idnits 2.17.1 draft-ietf-idr-aigp-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- 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 (March 29, 2011) is 4771 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-03) exists of draft-pmohapat-idr-fast-conn-restore-01 == Outdated reference: A later version (-05) exists of draft-ietf-idr-best-external-03 Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 1 comment (--).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 Rex Fernando 4 Intended Status: Proposed Standard Eric C. Rosen 5 Expires: September 29, 2011 Cisco Systems, Inc. 7 James Uttaro 8 ATT 10 March 29, 2011 12 The Accumulated IGP Metric Attribute for BGP 14 draft-ietf-idr-aigp-05.txt 16 Abstract 18 Routing protocols that have been designed to run within a single 19 administrative domain ("IGPs") generally do so by assigning a metric 20 to each link, and then choosing as the installed path between two 21 nodes the path for which the total distance (sum of the metric of 22 each link along the path) is minimized. BGP, designed to provide 23 routing over a large number of independent administrative domains 24 ("autonomous systems"), does not make its path selection decisions 25 through the use of a metric. It is generally recognized that any 26 attempt to do so would incur significant scalability problems, as 27 well as inter-administration coordination problems. However, there 28 are deployments in which a single administration runs several 29 contiguous BGP networks. In such cases, it can be desirable, within 30 that single administrative domain, for BGP to select paths based on a 31 metric, just as an IGP would do. The purpose of this document is to 32 provide a specification for doing so. 34 Status of this Memo 36 This Internet-Draft is submitted to IETF in full conformance with the 37 provisions of BCP 78 and BCP 79. 39 Internet-Drafts are working documents of the Internet Engineering 40 Task Force (IETF), its areas, and its working groups. Note that 41 other groups may also distribute working documents as Internet- 42 Drafts. 44 Internet-Drafts are draft documents valid for a maximum of six months 45 and may be updated, replaced, or obsoleted by other documents at any 46 time. It is inappropriate to use Internet-Drafts as reference 47 material or to cite them other than as "work in progress." 49 The list of current Internet-Drafts can be accessed at 50 http://www.ietf.org/ietf/1id-abstracts.txt. 52 The list of Internet-Draft Shadow Directories can be accessed at 53 http://www.ietf.org/shadow.html. 55 Copyright and License Notice 57 Copyright (c) 2011 IETF Trust and the persons identified as the 58 document authors. All rights reserved. 60 This document is subject to BCP 78 and the IETF Trust's Legal 61 Provisions Relating to IETF Documents 62 (http://trustee.ietf.org/license-info) in effect on the date of 63 publication of this document. Please review these documents 64 carefully, as they describe your rights and restrictions with respect 65 to this document. Code Components extracted from this document must 66 include Simplified BSD License text as described in Section 4.e of 67 the Trust Legal Provisions and are provided without warranty as 68 described in the Simplified BSD License. 70 Table of Contents 72 1 Specification of requirements ......................... 3 73 2 Introduction .......................................... 3 74 3 AIGP Attribute ........................................ 5 75 3.1 Applicability Restrictions and Cautions ............... 6 76 3.2 Restrictions on Sending/Receiving ..................... 6 77 3.3 Creating and Modifying the AIGP Attribute ............. 7 78 3.3.1 Originating the AIGP Attribute ........................ 7 79 3.3.2 Modifications by the Originator ....................... 8 80 3.3.3 Modifications by a Non-Originator ..................... 8 81 4 Decision Process ...................................... 10 82 4.1 When a Route has an AIGP Attribute .................... 10 83 4.2 When the Route to the Next Hop has an AIGP attribute .. 11 84 5 Deployment Considerations ............................. 12 85 6 IANA Considerations ................................... 12 86 7 Security Considerations ............................... 12 87 8 Acknowledgments ....................................... 12 88 9 Authors' Addresses .................................... 13 89 10 Normative References .................................. 13 90 11 Informative References ................................ 14 92 1. Specification of requirements 94 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 95 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 96 document are to be interpreted as described in [RFC2119]. 