idnits 2.17.1 draft-kini-mpls-spring-entropy-label-03.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 3, 2015) is 3332 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'I-D.filsfils-spring-segment-routing-use-cases' is defined on line 451, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-isis-segment-routing-extensions' is defined on line 459, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-ospf-segment-routing-extensions' is defined on line 465, but no explicit reference was found in the text == Outdated reference: A later version (-02) exists of draft-xu-isis-mpls-elc-01 == Outdated reference: A later version (-25) exists of draft-ietf-isis-segment-routing-extensions-03 == Outdated reference: A later version (-27) exists of draft-ietf-ospf-segment-routing-extensions-04 Summary: 0 errors (**), 0 flaws (~~), 7 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group S. Kini, Ed. 3 Internet-Draft Ericsson 4 Intended status: Informational K. Kompella 5 Expires: September 4, 2015 Juniper 6 S. Sivabalan 7 Cisco 8 S. Litkowski 9 Orange 10 R. Shakir 11 B.T. 12 X. Xu 13 Huawei 14 W. Hendrickx 15 Alcatel-Lucent 16 J. Tantsura 17 Ericsson 18 March 3, 2015 20 Entropy labels for source routed stacked tunnels 21 draft-kini-mpls-spring-entropy-label-03 23 Abstract 25 Source routed tunnel stacking is a technique that can be leveraged to 26 provide a method to steer a packet through a controlled set of 27 segments. This can be applied to the Multi Protocol Label Switching 28 (MPLS) data plane. Entropy label (EL) is a technique used in MPLS to 29 improve load balancing. This document examines and describes how ELs 30 are to be applied to source routed stacked tunnels. 32 Status of This Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at http://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on September 4, 2015. 49 Copyright Notice 51 Copyright (c) 2015 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents 56 (http://trustee.ietf.org/license-info) in effect on the date of 57 publication of this document. Please review these documents 58 carefully, as they describe your rights and restrictions with respect 59 to this document. Code Components extracted from this document must 60 include Simplified BSD License text as described in Section 4.e of 61 the Trust Legal Provisions and are provided without warranty as 62 described in the Simplified BSD License. 64 Table of Contents 66 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 67 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3 68 2. Abbreviations and Terminology . . . . . . . . . . . . . . . . 3 69 3. Use-case requiring multipath load balancing in source stacked 70 tunnels . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 71 4. Recommended EL solution for SPRING . . . . . . . . . . . . . 5 72 5. Options considered . . . . . . . . . . . . . . . . . . . . . 6 73 5.1. Single EL at the bottom of the stack of tunnels . . . . . 6 74 5.2. An EL per tunnel in the stack . . . . . . . . . . . . . . 7 75 5.3. A re-usable EL for a stack of tunnels . . . . . . . . . . 7 76 5.3.1. EL at top of stack . . . . . . . . . . . . . . . . . 8 77 5.4. ELs at readable label stack depths . . . . . . . . . . . 8 78 6. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 9 79 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 80 8. Security Considerations . . . . . . . . . . . . . . . . . . . 9 81 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 9 82 9.1. Normative References . . . . . . . . . . . . . . . . . . 9 83 9.2. Informative References . . . . . . . . . . . . . . . . . 10 84 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 11 86 1. Introduction 88 The source routed stacked tunnels paradigm is leveraged by techniques 89 such as Segment Routing (SR) [I-D.filsfils-spring-segment-routing] to 90 steer a packet through a set of segments. This can be directly 91 applied to the MPLS data plane, but it has implications on label 92 stack depth. 94 Clarifying statements on label stack depth have been provided in 95 [RFC7325] but they do not address the case of source routed stacked 96 MPLS tunnels as described in [I-D.gredler-spring-mpls] or 98 [I-D.filsfils-spring-segment-routing] where deeper label stacks are 99 more prevalent. 