva Popper, "Explicit Address 1788 Mappings for Stateless IP/ICMP Translation", RFC 7757, 1789 DOI 10.17487/RFC7757, February 2016, 1790 . 1792 [RFC7915] Bao, C., Li, X., Baker, F., Anderson, T., and F. Gont, 1793 "IP/ICMP Translation Algorithm", RFC 7915, 1794 DOI 10.17487/RFC7915, June 2016, 1795 . 1797 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1798 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1799 May 2017, . 1801 [RFC8273] Brzozowski, J. and G. Van de Velde, "Unique IPv6 Prefix 1802 per Host", RFC 8273, DOI 10.17487/RFC8273, December 2017, 1803 . 1805 [RFC8305] Schinazi, D. and T. Pauly, "Happy Eyeballs Version 2: 1806 Better Connectivity Using Concurrency", RFC 8305, 1807 DOI 10.17487/RFC8305, December 2017, 1808 . 1810 [RFC8375] Pfister, P. and T. Lemon, "Special-Use Domain 1811 'home.arpa.'", RFC 8375, DOI 10.17487/RFC8375, May 2018, 1812 . 1814 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 1815 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 1816 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 1817 RFC 8415, DOI 10.17487/RFC8415, November 2018, 1818 . 1820 [RFC8484] Hoffman, P. and P. McManus, "DNS Queries over HTTPS 1821 (DoH)", RFC 8484, DOI 10.17487/RFC8484, October 2018, 1822 . 1824 17.2. Informative References 1826 [About-DNS64] 1827 Linkova, J., "Let's talk about IPv6 DNS64 & DNSSEC", 2016, 1828 . 1831 [DNS64-Benchm] 1832 Lencse, G. and Y. Kadobayashi, "Benchmarking DNS64 1833 Implementations: Theory and Practice", Computer 1834 Communications , vol. 127, no. 1, pp. 61-74, 1835 DOI 10.1016/j.comcom.2018.05.005, September 2018. 1837 [DNS64-BM-Meth] 1838 Lencse, G., Georgescu, M., and Y. Kadobayashi, 1839 "Benchmarking Methodology for DNS64 Servers", Computer 1840 Communications , vol. 109, no. 1, pp. 162-175, 1841 DOI 10.1016/j.comcom.2017.06.004, September 2017. 1843 [I-D.bp-v6ops-ipv6-ready-dns-dnssec] 1844 Byrne, C. and J. Palet, "IPv6-Ready DNS/DNSSSEC 1845 Infrastructure", draft-bp-v6ops-ipv6-ready-dns-dnssec-00 1846 (work in progress), October 2018. 1848 [I-D.huitema-quic-dnsoquic] 1849 Huitema, C., Shore, M., Mankin, A., Dickinson, S., and J. 1850 Iyengar, "Specification of DNS over Dedicated QUIC 1851 Connections", draft-huitema-quic-dnsoquic-06 (work in 1852 progress), March 2019. 1854 [I-D.ietf-6man-ra-pref64] 1855 Colitti, L., Kline, E., and J. Linkova, "Discovering 1856 PREF64 in Router Advertisements", draft-ietf-6man-ra- 1857 pref64-00 (work in progress), March 2019. 1859 [I-D.ietf-v6ops-transition-ipv4aas] 1860 Palet, J., Liu, H., and M. Kawashima, "Requirements for 1861 IPv6 Customer Edge Routers to Support IPv4 Connectivity 1862 as-a-Service", draft-ietf-v6ops-transition-ipv4aas-15 1863 (work in progress), January 2019. 1865 [I-D.li-intarea-nat64-prefix-dhcp-option] 1866 Li, L., Cui, Y., Liu, C., Wu, J., Baker, F., and J. Palet, 1867 "DHCPv6 Options for Discovery NAT64 Prefixes", draft-li- 1868 intarea-nat64-prefix-dhcp-option-01 (work in progress), 1869 March 2017. 