idnits 2.17.1 draft-ietf-bmwg-bgpbas-01.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: ---------------------------------------------------------------------------- ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? == No 'Intended status' indicated for this document; assuming Proposed Standard == The page length should not exceed 58 lines per page, but there was 15 longer pages, the longest (page 2) being 59 lines Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack a Security Considerations section. ** 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. == There are 30 instances 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 (February 2002) is 8104 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: '3' is defined on line 618, but no explicit reference was found in the text == Unused Reference: '7' is defined on line 637, but no explicit reference was found in the text == Unused Reference: '8' is defined on line 642, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. '3' ** Downref: Normative reference to an Informational RFC: RFC 2544 (ref. '5') ** Downref: Normative reference to an Informational RFC: RFC 2928 (ref. '8') ** Downref: Normative reference to an Informational RFC: RFC 3222 (ref. '9') ** Obsolete normative reference: RFC 1771 (ref. '10') (Obsoleted by RFC 4271) ** Obsolete normative reference: RFC 2385 (ref. '11') (Obsoleted by RFC 5925) == Outdated reference: A later version (-13) exists of draft-ietf-idr-restart-02 == Outdated reference: A later version (-06) exists of draft-ietf-bmwg-conterm-01 ** Downref: Normative reference to an Informational draft: draft-ietf-bmwg-conterm (ref. '13') -- Possible downref: Non-RFC (?) normative reference: ref. '14' Summary: 11 errors (**), 0 flaws (~~), 9 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Benchmarking Working Group H.Berkowitz, Gett Communications 3 Internet Draft S.Hares, Nexthop 4 Document: draft-ietf-bmwg-bgpbas-01.txt A.Retana, Cisco 5 Expires August 2002 P.Krishnaswamy 6 M.Lepp, Juniper Networks 7 E.Davies, Nortel Networks 8 February 2002 10 Benchmarking Methodology for Basic BGP Device Convergence 12 Status of this Memo 14 This document is an Internet-Draft and is in full conformance with 15 all provisions of Section 10 of RFC2026 [1]. 17 Internet-Drafts are working documents of the Internet Engineering 18 Task Force (IETF), its areas, and its working groups. Note that other 19 groups may also distribute working documents as Internet-Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet- Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt 27 The list of Internet-Draft Shadow Directories can be accessed at 28 http://www.ietf.org/shadow.html. 30 A revised version of this draft document will be submitted to the RFC 31 editor as a Informational document for the Internet Community. 32 Discussion and suggestions for improvement are requested. 33 This document will expire before August 2002. Distribution of this 34 draft is unlimited. 36 Abstract 38 This draft begins the process of establishing standards for measuring 39 the performance of the BGP routing subsystem in a network. Its 40 initial emphasis is on the control plane of single BGP devices. We 41 do not address forwarding plane performance. 43 Conventions used in this document 45 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 46 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 47 document are to be interpreted as described in RFC-2119 [2]. 49 Berkowitz, et al 1 50 Table of Contents 52 1. Introduction....................................................3 53 1.1 Overview and Roadmap.......................................3 54 1.2 Scope......................................................4 55 1.3 Types of Single-Device Convergence.........................4 56 2. Reference Configurations........................................5 57 3. Basic eBGP tests................................................6 58 3.1 Connection Conditions......................................6 59 3.2 Test Streams...............................................7 60 3.3 Route Mixtures.............................................7 61 3.4 Order of Received Updates..................................8 62 3.5 Initial