idnits 2.17.1 draft-ietf-teas-pce-native-ip-09.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 (June 8, 2020) is 1416 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 7752 (Obsoleted by RFC 9552) == Outdated reference: A later version (-30) exists of draft-ietf-pce-pcep-extension-native-ip-05 Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 TEAS Working Group A. Wang 3 Internet-Draft China Telecom 4 Intended status: Experimental B. Khasanov 5 Expires: December 10, 2020 Huawei Technologies 6 Q. Zhao 7 Etheric Networks 8 H. Chen 9 Futurewei 10 June 8, 2020 12 PCE in Native IP Network 13 draft-ietf-teas-pce-native-ip-09 15 Abstract 17 This document defines the framework for traffic engineering within 18 native IP network, using multiple BGP sessions strategy and PCE 19 -based central control architecture. The procedures described in 20 this document are experimental. The experiment is intended to enable 21 research for the usage of PCE in native IP scenarios. For this 22 purpose, this document describe the Central Control Dynamic Routing 23 (CCDR) framework and the PCEP extension is specified in draft ietf- 24 pce-pcep-extension-native-ip. 26 Status of This Memo 28 This Internet-Draft is submitted in full conformance with the 29 provisions of BCP 78 and BCP 79. 31 Internet-Drafts are working documents of the Internet Engineering 32 Task Force (IETF). Note that other groups may also distribute 33 working documents as Internet-Drafts. The list of current Internet- 34 Drafts is at https://datatracker.ietf.org/drafts/current/. 36 Internet-Drafts are draft documents valid for a maximum of six months 37 and may be updated, replaced, or obsoleted by other documents at any 38 time. It is inappropriate to use Internet-Drafts as reference 39 material or to cite them other than as "work in progress." 41 This Internet-Draft will expire on December 10, 2020. 43 Copyright Notice 45 Copyright (c) 2020 IETF Trust and the persons identified as the 46 document authors. All rights reserved. 48 This document is subject to BCP 78 and the IETF Trust's Legal 49 Provisions Relating to IETF Documents 50 (https://trustee.ietf.org/license-info) in effect on the date of 51 publication of this document. Please review these documents 52 carefully, as they describe your rights and restrictions with respect 53 to this document. Code Components extracted from this document must 54 include Simplified BSD License text as described in Section 4.e of 55 the Trust Legal Provisions and are provided without warranty as 56 described in the Simplified BSD License. 58 Table of Contents 60 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 61 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 62 3. CCDR Framework in Simple Topology . . . . . . . . . . . . . . 4 63 4. CCDR Framework in Large Scale Topology . . . . . . . . . . . 5 64 5. CCDR Multiple BGP Sessions Strategy . . . . . . . . . . . . . 6 65 6. PCEP Extension for Key Parameters Delivery . . . . . . . . . 8 66 7. Deployment Consideration . . . . . . . . . . . . . . . . . . 9 67 7.1. Scalability . . . . . . . . . . . . . . . . . . . . . . . 9 68 7.2. High Availability . . . . . . . . . . . . . . . . . . . . 10 69 7.3. Incremental deployment . . . . . . . . . . . . . . . . . 10 70 8. Security Considerations . . . . . . . . . . . . . . . . . . . 10 71 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 11 72 10. Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . 11 73 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 11 74 11.1. Normative References . . . . . . . . . . . . . . . . . . 11 75 11.2. Informative References . . . . . . . . . . . . . . . . . 12 76 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 12 78 1. Introduction 80 [RFC8735] describes the scenarios and simulation results for traffic 81 engineering in the native IP network to provide End-to-End (E2E) 82 performance assurance and QoS using PCE based central control, 83 referred to as Centralized Control Dynamic Routing (CCDR). Based on 84 the various scenarios and analysis as per [RFC8735], the solution for 85 traffic engineering in native IP network should have the following 86 criteria: 88 o Support native IPv4 and IPv6 traffic simultaneously, no complex 89 signaling procedures among network nodes like MPLS-TE and MPLS 90 data plane. 92 o Same deployment guidline for intra-domain and inter-domain 93 scenarios. 95 o Achieve End to End traffic assurance, determined QoS behavior. 97 o No upgrade to forwarding behavior of the router. 