idnits 2.17.1 draft-ietf-grow-bgp-gshut-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, you can and should remove the disclaimer. Otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (January 28, 2014) is 3740 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'RFC4360' is defined on line 417, but no explicit reference was found in the text == Unused Reference: 'BGPWKC' is defined on line 425, but no explicit reference was found in the text == Outdated reference: A later version (-15) exists of draft-ietf-idr-add-paths-09 Summary: 2 errors (**), 0 flaws (~~), 5 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Pierre Francois 3 Internet-Draft Institute IMDEA Networks 4 Intended status: Informational Bruno Decraene 5 Expires: August 1, 2014 Orange 6 Cristel Pelsser 7 Internet Initiative Japan 8 Keyur Patel 9 Clarence Filsfils 10 Cisco Systems 11 January 28, 2014 13 Graceful BGP session shutdown 14 draft-ietf-grow-bgp-gshut-05 16 Abstract 18 This draft describes operational procedures aimed at reducing the 19 amount of traffic lost during planned maintenances of routers or 20 links, involving the shutdown of BGP peering sessions. 22 Status of this Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at http://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on August 1, 2014. 39 Copyright Notice 41 Copyright (c) 2014 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents 46 (http://trustee.ietf.org/license-info) in effect on the date of 47 publication of this document. Please review these documents 48 carefully, as they describe your rights and restrictions with respect 49 to this document. Code Components extracted from this document must 50 include Simplified BSD License text as described in Section 4.e of 51 the Trust Legal Provisions and are provided without warranty as 52 described in the Simplified BSD License. 54 This document may contain material from IETF Documents or IETF 55 Contributions published or made publicly available before November 56 10, 2008. The person(s) controlling the copyright in some of this 57 material may not have granted the IETF Trust the right to allow 58 modifications of such material outside the IETF Standards Process. 59 Without obtaining an adequate license from the person(s) controlling 60 the copyright in such materials, this document may not be modified 61 outside the IETF Standards Process, and derivative works of it may 62 not be created outside the IETF Standards Process, except to format 63 it for publication as an RFC or to translate it into languages other 64 than English. 66 Table of Contents 68 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 69 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 70 3. Packet loss upon manual eBGP session shutdown . . . . . . . . 5 71 4. Practices to avoid packet losses . . . . . . . . . . . . . . . 5 72 4.1. Improving availability of alternate paths . . . . . . . . 5 73 4.2. Make before break convergence: g-shut . . . . . . . . . . 6 74 4.2.1. eBGP g-shut . . . . . . . . . . . . . . . . . . . . . 6 75 4.2.2. iBGP g-shut . . . . . . . . . . . . . . . . . . . . . 7 76 4.2.3. Router g-shut . . . . . . . . . . . . . . . . . . . . 7 77 5. Forwarding modes and transient forwarding loops during 78 convergence . . . . . . . . . . . . . . . . . . . . . . . . . 8 79 6. Link Up cases . . . . . . . . . . . . . . . . . . . . . . . . 8 80 6.1. Unreachability local to the ASBR . . . . . . . . . . . . . 8 81 6.2. iBGP convergence . . . . . . . . . . . . . . . . . . . . . 9 82 7. IANA assigned g-shut BGP community . . . . . . . . . . . . . . 9 83 8. Security Considerations . . . . . . . . . . . . . . . . . . . 10 84 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 10 85 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 10 86 Appendix A. Alternative techniques with limited applicability . . 11 87 A.1. Multi Exit Discriminator tweaking . . . . . . . . . . . . 11 88 A.2. IGP distance Poisoning . . . . . . . . . . . . . . . . . . 11 89 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 11 91 1. Introduction 93 Routing changes in BGP can be caused by planned, maintenance 94 operations. This document discusses operational procedures to be 95 applied in order to reduce or eliminate losses of packets during the 96 maintenance. These losses come from the transient lack of 97 reachability during the BGP convergence following the shutdown of an 98 eBGP peering session between two Autonomous System Border Routers 99 (ASBR). 