Convergence........................................9 63 3.6 Incremental Re-convergence with a Single Peer and a Single 64 Update.................................................10 65 3.7 Incremental Reconvergence with a Single Peer and Small Packet 66 Trains.................................................10 67 3.8 Incremental Re-convergence with Multiple Peers............11 68 4. Route Flaps....................................................12 69 4.1 Flap Isolation Test.......................................12 70 4.2 Authentication............................................12 71 5. Acknowledgements...............................................12 72 6. References.....................................................13 73 7. Acknowledgments................................................14 74 8. Author's Addresses.............................................14 75 Appendix A. Representative Scenarios..............................15 76 A.1 Default-free interprovider peering........................15 77 A.2 Interprovider peering with transit........................15 78 A.3 Provider edge device......................................15 79 A.4 Multihomed subscriber edge device.........................15 81 Berkowitz, et al Expires: August 2002 2 82 1. Introduction 84 This document describes a specific set of tests aimed at 85 characterizing the convergence performance of BGP-4 processes in 86 devices that incorporate BGP functionality as described in [10] and 87 subsequent additions. Such devices include conventional routers, 88 route servers, "layer 3 switches" with external path determination 89 engines (e.g. Ethernet Switch/Routers), and controllers of sub-IP 90 path creation and management. A key objective is to propose 91 methodology that will standardize the conducting and reporting of 92 convergence-related measurements. 94 The convergence performance of BGP-4 processes is important to the 95 effectiveness and efficiency of IP networks. Poor performance can 96 make the network slow to respond to changes and failures, unnecessary 97 processing when updates are delivered over a longer period than is 98 desirable and consequent misdirection or delay of traffic. 100 Although both convergence and forwarding are essential to basic 101 device operation, this document does not consider the forwarding 102 performance in the Device Under Test (DUT), for two reasons: 103 - Forwarding performance is being treated separately: Methodology 104 for forwarding performance is the primary focus in [5] and it is 105 expected to be further covered in work that ensues from the 106 definitions of terminology for Forwarding Information Bases in 107 [9]. 108 - The additional time taken to establish new forwarding behavior 109 after the BGP-4 processes have determined new routes and generated 110 adverts to downstream devices should not affect the overall time 111 for convergence of the network 113 Further, as convergence characterization is a complex process, we 114 deliberately restrict this document to basic measurements aimed at 115 characterizing BGP convergence in an isolated device receiving inputs 116 on two interfaces and generating outputs on a third interface. 118 Subsequent versions of this document and other document will explore 119 more complex interconnections, interactions of several devices and 120 the more intricate aspects of convergence measurement, such as the 121 presence of policy processing, simultaneous traffic on the control 122 and data paths within the DUT, and other realistic performance 123 modifiers. Convergence of Interior Gateway Protocols will be 124 considered in separate drafts. 