idnits 2.17.1 draft-ietf-idr-aigp-10.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 (May 23, 2013) is 3990 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 (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Pradosh Mohapatra 3 Internet Draft Cumulus Networks 4 Intended Status: Proposed Standard 5 Expires: November 23, 2013 Rex Fernando 6 Eric C. Rosen 7 Cisco Systems, Inc. 9 James Uttaro 10 ATT 12 May 23, 2013 14 The Accumulated IGP Metric Attribute for BGP 16 draft-ietf-idr-aigp-10.txt 18 Abstract 20 Routing protocols that have been designed to run within a single 21 administrative domain ("IGPs") generally do so by assigning a metric 22 to each link, and then choosing as the installed path between two 23 nodes the path for which the total distance (sum of the metric of 24 each link along the path) is minimized. BGP, designed to provide 25 routing over a large number of independent administrative domains 26 ("autonomous systems"), does not make its path selection decisions 27 through the use of a metric. It is generally recognized that any 28 attempt to do so would incur significant scalability problems, as 29 well as inter-administration coordination problems. However, there 30 are deployments in which a single administration runs several 31 contiguous BGP networks. In such cases, it can be desirable, within 32 that single administrative domain, for BGP to select paths based on a 33 metric, just as an IGP would do. The purpose of this document is to 34 provide a specification for doing so. 36 Status of this Memo 38 This Internet-Draft is submitted to IETF in full conformance with the 39 provisions of BCP 78 and BCP 79. 41 Internet-Drafts are working documents of the Internet Engineering 42 Task Force (IETF), its areas, and its working groups. Note that 43 other groups may also distribute working documents as Internet- 44 Drafts. 46 Internet-Drafts are draft documents valid for a maximum of six months 47 and may be updated, replaced, or obsoleted by other documents at any 48 time. It is inappropriate to use Internet-Drafts as reference 49 material or to cite them other than as "work in progress." 51 The list of current Internet-Drafts can be accessed at 52 http://www.ietf.org/ietf/1id-abstracts.txt. 54 The list of Internet-Draft Shadow Directories can be accessed at 55 http://www.ietf.org/shadow.html. 57 Copyright and License Notice 59 Copyright (c) 2013 IETF Trust and the persons identified as the 60 document authors. All rights reserved. 62 This document is subject to BCP 78 and the IETF Trust's Legal 63 Provisions Relating to IETF Documents 64 (http://trustee.ietf.org/license-info) in effect on the date of 65 publication of this document. Please review these documents 66 carefully, as they describe your rights and restrictions with respect 67 to this document. Code Components extracted from this document must 68 include Simplified BSD License text as described in Section 4.e of 69 the Trust Legal Provisions and are provided without warranty as 70 described in the Simplified BSD License. 72 Table of Contents 74 1 Specification of requirements ......................... 3 75 2 Introduction .......................................... 3 76 3 AIGP Attribute ........................................ 5 77 3.1 Applicability Restrictions and Cautions ............... 6 78 3.2 Restrictions on Sending/Receiving ..................... 6 79 3.3 Creating and Modifying the AIGP Attribute ............. 7 80 3.3.1 Originating the AIGP Attribute ........................ 7 81 3.3.2 Modifications by the Originator ....................... 8 82 3.3.3 Modifications by a Non-Originator ..................... 8 83 4 Decision Process ...................................... 10 84 4.1 When a Route has an AIGP Attribute .................... 10 85 4.2 When the Route to the Next Hop has an AIGP attribute .. 11 86 5 Deployment Considerations ............................. 12 87 6 IANA Considerations ................................... 12 88 7 Security Considerations ............................... 12 89 8 Acknowledgments ....................................... 12 90 9 Authors' Addresses .................................... 13 91 10 Normative References .................................. 13 92 11 Informative References ................................ 13 94 1. Specification of requirements 96 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 97 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 98 document are to be interpreted as described in [RFC2119]. 