idnits 2.17.1 draft-ietf-mpls-ldp-ipv6-14.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 3 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 4 instances 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 (Using the creation date from RFC5036, updated by this document, for RFC5378 checks: 2004-10-12) -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, you can and should remove the disclaimer; otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (October 2, 2014) is 3491 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) -- Obsolete informational reference (is this intentional?): RFC 7321 (Obsoleted by RFC 8221) Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 MPLS Working Group Rajiv Asati 2 Internet Draft Carlos Pignataro 3 Updates: 5036, 6720 (if approved) Kamran Raza 4 Intended status: Standards Track Cisco 5 Expires: April 2015 6 Vishwas Manral 7 Hewlett-Packard, Inc 9 Rajiv Papneja 10 Huawei 12 October 2, 2014 14 Updates to LDP for IPv6 15 draft-ietf-mpls-ldp-ipv6-14 17 Status of this Memo 19 This Internet-Draft is submitted in full conformance with the 20 provisions of BCP 78 and BCP 79. 22 Internet-Drafts are working documents of the Internet Engineering 23 Task Force (IETF). Note that other groups may also distribute 24 working documents as Internet-Drafts. The list of current Internet- 25 Drafts is at http://datatracker.ietf.org/drafts/current/. 27 Internet-Drafts are draft documents valid for a maximum of six 28 months and may be updated, replaced, or obsoleted by other documents 29 at any time. It is inappropriate to use Internet-Drafts as 30 reference material or to cite them other than as "work in progress." 32 This Internet-Draft will expire on April 2, 2015. 34 Copyright Notice 36 Copyright (c) 2014 IETF Trust and the persons identified as the 37 document authors. All rights reserved. 39 This document is subject to BCP 78 and the IETF Trust's Legal 40 Provisions Relating to IETF Documents 41 (http://trustee.ietf.org/license-info) in effect on the date of 42 publication of this document. Please review these documents 43 carefully, as they describe your rights and restrictions with 44 respect to this document. Code Components extracted from this 45 document must include Simplified BSD License text as described in 46 Section 4.e of the Trust Legal Provisions and are provided without 47 warranty as described in the Simplified BSD License. 49 This document may contain material from IETF Documents or IETF 50 Contributions published or made publicly available before November 51 10, 2008. The person(s) controlling the copyright in some of this 52 material may not have granted the IETF Trust the right to allow 53 modifications of such material outside the IETF Standards Process. 54 Without obtaining an adequate license from the person(s) controlling 55 the copyright in such materials, this document may not be modified 56 outside the IETF Standards Process, and derivative works of it may 57 not be created outside the IETF Standards Process, except to format 58 it for publication as an RFC or to translate it into languages other 59 than English. 61 Abstract 63 The Label Distribution Protocol (LDP) specification defines 64 procedures to exchange label bindings over either IPv4, or IPv6 or 65 both networks. This document corrects and clarifies the LDP behavior 66 when IPv6 network is used (with or without IPv4). This document 67 updates RFC 5036 and RFC 6720. 69 Table of Contents 71 1. Introduction...................................................3 72 1.1. Topology Scenarios for Dual-stack Environment.............4 73 1.2. Single-hop vs. Multi-hop LDP Peering......................5 74 2. Specification Language.........................................6 75 3. LSP Mapping....................................................7 76 4. LDP Identifiers................................................7 77 5. Neighbor Discovery.............................................8 78 5.1. Basic Discovery Mechanism.................................8 79 5.1.1. Maintaining Hello Adjacencies........................9 80 5.2. Extended Discovery Mechanism..............................9 81 6. LDP Session Establishment and Maintenance......................9 82 6.1. Transport connection establishment.......................10 83 6.1.1. Determining Transport connection Roles..............11 84 6.2. LDP Sessions Maintenance.................................14 85 7. Binding Distribution..........................................14 86 7.1. Address Distribution.....................................15 87 7.2. Label Distribution.......................................15 89 8. LDP Identifiers and Duplicate Next Hop Addresses..............16 90 9. LDP TTL Security..............................................17 91 10. IANA Considerations..........................................18 92 11. Security Considerations......................................18 93 12. Acknowledgments..............................................19 94 13. Additional Contributors......................................19 95 14. References...................................................20 96 14.1. Normative References....................................20 97 14.2. Informative References..................................20 98 Appendix A.......................................................22 99 A.1. LDPv6 and LDPv4 Interoperability Safety Net..............22 100 A.2. Accommodating Non-RFC5036-compliant implementations......22 101 A.3. Why prohibit IPv4-mapped IPv6 addresses in LDP...........23 102 A.4. Why 32-bit value even for IPv6 LDP Router ID.............23 103 Author's Addresses...............................................24 105 1. Introduction 107 The LDP [RFC5036] specification defines procedures and messages for 108 exchanging FEC-label bindings over either IPv4 or IPv6 or both (e.g. 109 Dual-stack) networks. 