126 1.1 Overview and Roadmap 128 Measurements of protocols can be classified either as internal or 129 external. Internal measurements are derived from time-stamps applied 130 within the Device Under Test (DUT). External measurements infer the 131 timing of a process in the DUT from measurements made on externally 132 observable phenomena such as transmission of packets to or reception 133 of packets from the DUT by connected test devices. In the case of 134 BGP convergence, the DUT is stimulated by sending BGP UPDATE packets 135 to one or more of its interfaces: If measurements of control plane 137 Berkowitz, et al Expires: August 2002 3 138 behavior are required, the progress of the BGP processes can be 139 gauged by observing the UPDATEs generated and transmitted by the DUT 140 in response to the stimuli as they are received by test receivers 141 connected to the interfaces. An alternative type of external 142 measurement, involving both the data and control planes, is to test 143 for data forwarded to the downstream device that relies upon the new 144 route(s) just installewd in the FIB of the Device Under Test. 146 We focus here on external measurements based only on control plane 147 phenomena, thus facilitating black box comparisons of the routing 148 subsystem in devices with diverse internal architectures and 149 functions. 151 If alternative internal measurements were adopted, correlating the 152 DUT's time stamps with those from the rest of the test system is a 153 key problem: The requisite Network Time Protocol (NTP) functionality 154 may not be present and it may be difficult to reach the precision 155 needed for these measurements. 157 For the purposes of this paper, the external technique is more 158 readily applicable. However, external measurements have their own 159 problems because they include the time to advertise the new route 160 downstream and transmission times for the advertisement within the 161 device under test. If data forwarding were to feature in the 162 measurement methodology it too would include some extraneous latency 163 - that of the forwarding lookup process in the DUT at the minimum. 164 This document deals only with external measurements limited to route 165 propagation. 167 Characterization of the BGP convergence performance of a device 168 SHOULD take into account all distinct stages and aspects of BGP 169 functionality. This requires that the relevant terms and metrics be 170 as specific as possible. A terminology that meets this objective was 171 presented in "Terinology for Benchmarking External Routing 172 Convergence Measurements" [13]. 174 1.2 Scope 176 This document deals with eBGP convergence of a single deviceDevice 177 Under Test (DUT). It restricts the measurement of convergence to 178 events in the control plane, and does not consider the interactions 179 of convergence and forwarding. 181 Convergence measurements among multiple iBGP-connected devices in an 182 AS, and Internet-wide convergence measurements, are also outside the 183 scope of this document. 185 These additional topics are unquestionably of interest, and it is the 186 intention of this document to form a stepping stone toward them 188 1.3 Types of Single-Device Convergence 190 There are two major types of convergence time that tend to be lumped 191 together in product specifications: 193 Berkowitz, et al Expires: August 2002 4 194 - The time needed for a BGP speaker to build a full table after 195 initialization, or for a particular peering session to rebuild its 196 table after a hard reset (see [12],[13] and section 3.5). 197 - The time needed for a device to respond to a new announcement or 198 withdrawal. This second time has two subtypes: the time to 199 reconverge a update with a single prefix, and the time to 200 reconverge after receiving a small train of updates. See sections 201 3.6 and 3.7. 203 External measurements start with the delivery of a stimulus or the 204 first of several route advertisement stimuli from one or more 205 "upstream" devices (identified as TD1 to Tdn) and end when the BGP 206 process(es) in the device have returned to equilibrium as indicated 207 by all advertisements resulting from the stimuli having been sent to 208 a "downstream" peer (TDrx). In the reference configurations below, 209 external measurements are defined with respect to TDrx as the 210 downstream device. 212 2. Reference Configurations 214 For tests when the number of peers is not a performance parameter 215 of interest, use the configuration in Figure 1: 217 +---------+ 218 TD1==========| |==========TDrx 219 | | | 220 D1 | | 221 | | DUT | 222 TD2==========| | 223 +---------+ 225 Figure 1. Basic Test Configuration. 