idnits 2.17.1 draft-ietf-idr-aigp-11.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 (November 21, 2013) is 3799 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-19) exists of draft-ietf-idr-error-handling-04 Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Pradosh Mohapatra 3 Internet Draft Cumulus Networks 4 Intended Status: Proposed Standard 5 Expires: May 21, 2014 Rex Fernando 6 Eric C. Rosen 7 Cisco Systems, Inc. 9 James Uttaro 10 ATT 12 November 21, 2013 14 The Accumulated IGP Metric Attribute for BGP 16 draft-ietf-idr-aigp-11.txt 18 Abstract 20 Routing protocols that have been designed to run within a single 21 administrative domain ("IGPs") generally do so by assigning a metric 22 to each link, and then choosing as the installed path between two 23 nodes the path for which the total distance (sum of the metric of 24 each link along the path) is minimized. BGP, designed to provide 25 routing over a large number of independent administrative domains 26 ("autonomous systems"), does not make its path selection decisions 27 through the use of a metric. It is generally recognized that any 28 attempt to do so would incur significant scalability problems, as 29 well as inter-administration coordination problems. However, there 30 are deployments in which a single administration runs several 31 contiguous BGP networks. In such cases, it can be desirable, within 32 that single administrative domain, for BGP to select paths based on a 33 metric, just as an IGP would do. The purpose of this document is to 34 provide a specification for doing so. 36 Status of this Memo 38 This Internet-Draft is submitted to IETF in full conformance with the 39 provisions of BCP 78 and BCP 79. 41 Internet-Drafts are working documents of the Internet Engineering 42 Task Force (IETF), its areas, and its working groups. Note that 43 other groups may also distribute working documents as Internet- 44 Drafts. 46 Internet-Drafts are draft documents valid for a maximum of six months 47 and may be updated, replaced, or obsoleted by other documents at any 48 time. It is inappropriate to use Internet-Drafts as reference 49 material or to cite them other than as "work in progress." 51 The list of current Internet-Drafts can be accessed at 52 http://www.ietf.org/ietf/1id-abstracts.txt. 54 The list of Internet-Draft Shadow Directories can be accessed at 55 http://www.ietf.org/shadow.html. 57 Copyright and License Notice 59 Copyright (c) 2013 IETF Trust and the persons identified as the 60 document authors. All rights reserved. 62 This document is subject to BCP 78 and the IETF Trust's Legal 63 Provisions Relating to IETF Documents 64 (http://trustee.ietf.org/license-info) in effect on the date of 65 publication of this document. Please review these documents 66 carefully, as they describe your rights and restrictions with respect 67 to this document. Code Components extracted from this document must 68 include Simplified BSD License text as described in Section 4.e of 69 the Trust Legal Provisions and are provided without warranty as 70 described in the Simplified BSD License. 72 Table of Contents 74 1 Specification of requirements ......................... 3 75 2 Introduction .......................................... 3 76 3 AIGP Attribute ........................................ 5 77 3.1 Applicability Restrictions and Cautions ............... 6 78 3.2 Handling a Malformed AIGP Attribute ................... 6 79 3.3 Restrictions on Sending/Receiving ..................... 7 80 3.4 Creating and Modifying the AIGP Attribute ............. 7 81 3.4.1 Originating the AIGP Attribute ........................ 7 82 3.4.2 Modifications by the Originator ....................... 9 83 3.4.3 Modifications by a Non-Originator ..................... 9 84 4 Decision Process ...................................... 11 85 4.1 When a Route has an AIGP Attribute .................... 11 86 4.2 When the Route to the Next Hop has an AIGP attribute .. 12 87 5 Deployment Considerations ............................. 12 88 6 IANA Considerations ................................... 13 89 7 Security Considerations ............................... 13 90 8 Acknowledgments ....................................... 13 91 9 Authors' Addresses .................................... 13 92 10 Normative References .................................. 14 93 11 Informative References ................................ 