98 2. Introduction 100 There are many routing protocols that have been designed to run 101 within a single administrative domain. These are known collectively 102 as "Interior Gateway Protocols" (IGPs). Typically, each link is 103 assigned a particular "metric" value. The path between two nodes can 104 then be assigned a "distance", which is the sum of the metrics of all 105 the links that belong to that path. An IGP selects the "shortest" 106 (minimal distance) path between any two nodes, perhaps subject to the 107 constraint that if the IGP provides multiple "areas", it may prefer 108 the shortest path within an area to a path that traverses more than 109 one area. Typically the administration of the network has some 110 routing policy which can be approximated by selecting shortest paths 111 in this way. 113 BGP, as distinguished from the IGPs, was designed to run over an 114 arbitrarily large number of administrative domains ("autonomous 115 systems", or "ASes") with limited coordination among the various 116 administrations. BGP does not make its path selection decisions 117 based on a metric; there is no such thing as an "inter-AS metric". 118 There are two fundamental reasons for this: 120 - The distance between two nodes in a common administrative domain 121 may change at any time due to events occurring in that domain. 122 These changes are not propagated around the Internet unless they 123 actually cause the border routers of the domain to select routes 124 with different BGP attributes for some set of address prefixes. 125 This accords with a fundamental principle of scaling, viz., that 126 changes with only local significance must not have global 127 effects. If local changes in distance were always propagated 128 around the Internet, this principle would be violated. 130 - A basic principle of inter-domain routing is that the different 131 administrative domains may have their own policies, which do not 132 have to be revealed to other domains, and which certainly do not 133 have to be agreed to by other domains. Yet the use of inter-AS 134 metric in the Internet would have exactly these effects. 136 There are, however, deployments in which a single administration runs 137 a network which has been sub-divided into multiple, contiguous ASes, 138 each running BGP. There are several reasons why a single 139 administrative domain may be broken into several ASes (which, in this 140 case, are not really "autonomous".) It may be that the existing IGPs 141 do not scale well in the particular environment; it may be that a 142 more generalized topology is desired than could be obtained by use of 143 a single IGP domain; it may be that a more finely grained routing 144 policy is desired than can be supported by an IGP. In such 145 deployments, it can be useful to allow BGP to make its routing 146 decisions based on the IGP metric, so that BGP chooses the "shortest" 147 path between two nodes, even if the nodes are in two different ASes 148 within that same administrative domain. We will refer to the set of 149 ASes in a common administrative domain as an "AIGP Administrative 150 Domain". 152 There are in fact some implementations that already do something like 153 this, using BGP's MULTI_EXIT_DISC (MED) attribute to carry a value 154 based on IGP metrics. However, that doesn't really provide IGP-like 155 "shortest path" routing, as the BGP decision process gives priority 156 to other factors, such as the AS_PATH length. Also, the standard 157 procedures for use of the MED do not ensure that the IGP metric is 158 properly accumulated so that it covers all the links along the path. 160 In this document, we define a new optional, non-transitive BGP 161 attribute, called the "Accumulated IGP Metric Attribute", or "AIGP 162 attribute", and specify the procedures for using it. 164 The specified procedures prevent the AIGP attribute from "leaking 165 out" past an AIGP administrative domain boundary into the Internet. 167 The specified procedures also ensure that the value in the AIGP 168 attribute has been accumulated all along the path from the 169 destination, i.e., that the AIGP attribute does not appear when there 170 are "gaps" along the path where the IGP metric is unknown. 172 3. AIGP Attribute 174 The AIGP Attribute is an optional non-transitive BGP Path Attribute. 175 The attribute type code for the AIGP Attribute is 26. The value 176 field of the AIGP Attribute is defined here to be a set of TLVs 177 (elements encoded as "Type/Length/Value"). However, this document 178 defines only a single such TLV, the AIGP TLV, that contains the 179 Accumulated IGP Metric. The AIGP TLV is encoded as shown in Figure 180 1. An AIGP Attribute MUST NOT contain more than one AIGP TLV. 182 0 1 2 3 183 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 1 184 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 185 | Type=1 | Length | | 186 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 187 ~ ~ 188 | Accumulated IGP Metric | 189 | +-+-+-+-+-+-+-+-+ 190 | | 191 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 193 AIGP Attribute 194 Figure 1 196 - Type: A single octet encoding the AIGP Attribute Type. Only type 197 1 is defined in this document. 