idnits 2.17.1 draft-ietf-rtgwg-remote-lfa-09.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 637 has weird spacing: '...achable via a...' -- The document date (December 16, 2014) is 3412 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-02) exists of draft-ietf-ospf-routable-ip-address-01 == Outdated reference: A later version (-11) exists of draft-ietf-rtgwg-lfa-manageability-04 == Outdated reference: A later version (-13) exists of draft-ietf-rtgwg-rlfa-node-protection-01 == Outdated reference: A later version (-04) exists of draft-litkowski-rtgwg-uloop-delay-03 Summary: 0 errors (**), 0 flaws (~~), 6 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group S. Bryant 3 Internet-Draft C. Filsfils 4 Intended status: Standards Track S. Previdi 5 Expires: June 19, 2015 Cisco Systems 6 M. Shand 7 Independent Contributor 8 N. So 9 Vinci Systems 10 December 16, 2014 12 Remote LFA FRR 13 draft-ietf-rtgwg-remote-lfa-09 15 Abstract 17 This draft describes an extension to the basic IP fast re-route 18 mechanism described in RFC5286, that provides additional backup 19 connectivity for point to point link failures when none can be 20 provided by the basic mechanisms. 22 Requirements Language 24 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 25 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 26 document are to be interpreted as described in RFC2119 [RFC2119]. 28 Status of This Memo 30 This Internet-Draft is submitted in full conformance with the 31 provisions of BCP 78 and BCP 79. 33 Internet-Drafts are working documents of the Internet Engineering 34 Task Force (IETF). Note that other groups may also distribute 35 working documents as Internet-Drafts. The list of current Internet- 36 Drafts is at http://datatracker.ietf.org/drafts/current/. 38 Internet-Drafts are draft documents valid for a maximum of six months 39 and may be updated, replaced, or obsoleted by other documents at any 40 time. It is inappropriate to use Internet-Drafts as reference 41 material or to cite them other than as "work in progress." 43 This Internet-Draft will expire on June 19, 2015. 45 Copyright Notice 47 Copyright (c) 2014 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (http://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 63 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 64 3. Repair Paths . . . . . . . . . . . . . . . . . . . . . . . . 5 65 3.1. Tunnels as Repair Paths . . . . . . . . . . . . . . . . . 6 66 3.2. Tunnel Requirements . . . . . . . . . . . . . . . . . . . 6 67 4. Construction of Repair Paths . . . . . . . . . . . . . . . . 7 68 4.1. Identifying Required Tunneled Repair Paths . . . . . . . 7 69 4.2. Determining Tunnel End Points . . . . . . . . . . . . . . 8 70 4.2.1. Computing Repair Paths . . . . . . . . . . . . . . . 8 71 4.2.2. Selecting Repair Paths . . . . . . . . . . . . . . . 11 72 4.3. A Cost Based RLFA Algorithm . . . . . . . . . . . . . . . 11 73 4.4. Interactions with IS-IS Overload, RFC6987, and Costed 74 Out Links . . . . . . . . . . . . . . . . . . . . . . . . 16 75 5. Example Application of Remote LFAs . . . . . . . . . . . . . 17 76 6. Node Failures . . . . . . . . . . . . . . . . . . . . . . . . 18 77 7. Operation in an LDP environment . . . . . . . . . . . . . . . 19 78 8. Analysis of Real World Topologies . . . . . . . . . . . . . . 20 79 8.1. Topology Details . . . . . . . . . . . . . . . . . . . . 21 80 8.2. LFA only . . . . . . . . . . . . . . . . . . . . . . . . 21 81 8.3. RLFA . . . . . . . . . . . . . . . . . . . . . . . . . . 22 82 8.4. Comparison of LFA an RLFA results . . . . . . . . . . . . 23 83 9. Management Considerations . . . . . . . . . . . . . . . . . . 24 84 10. Historical Note . . . . . . . . . . . . . . . . . . . . . . . 25 85 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 25 86 12. Security Considerations . . . . . . . . . . . . . . . . . . . 25 87 13. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 25 88 14. References . . . . . . . . . . . . . . . . . . . . . . . . . 26 89 14.1. Normative References . . . . . . . . . . . . . . . . . . 26 90 14.2. Informative References . . . . . . . . . . . . . . . . . 26 91 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 28 93 1. Terminology 95 This draft uses the terms defined in [RFC5714]. This section defines 96 additional terms used in this draft. 98 FIB Forwarding Information (data)Base. The database used 99 by a packet forwarder to determine the actions it 100 should take on a packet it is processing. 102 Repair tunnel A tunnel established for the purpose of providing a 103 virtual neighbor which is a Loop Free Alternate. 105 P-space P-space is the set of routers reachable from a 106 specific router using the normal FIB, without any path 107 (including equal cost path splits) transiting the 108 protected link. 110 For example, the P-space of S with respect to link 111 S-E, is the set of routers that S can reach without 112 using the protected link S-E. 114 Extended P-space 116 The union of the P-space of the neighbours of a 117 specific router with respect to the protected link 118 (see Section 4.2.1.2). 120 Q-space Q-space is the set of routers from which a specific 121 router can be reached without any path (including 122 equal cost path splits) transiting the protected link. 124 PQ node A node which is a member of both the P-space and the 125 Q-space. Where extended P-space is in use it is a 126 node which is a member of both the extended P-space 127 and the Q-space. In remote LFA this is used as the 128 repair tunnel endpoint. 130 Remote LFA (RLFA) The use of a PQ node rather than a neighbour of 131 the repairing node as the next hop in an LFA repair. 133 In this document we use the notation X-Y to mean the path from X to Y 134 over the link directly connecting X and Y, whilst the notation X->Y 135 refers to the shortest path from X to Y via some set of unspecified 136 nodes including the null set (i.e. including over a link directly 137 connecting X and Y). 139 2. Introduction 141 RFC 5714 [RFC5714] describes a framework for IP Fast Re-route and 142 provides a summary of various proposed IPFRR solutions. A basic 143 mechanism using loop-free alternates (LFAs) is described in [RFC5286] 144 that provides good repair coverage in many topologies [RFC6571], 145 especially those that are highly meshed. However, some topologies, 146 notably ring based topologies are not well protected by LFAs alone. 147 This is illustrated in Figure 1 below. 