101 This document presents procedures for the cases where the forwarding 102 plane is impacted by the maintenance, hence when the use of Graceful 103 Restart does not apply. 105 The procedures described in this document can be applied to reduce or 106 avoid packet loss for outbound and inbound traffic flows initially 107 forwarded along the peering link to be shut down. These procedures 108 trigger, in both involved ASes, rerouting to the alternate path, 109 while allowing routers to keep using old paths until alternate ones 110 are learned, installed in the RIB and in the FIB. This ensures that 111 routers always have a valid route available during the convergence 112 process. 114 The goal of the document is to meet the requirements described in 115 [REQS] at best, without changing the BGP protocol. 117 Still, it explains why reserving a community value for the purpose of 118 BGP session graceful shutdown would reduce the management overhead 119 bound with the solution. It would also allow vendors to provide an 120 automatic graceful shutdown mechanism that does not require any 121 router reconfiguration at maintenance time. 123 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 124 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 125 document are to be interpreted as described in RFC 2119 [RFC2119]. 127 2. Terminology 129 g-shut initiator: a router on which the session shutdown is performed 130 for the maintenance. 132 g-shut neighbor: a router that peers with the g-shut initiator via 133 (one of) the session(s) to be shut down. 135 Initiator AS: the Autonomous System of the g-shut initiator. 137 Neighbor AS: the Autonomous System of the g-shut neighbor. 139 Loss of Connectivity (LoC: the state when a router has no path 140 towards an affected prefix. 142 3. Packet loss upon manual eBGP session shutdown 144 Packets can be lost during a manual shutdown of an eBGP session for 145 two reasons. 147 First, routers involved in the convergence process can transiently 148 lack of paths towards an affected prefix, and drop traffic destined 149 to this prefix. This is because alternate paths can be hidden by 150 nodes of an AS. This happens when the paths are not selected as best 151 by the ASBR that receive them on an eBGP session, or by Route 152 Reflectors that do not propagate them further in the iBGP topology 153 because they do not select them as best. 155 Second, within the AS, the FIB of routers can be transiently 156 inconsistent during the BGP convergence and packets towards affected 157 prefixes can loop and be dropped. Note that these loops only happen 158 when ASBR-to-ASBR encapsulation is not used within the AS. 160 This document only addresses the first reason. 162 4. Practices to avoid packet losses 164 This section describes means for an ISP to reduce the transient loss 165 of packets upon a manual shutdown of a BGP session. 167 4.1. Improving availability of alternate paths 169 All solutions that increase the availability of alternate BGP paths 170 at routers performing packet lookups in BGP tables such as 171 [BestExternal] and [AddPath] help in reducing the LoC bound with 172 manual shutdown of eBGP sessions. 174 One of such solutions increasing diversity in such a way that, at any 175 single step of the convergence process following the eBGP session 176 shutdown, a BGP router does not receive a message withdrawing the 177 only path it currently knows for a given NLRI, allows for a 178 simplified g-shut procedure. 180 Note that the LoC for the inbound traffic of the maintained router, 181 induced by a lack of alternate path propagation within the iBGP 182 topology of a neighboring AS is not under the control of the operator 183 performing the maintenance. The part of the procedure aimed at 184 avoiding LoC for incoming paths can thus be applied even if no LoC 185 are expected for the outgoing paths. 187 4.2. Make before break convergence: g-shut 189 This section describes configurations and actions to be performed for 190 the graceful shutdown of eBGP peering links. 