idnits 2.17.1 draft-ietf-grow-bgp-med-considerations-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by 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 == The page length should not exceed 58 lines per page, but there was 12 longer pages, the longest (page 13) being 70 lines == It seems as if not all pages are separated by form feeds - found 0 form feeds but 14 pages 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. ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 104: '...ith lower metric SHOULD be preferred. ...' RFC 2119 keyword, line 105: '...T_DISC attribute MAY be propagated ove...' RFC 2119 keyword, line 107: '... neighboring AS MUST NOT be propagate...' RFC 2119 keyword, line 110: '... A BGP speaker MUST IMPLEMENT a mech...' RFC 2119 keyword, line 112: '...m a route. This MAY be done prior to ...' (6 more instances...) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == Line 371 has weird spacing: '...us more pre...' -- 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 (July 2004) is 7223 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) == Missing Reference: 'RFC 2119' is mentioned on line 35, but not defined == Unused Reference: 'RFC 1519' is defined on line 448, 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) ** Downref: Normative reference to an Informational RFC: RFC 3345 -- Possible downref: Non-RFC (?) normative reference: ref. 'BGP4' Summary: 9 errors (**), 0 flaws (~~), 7 warnings (==), 3 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: January 2005 July 2004 8 BGP MED Considerations 9 11 Status of this Document 13 This document is an Internet-Draft and is in full conformance with 14 all provisions of Section 10 of RFC2026. 16 Internet-Drafts are working documents of the Internet Engineering 17 Task Force (IETF), its areas, and its working groups. Note that 18 other groups may also distribute working documents as Internet- 19 Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at 27 http://www.ietf.org/ietf/1id-abstracts.txt 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 The key words "MUST"", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 33 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 34 document are to be interpreted as described in RFC 2119 [RFC 2119]. 36 This document is a product of an individual. Comments are solicited 37 and should be addressed to the author(s). 39 Copyright Notice 41 Copyright (C) The Internet Society (2004). All Rights Reserved. 43 Abstract 45 The BGP MED attribute provides a mechanism for BGP speakers to convey 46 to an adjacent AS the optimal entry point into the local AS. While 47 BGP MEDs function correctly in many scenarios, there are a number of 48 issues which may arise when utilizing MEDs in dynamic or complex 49 topologies. 51 This document discusses implementation and deployment considerations 52 regarding BGP MEDs and provides information which implementors and 53 network operators should be familiar with. 55 Table of Contents 57 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 58 1.1. About the MULTI_EXIT_DISC (MED) Attribute . . . . . . . . . 4 59 1.2. MEDs and Potatoes . . . . . . . . . . . . . . . . . . . . . 5 60 2. Implementation and Protocol Considerations . . . . . . . . . . 6 61 2.1. MULTI_EXIT_DISC is a Optional Non-Transitive 62 Attribute. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 63 2.2. MED Values and Preferences. . . . . . . . . . . . . . . . . 7 64 2.3. Comparing MEDs Between Different Autonomous 65 Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 66 2.4. MEDs, Route Reflection and AS Confederations 67 for BGP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 68 2.5. Route Flap Damping and MED Churn. . . . . . . . . . . . . . 9 69 2.6. Effects of MEDs on Update Packing Efficiency. . . . . . . . 9 70 2.7. Temporal Route Selection. . . . . . . . . . . . . . . . . . 10 71 3. Deployment Considerations. . . . . . . . . . . . . . . . . . . 10 72 3.1. Comparing MEDs Between Different Autonomous 73 Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 74 3.2. Effects of Aggregation on MEDs` . . . . . . . . . . . . . . 11 75 4. Security Considerations. . . . . . . . . . . . . . . . . . . . 12 76 4.1. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . 12 77 5. References . . . . . . . . . . . . . . . . . . . . . . . . . . 13 78 6. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 13 79 7. Full Copyright Statement . . . . . . . . . . . . . . . . . . . 13 81 1. Introduction 83 The BGP MED attribute provides a mechanism for BGP speakers to convey 84 to an adjacent AS the optimal entry point into the local AS. While 85 BGP MEDs function correctly in many scenarios, there are a number of 86 issues which may arise when utilizing MEDs in dynamic or complex 87 topologies. 