idnits 2.17.00 (12 Aug 2021) /tmp/idnits57077/draft-ietf-rtgwg-ipfrr-spec-base-09.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 16. -- Found old boilerplate from RFC 3978, Section 5.5, updated by RFC 4748 on line 1319. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1330. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1337. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1343. 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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (Sept 18, 2007) is 5632 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. 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 normative reference: RFC 2740 (Obsoleted by RFC 5340) ** Obsolete normative reference: RFC 3036 (Obsoleted by RFC 5036) == Outdated reference: draft-ietf-isis-link-attr has been published as RFC 5029 == Outdated reference: draft-ietf-isis-wg-multi-topology has been published as RFC 5120 == Outdated reference: draft-ietf-ospf-ospfv3-update has been published as RFC 5340 == Outdated reference: draft-ietf-rtgwg-ipfrr-framework has been published as RFC 5714 -- Obsolete informational reference (is this intentional?): RFC 2966 (Obsoleted by RFC 5302) -- Obsolete informational reference (is this intentional?): RFC 3137 (Obsoleted by RFC 6987) -- Obsolete informational reference (is this intentional?): RFC 4205 (Obsoleted by RFC 5307) Summary: 3 errors (**), 0 flaws (~~), 5 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Routing Area Working Group A. Atlas, Ed. 3 Internet-Draft Google, Inc. 4 Intended status: Standards Track A. Zinin, Ed. 5 Expires: March 21, 2008 Alcatel 6 Sept 18, 2007 8 Basic Specification for IP Fast-Reroute: Loop-free Alternates 9 draft-ietf-rtgwg-ipfrr-spec-base-09 11 Status of this Memo 13 By submitting this Internet-Draft, each author represents that any 14 applicable patent or other IPR claims of which he or she is aware 15 have been or will be disclosed, and any of which he or she becomes 16 aware will be disclosed, in accordance with Section 6 of BCP 79. 18 Internet-Drafts are working documents of the Internet Engineering 19 Task Force (IETF), its areas, and its working groups. Note that 20 other groups may also distribute working documents as Internet- 21 Drafts. 23 Internet-Drafts are draft documents valid for a maximum of six months 24 and may be updated, replaced, or obsoleted by other documents at any 25 time. It is inappropriate to use Internet-Drafts as reference 26 material or to cite them other than as "work in progress." 28 The list of current Internet-Drafts can be accessed at 29 http://www.ietf.org/ietf/1id-abstracts.txt. 31 The list of Internet-Draft Shadow Directories can be accessed at 32 http://www.ietf.org/shadow.html. 34 This Internet-Draft will expire on March 21, 2008. 36 Copyright Notice 38 Copyright (C) The IETF Trust (2007). 40 Abstract 42 This document describes the use of loop-free alternates to provide 43 local protection for unicast traffic in pure IP and MPLS/LDP networks 44 in the event of a single failure, whether link, node or shared risk 45 link group (SRLG). The goal of this technology is to reduce the 46 micro-looping and packet loss that happens while routers converge 47 after a topology change due to a failure. Rapid failure repair is 48 achieved through use of precalculated backup next-hops that are loop- 49 free and safe to use until the distributed network convergence 50 process completes. 52 Table of Contents 54 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 55 1.1. Failure Scenarios . . . . . . . . . . . . . . . . . . . . 5 56 2. Applicability of Described Mechanisms . . . . . . . . . . . . 8 57 3. Alternate Next-Hop Calculation . . . . . . . . . . . . . . . . 9 58 3.1. Basic Loop-free Condition . . . . . . . . . . . . . . . . 10 59 3.2. Node-Protecting Alternate Next-Hops . . . . . . . . . . . 10 60 3.3. Broadcast and NBMA Links . . . . . . . . . . . . . . . . . 11 61 3.4. ECMP and Alternates . . . . . . . . . . . . . . . . . . . 12 62 3.5. Interactions with ISIS Overload, RFC 3137 and Costed 63 Out Links . . . . . . . . . . . . . . . . . . . . . . . . 13 64 3.5.1. Interactions with ISIS Link Attributes . . . . . . . . 14 65 3.6. Selection Procedure . . . . . . . . . . . . . . . . . . . 14 66 3.7. A Simplification: Per-Next-Hop LFAs . . . . . . . . . . . 18 67 4. Using an Alternate . . . . . . . . . . . . . . . . . . . . . . 19 68 4.1. Terminating Use of Alternate . . . . . . . . . . . . . . . 19 69 5. Requirements on LDP Mode . . . . . . . . . . . . . . . . . . . 21 70 6. Routing Aspects . . . . . . . . . . . . . . . . . . . . . . . 21 71 6.1. Multi-Homed Prefixes . . . . . . . . . . . . . . . . . . . 21 72 6.2. ISIS . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 73 6.3. OSPF . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 74 6.3.1. OSPF External Routing . . . . . . . . . . . . . . . . 23 75 6.3.2. OSPF Multi-Topology . . . . . . . . . . . . . . . . . 23 76 6.4. BGP Next-Hop Synchronization . . . . . . . . . . . . . . . 24 77 6.5. Multicast Considerations . . . . . . . . . . . . . . . . . 24 78 7. Security Considerations . . . . . . . . . . . . . . . . . . . 24 79 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 24 80 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 25 81 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 25 82 10.1. Normative References . . . . . . . . . . . . . . . . . . . 25 83 10.2. Informative References . . . . . . . . . . . . . . . . . . 25 84 Appendix A. OSPF Example Where LFA Based on Local Area 85 Topology is Insufficient . . . . . . . . . . . . . . 26 86 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 27 87 Intellectual Property and Copyright Statements . . . . . . . . . . 30 89 1. Introduction 91 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 92 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 93 document are to be interpreted as described in RFC 2119. [RFC2119] 95 Applications for interactive multimedia services such as VoIP and 96 pseudo-wires can be very sensitive to traffic loss, such as occurs 97 when a link or router in the network fails. A router's convergence 98 time is generally on the order of hundreds of milliseconds; the 99 application traffic may be sensitive to losses greater than tens of 100 milliseconds. 102 As discussed in [I-D.ietf-rtgwg-ipfrr-framework], minimizing traffic 103 loss requires a mechanism for the router adjacent to a failure to 104 rapidly invoke a repair path, which is minimally affected by any 105 subsequent re-convergence. This specification describes such a 106 mechanism which allows a router whose local link has failed to 107 forward traffic to a pre-computed alternate until the router installs 108 the new primary next-hops based upon the changed network topology. 109 The terminology used in this specification is given in 110 [I-D.ietf-rtgwg-ipfrr-framework]. The described mechanism assumes 111 that routing in the network is performed using a link-state routing 112 protocol-- OSPF [RFC2328] [RFC2740] [I-D.ietf-ospf-ospfv3-update] or 113 ISIS [RFC1195] [RFC2966] (for IPv4 or IPv6). The mechanism also 114 assumes that both the primary path and the alternate path are in the 115 same routing area. 117 When a local link fails, a router currently must signal the event to 118 its neighbors via the IGP, recompute new primary next-hops for all 119 affected prefixes, and only then install those new primary next-hops 120 into the forwarding plane. Until the new primary next-hops are 121 installed, traffic directed towards the affected prefixes is 122 discarded. This process can take hundreds of milliseconds. 