idnits 2.17.1 draft-ietf-grow-bgp-med-considerations-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 16. -- Found old boilerplate from RFC 3978, Section 5.5 on line 542. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 519. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 526. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 532. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 548), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 35. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 3 instances of too long lines in the document, the longest one being 1 character in excess of 72. == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (December 2005) is 6705 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: 'RFC 1519' is defined on line 471, but no explicit reference was found in the text ** Obsolete normative reference: RFC 1519 (Obsoleted by RFC 4632) ** Obsolete normative reference: RFC 1771 (Obsoleted by RFC 4271) ** Obsolete normative reference: RFC 2796 (Obsoleted by RFC 4456) ** Obsolete normative reference: RFC 3065 (Obsoleted by RFC 5065) -- Possible downref: Non-RFC (?) normative reference: ref. 'BGP4' Summary: 9 errors (**), 0 flaws (~~), 4 warnings (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 INTERNET-DRAFT Danny McPherson 2 Arbor Networks, Inc. 3 Vijay Gill 4 AOL 5 Category Informational 6 Expires: June 2006 December 2005 8 BGP MULTI_EXIT_DISC (MED) Considerations 9 11 Status of this Memo 13 By submitting this Internet-Draft, each author represents that any 14 applicable patent or other IPR claims of which he or she is aware have 15 been or will be disclosed, and any of which he or she becomes aware 16 will be disclosed, in accordance with Section 6 of BCP 79. 18 Internet-Drafts are working documents of the Internet Engineering 19 Task Force (IETF), its areas, and its working groups. Note that other 20 groups may also distribute working documents as Internet-Drafts. 22 Internet-Drafts are draft documents valid for a maximum of six 23 months and may be updated, replaced, or obsoleted by other documents 24 at any time. It is inappropriate to use Internet-Drafts as reference 25 material or to cite them other than as "work in progress". 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/1id-abstracts.html 30 The list of Internet-Draft Shadow Directories can be accessed at 31 http://www.ietf.org/shadow.html. 33 Copyright Notice 35 Copyright (C) The Internet Society (2005). All Rights Reserved. 37 Abstract 39 The BGP MULTI_EXIT_DISC (MED) attribute provides a mechanism for BGP 40 speakers to convey to an adjacent AS the optimal entry point into the 41 local AS. While BGP MEDs function correctly in many scenarios, there 42 are a number of issues which may arise when utilizing MEDs in dynamic 43 or complex topologies. 45 This document discusses implementation and deployment considerations 46 regarding BGP MEDs and provides information which implementors and 47 network operators should be familiar with. 49 Table of Contents 51 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 52 2. Specification of Requirements. . . . . . . . . . . . . . . . . 4 53 2.1. About the MULTI_EXIT_DISC (MED) Attribute . . . . . . . . . 4 54 2.2. MEDs and Potatos. . . . . . . . . . . . . . . . . . . . . . 6 55 3. Implementation and Protocol Considerations . . . . . . . . . . 7 56 3.1. MULTI_EXIT_DISC is a Optional Non-Transitive 57 Attribute. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 58 3.2. MED Values and Preferences. . . . . . . . . . . . . . . . . 7 59 3.3. Comparing MEDs Between Different Autonomous 60 Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 61 3.4. MEDs, Route Reflection and AS Confederations 62 for BGP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 63 3.5. Route Flap Damping and MED Churn. . . . . . . . . . . . . . 9 64 3.6. Effects of MEDs on Update Packing Efficiency. . . . . . . . 10 65 3.7. Temporal Route Selection. . . . . . . . . . . . . . . . . . 11 66 4. Deployment Considerations. . . . . . . . . . . . . . . . . . . 11 67 4.1. Comparing MEDs Between Different Autonomous 68 Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 69 4.2. Effects of Aggregation on MEDs` . . . . . . . . . . . . . . 12 70 5. IANA Considerations. . . . . . . . . . . . . . . . . . . . . . 12 71 6. Security Considerations. . . . . . . . . . . . . . . . . . . . 12 72 7. Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . 13 73 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 14 74 8.1. Normative References. . . . . . . . . . . . . . . . . . . . 15 75 8.2. Informative References. . . . . . . . . . . . . . . . . . . 