89 This document discusses implementation and deployment considerations 90 regarding BGP MEDs and provides information which implementors and 91 network operators should be familiar with. 93 1.1. About the MULTI_EXIT_DISC (MED) Attribute 95 The BGP MUTLI_EXIT_DISC (MED) attribute, formerly known as the 96 INTER_AS_METRIC, is currently defined in section 5.1.4 of [BGP4], as 97 follows: 99 MULTI_EXIT_DISC is an optional non-transitive attribute which is 100 intended to be used on external (inter-AS) links to discriminate 101 among multiple exit or entry points to the same neighboring AS. 102 The MULTI_EXIT_DISC attribute is a four octet unsigned number which 103 is called a metric. All other factors being equal, the exit point 104 with lower metric SHOULD be preferred. If received over EBGP, the 105 MULTI_EXIT_DISC attribute MAY be propagated over IBGP to other BGP 106 speakers within the same AS. An MED attribute received from a 107 neighboring AS MUST NOT be propagated to other neighboring 108 autonomous systems. 110 A BGP speaker MUST IMPLEMENT a mechanism based on local 111 configuration which allows the MULTI_EXIT_DISC attribute to be 112 removed from a route. This MAY be done prior to determining the 113 degree of preference of the route and performing route selection 114 (decision process phases 1 and 2). 116 An implementation MAY also (based on local configuration) alter the 117 value of the MULTI_EXIT_DISC attribute received over EBGP. This 118 MAY be done prior to determining the degree of preference of the 119 route and performing route selection (decision process phases 1 and 120 2). 122 Section 9.1.2.2 (c) of [BGP4] defines the following route selection 123 criteria regarding MEDs: 125 Remove from consideration routes with less-preferred 126 MULTI_EXIT_DISC attributes. MULTI_EXIT_DISC is only comparable 127 between routes learned from the same neighboring AS (the 128 neighboring AS is determined from the AS_PATH attribute). Routes 129 which do not have the MULTI_EXIT_DISC attribute are considered to 130 have the lowest possible MULTI_EXIT_DISC value. 132 This is also described in the following procedure: 134 for m = all routes still under consideration 135 for n = all routes still under consideration 136 if (neighborAS(m) == neighborAS(n)) and (MED(n) < MED(m)) 137 remove route m from consideration 139 In the pseudo-code above, MED(n) is a function which returns the 140 value of route n's MULTI_EXIT_DISC attribute. If route n has no 141 MULTI_EXIT_DISC attribute, the function returns the lowest possible 142 MULTI_EXIT_DISC value, i.e. 0. 144 If a MULTI_EXIT_DISC attribute is removed before re- advertising a 145 route into IBGP, then comparison based on the received EBGP 146 MULTI_EXIT_DISC attribute MAY still be performed. If an 147 implementation chooses to remove MULTI_EXIT_DISC, then the optional 148 comparison on MULTI_EXIT_DISC if performed at all MUST be performed 149 only among EBGP learned routes. The best EBGP learned route may 150 then be compared with IBGP learned routes after the removal of the 151 MULTI_EXIT_DISC attribute. If MULTI_EXIT_DISC is removed from a 152 subset of EBGP learned routes and the selected "best" EBGP learned 153 route will not have MULTI_EXIT_DISC removed, then the 154 MULTI_EXIT_DISC must be used in the comparison with IBGP learned 155 routes. For IBGP learned routes the MULTI_EXIT_DISC MUST be used in 156 route comparisons which reach this step in the decision process. 157 Including the MULTI_EXIT_DISC of an EBGP learned route in the 158 comparison with an IBGP learned route, then removing the 159 MULTI_EXIT_DISC attribute and advertising the route has been proven 160 to cause route loops. 162 Routes that have different MULTI_EXIT_DISC attribute SHALL NOT be 163 aggregated. 165 1.2. MEDs and Potatoes 167 In a situation where traffic flows between a pair of hosts, each 168 connected to different transit networks, which are themselves 169 interconnected at two or more locations, each transit network has the 170 choice of either sending traffic to the closest peering to the 171 adjacent transit network or passing traffic to the interconnection 172 location which advertises the least cost path to the destination 173 host. 175 The former method is called "hot potato routing" (or closest-exit) 176 because like a hot potato held in bare hands, whoever has it tries to 177 get rid of it quickly. Hot potato routing is accomplished by not 178 passing the EGBP learned MED into IBGP. This minimizes transit 179 traffic for the provider routing the traffic. Far less common is 180 "cold potato routing" (or best-exit) where the transit provider uses 181 their own transit capacity to get the traffic to the point that 182 adjacent transit provider advertised as being closest to the 183 destination. Cold potato routing is accomplished by passing the EBGP 184 learned MED into IBGP. 186 If one transit provider uses hot potato routing and another uses cold 187 potato, traffic between the two tends to be more symmetric. 188 Depending on the business relationships, if one provider has more 189 capacity or a significantly less congested backbone network, then 190 that provider may use cold potato routing. An example of widespread 191 use of cold potato routing was the NSF funded NSFNET backbone and NSF 192 funded regional networks in the mid 1990s. 194 In some cases a provider may use hot potato routing for some 195 destinations for a given peer AS and cold potato routing for others. 