14 95 1. Specification of requirements 97 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 98 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 99 document are to be interpreted as described in [RFC2119]. 101 2. Introduction 103 There are many routing protocols that have been designed to run 104 within a single administrative domain. These are known collectively 105 as "Interior Gateway Protocols" (IGPs). Typically, each link is 106 assigned a particular "metric" value. The path between two nodes can 107 then be assigned a "distance", which is the sum of the metrics of all 108 the links that belong to that path. An IGP selects the "shortest" 109 (minimal distance) path between any two nodes, perhaps subject to the 110 constraint that if the IGP provides multiple "areas", it may prefer 111 the shortest path within an area to a path that traverses more than 112 one area. Typically the administration of the network has some 113 routing policy which can be approximated by selecting shortest paths 114 in this way. 116 BGP, as distinguished from the IGPs, was designed to run over an 117 arbitrarily large number of administrative domains ("autonomous 118 systems", or "ASes") with limited coordination among the various 119 administrations. BGP does not make its path selection decisions 120 based on a metric; there is no such thing as an "inter-AS metric". 121 There are two fundamental reasons for this: 123 - The distance between two nodes in a common administrative domain 124 may change at any time due to events occurring in that domain. 125 These changes are not propagated around the Internet unless they 126 actually cause the border routers of the domain to select routes 127 with different BGP attributes for some set of address prefixes. 128 This accords with a fundamental principle of scaling, viz., that 129 changes with only local significance must not have global 130 effects. If local changes in distance were always propagated 131 around the Internet, this principle would be violated. 133 - A basic principle of inter-domain routing is that the different 134 administrative domains may have their own policies, which do not 135 have to be revealed to other domains, and which certainly do not 136 have to be agreed to by other domains. Yet the use of inter-AS 137 metric in the Internet would have exactly these effects. 139 There are, however, deployments in which a single administration runs 140 a network which has been sub-divided into multiple, contiguous ASes, 141 each running BGP. There are several reasons why a single 142 administrative domain may be broken into several ASes (which, in this 143 case, are not really "autonomous".) It may be that the existing IGPs 144 do not scale well in the particular environment; it may be that a 145 more generalized topology is desired than could be obtained by use of 146 a single IGP domain; it may be that a more finely grained routing 147 policy is desired than can be supported by an IGP. In such 148 deployments, it can be useful to allow BGP to make its routing 149 decisions based on the IGP metric, so that BGP chooses the "shortest" 150 path between two nodes, even if the nodes are in two different ASes 151 within that same administrative domain. We will refer to the set of 152 ASes in a common administrative domain as an "AIGP Administrative 153 Domain". 155 There are in fact some implementations that already do something like 156 this, using BGP's MULTI_EXIT_DISC (MED) attribute to carry a value 157 based on IGP metrics. However, that doesn't really provide IGP-like 158 "shortest path" routing, as the BGP decision process gives priority 159 to other factors, such as the AS_PATH length. Also, the standard 160 procedures for use of the MED do not ensure that the IGP metric is 161 properly accumulated so that it covers all the links along the path. 163 In this document, we define a new optional, non-transitive BGP 164 attribute, called the "Accumulated IGP Metric Attribute", or "AIGP 165 attribute", and specify the procedures for using it. 167 The specified procedures prevent the AIGP attribute from "leaking 168 out" past an AIGP administrative domain boundary into the Internet. 170 The specified procedures also ensure that the value in the AIGP 171 attribute has been accumulated all along the path from the 172 destination, i.e., that the AIGP attribute does not appear when there 173 are "gaps" along the path where the IGP metric is unknown. 