idnits 2.17.1 draft-ietf-idr-error-handling-06.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 draft header indicates that this document updates RFC4360, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC4760, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC5701, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC1997, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC4456, but the abstract doesn't seem to mention this, which it should. -- The draft header indicates that this document updates RFC4271, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC1997, updated by this document, for RFC5378 checks: 1996-04-10) -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, 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 (February 14, 2014) is 3696 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) No issues found here. Summary: 0 errors (**), 0 flaws (~~), 1 warning (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force E. Chen, Ed. 3 Internet-Draft Cisco Systems, Inc. 4 Updates: 1997, 4271, 4360, 4456, 4760, 5701 (if approved)J. Scudder, Ed. 5 Intended status: Standards Track Juniper Networks 6 Expires: August 18, 2014 P. Mohapatra 7 Cumulus Networks, Inc. 8 K. Patel 9 Cisco Systems, Inc. 10 February 14, 2014 12 Revised Error Handling for BGP UPDATE Messages 13 draft-ietf-idr-error-handling-06 15 Abstract 17 According to the base BGP specification, a BGP speaker that receives 18 an UPDATE message containing a malformed attribute is required to 19 reset the session over which the offending attribute was received. 20 This behavior is undesirable as a session reset would impact not only 21 routes with the offending attribute, but also other valid routes 22 exchanged over the session. This document partially revises the 23 error handling for UPDATE messages, and provides guidelines for the 24 authors of documents defining new attributes. Finally, it revises 25 the error handling procedures for a number of existing attributes. 27 This document updates error handling for RFCs 1997, 4271, 4360, 4456, 28 4760, and 5701. 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at http://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on August 18, 2014. 47 Copyright Notice 49 Copyright (c) 2014 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (http://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 This document may contain material from IETF Documents or IETF 63 Contributions published or made publicly available before November 64 10, 2008. The person(s) controlling the copyright in some of this 65 material may not have granted the IETF Trust the right to allow 66 modifications of such material outside the IETF Standards Process. 67 Without obtaining an adequate license from the person(s) controlling 68 the copyright in such materials, this document may not be modified 69 outside the IETF Standards Process, and derivative works of it may 70 not be created outside the IETF Standards Process, except to format 71 it for publication as an RFC or to translate it into languages other 72 than English. 74 Table of Contents 76 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 77 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 78 2. Revision to Base Specification . . . . . . . . . . . . . . . 4 79 3. Parsing of NLRI Fields . . . . . . . . . . . . . . . . . . . 6 80 3.1. Inconsistency of Attribute Length Fields . . . . . . . . 6 81 3.2. Syntactic Correctness of NLRI Fields . . . . . . . . . . 7 82 4. Operational Considerations . . . . . . . . . . . . . . . . . 7 83 5. Error Handling Procedures for Existing Attributes . . . . . . 8 84 5.1. ORIGIN . . . . . . . . . . . . . . . . . . . . . . . . . 8 85 5.2. AS_PATH . . . . . . . . . . . . . . . . . . . . . . . . . 8 86 5.3. NEXT_HOP . . . . . . . . . . . . . . . . . . . . . . . . 9 87 5.4. MULTI_EXIT_DESC . . . . . . . . . . . . . . . . . . . . . 9 88 5.5. LOCAL_PREF . . . . . . . . . . . . . . . . . . . . . . . 9 89 5.6. ATOMIC_AGGREGATE . . . . . . . . . . . . . . . . . . . . 9 90 5.7. AGGREGATOR . . . . . . . . . . . . . . . . . . . . . . . 9 91 5.8. Community . . . . . . . . . . . . . . . . . . . . . . . . 10 92 5.9. Extended Community . . . . . . . . . . . . . . . . . . . 10 93 5.10. IPv6 Address Specific BGP Extended Community Attribute . 10 94 5.11. ORIGINATOR_ID . . . . . . . . . . . . . . . . . . . . . . 11 95 5.12. CLUSTER_LIST . . . . . . . . . . . . . . . . . . . . . . 