101 Entropy label (EL) [RFC6790] is a technique used in the MPLS data 102 plane to provide entropy for load balancing. When using LSP 103 hierarchies there are implications on how [RFC6790] should be 104 applied. One such issue is addressed by 105 [I-D.ravisingh-mpls-el-for-seamless-mpls] but that is when different 106 levels of the hierarchy are created at different LSRs. The current 107 document addresses the case where the hierarchy is created at a 108 single LSR as required by source stacked tunnels. 110 A use-case requiring load balancing with source stacked tunnels is 111 given in Section 3. A recommended solution is described in Section 4 112 keeping in consideration the limitations of implementations when 113 applying [RFC6790] to deeper label stacks. Options that were 114 considered to arrive at the recommended solution are documented for 115 historical purposes in Section 5. 117 1.1. Requirements Language 119 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 120 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 121 document are to be interpreted as described in RFC 2119 [RFC2119]. 123 Although this document is not a protocol specification, the use of 124 this language clarifies the instructions to protocol designers 125 producing solutions that satisfy the requirements set out in this 126 document. 128 2. Abbreviations and Terminology 130 EL - Entropy Label 132 ELI - Entropy Label Identifier 134 ELC - Entropy Label Capability 136 SR - Segment Routing 138 ECMP - Equal Cost Multi Paths 140 MPLS - Multiprotocol Label Switching 142 SID - Segment Identifier 144 RLD - Readable Label Depth 145 OAM - Operation, Administration and Maintenance 147 3. Use-case requiring multipath load balancing in source stacked 148 tunnels 150 +------+ 151 | | 152 +-------| P3 |-----+ 153 | +-----| |---+ | 154 L3| |L4 +------+ L1| |L2 +----+ 155 | | | | +--| P4 |--+ 156 +-----+ +-----+ +-----+ | +----+ | +-----+ 157 | S |-----| P1 |------------| P2 |--+ +--| D | 158 | | | | | |--+ +--| | 159 +-----+ +-----+ +-----+ | +----+ | +-----+ 160 +--| P5 |--+ 161 +----+ 163 S=Source LSR, D=Destination LSR, P1,P2,P3,P4,P5=Transit LSRs, 164 L1,L2,L3,L4=Links 166 Figure 1: Traffic engineering use-case 168 Traffic-engineering (TE) is one of the applications of MPLS and is 169 also a requirement for source stacked tunnels. Consider the topology 170 shown in Figure 1. Lets say the LSR P1 has a limitation that it can 171 only look four labels deep in the stack to do multipath decisions. 172 All other transit LSRs in the figure can read deep label stacks and 173 the LSR S can insert as many pairs as needed. The LSR S 174 requires data to be sent to LSR D along a traffic-engineered path 175 that goes over the link L1. Good load balancing is also required 176 across equal cost paths (including parallel links). To engineer 177 traffic along a path that takes link L1, the label stack that LSR S 178 creates consists of a label to the node SID of LSR P3, stacked over 179 the label for the adjacency SID of link L1 and that in turn is 180 stacked over the label to the node SID of LSR D. For simplicity lets 181 assume that all LSRs use the same label space for source stacked 182 tunnels. Lets L_N-P denote the label to be used to reach the node 183 SID of LSR P. Let L_A-Ln denote the label used for the adjacency SID 184 for link Ln. The LSR S must use the label stack for traffic-engineering. However to achieve good load 186 balancing over the equal cost paths P2-P4-D, P2-P5-D and the parallel 187 links L3, L4, a mechanism such as Entropy labels [RFC6790] should be 188 adapted for source stacked tunnels. Multiple ways to apply entropy 189 labels were considered and are documented in Section 5 along with 190 their tradeoffs. A recommended solution is described in Section 4. 192 4. Recommended EL solution for SPRING 194 The solution described in this section follows [RFC6790]. 196 An LSR may have a limitation in its ability to read and process the 197 label stack in order to do multipath load balancing. This limitation 198 expressed in terms of the number of label stack entries that the LSR 199 can read is henceforth referred to as the Readable Label Depth (RLD) 200 capability of that LSR. If an EL does not occur within the RLD of an 201 LSR in the label stack of the MPLS packet that it receives, then it 202 would lead to poor load balancing at that LSR. The RLD of an LSR is 203 a characteristic of the forwarding plane of that LSR's implementation 204 and determining it is outside the scope of this document. 