idnits 2.17.1 draft-ietf-rtgwg-lfa-manageability-02.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 -- The document date (February 3, 2014) is 3734 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) == Unused Reference: 'I-D.litkowski-rtgwg-lfa-rsvpte-cooperation' is defined on line 911, but no explicit reference was found in the text == Unused Reference: 'I-D.psarkar-isis-node-admin-tag' is defined on line 917, but no explicit reference was found in the text == Unused Reference: 'RFC3630' is defined on line 931, but no explicit reference was found in the text == Unused Reference: 'RFC3906' is defined on line 935, but no explicit reference was found in the text == Unused Reference: 'RFC4090' is defined on line 939, but no explicit reference was found in the text == Unused Reference: 'RFC5305' is defined on line 943, but no explicit reference was found in the text == Unused Reference: 'RFC5714' is defined on line 946, but no explicit reference was found in the text == Unused Reference: 'RFC5715' is defined on line 949, but no explicit reference was found in the text ** Obsolete normative reference: RFC 4205 (Obsoleted by RFC 5307) == Outdated reference: A later version (-11) exists of draft-ietf-rtgwg-remote-lfa-04 == Outdated reference: A later version (-03) exists of draft-psarkar-isis-node-admin-tag-00 == Outdated reference: A later version (-05) exists of draft-psarkar-rtgwg-rlfa-node-protection-03 Summary: 1 error (**), 0 flaws (~~), 12 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Routing Area Working Group S. Litkowski 3 Internet-Draft B. Decraene 4 Intended status: Standards Track Orange 5 Expires: August 7, 2014 C. Filsfils 6 K. Raza 7 Cisco Systems 8 M. Horneffer 9 Deutsche Telekom 10 P. Sarkar 11 Juniper Networks 12 February 3, 2014 14 Operational management of Loop Free Alternates 15 draft-ietf-rtgwg-lfa-manageability-02 17 Abstract 19 Loop Free Alternates (LFA), as defined in RFC 5286 is an IP Fast 20 ReRoute (IP FRR) mechanism enabling traffic protection for IP traffic 21 (and MPLS LDP traffic by extension). Following first deployment 22 experiences, this document provides operational feedback on LFA, 23 highlights some limitations, and proposes a set of refinements to 24 address those limitations. It also proposes required management 25 specifications. 27 This proposal is also applicable to remote LFA solution. 29 Requirements Language 31 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 32 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 33 document are to be interpreted as described in [RFC2119]. 35 Status of this Memo 37 This Internet-Draft is submitted in full conformance with the 38 provisions of BCP 78 and BCP 79. 40 Internet-Drafts are working documents of the Internet Engineering 41 Task Force (IETF). Note that other groups may also distribute 42 working documents as Internet-Drafts. The list of current Internet- 43 Drafts is at http://datatracker.ietf.org/drafts/current/. 45 Internet-Drafts are draft documents valid for a maximum of six months 46 and may be updated, replaced, or obsoleted by other documents at any 47 time. It is inappropriate to use Internet-Drafts as reference 48 material or to cite them other than as "work in progress." 49 This Internet-Draft will expire on August 7, 2014. 51 Copyright Notice 53 Copyright (c) 2014 IETF Trust and the persons identified as the 54 document authors. All rights reserved. 56 This document is subject to BCP 78 and the IETF Trust's Legal 57 Provisions Relating to IETF Documents 58 (http://trustee.ietf.org/license-info) in effect on the date of 59 publication of this document. Please review these documents 60 carefully, as they describe your rights and restrictions with respect 61 to this document. Code Components extracted from this document must 62 include Simplified BSD License text as described in Section 4.e of 63 the Trust Legal Provisions and are provided without warranty as 64 described in the Simplified BSD License. 66 Table of Contents 68 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 69 2. Operational issues with default LFA tie breakers . . . . . . . 4 70 2.1. Case 1: Edge router protecting core failures . . . . . . . 5 71 2.2. Case 2: Edge router choosen to protect core failures 72 while core LFA exists . . . . . . . . . . . . . . . . . . 6 73 2.3. Case 3: suboptimal core alternate choice . . . . . . . . . 7 74 2.4. Case 4: ISIS overload bit on LFA computing node . . . . . 8 75 3. Need for coverage monitoring . . . . . . . . . . . . . . . . . 8 76 4. Need for LFA activation granularity . . . . . . . . . . . . . 9 77 5. Configuration requirements . . . . . . . . . . . . . . . . . . 9 78 5.1. LFA enabling/disabling scope . . . . . . . . . . . . . . . 9 79 5.2. Policy based LFA selection . . . . . . . . . . . . . . . . 