1871 [I-D.lmhp-v6ops-transition-comparison] 1872 Lencse, G., Palet, J., Howard, L., Patterson, R., and I. 1873 Farrer, "Pros and Cons of IPv6 Transition Technologies for 1874 IPv4aaS", draft-lmhp-v6ops-transition-comparison-02 (work 1875 in progress), January 2019. 1877 [I-D.palet-v6ops-464xlat-opt-cdn-caches] 1878 Palet, J. and A. D'Egidio, "464XLAT Optimization for CDNs/ 1879 Caches", draft-palet-v6ops-464xlat-opt-cdn-caches-01 (work 1880 in progress), March 2019. 1882 [I-D.vixie-dns-rpz] 1883 Vixie, P. and V. Schryver, "DNS Response Policy Zones 1884 (RPZ)", draft-vixie-dns-rpz-04 (work in progress), 1885 December 2016. 1887 [RFC6889] Penno, R., Saxena, T., Boucadair, M., and S. Sivakumar, 1888 "Analysis of Stateful 64 Translation", RFC 6889, 1889 DOI 10.17487/RFC6889, April 2013, 1890 . 1892 [RFC6950] Peterson, J., Kolkman, O., Tschofenig, H., and B. Aboba, 1893 "Architectural Considerations on Application Features in 1894 the DNS", RFC 6950, DOI 10.17487/RFC6950, October 2013, 1895 . 1897 [RFC7051] Korhonen, J., Ed. and T. Savolainen, Ed., "Analysis of 1898 Solution Proposals for Hosts to Learn NAT64 Prefix", 1899 RFC 7051, DOI 10.17487/RFC7051, November 2013, 1900 . 1902 [RFC7269] Chen, G., Cao, Z., Xie, C., and D. Binet, "NAT64 1903 Deployment Options and Experience", RFC 7269, 1904 DOI 10.17487/RFC7269, June 2014, 1905 . 1907 [RFC7849] Binet, D., Boucadair, M., Vizdal, A., Chen, G., Heatley, 1908 N., Chandler, R., Michaud, D., Lopez, D., and W. Haeffner, 1909 "An IPv6 Profile for 3GPP Mobile Devices", RFC 7849, 1910 DOI 10.17487/RFC7849, May 2016, 1911 . 1913 [RFC7858] Hu, Z., Zhu, L., Heidemann, J., Mankin, A., Wessels, D., 1914 and P. Hoffman, "Specification for DNS over Transport 1915 Layer Security (TLS)", RFC 7858, DOI 10.17487/RFC7858, May 1916 2016, . 1918 [RFC8094] Reddy, T., Wing, D., and P. Patil, "DNS over Datagram 1919 Transport Layer Security (DTLS)", RFC 8094, 1920 DOI 10.17487/RFC8094, February 2017, 1921 . 1923 [RFC8219] Georgescu, M., Pislaru, L., and G. Lencse, "Benchmarking 1924 Methodology for IPv6 Transition Technologies", RFC 8219, 1925 DOI 10.17487/RFC8219, August 2017, 1926 . 1928 [RIPE-690] 1929 RIPE, "Best Current Operational Practice for Operators: 1930 IPv6 prefix assignment for end-users - persistent vs non- 1931 persistent, and what size to choose", October 2017, 1932 . 1934 [Threat-DNS64] 1935 Lencse, G. and Y. Kadobayashi, "Methodology for the 1936 identification of potential security issues of different 1937 IPv6 transition technologies: Threat analysis of DNS64 and 1938 stateful NAT64", Computers & Security , vol. 77, no. 1, 1939 pp. 397-411, DOI 10.1016/j.cose.2018.04.012, August 2018. 1941 Author's Address 1943 Jordi Palet Martinez 1944 The IPv6 Company 1945 Molino de la Navata, 75 1946 La Navata - Galapagar, Madrid 28420 1947 Spain 1949 Email: jordi.palet@theipv6company.com 1950 URI: http://www.theipv6company.com/