idnits 2.17.1 draft-ietf-bmwg-ospfconv-applicability-07.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3667, Section 5.1 on line 17. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 479. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 500), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 37. ** The document seems to lack an RFC 3978 Section 5.1 IPR Disclosure Acknowledgement -- however, there's a paragraph with a matching beginning. Boilerplate error? ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** The document seems to lack an RFC 3978 Section 5.5 (updated by RFC 4748) Disclaimer -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack an RFC 3979 Section 5, para. 1 IPR Disclosure Acknowledgement -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack an RFC 3979 Section 5, para. 3 IPR Disclosure Invitation -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document uses RFC 3667 boilerplate or RFC 3978-like boilerplate instead of verbatim RFC 3978 boilerplate. After 6 May 2005, submission of drafts without verbatim RFC 3978 boilerplate is not accepted. The following non-3978 patterns matched text found in the document. That text should be removed or replaced: By submitting this Internet-Draft, I certify that any applicable patent or other IPR claims of which I am aware have been disclosed, or will be disclosed, and any of which I become aware will be disclosed, in accordance with RFC 3668. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity. == No 'Intended status' indicated for this document; assuming Proposed Standard == The page length should not exceed 58 lines per page, but there was 11 longer pages, the longest (page 2) being 60 lines == It seems as if not all pages are separated by form feeds - found 0 form feeds but 12 pages Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. 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 (June 2004) is 7252 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: 'TERM' is mentioned on line 68, but not defined == Missing Reference: 'OSPF' is mentioned on line 400, but not defined == Unused Reference: 'RFC2119' is defined on line 417, but no explicit reference was found in the text -- No information found for draft-bmwg-ospfconv-intraarea - is the name correct? -- Possible downref: Normative reference to a draft: ref. 'BENCHMARK' == Outdated reference: A later version (-06) exists of draft-ietf-isis-igp-p2p-over-lan-03 Summary: 10 errors (**), 0 flaws (~~), 8 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Network Working Group Vishwas Manral 2 Internet Draft Netplane Systems 3 Russ White 4 Cisco Systems 5 Aman Shaikh 6 Expiration Date: December 2004 University of California 7 File Name: draft-ietf-bmwg-ospfconv-applicability-07.txt June 2004 9 Considerations When Using Basic OSPF Convergence Benchmarks 10 draft-ietf-bmwg-ospfconv-applicability-07.txt 12 Status of this Memo 14 By submitting this Internet-Draft, I certify that any applicable 15 patent or other IPR claims of which I am aware have been disclosed, 16 and any of which I become aware will be disclosed, in accordance with 17 RFC 3668. 19 Internet Drafts are working documents of the Internet Engineering 20 Task Force (IETF), its Areas, and its Working Groups. Note that other 21 groups may also distribute working documents as Internet Drafts. 23 Internet Drafts are draft documents valid for a maximum of six 24 months. Internet Drafts may be updated, replaced, or obsoleted by 25 other documents at any time. It is not appropriate to use Internet 26 Drafts as reference material or to cite them other than as a "working 27 draft" or "work in progress". 29 The list of current Internet-Drafts can be accessed at 30 http://www.ietf.org/ietf/1id-abstracts.txt 32 The list of Internet-Draft Shadow Directories can be accessed at 33 http://www.ietf.org/shadow.html. 35 Copyright Notice 37 Copyright (C) The Internet Society (2004). All Rights Reserved. 39 Abstract 41 This document discusses the applicability of various tests for 42 measuring single router control plane convergence, specifically in 43 regards to the Open Shortest First (OSPF) protocol. There are two 44 general sections in this document, the first discussing specific 45 advantages and limitations of specific OSPF convergence tests, and 46 the second discussing more general pitfalls to be considered when 47 testing routing protocols convergence testing. 49 1. Introduction 51 There is a growing interest in testing single router control plane 52 convergence for routing protocols, with many people looking at 53 testing methodologies which can provide information on how long it 54 takes for a network to converge after various network events occur. 