196 An example of this is the different treatment of commercial and 197 research traffic in the NSFNET in the mid 1990s. Today many 198 commercial networks exchange MEDs with customers but not bilateral 199 peers. However, commercial use of MEDs varies widely, from 200 ubiquitous use of MEDs to no use of MEDs at all. 202 In addition, many deployments of MEDs today are likely behaving 203 differently (e.g., resulting is sub-optimal routing) than the network 204 operator intended, thereby resulting not in hot or cold potatoes, but 205 mashed potatoes! More information on unintended behavior resulting 206 from MEDs is provided throughout this document. 208 2. Implementation and Protocol Considerations 210 There are a number of implementation and protocol peculiarities 211 relating to MEDs that have been discovered that may affect network 212 behavior. The following sections provide information on these 213 issues. 215 2.1. MULTI_EXIT_DISC is a Optional Non-Transitive Attribute 217 MULTI_EXIT_DISC is a non-transitive optional attribute whose 218 advertisement to both IBGP and EBGP peers is discretionary. As a 219 result, some implementations enable sending of MEDs to IBGP peers by 220 default, while others do not. This behavior may result in sub- 221 optimal route selection within an AS. In addition, some 222 implementations send MEDs to EBGP peers by default, while others do 223 not. This behavior may result in sub-optimal inter-domain route 224 selection. 226 2.2. MED Values and Preferences 228 Some implementations consider an MED value of zero as less preferable 229 than no MED value. This behavior resulted in path selection 230 inconsistencies within an AS. The current draft version of the BGP 231 specification [BGP4] removes ambiguities that existed in [RFC 1771] 232 by stating that if route n has no MULTI_EXIT_DISC attribute, the 233 lowest possible MULTI_EXIT_DISC value (i.e. 0) should be assigned to 234 the attribute. 236 It is apparent that different implementations and different versions 237 of the BGP draft specification have been all over the map with 238 interpretation of missing-MED. For example, earlier versions of the 239 specification called for a missing MED to be assigned the highest 240 possible MED value (i.e., 2^32-1). 242 In addition, some implementations have been shown to internally 243 employ a maximum possible MED value (2^32-1) as an "infinity" metric 244 (i.e., the MED value is used to tag routes as unfeasible), and would 245 upon on receiving an update with an MED value of 2^32-1 rewrite the 246 value to 2^32-2. Subsequently, the new MED value would be propagated 247 and could result in routing inconsistencies or unintended path 248 selections. 250 As a result of implementation inconsistencies and protocol revision 251 variances, many network operators today explicitly reset all MED 252 values on ingress to conform to their internal routing policies 253 (i.e., to include policy that requires that MED values of 0 and 254 2^32-1 NOT be used in configurations, whether the MEDs are directly 255 computed or configured), so as to not have to rely on all their 256 routers having the same missing-MED behavior. 258 2.3. Comparing MEDs Between Different Autonomous Systems 260 The MED was intended to be used on external (inter-AS) links to 261 discriminate among multiple exit or entry points to the same 262 neighboring AS. However, a large number of MED applications now 263 employ MEDs for the purpose of determining route preference between 264 like routes received from different autonomous systems. 266 A large number of implementations provide the capability to enable 267 comparison of MEDs between routes received from different neighboring 268 autonomous systems. While this capability has demonstrated some 269 benefit (e.g., that described in [RFC 3345]), operators should be 270 wary of the potential side effects with enabling such a function. 271 The deployment section below provides some examples as to why this 272 may result in undesirable behavior. 274 2.4. MEDs, Route Reflection and AS Confederations for BGP 276 In particular configurations, the BGP scaling mechanisms defined in 277 "BGP Route Reflection - An Alternative to Full Mesh IBGP" [RFC 2796] 278 and "Autonomous System Confederations for BGP" [RFC 3065] will 279 introduce persistent BGP route oscillation [RFC 3345]. The problem 280 is inherent in the way BGP works: a conflict exists between 281 information hiding/hierarchy and the non-hierarchical selection 282 process imposed by lack of total ordering caused by the MED rules. 283 Given current practices, we see the problem most frequently manifest 284 itself in the context of MED + route reflectors or confederations. 286 One potential way to avoid this is by configuring inter-Member-AS or 287 inter-cluster IGP metrics higher than intra-Member-AS IGP metrics 288 and/or using other tie breaking policies to avoid BGP route selection 289 based on incomparable MEDs. Of course, IGP metric constraints may be 290 unreasonably onerous for some applications. 