199 - Length: Two octets encoding the length in octets of the attribute, 200 including the type and length fields. The length is encoded as an 201 unsigned binary integer. 203 The length of the AIGP TLV is always 11. 205 - Accumulated IGP Metric: For a type 1 AIGP attribute, the value 206 field is always 8 bytes long. IGP metrics are frequently 207 expressed as 4-octet values, and this ensures that the AIGP 208 attribute can be used to hold the sum of an arbitrary number of 209 4-octet values. 211 3.1. Applicability Restrictions and Cautions 213 This document only considers the use of the AIGP attribute in 214 networks where each router uses tunneling of some sort to deliver a 215 packet to its BGP next hop. Use of the AIGP attribute in networks 216 that do not use tunneling is outside the scope of this document. 218 If a Route Reflector supports the AIGP attribute, but some of its 219 clients do not, then the routing choices that result may not all 220 reflect the intended routing policy. 222 3.2. Restrictions on Sending/Receiving 224 An implementation that supports the AIGP attribute MUST support a 225 per-session configuration item, AIGP_SESSION, that indicates whether 226 the attribute is enabled or disabled for use on that session. 228 - The default value of AIGP_SESSION, for EBGP sessions, MUST be 229 "disabled". 231 - The default value of AIGP_SESSION, for IBGP and confederation- 232 EBGP sessions, MUST be "enabled." 234 The AIGP attribute MUST NOT be sent on any BGP session for which 235 AIGP_SESSION is disabled. 237 If an AIGP attribute is received on a BGP session for which 238 AIGP_SESSION is disabled, the attribute MUST be treated exactly as if 239 it were an unrecognized non-transitive attribute. That is, "it MUST 240 be quietly ignored and not passed along to other BGP peers" (see 241 [BGP], section 5). 243 3.3. Creating and Modifying the AIGP Attribute 245 3.3.1. Originating the AIGP Attribute 247 An implementation that supports the AIGP attribute MUST support a 248 configuration item, AIGP_ORIGINATE, that enables or disables its 249 creation and attachment to routes. The default value of 250 AIGP_ORIGINATE MUST be "disabled". 252 A BGP speaker MUST NOT add the AIGP attribute to any route whose path 253 leads outside the "AIGP administrative domain" to which the BGP 254 speaker belongs. It may be added only to routes that satisfy one of 255 the following conditions: 257 - The route is a static route that is being redistributed into BGP 259 - The route is an IGP route that is being redistributed into BGP 261 - The route is an IBGP-learned route whose AS_PATH attribute is 262 empty. 264 - The route is an EBGP-learned route whose AS_PATH contains only 265 ASes that are in the same AIGP Administrative Domain as the BGP 266 speaker. 268 A BGP speaker MUST NOT add the AIGP attribute to any route for which 269 it has not set itself as the next hop. 271 It SHOULD be possible to set AIGP_ORIGINATE to "enabled for the 272 routes of a particular IGP that are redistributed into BGP" (where "a 273 particular IGP" might be "OSPF" or "ISIS"). Other policies 274 determining when and whether to originate an AIGP attribute are also 275 possible, depending on the needs of a particular deployment scenario. 277 When originating an AIGP attribute for a BGP route to address prefix 278 P, the value of the attribute is set according to policy. There are 279 a number of useful policies, some of which are in the following list: 281 - When a BGP speaker is redistributing into BGP an IGP route to 282 address prefix P, the IGP will have computed a "distance" from R 283 to P. This distance MAY be assigned as the value of AIGP 284 attribute. 286 - A BGP speaker may be redistributing into BGP a static route to 287 address prefix P, for which a "distance" from R to P has been 288 configured. This distance MAY be assigned as the value of AIGP 289 attribute. 291 - A BGP speaker R may have received and installed a BGP-learned 292 route to prefix P, with next hop N. Or it may be redistributing 293 a static route to P, with next hop N. The "distance" from R to N 294 MAY be assigned as the value of the AIGP attribute of the route 295 to P. 297 * If R has an IGP route to N, the IGP-computed distance from R 298 to N MAY be used. 300 * If R has a BGP route to N, and an AIGP attribute value has 301 been computed for that route (see section 3.3.3), that value 302 MAY be used as the AIGP attribute value of the route to P. 304 3.3.2. Modifications by the Originator 306 If BGP speaker R is the originator of the AIGP attribute of prefix P, 307 and at some point the "distance" from R to P changes, R SHOULD issue 308 a new BGP update containing the new value of the AIGP attribute. 