124 <-- 125 +-----+ 126 /------| S |--\ 127 / +-----+ \ 128 / 5 8 \ 129 / \ 130 +-----+ +-----+ 131 | E | | N_1 | 132 +-----+ +-----+ 133 \ / 134 \ \ 4 3 / / 135 \| \ / |/ 136 -+ \ +-----+ / +- 137 \---| D |---/ 138 +-----+ 140 Figure 1: Basic Topology 142 The goal of IP Fast-Reroute is to reduce failure reaction time to 10s 143 of milliseconds by using a pre-computed alternate next-hop, in the 144 event that the currently selected primary next-hop fails, so that the 145 alternate can be rapidly used when the failure is detected. A 146 network with this feature experiences less traffic loss and less 147 micro-looping of packets than a network without IPFRR. There are 148 cases where micro-looping is still a possibility since IPFRR coverage 149 varies but in the worst possible situation a network with IPFRR is 150 equivalent with respect to traffic convergence to a network without 151 IPFRR. 153 To clarify the behavior of IP Fast-Reroute, consider the simple 154 topology in Figure 1. When router S computes its shortest path to 155 router D, router S determines to use the link to router E as its 156 primary next-hop. Without IP Fast-Reroute, that link is the only 157 next-hop that router S computes to reach D. With IP Fast-Reroute, S 158 also looks for an alternate next-hop to use. In this example, S 159 would determine that it could send traffic destined to D by using the 160 link to router N_1 and therefore S would install the link to N_1 as 161 its alternate next-hop. At some later time, the link between router 162 S and router E could fail. When that link fails, S and E will be the 163 first to detect it. On detecting the failure, S will stop sending 164 traffic destined for D towards E via the failed link, and instead 165 send the traffic to S's pre-computed alternate next-hop, which is the 166 link to N_1, until a new SPF is run and its results are installed. 167 As with the primary next-hop, an alternate next-hop is computed for 168 each destination. The process of computing an alternate next-hop 169 does not alter the primary next-hop computed via a standard SPF. 171 If in the example of Figure 1, the link cost from N_1 to D increased 172 to 30 from 3, then N_1 would not be a loop-free alternate, because 173 the cost of the path from N_1 to D via S would be 17 while the cost 174 from N_1 directly to D would be 30. In real networks, we may often 175 face this situation. The existence of a suitable loop-free alternate 176 next-hop is dependent on the topology and the nature of the failure 177 the alternate is calculated for. 179 This specification uses the terminology introduced in 180 [I-D.ietf-rtgwg-ipfrr-framework]. In particular, it uses 181 Distance_opt(X,Y), abbreviated to D_opt(X,Y), to indicate the 182 shortest distance from X to Y. S is used to indicate the calculating 183 router. N_i is a neighbor of S; N is used as an abbreviation when 184 only one neighbor is being discussed. D is the destination under 185 consideration. 187 A neighbor N can provide a loop-free alternate (LFA) if and only if 189 Distance_opt(N, D) < Distance_opt(N, S) + Distance_opt(S, D) 191 Inequality 1: Loop-Free Criterion 193 A sub-set of loop-free alternate are downstream paths which must meet 194 a more restrictive condition that is applicable to more complex 195 failure scenarios: 197 Distance_opt(N, D) < Distance_opt(S, D) 199 Inequality 2: Downstream Path Criterion 201 1.1. Failure Scenarios 203 The alternate next-hop can protect against a single link failure, a 204 single node failure, one or more shared risk link group failures, or 205 a combination of these. Whenever a failure occurs that is more 206 extensive than what the alternate was intended to protect, there is 207 the possibility of temporarily looping traffic (note again, that such 208 a loop would only last until the next complete SPF calculation). The 209 example where a node fails when the alternate provided only link 210 protection is illustrated below. If unexpected simultaneous failures 211 occur, then micro-looping may occur since the alternates are not pre- 212 computed to avoid the set of failed links. 214 If only link protection is provided and the node fails, it is 215 possible for traffic using the alternates to experience micro- 216 looping. This issue is illustrated in Figure 2. If Link(S->E) 217 fails, then the link-protecting alternate via N will work correctly. 218 However, if router E fails, then both S and N will detect a failure 219 and switch to their alternates. In this example, that would cause S 220 to redirect the traffic to N and N to redirect the traffic to S and 221 thus causing a forwarding loop. Such a scenario can arise because 222 the key assumption, that all other routers in the network are 223 forwarding based upon the shortest path, is violated because of a 224 second simultaneous correlated failure - another link connected to 225 the same primary neighbor. If there are not other protection 226 mechanisms a node failure is still a concern when only using link 227 protection. 229 <@@@ 230 @@@> 231 +-----+ +-----+ 232 | S |-------| N | 233 +-+---+ 5 +-----+ 234 | | 235 | 5 4 | | 236 | | | \|/ 237 \|/ | | 238 | +-----+ | 239 +----| E |---+ 240 +--+--+ 241 | 242 | 243 | 10 244 | 245 +--+--+ 246 | D | 247 +-----+ 249 Figure 2: Link-Protecting Alternates Causing Loop on Node Failure 251 Micro-looping of traffic via the alternates caused when a more 252 extensive failure than planned for occurs can be prevented via 253 selection of only downstream paths as alternates. In Figure 2, S 254 would be able to use N as an alternate, but N could not use S; 255 therefore N would have no alternate and would discard the traffic, 256 thus avoiding the micro-loop. A micro-loop due to the use of 257 alternates can be avoided by using downstream paths because each 258 succeeding router in the path to the destination must be closer to 259 the destination than its predecessor (according to the topology prior 260 to the failures). Although use of downstream paths ensures that the 261 micro-looping via alternates does not occur, such a restriction can 262 severely limit the coverage of alternates. 264 As shown above, the use of either a node protecting LFA or a 265 downstream path provides protection against micro-looping in the 266 event of node failure. There are topologies where there may be 267 either a node portecting LFA, a downstream path, both or neither. A 268 node may select either a node protecting LFA or a downstream path 269 without risk of causing micro-loops in the event of node failure. 270 While a link-and-node-protecting LFA guarantees protection against 271 either link or node failure, a downstream path provides protection 272 only against a link failure and may or may not provide protection 273 against a node failure depending on the protection available at the 274 downstream node, but it cannot cause a micro-loop. For example in 275 Figure 2, if S uses N as a downstream path, although no looping can 276 occur, the traffic will not be protected in the event of the failure 277 of node E because N has no viable repair path, and it will simply 278 discard the packet. However, if N had a link and node protecting LFA 279 or downstream path via some other path (not shown), then the repair 280 may succeed. 