175 3. AIGP Attribute 177 The AIGP Attribute is an optional non-transitive BGP Path Attribute. 178 The attribute type code for the AIGP Attribute is 26. 180 The value field of the AIGP Attribute is defined here to be a set of 181 elements encoded as "Type/Length/Value" (i.e., a set of "TLVs"). 182 Each such TLV is encoded as shown in Figure 1. 184 0 1 2 3 185 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 186 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 187 | Type | Length | | 188 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 189 ~ ~ 190 | Value | 191 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+.......................... 193 AIGP TLV 194 Figure 1 196 - Type: A single octet encoding the TLV Type. Only type 1, "AIGP 197 TLV", is defined in this document. Use of other TLV types is 198 outside the scope of this document. 200 - Length: Two octets encoding the length in octets of the TLV, 201 including the type and length fields. The length is encoded as an 202 unsigned binary integer. (Note that the minimum length is 3, 203 indicating that no value field is present.) 205 - A value field containing zero or more octets. 207 This document defines only a single such TLV, the "AIGP TLV". The 208 AIGP TLV is encoded as follows: 210 - Type: 1 212 - Length: 11 214 - Accumulated IGP Metric. 216 The value field of the AIGP TLV is always 8 bytes long. IGP 217 metrics are frequently expressed as 4-octet values, and this 218 ensures that the AIGP attribute can be used to hold the sum of an 219 arbitrary number of 4-octet values. 221 When an AIGP attribute is created, it SHOULD contain no more than one 222 AIGP TLV. However, if it contains more than one AIGP TLV, only the 223 first one is is used as described in Section 3.4 and Section 4. In 224 the remainder of this document, we will use the term "value of the 225 AIGP TLV" to mean the value of the first AIGP TLV in the AIGP 226 attribute. Any other AIGP TLVs in the AIGP attribute MUST be passed 227 along unchanged if the AIGP attribute is passed along. 229 3.1. Applicability Restrictions and Cautions 231 This document only considers the use of the AIGP attribute in 232 networks where each router uses tunneling of some sort to deliver a 233 packet to its BGP next hop. Use of the AIGP attribute in networks 234 that do not use tunneling is outside the scope of this document. 236 If a Route Reflector supports the AIGP attribute, but some of its 237 clients do not, then the routing choices that result may not all 238 reflect the intended routing policy. 240 3.2. Handling a Malformed AIGP Attribute 242 When receiving a BGP Update message containing a malformed AIGP 243 attribute, the attribute MUST be treated exactly as if it were an 244 unrecognized non-transitive attribute. That is, "it MUST be quietly 245 ignored and not passed along to other BGP peers". This is equivalent 246 to the "attribute discard" action specified in [BGP-ERROR]. 248 Note that an AIGP attribute MUST NOT be considered to be malformed 249 because it contains more than one TLV of a given type, or because it 250 contains TLVs of unknown types. 252 If a BGP Path Attribute is received that has the AIGP attribute 253 codepoint, but also has the transitive bit set, the attribute MUST be 254 considered to be a malformed AIGP attribute, and MUST be discarded as 255 specified in this section. 257 If an AIGP attribute is received, and its first AIGP TLV contains the 258 maximum value 0xffffffffffffffff, the attribute SHOULD be considered 259 to be malformed, and discarded as specified in this section. (Since 260 the TLV value cannot be increased any further, it is not useful for 261 metric-based path selection.) 263 3.3. Restrictions on Sending/Receiving 265 An implementation that supports the AIGP attribute MUST support a 266 per-session configuration item, AIGP_SESSION, that indicates whether 267 the attribute is enabled or disabled for use on that session. 269 - The default value of AIGP_SESSION, for EBGP sessions, MUST be 270 "disabled". 272 - The default value of AIGP_SESSION, for IBGP and confederation- 273 EBGP sessions, SHOULD be "enabled." 275 The AIGP attribute MUST NOT be sent on any BGP session for which 276 AIGP_SESSION is disabled. 278 If an AIGP attribute is received on a BGP session for which 279 AIGP_SESSION is disabled, the attribute MUST be treated exactly as if 280 it were an unrecognized non-transitive attribute. That is, "it MUST 281 be quietly ignored and not passed along to other BGP peers" (see 282 [BGP], section 5). 