55 It is important to consider the framework within which any given 56 convergence test is executed when attempting to apply the results of 57 the testing, since the framework can have a major impact on the 58 results. For instance, determining when a network is converged, what 59 parts of the router's operation are considered within the testing, 60 and other such things will have a major impact on what apparent 61 performance routing protocols provide. 63 This document describes in detail the various benefits and pitfalls 64 of tests described in [BENCHMARK]. It also explains how such 65 measurements can be useful for providers and the research community. 67 NOTE: The word convergence within this document refers to single 68 router control plane convergence [TERM]. 70 2. Advantages of Such Measurement 72 o To be able to compare the iterations of a protocol implementa- 73 tion. It is often useful to be able to compare the performance 74 of two iterations of a given implementation of a protocol to 75 determine where improvements have been made and where further 76 improvements can be made. 78 o To understand, given a set of parameters (network conditions), 79 how a particular implementation on a particular device is going 80 to perform. For instance, if you were trying to decide the pro- 81 cessing power (size of device) required in a certain location 82 within a network, you can emulate the conditions which are going 83 to exist at that point in the network and use the test described 84 to measure the performance of several different routers. The 85 results of these tests can provide one possible data point for 86 an intelligent decision. 88 If the device being tested is to be deployed in a running net- 89 work, using routes taken from the network where the equipment is 90 to be deployed rather than some generated topology in these 91 tests will give results which are closer to the real performance 92 of the device. Care should be taken to emulate or take routes 93 from the actual location in the network where the device will be 94 (or would be) deployed. For instance, one set of routes may be 95 taken from an ABR, one set from an area 0 only router, various 96 sets from stub area, another set from various normal areas, etc. 98 o To measure the performance of an OSPF implementation in a wide 99 variety of scenarios. 101 o To be used as parameters in OSPF simulations by researchers. It 102 may some times be required for certain kinds of research to 103 measure the individual delays of each parameter within an OSPF 104 implementation. These delays can be measured using the methods 105 defined in [BENCHMARK]. 107 o To help optimize certain configurable parameters. It may some 108 times be helpful for operators to know the delay required for 109 individual tasks so as to optimize the resource usage in the 110 network i.e. if it is found that the processing time is x 111 seconds on an router, it would be helpful to determine the rate 112 at which to flood LSA's to that router so as to not overload the 113 network. 115 3. Assumptions Made and Limitations of such measurements 117 o The interactions of convergence and forwarding; testing is res- 118 tricted to events occurring within the control plane. Forwarding 119 performance is the primary focus in [INTERCONNECT] and it is 120 expected to be dealt with in work that ensues from [FIB-TERM]. 122 o Duplicate LSAs are Acknowledged Immediately. A few tests rely on 123 the property that duplicate LSA Acknowledgements are not delayed 124 but are done immediately. However if some implementation does 125 not acknowledge duplicate LSAs immediately on receipt, the test- 126 ing methods presented in [BENCHMARK] could give inaccurate meas- 127 urements. 129 o It is assumed that SPF is non-preemptive. If SPF is implemented 130 so that it can (and will be) preempted, the SPF measurements 131 taken in [BENCHMARK] would include the times that the SPF pro- 132 cess is not running ([BENCHMARK] measures the total time taken 133 for SPF to run, not the amount of time that SPF actually spends 134 on the device's processor), thus giving inaccurate measurements. 136 o Some implementations may be multithreaded or use a 137 multiprocess/multirouter model of OSPF. If because of this any 138 of the assumptions taken in measurement are violated in such a 139 model, it could lead to inaccurate measurements. 141 o The measurements resulting from the tests in [BENCHMARK] may not 142 provide the information required to deploy a device in a large 143 scale network. The tests described focus on individual com- 144 ponents of an OSPF implementation's performance, and it may be 145 difficult to combine the measurements in a way which accurately 146 depicts a device's performance in a large scale network. Further 147 research is required in this area. 149 o The measurements described in [BENCHMARK] should be used with 150 great care when comparing two different implementations of OSPF 151 from two different vendors. For instance, there are many other 152 factors than convergence speed that need to be taken into con- 153 sideration when comparing different vendor's products, and it's 154 difficult to align the resources available on one device to the 155 resources available on another device. 