282 Since the functionality of link and node protecting LFAs is greater 283 than that of downstream paths, a router SHOULD select a link and node 284 protecting LFA over a downstream path. If there are any destinations 285 for which a link and node protecting LFA is not available, then by 286 definition the path to all of those destinations from any neighbor of 287 the computing router (S) must be through the node (E) being protected 288 (otherwise there would be a node protecting LFA for that 289 destination). Consequently, if there exists a downstream path to the 290 protected node as destination, then that downstream path may be used 291 for all those destinations for which a link and node protecting LFA 292 is not available; the existence of a downstream path can be 293 determined by a single check of the condition Distance_opt(N, E) < 294 Distance_opt(S, E). 296 It may be desirable to find an alternate that can protect against 297 other correlated failures (of which node failure is a specific 298 instance). In the general case, these are handled by shared risk 299 link groups (SRLGs) where any links in the network can belong to the 300 SRLG. General SRLGs may add unacceptably to the computational 301 complexity of finding a loop-free alternate. 303 However, a sub-category of SRLGs is of interest and can be applied 304 only during the selection of an acceptable alternate. This sub- 305 category is to express correlated failures of links that are 306 connected to the same router. For example, if there are multiple 307 logical sub-interfaces on the same physical interface, such as VLANs 308 on an Ethernet interface, if multiple interfaces use the same 309 physical port because of channelization, or if multiple interfaces 310 share a correlated failure because they are on the same line-card. 311 This sub-category of SRLGs will be referred to as local-SRLGs. A 312 local-SRLG has all of its member links with one end connected to the 313 same router. Thus, router S could select a loop-free alternate which 314 does not use a link in the same local-SRLG as the primary next-hop. 316 The local-SRLGs belonging to E can be protected against via node- 317 protection; i.e. picking a loop-free node-protecting alternate. 319 Where SRLG protection is provided, it is in the context of the 320 particular OSPF or ISIS area, whose topology is used in the SPF 321 computations to compute the loop-free alternates. If an SRLG 322 contains links in multiple areas, then separate SRLG-protecting 323 alternates would be required in each area that is traversed by the 324 affected traffic. 326 2. Applicability of Described Mechanisms 328 IP Fast Reroute mechanisms described in this memo cover intra-domain 329 routing only, with OSPF[RFC2328] or ISIS [RFC1195] [RFC2966] as the 330 IGP. Specifically, Fast Reroute for BGP inter-domain routing is not 331 part of this specification. 333 Certain aspects of OSPF inter-area routing behavior explained in 334 Section 6.3 and Appendix A impact the ability of the router 335 calculating the backup next-hops to assess traffic trajectories. In 336 order to avoid micro-looping and ensure required coverage, certain 337 constraints are applied to multi-area OSPF networks: 339 a. Loop-free alternates should not be used in the backbone area if 340 there are any virtual links configured unless, for each transit 341 area, there is a full mesh of virtual links between all ABRs in 342 that area. Loop-free alternates may be used in non-backbone 343 areas regardless of whether there are virtual links configured. 345 b. Loop-free alternates should not be used for inter-area routes in 346 an area that contains more than one alternate ABR. [RFC3509] 348 c. Loop-free alternates should not be used for AS External routes or 349 ASBR routes in a non-backbone area of a network where there 350 exists an ABR that is announced as an ASBR in multiple non- 351 backbone areas and there exists another ABR that is in at least 352 two of the same non-backbone areas. 354 d. Loop-free alternates should not be used in a non-backbone area of 355 a network for AS External routes where an AS External prefix is 356 advertised with the same type of external metric by multiple 357 ASBRs, which are in different non-backbone areas, with a 358 forwarding address of 0.0.0.0 or by one or more ASBRs with 359 forwarding addresses in multiple non-backbone areas when an ABR 360 exists simultaneously in two or more of those non-backbone areas. 362 3. Alternate Next-Hop Calculation 364 In addition to the set of primary next-hops obtained through a 365 shortest path tree (SPT) computation that is part of standard link- 366 state routing functionality, routers supporting IP Fast Reroute also 367 calculate a set of backup next hops that are engaged when a local 368 failure occurs. These backup next hops are calculated to provide the 369 required type of protection (i.e. link-protecting and/or node- 370 protecting) and to guarantee that when the expected failure occurs, 371 forwarding traffic through them will not result in a loop. Such next 372 hops are called loop-free alternates or LFAs throughout this 373 specification. 375 In general, to be able to calculate the set of LFAs for a specific 376 destination D, a router needs to know the following basic pieces of 377 information: 379 o Shortest-path distance from the calculating router to the 380 destination (Distance_opt(S, D)) 382 o Shortest-path distance from the router's IGP neighbors to the 383 destination (Distance_opt(N, D)) 385 o Shortest path distance from the router's IGP neighbors to itself 386 (Distance_opt(N, S)) 388 o Distance_opt(S, D) is normally available from the regular SPF 389 calculation performed by the link-state routing protocols. 390 Distance_opt(N, D) and Distance_opt(N, S) can be obtained by 391 performing additional SPF calculations from the perspective of 392 each IGP neighbor (i.e. considering the neighbor's vertex as the 393 root of the SPT--called SPT(N) hereafter--rather than the 394 calculating router's one, called SPT(S)). 396 This specification defines a form of SRLG protection limited to those 397 SRLGs that include a link that the calculating router is directly 398 connected to. Only that set of SRLGs could cause a local failure; 399 the calculating router only computes alternates to handle a local 400 failure. Information about local link SRLG membership is manually 401 configured. Information about remote link SRLG membership may be 402 dynamically obtained using [RFC4205] or [RFC4203]. Define 403 SRLG_local(S) to be the set of SRLGs that include a link that the 404 calculating router S is directly connected to. Only SRLG_local(S) is 405 of interest during the calculation, but the calculating router must 406 correctly handle changes to SRLG_local(S) triggered by local link 407 SRLG membership changes. 409 In order to choose among all available LFAs that provide required 410 SRLG protection for a given destination, the calculating router needs 411 to track the set of SRLGs in SRLG_local(S) that the path through a 412 specific IGP neighbor involves. To do so, each node D in the network 413 topology is associated with SRLG_set(N, D), which is the set of SRLGs 414 that would be crossed if traffic to D was forwarded through N. To 415 calculate this set, the router initializes SRLG_set(N, N) for each of 416 its IGP neighbors to be empty. During the SPT(N) calculation, when a 417 new vertex V is added to the SPT, its SRLG_set(N, V) is set to the 418 union of SRLG sets associated with its parents, and the SRLG sets in 419 SRLG_local(S) that are associated with the links from V's parents to 420 V. The union of the set of SRLG associated with a candidate alternate 421 next-hop and the SRLG_set(N, D) for the neighbor reached via that 422 candidate next-hop is used to determine SRLG protection. 424 The following sections provide information required for calculation 425 of LFAs. Sections Section 3.1 through Section 3.4 define different 426 types of LFA conditions. Section 3.5 describes constraints imposed 427 by the IS-IS overload and OSPF stub router functionality. 