284 3.4. Creating and Modifying the AIGP Attribute 286 3.4.1. Originating the AIGP Attribute 288 An implementation that supports the AIGP attribute MUST support a 289 configuration item, AIGP_ORIGINATE, that enables or disables its 290 creation and attachment to routes. The default value of 291 AIGP_ORIGINATE MUST be "disabled". 293 A BGP speaker MUST NOT add the AIGP attribute to any route whose path 294 leads outside the "AIGP administrative domain" to which the BGP 295 speaker belongs. It may be added only to routes that satisfy one of 296 the following conditions: 298 - The route is a static route that is being redistributed into BGP 300 - The route is an IGP route that is being redistributed into BGP 302 - The route is an IBGP-learned route whose AS_PATH attribute is 303 empty. 305 - The route is an EBGP-learned route whose AS_PATH contains only 306 ASes that are in the same AIGP Administrative Domain as the BGP 307 speaker. 309 A BGP speaker R MUST NOT add the AIGP attribute to any route for 310 which R does not set itself as the next hop. 312 It SHOULD be possible to set AIGP_ORIGINATE to "enabled for the 313 routes of a particular IGP that are redistributed into BGP" (where "a 314 particular IGP" might be "OSPF" or "ISIS"). Other policies 315 determining when and whether to originate an AIGP attribute are also 316 possible, depending on the needs of a particular deployment scenario. 318 When originating an AIGP attribute for a BGP route to address prefix 319 P, the value of the AIGP TLV is set according to policy. There are a 320 number of useful policies, some of which are in the following list: 322 - When a BGP speaker R is redistributing into BGP an IGP route to 323 address prefix P, the IGP will have computed a "distance" from R 324 to P. This distance MAY be assigned as the value of the AIGP 325 TLV. 327 - A BGP speaker R may be redistributing into BGP a static route to 328 address prefix P, for which a "distance" from R to P has been 329 configured. This distance MAY be assigned as the value of the 330 AIGP TLV. 332 - A BGP speaker R may have received and installed a BGP-learned 333 route to prefix P, with next hop N. Or it may be redistributing 334 a static route to P, with next hop N. Then: 336 * If R has an IGP route to N, the IGP-computed distance from R 337 to N MAY be used as the value of the AIGP TLV of the route to 338 P. 340 * If R has a BGP route to N, and an AIGP TLV attribute value 341 has been computed for that route (see section 3.4.3), that 342 value MAY be used as the AIGP TLV value of the route to P. 344 3.4.2. Modifications by the Originator 346 If BGP speaker R is the originator of the AIGP attribute of prefix P, 347 and at some point the "distance" from R to P changes, R SHOULD issue 348 a new BGP update containing the new value of the AIGP TLV of the AIGP 349 attribute. (Here we use the term "distance" to refer to whatever 350 value the originator assigns to the AIGP TLV, however it is computed; 351 see section 3.4.1.) However, if the difference between the new 352 distance and the distance advertised in the AIGP TLV is less than a 353 configurable threshold, the update MAY be suppressed. 355 3.4.3. Modifications by a Non-Originator 357 Suppose a BGP speaker R1 receives a route with an AIGP attribute 358 whose value is A, and a Next Hop whose value is R2. Suppose also 359 that R1 is about to redistribute that route on a BGP session that is 360 enabled for sending/receiving the attribute. 362 If R1 does not change the Next Hop of the route, then R1 MUST NOT 363 change the AIGP attribute value of the route. 365 In all the computations discussed in this section, the AIGP value 366 MUST be capped at its maximum unsigned value 0xffffffffffffffff. 367 Increasing the AIGP value MUST NOT cause the value to wrap around. 369 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 370 route to R2 is an IGP-learned route, or a static route that does not 371 require recursive next hop resolution, then R1 MUST increase the 372 value of the AIGP TLV by adding to A the distance from R1 to R2. 373 This distance is either the IGP-computed distance from R1 to R2, or 374 some value determined by policy. However, A MUST be increased by a 375 non-zero amount. 