100 2. Introduction 102 There are many routing protocols that have been designed to run 103 within a single administrative domain. These are known collectively 104 as "Interior Gateway Protocols" (IGPs). Typically, each link is 105 assigned a particular "metric" value. The path between two nodes can 106 then be assigned a "distance", which is the sum of the metrics of all 107 the links that belong to that path. An IGP selects the "shortest" 108 (minimal distance) path between any two nodes, perhaps subject to the 109 constraint that if the IGP provides multiple "areas", it may prefer 110 the shortest path within an area to a path that traverses more than 111 one area. Typically the administration of the network has some 112 routing policy which can be approximated by selecting shortest paths 113 in this way. 115 BGP, as distinguished from the IGPs, was designed to run over an 116 arbitrarily large number of administrative domains ("autonomous 117 systems", or "ASes") with limited coordination among the various 118 administrations. BGP does not make its path selection decisions 119 based on a metric; there is no such thing as an "inter-AS metric". 120 There are two fundamental reasons for this: 122 - The distance between two nodes in a common administrative domain 123 may change at any time due to events occurring in that domain. 124 These changes are not propagated around the Internet unless they 125 actually cause the border routers of the domain to select routes 126 with different BGP attributes for some set of address prefixes. 127 This accords with a fundamental principle of scaling, viz., that 128 changes with only local significance must not have global 129 effects. If local changes in distance were always propagated 130 around the Internet, this principle would be violated. 132 - A basic principle of inter-domain routing is that the different 133 administrative domains may have their own policies, which do not 134 have to be revealed to other domains, and which certainly do not 135 have to be agreed to by other domains. Yet the use of inter-AS 136 metric in the Internet would have exactly these effects. 138 There are, however, deployments in which a single administration runs 139 a network which has been sub-divided into multiple, contiguous ASes, 140 each running BGP. There are several reasons why a single 141 administrative domain may be broken into several ASes (which, in this 142 case, are not really "autonomous".) It may be that the existing IGPs 143 do not scale well in the particular environment; it may be that a 144 more generalized topology is desired than could be obtained by use of 145 a single IGP domain; it may be that a more finely grained routing 146 policy is desired than can be supported by an IGP. In such 147 deployments, it can be useful to allow BGP to make its routing 148 decisions based on the IGP metric, so that BGP chooses the "shortest" 149 path between two nodes, even if the nodes are in two different ASes 150 within that same administrative domain. We will refer to the set of 151 ASes in a common administrative domain as an "AIGP Administrative 152 Domain". 154 There are in fact some implementations that already do something like 155 this, using BGP's MULTI_EXIT_DISC (MED) attribute to carry a value 156 based on IGP metrics. However, that doesn't really provide IGP-like 157 "shortest path" routing, as the BGP decision process gives priority 158 to other factors, such as the AS_PATH length. Also, the standard 159 procedures for use of the MED do not ensure that the IGP metric is 160 properly accumulated so that it covers all the links along the path. 162 In this document, we define a new optional, non-transitive BGP 163 attribute, called the "Accumulated IGP Metric Attribute", or "AIGP 164 attribute", and specify the procedures for using it. 166 The specified procedures prevent the AIGP attribute from "leaking 167 out" past an AIGP administrative domain boundary into the Internet. 169 The specified procedures also ensure that the value in the AIGP 170 attribute has been accumulated all along the path from the 171 destination, i.e., that the AIGP attribute does not appear when there 172 are "gaps" along the path where the IGP metric is unknown. 174 3. AIGP Attribute 176 The AIGP Attribute is an optional non-transitive BGP Path Attribute. 177 The attribute type code for the AIGP Attribute is 26. 179 The value field of the AIGP Attribute is defined here to be a set of 180 elements encoded as "Type/Length/Value" (i.e., a set of "TLVs"). 181 Each such TLV is encoded as shown in Figure 1. 183 0 1 2 3 184 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 1 185 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 186 | Type | Length | | 187 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 188 ~ ~ 189 | Value | 190 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+.......................... 192 AIGP TLV 193 Figure 1 195 - Type: A single octet encoding the TLV Type. Only type 1, "AIGP 196 TLV", is defined in this document. 198 - Length: Two octets encoding the length in octets of the TLV, 199 including the type and length fields. The length is encoded as an 200 unsigned binary integer. (Note that the minimum length is 3, 201 indicating that no value field is present.) 