157 4. Observations on the Tests Described in [BENCHMARK] 159 Some observations taken while implementing the tests described in 160 [BENCHMARK] are noted in this section. 162 4.1. Measuring the SPF Processing Time Externally 164 The most difficult test to perform is the external measurement of the 165 time required to perform an SPF calculation, since the amount of time 166 between the first LSA which indicates a topology change and the 167 duplicate LSA is critical. If the duplicate LSA is sent too quickly, 168 it may be received before the device under test actually begins run- 169 ning SPF on the network change information. If the delay between the 170 two LSAs is too long, the device under test may finish SPF processing 171 before receiving the duplicate LSA. It is important to closely inves- 172 tigate any delays between the receipt of an LSA and the beginning of 173 an SPF calculation in the device under test; multiple tests with 174 various delays might be required to determine what delay needs to be 175 used to accurately measure the SPF calculation time. 177 Some implementations may force two intervals, the SPF hold time and 178 the SPF delay, between successive SPF calculations. If an SPF hold 179 time exists, it should be subtracted from the total SPF execution 180 time. If an SPF delay exists, it should be noted in the test results. 182 4.2. Noise in the Measurement Device 184 The device on which measurements are taken (not the device under 185 test) also adds noise to the test results, primarily in the form of 186 delay in packet processing and measurement output. The largest source 187 of noise is generally the delay between the receipt of packets by the 188 measuring device and the information about the packet reaching the 189 device's output, where the event can be measured. The following steps 190 may be taken to reduce this sampling noise: 192 o Increasing the number of samples taken will generally improve 193 the tester's ability to determine what is noise, and remove it 194 from the results. 196 o Try to take time-stamp for a packet as early as possible. 197 Depending on the operating system being used on the box, one can 198 instrument the kernel to take the time-stamp when the interrupt 199 is processed. This does not eliminate the noise completely, but 200 at least reduces it. 202 o Keep the measurement box as lightly loaded as possible. 204 o Having an estimate of noise can also be useful. 206 The DUT also adds noise to the measurement. Points (a) and (c) 207 apply to the DUT as well. 209 4.3. Gaining an Understanding of the Implementation Improves Measure- 210 ments 212 While the tester will (generally) not have access to internal infor- 213 mation about the OSPF implementation being tested using [BENCHMARK], 214 the more thorough the tester's knowledge of the implementation is, 215 the more accurate the results of the tests will be. For instance, in 216 some implementations, the installation of routes in local routing 217 tables may occur while the SPF is being calculated, dramatically 218 impacting the time required to calculate the SPF. 220 4.4. Gaining an Understanding of the Tests Improves Measurements 222 One method which can be used to become familiar with the tests 223 described in [BENCHMARK] is to perform the tests on an OSPF implemen- 224 tation for which all the internal details are available, such as 225 [GATED]. While there is no assurance that any two implementations 226 will be similar, this will provide a better understanding of the 227 tests themselves. 229 5. LSA and Destination mix 231 In many OSPF benchmark tests, a generator injecting a number of LSAs 232 is called for. There are several areas in which injected LSAs can be 233 varied in testing: 235 o The number of destinations represented by the injected LSAs 237 Each destination represents a single reachable IP network; these 238 will be leaf nodes on the shortest path tree. The primary impact 239 to performance should be the time required to insert destina- 240 tions in the local routing table and handling the memory 241 required to store the data. 243 o The types of LSAs injected 245 There are several types of LSAs which would be acceptable under 246 different situations; within an area, for instance, type 1, 2, 247 3, 4, and 5 are likely to be received by a router. Within a 248 not-so-stubby area, however, type 7 LSAs would replace the type 249 5 LSAs received. These sorts of characterizations are important 250 to note in any test results. 