377 Note that if R1 and R2 above are EBGP neighbors, and there is a 378 direct link between them on which no IGP is running, then when R1 379 changes the next hop of a route from R2 to R1, the AIGP TLV value 380 MUST be increased by a non-zero amount. The amount of the increase 381 SHOULD be such that it is properly comparable to the IGP metrics. 382 E.g., if the IGP metric is a function of latency, then the amount of 383 the increase should be a function of the latency from R1 to R2. 385 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 386 route to R2 is a BGP-learned route, or a static route that requires 387 recursive next hop resolution, then the AIGP TLV value needs to be 388 increased in several steps, according to the following procedure. 389 (Note that this procedure is ONLY used when recursive next hop 390 resolution is needed.) 391 1. Let Xattr be the new AIGP TLV value. 393 2. Initialize Xattr to A. 395 3. Set the XNH to R2. 397 4. Find the route to XNH. 399 5. If the route to XNH does not require recursive next hop 400 resolution, get the distance D from R1 to XNH. (Note that this 401 condition cannot be satisfied the first time through this 402 procedure.) If D is above a configurable threshold, set the 403 AIGP TLV value to Xattr+D. If D is below a configurable 404 threshold, set the AIGP TLV value to Xattr. In either case, 405 exit this procedure. 407 6. If the route to XNH is a BGP-learned route, and the route does 408 NOT have an AIGP attribute, then exit this procedure and do not 409 pass on any AIGP attribute. If the route has an AIGP attribute 410 without an AIGP TLV, then the AIGP attribute MAY be passed 411 along unchanged. 413 7. If the route to XNH is a BGP-learned route, and the route has 414 an AIGP TLV value of Y, then set Xattr=Xattr+Y, and set XNH to 415 the next hop of this route. (The intention here is that Y is 416 the AIGP TLV value of the route as it was received by R1, 417 without having been modified by R1.) 419 8. Go to step 4. 421 The AIGP TLV value of a given route depends on (a) the AIGP TLV 422 values of all the next hops that are recursively resolved during this 423 procedure, and (b) the IGP distance to any next hop that is not 424 recursively resolved. Any change due to (a) in any of these values 425 MUST trigger a new AIGP computation for that route. Whether a change 426 due to (b) triggers a new AIGP computation depends upon whether the 427 change in IGP distance exceeds a configurable threshold. 429 If the AIGP attribute is carried across several ASes, each with its 430 own IGP domain, it is clear that these procedures are unlikely to 431 give a sensible result if the IGPs are different (e.g., some OSPF and 432 some IS-IS), or if the meaning of the metrics is different in the 433 different IGPs (e.g., if the metric represents bandwidth in some IGP 434 domains but represents latency in others). These procedures also are 435 unlikely to give a sensible result if the metric assigned to inter-AS 436 BGP links (on which no IGP is running) or to static routes is not 437 comparable to the IGP metrics. All such cases are outside the scope 438 of the current document. 440 4. Decision Process 442 Support for the AIGP attribute involves several modifications to the 443 tie breaking procedures of the BGP "phase 2" decision described in 444 [BGP], section 9.1.2.2. These modifications are described below in 445 sections 4.1 and 4.2. 447 In some cases, the BGP decision process may install a route without 448 executing any tie breaking procedures. This may happen, e.g., if 449 only one route to a given prefix has the highest degree of preference 450 (as defined in [BGP] section 9.1.1). In this case, the AIGP 451 attribute is not considered. 453 In other cases, some routes may be eliminated before the tie breaking 454 procedures are invoked, e.g., routes with AS-PATH attributes 455 indicating a loop, or routes with unresolvable next hops. In these 456 cases, the AIGP attributes of the eliminated routes are not 457 considered. 459 4.1. When a Route has an AIGP Attribute 461 Assuming that the BGP decision process invokes the tie breaking 462 procedures, the procedures in this section MUST be executed BEFORE 463 any of the tie breaking procedures described in [BGP] section 9.1.2.2 464 are executed. 