11 96 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11 97 7. Security Considerations . . . . . . . . . . . . . . . . . . . 11 98 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 11 99 9. Normative References . . . . . . . . . . . . . . . . . . . . 12 100 Appendix A. Why not discard UPDATE messages? . . . . . . . . . . 12 101 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 13 103 1. Introduction 105 According to the base BGP specification [RFC4271], a BGP speaker that 106 receives an UPDATE message containing a malformed attribute is 107 required to reset the session over which the offending attribute was 108 received. This behavior is undesirable as a session reset would 109 impact not only routes with the offending attribute, but also other 110 valid routes exchanged over the session. In the case of optional 111 transitive attributes, the behavior is especially troublesome and may 112 present a potential security vulnerability. The reason is that such 113 attributes may have been propagated without being checked by 114 intermediate routers that do not recognize the attributes -- in 115 effect the attribute may have been tunneled, and when they do reach a 116 router that recognizes and checks them, the session that is reset may 117 not be associated with the router that is at fault. 119 The goal for revising the error handling for UPDATE messages is to 120 minimize the impact on routing by a malformed UPDATE message, while 121 maintaining protocol correctness to the extent possible. This can be 122 achieved largely by maintaining the established session and keeping 123 the valid routes exchanged, but removing the routes carried in the 124 malformed UPDATE from the routing system. 126 This document partially revises the error handling for UPDATE 127 messages, and provides guidelines for the authors of documents 128 defining new attributes. Finally, it revises the error handling 129 procedures for a number of existing attributes. Specifically, the 130 error handling procedures of, [RFC1997], [RFC4271], [RFC4360], 131 [RFC4456], [RFC4760] and [RFC5701] are revised. 133 1.1. Requirements Language 135 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 136 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 137 document are to be interpreted as described in RFC 2119 [RFC2119]. 139 2. Revision to Base Specification 141 The first paragraph of Section 6.3 of [RFC4271] is revised as 142 follows: 144 Old Text: 146 All errors detected while processing the UPDATE message MUST be 147 indicated by sending the NOTIFICATION message with the Error Code 148 UPDATE Message Error. The error subcode elaborates on the 149 specific nature of the error. 151 New text: 153 An error detected while processing the UPDATE message for which a 154 session reset is specified MUST be indicated by sending the 155 NOTIFICATION message with the Error Code UPDATE Message Error. 156 The error subcode elaborates on the specific nature of the error. 158 The error handling of the following case described in Section 6.3 of 159 [RFC4271] remains unchanged: 161 If the Withdrawn Routes Length or Total Attribute Length is too 162 large (i.e., if Withdrawn Routes Length + Total Attribute Length + 163 23 exceeds the message Length), then the Error Subcode MUST be set 164 to Malformed Attribute List. 166 The error handling of the following case described in Section 6.3 of 167 [RFC4271] is revised 169 If any recognized attribute has Attribute Flags that conflict with 170 the Attribute Type Code, then the Error Subcode MUST be set to 171 Attribute Flags Error. The Data field MUST contain the erroneous 172 attribute (type, length, and value). 174 as follows: 176 If any recognized attribute has Attribute Flags that conflict with 177 the Attribute Type Code, then the attribute MUST be treated as 178 malformed and the treat-as-withdraw approach (see below) used, 179 unless the specification for the attribute mandates different 180 handling for incorrect Attribute Flags. 182 The error handling of all other cases involving path attributes as 183 described in Section 6.3 of [RFC4271] that specify a session reset is 184 revised as follows. 