203 - A value field containing zero or more octets. 205 This document defines only a single such TLV, the "AIGP TLV". The 206 AIGP TLV is encoded as follows: 208 - Type: 1 210 - Length: 11 212 - Accumulated IGP Metric. 214 The value field of the AIGP TLV is always 8 bytes long. IGP 215 metrics are frequently expressed as 4-octet values, and this 216 ensures that the AIGP attribute can be used to hold the sum of an 217 arbitrary number of 4-octet values. 219 3.1. Applicability Restrictions and Cautions 221 This document only considers the use of the AIGP attribute in 222 networks where each router uses tunneling of some sort to deliver a 223 packet to its BGP next hop. Use of the AIGP attribute in networks 224 that do not use tunneling is outside the scope of this document. 226 If a Route Reflector supports the AIGP attribute, but some of its 227 clients do not, then the routing choices that result may not all 228 reflect the intended routing policy. 230 3.2. Restrictions on Sending/Receiving 232 An implementation that supports the AIGP attribute MUST support a 233 per-session configuration item, AIGP_SESSION, that indicates whether 234 the attribute is enabled or disabled for use on that session. 236 - The default value of AIGP_SESSION, for EBGP sessions, MUST be 237 "disabled". 239 - The default value of AIGP_SESSION, for IBGP and confederation- 240 EBGP sessions, MUST be "enabled." 242 The AIGP attribute MUST NOT be sent on any BGP session for which 243 AIGP_SESSION is disabled. 245 If an AIGP attribute is received on a BGP session for which 246 AIGP_SESSION is disabled, the attribute MUST be treated exactly as if 247 it were an unrecognized non-transitive attribute. That is, "it MUST 248 be quietly ignored and not passed along to other BGP peers" (see 250 [BGP], section 5). 252 3.3. Creating and Modifying the AIGP Attribute 254 3.3.1. Originating the AIGP Attribute 256 An implementation that supports the AIGP attribute MUST support a 257 configuration item, AIGP_ORIGINATE, that enables or disables its 258 creation and attachment to routes. The default value of 259 AIGP_ORIGINATE MUST be "disabled". 261 A BGP speaker MUST NOT add the AIGP attribute to any route whose path 262 leads outside the "AIGP administrative domain" to which the BGP 263 speaker belongs. It may be added only to routes that satisfy one of 264 the following conditions: 266 - The route is a static route that is being redistributed into BGP 268 - The route is an IGP route that is being redistributed into BGP 270 - The route is an IBGP-learned route whose AS_PATH attribute is 271 empty. 273 - The route is an EBGP-learned route whose AS_PATH contains only 274 ASes that are in the same AIGP Administrative Domain as the BGP 275 speaker. 277 A BGP speaker R MUST NOT add the AIGP attribute to any route for 278 which R does not set itself as the next hop. 280 It SHOULD be possible to set AIGP_ORIGINATE to "enabled for the 281 routes of a particular IGP that are redistributed into BGP" (where "a 282 particular IGP" might be "OSPF" or "ISIS"). Other policies 283 determining when and whether to originate an AIGP attribute are also 284 possible, depending on the needs of a particular deployment scenario. 286 When originating an AIGP attribute for a BGP route to address prefix 287 P, the value of the attribute is set according to policy. There are 288 a number of useful policies, some of which are in the following list: 290 - When a BGP speaker R is redistributing into BGP an IGP route to 291 address prefix P, the IGP will have computed a "distance" from R 292 to P. This distance MAY be assigned as the value of AIGP 293 attribute. 295 - A BGP speaker R may be redistributing into BGP a static route to 296 address prefix P, for which a "distance" from R to P has been 297 configured. This distance MAY be assigned as the value of AIGP 298 attribute. 300 - A BGP speaker R may have received and installed a BGP-learned 301 route to prefix P, with next hop N. Or it may be redistributing 302 a static route to P, with next hop N. Then: 304 * If R has an IGP route to N, the IGP-computed distance from R 305 to N MAY be used as the AIGP attribute value of the route to 306 P. 308 * If R has a BGP route to N, and an AIGP attribute value has 309 been computed for that route (see section 3.3.3), that value 310 MAY be used as the AIGP attribute value of the route to P. 312 3.3.2. Modifications by the Originator 314 If BGP speaker R is the originator of the AIGP attribute of prefix P, 315 and at some point the "distance" from R to P changes, R SHOULD issue 316 a new BGP update containing the new value of the AIGP attribute. 