16 76 9. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 16 78 1. Introduction 80 The BGP MED attribute provides a mechanism for BGP speakers to convey 81 to an adjacent AS the optimal entry point into the local AS. While 82 BGP MEDs function correctly in many scenarios, there are a number of 83 issues which may arise when utilizing MEDs in dynamic or complex 84 topologies. 86 While reading this document it's important to keep in mind that the 87 goal is to discuss both implementation and deployment considerations 88 regarding BGP MEDs. In addition, the intention is to provide 89 guidance that both implementors and network operators should be 90 familiar with. In some instances implementation advice varies from 91 deployment advice. 93 2. Specification of Requirements 95 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 96 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 97 document are to be interpreted as described in [RFC 2119]. 99 2.1. About the MULTI_EXIT_DISC (MED) Attribute 101 The BGP MULTI_EXIT_DISC (MED) attribute, formerly known as the 102 INTER_AS_METRIC, is currently defined in section 5.1.4 of [BGP4], as 103 follows: 105 The MULTI_EXIT_DISC is an optional non-transitive attribute which 106 is intended to be used on external (inter-AS) links to discriminate 107 among multiple exit or entry points to the same neighboring AS. 108 The value of the MULTI_EXIT_DISC attribute is a four octet unsigned 109 number which is called a metric. All other factors being equal, the 110 exit point with lower metric SHOULD be preferred. If received over 111 EBGP, the MULTI_EXIT_DISC attribute MAY be propagated over IBGP to 112 other BGP speakers within the same AS (see also 9.1.2.2). The 113 MULTI_EXIT_DISC attribute received from a neighboring AS MUST NOT 114 be propagated to other neighboring ASs. 116 A BGP speaker MUST implement a mechanism based on local 117 configuration which allows the MULTI_EXIT_DISC attribute to be 118 removed from a route. If a BGP speaker is configured to remove the 119 MULTI_EXIT_DISC attribute from a route, then this removal MUST be 120 done prior to determining the degree of preference of the route and 121 performing route selection (Decision Process phases 1 and 2). 123 An implementation MAY also (based on local configuration) alter the 124 value of the MULTI_EXIT_DISC attribute received over EBGP. If a 125 BGP speaker is configured to alter the value of the MULTI_EXIT_DISC 126 attribute received over EBGP, then altering the value MUST be done 127 prior to determining the degree of preference of the route and 128 performing route selection (Decision Process phases 1 and 2). See 129 Section 9.1.2.2 of BGP4] for necessary restrictions on this. 131 Section 9.1.2.2 (c) of [BGP4] defines the following route selection 132 criteria regarding MEDs: 134 c) Remove from consideration routes with less-preferred 135 MULTI_EXIT_DISC attributes. MULTI_EXIT_DISC is only comparable 136 between routes learned from the same neighboring AS (the neighbor- 137 ing AS is determined from the AS_PATH attribute). Routes which do 138 not have the MULTI_EXIT_DISC attribute are considered to have the 139 lowest possible MULTI_EXIT_DISC value. 141 This is also described in the following procedure: 143 for m = all routes still under consideration 144 for n = all routes still under consideration 145 if (neighborAS(m) == neighborAS(n)) and (MED(n) < MED(m)) 146 remove route m from consideration 148 In the pseudo-code above, MED(n) is a function which returns the 149 value of route n's MULTI_EXIT_DISC attribute. If route n has no 150 MULTI_EXIT_DISC attribute, the function returns the lowest possi- 151 ble MULTI_EXIT_DISC value, i.e. 0. 153 If a MULTI_EXIT_DISC attribute is removed before re-advertising a 154 route into IBGP, then comparison based on the received EBGP 155 MULTI_EXIT_DISC attribute MAY still be performed. If an 156 implementation chooses to remove MULTI_EXIT_DISC, then the optional 157 comparison on MULTI_EXIT_DISC if performed at all MUST be performed 158 only among EBGP learned routes. The best EBGP learned route may 159 then be compared with IBGP learned routes after the removal of the 160 MULTI_EXIT_DISC attribute. If MULTI_EXIT_DISC is removed from a 161 subset of EBGP learned routes and the selected "best" EBGP learned 162 route will not have MULTI_EXIT_DISC removed, then the 163 MULTI_EXIT_DISC must be used in the comparison with IBGP learned 164 routes. For IBGP learned routes the MULTI_EXIT_DISC MUST be used in 165 route comparisons which reach this step in the Decision Process. 