111 However, RFC5036 specification has the following deficiency (or 112 lacks details) in regards to IPv6 usage (with or without IPv4): 114 1) LSP Mapping: No rule for mapping a particular packet to a 115 particular LSP that has an Address Prefix FEC element containing 116 IPv6 address of the egress router 118 2) LDP Identifier: No details specific to IPv6 usage 120 3) LDP Discovery: No details for using a particular IPv6 destination 121 (multicast) address or the source address 123 4) LDP Session establishment: No rule for handling both IPv4 and 124 IPv6 transport address optional objects in a Hello message, and 125 subsequently two IPv4 and IPv6 transport connections 127 5) LDP Address Distribution: No rule for advertising IPv4 or/and 128 IPv6 Address bindings over an LDP session 130 6) LDP Label Distribution: No rule for advertising IPv4 or/and IPv6 131 FEC-label bindings over an LDP session, and for handling the co- 132 existence of IPv4 and IPv6 FEC Elements in the same FEC TLV 134 7) Next Hop Address Resolution: No rule for accommodating the usage 135 of duplicate link-local IPv6 addresses 137 8) LDP TTL Security: No rule for built-in Generalized TTL Security 138 Mechanism (GTSM) in LDP with IPv6 (this is a deficiency in 139 RFC6720) 141 This document addresses the above deficiencies by specifying the 142 desired behavior/rules/details for using LDP in IPv6 enabled 143 networks (IPv6-only or Dual-stack networks). 145 Note that this document updates RFC5036 and RFC6720. 147 1.1. Topology Scenarios for Dual-stack Environment 149 Two LSRs may involve basic and/or extended LDP discovery in IPv6 150 and/or IPv4 address-families in various topology scenarios. 152 This document addresses the following 3 topology scenarios in which 153 the LSRs may be connected via one or more Dual-stack LDP enabled 154 interfaces (figure 1), or one or more Single-stack LDP enabled 155 interfaces (figure 2 and figure 3): 157 R1------------------R2 158 IPv4+IPv6 160 Figure 1 LSRs connected via a Dual-stack Interface 162 IPv4 163 R1=================R2 164 IPv6 166 Figure 2 LSRs connected via two Single-stack Interfaces 167 R1------------------R2---------------R3 168 IPv4 IPv6 170 Figure 3 LSRs connected via a Single-stack Interface 172 Note that the topology scenario illustrated in figure 1 also covers 173 the case of a Single-stack LDP enabled interface (IPv4, say) being 174 converted to a Dual-stacked LDP enabled interface (by enabling IPv6 175 routing as well as IPv6 LDP), even though the LDPoIPv4 session may 176 already be established between the LSRs. 178 Note that the topology scenario illustrated in figure 2 also covers 179 the case of two routers getting connected via an additional Single- 180 stack LDP enabled interface (IPv6 routing and IPv6 LDP), even though 181 the LDPoIPv4 session may already be established between the LSRs 182 over the existing interface(s). 184 This document also addresses the scenario in which the LSRs do the 185 extended discovery in IPv6 and/or IPv4 address-families: 187 IPv4 188 R1-------------------R2 189 IPv6 191 Figure 4 LSRs involving IPv4 and IPv6 address-families 193 1.2. Single-hop vs. Multi-hop LDP Peering 195 LDP TTL Security mechanism specified by this document applies only 196 to single-hop LDP peering sessions, but not to multi-hop LDP peering 197 sessions, in line with Section 5.5 of [RFC5082] that describes 198 Generalized TTL Security Mechanism (GTSM). 200 As a consequence, any LDP feature that relies on multi-hop LDP 201 peering session would not work with GTSM and will warrant 202 (statically or dynamically) disabling GTSM. Please see section 10. 204 2. Specification Language 206 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 207 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 208 document are to be interpreted as described in [RFC2119]. 210 Abbreviations: 212 LDP - Label Distribution Protocol 214 LDPoIPv4 - LDP over IPv4 transport connection 216 LDPoIPv6 - LDP over IPv6 transport connection 218 FEC - Forwarding Equivalence Class 220 TLV - Type Length Value 222 LSR - Label Switching Router 224 LSP - Label Switched Path 226 LSPv4 - IPv4-signaled Label Switched Path [RFC4798] 228 LSPv6 - IPv6-signaled Label Switched Path [RFC4798] 230 AFI - Address Family Identifier 232 LDP Id - LDP Identifier 234 Single-stack LDP - LDP supporting just one address family (for 235 discovery, session setup, address/label binding 236 exchange etc.) 238 Dual-stack LDP - LDP supporting two address families (for 239 discovery, session setup, address/label binding 240 exchange etc.) 242 Dual-stack LSR - LSR supporting Dual-stack LDP for a peer 244 Single-stack LSR - LSR supporting Single-stack LDP for a peer 246 Note that an LSR can be a Dual-stack and Single-stack LSR at the 247 same time for different peers. This document loosely uses the term 248 address family to mean IP address family. 