149 S---E 150 / \ 151 A D 152 \ / 153 B---C 155 Figure 1: A simple ring topology 157 If all link costs are equal, the link S-E cannot be fully protected 158 by LFAs. The destination C is an ECMP from S, and so can be 159 protected when S-E fails, but D and E are not protectable using LFAs. 161 This draft describes extensions to the basic repair mechanism in 162 which tunnels are used to provide additional logical links which can 163 then be used as loop free alternates where none exist in the original 164 topology. In Figure 1 S can reach A, B, and C without going via E; 165 these form S's extended P-space. The routers that can reach E 166 without going through S-E will be E's Q-space; these are D and C. B 167 has equal-cost paths via B-A-S-E and B-C-D-E and so may go through 168 S-E. The single node in both S's extended P-space and E's Q-space is 169 C; thus node C is selected as the repair tunnel's end-point. Thus, 170 if a tunnel is provided between S and C as shown in Figure 2 then C, 171 now being a direct neighbor of S would become an LFA for D and E. 172 The definition of (extended-)P space and Q space are provided in 173 Section 1 and details of the calculation of the tunnel end points is 174 provided in Section 4.2. 176 The non-failure traffic distribution is not disrupted by the 177 provision of such a tunnel since it is only used for repair traffic 178 and MUST NOT be used for normal traffic. Note that OAM traffic 179 specifically to verify the viability of the repair MAY traverse the 180 tunnel prior to a failure. 182 S---E 183 / \ \ 184 A \ D 185 \ \ / 186 B---C 188 Figure 2: The addition of a tunnel 190 The use of this technique is not restricted to ring based topologies, 191 but is a general mechanism which can be used to enhance the 192 protection provided by LFAs. A study of the protection achieved 193 using remote LFA in typical service provider core networks is 194 provided in Section 8, and a side by side comparison between LFA and 195 remote LFA is provided in Section 8.4. 197 Remote LFA is suitable for incremental deployment within a network, 198 including a network that is already deploying LFA. Computation of 199 the repair path requires acceptable CPU resources, and takes place 200 exclusively on the repairing node. In MPLS networks the targeted LDP 201 protocol needed to learn the label binding at the repair tunnel 202 endpoint is a well understood and widely deployed technology. 204 The technique described in this document is directed at providing 205 repairs in the case of link failures. Considerations regarding node 206 failures are discussed in Section 6. This memo describes a solution 207 to the case where the failure occurs on a point to point link. It 208 covers the case where the repair first hop is reached via a broadcast 209 or non-broadcast multi-access (NBMA) link such as a LAN, and the case 210 where the P or Q node is attached via such a link. It does not 211 however cover the more complicated case where the failed interface is 212 a broadcast or non-broadcast multi-access (NBMA) link. 214 This document considers the case when the repair path is confined to 215 either a single area or to the level two routing domain. In all 216 other cases, the chosen PQ node should be regarded as a tunnel 217 adjacency of the repairing node, and the considerations described in 218 Section 6 of [RFC5286] taken into account. 220 3. Repair Paths 222 As with LFA FRR, when a router detects an adjacent link failure, it 223 uses one or more repair paths in place of the failed link. Repair 224 paths are pre-computed in anticipation of later failures so they can 225 be promptly activated when a failure is detected. 227 A tunneled repair path tunnels traffic to some staging point in the 228 network from which it is known that, in the absence of a worse than 229 anticipated failure, the traffic will travel to its destination using 230 normal forwarding without looping back. This is equivalent to 231 providing a virtual loop-free alternate to supplement the physical 232 loop-free alternates. Hence the name "Remote LFA FRR". In its 233 simplest form, when a link cannot be entirely protected with local 234 LFA neighbors, the protecting router seeks the help of a remote LFA 235 staging point. Network manageability considerations may lead to a 236 repair strategy that uses a remote LFA more frequently 237 [I-D.ietf-rtgwg-lfa-manageability]. 239 Examples of worse failures are node failures (see Section 6 ), the 240 failure of a shared risk link group (SRLG), the independent 241 concurrent failures of multiple links, broadcast or non-broadcast 242 multi-access (NBMA) links Section 2 ; protecting against such worse 243 failures is out of scope for this specification. 245 3.1. Tunnels as Repair Paths 247 Consider an arbitrary protected link S-E. In LFA FRR, if a path to 248 the destination from a neighbor N of S does not cause a packet to 249 loop back over the link S-E (i.e. N is a loop-free alternate), then 250 S can send the packet to N and the packet will be delivered to the 251 destination using the pre-failure forwarding information. If there 252 is no such LFA neighbor, then S may be able to create a virtual LFA 253 by using a tunnel to carry the packet to a point in the network which 254 is not a direct neighbor of S from which the packet will be delivered 255 to the destination without looping back to S. In this document such 256 a tunnel is termed a repair tunnel. The tail-end of this tunnel (the 257 repair tunnel endpoint) is a "PQ node" and the repair mechanism is a 258 "remote LFA". This tunnel MUST NOT traverse the link S-E. 260 Note that the repair tunnel terminates at some intermediate router 261 between S and E, and not E itself. This is clearly the case, since 262 if it were possible to construct a tunnel from S to E then a 263 conventional LFA would have been sufficient to effect the repair. 265 3.2. Tunnel Requirements 267 There are a number of IP in IP tunnel mechanisms that may be used to 268 fulfil the requirements of this design, such as IP-in-IP [RFC1853] 269 and GRE[RFC1701] . 271 In an MPLS enabled network using LDP[RFC5036], a simple label 272 stack[RFC3032] may be used to provide the required repair tunnel. In 273 this case the outer label is S's neighbor's label for the repair 274 tunnel end point, and the inner label is the repair tunnel end 275 point's label for the packet destination. In order for S to obtain 276 the correct inner label it is necessary to establish a targeted LDP 277 session[RFC5036] to the tunnel end point. 279 The selection of the specific tunnelling mechanism (and any necessary 280 enhancements) used to provide a repair path is outside the scope of 281 this document. The deployment in an MPLS/LDP environment is 282 relatively simple in the data plane as an LDP LSP from S to the 283 repair tunnel endpoint (the selected PQ node) is readily available, 284 and hence does not require any new protocol extension or design 285 change. This LSP is automatically established as a basic property of 286 LDP behavior. The performance of the encapsulation and decapsulation 287 is efficient as encapsulation is just a push of one label (like 288 conventional MPLS TE FRR) and the decapsulation is normally 289 configured to occur at the penultimate hop before the repair tunnel 290 endpoint. In the control plane, a targeted LDP (TLDP) session is 291 needed between the repairing node and the repair tunnel endpoint, 292 which will need to be established and the labels processed before the 293 tunnel can be used. The time to establish the TLDP session and 294 acquire labels will limit the speed at which a new tunnel can be put 295 into service. This is not anticipated to be a problem in normal 296 operation since the managed introduction and removal of links is 297 relatively rare as is the incidence of failure in a well managed 298 network. 