10 80 5.2.1. Connected vs remote alternates . . . . . . . . . . . . 10 81 5.2.2. Mandatory criteria . . . . . . . . . . . . . . . . . . 11 82 5.2.3. Enhanced criteria . . . . . . . . . . . . . . . . . . 11 83 5.2.4. Retrieving alternate path attributes . . . . . . . . . 11 84 5.2.5. ECMP LFAs . . . . . . . . . . . . . . . . . . . . . . 13 85 5.2.6. SRLG . . . . . . . . . . . . . . . . . . . . . . . . . 14 86 5.2.7. Link coloring . . . . . . . . . . . . . . . . . . . . 15 87 5.2.8. Bandwidth . . . . . . . . . . . . . . . . . . . . . . 16 88 5.2.9. Neighbor preference . . . . . . . . . . . . . . . . . 17 89 6. Operational aspects . . . . . . . . . . . . . . . . . . . . . 18 90 6.1. ISIS overload bit on LFA computing node . . . . . . . . . 18 91 6.2. Manual triggering of FRR . . . . . . . . . . . . . . . . . 18 92 6.3. Required local information . . . . . . . . . . . . . . . . 19 93 6.4. Coverage monitoring . . . . . . . . . . . . . . . . . . . 20 94 6.5. LFA and network planning . . . . . . . . . . . . . . . . . 20 95 7. Security Considerations . . . . . . . . . . . . . . . . . . . 21 96 8. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 21 97 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 21 98 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 21 99 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 21 100 11.1. Normative References . . . . . . . . . . . . . . . . . . . 21 101 11.2. Informative References . . . . . . . . . . . . . . . . . . 21 102 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 23 104 1. Introduction 106 Following the first deployments of Loop Free Alternates (LFA), this 107 document provides feedback to the community about the management of 108 LFA. 110 Section 2 provides real uses cases illustrating some limitations 111 and suboptimal behavior. 113 Section 3 proposes requirements for activation granularity and 114 policy based selection of the alternate. 116 Section 4 express requirements for the operational management of 117 LFA. 119 2. Operational issues with default LFA tie breakers 121 [RFC5286] introduces the notion of tie breakers when selecting the 122 LFA among multiple candidate alternate next-hops. When multiple LFA 123 exist, RFC 5286 has favored the selection of the LFA providing the 124 best coverage of the failure cases. While this is indeed a goal, 125 this is one among multiple and in some deployment this lead to the 126 selection of a suboptimal LFA. The following sections details real 127 use cases of such limitations. 129 Note that the use case of per-prefix LFA is assumed throughout this 130 analysis. 132 2.1. Case 1: Edge router protecting core failures 134 R1 --------- R2 ---------- R3 --------- R4 135 | 1 100 1 | 136 | | 137 | 100 | 100 138 | | 139 | 1 100 1 | 140 R5 --------- R6 ---------- R7 --------- R8 -- R9 - PE1 141 | | | | 142 | 5k | 5k | 5k | 5k 143 | | | | 144 +--- n*PEx ---+ +---- PE2 ----+ 145 | 146 | 147 PEy 149 Figure 1 151 Rx routers are core routers using n*10G links. PEs are connected 152 using links with lower bandwidth. PEx are a set of PEs connected to 153 R5 and R6. 155 In figure 1, let us consider the traffic flowing from PE1 to PEx. 156 The nominal path is R9-R8-R7-R6-PEx. Let us consider the failure of 157 link R7-R8. For R8, R4 is not an LFA and the only available LFA is 158 PE2. 160 When the core link R8-R7 fails, R8 switches all traffic destined to 161 all the PEx towards the edge node PE2. Hence an edge node and edge 162 links are used to protect the failure of a core link. Typically, 163 edge links have less capacity than core links and congestion may 164 occur on PE2 links. Note that although PE2 was not directly affected 165 by the failure, its links become congested and its traffic will 166 suffer from the congestion. 168 In summary, in case of failure, the impact on customer traffic is: 170 o From PE2 point of view : 172 * without LFA: no impact 174 * with LFA: traffic is partially dropped (but possibly 175 prioritized by a QoS mechanism). It must be highlighted that 176 in such situation, traffic not affected by the failure may be 177 affected by the congestion. 179 o From R8 point of view: 181 * without LFA: traffic is totally dropped until convergence 182 occurs. 184 * with LFA: traffic is partially dropped (but possibly 185 prioritized by a QoS mechanism). 187 Besides the congestion aspects of using an Edge router as an 188 alternate to protect a core failure, a service provider may consider 189 this as a bad routing design and would like to prevent it. 191 2.2. Case 2: Edge router choosen to protect core failures while core 192 LFA exists 194 R1 --------- R2 ------------ R3 --------- R4 195 | 1 100 | 1 | 196 | | | 197 | 100 | 30 | 30 198 | | | 199 | 1 50 50 | 10 | 200 R5 -------- R6 ---- R10 ---- R7 -------- R8 --- R9 - PE1 201 | | \ | 202 | 5000 | 5000 \ 5000 | 5000 203 | | \ | 204 +--- n*PEx --+ +----- PE2 ----+ 205 | 206 | 207 PEy 209 Figure 2 211 Rx routers are core routers meshed with n*10G links. PEs are meshed 212 using links with lower bandwidth. 