428 Section 3.6 defines the summarized algorithm for LFA calculation 429 using the definitions in the previous sections. 431 3.1. Basic Loop-free Condition 433 Alternate next hops used by implementations following this 434 specification MUST conform to at least the loop-freeness condition 435 stated above in Inequality 1. This condition guarantees that 436 forwarding traffic to an LFA will not result in a loop after a link 437 failure. 439 Further conditions may be applied when determining link-protecting 440 and/or node-protecting alternate next-hops as described in Sections 441 Section 3.2 and Section 3.3. 443 3.2. Node-Protecting Alternate Next-Hops 445 For an alternate next-hop N to protect against node failure of a 446 primary neighbor E for destination D, N must be loop-free with 447 respect to both E and D. In other words, N's path to D must not go 448 through E. This is the case if Inequality 3 is true, where N is the 449 neighbor providing a loop-free alternate. 451 Distance_opt(N, D) < Distance_opt(N, E) + Distance_opt(E, D) 453 Inequality 3: Criteria for a Node-Protecting Loop-Free Alternate 455 If Distance_opt(N,D) = Distance_opt(N, E) + Distance_opt(E, D), it is 456 possible that N has equal-cost paths and one of those could provide 457 protection against E's node failure. However, it is equally possible 458 that one of N's paths goes through E, and the calculating router has 459 no way to influence N's decision to use it. Therefore, it SHOULD be 460 assumed that an alternate next-hop does not offer node protection if 461 Inequality 3 is not met. 463 3.3. Broadcast and NBMA Links 465 Verification of the link-protection property of a next hop in the 466 case of a broadcast link is more elaborate than for a point-to-point 467 link. This is because a broadcast link is represented as a pseudo- 468 node with zero-cost links connecting it to other nodes. 470 Because failure of an interface attached to a broadcast segment may 471 mean loss of connectivity of the whole segment, the condition 472 described for broadcast link protection is pessimistic and requires 473 that the alternate is loop-free with regard to the pseudo-node. 474 Consider the example in Figure 3. 476 +-----+ 15 477 | S |-------- 478 +-----+ | 479 | 5 | 480 | | 481 | 0 | 482 /----\ 0 5 +-----+ 483 | PN |-----| N | 484 \----/ +-----+ 485 | 0 | 486 | | 8 487 | 5 | 488 +-----+ 5 +-----+ 489 | E |----| D | 490 +-----+ +-----+ 492 Figure 3: Loop-Free Alternate that is Link-Protecting 494 In Figure 3, N offers a loop-free alternate which is link-protecting. 495 If the primary next-hop uses a broadcast link, then an alternate 496 SHOULD be loop-free with respect to that link's pseudo-node to 497 provide link protection. This requirement is described in 498 Inequality 4 below. 500 D_opt(N, D) < D_opt(N, pseudo) + D_opt(pseudo, D) 502 Inequality 4: Loop-Free Link-Protecting Criterion for Broadcast Links 504 Because the shortest path from the pseudo-node goes through E, if a 505 loop-free alternate from a neighbor N is node-protecting, the 506 alternate will also be link-protecting unless the router S can only 507 reach the alternate neighbor N via the same pseudo-node. Because S 508 can direct the traffic away from the shortest path to use the 509 alternate N, traffic might pass through the same broadcast link as it 510 would when S sent the traffic to the primary E. Thus, an LFA from N 511 that is node-protecting is not automatically link-protecting. 513 To obtain link protection, it is necessary both that the path from 514 the selected alternate next-hop does not traverse the link of 515 interest and that the link used from S to reach that alternate next- 516 hop is not the link of interest. The latter can only occur with non- 517 point-to-point links. Therefore, if the primary next-hop is across a 518 broadcast or NBMA interface, it is necessary to consider link 519 protection during the alternate selection. To clarify, consider the 520 topology in Figure 3. For N to provide link-protection, it is first 521 necessary that N's shortest path to D does not traverse the pseudo- 522 node PN. Second, it is necessary that the alternate next-hop 523 selected by S does not traverse PN. In this example, S's shortest 524 path to N is via the pseudo-node. Thus, to obtain link-protection, S 525 must find a next-hop to N (the point-to-point link from S to N in 526 this example) that avoids the pseudo-node PN. 528 Similar consideration of the link from S to the selected alternate 529 next-hop as well as the path from the selected alternate next-hop is 530 also necessary for SRLG protection. S's shortest path to the 531 selected neighbor N may not be acceptable as an alternate next-hop to 532 provide SRLG protection, even if the path from N to D can provide 533 SRLG protection. 535 3.4. ECMP and Alternates 537 With equal-cost multi-path, a prefix may have multiple primary next- 538 hops that are used to forward traffic. When a particular primary 539 next-hop fails, alternate next-hops should be used to preserve the 540 traffic. These alternate next-hops may themselves also be primary 541 next-hops, but need not be. Other primary next-hops are not 542 guaranteed to provide protection against the failure scenarios of 543 concern. 545 20 L1 L3 3 546 [ N ]----[ S ]--------[ E3 ] 547 | | | 548 | 5 | L2 | 549 20 | | | 550 | --------- | 2 551 | 5 | | 5 | 552 | [ E1 ] [ E2 ]-----| 553 | | | 554 | 10 | 10 | 555 |---[ A ] [ B ] 556 | | 557 2 |--[ D ]-| 2 559 Figure 4: ECMP where Primary Next-Hops Provide Limited Protection 561 In Figure 4 S has three primary next-hops to reach D; these are L2 to 562 E1, L2 to E2 and L3 to E3. The primary next-hop L2 to E1 can obtain 563 link and node protection from L3 to E3, which is one of the other 564 primary next-hops; L2 to E1 cannot obtain link protection from the 565 other primary next-hop L2 to E2. Similarly, the primary next-hop L2 566 to E2 can only get node protection from L2 to E1 and can only get 567 link protection from L3 to E3. The third primary next-hop L3 to E3 568 can obtain link and node protection from L2 to E1 and from L2 to E2. 569 It is possible for both the primary next-hop L2 to E2 and the primary 570 next-hop L2 to E1 to obtain an alternate next-hop that provides both 571 link and node protection by using L1. 573 Alternate next-hops are determined for each primary next-hop 574 separately. As with alternate selection in the non-ECMP case, these 575 alternate next-hops should maximize the coverage of the failure 576 cases. 578 3.5. Interactions with ISIS Overload, RFC 3137 and Costed Out Links 580 As described in [RFC3137], there are cases where it is desirable not 581 to have a router used as a transit node. For those cases, it is also 582 desirable not to have the router used on an alternate path. 584 For computing an alternate, a router MUST NOT use an alternate next- 585 hop that is along a link whose cost or reverse cost is LSInfinity 586 (for OSPF) or the maximum cost (for ISIS) or which has the overload 587 bit set (for ISIS). For a broadcast link, the reverse cost 588 associated with a potential alternate next-hop is the cost towards 589 the pseudo-node advertised by the next-hop router. For point-to- 590 point links, if a specific link from the next-hop router cannot be 591 associated with a particular link, then the reverse cost considered 592 is that of the minimum cost link from the next-hop router back to S. 594 In the case of OSPF, if all links from router S to a neighbor N_i 595 have a reverse cost of LSInfinity, then router S MUST NOT use N_i as 596 an alternate. 