186 When a path attribute (other than the MP_REACH_NLRI attribute 187 [RFC4760] or the MP_UNREACH_NLRI attribute [RFC4760]) in an UPDATE 188 message is determined to be malformed, the UPDATE message containing 189 that attribute MUST be treated as though all contained routes had 190 been withdrawn just as if they had been listed in the WITHDRAWN 191 ROUTES field (or in the MP_UNREACH_NLRI attribute if appropriate) of 192 the UPDATE message, thus causing them to be removed from the Adj-RIB- 193 In according to the procedures of [RFC4271]. In the case of an 194 attribute which has no effect on route selection or installation, the 195 malformed attribute MAY instead be discarded and the UPDATE message 196 continue to be processed. For the sake of brevity, the former 197 approach is termed "treat-as-withdraw", and the latter as "attribute 198 discard". 200 If any of the well-known mandatory attributes are not present in an 201 UPDATE message, then the approach of "treat-as-withdraw" MUST be used 202 for the error handling. 204 The approach of "treat-as-withdraw" MUST be used for the error 205 handling of the cases described in Section 6.3 of [RFC4271] that 206 specify a session reset and involve any of the following attributes: 207 ORIGIN, AS_PATH, NEXT_HOP, MULTI_EXIT_DISC, and LOCAL_PREF. 209 The approach of "attribute discard" MUST be used for the error 210 handling of the cases described in Section 6.3 of [RFC4271] that 211 specify a session reset and involve any of the following attributes: 212 ATOMIC_AGGREGATE and AGGREGATOR. 214 If the MP_REACH_NLRI attribute or the MP_UNREACH_NLRI attribute 215 appears more than once in the UPDATE message, then a NOTIFICATION 216 message MUST be sent with the Error Subcode "Malformed Attribute 217 List". If any other attribute appears more than once in an UPDATE 218 message, then all the occurrences of the attribute other than the 219 first one SHALL be discarded and the UPDATE message continue to be 220 processed. 222 When multiple attribute errors exist in an UPDATE message, if the 223 same approach (either "session reset", or "treat-as-withdraw" or 224 "attribute discard") is specified for the handling of these malformed 225 attributes, then the specified approach MUST be used. Otherwise the 226 approach with the strongest action MUST be used following the order 227 of "session reset", "treat-as-withdraw" and "attribute discard" from 228 the strongest to the weakest. 230 A document which specifies a new attribute MUST provide specifics 231 regarding what constitutes an error for that attribute and how that 232 error is to be handled. 234 Finally, we observe that in order to use the approach of "treat-as- 235 withdraw", the entire NLRI field and/or the MP_REACH_NLRI and 236 MP_UNREACH_NLRI attributes need to be successfully parsed. If this 237 is not possible, the procedures of [RFC4271] continue to apply. 238 Alternatively the error handling procedures specified in [RFC4760] 239 for disabling a particular AFI/SAFI MAY be followed. One notable 240 case where it would be not possible to successfully parse the NLRI is 241 if the NLRI field is found to be "syntactically incorrect" (see 242 Section 3.2). It can be seen that therefore, this part of [RFC4271] 243 Section 6.3 necessarily continues to apply: 245 The NLRI field in the UPDATE message is checked for syntactic 246 validity. If the field is syntactically incorrect, then the Error 247 Subcode MUST be set to Invalid Network Field. 249 Furthermore, this document extends RFC 4271 by mandating that the 250 Withdrawn Routes field SHALL be checked for syntactic correctness in 251 the same manner as the NLRI field. 253 3. Parsing of NLRI Fields 255 To facilitate the determination of the NLRI field in an UPDATE with a 256 malformed attribute, the MP_REACH_NLRI or MP_UNREACH_NLRI attribute 257 (if present) SHALL be encoded as the very first path attribute in an 258 UPDATE. An implementation, however, MUST still be prepared to 259 receive these fields in any position. 261 If the encoding of [RFC4271] is used, the NLRI field for the IPv4 262 unicast address family is carried immediately following all the 263 attributes in an UPDATE. When such an UPDATE is received, we observe 264 that the NLRI field can be determined using the "Message Length", 265 "Withdrawn Route Length" and "Total Attribute Length" (when they are 266 consistent) carried in the message instead of relying on the length 267 of individual attributes in the message. 