300 When a failure is detected, it is necessary to immediately redirect 301 traffic to the repair path. Consequently, the repair tunnel used 302 MUST be provisioned beforehand in anticipation of the failure. Since 303 the location of the repair tunnels is dynamically determined it is 304 necessary to automatically establish the repair tunnels. Multiple 305 repairs MAY share a tunnel end point. 307 4. Construction of Repair Paths 309 4.1. Identifying Required Tunneled Repair Paths 311 Not all links will require protection using a tunneled repair path. 312 Referring to Figure 1, if E can already be protected via an LFA, S-E 313 does not need to be protected using a repair tunnel, since all 314 destinations normally reachable through E must therefore also be 315 protectable by an LFA. Such an LFA is frequently termed a "link 316 LFA". Tunneled repair paths (which may be calculated per-prefix) are 317 only required for links which do not have a link or per-prefix LFA. 319 It should be noted that using the Q-space of E as a proxy for the 320 Q-space of each destination can result in failing to identify valid 321 remote LFAs. The extent to which this reduces the effective 322 protection coverage is topology dependent. 324 4.2. Determining Tunnel End Points 326 The repair tunnel endpoint needs to be a node in the network 327 reachable from S without traversing S-E. In addition, the repair 328 tunnel end point needs to be a node from which packets will normally 329 flow towards their destination without being attracted back to the 330 failed link S-E. 332 Note that once released from the tunnel, the packet will be 333 forwarded, as normal, on the shortest path from the release point to 334 its destination. This may result in the packet traversing the router 335 E at the far end of the protected link S-E, but this is obviously not 336 required. 338 The properties that are required of repair tunnel end points are 339 therefore: 341 o The repair tunneled point MUST be reachable from the tunnel source 342 without traversing the failed link; and 344 o When released from the tunnel, packets MUST proceed towards their 345 destination without being attracted back over the failed link. 347 Provided both these requirements are met, packets forwarded over the 348 repair tunnel will reach their destination, and will not loop after a 349 single link failure. 351 In some topologies it will not be possible to find a repair tunnel 352 endpoint that exhibits both the required properties. For example if 353 the ring topology illustrated in Figure 1 had a cost of 4 for the 354 link B-C, while the remaining links were cost 1, then it would not be 355 possible to establish a tunnel from S to C (without resorting to some 356 form of source routing). 358 4.2.1. Computing Repair Paths 360 To compute the repair path for link S-E we need to determine the set 361 of routers which can be reached from S without traversing S-E, and 362 match this with the set of routers from which the node E can be 363 reached, by normal forwarding, without traversing the link S-E. 365 The approach described in this memo is as follows: 367 o We describe how to compute the set of routers which can be reached 368 from S on the shortest path tree without traversing S-E. We call 369 this the S's P-space with respect to the failure of link S-E. 371 o We show how to extend the distance of the tunnel endpoint from the 372 point of local repair (PLR) by noting that S is able to use the 373 P-Space of its neighbours since S can determine which neighbour it 374 will use as the next hop for the repair. We call this the S's 375 Extended P-Space with respect to the failure of link S-E. The use 376 of extended P-Space allows greater repair coverage and is the 377 preferred approach. 379 o Finally we show how to compute the set of routers from which the 380 node E can be reached, by normal forwarding, without traversing 381 the link S-E. This is called the Q-space of E with respect to the 382 link S-E. 384 The selection of the preferred node from the set of nodes that an in 385 both Extended P-Space and Q-Space is described in Section 4.2.2. 387 A suitable cost based algorithm to compute the set of nodes common to 388 both extended P-space and Q-space is provided in Section 4.3. 390 4.2.1.1. P-space 392 The set of routers which can be reached from S on the shortest path 393 tree without traversing S-E is termed the P-space of S with respect 394 to the link S-E. The P-space can be obtained by computing a shortest 395 path tree (SPT) rooted at S and excising the sub-tree reached via the 396 link S-E (including those routers which are members of an ECMP that 397 includes link S-E). The exclusion of routers reachable via an ECMP 398 that includes S-E prevents the forwarding subsystem from attempting 399 to a repair endpoint via the failed link S-E. Thus for example, if 400 the SPF computation stores at each node the next-hops to be used to 401 reach that node from S, then the node can be added to P-space if none 402 of its next-hops are S-E. In the case of Figure 1 the P-space 403 comprises nodes A and B only. Expressed in cost terms the set of 404 routers {P} are those for which the shortest path cost S->P is 405 strictly less than the shortest path cost S->E->P. 407 4.2.1.2. Extended P-space 409 The description in Section 4.2.1.1 calculated router S's P-space 410 rooted at S itself. However, since router S will only use a repair 411 path when it has detected the failure of the link S-E, the initial 412 hop of the repair path need not be subject to S's normal forwarding 413 decision process. Thus we introduce the concept of extended P-space. 