309 (Here we use the term "distance" to refer to whatever value the 310 originator assigns to the AIGP attribute, however it is computed; see 311 section 3.3.1.) However, if the difference between the new distance 312 and the distance advertised in the AIGP attribute is less than a 313 configurable threshold, the update MAY be suppressed. 315 3.3.3. Modifications by a Non-Originator 317 Suppose a BGP speaker R1 receives a route with an AIGP attribute 318 whose value is A, and a Next Hop whose value is R2. Suppose also 319 that R1 is about to redistribute that route on a BGP session that is 320 enabled for sending/receiving the attribute. 322 If R1 does not change the Next Hop of the route, then R1 MUST NOT 323 change the AIGP attribute value of the route. 325 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 326 route to R2 is an IGP-learned route, or a static route that does not 327 require recursive next hop resolution, then R1 must increase the 328 value of the AIGP attribute by adding to A the distance from R1 to 329 R2. This distance is either the IGP-computed distance from R1 to R2, 330 or some value determined by policy. However, A MUST be increased by 331 a non-zero amount. 333 Note that if R1 and R2 above are EBGP neighbors, and there is a 334 direct link between them on which no IGP is running, then when R1 335 changes the next hop of a route from R2 to R1, the AIGP metric value 336 MUST be increased by a non-zero amount. The amount of the increase 337 SHOULD be such that it is properly comparable to the IGP metrics. 338 E.g., if the IGP metric is a function of latency, then the amount of 339 the increase should be a function of the latency from R1 to R2. 341 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 342 route to R2 is a BGP-learned route, or a static route that requires 343 recursive next hop resolution, then the AIGP attribute value needs to 344 be increased in several steps, according to the following procedure. 345 (Note that this procedure is ONLY used when recursive next hop 346 resolution is needed.) 348 1. Let Xattr be the new AIGP attribute value. 350 2. Initialize Xattr to A. 352 3. Set the XNH to R2. 354 4. Find the route to XNH. 356 5. If the route to XNH does not require recursive next hop 357 resolution, get the distance D from R1 to XNH. (Note that this 358 condition cannot be satisfied the first time through this 359 procedure.) If D is above a configurable threshold, set the 360 AIGP attribute value to Xattr+D. If D is below a configurable 361 threshold, set the AIGP attribute value to Xattr. In either 362 case, exit this procedure. 364 6. If the route to XNH is a BGP-learned route, and the route does 365 NOT have an AIGP attribute, then exit this procedure and do not 366 pass on any AIGP attribute. 368 7. If the route to XNH is a BGP-learned route, and the route has 369 an AIGP attribute value of Y, then set Xattr=Xattr+Y, and set 370 XNH to the next hop of this route. (The intention here is that 371 Y is the AIGP value of the route as it was received by R1, 372 without having been modified by R1.) 374 8. Go to step 4. 376 The AIGP value of a given route depends on (a) the AIGP values of all 377 the next hops that are recursively resolved during this procedure, 378 and (b) the IGP distance to any next hop that is not recursively 379 resolved. Any change due to (a) in any of these values MUST trigger 380 a new AIGP computation for that route. Whether a change due to (b) 381 triggers a new AIGP computation depends upon whether the change in 382 IGP distance exceeds a configurable threshold. 384 If the AIGP attribute is carried across several ASes, each with its 385 own IGP domain, it is clear that these procedures are unlikely to 386 give a sensible result if the IGPs are different (e.g., some OSPF and 387 some IS-IS), or if the meaning of the metrics is different in the 388 different IGPs (e.g., if the metric represents bandwidth in some IGP 389 domains but represents latency in others). These procedures also are 390 unlikely to give a sensible result if the metric assigned to inter-AS 391 BGP links (on which no IGP is running) or to static routes is not 392 comparable to the IGP metrics. All such cases are outside the scope 393 of the current document. 395 4. Decision Process 397 Support for the AIGP attribute involves several modifications to the 398 tie breaking procedures of the BGP "phase 2" decision described in 399 [BGP], section 9.1.2.2. These modifications are described below in 400 sections 4.1 and 4.2. 402 In some cases, the BGP decision process may install a route without 403 executing any tie breaking procedures. This may happen, e.g., if 404 only one route to a given prefix has the highest degree of preference 405 (as defined in [BGP] section 9.1.1). In this case, the AIGP 406 attribute is not considered. 