192 The goal of this procedure is to let the paths being shutdown 193 visible, but with a lower LOCAL_PREF value, while alternate paths 194 spread through the iBGP topology. Instead of withdrawing the path, 195 routers of an AS will keep on using it until they become aware of 196 alternate paths. 198 4.2.1. eBGP g-shut 200 4.2.1.1. Pre-configuration 202 On each ASBR supporting the g-shut procedure, an outbound BGP route 203 policy is applied on all iBGP sessions of the ASBR, that: 204 o matches the g-shut community 205 o sets the LOCAL_PREF attribute of the paths tagged with the 206 g-shut community to a low value 207 o removes the g-shut community from the paths. 208 o optionally, adds an AS specific g-shut community on these paths 209 to indicate that these are to be withdrawn soon. If some 210 ingress ASBRs reset the LOCAL_PREF attribute, this AS specific 211 g-shut community will be used to override other LOCAL_PREF 212 preference changes. 214 Note that in the case where an AS is aggregating multiple routes 215 under a covering prefix, it is recommended to filter out the g-shut 216 community from the resulting aggregate BGP route. By doing so, the 217 setting of the g-shut community on one of the aggregated routes will 218 not let the entire aggregate inherit the community. Not doing so 219 would let the entire aggregate undergo the g-shut behavior. 221 4.2.1.2. Operations at maintenance time 223 On the g-shut initiator, upon maintenance time, it is required to: 224 o apply an outbound BGP route policy on the maintained eBGP session 225 to tag the paths propagated over the session with the g-shut 226 community. This will trigger the BGP implementation to re- 227 advertise all active routes previously advertised, and tag them 228 with the g-shut community. 229 o apply an inbound BGP route policy on the maintained eBGP session 230 to tag the paths received over the session with the g-shut 231 community. 233 o wait for convergence to happen. 234 o perform a BGP session shutdown. 236 4.2.1.3. BGP implementation support for G-Shut 238 A BGP router implementation MAY provide features aimed at automating 239 the application of the graceful shutdown procedures described above. 241 Upon a session shutdown specified as graceful by the operator, a BGP 242 implementation supporting a g-shut feature SHOULD: 244 1. On the eBGP side, update all the paths propagated over the 245 corresponding eBGP session, tagging the GSHUT community to them. 246 Any subsequent update sent to the session being gracefully shut 247 down would be tagged with the GSHUT community. 248 2. On the iBGP side, lower the LOCAL_PREF value of the paths 249 received over the eBGP session being shut down, upon their 250 propagation over iBGP sessions. Optionally, also tag these 251 paths with an AS specific g-shut community. Note that 252 alternatively, the LOCAL_PREF of the paths received over the 253 eBGP session can be lowered on the g-shut initiator itself, 254 instead of only when propagating over its iBGP sessions. 255 3. Optionally shut down the session after a configured time. 256 4. Prevent the GSHUT community from being inherited by a path that 257 would aggregate some paths tagged with the GSHUT community. 258 This behavior avoids the GSHUT procedure to be applied to the 259 aggregate upon the graceful shutdown of one of its covered 260 prefixes. 262 A BGP implementation supporting a g-shut feature SHOULD also 263 automatically install the BGP policies that are supposed to be 264 configured, as decribed in Section 4.2.1.1 for sessions over which 265 g-shut is to be supported. 267 4.2.2. iBGP g-shut 269 If the iBGP topology is viable after the maintenance of the session, 270 i.e, if all BGP speakers of the AS have an iBGP signaling path for 271 all prefixes advertised on this g-shut iBGP session, then the 272 shutdown of an iBGP session does not lead to transient 273 unreachability. 275 4.2.3. Router g-shut 277 In the case of a shutdown of a router, a reconfiguration of the 278 outbound BGP route policies of the g-shut initiator SHOULD be 279 performed to set a low LOCAL_PREF value for the paths originated by 280 the g-shut initiator (e.g, BGP aggregates redistributed from other 281 protocols, including static routes). 