250 3. LSP Mapping 252 Section 2.1 of [RFC5036] specifies the procedure for mapping a 253 particular packet to a particular LSP using three rules. Quoting the 254 3rd rule from RFC5036: 256 "If it is known that a packet must traverse a particular egress 257 router, and there is an LSP that has an Address Prefix FEC element 258 that is a /32 address of that router, then the packet is mapped to 259 that LSP." 261 This rule is correct for IPv4, but not for IPv6, since an IPv6 262 router may even have a /64 or /96 or /128 (or whatever prefix 263 length) address. Hence, it is reasonable to say IPv4 or IPv6 address 264 instead of /32 or /128 addresses as shown below in the updated rule: 266 "If it is known that a packet must traverse a particular egress 267 router, and there is an LSP that has an Address Prefix FEC element 268 that is an IPv4 or IPv6 address of that router, then the packet is 269 mapped to that LSP." 271 4. LDP Identifiers 273 In line with section 2.2.2 of [RFC5036], this document specifies the 274 usage of 32-bit (unsigned non-zero integer) LSR Id on an IPv6 275 enabled LSR (with or without Dual-stacking). 277 This document also qualifies the first sentence of last paragraph of 278 Section 2.5.2 of [RFC5036] to be per address family and therefore 279 updates that sentence to the following: 281 "For a given address family, an LSR MUST advertise the same 282 transport address in all Hellos that advertise the same label 283 space." 285 This rightly enables the per-platform label space to be shared 286 between IPv4 and IPv6. 288 In summary, this document mandates the usage of a common LDP 289 identifier (same LSR Id aka LDP Router Id as well as a common Label 290 space id) for both IPv4 and IPv6 address families. 292 5. Neighbor Discovery 294 If Dual-stack LDP is enabled (e.g. LDP enabled in both IPv6 and IPv4 295 address families) on an interface or for a targeted neighbor, then 296 the LSR MUST transmit both IPv6 and IPv4 LDP (Link or targeted) 297 Hellos and include the same LDP Identifier (assuming per-platform 298 label space usage) in them. 300 If Single-stack LDP is enabled (e.g. LDP enabled in either IPv6 or 301 IPv4 address family), then the LSR MUST transmit either IPv6 or IPv4 302 LDP (Link or targeted) Hellos respectively. 304 5.1. Basic Discovery Mechanism 306 Section 2.4.1 of [RFC5036] defines the Basic Discovery mechanism for 307 directly connected LSRs. Following this mechanism, LSRs periodically 308 send LDP Link Hellos destined to "all routers on this subnet" group 309 multicast IP address. 311 Interesting enough, per the IPv6 addressing architecture [RFC4291], 312 IPv6 has three "all routers on this subnet" multicast addresses: 314 FF01:0:0:0:0:0:0:2 = Interface-local scope 316 FF02:0:0:0:0:0:0:2 = Link-local scope 318 FF05:0:0:0:0:0:0:2 = Site-local scope 320 [RFC5036] does not specify which particular IPv6 'all routers on 321 this subnet' group multicast IP address should be used by LDP Link 322 Hellos. 324 This document specifies the usage of link-local scope e.g. 325 FF02:0:0:0:0:0:0:2 as the destination multicast IP address in IPv6 326 LDP Link Hellos. An LDP Link Hello packet received on any of the 327 other destination addresses MUST be dropped. Additionally, the link- 328 local IPv6 address MUST be used as the source IP address in IPv6 LDP 329 Link Hellos. 331 Also, the LDP Link Hello packets MUST have their IPv6 Hop Limit set 332 to 255, be checked for the same upon receipt (before any LDP 333 specific processing) and be handled as specified in Generalized TTL 334 Security Mechanism (GTSM) section 3 of [RFC5082]. The built-in 335 inclusion of GTSM automatically protects IPv6 LDP from off-link 336 attacks. 338 More importantly, if an interface is a Dual-stack LDP interface 339 (e.g. LDP enabled in both IPv6 and IPv4 address families), then the 340 LSR MUST periodically transmit both IPv6 and IPv4 LDP Link Hellos 341 (using the same LDP Identifier per section 4) on that interface and 342 be able to receive them. This facilitates discovery of IPv6-only, 343 IPv4-only and Dual-stack peers on the interface's subnet and ensures 344 successful subsequent peering using the appropriate (address family) 345 transport on a multi-access or broadcast interface. 347 An implementation MUST transmit IPv6 LDP link Hellos before IPv4 LDP 348 Link Hellos on a Dual-stack interface, particularly during the 349 interface coming into service or configuration time. 351 5.1.1. Maintaining Hello Adjacencies 353 In case of Dual-stack LDP enabled interface, the LSR SHOULD maintain 354 link Hello adjacencies for both IPv4 and IPv6 address families. This 355 document, however, allows an LSR to maintain Rx-side Link Hello 356 adjacency only for the address family that has been used for the 357 establishment of the LDP session (whether LDPoIPv4 or LDPoIPv6 358 session). 