227 D1 is a prefix reachable by both TD1 and TD2. Neither TD1 nor TD2 is 228 the originating AS for the announcement of D1. Stimuli will 229 typically be generated by one or both of TD1 and TD2 according to a 230 timed schedule. The DUT will propagate consequent adverts towards 231 TDrx where their arrival will be recorded and timed. 233 More complex peering arrangements will involve up to n Test Routers, 234 as shown in Figure 2. It is recommended that the Figure 1 235 configuration always be tested as a baseline, and then additional 236 reports made that show the effect on performance of increasing the 237 number of peers. Again stimuli would be expected from one or more 238 of the TD1 to TDn. 240 Berkowitz, et al Expires: August 2002 5 241 +---------+ 242 TD1==========| |==========TDrx 243 | | | 244 D1 | | 245 | | DUT | 246 TD2==========| | 247 | | 248 ... 249 TDn==========+---------+ 251 Figure 2. Test Configuration with n Peers. 253 Interface speeds and types MUST be specified as part of the test 254 report. At least 100 Mbps is recommended, so media delays are not a 255 significant component of convergence times. 257 In the absence of other route selection criteria, TD1 SHALL have an 258 IP address that makes it most preferred. 260 3. Basic eBGP tests 262 All devices in this configuration SHOULD have a policy of ADVERTISE 263 ALL/ACCEPT ALL [6]. Tests with prefix filtering, community-based 264 preferences, authentication, etc., as well as performance under route 265 flap conditions are TBD. 267 Not all eBGP applications are alike. While the tests in this section 268 are applicable to a wide range of configurations, testers MAY select 269 configurations that are most relevant to the intended product use. 270 Such configurations include: 272 1. Interprovider peering, characterized by an exchange of customer 273 routes, which, in the case of major providers, may be in the tens 274 of thousands of routes but smaller than the full default-free 275 table. 277 2. Provider/Subscriber edge peering, where transit service implies 278 the subscriber advertises relatively few routes to the provider 279 but may take, variously, a full set of default-free routes, a 280 limited subset of the full set, or just a default route from the 281 provider. 283 3.1 Connection Conditions 285 The DUT SHOULD be physically connected to the test devices over a 286 medium sufficiently fast that propagation time is not a significant 287 factor. A medium of at least 100 Mbps is recommended. 289 Multiple peers MAY be connected to a single physical interface using 290 802.1q VLANs or another appropriate multiplexing scheme, such as a 291 channelized interface. If so, this MUST be documented in the test 292 results because it may change the arrival times of UPDATEs by 294 Berkowitz, et al Expires: August 2002 6 295 serializing packets which might otherwise arrive in parallel where 296 truly separate, asynchronous interfaces are used. 298 TCP connections SHOULD NOT use slow start. Any nonstandard initial 299 or maximum window sizes SHALL be indicated in the test report. 301 3.2 Test Streams 303 Update Packet trains presented to the DUT SHOULD in general be random 304 (see definition of random update train in [13]) with respect to 305 selection of prefixes, prefix length, ordering of prefixes, and time 306 of delivery to DUT. Note that this does not preclude prior, offline 307 creation of sets of update trains with the required randomness that 308 can then be used in running the tests multiple times to determine the 309 reproducibility of results, and for use in comparison tests between 310 products. There may also be circumstances such as are described in 311 Section 3.3 where specific ordering of prefixes may be appropriate 312 for some tests. 314 The degree of update packing SHALL be specified. When long update 315 trains are being sent, the usual case is that the maximum possible 316 number of prefixes are packed into an UPDATE packet subject to the 317 MTU size of the link over which they are being sent. 319 3.3 Route Mixtures 321 As shown by measurements of routers in actual deployment, such as are 322 documented in 'The CIDR Report' [14] and similar monitoring projects, 323 both the prefix length distribution and the clustering of prefixes 324 take on characteristic values in the mixture of routes seen. 326 There are two sets of statistics which exhibit related but different 327 characteristics: 328 - The distribution in typical default free routing table 329 - The distribution in the dynamic UPDATEs arriving at a device 331 The characteristics are reasonably consistent although there are 332 significant bursts of activity from time to time that distort the 333 normal situation. 