598 Similarly in the case of ISIS, if N_i has the overload bit set, then 599 S MUST NOT consider using N_i as an alternate. 601 This preserves the desired behavior of diverting traffic away from a 602 router which is following [RFC3137] and it also preserves the desired 603 behavior when an operator sets the cost of a link to LSInfinity for 604 maintenance which is not permitting traffic across that link unless 605 there is no other path. 607 If a link or router which is costed out was the only possible 608 alternate to protect traffic from a particular router S to a 609 particular destination, then there should be no alternate provided 610 for protection. 612 3.5.1. Interactions with ISIS Link Attributes 614 [I-D.ietf-isis-link-attr] describes several flags whose interactions 615 with LFAs needs to be defined. A router SHOULD NOT specify the 616 "local protection available" flag as a result of having LFAs. A 617 router SHOULD NOT use an alternate next-hop that is along a link for 618 which the link has been advertised with the attribute "link excluded 619 from local protection path" or with the attribute "local maintenance 620 required". 622 3.6. Selection Procedure 624 A router supporting this specification SHOULD attempt to select at 625 least one loop-free alternate next-hop for each primary next-hop used 626 for a given prefix. A router MAY decide to not use an available 627 loop-free alternate next-hop. A reason for such a decision might be 628 that the loop-free alternate next-hop does not provide protection for 629 the failure scenario of interest. 631 The alternate selection should maximize the coverage of the failure 632 cases. 634 When calculating alternate next-hops, the calculating router S 635 applies the following rules. 637 1. S SHOULD select a loop-free node-protecting alternate next-hop, 638 if one is available. If no loop-free node-protecting alternate 639 is available, then S MAY select a loop-free link-protecting 640 alternate. 642 2. If S has a choice between a loop-free link-protecting node- 643 protecting alternate and a loop-free node-protecting alternate 644 which is not link-protecting, S SHOULD select a loop-free node- 645 protecting alternate which is also link-protecting. This can 646 occur as explained in Section 3.3. 648 3. If S has multiple primary next-hops, then S SHOULD select as a 649 loop-free alternate either one of the other primary next-hops or 650 a loop-free node-protecting alternate if available. If no loop- 651 free node-protecting alternate is available and no other primary 652 next-hop can provide link-protection, then S SHOULD select a 653 loop-free link-protecting alternate. 655 4. Implementations SHOULD support a mode where other primary next- 656 hops satisfying the basic loop-free condition and providing at 657 least link or node protection are preferred over any non-primary 658 alternates. This mode is provided to allow the administrator to 659 preserve traffic patterns based on regular ECMP behavior. 661 5. Implementations considering SRLGs MAY use SRLG-protection to 662 determine that a node-protecting or link-protecting alternate is 663 not available for use. 665 Following the above rules maximizes the level of protection and use 666 of primary (ECMP) next-hops. 668 Each next-hop is associated with a set of non-mutually-exclusive 669 characteristics based on whether it is used as a primary next-hop to 670 a particular destination D, and the type of protection it can provide 671 relative to a specific primary next-hop E: 673 Primary Path - The next-hop is used by S as primary. 675 Loop-Free Node-Protecting Alternate - This next-hop satisfies 676 Inequality 1 and Inequality 3. The path avoids S, S's primary 677 neighbor E, and the link from S to E. 679 Loop-Free Link-Protecting Alternate - This next-hop satisfies 680 Inequality 1 but not Inequality 3. If the primary next-hop uses a 681 broadcast link, then this next-hop satisfies Inequality 4. 683 An alternate path may also provide none, some or complete SRLG 684 protection as well as node and link or link protection. For 685 instance, a link may belong to two SRLGs G1 and G2. The alternate 686 path might avoid other links in G1 but not G2, in which case the 687 alternate would only provide partial SRLG protection. 689 Below is an algorithm that can be used to calculate loop-free 690 alternate next-hops. The algorithm is given for informational 691 purposes and implementations are free to use any other algorithm as 692 long as it satisfies the rules described above. 694 The following procedure describes how to select an alternate next- 695 hop. The procedure is described to determine alternate next-hops to 696 use to reach each router in the topology. Prefixes that are 697 advertised by a single router can use the alternate next-hop computed 698 for the router to which they are attached. The same procedure can be 699 used to reach a prefix that is advertised by more than one router 700 when the logical topological transformation described in Section 6.1 701 is used. 703 S is the computing router and has candidate next-hops H_1 through 704 H_k. N_i and N_j are used to refer to neighbors of S. For a next-hop 705 to be a candidate, the next-hop must be associated with a bi- 706 directional link, as is determined by the IGP. For a particular 707 destination router D, let S have already computed D_opt(S, D), and 708 for each neighbor N_i, D_opt(N_i, D), D_opt(N_i, S), and D_opt(N_i, 709 N_j), the distance from N_i to each other neighbor N_j, and the set 710 of SRLGs traversed by the path D_opt(N_i, D). S should follow the 711 below procedure for every primary next-hop selected to reach D. This 712 set of primary next-hops is represented P_1 to P_p. This procedure 713 finds the alternate next-hop(s) for P_i. 715 First, initialize the alternate information for P_i as follows: 717 P_i.alt_next_hops = {} 718 P_i.alt_type = NONE 719 P_i.alt_link-protect = FALSE 720 P_i.alt_node-protect = FALSE 721 P_i.alt_srlg-protect = {} 723 For each candidate next-hop H_h, 725 1. Initialize variables as follows: 727 cand_type = NONE 728 cand_link-protect = FALSE 729 cand_node-protect = FALSE 730 cand_srlg-protect = {} 732 2. If H_h is P_i, skip it and continue to the next candidate next- 733 hop. 735 3. If H_h.link is administratively allowed to be used as an 736 alternate, 737 and the cost of H_h.link is less than the maximum, 738 and the reverse cost of H_h is less than the maximum, 739 and H_h.neighbor is not overloaded (for ISIS), 740 and H_h.link is bi-directional, 742 then H_h can be considered as an alternate. Otherwise, skip it 743 and continue to the next candidate next-hop. 745 4. If D_opt( H_h.neighbor, D) >= D_opt( H_h.neighbor, S) + D_opt(S, 746 D), then H_h is not loop-free. Skip it and continue to the next 747 candidate next-hop. 749 5. cand_type = LOOP-FREE. 751 6. If H_h is a primary next-hop, set cand_type to PRIMARY. 753 7. If H_h.link is not P_i.link, set cand_link-protect to TRUE. 755 8. If D_opt(H_h.neighbor, D) < D_opt(H_h.neighbor, P_i.neighbor) + 756 D_opt(P_i.neighbor, D), set cand_node-protect to TRUE. 758 9. If the router considers SRLGs, then set the cand_srlg-protect to 759 the set of SRLGs traversed on the path from S via P_i to 760 P_i.neighbor. Remove the set of SRLGs to which H_h belongs from 761 cand_srlg-protect. Remove from cand_srlg-protect the set of 762 SRLGs traversed on the path from H_h.neighbor to D. Now 763 cand_srlg-protect holds the set of SRLGs to which P_i belongs 764 and that are not traversed on the path from S via H_h to D. 766 10. If cand_type is PRIMARY, the router prefers other primary next- 767 hops for use as the alternate, and the P_i.alt_type is not 768 PRIMARY, goto Step 19. 