167 Including the MULTI_EXIT_DISC of an EBGP learned route in the 168 comparison with an IBGP learned route, then removing the 169 MULTI_EXIT_DISC attribute and advertising the route has been proven 170 to cause route loops. 172 2.2. MEDs and Potatos 174 In a situation where traffic flows between a pair of hosts, each 175 connected to different transit networks, which are themselves 176 interconnected at two or more locations, each transit network has the 177 choice of either sending traffic to the closest peering to the 178 adjacent transit network or passing traffic to the interconnection 179 location which advertises the least cost path to the destination 180 host. 182 The former method is called "hot potato routing" (or closest-exit) 183 because like a hot potato held in bare hands, whoever has it tries to 184 get rid of it quickly. Hot potato routing is accomplished by not 185 passing the EGBP learned MED into IBGP. This minimizes transit 186 traffic for the provider routing the traffic. Far less common is 187 "cold potato routing" (or best-exit) where the transit provider uses 188 their own transit capacity to get the traffic to the point that 189 adjacent transit provider advertised as being closest to the 190 destination. Cold potato routing is accomplished by passing the EBGP 191 learned MED into IBGP. 193 If one transit provider uses hot potato routing and another uses cold 194 potato, traffic between the two tends to be more symmetric. However, 195 if both providers employ cold potato routing, or both providers 196 employ hot potato routing between their networks, it's likely that a 197 larger amount of asymmetry would exist. 199 Depending on the business relationships, if one provider has more 200 capacity or a significantly less congested backbone network, then 201 that provider may use cold potato routing. An example of widespread 202 use of cold potato routing was the NSF funded NSFNET backbone and NSF 203 funded regional networks in the mid 1990s. 205 In some cases a provider may use hot potato routing for some 206 destinations for a given peer AS and cold potato routing for others. 207 An example of this is the different treatment of commercial and 208 research traffic in the NSFNET in the mid 1990s. Today many 209 commercial networks exchange MEDs with customers but not bilateral 210 peers. However, commercial use of MEDs varies widely, from 211 ubiquitous use of MEDs to no use of MEDs at all. 213 In addition, many deployments of MEDs today are likely behaving 214 differently (e.g., resulting is sub-optimal routing) than the network 215 operator intended, thereby resulting not in hot or cold potatos, but 216 mashed potatos! More information on unintended behavior resulting 217 from MEDs is provided throughout this document. 219 3. Implementation and Protocol Considerations 221 There are a number of implementation and protocol peculiarities 222 relating to MEDs that have been discovered that may affect network 223 behavior. The following sections provide information on these 224 issues. 226 3.1. MULTI_EXIT_DISC is a Optional Non-Transitive Attribute 228 MULTI_EXIT_DISC is a non-transitive optional attribute whose 229 advertisement to both IBGP and EBGP peers is discretionary. As a 230 result, some implementations enable sending of MEDs to IBGP peers by 231 default, while others do not. This behavior may result in sub- 232 optimal route selection within an AS. In addition, some 233 implementations send MEDs to EBGP peers by default, while others do 234 not. This behavior may result in sub-optimal inter-domain route 235 selection. 237 3.2. MED Values and Preferences 239 Some implementations consider an MED value of zero as less preferable 240 than no MED value. This behavior resulted in path selection 241 inconsistencies within an AS. The current draft version of the BGP 242 specification [BGP4] removes ambiguities that existed in [RFC 1771] 243 by stating that if route n has no MULTI_EXIT_DISC attribute, the 244 lowest possible MULTI_EXIT_DISC value (i.e. 0) should be assigned to 245 the attribute. 247 It is apparent that different implementations and different versions 248 of the BGP draft specification have been all over the map with 249 interpretation of missing-MED. For example, earlier versions of the 250 specification called for a missing MED to be assigned the highest 251 possible MED value (i.e., 2^32-1). 253 In addition, some implementations have been shown to internally 254 employ a maximum possible MED value (2^32-1) as an "infinity" metric 255 (i.e., the MED value is used to tag routes as unfeasible), and would 256 upon on receiving an update with an MED value of 2^32-1 rewrite the 257 value to 2^32-2. Subsequently, the new MED value would be propagated 258 and could result in routing inconsistencies or unintended path 259 selections. 