335 In creating update trains as test stimuli, these characteristics 336 SHOULD be used to drive: 337 - The distribution of prefix lengths 338 - The clustering of prefixes in the total prefix space 340 The characteristics used should be appropriate for the sort of test 341 in which the update train is to be used. Initial table load trains 342 should reflect the structure of a default free routing table whereas 343 trains for incremental updates should typically reflect the 344 characteristics of the dynamic UPDATEs. 346 Future versions of this document may suggest specific profiles for 347 these characteristics, but these remain TBD at present. 349 Berkowitz, et al Expires: August 2002 7 350 3.4 Order of Received Updates 352 For the fairest testing of update trains the order of the prefixes 353 SHALL include one randomized test. It Should also include one test 354 sorted by prefix size, and one radix tree implementation. 356 Assume we have a Adj-RIB-out that consists of 358 1.0.0.0/8 359 2.0.0.0/8 360 3.0.0.0/8 361 1.1.0.0/16 362 2.1.0.0/16 363 3.1.0.0/16 364 3.2.0.0/16 365 1.1.1.0/24 366 1.1.2.0/24 367 2.1.2.0/24 369 If it were sent in this order, top to bottom, it would be sorted by 370 prefix size and prefix value within size. A radix tree 371 implementation might like to receive this very much. 373 But if it were sent out in the following order 375 1.0.0.0/8 376 1.1.0.0/16 377 1.1.1.0/24 378 1.1.2.0/24 379 2.0.0.0/8 380 2.1.0.0/16 381 2.1.2.0/24 382 3.0.0.0/8 383 3.1.0.0/16 384 3.2.0.0/16 386 It would favor an implementation that orders its routing table as a 387 strict tree, implemented as a linked list. 389 A 'fair' test train would be 391 1.0.0.0/8 392 2.1.0.0/16 393 1.1.0.0/16 394 3.0.0.0/8 395 1.1.1.0/24 396 2.0.0.0/8 397 1.1.2.0/24 398 3.1.0.0/16 399 2.1.2.0/24 400 3.2.0.0/16 402 which is random, and equally fair to any particular implementation. 404 Berkowitz, et al Expires: August 2002 8 405 On the other hand, when dealing with a network of devices from a 406 single vendor, in the updates forwarded from a device as a result of 407 a set of stimuli, particularly during a complete table load, the 408 prefixes may be ordered, both within the route packing in a single 409 UPDATE and across the update train which results from the stimuli, in 410 such a way as to be advantageous to downstream devices of the same 411 type: Hence, it MAY be desirable to measure both randomized and 412 'friendly' orders so as to get a more realistic view of the real 413 world behavior of the devices. Note that this can only apply to 414 update trains where the individual update packets are delivered close 415 together in time. If the spacing is too great(greater than the 416 MIN_ADVERT_TIME) the packets will become separate stimuli that are 417 processed individually. 419 Measurement units: A metric of randomness,TBD 421 3.5 Initial Convergence 423 While this is relatively simple to measure, and often is the basis of 424 product specifications, it is operationally far less significant than 425 reconvergence after changes. A "carrier-grade" device should not 426 initialize often, and the proposed soft reset option reduces the need 427 to rebuild views. The initialization time, therefore, can be 428 amortized over a long period of time and may disappear into the noise 429 when compared to reconvergence (See [12] for details of soft restart 430 standards proposals. Proprietary implementations already exist). 432 3.5.1 Single Peer Initial Convergence Time 434 This basic reference test uses a representatively sized and populated 435 target RIB and the device SHOULD be configured for as basic behavior 436 as possible, thus minimizing variable influences (eg authentication 437 off, filters off, no policy, slow start off). 