414 Router S's extended P-space is the union of the P-spaces of each of 415 S's neighbours (N). This may be calculated by computing an SPT at 416 each of S's neighbors (excluding E) and excising the subtree reached 417 via the path N->S->E. Note this will excise those routers which are 418 reachable through all ECMPs that includes link S-E. The use of 419 extended P-space may allow router S to reach potential repair tunnel 420 end points that were otherwise unreachable. In cost terms a router 421 (P) is in extended P-space if the shortest path cost N->P is strictly 422 less than the shortest path cost N->S->E->P. In other words, once 423 the packet it forced to N by S, it is lower cost for it to continue 424 on to P by any path except one that takes it back to S and then 425 across the S->E link. 427 Since in the case of Figure 1 node A is a per-prefix LFA for the 428 destination node C, the set of extended P-space nodes comprises nodes 429 A, B and C. Since node C is also in E's Q-space, there is now a node 430 common to both extended P-space and Q-space which can be used as a 431 repair tunnel end-point to protect the link S-E. 433 4.2.1.3. Q-space 435 The set of routers from which the node E can be reached, by normal 436 forwarding, without traversing the link S-E is termed the Q-space of 437 E with respect to the link S-E. The Q-space can be obtained by 438 computing a reverse shortest path tree (rSPT) rooted at E, with the 439 sub-tree which traverses the failed link excised (including those 440 which are members of an ECMP). The rSPT uses the cost towards the 441 root rather than from it and yields the best paths towards the root 442 from other nodes in the network. In the case of Figure 1 the Q-space 443 comprises nodes C and D only. Expressed in cost terms the set of 444 routers {Q} are those for which the shortest path cost Q<-E is 445 strictly less than the shortest path cost Q<-S<-E. In Figure 1 the 446 intersection of the E's Q-space with S's P-space defines the set of 447 viable repair tunnel end-points, known as "PQ nodes". As can be 448 seen, for the case of Figure 1 there is no common node and hence no 449 viable repair tunnel end-point. However when the extended the 450 extended P-space Section 4.2.1.2 at S is considered a suitable 451 intersection is found at C. 453 Note that the Q-space calculation could be conducted for each 454 individual destination and a per-destination repair tunnel end point 455 determined. However this would, in the worst case, require an SPF 456 computation per destination which is not currently considered to be 457 scalable. We therefore use the Q-space of E as a proxy for the 458 Q-space of each destination. This approximation is obviously correct 459 since the repair is only used for the set of destinations which were, 460 prior to the failure, routed through node E. This is analogous to 461 the use of link-LFAs rather than per-prefix LFAs. 463 4.2.2. Selecting Repair Paths 465 The mechanisms described above will identify all the possible repair 466 tunnel end points that can be used to protect a particular link. In 467 a well-connected network there are likely to be multiple possible 468 release points for each protected link. All will deliver the packets 469 correctly so, arguably, it does not matter which is chosen. However, 470 one repair tunnel end point may be preferred over the others on the 471 basis of path cost or some other selection criteria. 473 There is no technical requirement for the selection criteria to be 474 consistent across all routers, but such consistency may be desirable 475 from an operational point of view. In general there are advantages 476 in choosing the repair tunnel end point closest (shortest metric) to 477 S. Choosing the closest maximises the opportunity for the traffic to 478 be load balanced once it has been released from the tunnel. For 479 consistency in behavior, it is RECOMMENDED that the member of the set 480 of routers {PQ} with the lowest cost S->P be the default choice for 481 P. In the event of a tie the router with the lowest node identifier 482 SHOULD be selected. 484 It is a local matter whether the repair path selection policy used by 485 the router favours LFA repairs over RLFA repairs. An LFA repair has 486 the advantage of not requiring the use of tunnel, however network 487 manageability considerations may lead to a repair strategy that uses 488 a remote LFA more frequently [I-D.ietf-rtgwg-lfa-manageability]. 490 As described in [RFC5286], always selecting a PQ node that is 491 downstream to the destination with respect to the repairing node, 492 prevents the formation of loops when the failure is worse than 493 expected. The use of downstream nodes reduces the repair coverage, 494 and operators are advised to determine whether adequate coverage is 495 achieved before enabling this selection feature. 497 4.3. A Cost Based RLFA Algorithm 499 The preceding text has mostly described the computation of the remote 500 LFA repair target (PQ) in terms of the intersection of two 501 reachability graphs computed using SPFs. This section describes a 502 method of computing the remote LFA repair target for a specific 503 failed link using a cost based algorithm. The pseudo-code provided 504 in this section avoids unnecessary SPF computations, but for the sake 505 of readability, it does not otherwise try to optimize the code. The 506 algorithm covers the case where the repair first hop is reached via a 507 broadcast or non-broadcast multi-access (NBMA) link such as a LAN. 508 It also covers the case where the P or Q node is attached via such a 509 link. It does not cover the case where the failed interface is a 510 broadcast or non-broadcast multi-access (NBMA) link. To address that 511 case it is necessary to compute the Q space of each neighbor of the 512 repairing router reachable though the LAN, i.e. to treat the 513 pseudonode as a node failure. This is because the Q spaces of the 514 neighbors of the pseudonode may be disjoint requiring use of a 515 neighbor specific PQ node. The reader is referred to 516 [I-D.ietf-rtgwg-rlfa-node-protection] for further information on the 517 use of RLFA for node repairs. 519 The following notation is used: 521 o D_opt(a,b) is the shortest distance from node a to node b as 522 computed by the SPF. 524 o dest is the packet destination 526 o fail_intf is the failed interface (S-E in the example) 528 o fail_intf.remote_node is the node reachable over interface 529 fail_intf (node E in the example) 531 o intf.remote_node is the set of nodes reachable over interface intf 533 o root is the root of the SPF calculation 535 o self is the node carrying out the computation 537 o y is the node in the network under consideration 539 o y.pseudonode is true if y is a pseudonode 540 ////////////////////////////////////////////////////////////////// 541 // 542 // Main Function 544 ////////////////////////////////////////////////////////////////// 545 // 546 // We have already computed the forward SPF from self to all nodes 547 // y in network and thus we know D_opt (self, y). This is needed 548 // for normal forwarding. 549 // However for completeness. 551 Compute_and_Store_Forward_SPF(self) 553 // To extend P-space we compute the SPF at each neighbour except 554 // the neighbour that is reached via the link being protected. 555 // We will also need D_opt(fail_intf.remote_node,y) so compute 556 // that at the same time. 558 Compute_Neighbor_SPFs() 560 // Compute the set of nodes {P} reachable other than via the 561 // failed link 563 Compute_Extended_P_Space(fail_intf) 565 // Compute the set of nodes that can reach the node on the far 566 // side of the failed link without traversing the failed link. 568 Compute_Q_Space(fail_intf) 570 // Compute the set of candidate RLFA tunnel endpoints 572 Intersect_Extended_P_and_Q_Space() 574 // Make sure that we cannot get looping repairs when the 575 // failure is worse than expected. 