206 In order for the EL to occur within the RLD of LSRs along the path 207 corresponding to a label stack, multiple pairs MAY be 208 inserted in the label stack as long as the tunnel's label below which 209 they are inserted are advertised with entropy label capability 210 enabled. The LSR that inserts pairs MAY have limitations 211 on the number of such pairs that it can insert and also the depth at 212 which it can insert them. If due to any limitation, the inserted ELs 213 are at positions such that an LSR along the path receives an MPLS 214 packet without an EL in the label stack within that LSR's RLD, then 215 the load balancing performed by that LSR would be poor. Special 216 attention should be paid when a forwarding adjacency LSP (FA-LSP) 217 [RFC4206] is used as a link along the path of a source stacked LSP, 218 since the labels of the FA-LSP would additionally count towards the 219 depth of the label stack when calculating the appropriate positions 220 to insert the ELs. The recommendations for inserting pairs 221 are: 223 o An LSR that is limited in the number of pairs that it 224 can insert SHOULD insert such pairs deeper in the stack. 226 o An LSR SHOULD try to insert pairs at positions so that 227 for the maximum number of transit LSRs, the EL occurs within the 228 RLD of the incoming packet to that LSR. 230 o An LSR SHOULD try to insert the minimum number of such pairs while 231 trying to satisfy the above criteria. 233 A sample algorithm to insert ELs is shown below. Implementations can 234 choose any algorithm as long as it follows the above recommendations. 236 Initialize the current EL insertion point to the 237 bottommost label in the stack that is EL-capable 238 while (local-node can push more pairs OR 239 insertion point is not above label stack) { 240 insert an pair below current insertion point 241 move new insertion point up from current insertion point until 242 ((last inserted EL is below the RLD) AND (RLD > 2) 243 AND 244 (new insertion point is EL-capable)) 245 set current insertion point to new insertion point 246 } 248 Figure 2: Algorithm to insert pairs in a label stack 250 When this algorithm is applied to the example described in Section 3 251 it will result in ELs being inserted in two positions, one below the 252 label L_N-D and another below L_N-P3. Thus the resulting label stack 253 would be 255 The RLD can be advertised via protocols and those extensions would be 256 described in separate documents [I-D.xu-isis-mpls-elc] and 257 [I-D.xu-ospf-mpls-elc]. 259 The recommendations above are not expected to bring any additional 260 OAM considerations beyond those described in section 6 of [RFC6790]. 261 However, the OAM requirements and solutions for source stacked 262 tunnels are still under discussion and future revisions of this 263 document will address those if needed. 265 5. Options considered 267 5.1. Single EL at the bottom of the stack of tunnels 269 In this option a single EL is used for the entire label stack. The 270 source LSR S encodes the entropy label (EL) below the labels of all 271 the stacked tunnels. In the example described in Section 3 it will 272 result in the label stack at LSR S to look like . Note that the notation in 274 [RFC6790] is used to describe the label stack. An issue with this 275 approach is that as the label stack grows due an increase in the 276 number of SIDs, the EL goes correspondingly deeper in the label 277 stack. Hence transit LSRs have to access a larger number of bytes in 278 the packet header when making forwarding decisions. In the example 279 described in Section 3 the LSR P1 would poorly load-balance traffic 280 on the parallel links L3, L4 since the EL is below the RLD of the 281 packet received by P1. A load balanced network design using this 282 approach must ensure that all intermediate LSRs have the capability 283 to traverse the maximum label stack depth as required for that 284 application that uses source routed stacking. 286 In the case where the hardware is capable of pushing a single pair at any depth, this option is the same as the recommended 288 solution in Section 4. 290 This option was discounted since there exist a number of hardware 291 implementations which have a low maximum readable label depth. 292 Choosing this option can lead to a loss of load-balancing using EL in 293 a significant part of the network but that is a critical requirement 294 in a service provider network. 