252 o The Number of LSAs injected 254 Within any injected set of information, the number of each type 255 of LSA injected is also important. This will impact the shortest 256 path algorithms ability to handle large numbers of nodes, large 257 shortest path first trees, etc. 259 o The Order of LSA Injection 261 The order in which LSAs are injected should not favor any given 262 data structure used for storing the LSA database on the device 263 under test. For instance, AS-External LSA's have AS wide flood- 264 ing scope; any Type-5 LSA originated is immediately flooded to 265 all neighbors. However the Type-4 LSA which announces the ASBR 266 as a border router is originated in an area at SPF time (by ABRs 267 on the edge of the area in which the ASBR is). If SPF isn't 268 scheduled immediately on the ABRs originating the type 4 LSA, 269 the type-4 LSA is sent after the type-5 LSA's reach a router in 270 the adjacent area. So routes to the external destinations aren't 271 immediately added to the routers in the other areas. When the 272 routers which already have the type 5's receive the type-4 LSA, 273 all the external routes are added to the tree at the same time. 274 This timing could produce different results than a router 275 receiving a type 4 indicating the presence of a border router, 276 followed by the type 5's originated by that border router. 278 The ordering can be changed in various tests to provide insight 279 on the efficiency of storage within the DUT. Any such changes in 280 ordering should be noted in test results. 282 6. Tree Shape and the SPF Algorithm 284 The complexity of Dijkstra's algorithm depends on the data structure 285 used for storing vertices with their current minimum distances from 286 the source, with the simplest structure being a list of vertices 287 currently reachable from the source. In a simple list of vertices, 288 finding the minimum cost vertex then would take O(size of the list). 289 There will be O(n) such operations if we assume that all the vertices 290 are ultimately reachable from the source. Moreover, after the vertex 291 with min cost is found, the algorithm iterates thru all the edges of 292 the vertex and updates cost of other vertices. With an adjacency list 293 representation, this step when iterated over all the vertices, would 294 take O(E) time, with E being the number of edges in the graph. Thus, 295 overall running time is: 297 O(sum(i:1, n)(size(list at level i) + E). 299 So, everything boils down to the size (list at level i). 301 If the graph is linear: 303 root 304 | 305 1 306 | 307 2 308 | 309 3 310 | 311 4 312 | 313 5 314 | 315 6 317 and source is a vertex on the end, then size(list at level i) = 1 318 for all i. Moreover, E = n - 1. Therefore, running time is O(n). 320 If graph is a balanced binary tree: 322 root 323 / \ 324 1 2 325 / \ / \ 326 3 4 5 6 328 size(list at level i) is a little complicated. First it increases 329 by 1 at each level up to a certain number, and then goes down by 330 1. If we assumed that tree is a complete tree (like the one in the 331 draft) with k levels (1 to k), then size(list) goes on like this: 332 1, 2, 3, 334 Then the number of edges E is still n - 1. It then turns out that 335 the run-time is O(n^2) for such a tree. 337 If graph is a complete graph (fully-connected mesh), then 338 size(list at level i) = n - i. Number of edges E = O(n^2). There- 339 fore, run-time is O(n^2). 341 So, the performance of the shortest path first algorithm used to 342 compute the best paths through the network is dependant o the con- 343 struction of the tree The best practice would be to try and make 344 any emulated network look as much like a real network as possible, 345 especially in the area of the tree depth, the meshiness of the 346 network, the number of stub links versus transit links, and the 347 number of connections and nodes to process at each level within 348 the original tree. 350 7. Topology Generation 352 As the size of networks grows, it becomes more and more difficult to 353 actually create a large scale network on which to test the properties 354 of routing protocols and their implementations. In general, network 355 emulators are used to provide emulated topologies which can be adver- 356 tised to a device with varying conditions. Route generators either 357 tend to be a specialized device, a piece of software which runs on a 358 router, or a process that runs on another operating system, such as 359 Linux or another variant of Unix. 361 Some of the characteristics of this device should be: 363 o The ability to connect to the several devices using both point- 364 to-point and broadcast high speed media. Point-to-point links 365 can be emulated with high speed Ethernet as long as there is no 366 hub or other device in between the DUT and the route generator, 367 and the link is configured as a point-to-point link within OSPF 368 [BROADCAST-P2P]. 