577 if (guarantee_no_looping_on_worse_than_protected_failure) 578 Apply_Downstream_Constraint() 580 // 581 // End of Main Function 582 // 583 ////////////////////////////////////////////////////////////////// 584 ////////////////////////////////////////////////////////////////// 585 // 586 // Procedures 587 // 589 ///////////////////////////////////////////////////////////////// 590 // 591 // This computes the SPF from root, and stores the optimum 592 // distance from root to each node y 594 Compute_and_Store_Forward_SPF(root) 595 Compute_Forward_SPF(root) 596 foreach node y in network 597 store D_opt(root,y) 599 ///////////////////////////////////////////////////////////////// 600 // 601 // This computes the optimum distance from each neighbour (other 602 // than the neighbour reachable through the failed link) and 603 // every other node in the network 605 Compute_Neighbor_SPFs() 606 foreach interface intf in self 607 Compute_and_Store_Forward_SPF(intf.remote_node) 609 ///////////////////////////////////////////////////////////////// 610 // 611 // The reverse SPF computes the cost from each remote node to 612 // root. This is achieved by running the normal SPF algorithm, 613 // but using the link cost in the direction from the next hop 614 // back towards root in place of the link cost in the direction 615 // away from root towards the next hop. 617 Compute_and_Store_Reverse_SPF(root) 618 Compute_Reverse_SPF(root) 619 foreach node y in network 620 store D_opt(y,root) 622 ///////////////////////////////////////////////////////////////// 623 // 624 // Calculate extended P-space 625 // 626 // Note the strictly less than operator is needed to 627 // avoid ECMP issues. 629 Compute_Extended_P_Space(fail_intf) 630 foreach node y in network 631 y.in_extended_P_space = false 632 // Extend P-space to the P-spaces of all reachable 633 // neighbours 634 foreach interface intf in self 635 // Exclude failed interface, noting that 636 // the node reachable via that interface may be 637 // reachable via another interface (parallel path) 638 if (intf != fail_intf) 639 foreach neighbor n in intf.remote_node 640 // Apply RFC5286 Inequality 1 641 if ( D_opt(n, y) < 642 D_opt(n,self) + D_opt)(self, y) 643 y.in_extended_P_space = true 645 ///////////////////////////////////////////////////////////////// 646 // 647 // Compute the nodes in Q-space 648 // 650 Compute_Q_Space(fail_intf) 651 // Compute the cost from every node the network to the 652 // node normally reachable across the failed link 653 Compute_and_Store_Reverse_SPF(fail_intf.remote_node) 655 // Compute the cost from every node the network to self 656 Compute_and_Store_Reverse_SPF(self) 658 foreach node y in network 659 if ( D_opt(y,fail_intf.remote_node) < D_opt(y,self) + 660 D_opt(self,fail_intf.remote_node) ) 661 y.in_Q_space = true 662 else 663 y.in_Q_space = false 665 ///////////////////////////////////////////////////////////////// 666 // 667 // Compute set of nodes in both extended P-space and in Q-space 669 Intersect_Extended_P_and_Q_Space() 670 foreach node y in network 671 if ( y.in_extended_P_space && y.in_Q_space && 672 y.pseudonode == False) 673 y.valid_tunnel_endpoint = true 674 else 675 y.valid_tunnel_endpoint = false 677 ///////////////////////////////////////////////////////////////// 678 // 679 // A downstream route is one where the next hop is strictly 680 // closer to the destination. By sending the packet to a 681 // PQ node that is downstream, we know that if the PQ node 682 // detects a failure, it will not loop the packet back to self. 683 // This is useful when there are two failures, or a node has 684 // failed rather than a link. 686 Apply_Downstream_Constraint() 687 foreach node y in network 688 if (y.valid_tunnel_endpoint) 689 Compute_and_Store_Forward_SPF(y) 690 if ((D_opt(y,dest) < D_opt(self,dest)) 691 y.valid_tunnel_endpoint = true 692 else 693 y.valid_tunnel_endpoint = false 695 // 696 ///////////////////////////////////////////////////////////////// 698 4.4. Interactions with IS-IS Overload, RFC6987, and Costed Out Links 700 Since normal link state routing takes into account the IS-IS overload 701 bit, [RFC6987], and costing out of links as described in Section 3.5 702 of [RFC5286], the forward SPFs performed by the PLR rooted at the 703 neighbors of the PLR also need to take this into account. A repair 704 tunnel path from a neighbor of the PLR to a repair tunnel endpoint 705 will generally avoid the nodes and links excluded by the IGP 706 overload/costing out rules. However, there are two situations where 707 this behavior may result in a repair path traversing a link or router 708 that should be excluded: 710 1. When the first hop on the repair tunnel path (from the PLR to a 711 direct neighbor) does not follow the IGP shortest path. In this 712 case, the PLR MUST NOT use a repair tunnel path whose first hop 713 is along a link whose cost or reverse cost is MaxLinkMetric (for 714 OSPF) or the maximum cost (for IS-IS) or, has the overload bit 715 set (for IS-IS). 717 2. The IS-IS overload bit and the mechanism of [RFC6987] only 718 prevent transit traffic from traversing a node. They do not 719 prevent traffic destined to a node. The per-neighbor forward 720 SPFs using the standard IGP overload rules will not prevent a PLR 721 from choosing a repair tunnel endpoint that is advertising a 722 desire to not carry transit traffic. Therefore, the PLR MUST NOT 723 use a repair tunnel endpoint with the IS-IS overload bit set, or 724 where all outgoing interfaces have the cost set to MaxLinkMetric 725 for OSPF. 727 5. Example Application of Remote LFAs 729 An example of a commonly deployed topology which is not fully 730 protected by LFAs alone is shown in Figure 3. PE1 and PE2 are 731 connected in the same site. P1 and P2 may be geographically 732 separated (inter-site). In order to guarantee the lowest latency 733 path from/to all other remote PEs, normally the shortest path follows 734 the geographical distance of the site locations. Therefore, to 735 ensure this, a lower IGP metric (5) is assigned between PE1 and PE2. 736 A high metric (1000) is set on the P-PE links to prevent the PEs 737 being used for transit traffic. The PEs are not individually dual- 738 homed in order to reduce costs. 740 This is a common topology in SP networks. 742 When a failure occurs on the link between PE1 and P1, PE1 does not 743 have an LFA for traffic reachable via P1. Similarly, by symmetry, if 744 the link between PE2 and P2 fails, PE2 does not have an LFA for 745 traffic reachable via P2. 747 Increasing the metric between PE1 and PE2 to allow the LFA would 748 impact the normal traffic performance by potentially increasing the 749 latency. 