296 5.2. An EL per tunnel in the stack 298 In this option each tunnel in the stack can be given its own EL. The 299 source LSR pushes an before pushing a tunnel label when 300 load balancing is required to direct traffic on that tunnel. In the 301 example described in Section 3, the source LSR S encoded label stack 302 would be where all the ELs 303 can be the same. Accessing the EL at an intermediate LSR is 304 independent of the depth of the label stack and hence independent of 305 the specific application that uses source stacking on that network. 306 A drawback is that the depth of the label stack grows significantly, 307 almost 3 times as the number of labels in the label stack. The 308 network design should ensure that source LSRs should have the 309 capability to push such a deep label stack. Also, the bandwidth 310 overhead and potential MTU issues of deep label stacks should be 311 accounted for in the network design. 313 In the case where the RLD is the minimum value (3) for all LSRs, all 314 LSRs are EL capable and the LSR that is inserting pairs has 315 no limit on how many it can insert then this option is the same as 316 the recommended solution in Section 4. 318 This option was discounted due to the existence of hardware 319 implementations that can push a limited number of labels on the label 320 stack. Choosing this option would result in a hardware requirement 321 to push two additional labels per tunnel label. Hence it would 322 restrict the number of tunnels that can form a LSP and constrain the 323 types of LSPs that can be created. This was considered unacceptable. 325 5.3. A re-usable EL for a stack of tunnels 327 In this option an LSR that terminates a tunnel re-uses the EL of the 328 terminated tunnel for the next inner tunnel. It does this by storing 329 the EL from the outer tunnel when that tunnel is terminated and re- 330 inserting it below the next inner tunnel label during the label swap 331 operation. The LSR that stacks tunnels SHOULD insert an EL below the 332 outermost tunnel. It SHOULD NOT insert ELs for any inner tunnels. 333 Also, the penultimate hop LSR of a segment MUST NOT pop the ELI and 334 EL even though they are exposed as the top labels since the 335 terminating LSR of that segment would re-use the EL for the next 336 segment. 338 In Section 3 above, the source LSR S encoded label stack would be 339 . At P1 the outgoing label stack 340 would be after it has load balanced 341 to one of the links L3 or L4. At P3 the outgoing label stack would 342 be . At P2 the outgoing label stack would be and it would load balance to one of the nexthop LSRs P4 or 344 P5. Accessing the EL at an intermediate LSR (e.g. P1) is 345 independent of the depth of the label stack and hence independent of 346 the specific use-case to which the stacked tunnels are applied. 348 This option was discounted due to the significant change in label 349 swap operations that would be required for existing hardware. 351 5.3.1. EL at top of stack 353 A slight variant of the re-usable EL option is to keep the EL at the 354 top of the stack rather than below the tunnel label. In this case 355 each LSR that is not terminating a segment should continue to keep 356 the received EL at the top of the stack when forwarding the packet 357 along the segment. An LSR that terminates a segment should use the 358 EL from the terminated segment at the top of the stack when 359 forwarding onto the next segment. 361 This option was discounted due to the significant change in label 362 swap operations that would be required for existing hardware. 364 5.4. ELs at readable label stack depths 366 In this option the source LSR inserts ELs for tunnels in the label 367 stack at depths such that each LSR along the path that must load 368 balance is able to access at least one EL. Note that the source LSR 369 may have to insert multiple ELs in the label stack at different 370 depths for this to work since intermediate LSRs may have differing 371 capabilities in accessing the depth of a label stack. The label 372 stack depth access value of intermediate LSRs must be known to create 373 such a label stack. How this value is determined is outside the 374 scope of this document. This value can be advertised using a 375 protocol such as an IGP. For the same Section 3 above, if LSR P1 376 needs to have the EL within a depth of 4, then the source LSR S 377 encoded label stack would be where all the ELs would typically have the same value. 