261 As a result of implementation inconsistencies and protocol revision 262 variances, many network operators today explicitly reset (i.e., set 263 to zero or some other 'fixed' value) all MED values on ingress to 264 conform to their internal routing policies (i.e., to include policy 265 that requires that MED values of 0 and 2^32-1 not be used in 266 configurations, whether the MEDs are directly computed or 267 configured), so as to not have to rely on all their routers having 268 the same missing-MED behavior. 270 Because implementations don't normally provide a mechanism to disable 271 MED comparisons in the decision algorithm, "not using MEDs" usually 272 entails explicitly setting all MEDs to some fixed value upon ingress 273 to the routing domain. By assigning a fixed MED value consistently 274 to all routes across the network, MEDs are a effectively a non-issue 275 in the decision algorithm. 277 3.3. Comparing MEDs Between Different Autonomous Systems 279 The MED was intended to be used on external (inter-AS) links to 280 discriminate among multiple exit or entry points to the same 281 neighboring AS. However, a large number of MED applications now 282 employ MEDs for the purpose of determining route preference between 283 like routes received from different autonomous systems. 285 A large number of implementations provide the capability to enable 286 comparison of MEDs between routes received from different neighboring 287 autonomous systems. While this capability has demonstrated some 288 benefit (e.g., that described in [RFC 3345]), operators should be 289 wary of the potential side effects with enabling such a function. 290 The deployment section below provides some examples as to why this 291 may result in undesirable behavior. 293 3.4. MEDs, Route Reflection and AS Confederations for BGP 295 In particular configurations, the BGP scaling mechanisms defined in 296 "BGP Route Reflection - An Alternative to Full Mesh IBGP" [RFC 2796] 297 and "Autonomous System Confederations for BGP" [RFC 3065] will 298 introduce persistent BGP route oscillation [RFC 3345]. The problem 299 is inherent in the way BGP works: a conflict exists between 300 information hiding/hierarchy and the non-hierarchical selection 301 process imposed by lack of total ordering caused by the MED rules. 302 Given current practices, we see the problem most frequently manifest 303 itself in the context of MED + route reflectors or confederations. 305 One potential way to avoid this is by configuring inter-Member-AS or 306 inter-cluster IGP metrics higher than intra-Member-AS IGP metrics 307 and/or using other tie breaking policies to avoid BGP route selection 308 based on incomparable MEDs. Of course, IGP metric constraints may be 309 unreasonably onerous for some applications. 311 Not comparing MEDs between multiple paths for a prefix learned from 312 different adjacent autonomous systems, as discussed in section 2.3), 313 or not utilizing MEDs at all, significantly decreases the probability 314 of introducing potential route oscillation conditions into the 315 network. 317 Although perhaps "legal" as far as current specifications are 318 concerned, modifying MED attributes received on any type of IBGP 319 session (e.g., standard IBGP, AS confederations EIBGP, route 320 reflection, etc..) is not recommended. 322 3.5. Route Flap Damping and MED Churn 324 MEDs are often derived dynamically from IGP metrics or additive costs 325 associated with an IGP metric to a given BGP NEXT_HOP. This 326 typically provides an efficient model for ensuring that the BGP MED 327 advertised to peers used to represent the best path to a given 328 destination within the network is aligned with that of the IGP within 329 a given AS. 331 The consequence with dynamically derived IGP-based MEDs is that 332 instability within an AS, or even on a single given link within the 333 AS, can result in wide-spread BGP instability or BGP route 334 advertisement churn that propagates across multiple domains. In 335 short, if your MED "flaps" every time your IGP metric flaps, you're 336 routes are likely going to be suppressed as a result of BGP Route 337 Flap Damping [RFC 2439]. 339 Employment of MEDs may compound the adverse effects of BGP flap 340 dampening behavior because it many cause routes to be re- advertised 341 solely to reflect an internal topology change. 