770 11. If cand_node-protect is TRUE and P_i.alt_node-protect is FALSE, 771 goto Paragraph 19. 773 12. If cand_link-protect is TRUE and P_i.alt_link-protect is FALSE, 774 goto Step 19. 776 13. If cand_srlg-protect has a better set of SRLGs than 777 P_i.alt_srlg-protect, goto Step 19. 779 14. If cand_srlg-protect is different from P_i.alt_srlg-protect, 780 then select between H_h and P_i.alt_next_hops based upon 781 distance, IP addresses, or any router-local tie-breaker. If H_h 782 is preferred, then goto Step 19. Otherwise, skip H_h and 783 continue to the next candidate next-hop. 785 15. If D_opt(H_h.neighbor, D) < D_opt(P_i.neighbor, D) and 786 D_opt(P_i.alt_next_hops, D) >= D_opt(P_i.neighbor, D), then H_h 787 is a downstream alternate and P_i.alt_next_hops is simply an 788 LFA. Prefer H_h and goto Step 19. 790 16. Based upon the alternate types, the alternate distances, IP 791 addresses, or other tie-breakers, decide if H_h is preferred to 792 P_i.alt_next_hops. If so, goto Step 19. 794 17. Decide if P_i.alt_next_hops is preferred to H_h. If so, then 795 skip H_h and continue to the next candidate next-hop. 797 18. Add H_h into P_i.alt_next_hops. Set P_i.alt_type to the better 798 type of H_h.alt_type and P_i.alt_type. Continue to the next 799 candidate next-hop. 801 19. Replace the P_i alternate next-hop set with H_h as follows: 803 P_i.alt_next_hops = {H_h} 804 P_i.alt_type = cand_type 805 P_i.alt_link-protect = cand_link-protect 806 P_i.alt_node-protect = cand_node-protect 807 P_i.alt_srlg-protect = cand_srlg-protect 809 Continue to the next candidate next-hop. 811 3.7. A Simplification: Per-Next-Hop LFAs 813 It is possible to simplify the computation and use of LFAs when 814 solely link protection is desired by considering and computing only 815 one link-protecting LFA for each next-hop connected to the router. 816 All prefixes that use that next-hop as a primary will use the LFA 817 computed for that next-hop as its LFA. 819 Even a prefix with multiple primary next-hops will have each primary 820 next-hop protected individually by the primary next-hop's associated 821 LFA. That associated LFA might or might not be another of the 822 primary next-hops of the prefix. 824 This simplification may reduce coverage in a network. In addition to 825 limiting protection for multi-homed prefixes (see Section 6.1), the 826 computation per next-hop may also not find an LFA when one could be 827 found for some of the prefixes that use that next-hop. 829 For example, consider Figure 4 where S has 3 ECMP next-hops, E1, E2, 830 and E3 to reach D. For the prefix D, E3 can give link protection for 831 the next-hops E1 and E2; E1 and E2 can give link protection for the 832 next-hops E3. However, if one uses this simplification to compute 833 LFAs for E1, E2 and E3 individually, there is no link-protecting LFA 834 for E1. E3 and E2 can protect each other. 836 4. Using an Alternate 838 If an alternate next-hop is available, the router redirects traffic 839 to the alternate next-hop in case of a primary next-hop failure as 840 follows. 842 When a next-hop failure is detected via a local interface failure or 843 other failure detection mechanisms (see 844 [I-D.ietf-rtgwg-ipfrr-framework]), the router SHOULD: 846 1. Remove the primary next-hop associated with the failure. 848 2. Install the loop-free alternate calculated for the failed next- 849 hop if it is not already installed (e.g. the alternate is also a 850 primary next-hop). 852 Note that the router MAY remove other next-hops if it believes (via 853 SRLG analysis) that they may have been affected by the same failure, 854 even if it is not visible at the time of failure detection. 856 The alternate next-hop MUST be used only for traffic types which are 857 routed according to the shortest path. Multicast traffic is 858 specifically out of scope for this specification. 860 4.1. Terminating Use of Alternate 862 A router MUST limit the amount of time an alternate next-hop is used 863 after the primary next-hop has become unavailable. This ensures that 864 the router will start using the new primary next-hops. It ensures 865 that all possible transient conditions are removed and the network 866 converges according to the deployed routing protocol. 868 A router that implements [I-D.ietf-rtgwg-microloop-analysis] SHOULD 869 follow the rules given there for terminating the use of an alternate. 871 A router that implements [I-D.francois-ordered-fib] SHOULD follow the 872 rules given there for terminating the use of an alternate. 874 It is desirable to avoid micro-forwarding loops involving S. An 875 example illustrating the problem is given in Figure 5. If the link 876 from S to E fails, S will use N1 as an alternate and S will compute 877 N2 as the new primary next-hop to reach D. If S starts using N2 as 878 soon as S can compute and install its new primary, it is probable 879 that N2 will not have yet installed its new primary next-hop. This 880 would cause traffic to loop and be dropped until N2 has installed the 881 new topology. This can be avoided by S delaying its installation and 882 leaving traffic on the alternate next-hop. 884 +-----+ 885 | N2 |-------- | 886 +-----+ 1 | \|/ 887 | | 888 | +-----+ @@> +-----+ 889 | | S |---------| N1 | 890 10 | +-----+ 10 +-----+ 891 | | | 892 | 1 | | | 893 | | \|/ 10 | 894 | +-----+ | | 895 | | E | | \|/ 896 | +-----+ | 897 | | | 898 | 1 | | | 899 | | \|/ | 900 | +-----+ | 901 |----| D |-------------- 902 +-----+ 904 Figure 5: Example where Continued Use of Alternate is Desirable 906 This is an example of a case where the new primary is not a loop-free 907 alternate before the failure and therefore may have been forwarding 908 traffic through S. This will occur when the path via a previously 909 upstream node is shorter than the the path via a loop-free alternate 910 neighbor. In these cases, it is useful to give sufficient time to 911 ensure that the new primary neighbor and other nodes on the new 912 primary path have switched to the new route. 914 If the newly selected primary was loop-free before the failure, then 915 it is safe to switch to that new primary immediately; the new primary 916 wasn't dependent on the failure and therefore its path will not have 917 changed. 919 Given that there is an alternate providing appropriate protection and 920 while the assumption of a single failure holds, it is safe to delay 921 the installation of the new primaries; this will not create 922 forwarding loops because the alternate's path to the destination is 923 known to not go via S or the failed element and will therefore not be 924 affected by the failure. 926 An implementation SHOULD continue to use the alternate next-hops for 927 packet forwarding even after the new routing information is available 928 based on the new network topology. The use of the alternate next- 929 hops for packet forwarding SHOULD terminate: 931 a. if the new primary next-hop was loop-free prior to the topology 932 change, or 934 b. if a configured hold-down, which represents a worst-case bound on 935 the length of the network convergence transition, has expired, or 937 c. if notification of an unrelated topological change in the network 938 is received. 940 5. Requirements on LDP Mode 942 Since LDP [RFC3036] traffic will follow the path specified by the 943 IGP, it is also possible for the LDP traffic to follow the loop-free 944 alternates indicated by the IGP. To do so, it is necessary for LDP 945 to have the appropriate labels available for the alternate so that 946 the appropriate out-segments can be installed in the forwarding plane 947 before the failure occurs. 