360 5.2. Extended Discovery Mechanism 362 The extended discovery mechanism (defined in section 2.4.2 of 363 [RFC5036]), in which the targeted LDP Hellos are sent to a unicast 364 IPv6 address destination, requires only one IPv6 specific 365 consideration: the link-local IPv6 addresses MUST NOT be used as the 366 targeted LDP hello packet's source or destination addresses. 368 6. LDP Session Establishment and Maintenance 370 Section 2.5.1 of [RFC5036] defines a two-step process for LDP 371 session establishment, once the neighbor discovery has completed 372 (i.e. LDP Hellos have been exchanged): 374 1. Transport connection establishment 375 2. Session initialization 377 The forthcoming sub-section 6.1 discusses the LDP consideration for 378 IPv6 and/or Dual-stacking in the context of session establishment, 379 whereas sub-section 6.2 discusses the LDP consideration for IPv6 380 and/or Dual-stacking in the context of session maintenance. 382 6.1. Transport connection establishment 384 Section 2.5.2 of [RFC5036] specifies the use of an optional 385 transport address object (TLV) in LDP Hello message to convey the 386 transport (IP) address, however, it does not specify the behavior of 387 LDP if both IPv4 and IPv6 transport address objects (TLV) are sent 388 in a Hello message or separate Hello messages. More importantly, it 389 does not specify whether both IPv4 and IPv6 transport connections 390 should be allowed, if both IPv4 and IPv6 Hello adjacencies were 391 present prior to the session establishment. 393 This document specifies that: 395 1. An LSR MUST NOT send a Hello message containing both IPv4 and 396 IPv6 transport address optional objects. In other words, there 397 MUST be at most one optional Transport Address object in a 398 Hello message. An LSR MUST include only the transport address 399 whose address family is the same as that of the IP packet 400 carrying the Hello message. 402 2. An LSR SHOULD accept the Hello message that contains both IPv4 403 and IPv6 transport address optional objects, but MUST use only 404 the transport address whose address family is the same as that 405 of the IP packet carrying the Hello message. An LSR SHOULD 406 accept only the first transport object for a given address 407 family in the received Hello message, and ignore the rest, if 408 the LSR receives more than one transport object for a given 409 address family. 411 3. An LSR MUST send separate Hello messages (each containing 412 either IPv4 or IPv6 transport address optional object) for each 413 IP address family, if Dual-stack LDP was enabled. 415 4. An LSR MUST use a global unicast IPv6 address in IPv6 transport 416 address optional object of outgoing targeted Hellos, and check 417 for the same in incoming targeted hellos (i.e. MUST discard the 418 targeted hello, if it failed the check). 420 5. An LSR MUST prefer using a global unicast IPv6 address in IPv6 421 transport address optional object of outgoing Link Hellos, if 422 it had to choose between global unicast IPv6 address and 423 unique-local or link-local IPv6 address. 425 6. A Dual-stack LSR MUST NOT initiate (or accept the request for) 426 a TCP connection for a new LDP session with a remote LSR, if 427 they already have an LDPoIPv4 or LDPoIPv6 session (for the same 428 LDP Identifier) established. 430 This means that only one transport connection is established 431 regardless of IPv6 or/and IPv4 Hello adjacencies presence 432 between two LSRs. 434 7. A Dual-stack LSR MUST prefer establishing LDPoIPv6 session with 435 a remote LSR by following the 'transport connection role' 436 determination logic in section 6.1.1. 438 8. A Single-stack LSR MUST establish LDPoIPv4 or LDPoIPv6 session 439 with a remote LSR as per the enabled address-family. 441 6.1.1. Determining Transport connection Roles 443 Section 2.5.2 of [RFC5036] specifies the rules for determining 444 active/passive roles in setting up TCP connection. These rules are 445 clear for a Single-stack LDP, but not for a Dual-stack LDP, in which 446 an LSR may assume different roles for different address families, 447 causing LDP session to not get established. 449 To ensure deterministic transport connection (active/passive) role 450 in case of Dual-stack LDP, this document specifies that the Dual- 451 stack LSR convey its transport connection preference in every LDP 452 Hello message. This preference is encoded in a new TLV, named Dual- 453 stack capability TLV, as defined below: 455 0 1 2 3 456 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 9 0 1 457 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 458 |1|0| Dual-stack capability | Length | 459 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 460 |TR | Reserved | MBZ | 461 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 463 Figure 5 Dual-stack capability TLV 465 Where: 467 U and F bits: 1 and 0 (as specified by RFC5036) 469 Dual-stack capability: TLV code point (to be assigned by IANA). 471 TR, Transport Connection Preference. 473 This document defines the following 2 values: 475 0100: LDPoIPv4 connection 477 0110: LDPoIPv6 connection 479 Reserved 481 This field is reserved. It MUST be set to zero on 482 transmission and ignored on receipt. 