269 3.1. Inconsistency of Attribute Length Fields 271 There are two error cases in which the Total Attribute Length value 272 can be in conflict with the enclosed path attributes, which 273 themselves carry length values. In the "overrun" case, as the 274 enclosed path attributes are parsed, the length of the last 275 encountered path attribute would cause the Total Attribute Length to 276 be exceeded. In the "underrun" case, as the enclosed path attributes 277 are parsed, after the last successfully-parsed attribute, fewer than 278 three bytes remain, or fewer than four bytes, if the Attribute Flags 279 field has the Extended Length bit set -- that is, there remains 280 unconsumed data in the path attributes but yet insufficient data to 281 encode a single minimum-sized path attribute. In either of these 282 cases an error condition exists and the treat-as-withdraw approach 283 MUST be used (unless some other, more severe error is encountered 284 dictating a stronger approach), and the Total Attribute Length MUST 285 be relied upon to enable the beginning of the NLRI field to be 286 located. 288 3.2. Syntactic Correctness of NLRI Fields 290 The NLRI field or Withdrawn Routes field SHALL be considered 291 "syntactically incorrect" if either of the following are true: 293 o The length of any of the included NLRI is greater than 32, 295 o When parsing NLRI contained in the field, the length of the last 296 NLRI found exceeds the amount of unconsumed data remaining in the 297 field. 299 Similarly, the MP_REACH or MP_UNREACH attribute of an update SHALL be 300 considered to be incorrect if any of the following are true: 302 o The length of any of the included NLRI is inconsistent with the 303 given AFI/SAFI (for example, if an IPv4 NLRI has a length greater 304 than 32 or an IPv6 NLRI has a length greater than 128), 306 o When parsing NLRI contained in the attribute, the length of the 307 last NLRI found exceeds the amount of unconsumed data remaining in 308 the attribute. 310 4. Operational Considerations 312 Although the "treat-as-withdraw" error-handling behavior defined in 313 Section 2 makes every effort to preserve BGP's correctness, we note 314 that if an UPDATE received on an IBGP session is subjected to this 315 treatment, inconsistent routing within the affected Autonomous System 316 may result. The consequences of inconsistent routing can include 317 long-lived forwarding loops and black holes. While lamentable, this 318 issue is expected to be rare in practice, and more importantly is 319 seen as less problematic than the session-reset behavior it replaces. 321 When a malformed attribute is indeed detected over an IBGP session, 322 we RECOMMEND that routes with the malformed attribute be identified 323 and traced back to the ingress router in the network where the routes 324 were sourced or received externally, and then a filter be applied on 325 the ingress router to prevent the routes from being sourced or 326 received. This will help maintain routing consistency in the 327 network. 329 Even if inconsistent routing does not arise, the "treat-as-withdraw" 330 behavior can cause either complete unreachability or sub-optimal 331 routing for the destinations whose routes are carried in the affected 332 UPDATE message. 334 Note that "treat-as-withdraw" is different from discarding an UPDATE 335 message. The latter violates the basic BGP principle of incremental 336 update, and could cause invalid routes to be kept. (See also 337 Appendix A.) 339 For any malformed attribute which is handled by the "attribute 340 discard" instead of the "treat-as-withdraw" approach, it is critical 341 to consider the potential impact of doing so. In particular, if the 342 attribute in question has or may have an effect on route selection or 343 installation, the presumption is that discarding it is unsafe, unless 344 careful analysis proves otherwise. The analysis should take into 345 account the tradeoff between preserving connectivity and potential 346 side effects. 348 Because of these potential issues, a BGP speaker MUST provide 349 debugging facilities to permit issues caused by a malformed attribute 350 to be diagnosed. At a minimum, such facilities MUST include logging 351 an error listing the NLRI involved, and containing the entire 352 malformed UPDATE message when such an attribute is detected. The 353 malformed UPDATE message SHOULD be analyzed, and the root cause 354 SHOULD be investigated. 356 5. Error Handling Procedures for Existing Attributes 358 5.1. ORIGIN 360 The attribute is considered malformed if its length is not 1, or it 361 has an undefined value [RFC4271]. 