214 In the figure 2, let us consider the traffic coming from PE1 to PEx. 215 Nominal path is R9-R8-R7-R10-R6-PEx. Let us consider the failure of 216 the link R7-R8. For R8, R4 is a link-protecting LFA and PE2 is a 217 node-protecting LFA. PE2 is chosen as best LFA due to its better 218 protection type. Just like in case 1, this may lead to congestion on 219 PE2 links upon LFA activation. 221 2.3. Case 3: suboptimal core alternate choice 223 +--- PE3 --+ 224 / \ 225 1000 / \ 1000 226 / \ 227 +----- R1 ---------------- R2 ----+ 228 | | 500 | | 229 | 10 | | | 10 230 | | | | 231 R5 | 10 | 10 R7 232 | | | | 233 | 10 | | | 10 234 | | 500 | | 235 +---- R3 ---------------- R4 -----+ 236 \ / 237 1000 \ / 1000 238 \ / 239 +--- PE1 ---+ 241 Figure 3 243 Rx routers are core routers. R1-R2 and R3-R4 links are 1G links. 244 All others inter Rx links are 10G links. 246 In the figure above, let us consider the failure of link R1-R3. For 247 destination PE3, R3 has two possible alternates: 249 o R4, which is node-protecting 251 o R5, which is link-protecting 253 R4 is chosen as best LFA due to its better protection type. However, 254 it may not be desirable to use R4 for bandwidth capacity reason. A 255 service provider may prefer to use high bandwidth links as prefered 256 LFA. In this example, prefering shortest path over protection type 257 may achieve the expected behavior, but in cases where metric are not 258 reflecting bandwidth, it would not work and some other criteria would 259 need to be involved when selecting the best LFA. 261 2.4. Case 4: ISIS overload bit on LFA computing node 263 P1 P2 264 | \ / | 265 50 | 50 \/ 50 | 50 266 | /\ | 267 PE1-+ +-- PE2 268 \ / 269 45 \ / 45 270 -PE3-+ 271 (OL set) 273 Figure 4 275 In the figure above, PE3 has its overload bit set (permanently, for 276 design reason) and wants to protect traffic using LFA for destination 277 PE2. 279 On PE3, the loopfree condition is not satisified : 100 !< 45 + 45. 280 PE1 is thus not considered as an LFA. However thanks to the overload 281 bit set on PE3, we know that PE1 is loopfree so PE1 is an LFA to 282 reach PE2. 284 In case of overload condition set on a node, LFA behavior must be 285 clarified. 287 3. Need for coverage monitoring 289 As per [RFC6571], LFA coverage highly depends on the used network 290 topology. Even if remote LFA ([I-D.ietf-rtgwg-remote-lfa]) extends 291 significantly the coverage of the basic LFA specification, there is 292 still some cases where protection would not be available. As network 293 topologies are constantly evolving (network extension, capacity 294 addings, latency optimization ...), the protection coverage may 295 change. Fast reroute functionality may be critical for some services 296 supported by the network, a service provider must constantly know 297 what protection coverage is currently available on the network. 298 Moreover, predicting the protection coverage in case of network 299 topology change is mandatory : using network simulation tool and 300 whatif scenarios functionnality, a service provider may be able to 301 evaluate protection coverage after a topology change and may be able 302 to adjust the topology change to cover the primary need (e.g. latency 303 optimization or bandwidth increase) as well as LFA protection. 305 4. Need for LFA activation granularity 307 As all FRR mechanism, LFA installs backup paths in FIB. Depending of 308 the hardware used by a service provider, FIB ressource may be 309 critical. Activating LFA, by default, on all available components 310 (IGP topologies, interface, address families ...) may lead to waste 311 of FIB ressource as generally in a network only few destinations 312 should be protected (e.g. loopback addresses supporting MPLS 313 services) compared to the amount of destinations in RIB. 315 Moreover a service provider may implement multiple different FRR 316 mechanism in its networks for different usages (MRT, TE FRR), 317 computing LFAs for prefixes or interfaces that are already protected 318 by another mechanism is useless. 320 5. Configuration requirements 322 Controlling best alternate and LFA activation granularity is a 323 requirement for Service Providers. This section defines 324 configuration requirements for LFA. 326 5.1. LFA enabling/disabling scope 328 The granularity of LFA activation should be controlled (as alternate 329 nexthop consume memory in forwarding plane). 331 An implementation of LFA SHOULD allow its activation with the 332 following criteria: 334 o Per address-family : ipv4 unicast, ipv6 unicast, LDP IPv4 unicast, 335 LDP IPv6 unicast ... 