292 Comparing MEDs between differing adjacent autonomous systems (which 293 will be discussed in later sections), or not utilizing MEDs at all, 294 significantly decreases the probability of introducing potential 295 route oscillation conditions into the network. 297 Although perhaps "legal" as far as current specifications are 298 concerned, modifying MED attributes received on any type of IBGP 299 session (e.g., standard IBGP, AS confederations EIBGP, route 300 reflection, etc..) is NOT recommended. 302 2.5. Route Flap Damping and MED Churn 304 MEDs are often derived dynamically from IGP metrics or additive costs 305 associated with an IGP metric to a given BGP NEXT_HOP. This 306 typically provides an efficient model for ensuring that the BGP MED 307 advertised to peers used to represent the best path to a given 308 destination within the network is aligned with that of the IGP within 309 a given AS. 311 The consequence with dynamically derived IGP-based MEDs is that 312 instability within an AS, or even on a single given link within the 313 AS, can result in wide-spread BGP instability or BGP route 314 advertisement churn that propagates across multiple domains. In 315 short, if your MED "flaps" every time your IGP metric flaps, you're 316 routes are likely going to be suppressed as a result of BGP Route 317 Flap Damping [RFC 2439]. 319 Employment of MEDs may compound the adverse effects of BGP flap 320 dampening behavior because it many cause routes to be re- advertised 321 solely to reflect an internal topology change. 323 Many implementations don't have a practical problem with IGP 324 flapping, they either latch their IGP metric upon first advertisement 325 or they employ some internal suppression mechanism. Some 326 implementations regard BGP attribute changes as less significant than 327 route withdrawals and announcements to attempt to mitigate the impact 328 of this type of event. 330 2.6. Effects of MEDs on Update Packing Efficiency 332 Multiple unfeasible routes can be advertised in a single BGP Update 333 message. In addition, one or more feasible routes can be advertised 334 in a single Update message so long as all prefixes share a common 335 attribute set. 337 The BGP4 protocol permits advertisement of multiple prefixes with a 338 common set of path attributes to be advertised in a single update 339 message, this is commonly referred to as "update packing". When 340 possible, update packing is recommended as it provides a mechanism 341 for more efficient behavior in a number of areas, to include: 343 o Reduction in system overhead due to generation or receipt of 344 fewer Update messages. 346 o Reduction in network overhead as a result of fewer packets and 347 lower bandwidth consumption. 349 o Allows processing of path attributes and searches for matching 350 sets in your AS_PATH database (if you have one) less frequently. 351 Consistent ordering of the path attributes allows for ease of 352 matching in the database as you don't have different 353 representations 354 of the same data. 356 Update packing requires that all feasible routes within a single 357 update message share a common attribute set, to include a common 358 MULTI_EXIT_DISC value. As such, potential wide-scale variance in MED 359 values introduces another variable and may resulted in a marked 360 decrease in update packing efficiency. 362 2.7. Temporal Route Selection 364 Some implementations have had bugs which lead to temporal behavior in 365 MED-based best path selection. These usually involved methods used 366 to store the oldest route along with ordering routes for MED in 367 earlier implementations that cause non-deterministic behavior on 368 whether the oldest route would truly be selected or not. 370 The reasoning for this is that "older" paths are presumably more 371 stable, and thus more preferable. However, temporal behavior in 372 route selection results in non-deterministic behavior, and as such, 373 is often undesirable. 375 3. Deployment Considerations 377 Empirical data [MFN/Ixia Monitoring Project] has shown that accepting 378 MEDs from other autonomous systems have the potential to cause 379 traffic flow churns in the network. Some implementations only 380 ratchet down the MED and never move it back up to prevent excessive 381 churn. 383 However, if that session is reset, the MEDs being advertised have the 384 potential of changing. If an network is relying on received MEDs to 385 route traffic properly, the traffic patterns have the potential for 386 changing dramatically, potentially resulting in congestion on the 387 network. Essentially, accepting and routing traffic based on MEDs 388 allows other people to traffic engineer your network. This may or may 389 not be acceptable to you. 391 As previously discussed, many network operators choose to reset MED 392 values on ingress. In addition, many operators explicitly do not 393 employ MED values of 0 or 2^32-1 in order to avoid inconsistencies 394 with implementations and various revisions of the BGP specification. 