751 | 100 | 752 -P1---------P2- 753 \ / 754 1000 \ / 1000 755 PE1---PE2 756 5 758 Figure 3: Example SP topology 760 Clearly, full protection can be provided, using the techniques 761 described in this draft, by PE1 choosing P2 as the remote LFA repair 762 target node, and PE2 choosing P1 as the remote LFA repair target. 764 6. Node Failures 766 When the failure is a node failure rather than a point-to-point link 767 failure there is a danger that the RLFA repair will loop. This is 768 discussed in detail in [I-D.bryant-ipfrr-tunnels]. In summary the 769 problem is that two of more of E's neighbors each with E as the next 770 hop to some destination D may attempt to repair a packet addressed to 771 destination D via the other neighbor and then E, thus causing a loop 772 to form. A similar problem exists in the case of a shared risk link 773 group failure where the PLR for each failure attempts to repair via 774 the other failure. As will be noted from [I-D.bryant-ipfrr-tunnels], 775 this can rapidly become a complex problem to address. 777 There are a number of ways to minimize the probability of a loop 778 forming when a node failure occurs and there exists the possibility 779 that two of E's neighbors may form a mutual repair. 781 1. Detect when a packet has arrived on some interface I that is also 782 the interface used to reach the first hop on the RLFA path to the 783 remote LFA repair target, and drop the packet. This is useful in 784 the case of a ring topology. 786 2. Require that the path from the remote LFA repair target to 787 destination D never passes through E (including in the ECMP 788 case), i.e. only use node protecting paths in which the cost from 789 the remote LFA repair target to D is strictly less than the cost 790 from the remote LFA repair target to E plus the cost E to D. 792 3. Require that where the packet may pass through another neighbor 793 of E, that node is down stream (i.e. strictly closer to D than 794 the repairing node). This means that some neighbor of E (X) can 795 repair via some other neighbor of E (Y), but Y cannot repair via 796 X. 798 Case 1 accepts that loops may form and suppresses them by dropping 799 packets. Dropping packets may be considered less detrimental than 800 looping packets. This approach may also lead to dropping some 801 legitimate packets. Cases 2 and 3 above prevent the formation of a 802 loop, but at the expense of a reduced repair coverage and at the cost 803 of additional complexity in the algorithm to compute the repair path. 804 Alternatively one might choose to assume that the probability of a 805 node failure is sufficiently rare that the issue of looping RLFA 806 repairs can be ignored. 808 The probability of a node failure and the consequences of node 809 failure in any particular topology will depend on the node design, 810 the particular topology in use, and the strategy adopted under node 811 failure. It is recommended that a network operator perform an 812 analysis of the consequences and probability of node failure in their 813 network, and determine whether the incidence and consequence of 814 occurrence are acceptable. 816 This topic is further discussed in 817 [I-D.ietf-rtgwg-rlfa-node-protection]. 819 7. Operation in an LDP environment 821 Where this technique is used in an MPLS network using LDP [RFC5036], 822 and S is a transit node, S will need to swap the top label in the 823 stack for the remote LFA repair target's (PQ's) label to the 824 destination, and to then push its own label for the remote LFA repair 825 target. 827 In the example Figure 2 S already has the first hop (A) label for the 828 remote LFA repair target (C) as a result of the ordinary operation of 829 LDP. To get the remote LFA repair target's label (C's label) for the 830 destination (D), S needs to establish a targeted LDP session with C. 831 The label stack for normal operation and RLFA operation is shown 832 below in Figure 4. 834 +-----------------+ +-----------------+ +-----------------+ 835 | datalink | | datalink | | datalink | 836 +-----------------+ +-----------------+ +-----------------+ 837 | S's label for D | | E's label for D | | A's label for C | 838 +-----------------+ +-----------------+ +-----------------+ 839 | Payload | | Payload | | C's label for D | 840 +-----------------+ +-----------------+ +-----------------+ 841 X Y | Payload | 842 +-----------------+ 843 Z 845 X = Normal label stack packet arriving at S 846 Y = Normal label stack packet leaving S 847 Z = RLFA label stack to D via C as the remote LFA repair target. 849 Figure 4 851 To establish an targeted LDP session with a candidate remote LFA 852 repair target node the repairing node (S) needs to know what IP 853 address that the remote LFA repair target is willing to use for 854 targeted LDP sessions. Ideally this is provided by the remote LFA 855 repair target advertising this address in the IGP in use. Which 856 address is used, how this is advertised in the IGP, and whether this 857 is a special IP address or an IP address also used for some other 858 purpose is out of scope for this document and must be specified in an 859 IGP specific RFC. 861 In the absence of a protocol to learn the preferred IP address for 862 targeted LDP, an LSR should attempt a targeted LDP session with the 863 Router ID [RFC2328] [RFC5305] [RFC5340] [RFC6119] 864 [I-D.ietf-ospf-routable-ip-address], unless it is configured 865 otherwise. 867 No protection is available until the TLDP session has been 868 established and a label for the destination has been learned from the 869 remote LFA repair target. If for any reason the TLDP session cannot 870 not be established, an implementation SHOULD advise the operator 871 about the protection setup issue using any well known mechanism such 872 as Syslog [RFC5424] or SNMP [RFC3411]. 874 8. Analysis of Real World Topologies 876 This section gives the results of analysing a number of real world 877 service provider topologies collected between the end of 2012 and 878 early 2013 880 8.1. Topology Details 882 The figure below characterises each topology (topo) studied in terms 883 of : 885 o The number of nodes (# nodes) excluding pseudonodes. 887 o The number of bidirectional links ( # links) including parallel 888 links and links to and from pseudonodes. 890 o The number of node pairs that are connected by one or more links 891 (# pairs). 893 o The number of node pairs that are connected by more than one (i.e. 894 parallel) link ( # para). 896 o The number of links (excluding pseudonode links, which are by 897 definition asymmetric) that have asymmetric metrics (#asym). 