949 This means that a Label Switched Router (LSR) running LDP must 950 distribute its labels for the FECs it can provide to all its 951 neighbors, regardless of whether or not they are upstream. 952 Additionally, LDP must be acting in liberal label retention mode so 953 that the labels which correspond to neighbors that aren't currently 954 the primary neighbor are stored. Similarly, LDP should be in 955 downstream unsolicited mode, so that the labels for the FEC are 956 distributed other than along the SPT. 958 If these requirements are met, then LDP can use the loop-free 959 alternates without requiring any targeted sessions or signaling 960 extensions for this purpose. 962 6. Routing Aspects 964 6.1. Multi-Homed Prefixes 966 An SPF-like computation is run for each topology, which corresponds 967 to a particular OSPF area or ISIS level. The IGP needs to determine 968 loop-free alternates to multi-homed routes. Multi-homed routes occur 969 for routes obtained from outside the routing domain by multiple 970 routers, for subnets on links where the subnet is announced from 971 multiple ends of the link, and for routes advertised by multiple 972 routers to provide resiliency. 974 Figure 6 demonstrates such a topology. In this example, the shortest 975 path to reach the prefix p is via E. The prefix p will have the link 976 to E as its primary next-hop. If the alternate next-hop for the 977 prefix p is simply inherited from the router advertising it on the 978 shortest path to p, then the prefix p's alternate next-hop would be 979 the link to C. This would provide link protection, but not the node 980 protection that is possible via A. 982 5 +---+ 4 +---+ 5 +---+ 983 ------| S |------| A |-----| B | 984 | +---+ +---+ +---+ 985 | | | 986 | 5 | 5 | 987 | | | 988 +---+ 5 +---+ 5 7 +---+ 989 | C |---| E |------ p -------| F | 990 +---+ +---+ +---+ 992 Figure 6: Multi-homed prefix 994 To determine the best protection possible, the prefix p can be 995 treated in the SPF computations as a node with uni-directional links 996 to it from those routers that have advertised the prefix. Such a 997 node need never have its links explored, as it has no out-going 998 links. 1000 If there exist multiple multi-homed prefixes that share the same 1001 connectivity and the difference in metrics to those routers, then a 1002 single node can be used to represent the set. For instance, if in 1003 Figure 6 there were another prefix X that was connected to E with a 1004 metric of 1 and to F with a metric of 3, then that prefix X could use 1005 the same alternate next-hop as was computed for prefix p. 1007 A router SHOULD compute the alternate next-hop for an IGP multi-homed 1008 prefix by considering alternate paths via all routers that have 1009 announced that prefix. 1011 6.2. ISIS 1013 The applicability and interactions of LFAs with multi-topology ISIS 1014 [I-D.ietf-isis-wg-multi-topology] is out of scope for this 1015 specification. 1017 6.3. OSPF 1019 OSPF introduces certain complications because it is possible for the 1020 traffic path to exit an area and then re-enter that area. This can 1021 occur whenever a router considers the same route from multiple areas. 1022 There are several cases where issues such as this can occur. They 1023 happen when another area permits a shorter path to connect two ABRs 1024 than is available in the area where the LFA has been computed. To 1025 clarify, an example topology is given in Appendix A. 1027 a. Virtual Links: These allow paths to leave the backbone area and 1028 traverse the transit area. The path provided via the transit 1029 area can exit via any ABR. The path taken is not the shortest 1030 path determined by doing an SPF in the backbone area. 1032 b. Alternate ABR[RFC3509]: When an ABR is not connected to the 1033 backbone, it considers the inter-area summaries from multiple 1034 areas. The ABR A may determine to use area 2 but that path could 1035 traverse another alternate ABR B that determines to use area 1. 1036 This can lead to scenarios similar to that illustrated in 1037 Figure 7. 1039 c. ASBR Summaries: An ASBR may itself be an ABR and can be announced 1040 into multiple areas. This presents other ABRs with a decision as 1041 to which area to use. This is the example illustrated in 1042 Figure 7. 1044 d. AS External Prefixes: A prefix may be advertised by multiple 1045 ASBRs in different areas and/or with multiple forwarding 1046 addresses that are in different areas, which are connected via at 1047 least one common ABR. This presents such ABRs with a decision as 1048 to which area to use to reach the prefix. 1050 Loop-free alternates should not be used in an area where one of the 1051 above issues affects that area. 1053 6.3.1. OSPF External Routing 1055 When a forwarding address is set in an OSPF AS-external LSA, all 1056 routers in the network calculate their next-hops for the external 1057 prefix by doing a lookup for the forwarding address in the routing 1058 table, rather than using the next-hops calculated for the ASBR. In 1059 this case, the alternate next-hops SHOULD be computed by selecting 1060 among the alternate paths to the forwarding link(s) instead of among 1061 alternate paths to the ASBR. 1063 6.3.2. OSPF Multi-Topology 1065 The applicabilty and interactions of LFAs with multi-topology OSPF 1066 [RFC4915] [I-D.ietf-ospf-mt-ospfv3] is out of scope for this 1067 specification. 1069 6.4. BGP Next-Hop Synchronization 1071 Typically BGP prefixes are advertised with AS exit routers router-id 1072 as the BGP next-hop, and AS exit routers are reached by means of IGP 1073 routes. BGP resolves its advertised next-hop to the immediate next- 1074 hop by potential recursive lookups in the routing database. IP Fast- 1075 Reroute computes the alternate next-hops to all IGP destinations, 1076 which include alternate next-hops to the AS exit router's router-id. 1077 BGP simply inherits the alternate next-hop from IGP. The BGP 1078 decision process is unaltered; BGP continues to use the IGP optimal 1079 distance to find the nearest exit router. MBGP routes do not need to 1080 copy the alternate next hops. 1082 It is possible to provide ASBR protection if BGP selected a set of 1083 BGP next-hops and allowed the IGP to determine the primary and 1084 alternate next-hops as if the BGP route were a multi-homed prefix. 1085 This is for future study. 1087 6.5. Multicast Considerations 1089 Multicast traffic is out of scope for this specification of IP Fast- 1090 Reroute. The alternate next-hops SHOULD NOT be used for multicast 1091 RPF checks. 1093 7. Security Considerations 1095 The mechanism described in this document does not modify any routing 1096 protocol messages, and hence no new threats related to packet 1097 modifications or replay attacks are introduced. Traffic to certain 1098 destinations can be temporarily routed via next-hop routers that 1099 would not be used with the same topology change if this mechanism 1100 wasn't employed. However, these next-hop routers can be used anyway 1101 when a different topological change occurs, and hence this can't be 1102 viewed as a new security threat. 1104 In LDP, the wider distribution of FEC label information is still to 1105 neighbors with whom a trusted LDP session has been established. This 1106 wider distribution and the recommendation of using liberal label 1107 retention mode are believed to have no significant security impact. 1109 8. IANA Considerations 1111 This document requires no IANA considerations. 1113 9. Acknowledgements 1115 The authors would like to thank Joel Halpern, Mike Shand, Stewart 1116 Bryant, and Stefano Previdi for their assistance and useful review. 