343 Many implementations don't have a practical problem with IGP 344 flapping, they either latch their IGP metric upon first advertisement 345 or they employ some internal suppression mechanism. Some 346 implementations regard BGP attribute changes as less significant than 347 route withdrawals and announcements to attempt to mitigate the impact 348 of this type of event. 350 3.6. Effects of MEDs on Update Packing Efficiency 352 Multiple unfeasible routes can be advertised in a single BGP Update 353 message. The BGP4 protocol also permits advertisement of multiple 354 prefixes with a common set of path attributes to be advertised in a 355 single update message, this is commonly referred to as "update 356 packing". When possible, update packing is recommended as it 357 provides a mechanism for more efficient behavior in a number of 358 areas, to include: 360 o Reduction in system overhead due to generation or receipt of 361 fewer Update messages. 363 o Reduction in network overhead as a result of fewer packets and 364 lower bandwidth consumption. 366 o Allows processing of path attributes and searches for matching 367 sets in your AS_PATH database (if you have one) less frequently. 368 Consistent ordering of the path attributes allows for ease of 369 matching in the database as you don't have different 370 representations 371 of the same data. 373 Update packing requires that all feasible routes within a single 374 update message share a common attribute set, to include a common 375 MULTI_EXIT_DISC value. As such, potential wide-scale variance in MED 376 values introduces another variable and may resulted in a marked 377 decrease in update packing efficiency. 379 3.7. Temporal Route Selection 381 Some implementations have had bugs which lead to temporal behavior in 382 MED-based best path selection. These usually involved methods used 383 to store the oldest route along with ordering routes for MED in 384 earlier implementations that cause non-deterministic behavior on 385 whether the oldest route would truly be selected or not. 387 The reasoning for this is that older paths are presumably more 388 stable, and thus more preferable. However, temporal behavior in 389 route selection results in non-deterministic behavior, and as such, 390 is often undesirable. 392 4. Deployment Considerations 394 It has been discussed that accepting MEDs from other autonomous 395 systems have the potential to cause traffic flow churns in the 396 network. Some implementations only ratchet down the MED and never 397 move it back up to prevent excessive churn. 399 However, if a session is reset, the MEDs being advertised have the 400 potential of changing. If an network is relying on received MEDs to 401 route traffic properly, the traffic patterns have the potential for 402 changing dramatically, potentially resulting in congestion on the 403 network. Essentially, accepting and routing traffic based on MEDs 404 allows other people to traffic engineer your network. This may or may 405 not be acceptable to you. 407 As previously discussed, many network operators choose to reset MED 408 values on ingress. In addition, many operators explicitly do not 409 employ MED values of 0 or 2^32-1 in order to avoid inconsistencies 410 with implementations and various revisions of the BGP specification. 412 4.1. Comparing MEDs Between Different Autonomous Systems 414 Although the MED was meant to only be used when comparing paths 415 received from different external peers in the same AS, many 416 implementations provide the capability to compare MEDs between 417 different autonomous systems as well. AS operators often use 418 LOCAL_PREF to select the external preferences (primary, secondary 419 upstreams, peers, customers, etc.), using MED instead of LOCAL_PREF 420 would possibility lead to an inconsistent distribution of best routes 421 as MED is compared only after the AS_PATH length. 423 Though this may seem a fine idea for some configurations, care must 424 be taken when comparing MEDs between different autonomous systems. 425 BGP speakers often derive MED values by obtaining the IGP metric 426 associated with reaching a given BGP NEXT_HOP within the local AS. 427 This allows MEDs to reasonably reflect IGP topologies when 428 advertising routes to peers. While this is fine when comparing MEDs 429 between multiple paths learned from a single AS, it can result in 430 potentially "weighted" decisions when comparing MEDs between 431 different autonomous systems. This is most typically the case when 432 the autonomous systems use different mechanisms to derive IGP 433 metrics, BGP MEDs, or perhaps even use different IGP protocols with 434 vastly contrasting metric spaces (e.g., OSPF v. traditional metric 435 space in IS-IS). 