337 o Per routing context : VRF, virtual/logical router, global routing 338 table, ... 340 o Per interface 342 o Per protocol instance, topology, area 344 o Per prefixes: prefix protection SHOULD have a better priority 345 compared to interface protection. This means that if a specific 346 prefix must be protected due to a configuration request, LFA must 347 be computed and installed for this prefix even if the primary 348 outgoing interface is not configured for protection. 350 5.2. Policy based LFA selection 352 When multiple alternates exist, LFA selection algorithm is based on 353 tie breakers. Current tie breakers do not provide sufficient control 354 on how the best alternate is chosen. This document proposes an 355 enhanced tie breaker allowing service providers to manage all 356 specific cases: 358 1. An implementation of LFA SHOULD support policy-based decision for 359 determining the best LFA. 361 2. Policy based decision SHOULD be based on multiple criterions, 362 with each criteria having a level of preference. 364 3. If the defined policy does not permit to determine a unique best 365 LFA, an implementation SHOULD pick only one based on its own 366 decision, as a default behavior. An implementation SHOULD also 367 support election of multiple LFAs, for loadbalancing purposes. 369 4. Policy SHOULD be applicable to a protected interface or to a 370 specific set of destinations. In case of application on the 371 protected interface, all destinations primarily routed on this 372 interface SHOULD use the interface policy. 374 5. It is an implementation choice to reevaluate policy dynamically 375 or not (in case of policy change). If a dynamic approach is 376 chosen, the implementation SHOULD recompute the best LFAs and 377 reinstall them in FIB, without service disruption. If a non- 378 dynamic approach is chosen, the policy would be taken into 379 account upon the next IGP event. In this case, the 380 implementation SHOULD support a command to manually force the 381 recomputation/reinstallation of LFAs. 383 5.2.1. Connected vs remote alternates 385 In addition to direct LFAs, tunnels (e.g. IP, LDP or RSVP-TE) to 386 distant routers may be used to complement LFA coverage (tunnel tail 387 used as virtual neighbor). When a router has multiple alternate 388 candidates for a specific destination, it may have connected 389 alternates and remote alternates reachable via a tunnel. Connected 390 alternates may not always provide an optimal routing path and it may 391 be preferable to select a remote alternate over a connected 392 alternate. The usage of tunnels to extend LFA coverage is described 393 in [I-D.ietf-rtgwg-remote-lfa]. 395 In figure 1, there is no core alternate for R8 to reach PEs located 396 behind R6, so R8 is using PE2 as alternate, which may generate 397 congestion when FRR is activated. Instead, we could have a remote 398 core alternate for R8 to protect PEs destinations. For example, a 399 tunnel from R8 to R3 would ensure LFA protection without using an 400 edge router to protect a core router. 402 When selecting the best alternate, the selection algorithm MUST 403 consider all available alternates (connected or tunnel). Especially, 404 computation of PQ set ([I-D.ietf-rtgwg-remote-lfa]) SHOULD be 405 performed before best alternate selection. 407 5.2.2. Mandatory criteria 409 An implementation of LFA MUST support the following criteria: 411 o Non candidate link: A link marked as "non candidate" will never be 412 used as LFA. 414 o A primary nexthop being protected by another primary nexthop of 415 the same prefix (ECMP case). 417 o Type of protection provided by the alternate: link protection, 418 node protection. In case of node protection preference, an 419 implementation SHOULD support fallback to link protection if node 420 protection is not available. 422 o Shortest path: lowest IGP metric used to reach the destination. 424 o SRLG (as defined in [RFC5286] Section 3). 426 5.2.3. Enhanced criteria 428 An implementation of LFA SHOULD support the following enhanced 429 criteria: 431 o Downstreamness of a neighbor : preference of a downstream path 432 over a non downstream path SHOULD be configurable. 434 o Link coloring with : include, exclude and preference based system. 436 o Link Bandwidth. 438 o Neighbor preference. 440 5.2.4. Retrieving alternate path attributes 442 The policy to select the best alternate evaluate multiple criterions 443 (e.g. metric, SRLG, link colors ...) which first need to be computed 444 for each alternate.. In order to compare the different alternate 445 path, a router must retrieve the attributes of each alternate path. 447 The alternate path is composed of two distinct parts : PLR to 448 alternate and alternate to destination. 