380 In the case where the RLD has different values along the path and the 381 LSR that is inserting pairs has no limit on how many pairs 382 it can insert, and it knows the appropriate positions in the stack 383 where they should be inserted, then this option is the same as the 384 recommended solution in Section 4. 386 A variant of this solution was selected which balances the number of 387 labels that need to be pushed against the requirement for entropy. 389 6. Acknowledgements 391 The authors would like to thank John Drake, Loa Andersson, Curtis 392 Villamizar, Greg Mirsky, Markus Jork, Kamran Raza and Nobo Akiya for 393 their review comments and suggestions. 395 7. IANA Considerations 397 This memo includes no request to IANA. 399 8. Security Considerations 401 This document does not introduce any new security considerations 402 beyond those already listed in [RFC6790]. 404 9. References 406 9.1. Normative References 408 [I-D.filsfils-spring-segment-routing] 409 Filsfils, C., Previdi, S., Bashandy, A., Decraene, B., 410 Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., 411 Ytti, S., Henderickx, W., Tantsura, J., and E. Crabbe, 412 "Segment Routing Architecture", draft-filsfils-spring- 413 segment-routing-04 (work in progress), July 2014. 415 [I-D.gredler-spring-mpls] 416 Gredler, H., Rekhter, Y., Jalil, L., Kini, S., and X. Xu, 417 "Supporting Source/Explicitly Routed Tunnels via Stacked 418 LSPs", draft-gredler-spring-mpls-06 (work in progress), 419 May 2014. 421 [I-D.ravisingh-mpls-el-for-seamless-mpls] 422 Singh, R., Shen, Y., and J. Drake, "Entropy label for 423 seamless MPLS", draft-ravisingh-mpls-el-for-seamless- 424 mpls-04 (work in progress), October 2014. 426 [I-D.xu-isis-mpls-elc] 427 Xu, X., Kini, S., Sivabalan, S., Filsfils, C., and S. 428 Litkowski, "Signaling Entropy Label Capability Using IS- 429 IS", draft-xu-isis-mpls-elc-01 (work in progress), 430 September 2014. 432 [I-D.xu-ospf-mpls-elc] 433 Xu, X., Kini, S., Sivabalan, S., Filsfils, C., and S. 434 Litkowski, "Signaling Entropy Label Capability Using 435 OSPF", draft-xu-ospf-mpls-elc-01 (work in progress), 436 October 2014. 438 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 439 Requirement Levels", BCP 14, RFC 2119, March 1997. 441 [RFC4206] Kompella, K. and Y. Rekhter, "Label Switched Paths (LSP) 442 Hierarchy with Generalized Multi-Protocol Label Switching 443 (GMPLS) Traffic Engineering (TE)", RFC 4206, October 2005. 445 [RFC6790] Kompella, K., Drake, J., Amante, S., Henderickx, W., and 446 L. Yong, "The Use of Entropy Labels in MPLS Forwarding", 447 RFC 6790, November 2012. 449 9.2. Informative References 451 [I-D.filsfils-spring-segment-routing-use-cases] 452 Filsfils, C., Francois, P., Previdi, S., Decraene, B., 453 Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., 454 Ytti, S., Henderickx, W., Tantsura, J., Kini, S., and E. 455 Crabbe, "Segment Routing Use Cases", draft-filsfils- 456 spring-segment-routing-use-cases-01 (work in progress), 457 October 2014. 459 [I-D.ietf-isis-segment-routing-extensions] 460 Previdi, S., Filsfils, C., Bashandy, A., Gredler, H., 461 Litkowski, S., Decraene, B., and J. Tantsura, "IS-IS 462 Extensions for Segment Routing", draft-ietf-isis-segment- 463 routing-extensions-03 (work in progress), October 2014. 465 [I-D.ietf-ospf-segment-routing-extensions] 466 Psenak, P., Previdi, S., Filsfils, C., Gredler, H., 467 Shakir, R., Henderickx, W., and J. Tantsura, "OSPF 468 Extensions for Segment Routing", draft-ietf-ospf-segment- 469 routing-extensions-04 (work in progress), February 2015. 471 [RFC7325] Villamizar, C., Kompella, K., Amante, S., Malis, A., and 472 C. Pignataro, "MPLS Forwarding Compliance and Performance 473 Requirements", RFC 7325, August 2014. 475 Authors' Addresses 477 Sriganesh Kini (editor) 478 Ericsson 480 Email: sriganesh.kini@ericsson.com 482 Kireeti Kompella 483 Juniper 485 Email: kireeti@juniper.net 487 Siva Sivabalan 488 Cisco 490 Email: msiva@cisco.com 492 Stephane Litkowski 493 Orange 495 Email: stephane.litkowski@orange.com 497 Rob Shakir 498 B.T. 500 Email: rob.shakir@bt.com 502 Xiaohu Xu 503 Huawei 505 Email: xuxiaohu@huawei.com 507 Wim Hendrickx 508 Alcatel-Lucent 510 Email: wim.henderickx@alcatel-lucent.com 512 Jeff Tantsura 513 Ericsson 515 Email: jeff.tantsura@ericsson.com