1118 10. References 1120 10.1. Normative References 1122 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1123 Requirement Levels", BCP 14, RFC 2119, March 1997. 1125 [RFC2328] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998. 1127 [RFC2740] Coltun, R., Ferguson, D., and J. Moy, "OSPF for IPv6", 1128 RFC 2740, December 1999. 1130 [RFC3036] Andersson, L., Doolan, P., Feldman, N., Fredette, A., and 1131 B. Thomas, "LDP Specification", RFC 3036, January 2001. 1133 10.2. Informative References 1135 [I-D.francois-ordered-fib] 1136 Francois, P., "Loop-free convergence using oFIB", 1137 draft-francois-ordered-fib-02 (work in progress), 1138 October 2006. 1140 [I-D.ietf-isis-link-attr] 1141 Vasseur, J. and S. Previdi, "Definition of an IS-IS Link 1142 Attribute sub-TLV", draft-ietf-isis-link-attr-03 (work in 1143 progress), February 2007. 1145 [I-D.ietf-isis-wg-multi-topology] 1146 Przygienda, T., "M-ISIS: Multi Topology (MT) Routing in 1147 IS-IS", draft-ietf-isis-wg-multi-topology-11 (work in 1148 progress), October 2005. 1150 [I-D.ietf-ospf-mt-ospfv3] 1151 Mirtorabi, S. and A. Roy, "Multi-topology routing in 1152 OSPFv3 (MT-OSPFv3)", draft-ietf-ospf-mt-ospfv3-03 (work in 1153 progress), July 2007. 1155 [I-D.ietf-ospf-ospfv3-update] 1156 Ferguson, D., "OSPF for IPv6", 1157 draft-ietf-ospf-ospfv3-update-17 (work in progress), 1158 August 2007. 1160 [I-D.ietf-rtgwg-ipfrr-framework] 1161 Shand, M. and S. Bryant, "IP Fast Reroute Framework", 1162 draft-ietf-rtgwg-ipfrr-framework-07 (work in progress), 1163 July 2007. 1165 [I-D.ietf-rtgwg-microloop-analysis] 1166 Zinin, A., "Analysis and Minimization of Microloops in 1167 Link-state Routing Protocols", 1168 draft-ietf-rtgwg-microloop-analysis-01 (work in progress), 1169 October 2005. 1171 [RFC1195] Callon, R., "Use of OSI IS-IS for routing in TCP/IP and 1172 dual environments", RFC 1195, December 1990. 1174 [RFC2966] Li, T., Przygienda, T., and H. Smit, "Domain-wide Prefix 1175 Distribution with Two-Level IS-IS", RFC 2966, 1176 October 2000. 1178 [RFC3137] Retana, A., Nguyen, L., White, R., Zinin, A., and D. 1179 McPherson, "OSPF Stub Router Advertisement", RFC 3137, 1180 June 2001. 1182 [RFC3509] Zinin, A., Lindem, A., and D. Yeung, "Alternative 1183 Implementations of OSPF Area Border Routers", RFC 3509, 1184 April 2003. 1186 [RFC4203] Kompella, K. and Y. Rekhter, "OSPF Extensions in Support 1187 of Generalized Multi-Protocol Label Switching (GMPLS)", 1188 RFC 4203, October 2005. 1190 [RFC4205] Kompella, K. and Y. Rekhter, "Intermediate System to 1191 Intermediate System (IS-IS) Extensions in Support of 1192 Generalized Multi-Protocol Label Switching (GMPLS)", 1193 RFC 4205, October 2005. 1195 [RFC4915] Psenak, P., Mirtorabi, S., Roy, A., Nguyen, L., and P. 1196 Pillay-Esnault, "Multi-Topology (MT) Routing in OSPF", 1197 RFC 4915, June 2007. 1199 Appendix A. OSPF Example Where LFA Based on Local Area Topology is 1200 Insufficient 1202 This appendix provides an example scenario where the local area 1203 topology does not suffice to determine that an LFA is available. As 1204 described in Section 6.3, one problem scenario is for ASBR summaries 1205 where the ASBR is available in two areas via intra-area routes and 1206 there is at least one ABR or alternate ABR that is in both areas. 1208 The following Figure 7 illustrates this case. 1210 5 1211 [ F ]-----------[ C ] 1212 | | 1213 | | 5 1214 20 | 5 | 1 1215 | [ N ]-----[ A ]*****[ F ] 1216 | | # * 1217 | 40 | # 50 * 2 1218 | | 5 # 2 * 1219 | [ S ]-----[ B ]*****[ G ] 1220 | | * 1221 | 5 | * 15 1222 | | * 1223 | [ E ] [ H ] 1224 | | * 1225 | 5 | * 10** 1226 | | * 1227 |---[ X ]-----[ASBR] 1228 5 1230 ---- Link in Area 1 1231 **** Link in Area 2 1232 #### Link in Backbone Area 0 1234 Figure 7: Topology with Multi-area ASBR Causing Area Transiting 1236 In Figure 7, the ASBR is also an ABR and is announced into both area 1237 1 and area 2. A and B are both ABRs that are also connected to the 1238 backbone area. S determines that N can provide a loop-free alternate 1239 to reach the ASBR. N's path goes via A. A also sees an intra-area 1240 route to ASBR via Area 2; the cost of the path in area 2 is 30, which 1241 is less than 35, the cost of the path in area 1. Therefore, A uses 1242 the path from area 2 and directs traffic to F. The path from F in 1243 area 2 goes to B. B is also an ABR and learns the ASBR from both 1244 areas 1 and area 2; B's path via area 1 is shorter (cost 20) than B's 1245 path via area 2 (cost 25). Therefore, B uses the path from area 1 1246 that connects to S. 1248 Authors' Addresses 1250 Alia K. Atlas (editor) 1251 Google, Inc. 1252 One Broadway, 7th Floor 1253 Cambridge, MA 02142 1254 USA 1256 Email: akatlas@google.com 1258 Alex Zinin (editor) 1259 Alcatel 1260 701 E Middlefield Rd. 1261 Mountain View, CA 94043 1262 USA 1264 Email: alex.zinin@alcatel.com 1266 Raveendra Torvi 1267 FutureWei Technologies Inc. 1268 1700 Alma Dr. Suite 100 1269 Plano, TX 75075 1270 USA 1272 Email: traveendra@huawei.com 1274 Gagan Choudhury 1275 AT&T 1276 200 Laurel Avenue, Room D5-3C21 1277 Middletown, NJ 07748 1278 USA 1280 Phone: +1 732 420-3721 1281 Email: gchoudhury@att.com 1283 Christian Martin 1284 iPath Technologies 1286 Email: chris@ipath.net 1287 Brent Imhoff 1288 Juniper Networks 1289 1194 North Mathilda 1290 Sunnyvale, CA 94089 1291 USA 1293 Phone: +1 314 378 2571 1294 Email: bimhoff@planetspork.com 1296 Don Fedyk 1297 Nortel Networks 1298 600 Technology Park 1299 Billerica, MA 01821 1300 USA 1302 Phone: +1 978 288 3041 1303 Email: dwfedyk@nortelnetworks.com 1305 Full Copyright Statement 1307 Copyright (C) The IETF Trust (2007). 1309 This document is subject to the rights, licenses and restrictions 1310 contained in BCP 78, and except as set forth therein, the authors 1311 retain all their rights. 1313 This document and the information contained herein are provided on an 1314 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1315 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 1316 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 1317 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 1318 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1319 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1321 Intellectual Property 1323 The IETF takes no position regarding the validity or scope of any 1324 Intellectual Property Rights or other rights that might be claimed to 1325 pertain to the implementation or use of the technology described in 1326 this document or the extent to which any license under such rights 1327 might or might not be available; nor does it represent that it has 1328 made any independent effort to identify any such rights. Information 1329 on the procedures with respect to rights in RFC documents can be 1330 found in BCP 78 and BCP 79. 1332 Copies of IPR disclosures made to the IETF Secretariat and any 1333 assurances of licenses to be made available, or the result of an 1334 attempt made to obtain a general license or permission for the use of 1335 such proprietary rights by implementers or users of this 1336 specification can be obtained from the IETF on-line IPR repository at 1337 http://www.ietf.org/ipr. 1339 The IETF invites any interested party to bring to its attention any 1340 copyrights, patents or patent applications, or other proprietary 1341 rights that may cover technology that may be required to implement 1342 this standard. Please address the information to the IETF at 1343 ietf-ipr@ietf.org. 1345 Acknowledgment 1347 Funding for the RFC Editor function is provided by the IETF 1348 Administrative Support Activity (IASA).