450 5.2.4.1. Connected alternate 452 For alternate path using a connected alternate : 454 o attributes from PLR to alternate path are retrieved from the 455 interface connected to the alternate. 457 o attributes from alternate to destination path are retrieved from 458 SPF rooted at the alternate. As the alternate is a connected 459 alternate, the SPF has already been computed to find the 460 alternate, so there is no need of additional computation. 462 5.2.4.2. Remote alternate 464 For alternate path using a remote alternate (tunnel) : 466 o attributes from the PLR to alternate path are retrieved using the 467 PLR's primary SPF if P space is used or using the neighbor's SPF 468 if extended P space is used, combined with the attributes of the 469 link(s) to reach that neighbor. In both cases, no additional SPF 470 is required. 472 o attributes from alternate to destination path are retrieved from 473 SPF rooted at the remote alternate. An additional forward SPF is 474 required for each remote alternate as indicated in 475 [I-D.psarkar-rtgwg-rlfa-node-protection] section 3.2.. 477 The number of remote alternates may be very high, simulations shown 478 that hundred's of PQs may exist for a single interface being 479 protected. Running a forward SPF for every PQ-node in the network is 480 not scalable. 482 To handle this situation, it is needed to limit the number of remote 483 alternates to be evaluated to a finite number before collecting 484 alternate path attributes and running the policy evaluation. 485 [I-D.psarkar-rtgwg-rlfa-node-protection] Section 2.3.3 provides a way 486 to reduce the number of PQ to be evaluated. 488 Link Remote Remote 489 alternate alternate alternate 490 ------------- ------------------ ------------- 491 Alternates | LFA | | rLFA (PQs) | | Static | 492 sources | | | | | tunnels | 493 ------------- ------------------ ------------- 494 | | | 495 | | | 496 | ---------------------- | 497 | | Prune some PQs | | 498 | | (sorting strategy) | | 499 | ---------------------- | 500 | | | 501 | | | 502 ------------------------------------------------ 503 | Collect alternate attributes | 504 ------------------------------------------------ 505 | 506 | 507 ------------------------- 508 | Evaluate policy | 509 ------------------------- 510 | 511 | 512 Best alternates 514 5.2.5. ECMP LFAs 516 10 517 PE2 - PE3 518 | | 519 50 | 5 | 50 520 P1----P2 521 \\ // 522 50 \\ // 50 523 PE1 525 Figure 5 527 Links between P1 and PE1 are L1 and L2, links between P2 and PE1 are 528 L3 and L4 530 In the figure above, primary path from PE1 to PE2 is through P1 using 531 ECMP on two parallel links L1 and L2. In case of standard ECMP 532 behavior, if L1 is failing, postconvergence nexthop would become L2 533 and there would be no longer ECMP. If LFA is activated, as stated in 534 [RFC5286] Section 3.4., "alternate next-hops may themselves also be 535 primary next-hops, but need not be" and "alternate next-hops should 536 maximize the coverage of the failure cases". In this scenario there 537 is no alternate providing node protection, LFA will so prefer L2 as 538 alternate to protect L1 which makes sense compared to postconvergence 539 behavior. 541 Considering a different scenario using figure 5, where L1 and L2 are 542 configured as a layer 3 bundle using a local feature, as well as 543 L3/L4 being a second layer 3 bundle. Layer 3 bundles are configured 544 as if a link in the bundle is failing, the traffic must be rerouted 545 out of the bundle. Layer 3 bundles are generally introduced to 546 increase bandwidth between nodes. In nominal situation, ECMP is 547 still available from PE1 to PE2, but if L1 is failing, 548 postconvergence nexthop would become ECMP on L3 and L4. In this 549 case, LFA behavior SHOULD be adapted in order to reflect the 550 bandwidth requirement. 552 We would expect the following FIB entry on PE1 : 554 On PE1 : PE2 +--> ECMP -> L1 555 | | 556 | +----> L2 557 | 558 +--> LFA(ECMP) -> L3 559 | 560 +---------> L4 562 If L1 or L2 is failing, traffic must be switched on the LFA ECMP 563 bundle rather than using the other primary nexthop. 565 As mentioned in [RFC5286] Section 3.4., protecting a link within an 566 ECMP by another primary nexthop is not a MUST. Moreover, we already 567 presented in this document, that maximizing the coverage of the 568 failure case may not be the right approach and policy based choice of 569 alternate may be preferred. 571 An implementation SHOULD permit to prefer a primary nexthop by 572 another primary nexthop with the possibility to deactivate this 573 criteria. An implementation SHOULD permit to use an ECMP bundle as a 574 LFA. 576 5.2.6. SRLG 578 [RFC5286] Section 3. proposes to reuse GMPLS IGP extensions to encode 579 SRLGs ([RFC4205] and [RFC4203]). The section is also describing the 580 algorithm to compute SRLG protection. 