370 o The ability to create a set of LSAs which appear to be a logi- 371 cal, realistic topology. For instance, the generator should be 372 able to mix the number of point-to-point and broadcast links 373 within the emulated topology, and should be able to inject vary- 374 ing numbers of externally reachable destinations. 376 o The ability to withdraw and add routing information into and 377 from the emulated topology to emulate links flapping. 379 o The ability to randomly order the LSAs representing the emulated 380 topology as they are advertised. 382 o The ability to log or otherwise measure the time between packets 383 transmitted and received. 385 o The ability to change the rate at which OSPF LSAs are transmit- 386 ted. 388 o The generator and the collector should be fast enough so that 389 they are not bottle necks. The devices should also have a degree 390 of granularity of measurement at least as small as desired from 391 the test results. 393 8. IANA Considerations 395 This document requires no IANA considerations. 397 9. Security Considerations 399 This document does not modify the underlying security considerations 400 in [OSPF]. 402 10. Acknowledgements 404 Thanks to Howard Berkowitz, (hcb@clark.net) and the rest of the BGP 405 benchmarking team for their support and to Kevin 406 Dubray(kdubray@juniper.net) who realized the need of this draft. 408 11. Normative References 410 [BENCHMARK] 411 Manral, V., "Benchmarking Basic OSPF Single Router Control Plane 412 Convergence", draft-bmwg-ospfconv-intraarea-10, June 2004 414 [TERM]Manral, V., "OSPF Convergence Testing Terminology and Con- 415 cepts", draft-bmwg-ospfconv-term-10, May 2004 417 [RFC2119] 418 Bradner, S., "Key words for use in RFCs to Indicate Requirement 419 Levels", BCP 14, RFC 2119, March 1997 421 12. Informative References 423 [INTERCONNECT] 424 Bradner, S., McQuaid, J., "Benchmarking Methodology for Network 425 Interconnect Devices", RFC2544, March 1999. 427 [FIB-TERM] 428 Trotter, G., "Terminology for Forwarding Information Base (FIB) 429 based Router Performance", RFC3222, October 2001. 431 [BROADCAST-P2P] 432 Shen, Naiming, et al., "Point-to-point operation over LAN in 433 link-state routing protocols", draft-ietf-isis-igp-p2p-over- 434 lan-03.txt, August, 2003. 436 [GATED] 437 http://www.gated.org 439 13. Authors' Addresses 440 Vishwas Manral 441 Netplane Systems 442 189 Prashasan Nagar 443 Road number 72 444 Jubilee Hills 445 Hyderabad, India 447 vmanral@netplane.com 449 Russ White 450 Cisco Systems, Inc. 451 7025 Kit Creek Rd. 452 Research Triangle Park, NC 27709 454 riw@cisco.com 456 Aman Shaikh 457 University of California 458 School of Engineering 459 1156 High Street 460 Santa Cruz, CA 95064 462 aman@soe.ucsc.edu 464 Intellectual Property Statement 466 The IETF takes no position regarding the validity or scope of any Intel- 467 lectual Property Rights or other rights that might be claimed to pertain 468 to the implementation or use of the technology described in this docu- 469 ment or the extent to which any license under such rights might or might 470 not be available; nor does it represent that it has made any independent 471 effort to identify any such rights. Information on the procedures with 472 respect to rights in RFC documents can be found in BCP 78 and BCP 79. 474 Copies of IPR disclosures made to the IETF Secretariat and any 475 assurances of licenses to be made available, or the result of an attempt 476 made to obtain a general license or permission for the use of such 477 proprietary rights by implementers or users of this specification can be 478 obtained from the IETF on-line IPR repository at 479 http://www.ietf.org/ipr. 481 The IETF invites any interested party to bring to its attention any 482 copyrights, patents or patent applications, or other proprietary rights 483 that may cover technology that may be required to implement this stan- 484 dard. Please address the information to the IETF at ietf-ipr@ietf.org. 486 Disclaimer of Warranty 488 This document and the information contained herein are provided on an 489 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR 490 IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 491 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 492 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMA- 493 TION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF 494 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 496 Copyright Statement 498 Copyright (C) The Internet Society (2004). This document is subject to 499 the rights, licenses and restrictions contained in BCP 78, and except as 500 set forth therein, the authors retain all their rights.