439 The test begins with OPEN requests sent from TD1 and TD2 to the DUT. 440 Each Test Router sends a standard routing table with a number, NR, of 441 routes, designated first route (FR) to last route (LR). The value of 442 NR should be reported with every test. There are perfectly valid 443 reasons to test with a small NR, such as testing a device intended as 444 a small multihomed enterprise gateway to the Internet. In contrast, 445 a large NR would be appropriate for a device intended as a major 446 interprovider gateway. 448 Conceptually, the test ends when the DUT begins to advertise the last 449 route, LR, in the routing table to TDrx. Since individual 450 implementations may vary in the order in which they construct their 451 outgoing updates i.e., different ordering, packing, etc.), it is 452 possible that the received LR will not necessarily be the last update 453 advertised by the DUT. Note that the routes FR and LR are not in any 454 respect special. They are identified so that progress of the 455 stimulus generator can be monitored and the corresponding events that 456 might be logged in the DUT can be identified. 458 Berkowitz, et al Expires: August 2002 9 459 The test receiver (e.g TDrx) SHOULD record the time at which LR is 460 advertised, but also continue monitoring to see if additional routes 461 are advertised. Initial convergence time is the interval between 462 receipt of FR to the later of two events: the reception of re- 463 advertised LR, or the last update received after the stimulus of LR. 464 LR may be, and often will, be the last update, but that is not 465 guaranteed. 467 3.5.2 Multiple Peers 469 TBD 471 3.6 Incremental Re-convergence with a Single Peer and a Single Update 473 For all of these measurements, an update train with a single update 474 is used and the device SHOULD be configured for as basic behavior as 475 possible, thus minimizing variable influences (eg authentication off, 476 filters off, no policy, slow start off). 478 3.6.1 Explicit add of single new route 480 This test measures the time required to add a single route (D1) newly 481 advertised by a peer. At the start of the test the route does not 482 exist in the DUT's RIB, and hence the new route will not displace a 483 route in the RIB. 485 The DUT has been initialized, with no path to D1. Measurement time 486 begins when TD1 announces D1 to the DUT. 488 Measurement time stops when the DUT advertises D1 to TDrx. 490 3.6.2 Sequential withdraw and reannounce of a Single Prefix 492 The DUT has been initialized and has a path to D1 via TD1, but not 493 via TD2. Simultaneously, TD1 sends TDown(S,TD1) and TD2 announces the 494 new route with Tbest(TD2). 496 Measurement begins when Tbest is received at the DUT. Measurement 497 time stops when the DUT advertises the new route to D1 to TDrx. 499 3.6.3 Time to Change to Alternate Path after Explicit Withdrawal of a 500 Single Route 502 The DUT has been initialized and has paths to D1 via both TD1 and 503 TD2. TD1's path is preferred, but TD1 withdraws it with 504 TDown(S,TD1)). Re-convergence occurs when a route from the path(s) 505 advertised by TD2 becomes active. 507 Measurement time stops when the DUT advertises the new route to D1 508 via TD2 to TDrx. 510 3.7 Incremental Reconvergence with a Single Peer and Small Packet Trains 512 Berkowitz, et al Expires: August 2002 10 513 For all of these measurements, an update train with a small number of 514 updates is used and the device SHOULD be configured for as basic 515 behavior as possible, thus minimizing variable influences (eg 516 authentication off, filters off, no policy, slow start off). The 517 train SHOULD deliver the updates over a short period of time so that 518 the device may deal with them as a batch rather than reconverging 519 separately for each UPDATE packet received. 521 3.7.1 Explicit add of several new routes 523 This test measures the time required to add a number of routes (Dm) 524 newly advertised by a peer. At the start of the test these routes do 525 not exist in the DUT's RIB, and hence the new routes will not 526 displace any routes in the RIB. 528 The DUT has been initialized, with no path to any of Dm. Measurement 529 time begins when TD1 announces D1 to the DUT. 531 Measurement time stops when the DUT advertises the last of the routes 532 Dm to TDrx. 534 3.7.2 Sequential withdraw and reannounce for a small group of prefixes 536 The DUT has been initialized and has paths to each of Dm via TD1, but 537 not via TD2. Simultaneously, TD1 sends TDown(S,TD1) and TD2 announces 538 the new route with Tbest(S,TD2). 