484 A Dual-stack LSR MUST include "Dual-stack capability" TLV in all of 485 its LDP Hellos, and MUST set the "TR" field to announce its 486 preference for either LDPoIPv4 or LDPoIPv6 transport connection. The 487 default preference is LDPoIPv6. 489 Upon receiving the hello messages from the neighbor, a Dual-stack 490 LSR MUST check for the presence of "Dual-stack capability" TLV and 491 take appropriate actions as follows: 493 1. If "Dual-stack capability" TLV is present and remote preference 494 does not match with the local preference, then the LSR MUST 495 discard the hello message and log an error. 497 If LDP session was already in place, then LSR MUST send a fatal 498 Notification message with status code [Transport Connection 499 mismatch, IANA allocation TBD] and reset the session. 501 2. If "Dual-stack capability" TLV is present, and remote 502 preference matches with the local preference, then: 504 a) If TR=0100 (LDPoIPv4), then determine the active/passive 505 roles for TCP connection using IPv4 transport address as 506 defined in section 2.5.2 of RFC 5036. 508 b) If TR=0110 (LDPoIPv6), then determine the active/passive 509 roles for TCP connection by using IPv6 transport address 510 as defined in section 2.5.2 of RFC 5036. 512 3. If "Dual-stack capability" TLV is NOT present, and 513 a) Only IPv4 hellos are received, then the neighbor is deemed 514 as a legacy IPv4-only LSR (supporting Single-stack LDP), 515 hence, an LDPoIPv4 session SHOULD be established (similar 516 to that of 2a above). 518 However, if IPv6 hellos are also received at any time from 519 that neighbor, then the neighbor is deemed as a non- 520 compliant Dual-stack LSR (similar to that of 3c below), 521 resulting in any established LDPoIPv4 session being reset 522 and a fatal Notification message being sent (with status 523 code of 'Dual-Stack Non-Compliance', IANA allocation TBD). 525 b) Only IPv6 hellos are received, then the neighbor is deemed 526 as an IPv6-only LSR (supporting Single-stack LDP) and 527 LDPoIPv6 session SHOULD be established (similar to that of 528 2b above). 530 However, if IPv4 hellos are also received at any time from 531 that neighbor, then the neighbor is deemed as a non- 532 compliant Dual-stack LSR (similar to that of 3c below), 533 resulting in any established LDPoIPv6 session being reset 534 and a fatal Notification message being sent (with status 535 code of 'Dual-Stack Non-Compliance', IANA allocation TBD). 537 c) Both IPv4 and IPv6 hellos are received, then the neighbor 538 is deemed as a non-compliant Dual-stack neighbor, and is 539 not allowed to have any LDP session. 541 An LSR MUST convey the same transport connection preference ("TR" 542 field value) in all (link and targeted) Hellos that advertise the 543 same label space to the same peer and/or on same interface. This 544 ensures that two LSRs linked by multiple Hello adjacencies using the 545 same label spaces play the same connection establishment role for 546 each adjacency. 548 An implementation may provide an option to favor one AFI (IPv4, say) 549 over another AFI (IPv6, say) for the TCP transport connection, so as 550 to use the favored IP version for the LDP session, and force 551 deterministic active/passive roles. 553 Note - An alternative to this new Capability TLV could be a new Flag 554 value in LDP Hello message, however, it will get used even in a 555 Single-stack IPv6 LDP networks and linger on forever, even though 556 Dual-stack will not. Hence, this alternative is discarded. 558 6.2. LDP Sessions Maintenance 560 This document specifies that two LSRs maintain a single LDP session 561 regardless of number of Link or Targeted Hello adjacencies between 562 them, as described in section 6.1. This is independent of whether: 564 - they are connected via a Dual-stack LDP enabled interface(s) or 565 via two (or more) Single-stack LDP enabled interfaces; 566 - a Single-stack LDP enabled interface is converted to a Dual-stack 567 LDP enabled interface (e.g. figure 1) on either LSR; 568 - an additional Single-stack or Dual-stack LDP enabled interface is 569 added or removed between two LSRs (e.g. figure 2). 571 The procedures defined in section 6.1 SHOULD result in setting up 572 the LDP session in preferred AFI only after the loss of an existing 573 LDP session (because of link failure, node failure, reboot etc.). 575 If the last hello adjacency for a given address family goes down 576 (e.g. due to Dual-stack LDP enabled interfaces being converted into 577 a Single-stack LDP enabled interfaces on one LSR etc.), and that 578 address family is the same as the one used in the transport 579 connection, then the transport connection (LDP session) MUST be 580 reset. Otherwise, the LDP session MUST stay intact. 582 If the LDP session is torn down for whatever reason (LDP disabled 583 for the corresponding transport, hello adjacency expiry, preference 584 mismatch etc.), then the LSRs SHOULD initiate establishing a new LDP 585 session as per the procedures described in section 6.1 of this 586 document. 588 7. Binding Distribution 590 LSRs by definition can be enabled for Dual-stack LDP globally and/or 591 per peer so as to exchange the address and label bindings for both 592 IPv4 and IPv6 address-families, independent of LDPoIPv4 or LDPoIPV6 593 session between them. 595 However, there might be some legacy LSRs that are fully compliant 596 with RFC 5036 for IPv4, but non-compliant for IPv6 (say, section 597 3.5.5.1 of RFC 5036), causing them to reset the session upon 598 receiving IPv6 address bindings or IPv6 FEC (Prefix) label bindings. 