363 An UPDATE message with a malformed ORIGIN attribute SHALL be handled 364 using the approach of "treat-as-withdraw". 366 5.2. AS_PATH 368 The error conditions for the attribute have been defined in 369 [RFC4271]. 371 An UPDATE message with a malformed AS_PATH attribute SHALL be handled 372 using the approach of "treat-as-withdraw". 374 5.3. NEXT_HOP 376 The error conditions for the NEXT_HOP attribute have been defined in 377 [RFC4271]. 379 An UPDATE message with a malformed NEXT_HOP attribute SHALL be 380 handled using the approach of "treat-as-withdraw". 382 5.4. MULTI_EXIT_DESC 384 The attribute is considered malformed if its length is not 4 385 [RFC4271]. 387 An UPDATE message with a malformed MULTI_EXIT_DESC attribute SHALL be 388 handled using the approach of "treat-as-withdraw". 390 5.5. LOCAL_PREF 392 The attribute is considered malformed if its length is not 4 393 [RFC4271]. 395 An UPDATE message with a malformed LOCAL_PREF attribute SHALL be 396 handled as follows: 398 o using the approach of "attribute discard" if the UPDATE message is 399 received from an external neighbor, or 401 o using the approach of "treat-as-withdraw" if the UPDATE message is 402 received from an internal neighbor. 404 In addition, if the attribute is present in an UPDATE message from an 405 external neighbor, the approach of "attribute discard" SHALL be used 406 to handle the unexpected attribute in the message. 408 5.6. ATOMIC_AGGREGATE 410 The attribute SHALL be considered malformed if its length is not 0 411 [RFC4271]. 413 An UPDATE message with a malformed ATOMIC_AGGREGATE attribute SHALL 414 be handled using the approach of "attribute discard". 416 5.7. AGGREGATOR 418 The error conditions specified in [RFC4271] for the attribute are 419 revised as follows: 421 The AGGREGATOR attribute SHALL be considered malformed if any of the 422 following applies: 424 o Its length is not 6 (when the "4-octet AS number capability" is 425 not advertised to, or not received from the peer [RFC6793]). 427 o Its length is not 8 (when the "4-octet AS number capability" is 428 both advertised to, and received from the peer). 430 An UPDATE message with a malformed AGGREGATOR attribute SHALL be 431 handled using the approach of "attribute discard". 433 5.8. Community 435 The error handling of [RFC1997] is revised as follows: 437 The Community attribute SHALL be considered malformed if its length 438 is nonzero and is not a multiple of 4. 440 An UPDATE message with a malformed Community attribute SHALL be 441 handled using the approach of "treat-as-withdraw". 443 5.9. Extended Community 445 The error handling of [RFC4360] is revised as follows: 447 The Extended Community attribute SHALL be considered malformed if its 448 length is nonzero and is not a multiple of 8. 450 An UPDATE message with a malformed Extended Community attribute SHALL 451 be handled using the approach of "treat-as-withdraw". 453 Note that a BGP speaker MUST NOT treat an unrecognized Extended 454 Community Type or Sub-Type as an error. 456 5.10. IPv6 Address Specific BGP Extended Community Attribute 458 The error handling of [RFC5701] is revised as follows: 460 The IPv6 Address Specific Extended Community attribute SHALL be 461 considered malformed if its length is nonzero and is not a multiple 462 of 20. 464 An UPDATE message with a malformed IPv6 Address Specific Extended 465 Community attribute SHALL be handled using the approach of "treat-as- 466 withdraw". 468 Note that a BGP speaker MUST NOT treat an unrecognized IPv6 Address 469 Specific Extended Community Type or Sub-Type as an error. 471 5.11. ORIGINATOR_ID 473 The error handling of [RFC4456] is revised as follows. 475 o If the ORIGINATOR_ID attribute is received from an external 476 neighbor, it SHALL be discarded using the approach of "attribute 477 discard", or 479 o if received from an internal neighbor, it SHALL be considered 480 malformed if its length is not equal to 4. If malformed, the 481 UPDATE SHALL be handled using the approach of "treat-as-withdraw". 483 5.12. CLUSTER_LIST 485 The error handling of [RFC4456] is revised as follows. 487 o If the CLUSTER_LIST attribute is received from an external 488 neighbor, it SHALL be discarded using the approach of "attribute 489 discard", or 491 o if received from an internal neighbor, it SHALL be considered 492 malformed if its length is not a multiple 4. If malformed, the 493 UPDATE SHALL be handled using the approach of "treat-as-withdraw". 