466 If any routes have an AIGP attribute containing an AIGP TLV, remove 467 from consideration all routes that do not have an AIGP attribute 468 containing an AIGP TLV. 470 If router R is considering route T, where T has an AIGP attribute 471 with an AIGP TLV, 473 - then R must compute the value A, defined as follows: set A to the 474 sum of (a) T's AIGP TLV value and (b) the IGP distance from R to 475 T's next hop. 477 - remove from consideration all routes that are not tied for the 478 lowest value of A. 480 4.2. When the Route to the Next Hop has an AIGP attribute 482 Suppose that a given router R1 is comparing two BGP-learned routes, 483 such that either: 485 - the two routes have equal AIGP TLV values, or else 487 - neither of the two routes has an AIGP attribute containing an 488 AIGP TLV. 490 The BGP decision process as specified in [BGP] makes use, in its tie 491 breaker procedures, of "interior cost", defined as follows: 493 "interior cost of a route is determined by calculating the metric 494 to the NEXT_HOP for the route using the Routing Table." 496 Suppose route T has a next hop of N. We modify the notion of the 497 "interior cost" from node R1 to node N as follows: 499 - Let R2 be the BGP next hop of the route to N, after all recursive 500 resolution of the next hop is done. Let m be the IGP distance 501 (or in the case of a static route, the configured distance) from 502 R1 to R2. 504 - If the installed route to N has an AIGP attribute with an AIGP 505 TLV, set A to the AIGP value of the route to N, computing the 506 AIGP value of the route according to the procedure of section 507 3.4.3. 509 - If the installed route to N does not have an AIGP value, set A to 510 0. 512 - The "interior cost" of route T is the quantity A+m. 514 5. Deployment Considerations 516 Using the AIGP attribute to achieve a desired routing policy will be 517 more effective if each BGP speaker can use it to choose from among 518 multiple routes. Thus is it highly recommended that the procedures of 519 [BESTEXT] and [CONN-REST] be used in conjunction with the AIGP 520 Attribute. 522 If a Route Reflector does not pass all paths to its clients, then it 523 will tend to pass the paths for which the IGP distance from the Route 524 Reflector itself to the next hop is smallest. This may result in a 525 non-optimal choice by the clients. 527 6. IANA Considerations 529 IANA has assigned the codepoint 26 in the "BGP Path Attributes" 530 registry to the AIGP attribute. 532 IANA shall create a registry for "BGP AIGP Attribute Types". The 533 type field consists of a single octet, with possible values from 0 to 534 255. The allocation policy for this field is to be "Standards Action 535 with Early Allocation". Type 1 should be defined as "AIGP", and 536 should refer to this document. 538 7. Security Considerations 540 The spurious introduction, though error or malfeasance, of an AIGP 541 attribute, could result in the selection of paths other than those 542 desired. 544 Improper configuration on both ends of an EBGP connection could 545 result in an AIGP attribute being passed from one service provider to 546 another. This would likely result in an unsound selection of paths. 548 8. Acknowledgments 550 The authors would like to thank Waqas Alam, Rajiv Asati, Brian 551 Dickson, Clarence Filsfils, Pratima Kini, Thomas Mangin, Robert 552 Raszuk, Yakov Rekhter, Eric Rosenberg, Samir Saad, John Scudder, 553 Shyam Sethuram, and Ilya Varlashkin for their input. 555 9. Authors' Addresses 557 Rex Fernando 558 Cisco Systems, Inc. 559 170 Tasman Drive 560 San Jose, CA 95134 561 Email: rex@cisco.com 563 Pradosh Mohapatra 564 Cumulus Networks 565 Email: pmohapat@cumulusnetworks.com 566 Eric C. Rosen 567 Cisco Systems, Inc. 568 1414 Massachusetts Avenue 569 Boxborough, MA, 01719 570 Email: erosen@cisco.com 572 James Uttaro 573 AT&T 574 200 S. Laurel Avenue 575 Middletown, NJ 07748 576 Email: uttaro@att.com 578 10. Normative References 580 [BGP], "A Border Gateway Protocol 4 (BGP-4)", Y. Rekhter, T. Li, S. 581 Hares, RFC 4271, January 2006. 583 11. Informative References 585 [CONN-REST] "Fast Connectivity Restoration Using BGP Add-Path", P. 586 Mohapatra, R. Fernando, C. Filsfils, R. Raszuk, draft-pmohapat-idr- 587 fast-conn-restore-03.txt, January 2013. 589 [BESTEXT], "Advertisement of the Best External Route in BGP", P. 590 Marques, R. Fernando, E. Chen, P. Mohapatra, H. Gredler, draft-ietf- 591 idr-best-external-05.txt, January 2012. 593 [BGP-ERROR] "Revised Error Handling for BGP UPDATE Messages", J. 594 Scudder, E. Chen, P. Mohapatra, K. Patel, draft-ietf-idr-error- 595 handling-04.txt, June 2013. 597 [RFC2119] "Key words for use in RFCs to Indicate Requirement 598 Levels.", S. Bradner, March 1997.