599 This is somewhat undesirable, as clarified further Appendix A.1 and 600 A.2. 602 To help maintain backward compatibility (accommodate IPv4-only LDP 603 implementations that may not be compliant with RFC 5036 section 604 3.5.5.1), this specification requires that an LSR MUST NOT send any 605 IPv6 bindings to a peer if peer has been determined as a legacy LSR. 607 The 'Dual-stack capability' TLV, which is defined in section 6.1.1, 608 is also used to determine if a peer is a legacy (IPv4-only Single- 609 stack) LSR or not. 611 7.1. Address Distribution 613 An LSR MUST NOT advertise (via ADDRESS message) any IPv4-mapped IPv6 614 addresses (defined in section 2.5.5.2 of [RFC4291]), and ignore such 615 addresses, if ever received. Please see Appendix A.3. 617 If an LSR is enabled with Dual-stack LDP for a peer and 619 1. Is NOT able to find the Dual-stack capability TLV in the 620 incoming IPv4 LDP hello messages from that peer, then the LSR 621 MUST NOT advertise its local IPv6 Addresses to the peer. 623 2. Is able to find the Dual-stack capability in the incoming IPv4 624 (or IPv6) LDP Hello messages from that peer, then it MUST 625 advertise (via ADDRESS message) its local IPv4 and IPv6 626 addresses to that peer. 628 3. Is NOT able to find the Dual-stack capability in the incoming 629 IPv6 LDP Hello messages, then it MUST advertise (via ADDRESS 630 message) only its local IPv6 addresses to that peer. 632 The last point helps to maintain forward compatibility (no need 633 to require this TLV in case of IPv6 Single-stack LDP). 635 If an LSR is enabled with Single-stack LDP for any peer, then it 636 MUST advertise (via ADDRESS message) its local IP addresses as per 637 the enabled address family, and accept received Address messages 638 containing IP addresses as per the enabled address family. 640 7.2. Label Distribution 642 An LSR MUST NOT allocate and MUST NOT advertise FEC-Label bindings 643 for link-local or IPv4-mapped IPv6 addresses (defined in section 644 2.5.5.2 of [RFC4291]), and ignore such bindings, if ever received. 645 Please see Appendix A.3. 647 If an LSR enabled with Dual-stack LDP for a peer and 649 1. Is NOT able to find the Dual-stack capability TLV in the 650 incoming IPv4 LDP hello messages from that peer, then the LSR 651 MUST NOT advertise IPv6 FEC-label bindings to the peer. 653 2. Is able to find the Dual-stack capability in the incoming IPv4 654 (or IPv6) LDP Hello messages from that peer, then it MUST 655 advertise FEC-Label bindings for both IPv4 and IPv6 address 656 families to that peer. 658 3. Is NOT able to find the Dual-stack capability in the incoming 659 IPv6 LDP Hello messages, then it MUST advertise FEC-Label 660 bindings for IPv6 address families to that peer. 662 The last point helps to maintain forward compatibility (no need 663 to require this TLV for IPv6 Single-stack LDP). 665 If an LSR is enabled with Single-stack LDP for any peer, then it 666 MUST advertise (via ADDRESS message) FEC-Label bindings for the 667 enabled address family, and accept FEC-Label bindings for the 668 enabled address family. 670 An LSR MAY further constrain the advertisement of FEC-label bindings 671 for a particular address family by negotiating the IP Capability for 672 a given address family, as specified in [IPPWCap] document. This 673 allows an LSR pair to neither advertise nor receive the undesired 674 FEC-label bindings on a per address family basis to a peer. 676 If an LSR is configured to change an interface or peer from Single- 677 stack LDP to Dual-stack LDP, then an LSR SHOULD use Typed Wildcard 678 FEC procedures [RFC5918] to request the label bindings for the 679 enabled address family. This helps to relearn the label bindings 680 that may have been discarded before without resetting the session. 682 8. LDP Identifiers and Duplicate Next Hop Addresses 684 RFC5036 section 2.7 specifies the logic for mapping the IP routing 685 next-hop (of a given FEC) to an LDP peer so as to find the correct 686 label entry for that FEC. The logic involves using the IP routing 687 next-hop address as an index into the (peer Address) database (which 688 is populated by the Address message containing mapping between each 689 peer's local addresses and its LDP Identifier) to determine the LDP 690 peer. 692 However, this logic is insufficient to deal with duplicate IPv6 693 (link-local) next-hop addresses used by two or more peers. The 694 reason is that all interior IPv6 routing protocols (can) use link- 695 local IPv6 addresses as the IP routing next-hops, and 'IPv6 696 Addressing Architecture [RFC4291]' allows a link-local IPv6 address 697 to be used on more than one links. 699 Hence, this logic is extended by this specification to use not only 700 the IP routing next-hop address, but also the IP routing next-hop 701 interface to uniquely determine the LDP peer(s). The next-hop 702 address-based LDP peer mapping is to be done through LDP peer 703 address database (populated by Address messages received from the 704 LDP peers), whereas next-hop interface-based LDP peer mapping is to 705 be done through LDP hello adjacency/interface database (populated by 706 hello messages received from the LDP peers). 