495 6. IANA Considerations 497 This document makes no request of IANA. 499 7. Security Considerations 501 This specification addresses the vulnerability of a BGP speaker to a 502 potential attack whereby a distant attacker can generate a malformed 503 optional transitive attribute that is not recognized by intervening 504 routers (which thus propagate the attribute unchecked) but that 505 causes session resets when it reaches routers that do recognize the 506 given attribute type. 508 In other respects, this specification does not change BGP's security 509 characteristics. 511 8. Acknowledgements 513 The authors wish to thank Juan Alcaide, Ron Bonica, Mach Chen, Andy 514 Davidson, Bruno Decraene, Dong Jie, Rex Fernando, Joel Halpern, Akira 515 Kato, Miya Kohno, Tony Li, Alton Lo, Shin Miyakawa, Tamas Mondal, 516 Jonathan Oddy, Robert Raszuk, Yakov Rekhter, Eric Rosen, Rob Shakir, 517 Naiming Shen, Shyam Sethuram, Ananth Suryanarayana, Kaliraj 518 Vairavakkalai and Lili Wang for their observations and discussion of 519 this topic, and review of this document. 521 9. Normative References 523 [RFC1997] Chandrasekeran, R., Traina, P., and T. Li, "BGP 524 Communities Attribute", RFC 1997, August 1996. 526 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 527 Requirement Levels", BCP 14, RFC 2119, March 1997. 529 [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway 530 Protocol 4 (BGP-4)", RFC 4271, January 2006. 532 [RFC4360] Sangli, S., Tappan, D., and Y. Rekhter, "BGP Extended 533 Communities Attribute", RFC 4360, February 2006. 535 [RFC4456] Bates, T., Chen, E., and R. Chandra, "BGP Route 536 Reflection: An Alternative to Full Mesh Internal BGP 537 (IBGP)", RFC 4456, April 2006. 539 [RFC4760] Bates, T., Chandra, R., Katz, D., and Y. Rekhter, 540 "Multiprotocol Extensions for BGP-4", RFC 4760, January 541 2007. 543 [RFC5701] Rekhter, Y., "IPv6 Address Specific BGP Extended Community 544 Attribute", RFC 5701, November 2009. 546 [RFC6793] Vohra, Q. and E. Chen, "BGP Support for Four-Octet 547 Autonomous System (AS) Number Space", RFC 6793, December 548 2012. 550 Appendix A. Why not discard UPDATE messages? 552 A commonly asked question is "why not simply discard the UPDATE 553 message instead of treating it like a withdraw? Isn't that safer and 554 easier?" The answer is that it might be easier, but it would 555 compromise BGP's correctness so is unsafe. Consider the following 556 example of what might happen if UPDATE messages carrying bad 557 attributes were simply discarded: 559 AS1 ---- AS2 560 \ / 561 \ / 562 \ / 563 AS3 565 Figure 1 567 o AS1 prefers to reach AS3 directly, and advertises its route to 568 AS2. 570 o AS2 prefers to reach AS3 directly, and advertises its route to 571 AS1. 573 o Connections AS3-AS1 and AS3-AS2 fail simultaneously. 575 o AS1 switches to prefer AS2's route, and sends an update message 576 which includes a withdraw of its previous announcement. The 577 withdraw is bundled with some advertisements. It includes a bad 578 attribute. As a result, AS2 ignores the message. 580 o AS2 switches to prefer AS1's route, and sends an update message 581 which includes a withdraw of its previous announcement. The 582 withdraw is bundled with some advertisements. It includes a bad 583 attribute. As a result, AS1 ignores the message. 585 The end result is that AS1 forwards traffic for AS3 towards AS2, and 586 AS2 forwards traffic for AS3 towards AS1. This is a permanent (until 587 corrected) forwarding loop. 589 Although the example above discusses route withdraws, we observe that 590 in BGP the announcement of a route also withdraws the route 591 previously advertised. The implicit withdraw can be converted into a 592 real withdraw in a number of ways; for example, the previously- 593 announced route might have been accepted by policy, but the new 594 announcement might be rejected by policy. For this reason, the same 595 concerns apply even if explicit withdraws are removed from 596 consideration. 598 Authors' Addresses 600 Enke Chen (editor) 601 Cisco Systems, Inc. 603 Email: enkechen@cisco.com 604 John G. Scudder (editor) 605 Juniper Networks 607 Email: jgs@juniper.net 609 Pradosh Mohapatra 610 Cumulus Networks, Inc. 612 Email: pmohapat@cumulusnetworks.com 614 Keyur Patel 615 Cisco Systems, Inc. 617 Email: keyupate@cisco.com