540 Measurement begins when Tbest(S.TD1) is received at the DUT. 541 Measurement time stops when the DUT advertises the last of the new 542 routes to Dm to TDrx. 544 3.7.3 Time to Change to Alternate Path after Explicit Withdrawal of 545 several routes 547 The DUT has been initialized and has paths to each of Dm via both TD1 548 and TD2. TD1's path is preferred in each case, but TD1 withdraws it 549 with TDown(S,TD1). Re-convergence occurs when routes selected from 550 the path(s) advertised by TD2 become active. 552 Measurement time stops when the DUT advertises the last of the routes 553 to Dm via TD2 to TDrx. 555 3.8 Incremental Re-convergence with Multiple Peers 557 The number of routes per BGP peer is an obvious stressor to the 558 convergence process. The number, and relative proportion, of 559 multiple route instances and distinct routes being added or 560 withdrawn by each peer will affect the convergence process, as will 561 the mix of overlapping route instances advertised by teo or more 562 peers. 564 Berkowitz, et al Expires: August 2002 11 565 4. Route Flaps 567 The following tests evaluate convergence when route flap exists. 569 Let TDF be a device that will generate only flapping routes. 571 +---------+ 572 TD1==========| |==========TDrx 573 | | | 574 D1 | | 575 | | DUT | 576 TD2==========| | 577 | | 578 ... 579 TDF==========+---------+ 581 Figure 3. Test Diagram with a Router, TDF, flapping. 583 4.1 Flap Isolation Test 585 TDF will advertise a continuously flapping route i.e. repeated 586 advertisements and withdrawals of a single route sent at intervals 587 equal to the MIN_ADVERT_TIME. Repeat the eBGP convergence tests. The 588 objective is to determine whether one route flapping affects the 589 operation of the device. 591 If the DUT implements the BGP-4 route flap damping capability 592 described in [4], then the capability SHOULD be disabled for this 593 test. Testing of the route flap damping capability is FFS. 595 4.2 Authentication 597 Repeat all tests above with MD5 authentication if the DUT implements 598 the capabilities described in [11]. 600 Repeat all tests with Ipsec authentication turned on. 602 5. Acknowledgements 604 Thanks to Francis Ovenden for review and Abha Ahuja for 605 encouragement. Much appreciation to Jeff Haas, Matt Richardson, and 606 Shane Wright at Nexthop for comments and input. 608 Berkowitz, et al Expires: August 2002 12 609 6. References 611 [1] Bradner, S., "The Internet Standards Process -- 612 Revision 3", BCP 9, RFC 2026, October 1996 614 [2] Bradner, S., "Key words for use in RFCs to 615 Indicate Requirement Levels", BCP 14, RFC 2119, 616 March 1997 618 [3] Ahuja, A., Jahanian, F., Bose, A. and Labovitz, 619 C., 620 "An Experimental Study of Delayed Internet 621 Routing Convergence", RIPE 37 - Routing WG. 623 [4] Villamizar, C., Chandra, R. and Govindan, R., 624 "BGP Route Flap Damping", RFC 2439, 625 November 1998. 627 [5] Bradner, S. and McQuaid, J., "Benchmarking 628 Methodology for Network Interconnect Devices", 629 RFC 2544, March 1999 631 [6] Alaettinoglu, C., Villamizar, C., Gerich, E., 632 Kessens, D., Meyer, D., Bates, T., Karrenberg, 633 D. and Terpstra, M., "Routing Policy 634 Specification Language (RPSL)", RFC 2622, June 635 1999. 637 [7] Ferguson, P. and Senie, D., "Network Ingress 638 Filtering: Defeating Denial of Service Attacks 639 which employ IP Source Address Spoofing", 640 RFC 2827, May 2000 642 [8] Chen, E., "Route Refresh Capability for BGP-4", 643 RFC 2928, DATE NEEDED 645 [9] Trotter, G., "Terminology for Forwarding 646 Information Base(FIB)-based Router Performance", 647 RFC 3222, December 2001 649 [10] Rekhter, Y. and Li, T., "A Border Gateway 650 Protocol 4 (BGP-4)", RFC 1771, . March 1995. 652 [11] Heffernan, A., "Protection of BGP Sessions via 653 the TCP MD5 Signature Option", RFC 2385, August 654 1998. 656 [12] Ramachandra, S., Rekhter, Y., Fernando, R., 657 Scudder, J.G. and Chen, E., 658 "Graceful Restart Mechanism for BGP", 659 draft-ietf-idr-restart-02.txt, January 2002, work 660 in progress. 