899 +------+---------+---------+---------+--------+--------+ 900 | topo | # nodes | # links | # pairs | # para | # asym | 901 +------+---------+---------+---------+--------+--------+ 902 | 1 | 315 | 570 | 560 | 10 | 3 | 903 | 2 | 158 | 373 | 312 | 33 | 0 | 904 | 3 | 655 | 1768 | 1314 | 275 | 1195 | 905 | 4 | 1281 | 2326 | 2248 | 70 | 10 | 906 | 5 | 364 | 811 | 659 | 80 | 86 | 907 | 6 | 114 | 318 | 197 | 101 | 4 | 908 | 7 | 55 | 237 | 159 | 67 | 2 | 909 | 8 | 779 | 1848 | 1441 | 199 | 437 | 910 | 9 | 263 | 482 | 413 | 41 | 12 | 911 | 10 | 86 | 375 | 145 | 64 | 22 | 912 | 11 | 162 | 1083 | 351 | 201 | 49 | 913 | 12 | 380 | 1174 | 763 | 231 | 0 | 914 | 13 | 1051 | 2087 | 2037 | 48 | 64 | 915 | 14 | 92 | 291 | 204 | 64 | 2 | 916 +------+---------+---------+---------+--------+--------+ 918 8.2. LFA only 920 The figure below shows the percentage of protected destinations (% 921 prot) and percentage of guaranteed node protected destinations ( % 922 gtd N) for the set of topologies characterized in Section 8.1 923 achieved using only LFA repairs. 925 These statistics were generated by considering each node and then 926 considering each link to each next hop to each destination. The 927 percentage of such links across the entire network that are protected 928 against link failure was determined. This is the percentage of 929 protected destinations. If a link is protected against the failure 930 of the next hop node, this is considered guaranteed node protecting 931 (GNP) and percentage of guaranteed node protected destinations is 932 calculated using the same method used for calculating the link 933 protection coverage. 935 GNP is identical to Node-protecting as defined in [RFC5714] and does 936 not include the additional node protection coverage obtained by the 937 de facto node-protecting condition described in [RFC6571]. 939 +------+--------+---------+ 940 | topo | % prot | % gtd N | 941 +------+--------+---------+ 942 | 1 | 78.5 | 36.9 | 943 | 2 | 97.3 | 52.4 | 944 | 3 | 99.3 | 58 | 945 | 4 | 83.1 | 63.1 | 946 | 5 | 99 | 59.1 | 947 | 6 | 86.4 | 21.4 | 948 | 7 | 93.9 | 35.4 | 949 | 8 | 95.3 | 48.1 | 950 | 9 | 82.2 | 49.5 | 951 | 10 | 98.5 | 14.9 | 952 | 11 | 99.6 | 24.8 | 953 | 12 | 99.5 | 62.4 | 954 | 13 | 92.4 | 51.6 | 955 | 14 | 99.3 | 48.6 | 956 +------+--------+---------+ 958 8.3. RLFA 960 The figure below shows the percentage of protected destinations (% 961 prot) and % guaranteed node protected destinations ( % gtd N) for 962 RLFA protection in the topologies studies. In addition, it show the 963 percentage of destinations using an RLFA repair (% PQ) together with 964 the total number of unidirectional RLFA targeted LDP session 965 established (# PQ), the number of PQ sessions which would be required 966 for complete protection, but which could not be established because 967 there was no PQ node, i.e. the number of cases whether neither LFA or 968 RLFA protection was possible (no PQ). It also shows the 50 (p50), 90 969 (p90) and 100 (p100) percentiles for the number of individual LDP 970 sessions terminating at an individual node (whether used for TX, RX 971 or both). 973 For example, if there were LDP sessions required A->B, A->C, C->A, 974 C->D, these would be counted as 2, 1, 2, 1 at nodes A,B,C and D 975 respectively because:- 976 A has two sessions (to nodes B and C) 978 B has one session (to node A) 980 C has two sessions (to nodes A and D) 982 D has one session (to node D) 984 In this study, remote LFA is only used when necessary. i.e. when 985 there is at least one destination which is not reparable by a per 986 destination LFA, and a single remote LFA tunnel is used (if 987 available) to repair traffic to all such destinations. The remote 988 LFA repair target points are computed using extended P space and 989 choosing the PQ node which has the lowest metric cost from the 990 repairing node. 992 +------+--------+--------+------+------+-------+-----+-----+------+ 993 | topo | % prot |% gtd N | % PQ | # PQ | no PQ | p50 | p90 | p100 | 994 +------+--------+--------+------+------+-------+-----+-----+------+ 995 | 1 | 99.7 | 53.3 | 21.2 | 295 | 3 | 1 | 5 | 14 | 996 | 2 | 97.5 | 52.4 | 0.2 | 7 | 40 | 0 | 0 | 2 | 997 | 3 | 99.999 | 58.4 | 0.7 | 63 | 5 | 0 | 1 | 5 | 998 | 4 | 99 | 74.8 | 16 | 1424 | 54 | 1 | 3 | 23 | 999 | 5 | 99.5 | 59.5 | 0.5 | 151 | 7 | 0 | 2 | 7 | 1000 | 6 | 100 | 34.9 | 13.6 | 63 | 0 | 1 | 2 | 6 | 1001 | 7 | 99.999 | 40.6 | 6.1 | 16 | 2 | 0 | 2 | 4 | 1002 | 8 | 99.5 | 50.2 | 4.3 | 350 | 39 | 0 | 2 | 15 | 1003 | 9 | 99.5 | 55 | 17.3 | 428 | 5 | 1 | 2 | 67 | 1004 | 10 | 99.6 | 14.1 | 1 | 49 | 7 | 1 | 2 | 5 | 1005 | 11 | 99.9 | 24.9 | 0.3 | 85 | 1 | 0 | 2 | 8 | 1006 | 12 | 99.999 | 62.8 | 0.5 | 512 | 4 | 0 | 0 | 3 | 1007 | 13 | 97.5 | 54.6 | 5.1 | 1188 | 95 | 0 | 2 | 27 | 1008 | 14 | 100 | 48.6 | 0.7 | 79 | 0 | 0 | 2 | 4 | 1009 +------+--------+--------+------+------+-------+-----+-----+------+ 1011 Another study[ISOCORE2010] confirms the significant coverage increase 1012 provided by Remote LFAs. 1014 8.4. Comparison of LFA an RLFA results 1016 The table below provides a side by side comparison the LFA and the 1017 remote LFA results. This shows a significant improvement in the 1018 percentage of protected destinations and normally a modest 1019 improvement in the percentage of guaranteed node protected 1020 destinations. 1022 +------+--------+--------+---------+---------+ 1023 | topo | LFA | RLFA | LFA | RLFA | 1024 | | % prot | %prot | % gtd N | % gtd N | 1025 +------+--------+--------+---------+---------+ 1026 | 1 | 78.5 | 99.7 | 36.9 | 53.3 | 1027 | 2 | 97.3 | 97.5 | 52.4 | 52.4 | 1028 | 3 | 99.3 | 99.999 | 58 | 58.4 | 1029 | 4 | 83.1 | 99 | 63.1 | 74.8 | 1030 | 5 | 99 | 99.5 | 59.1 | 59.5 | 1031 | 6 | 86.4 |100 | 21.4 | 34.9 | 1032 | 7 | 93.9 | 99.999 | 35.4 | 40.6 | 1033 | 8 | 95.3 | 99.5 | 48.1 | 50.2 | 1034 | 9 | 82.2 | 99.5 | 49.5 | 55 | 1035 | 10 | 98.5 | 99.6 | 14.9 | 14.1 | 1036 | 11 | 99.6 | 99.9 | 24.8 | 24.9 | 1037 | 12 | 99.5 | 99.999 | 62.4 | 62.8 | 1038 | 13 | 92.4 | 97.5 | 51.6 | 54.6 | 1039 | 14 | 99.3 |100 | 48.6 | 48.6 | 1040 +------+--------+--------+---------+---------+ 1042 As shown in the table, remote LFA provides close to 100% prefix 1043 protection against link failure in 11 of the 14 topologies studied, 1044 and provides a significant improvement in two of the remaining three 1045 cases. Note that in an MPLS network the tunnels to the PQ nodes are 1046 always present as a property of an LDP-based deployment. 1048 In the small number of cases where there is no intersection between 1049 the (extended)P-space and the Q-space, a number of solutions to 1050 providing a suitable path between such disjoint regions in the 1051 network have been discussed in the working group. For example an 1052 explicitly routed LSP between P and Q might be set up using RSVP-TE 1053 or using Segment Routing [I-D.filsfils-spring-segment-routing]. Such 1054 extended repair methods are outside the scope of this document. 1056 9. Management Considerations 1058 The management of LFA and remote LFA is the subject of ongoing work 1059 withing the IETF [I-D.ietf-rtgwg-lfa-manageability] to which the 1060 reader is referred. Management considerations may lead to a 1061 preference for the use of a remote LFA over an available LFA. This 1062 preference is a matter for the network operator, and not a matter of 1063 protocol correctness. 