708 This extension solves the problem of two or more peers using the 709 same link-local IPv6 address (in other words, duplicate peer 710 addresses) as the IP routing next-hops. 712 Lastly, for better scale and optimization, an LSR may advertise only 713 the link-local IPv6 addresses in the Address message, assuming that 714 the peer uses only the link-local IPv6 addresses as static and/or 715 dynamic IP routing next-hops. 717 9. LDP TTL Security 719 This document recommends enabling Generalized TTL Security Mechanism 720 (GTSM) for LDP, as specified in [RFC6720], for the LDP/TCP transport 721 connection over IPv6 (i.e. LDPoIPv6). The GTSM inclusion is intended 722 to automatically protect IPv6 LDP peering session from off-link 723 attacks. 725 [RFC6720] allows for the implementation to statically 726 (configuration) and/or dynamically override the default behavior 727 (enable/disable GTSM) on a per-peer basis. Suffice to say that such 728 an option could be set on either LSR (since GTSM negotiation would 729 ultimately disable GTSM between LSR and its peer(s)). 731 LDP Link Hello packets MUST have their IPv6 Hop Limit set to 255, 732 and be checked for the same upon receipt before any further 733 processing, as per section 3 of [RFC5082]. 735 10. IANA Considerations 737 This document defines a new optional parameter for the LDP Hello 738 Message and two new status codes for the LDP Notification Message. 740 The 'Dual-Stack capability' parameter requires a code point from the 741 TLV Type Name Space. [RFC5036] partitions the TLV Type Name Space 742 into 3 regions: IETF Consensus region, First Come First Served 743 region, and Private Use region. The authors recommend that a code 744 point from the IETF Consensus range be assigned to the 'Dual-Stack 745 capability' TLV. 747 The 'Transport Connection Mismatch' status code requires a code 748 point from the Status Code Name Space. [RFC5036] partitions the 749 Status Code Name Space into 3 regions: IETF Consensus region, First 750 Come First Served region, and Private Use region. The authors 751 recommend that a code point from the IETF Consensus range be 752 assigned to the 'Transport Connection Mismatch ' status code. 754 The 'Dual-Stack Non-Compliance' status code requires a code point 755 from the Status Code Name Space. [RFC5036] partitions the Status 756 Code Name Space into 3 regions: IETF Consensus region, First Come 757 First Served region, and Private Use region. The authors recommend 758 that a code point from the IETF Consensus range be assigned to the 759 'Dual-Stack Non-Compliance' status code. 761 11. Security Considerations 763 The extensions defined in this document only clarify the behavior of 764 LDP, they do not define any new protocol procedures. Hence, this 765 document does not add any new security issues to LDP. 767 While the security issues relevant for the [RFC5036] are relevant 768 for this document as well, this document reduces the chances of off- 769 link attacks when using IPv6 transport connection by including the 770 use of GTSM procedures [RFC5082]. Please see section 9 for LDP TTL 771 Security details. 773 Moreover, this document allows the use of IPsec [RFC4301] for IPv6 774 protection, hence, LDP can benefit from the additional security as 775 specified in [RFC7321] as well as [RFC5920]. 777 12. Acknowledgments 779 We acknowledge the authors of [RFC5036], since some text in this 780 document is borrowed from [RFC5036]. 782 Thanks to Bob Thomas for providing critical feedback to improve this 783 document early on. 785 Many thanks to Eric Rosen, Lizhong Jin, Bin Mo, Mach Chen, Shane 786 Amante, Pranjal Dutta, Mustapha Aissaoui, Matthew Bocci, Mark Tinka, 787 Tom Petch, Kishore Tiruveedhula, Manoj Dutta, Vividh Siddha, Qin Wu, 788 Simon Perreault, Brian E Carpenter, Santosh Esale, Danial Johari and 789 Loa Andersson for thoroughly reviewing this document, and providing 790 insightful comments and multiple improvements. 792 This document was prepared using 2-Word-v2.0.template.dot. 794 13. Additional Contributors 796 The following individuals contributed to this document: 798 Kamran Raza 799 Cisco Systems, Inc. 800 2000 Innovation Drive 801 Kanata, ON K2K-3E8, Canada 802 Email: skraza@cisco.com 804 Nagendra Kumar 805 Cisco Systems, Inc. 806 SEZ Unit, Cessna Business Park, 807 Bangalore, KT, India 808 Email: naikumar@cisco.com 810 Andre Pelletier 811 Cisco Systems, Inc. 812 2000 Innovation Drive 813 Kanata, ON K2K-3E8, Canada 814 Email: apelleti@cisco.com 816 14. References 818 14.1. Normative References 820 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 821 Requirement Levels", BCP 14, RFC 2119, March 1997. 823 [RFC4291] Hinden, R. and S. Deering, "Internet Protocol Version 6 824 (IPv6) Addressing Architecture", RFC 4291, February 2006. 826 [RFC5036] Andersson, L., Minei, I., and Thomas, B., "LDP 827 Specification", RFC 5036, October 2007. 829 [RFC5082] Pignataro, C., Gill, V., Heasley, J., Meyer, D., and 830 Savola, P., "The Generalized TTL Security Mechanism 831 (GTSM)", RFC 5082, October 2007. 