582 When SRLG protection is computed, and implementation SHOULD permit to 583 : 585 o Exclude alternates violating SRLG. 587 o Maintain a preference system between alternates based on number of 588 SRLG violations : more violations = less preference. 590 When applying SRLG criteria, the SRLG violation check SHOULD be 591 performed on source to alternate as well as alternate to destination 592 paths. In the case of remote LFA, PQ to destination path attributes 593 would be retrieved from SPT rooted at PQ. 595 5.2.7. Link coloring 597 Link coloring is a powerful system to control the choice of 598 alternates. Protecting interfaces are tagged with colors. Protected 599 interfaces are configured to include some colors with a preference 600 level, and exclude others. 602 Link color information SHOULD be signalled in the IGP. How 603 signalling is done is out of scope of the document but it may be 604 useful to reuse existing admin-groups from traffic-engineering 605 extensions. 607 PE2 608 | +---- P4 609 | / 610 PE1 ---- P1 --------- P2 611 | 10Gb 612 1Gb | 613 | 614 P3 616 Figure 5 618 Example : P1 router is connected to three P routers and two PEs. 620 P1 is configured to protect the P1-P4 link. We assume that given the 621 topology, all neighbors are candidate LFA. We would like to enforce 622 a policy in the network where only a core router may protect against 623 the failure of a core link, and where high capacity links are 624 prefered. 626 In this example, we can use the proposed link coloring by: 628 o Marking PEs links with color RED 630 o Marking 10Gb CORE link with color BLUE 632 o Marking 1Gb CORE link with color YELLOW 634 o Configured the protected interface P1->P4 with : 636 * Include BLUE, preference 200 638 * Include YELLOW, preference 100 640 * Exclude RED 642 Using this, PE links will never be used to protect against P1-P4 link 643 failure and 10Gb link will be be preferred. 645 The main advantage of this solution is that it can easily be 646 duplicated on other interfaces and other nodes without change. A 647 Service Provider has only to define the color system (associate color 648 with a significance), as it is done already for TE affinities or BGP 649 communities. 651 An implementation of link coloring: 653 o SHOULD support multiple include and exclude colors on a single 654 protected interface. 656 o SHOULD provide a level of preference between included colors. 658 o SHOULD support multiple colors configuration on a single 659 protecting interface. 661 5.2.8. Bandwidth 663 As mentionned in previous sections, not taking into account bandwidth 664 of an alternate could lead to congestion during FRR activation. We 665 propose to base the bandwidth criteria on the link speed information 666 for the following reason : 668 o if a router S has a set of X destinations primarly forwarded to N, 669 using per prefix LFA may lead to have a subset of X protected by a 670 neighbor N1, another subset by N2, another subset by Nx ... 672 o S is not aware about traffic flows to each destination and is not 673 able to evaluate how much traffic will be sent to N1,N2, ... Nx 674 in case of FRR activation. 676 Based on this, it is not useful to gather available bandwidth on 677 alternate paths, as the router does not know how much bandwidth it 678 requires for protection. The proposed link speed approach provides a 679 good approximation with a small cost as information is easily 680 available. 682 The bandwidth criteria of the policy framework SHOULD work in two 683 ways : 685 o PRUNE : exclude a LFA if link speed to reach it is lower than the 686 link speed of the primary nexthop interface. 688 o PREFER : prefer a LFA based on his bandwidth to reach it compared 689 to the link speed of the primary nexthop interface. 691 5.2.9. Neighbor preference 693 Rather than tagging interface on each node (using link color) to 694 identify neighbor node type (as example), it would be helpful if 695 routers could be identified in the IGP. This would permit a grouped 696 processing on multiple nodes. As an implementation need to exclude 697 some specific neighbors (see Section 5.2.3), an implementation : 699 o SHOULD be able to give a preference to specific neighbor. 701 o SHOULD be able to give a preference to a group of neighbor. 703 o SHOULD be able to exclude a group of neighbor. 705 A specific neighbor may be identified by its interface, IP address or 706 router ID and group of neighbors may be identified by a marker (tag). 708 Consider the following network: 710 PE3 711 | 712 | 713 PE2 714 | +---- P4 715 | / 716 PE1 ---- P1 -------- P2 717 | 10Gb 718 1Gb | 719 | 720 P3 722 Figure 6 724 In the example above, each node is configured with a specific tag 725 flooded through the IGP. 727 o PE1,PE3: 200 (non candidate). 