662 Berkowitz, et al Expires: August 2002 13 664 [13] Berkowitz, H., Hares, S., Retana, A., 665 Krishnaswamy, P. and Lepp, M., "Terminology for 666 Benchmarking External Routing Convergence 667 Measurements", draft-ietf-bmwg-conterm-01.txt, 668 Febtruary 2002, Work in progress 670 [14] Bates, T., "The CIDR Report", 671 http://www.employees.org/~tbates/cidr-report.html 672 Internet statistics relevant to inter-domain 673 routing updated daily 675 7. Acknowledgments 677 Thanks to Francis Ovenden for review and Abha Ahuja for 678 encouragement. Much appreciation to Jeff Haas, Matt Richardson, and 679 Shane Wright at Nexthop for comments and input. Debby Stopp and Nick 680 Ambrose contributed the concept of route packing. Thanks to Martin 681 Biddiscombe for trying out the tests. 683 8. Author's Addresses 685 Howard Berkowitz 686 Gett Communications 687 5012 S. 25th St 688 Arlington VA 22206 689 Phone: +1 703 998-5819 690 Fax: +1 703 998-5058 691 EMail: hcb@gettcomm.com 693 Elwyn Davies 694 Nortel Networks 695 London Road 696 Harlow, Essex CM17 9NA 697 UK 698 Phone: +44-1279-405498 699 Email: elwynd@nortelnetworks.com 701 Susan Hares 702 Nexthop Technologies 703 517 W. William 704 Ann Arbor, Mi 48103 705 Phone: 706 Email: skh@nexthop.com 708 Padma Krishnaswamy 709 Email: kri1@earthlink.net 711 Marianne Lepp 712 Juniper Networks 713 51 Sawyer Road 714 Waltham, MA 02453 715 Phone: 617 645 9019 716 Email: mlepp@juniper.net 718 Berkowitz, et al Expires: August 2002 14 719 Alvaro Retana 720 Cisco Systems, Inc. 721 7025 Kit Creek Rd. 722 Research Triangle Park, NC 27709 723 Email: aretana@cisco.com 725 Appendix A. Representative Scenarios 727 The following describes sample BGP applications positioned at various 728 points in the network. 730 A.1 Default-free interprovider peering 732 The DUT exchanges 0.3 to 0.5 D with a small number of peers. 733 Typically, devices in this application are limited by bandwidth 734 rather than route processing 736 A.2 Interprovider peering with transit 738 The DUT exchanges 1.3 D routes with a small number of peers. 740 A.3 Provider edge device 742 The DUT has a large number (>10) of eBGP peers. 744 To 10% of the peers, the DUT advertises 1.3 D. 745 To 20% of the peers, the DUT advertises 0.3 D. 746 To 70% of the peers, the DUT advertises default. 748 50% of the peers advertise an aggregate and a more-specific route to 749 the DUT. 750 20% of the peers advertise 10 or more routes to the DUT. 752 30% of the peers advertise a single route to the DUT. 754 A.4 Multihomed subscriber edge device 756 The DUT connects to 2 peers. It advertises an aggregate and a more- 757 specific to each. 759 Full Copyright Statement 761 Copyright (C) The Internet Society (2002). All Rights Reserved. 763 This document and translations of it may be copied and furnished to 764 others, and derivative works that comment on or otherwise explain it 765 or assist in its implmentation may be prepared, copied, published and 766 distributed, in whole or in part, without restriction of any kind, 767 provided that the above copyright notice and this paragraph are 768 included on all such copies and derivative works. However, this 769 document itself may not be modified in any way, such as by removing 770 the copyright notice or references to the Internet Society or other 771 Internet organizations, except as needed for the purpose of 773 Berkowitz, et al Expires: August 2002 15 774 developing Internet standards in which case the procedures for 775 copyrights defined in the Internet Standards process must be 776 followed, or as required to translate it into languages other than 777 English. 779 The limited permissions granted above are perpetual and will not be 780 revoked by the Internet Society or its successors or assigns. 782 This document and the information contained herein is provided on an 783 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING 784 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDIN BUT 785 NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN 786 WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF 787 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 789 Berkowitz, et al Expires: August 2002 16