317 (Here we use the term "distance" to refer to whatever value the 318 originator assigns to the AIGP attribute, however it is computed; see 319 section 3.3.1.) However, if the difference between the new distance 320 and the distance advertised in the AIGP attribute is less than a 321 configurable threshold, the update MAY be suppressed. 323 3.3.3. Modifications by a Non-Originator 325 Suppose a BGP speaker R1 receives a route with an AIGP attribute 326 whose value is A, and a Next Hop whose value is R2. Suppose also 327 that R1 is about to redistribute that route on a BGP session that is 328 enabled for sending/receiving the attribute. 330 If R1 does not change the Next Hop of the route, then R1 MUST NOT 331 change the AIGP attribute value of the route. 333 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 334 route to R2 is an IGP-learned route, or a static route that does not 335 require recursive next hop resolution, then R1 must increase the 336 value of the AIGP attribute by adding to A the distance from R1 to 337 R2. This distance is either the IGP-computed distance from R1 to R2, 338 or some value determined by policy. However, A MUST be increased by 339 a non-zero amount. 341 Note that if R1 and R2 above are EBGP neighbors, and there is a 342 direct link between them on which no IGP is running, then when R1 343 changes the next hop of a route from R2 to R1, the AIGP metric value 344 MUST be increased by a non-zero amount. The amount of the increase 345 SHOULD be such that it is properly comparable to the IGP metrics. 346 E.g., if the IGP metric is a function of latency, then the amount of 347 the increase should be a function of the latency from R1 to R2. 349 If R1 changes the Next Hop of the route from R2 to R1, and if R1's 350 route to R2 is a BGP-learned route, or a static route that requires 351 recursive next hop resolution, then the AIGP attribute value needs to 352 be increased in several steps, according to the following procedure. 353 (Note that this procedure is ONLY used when recursive next hop 354 resolution is needed.) 356 1. Let Xattr be the new AIGP attribute value. 358 2. Initialize Xattr to A. 360 3. Set the XNH to R2. 362 4. Find the route to XNH. 364 5. If the route to XNH does not require recursive next hop 365 resolution, get the distance D from R1 to XNH. (Note that this 366 condition cannot be satisfied the first time through this 367 procedure.) If D is above a configurable threshold, set the 368 AIGP attribute value to Xattr+D. If D is below a configurable 369 threshold, set the AIGP attribute value to Xattr. In either 370 case, exit this procedure. 372 6. If the route to XNH is a BGP-learned route, and the route does 373 NOT have an AIGP attribute, then exit this procedure and do not 374 pass on any AIGP attribute. 376 7. If the route to XNH is a BGP-learned route, and the route has 377 an AIGP attribute value of Y, then set Xattr=Xattr+Y, and set 378 XNH to the next hop of this route. (The intention here is that 379 Y is the AIGP value of the route as it was received by R1, 380 without having been modified by R1.) 382 8. Go to step 4. 384 The AIGP value of a given route depends on (a) the AIGP values of all 385 the next hops that are recursively resolved during this procedure, 386 and (b) the IGP distance to any next hop that is not recursively 387 resolved. Any change due to (a) in any of these values MUST trigger 388 a new AIGP computation for that route. Whether a change due to (b) 389 triggers a new AIGP computation depends upon whether the change in 390 IGP distance exceeds a configurable threshold. 392 If the AIGP attribute is carried across several ASes, each with its 393 own IGP domain, it is clear that these procedures are unlikely to 394 give a sensible result if the IGPs are different (e.g., some OSPF and 395 some IS-IS), or if the meaning of the metrics is different in the 396 different IGPs (e.g., if the metric represents bandwidth in some IGP 397 domains but represents latency in others). These procedures also are 398 unlikely to give a sensible result if the metric assigned to inter-AS 399 BGP links (on which no IGP is running) or to static routes is not 400 comparable to the IGP metrics. All such cases are outside the scope 401 of the current document. 403 4. Decision Process 405 Support for the AIGP attribute involves several modifications to the 406 tie breaking procedures of the BGP "phase 2" decision described in 407 [BGP], section 9.1.2.2. These modifications are described below in 408 sections 4.1 and 4.2. 410 In some cases, the BGP decision process may install a route without 411 executing any tie breaking procedures. This may happen, e.g., if 412 only one route to a given prefix has the highest degree of preference 413 (as defined in [BGP] section 9.1.1). In this case, the AIGP 414 attribute is not considered. 