283 This behavior is equivalent to the recommended behavior for paths 284 "redistributed" from eBGP sessions to iBGP sessions in the case of 285 the shutdown of an ASBR. 287 5. Forwarding modes and transient forwarding loops during convergence 289 The g-shut procedure or the solutions improving the availability of 290 alternate paths, do not change the fact that BGP convergence and the 291 subsequent FIB updates are run independently on each router of the 292 ASes. If the AS applying the solution does not rely on encapsulation 293 to forward packets from the Ingress Border Router to the Egress 294 Border Router, then transient forwarding loops and consequent packet 295 losses can occur during the convergence process. If zero LoC is 296 required, encapsulation is required between ASBRs of the AS. 298 6. Link Up cases 300 We identify two potential causes for transient packet losses upon an 301 eBGP link up event. The first one is local to the g-no-shut 302 initiator, the second one is due to the BGP convergence following the 303 injection of new best paths within the iBGP topology. 305 6.1. Unreachability local to the ASBR 307 An ASBR that selects as best a path received over a newly brought up 308 eBGP session may transiently drop traffic. This can typically happen 309 when the nexthop attribute differs from the IP address of the eBGP 310 peer, and the receiving ASBR has not yet resolved the MAC address 311 associated with the IP address of that "third party" nexthop. 313 A BGP speaker implementation could avoid such losses by ensuring that 314 "third party" nexthops are resolved before installing paths using 315 these in the RIB. 317 If the link up event corresponds to an eBGP session that is being 318 manually brought up, over an already up multi-access link, then the 319 operator can ping third party nexthops that are expected to be used 320 before actually bringing the session up, or ping directed broadcast 321 the subnet IP address of the link. By proceeding like this, the MAC 322 addresses associated with these third party nexthops will be resolved 323 by the g-no-shut initiator. 325 6.2. iBGP convergence 327 Corner cases leading to LoC can occur during an eBGP link up event. 329 A typical example for such transient unreachability for a given 330 prefix is the following: 332 Let's consider 3 route reflectors RR1, RR2, RR3. There is a full 333 mesh of iBGP session between them. 335 1. RR1 is initially advertising the current best path to the 336 members of its iBGP RR full-mesh. It propagated that path 337 within its RR full-mesh. RR2 knows only that path towards the 338 prefix. 339 2. RR3 receives a new best path originated by the "g-no-shut" 340 initiator, being one of its RR clients. RR3 selects it as best, 341 and propagates an UPDATE within its RR full-mesh, i.e., to RR1 342 and RR2. 343 3. RR1 receives that path, reruns its decision process, and 344 picks this new path as best. As a result, RR1 withdraws its 345 previously announced best-path on the iBGP sessions of its RR 346 full-mesh. 347 4. If, for any reason, RR3 processes the withdraw generated in 348 step 3, before processing the update generated in step 2, RR3 349 transiently suffers from unreachability for the affected prefix. 351 The use of [BestExternal] among the RR of the iBGP full-mesh can 352 solve these corner cases by ensuring that within an AS, the 353 advertisement of a new route is not translated into the withdraw of a 354 former route. 356 Indeed, "best-external" ensures that an ASBR does not withdraw a 357 previously advertised (eBGP) path when it receives an additional, 358 preferred path over an iBGP session. Also, "best-intra-cluster" 359 ensures that a RR does not withdraw a previously advertised (iBGP) 360 path to its non clients (e.g. other RRs in a mesh of RR) when it 361 receives a new, preferred path over an iBGP session. 363 7. IANA assigned g-shut BGP community 365 Applying the g-shut procedure is rendered much easier with the use of 366 a single g-shut community value which could be used on all eBGP 367 sessions, for both inbound and outbound signaling. The community 368 value 0xFFFF0000 has been assigned by IANA for this purpose. 370 For Internet routes, a non transitive extended community will be 371 reserved from the pool defined in [EXT_POOL]. Using such a community 372 type allows for not leaking graceful signaling out of the AS 373 boundaries, without the need to explicitly configure filters to strip 374 the community off upon path propagation. 