1065 When the network re-converges, microloops [RFC5715] may form due to 1066 transient inconsistencies in the router FIBs. If it is determined 1067 that microloops are a significant issue in the deployment, then a 1068 suitable loop free convergence methods such as one of those described 1069 in [RFC5715], [RFC6976], or [I-D.litkowski-rtgwg-uloop-delay] should 1070 be implemented. 1072 10. Historical Note 1074 The basic concepts behind Remote LFA were invented in 2002 and were 1075 later included in [I-D.bryant-ipfrr-tunnels], submitted in 2004. 1077 [I-D.bryant-ipfrr-tunnels], targeted a 100% protection coverage and 1078 hence included additional mechanisms on top of the Remote LFA 1079 concept. The addition of these mechanisms made the proposal very 1080 complex and computationally intensive and it was therefore not 1081 pursued as a working group item. 1083 As explained in [RFC6571], the purpose of the LFA FRR technology is 1084 not to provide coverage at any cost. A solution for this already 1085 exists with MPLS TE FRR. MPLS TE FRR is a mature technology which is 1086 able to provide protection in any topology thanks to the explicit 1087 routing capability of MPLS TE. 1089 The purpose of LFA FRR technology is to provide for a simple FRR 1090 solution when such a solution is possible. The first step along this 1091 simplicity approach was "local" LFA [RFC5286]. This specification of 1092 "Remote LFA" is a natural second step. 1094 11. IANA Considerations 1096 There are no IANA considerations that arise from this architectural 1097 description of IPFRR. The RFC Editor may remove this section on 1098 publication. 1100 12. Security Considerations 1102 The security considerations of [RFC5286] also apply. 1104 Targeted LDP sessions and MPLS tunnels are normal features of an MPLS 1105 network and their use in this application raises no additional 1106 security concerns. 1108 To prevent their use as an attack vector IP repair tunnel endpoints 1109 (where used) SHOULD be assigned from a set of addresses that are not 1110 reachable from outside the routing domain. 1112 13. Acknowledgments 1114 The authors wish to thank Levente Csikor and Chris Bowers for their 1115 contribution to the cost based algorithm text. We thank Alia Atlas, 1116 Ross Callon, Stephane Litkowski, Bharath R, and Pushpasis Sarkar for 1117 their review of this document. 1119 14. References 1121 14.1. Normative References 1123 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1124 Requirement Levels", BCP 14, RFC 2119, March 1997. 1126 [RFC5286] Atlas, A. and A. Zinin, "Basic Specification for IP Fast 1127 Reroute: Loop-Free Alternates", RFC 5286, September 2008. 1129 14.2. Informative References 1131 [I-D.bryant-ipfrr-tunnels] 1132 Bryant, S., Filsfils, C., Previdi, S., and M. Shand, "IP 1133 Fast Reroute using tunnels", draft-bryant-ipfrr-tunnels-03 1134 (work in progress), November 2007. 1136 [I-D.filsfils-spring-segment-routing] 1137 Filsfils, C., Previdi, S., Bashandy, A., Decraene, B., 1138 Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., 1139 Ytti, S., Henderickx, W., Tantsura, J., and E. Crabbe, 1140 "Segment Routing Architecture", draft-filsfils-spring- 1141 segment-routing-04 (work in progress), July 2014. 1143 [I-D.ietf-ospf-routable-ip-address] 1144 Xu, X., Chunduri, U., and M. Bhatia, "Carrying Routable IP 1145 Addresses in OSPF RI LSA", draft-ietf-ospf-routable-ip- 1146 address-01 (work in progress), October 2014. 1148 [I-D.ietf-rtgwg-lfa-manageability] 1149 Litkowski, S., Decraene, B., Filsfils, C., Raza, K., 1150 Horneffer, M., and p. psarkar@juniper.net, "Operational 1151 management of Loop Free Alternates", draft-ietf-rtgwg-lfa- 1152 manageability-04 (work in progress), August 2014. 1154 [I-D.ietf-rtgwg-rlfa-node-protection] 1155 psarkar@juniper.net, p., Gredler, H., Hegde, S., Bowers, 1156 C., Litkowski, S., and H. Raghuveer, "Remote-LFA Node 1157 Protection and Manageability", draft-ietf-rtgwg-rlfa-node- 1158 protection-01 (work in progress), December 2014. 1160 [I-D.litkowski-rtgwg-uloop-delay] 1161 Litkowski, S., Decraene, B., Filsfils, C., and P. 1162 Francois, "Microloop prevention by introducing a local 1163 convergence delay", draft-litkowski-rtgwg-uloop-delay-03 1164 (work in progress), February 2014. 1166 [ISOCORE2010] 1167 So, N., Lin, T., and C. Chen, "LFA (Loop Free Alternates) 1168 Case Studies in Verizon's LDP Network", 2010. 1170 [RFC1701] Hanks, S., Li, T., Farinacci, D., and P. Traina, "Generic 1171 Routing Encapsulation (GRE)", RFC 1701, October 1994. 1173 [RFC1853] Simpson, W., "IP in IP Tunneling", RFC 1853, October 1995. 1175 [RFC2328] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998. 1177 [RFC3032] Rosen, E., Tappan, D., Fedorkow, G., Rekhter, Y., 1178 Farinacci, D., Li, T., and A. Conta, "MPLS Label Stack 1179 Encoding", RFC 3032, January 2001. 1181 [RFC3411] Harrington, D., Presuhn, R., and B. Wijnen, "An 1182 Architecture for Describing Simple Network Management 1183 Protocol (SNMP) Management Frameworks", STD 62, RFC 3411, 1184 December 2002. 1186 [RFC5036] Andersson, L., Minei, I., and B. Thomas, "LDP 1187 Specification", RFC 5036, October 2007. 1189 [RFC5305] Li, T. and H. Smit, "IS-IS Extensions for Traffic 1190 Engineering", RFC 5305, October 2008. 1192 [RFC5340] Coltun, R., Ferguson, D., Moy, J., and A. Lindem, "OSPF 1193 for IPv6", RFC 5340, July 2008. 1195 [RFC5424] Gerhards, R., "The Syslog Protocol", RFC 5424, March 2009. 1197 [RFC5714] Shand, M. and S. Bryant, "IP Fast Reroute Framework", RFC 1198 5714, January 2010. 1200 [RFC5715] Shand, M. and S. Bryant, "A Framework for Loop-Free 1201 Convergence", RFC 5715, January 2010. 1203 [RFC6119] Harrison, J., Berger, J., and M. Bartlett, "IPv6 Traffic 1204 Engineering in IS-IS", RFC 6119, February 2011. 1206 [RFC6571] Filsfils, C., Francois, P., Shand, M., Decraene, B., 1207 Uttaro, J., Leymann, N., and M. Horneffer, "Loop-Free 1208 Alternate (LFA) Applicability in Service Provider (SP) 1209 Networks", RFC 6571, June 2012. 1211 [RFC6976] Shand, M., Bryant, S., Previdi, S., Filsfils, C., 1212 Francois, P., and O. Bonaventure, "Framework for Loop-Free 1213 Convergence Using the Ordered Forwarding Information Base 1214 (oFIB) Approach", RFC 6976, July 2013. 1216 [RFC6987] Retana, A., Nguyen, L., Zinin, A., White, R., and D. 1217 McPherson, "OSPF Stub Router Advertisement", RFC 6987, 1218 September 2013. 1220 Authors' Addresses 1222 Stewart Bryant 1223 Cisco Systems 1224 250, Longwater, Green Park, 1225 Reading RG2 6GB, UK 1226 UK 1228 Email: stbryant@cisco.com 1230 Clarence Filsfils 1231 Cisco Systems 1232 De Kleetlaan 6a 1233 1831 Diegem 1234 Belgium 1236 Email: cfilsfil@cisco.com 1238 Stefano Previdi 1239 Cisco Systems 1241 Email: sprevidi@cisco.com 1243 Mike Shand 1244 Independent Contributor 1246 Email: imc.shand@gmail.com 1247 Ning So 1248 Vinci Systems 1250 Email: ning.so@vinci-systems.com