833 [RFC5918] Asati, R., Minei, I., and Thomas, B., "Label Distribution 834 Protocol (LDP) 'Typed Wildcard Forward Equivalence Class 835 (FEC)", RFC 5918, October 2010. 837 14.2. Informative References 839 [RFC4301] Kent, S. and K. Seo, "Security Architecture and Internet 840 Protocol", RFC 4301, December 2005. 842 [RFC7321] Manral, V., "Cryptographic Algorithm Implementation 843 Requirements for Encapsulating Security Payload (ESP) and 844 Authentication Header (AH)", RFC 7321, April 2007. 846 [RFC5920] Fang, L., "Security Framework for MPLS and GMPLS 847 Networks", RFC 5920, July 2010. 849 [RFC4798] De Clercq, et al., "Connecting IPv6 Islands over IPv4 MPLS 850 Using IPv6 Provider Edge Routers (6PE)", RFC 4798, 851 February 2007. 853 [IPPWCap] Raza, K., "LDP IP and PW Capability", draft-ietf-mpls-ldp- 854 ip-pw-capability, June 2011. 856 [RFC5340] Coltun, R., Ferguson, D., Moy, J., and A. Lindem, "OSPF 857 for IPv6", RFC 5340, July 2008. 859 [RFC6286] E. Chen, and J. Yuan, "Autonomous-System-Wide Unique BGP 860 Identifier for BGP-4", RFC 6286, June 2011. 862 [RFC6720] R. Asati, and C. Pignataro, "The Generalized TTL Security 863 Mechanism (GTSM) for the Label Distribution Protocol 864 (LDP)", RFC 6720, August 2012. 866 [RFC4038] M-K. Shin, Y-G. Hong, J. Hagino, P. Savola, and E. M. 867 Castro, "Application Aspects of IPv6 Transition", RFC 868 4038, March 2005. 870 Appendix A. 872 A.1. LDPv6 and LDPv4 Interoperability Safety Net 874 It is not safe to assume that RFC5036 compliant implementations have 875 supported handling IPv6 address family (IPv6 FEC label) in Label 876 Mapping message all along. 878 If a router upgraded with this specification advertised both IPv4 879 and IPv6 FECs in the same label mapping message, then an IPv4-only 880 peer (not knowing how to process such a message) may abort 881 processing the entire label mapping message (thereby discarding even 882 the IPv4 label FECs), as per the section 3.4.1.1 of RFC5036. 884 This would result in LDPv6 to be somewhat undeployable in existing 885 production networks. 887 The change proposed in section 8 of this document provides a good 888 safety net and makes LDPv6 incrementally deployable without making 889 any such assumption on the routers' support for IPv6 FEC processing 890 in current production networks. 892 A.2. Accommodating Non-RFC5036-compliant implementations 894 It is not safe to assume that implementations have been RFC5036 895 compliant in gracefully handling IPv6 address family (IPv6 Address 896 List TLV) in Address message all along. 898 If a router upgraded with this specification advertised IPv6 899 addresses (with or without IPv4 addresses) in Address message, then 900 an IPv4-only peer (not knowing how to process such a message) may 901 not follow section 3.5.5.1 of RFC5036, and tear down the LDP 902 session. 904 This would result in LDPv6 to be somewhat undeployable in existing 905 production networks. 907 The change proposed in section 7 of this document provides a good 908 safety net and makes LDPv6 incrementally deployable without making 909 any such assumption on the routers' support for IPv6 FEC processing 910 in current production networks. 912 A.3. Why prohibit IPv4-mapped IPv6 addresses in LDP 914 Per discussion with 6MAN and V6OPS working groups, the overwhelming 915 consensus was to not promote IPv4-mapped IPv6 addresses appear in 916 the routing table, as well as in LDP (address and label) databases. 918 Also, [RFC4038] section 4.2 suggests that IPv4-mapped IPv6 addressed 919 packets should never appear on the wire. 921 A.4. Why 32-bit value even for IPv6 LDP Router ID 923 The first four octets of the LDP identifier, the 32-bit LSR Id (e.g. 924 (i.e. LDP Router Id), identify the LSR and is a globally unique 925 value within the MPLS network. This is regardless of the address 926 family used for the LDP session. 928 Please note that 32-bit LSR Id value would not map to any IPv4- 929 address in an IPv6 only LSR (i.e., single stack), nor would there be 930 an expectation of it being IP routable, nor DNS-resolvable. In IPv4 931 deployments, the LSR Id is typically derived from an IPv4 address, 932 generally assigned to a loopback interface. In IPv6 only 933 deployments, this 32-bit LSR Id must be derived by some other means 934 that guarantees global uniqueness within the MPLS network, similar 935 to that of BGP Identifier [RFC6286] and OSPF router ID [RFC5340]. 937 This document reserves 0.0.0.0 as the LSR Id, and prohibits its 938 usage with IPv6, in line with OSPF router Id in OSPF version 3 939 [RFC5340]. 941 Author's Addresses 943 Vishwas Manral 944 Hewlet-Packard, Inc. 945 19111 Pruneridge Ave., Cupertino, CA, 95014 946 Phone: 408-447-1497 947 Email: vishwas.manral@hp.com 949 Rajiv Papneja 950 Huawei Technologies 951 2330 Central Expressway 952 Santa Clara, CA 95050 953 Phone: +1 571 926 8593 954 EMail: rajiv.papneja@huawei.com 956 Rajiv Asati 957 Cisco Systems, Inc. 958 7025 Kit Creek Road 959 Research Triangle Park, NC 27709-4987 960 Email: rajiva@cisco.com 962 Carlos Pignataro 963 Cisco Systems, Inc. 964 7200 Kit Creek Road 965 Research Triangle Park, NC 27709-4987 966 Email: cpignata@cisco.com