396 3.1. Comparing MEDs Between Different Autonomous Systems 398 Although the MED was meant to only be used when comparing paths 399 received from different external peers in the same AS, many 400 implementations provide the capability to compare MEDs between 401 different autonomous systems as well. 403 Though this may seem a fine idea for some configurations, care must 404 be taken when comparing MEDs between different autonomous systems. 405 BGP speakers often derive MED values by obtaining the IGP metric 406 associated with reaching a given BGP NEXT_HOP within the local AS. 407 This allows MEDs to reasonably reflect IGP topologies when 408 advertising routes to peers. While this is fine when comparing MEDs 409 between multiple paths learned from a single AS, it can result in 410 potentially "weighted" decisions when comparing MEDs between 411 different autonomous systems. This is most typically the case when 412 the autonomous systems use different mechanisms to derive IGP 413 metrics, BGP MEDs, or perhaps even use different IGP protocols with 414 vastly contrasting metric spaces (e.g., OSPF v. traditional metric 415 space in IS-IS). 417 3.2. Effects of Aggregation on MEDs` 419 Another MED deployment consideration involves the impact that 420 aggregation of BGP routing information has on MEDs. Aggregates are 421 often generated from multiple locations in an AS in order to 422 accommodate stability, redundancy and other network design goals. 423 When MEDs are derived from IGP metrics associated with said 424 aggregates the MED value advertised to peers can result in very 425 suboptimal routing. 427 4. Security Considerations 429 The MED was purposely designed to be a "weak" metric that would only 430 be used late in the best-path decision process. The BGP working 431 group was concerned that any metric specified by a remote operator 432 would only affect routing in a local AS IF no other preference was 433 specified. A paramount goal of the design of the MED was to ensure 434 that peers could not "shed" or "absorb" traffic for networks that 435 they advertise. As such, accepting MEDs from peers may in some sense 436 increase a network's susceptibility to exploitation by peers. 438 4.1. Acknowledgments 440 Thanks to John Scudder for applying his usual keen eye and 441 constructive insight. Also, thanks to Curtis Villamizar and JR 442 Mitchell. 444 Others to be supplied. 446 5. References 448 [RFC 1519] Fuller, V., Li. T., Yu J., and K. Varadhan, "Classless 449 Inter-Domain Routing (CIDR): an Address Assignment and 450 Aggregation Strategy", RFC 1519, September 1993. 452 [RFC 1771] Rekhter, Y., and T. Li, "A Border Gateway Protocol 4 453 (BGP-4)", RFC 1771, March 1995. 455 [RFC 2439] Villamizar, C. and Chandra, R., "BGP Route Flap Damping", 456 RFC 2439, November 1998. 458 [RFC 2796] Bates, T., Chandra, R., Chen, E., "BGP Route Reflection 459 - An Alternative to Full Mesh IBGP", RFC 2796, April 460 2000. 462 [RFC 3065] Traina, P., McPherson, D., Scudder, J.. "Autonomous System 463 Confederations for BGP", RFC 3065, February 2001. 465 [RFC 3345] McPherson, D., Gill, V., Walton, D., and Retana, A, "BGP 466 Persistent Route Oscillation Condition", RFC 3345, 467 August 2002. 469 [BGP4] Rekhter, Y., T. Li., and Hares. S, Editors, "A Border 470 Gateway Protocol 4 (BGP-4)", BGP Draft, Work in Progress. 472 [MFN/Ixia Monitoring Project] Vijay to Provide Pointer. 474 6. Authors' Addresses 476 Danny McPherson 477 Arbor Networks 478 Email: danny@arbor.net 480 Vijay Gill 481 AOL 482 Email: VijayGill9@aol.com 484 7. Full Copyright Statement 486 Copyright (C) The Internet Society (2004). All Rights Reserved. 488 This document and translations of it may be copied and furnished to 489 others, and derivative works that comment on or otherwise explain it 490 or assist in its implementation may be prepared, copied, published 491 and distributed, in whole or in part, without restriction of any 492 kind, provided that the above copyright notice and this paragraph are 493 included on all such copies and derivative works. However, this 494 document itself may not be modified in any way, such as by removing 495 the copyright notice or references to the Internet Society or other 496 Internet organizations, except as needed for the purpose of 497 developing Internet standards in which case the procedures for 498 copyrights defined in the Internet Standards process must be 499 followed, or as required to translate it into languages other than 500 English. 502 The limited permissions granted above are perpetual and will not be 503 revoked by the Internet Society or its successors or assigns. 505 This document and the information contained herein is provided on an 506 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING 507 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING 508 BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION 509 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF 510 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.