437 4.2. Effects of Aggregation on MEDs` 439 Another MED deployment consideration involves the impact that 440 aggregation of BGP routing information has on MEDs. Aggregates are 441 often generated from multiple locations in an AS in order to 442 accommodate stability, redundancy and other network design goals. 443 When MEDs are derived from IGP metrics associated with said 444 aggregates the MED value advertised to peers can result in very 445 suboptimal routing. 447 5. IANA Considerations 449 This document introduces no new IANA considerations. 451 6. Security Considerations 453 The MED was purposely designed to be a "weak" metric that would only 454 be used late in the best-path decision process. The BGP working 455 group was concerned that any metric specified by a remote operator 456 would only affect routing in a local AS IF no other preference was 457 specified. A paramount goal of the design of the MED was to ensure 458 that peers could not "shed" or "absorb" traffic for networks that 459 they advertise. As such, accepting MEDs from peers may in some sense 460 increase a network's susceptibility to exploitation by peers. 462 7. Acknowledgments 464 Thanks to John Scudder for applying his usual keen eye and 465 constructive insight. Also, thanks to Curtis Villamizar, JR Mitchell 466 and Pekka Savola for their valuable feedback. 468 8. References 469 8.1. Normative References 471 [RFC 1519] Fuller, V., Li. T., Yu J., and K. Varadhan, "Classless 472 Inter-Domain Routing (CIDR): an Address Assignment and 473 Aggregation Strategy", RFC 1519, September 1993. 475 [RFC 1771] Rekhter, Y., and T. Li, "A Border Gateway Protocol 4 476 (BGP-4)", RFC 1771, March 1995. 478 [RFC 2119] Bradner, S., "Key words for use in RFCs to Indicate 479 Requirement Levels", RFC 2119, March 1997. 481 [RFC 2796] Bates, T., Chandra, R., Chen, E., "BGP Route Reflection 482 - An Alternative to Full Mesh IBGP", RFC 2796, April 483 2000. 485 [RFC 3065] Traina, P., McPherson, D., Scudder, J.. "Autonomous System 486 Confederations for BGP", RFC 3065, February 2001. 488 [BGP4] Rekhter, Y., T. Li., and Hares. S, Editors, "A Border 489 Gateway Protocol 4 (BGP-4)", BGP Draft, Work in Progress. 491 8.2. Informative References 493 [RFC 2439] Villamizar, C. and Chandra, R., "BGP Route Flap Damping", 494 RFC 2439, November 1998. 496 [RFC 3345] McPherson, D., Gill, V., Walton, D., and Retana, A, "BGP 497 Persistent Route Oscillation Condition", RFC 3345, 498 August 2002. 500 9. Authors' Addresses 502 Danny McPherson 503 Arbor Networks 504 Email: danny@arbor.net 506 Vijay Gill 507 AOL 508 Email: VijayGill9@aol.com 510 Intellectual Property Statement 512 The IETF takes no position regarding the validity or scope of any 513 Intellectual Property Rights or other rights that might be claimed to 514 pertain to the implementation or use of the technology described in 515 this document or the extent to which any license under such rights 516 might or might not be available; nor does it represent that it has 517 made any independent effort to identify any such rights. Information 518 on the procedures with respect to rights in RFC documents can be 519 found in BCP 78 and BCP 79. 521 Copies of IPR disclosures made to the IETF Secretariat and any 522 assurances of licenses to be made available, or the result of an 523 attempt made to obtain a general license or permission for the use of 524 such proprietary rights by implementers or users of this 525 specification can be obtained from the IETF on-line IPR repository at 526 http://www.ietf.org/ipr. 528 The IETF invites any interested party to bring to its attention any 529 copyrights, patents or patent applications, or other proprietary 530 rights that may cover technology that may be required to implement 531 this standard. Please address the information to the IETF at 532 ietf-ipr@ietf.org. 534 Disclaimer of Validity 536 This document and the information contained herein are provided on an 537 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 538 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 539 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 540 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 541 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 542 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 544 Copyright Statement 546 Copyright (C) The Internet Society (2005). This document is subject 547 to the rights, licenses and restrictions contained in BCP 78, and 548 except as set forth therein, the authors retain all their rights. 550 Acknowledgment 552 Funding for the RFC Editor function is currently provided by the 553 Internet Society.