729 o PE2: 100 (edge/core). 731 o P1,P2,P3: 50 (core). 733 A simple policy could be configured on P1 to choose the best 734 alternate for P1->P4 based on router function/role as follows : 736 o criteria 1 -> neighbor preference: exclude tag 100 and 200. 738 o criteria 2 -> bandwidth. 740 6. Operational aspects 742 6.1. ISIS overload bit on LFA computing node 744 In [RFC5286], Section 3.5, the setting of the overload bit condition 745 in LFA computation is only taken into account for the case where a 746 neighbor has the overload bit set. 748 In addition to RFC 5286 inequality 1 Loop-Free Criterion 749 (Distance_opt(N, D) < Distance_opt(N, S) + Distance_opt(S, D)), the 750 IS-IS overload bit of the LFA calculating neighbor (S) SHOULD be 751 taken into account. Indeed, if it has the overload bit set, no 752 neighbor will loop back to traffic to itself. 754 6.2. Manual triggering of FRR 756 Service providers often perform manual link shutdown (using router 757 CLI) to perform some network changes/tests. A manual link shutdown 758 may be done at multiple level : physical interface, logical 759 interface, IGP interface, BFD session ... Especially testing or 760 troubleshooting FRR requires to perform the manual shutdown on the 761 remote end of the link as generally a local shutdown would not 762 trigger FRR. 764 To enhance such situation, an implementation SHOULD support 765 triggering/activating LFA Fast Reroute for a given link when a manual 766 shutdown is done on a component that currently supports FRR 767 activation. 769 For example : 771 o if an implementation supports FRR activation upon BFD session down 772 event, this implementation SHOULD support FRR activation when a 773 manual shutdown is done on the BFD session. But if an 774 implementation does not support FRR activation on BFD session 775 down, there is no need for this implementation to support FRR 776 activation on manual shutdown of BFD session. 778 o if an implementation supports FRR activation on physical link down 779 event (e.g. Rx laser Off detection, or error threshold raised 780 ...), this implementation SHOULD support FRR activation when a 781 manual shutdown at physical interface is done. But if an 782 implementation does not support FRR activation on physical link 783 down event, there is no need for this implementation to support 784 FRR activation on manual physical link shutdown. 786 6.3. Required local information 788 LFA introduction requires some enhancement in standard routing 789 information provided by implementations. Moreover, due to the non 790 100% coverage, coverage informations is also required. 792 Hence an implementation : 794 o MUST be able to display, for every prefixes, the primary nexthop 795 as well as the alternate nexthop information. 797 o MUST provide coverage information per activation domain of LFA 798 (area, level, topology, instance, virtual router, address family 799 ...). 801 o MUST provide number of protected prefixes as well as non protected 802 prefixes globally. 804 o SHOULD provide number of protected prefixes as well as non 805 protected prefixes per link. 807 o MAY provide number of protected prefixes as well as non protected 808 prefixes per priority if implementation supports prefix-priority 809 insertion in RIB/FIB. 811 o SHOULD provide a reason for chosing an alternate (policy and 812 criteria) and for excluding an alternate. 814 o SHOULD provide the list of non protected prefixes and the reason 815 why they are not protected (no protection required or no alternate 816 available). 818 6.4. Coverage monitoring 820 It is pretty easy to evaluate the coverage of a network in a nominal 821 situation, but topology changes may change the coverage. In some 822 situations, the network may no longer be able to provide the required 823 level of protection. Hence, it becomes very important for service 824 providers to get alerted about changes of coverage. 826 An implementation SHOULD : 828 o provide an alert system if total coverage (for a node) is below a 829 defined threshold or comes back to a normal situation. 831 o provide an alert system if coverage of a specific link is below a 832 defined threshold or comes back to a normal situation. 834 An implementation MAY : 836 o provide an alert system if a specific destination is not protected 837 anymore or when protection comes back up for this destination 839 Although the procedures for providing alerts are beyond the scope of 840 this document, we recommend that implementations consider standard 841 and well used mechanisms like syslog or SNMP traps. 