408 In other cases, some routes may be eliminated before the tie breaking 409 procedures are invoked, e.g., routes with AS-PATH attributes 410 indicating a loop, or routes with unresolvable next hops. In these 411 cases, the AIGP attributes of the eliminated routes are not 412 considered. 414 4.1. When a Route has an AIGP Attribute 416 Assuming that the BGP decision process invokes the tie breaking 417 procedures, the procedures in this section MUST be executed BEFORE 418 any of the tie breaking procedures described in [BGP] section 9.1.2.2 419 are executed. 421 If any routes have an AIGP attribute, remove from consideration all 422 routes that do not have an AIGP attribute. 424 If router R is considering route T, where T has an AIGP attribute, 425 - then R must compute the value A, defined as follows: set A to the 426 sum of (a) T's AIGP attribute value and (b) the IGP distance from 427 R to T's next hop. 429 - remove from consideration all routes that are not tied for the 430 lowest value of A. 432 4.2. When the Route to the Next Hop has an AIGP attribute 434 Suppose that a given router R1 is comparing two BGP-learned routes, 435 such that either: 437 - the two routes have equal AIGP attribute values, or else 439 - neither of the two routes has an AIGP attribute. The BGP 440 decision process as specified in [BGP] makes use, in its tie 441 breaker procedures, of "interior cost", defined as follows: 443 "interior cost of a route is determined by calculating the 444 metric to the NEXT_HOP for the route using the Routing 445 Table." 447 Suppose route T has a next hop of N. We modify the notion of the 448 "interior cost" from node R1 to node N as follows: 450 - Let R2 be the next hop of the route to N, after all recursive 451 resolution of the next hop is done. Let m be the IGP distance 452 (or in the case of a static route, the configured distance) from 453 R1 to R2. 455 - If the installed route to N has an AIGP attribute, set A to the 456 AIGP value of the route to N, computing the AIGP value of the 457 route according to the procedure of section 3.3.3. 459 - If the installed route to N does not have an AIGP value, set A to 460 0. 462 - The "interior cost" of route T is the quantity A+m. 464 5. Deployment Considerations 466 Using the AIGP attribute to achieve a desired routing policy will be 467 more effective if each BGP speaker can use it to choose from among 468 multiple routes. Thus is it highly recommended that the procedures of 469 [BESTEXT] and [ADDPATH] be used in conjunction with the AIGP 470 Attribute. 472 If a Route Reflector does not pass all paths to its clients, then it 473 will tend to pass the paths for which the IGP distance from the Route 474 Reflector itself to the next hop is smallest. This may result in a 475 non-optimal choice by the clients. 477 6. IANA Considerations 479 IANA has assigned the codepoint 26 in the "BGP Path Attributes" 480 registry to the AIGP attribute. 482 IANA shall create a registry for "BGP AIGP Attribute Types". The 483 type field consists of a single octet, with possible values from 0 to 484 255. The allocation policy for this field is to be "Standards Action 485 with Early Allocation". Type 1 should be defined as "AIGP", and 486 should refer to this document. 488 7. Security Considerations 490 The spurious introduction, though error or malfeasance, of an AIGP 491 attribute, could result in the selection of paths other than those 492 desired. 494 Improper configuration on both ends of an EBGP connection could 495 result in an AIGP attribute being passed from one service provider to 496 another. This would likely result in an unsound selection of paths. 498 8. Acknowledgments 500 The authors would like to thank Rajiv Asati, Clarence Filsfils, 501 Robert Raszuk, Yakov Rekhter, Samir Saad, and John Scudder for their 502 input. 504 9. Authors' Addresses 506 Rex Fernando 507 Cisco Systems, Inc. 508 170 Tasman Drive 509 San Jose, CA 95134 510 Email: rex@cisco.com 512 Pradosh Mohapatra 513 Cisco Systems, Inc. 514 170 Tasman Drive 515 San Jose, CA 95134 516 Email: pmohapat@cisco.com 518 Eric C. Rosen 519 Cisco Systems, Inc. 520 1414 Massachusetts Avenue 521 Boxborough, MA, 01719 522 Email: erosen@cisco.com 524 James Uttaro 525 AT&T 526 200 S. Laurel Avenue 527 Middletown, NJ 07748 528 Email: uttaro@att.com 530 10. Normative References 532 [BGP], "A Border Gateway Protocol 4 (BGP-4)", Y. Rekhter, T. Li, S. 533 Hares, RFC 4271, January 2006. 535 11. Informative References 537 [ADDPATH] "Fast Connectivity Restoration Using BGP Add-Path", P. 538 Mohapatra, R. Fernando, C. Filsfils, R. Raszuk, draft-pmohapat-idr- 539 fast-conn-restore-01.txt, March 2011. 541 [BESTEXT], "Advertisement of the Best External Route in BGP", P. 542 Marques, R. Fernando, E. Chen, P. Mohapatra, draft-ietf-idr-best- 543 external-03.txt, March 2011. 545 [RFC2119] "Key words for use in RFCs to Indicate Requirement 546 Levels.", S. Bradner, March 1997