376 8. Security Considerations 378 By providing the g-shut service to a neighboring AS, an ISP provides 379 means to this neighbor to lower the LOCAL_PREF value assigned to the 380 paths received from this neighbor. 382 The neighbor could abuse the technique and do inbound traffic 383 engineering by declaring some prefixes as undergoing a maintenance so 384 as to switch traffic to another peering link. 386 If this behavior is not tolerated by the ISP, it SHOULD monitor the 387 use of the g-shut community by this neighbor. 389 ASes using the regular (transitive) g-shut community SHOULD remove 390 the community from neighboring ASes that do not support the g-shut 391 procedure. Doing so prevents malignant remote ASes from using the 392 community through intermediate ASes that do not support the feature, 393 in order to perform inbound traffic engineering. ASes using the non- 394 transitive extended community do not need to do this as the community 395 is non transitive and hence cannot be used by remote ASes. 397 9. Acknowledgments 399 The authors wish to thank Olivier Bonaventure and Pradosh Mohapatra 400 for their useful comments on this work. 402 10. References 404 [AddPath] D. Walton, E. Chen, A. Retana, and J. Scudder, 405 "Advertisement of Multiple Paths in BGP", 406 draft-ietf-idr-add-paths-09.txt (work in progress). 408 [BestExternal] 409 Marques, P., Fernando, R., Chen, E., Mohapatra, P., and H. 410 Gredler, "Advertisement of the best-external route to 411 IBGP", draft-ietf-idr-best-external-05.txt. 413 [REQS] Decraene, B., Francois, P., Pelsser, C., Ahmad, Z., 414 Armengol, A., and T. Takeda, "Requirements for the 415 graceful shutdown of BGP sessions", RFC 6198. 417 [RFC4360] Sangli, S., Tappan, D., and Y. Rekhter, "BGP Extended 418 Communities Attribute", RFC 4360, February 2006. 420 [EXT_POOL] 421 Decraene, B. and P. Francois, "Assigned BGP extended 422 communities", 423 draft-ietf-idr-reserved-extended-communities-06. 425 [BGPWKC] "http://www.iana.org/assignments/ 426 bgp-well-known-communities". 428 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 429 Requirement Levels", BCP 14, RFC 2119, March 1997. 431 Appendix A. Alternative techniques with limited applicability 433 A few alternative techniques have been considered to provide g-shut 434 capabilities but have been rejected due to their limited 435 applicability. This section describe them for possible reference. 437 A.1. Multi Exit Discriminator tweaking 439 The MED attribute of the paths to be avoided can be increased so as 440 to force the routers in the neighboring AS to select other paths. 442 The solution only works if the alternate paths are as good as the 443 initial ones with respect to the Local-Pref value and the AS Path 444 Length value. In the other cases, increasing the MED value will not 445 have an impact on the decision process of the routers in the 446 neighboring AS. 448 A.2. IGP distance Poisoning 450 The distance to the BGP nexthop corresponding to the maintained 451 session can be increased in the IGP so that the old paths will be 452 less preferred during the application of the IGP distance tie-break 453 rule. However, this solution only works for the paths whose 454 alternates are as good as the old paths with respect to their Local- 455 Pref value, their AS Path length, and their MED value. 457 Also, this poisoning cannot be applied when nexthop self is used as 458 there is no nexthop specific to the maintained session to poison in 459 the IGP. 461 Authors' Addresses 463 Pierre Francois 464 Institute IMDEA Networks 465 Avda. del Mar Mediterraneo, 22 466 Leganese 28918 467 ES 469 Email: pierre.francois@imdea.org 471 Bruno Decraene 472 Orange 473 38-40 rue du General Leclerc 474 92794 Issy Moulineaux cedex 9 475 FR 477 Email: bruno.decraene@orange.com 479 Cristel Pelsser 480 Internet Initiative Japan 481 Jinbocho Mitsui Bldg. 482 1-105 Kanda Jinbo-cho 483 Tokyo 101-0051 484 JP 486 Email: cristel@iij.ad.jp 488 Keyur Patel 489 Cisco Systems 490 170 West Tasman Dr 491 San Jose, CA 95134 492 US 494 Email: keyupate@cisco.com 496 Clarence Filsfils 497 Cisco Systems 498 De kleetlaan 6a 499 Diegem 1831 500 BE 502 Email: cfilsfil@cisco.com