843 6.5. LFA and network planning 845 The operator may choose to run simulations in order to ensure full 846 coverage of a certain type for the whole network or a given subset of 847 the network. This is particularly likely if he operates the network 848 in the sense of the third backbone profiles described in [RFC6571], 849 that is, he seeks to design and engineer the network topology in a 850 way that a certain coverage is always achieved. Obviously a complete 851 and exact simulation of the IP FRR coverage can only be achieved, if 852 the behavior is deterministic and if the algorithm used is available 853 to the simulation tool. Thus, an implementation SHOULD: 855 o Behave deterministic in its selection LFA process. I.e. in the 856 same topology and with the same policy configuration, the 857 implementation MUST always choose the same alternate for a given 858 prefix. 860 o Document its behavior. The implementation SHOULD provide enough 861 documentation of its behavior that allows an implementer of a 862 simulation tool, to foresee the exact choice of the LFA 863 implementation for every prefix in a given topology. This SHOULD 864 take into account all possible policy configuration options. One 865 possible way to document this behavior is to disclose the 866 algorithm used to choose alternates. 868 7. Security Considerations 870 This document does not introduce any change in security consideration 871 compared to [RFC5286]. 873 8. Contributors 875 Significant contributions were made by Pierre Francois, Hannes 876 Gredler, Chris Bowers, Jeff Tantsura, Uma Chunduri and Mustapha 877 Aissaoui which the authors would like to acknowledge. 879 9. Acknowledgements 881 10. IANA Considerations 883 This document has no action for IANA. 885 11. References 887 11.1. Normative References 889 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 890 Requirement Levels", BCP 14, RFC 2119, March 1997. 892 [RFC4203] Kompella, K. and Y. Rekhter, "OSPF Extensions in Support 893 of Generalized Multi-Protocol Label Switching (GMPLS)", 894 RFC 4203, October 2005. 896 [RFC4205] Kompella, K. and Y. Rekhter, "Intermediate System to 897 Intermediate System (IS-IS) Extensions in Support of 898 Generalized Multi-Protocol Label Switching (GMPLS)", 899 RFC 4205, October 2005. 901 [RFC5286] Atlas, A. and A. Zinin, "Basic Specification for IP Fast 902 Reroute: Loop-Free Alternates", RFC 5286, September 2008. 904 11.2. Informative References 906 [I-D.ietf-rtgwg-remote-lfa] 907 Bryant, S., Filsfils, C., Previdi, S., Shand, M., and S. 908 Ning, "Remote LFA FRR", draft-ietf-rtgwg-remote-lfa-04 909 (work in progress), November 2013. 911 [I-D.litkowski-rtgwg-lfa-rsvpte-cooperation] 912 Litkowski, S., Decraene, B., Filsfils, C., and K. Raza, 913 "Interactions between LFA and RSVP-TE", 914 draft-litkowski-rtgwg-lfa-rsvpte-cooperation-02 (work in 915 progress), August 2013. 917 [I-D.psarkar-isis-node-admin-tag] 918 psarkar@juniper.net, p., Gredler, H., Hegde, S., 919 Raghuveer, H., Litkowski, S., and B. Decraene, 920 "Advertising Per-node Admin Tags in IS-IS", 921 draft-psarkar-isis-node-admin-tag-00 (work in progress), 922 October 2013. 924 [I-D.psarkar-rtgwg-rlfa-node-protection] 925 psarkar@juniper.net, p., Gredler, H., Hegde, S., 926 Raghuveer, H., cbowers@juniper.net, c., and S. Litkowski, 927 "Remote-LFA Node Protection and Manageability", 928 draft-psarkar-rtgwg-rlfa-node-protection-03 (work in 929 progress), December 2013. 931 [RFC3630] Katz, D., Kompella, K., and D. Yeung, "Traffic Engineering 932 (TE) Extensions to OSPF Version 2", RFC 3630, 933 September 2003. 935 [RFC3906] Shen, N. and H. Smit, "Calculating Interior Gateway 936 Protocol (IGP) Routes Over Traffic Engineering Tunnels", 937 RFC 3906, October 2004. 939 [RFC4090] Pan, P., Swallow, G., and A. Atlas, "Fast Reroute 940 Extensions to RSVP-TE for LSP Tunnels", RFC 4090, 941 May 2005. 943 [RFC5305] Li, T. and H. Smit, "IS-IS Extensions for Traffic 944 Engineering", RFC 5305, October 2008. 946 [RFC5714] Shand, M. and S. Bryant, "IP Fast Reroute Framework", 947 RFC 5714, January 2010. 949 [RFC5715] Shand, M. and S. Bryant, "A Framework for Loop-Free 950 Convergence", RFC 5715, January 2010. 952 [RFC6571] Filsfils, C., Francois, P., Shand, M., Decraene, B., 953 Uttaro, J., Leymann, N., and M. Horneffer, "Loop-Free 954 Alternate (LFA) Applicability in Service Provider (SP) 955 Networks", RFC 6571, June 2012. 957 Authors' Addresses 959 Stephane Litkowski 960 Orange 962 Email: stephane.litkowski@orange.com 964 Bruno Decraene 965 Orange 967 Email: bruno.decraene@orange.com 969 Clarence Filsfils 970 Cisco Systems 972 Email: cfilsfil@cisco.com 974 Kamran Raza 975 Cisco Systems 977 Email: skraza@cisco.com 979 Martin Horneffer 980 Deutsche Telekom 982 Email: Martin.Horneffer@telekom.de 984 Pushpasis Sarkar 985 Juniper Networks 987 Email: psarkar@juniper.net