416 In other cases, some routes may be eliminated before the tie breaking 417 procedures are invoked, e.g., routes with AS-PATH attributes 418 indicating a loop, or routes with unresolvable next hops. In these 419 cases, the AIGP attributes of the eliminated routes are not 420 considered. 422 4.1. When a Route has an AIGP Attribute 424 Assuming that the BGP decision process invokes the tie breaking 425 procedures, the procedures in this section MUST be executed BEFORE 426 any of the tie breaking procedures described in [BGP] section 9.1.2.2 427 are executed. 429 If any routes have an AIGP attribute, remove from consideration all 430 routes that do not have an AIGP attribute. 432 If router R is considering route T, where T has an AIGP attribute, 433 - then R must compute the value A, defined as follows: set A to the 434 sum of (a) T's AIGP attribute value and (b) the IGP distance from 435 R to T's next hop. 437 - remove from consideration all routes that are not tied for the 438 lowest value of A. 440 4.2. When the Route to the Next Hop has an AIGP attribute 442 Suppose that a given router R1 is comparing two BGP-learned routes, 443 such that either: 445 - the two routes have equal AIGP attribute values, or else 447 - neither of the two routes has an AIGP attribute. 449 The BGP decision process as specified in [BGP] makes use, in its tie 450 breaker procedures, of "interior cost", defined as follows: 452 "interior cost of a route is determined by calculating the metric 453 to the NEXT_HOP for the route using the Routing Table." 455 Suppose route T has a next hop of N. We modify the notion of the 456 "interior cost" from node R1 to node N as follows: 458 - Let R2 be the BGP next hop of the route to N, after all recursive 459 resolution of the next hop is done. Let m be the IGP distance 460 (or in the case of a static route, the configured distance) from 461 R1 to R2. 463 - If the installed route to N has an AIGP attribute, set A to the 464 AIGP value of the route to N, computing the AIGP value of the 465 route according to the procedure of section 3.3.3. 467 - If the installed route to N does not have an AIGP value, set A to 468 0. 470 - The "interior cost" of route T is the quantity A+m. 472 5. Deployment Considerations 474 Using the AIGP attribute to achieve a desired routing policy will be 475 more effective if each BGP speaker can use it to choose from among 476 multiple routes. Thus is it highly recommended that the procedures of 477 [BESTEXT] and [ADDPATH] be used in conjunction with the AIGP 478 Attribute. 480 If a Route Reflector does not pass all paths to its clients, then it 481 will tend to pass the paths for which the IGP distance from the Route 482 Reflector itself to the next hop is smallest. This may result in a 483 non-optimal choice by the clients. 485 6. IANA Considerations 487 IANA has assigned the codepoint 26 in the "BGP Path Attributes" 488 registry to the AIGP attribute. 490 IANA shall create a registry for "BGP AIGP Attribute Types". The 491 type field consists of a single octet, with possible values from 0 to 492 255. The allocation policy for this field is to be "Standards Action 493 with Early Allocation". Type 1 should be defined as "AIGP", and 494 should refer to this document. 496 7. Security Considerations 498 The spurious introduction, though error or malfeasance, of an AIGP 499 attribute, could result in the selection of paths other than those 500 desired. 502 Improper configuration on both ends of an EBGP connection could 503 result in an AIGP attribute being passed from one service provider to 504 another. This would likely result in an unsound selection of paths. 506 8. Acknowledgments 508 The authors would like to thank Waqas Alam, Rajiv Asati, Clarence 509 Filsfils, Robert Raszuk, Yakov Rekhter, Eric Rosenberg, Samir Saad, 510 John Scudder, and Shyam Sethuram for their input. 512 9. Authors' Addresses 514 Rex Fernando 515 Cisco Systems, Inc. 516 170 Tasman Drive 517 San Jose, CA 95134 518 Email: rex@cisco.com 520 Pradosh Mohapatra 521 Cumulus Networks 522 Email: pmohapat@cumulusnetworks.com 524 Eric C. Rosen 525 Cisco Systems, Inc. 526 1414 Massachusetts Avenue 527 Boxborough, MA, 01719 528 Email: erosen@cisco.com 530 James Uttaro 531 AT&T 532 200 S. Laurel Avenue 533 Middletown, NJ 07748 534 Email: uttaro@att.com 536 10. Normative References 538 [BGP], "A Border Gateway Protocol 4 (BGP-4)", Y. Rekhter, T. Li, S. 539 Hares, RFC 4271, January 2006. 541 11. Informative References 543 [ADDPATH] "Fast Connectivity Restoration Using BGP Add-Path", P. 544 Mohapatra, R. Fernando, C. Filsfils, R. Raszuk, draft-pmohapat-idr- 545 fast-conn-restore-03.txt, January 2013. 547 [BESTEXT], "Advertisement of the Best External Route in BGP", P. 548 Marques, R. Fernando, E. Chen, P. Mohapatra, H. Gredler, draft-ietf- 549 idr-best-external-05.txt, January 2012. 551 [RFC2119] "Key words for use in RFCs to Indicate Requirement 552 Levels.", S. Bradner, March 1997.