99 o Can exploit the power of centrally control and flexibility/ 100 robustness of distributed control protocol. 102 o Coping with the differentiation requirements for large amount 103 traffic and prefixes. 105 o Adjust the optimal path dynamically upon the change of network 106 status. No physical links resources planning in advance. 108 Stateful PCE [RFC8231] specifies a set of extensions to PCEP to 109 enable stateful control of paths such as MPLS-TE LSPs between and 110 across PCEP sessions in compliance with [RFC4657]. It includes 111 mechanisms to affect state synchronization between PCCs and PCEs, 112 delegation of control of LSPs to PCEs, and PCE control of timing and 113 sequence of path computations within and across PCEP sessions. 114 Furthermore, [RFC8281] specifies a mechanism to dynamically 115 instantiate LSPs on a PCC based on the requests from a stateful PCE 116 or a controller using stateful PCE. [RFC8283] introduces the 117 architecture for PCE as a central controller as an extension of the 118 architecture described in [RFC4655] and assumes the continued use of 119 PCEP as the protocol used between PCE and PCC.[RFC8283] further 120 examines the motivations and applicability for PCEP as a Southbound 121 Interface (SBI), and introduces the implications for the protocol. 123 This document defines the framework for traffic engineering within 124 native IP network, using multiple BGP session strategy, to meet the 125 above requirements in dynamical and centrally control mode. The 126 framework is referred as CCDR framework. It depends on the central 127 control (PCE) element to compute the optimal path for selected 128 traffic, and utilizes the dynamic routing behavior of traditional 129 IGP/BGP protocols to forward such traffic. 131 The control messages between PCE and underlying network node are 132 transmitted via Path Computation Element Communications Protocol 133 (PCEP) protocol. The related PCEP extensions are provided in draft 134 [I-D.ietf-pce-pcep-extension-native-ip]. 136 2. Terminology 138 This document uses the following terms defined in [RFC5440]: PCE, 139 PCEP 141 The following terms are used in this document: 143 o CCDR: Central Control Dynamic Routing 144 o E2E: End to End 146 o ECMP: Equal Cost Multi Path 148 o RR: Route Reflector 150 o SDN: Software Defined Network 152 3. CCDR Framework in Simple Topology 154 Figure 1 illustrates the CCDR framework for traffic engineering in 155 simple topology. The topology is comprised by four devices which are 156 SW1, SW2, R1, R2. There are multiple physical links between R1 and 157 R2. Traffic between prefix PF11(on SW1) and prefix PF21(on SW2) is 158 normal traffic, traffic between prefix PF12(on SW1) and prefix 159 PF22(on SW2) is priority traffic that should be treated differently. 161 In Intra-AS scenario, IGP and BGP are deployed between R1 and R2. In 162 inter-AS scenario, only native BGP protocol is deployed. The traffic 163 between each address pair may change in real time and the 164 corresponding source/destination addresses of the traffic may also 165 change dynamically. 167 The key ideas of the CCDR framework for this simple topology are the 168 followings: 170 o Build two BGP sessions between R1 and R2, via the different 171 loopback addresses on these routers. 173 o Send different prefixes via the established BGP sessions. For 174 example, PF11/PF21 via the BGP session 1 and PF12/PF22 via the BGP 175 session 2. 177 o Set the explicit peer route on R1 and R2 respectively for BGP next 178 hop to different physical link addresses between R1 and R2. Such 179 explicit peer route can be set in the format of static route to 180 BGP peer address, which is different from the route learned from 181 the IGP protocol. 183 After the above actions, the bi-direction traffic between the PF11 184 and PF21, and the bi-direction traffic between PF12 and PF22 will go 185 through different physical links between R1 and R2, each set of 186 traffic pass through different dedicated physical links. 188 If there is more traffic between PF12 and PF22 that needs to be 189 assured , one can add more physical links between R1 and R2 to reach 190 the the next hop for BGP session 2. In this cases the prefixes that 191 advertised by the BGP peers need not be changed. 193 If, for example, there is bi-direction traffic from another address 194 pair that needs to be assured (for example prefix PF13/PF23), and the 195 total volume of assured traffic does not exceed the capacity of the 196 previously provisioned physical links, one need only to advertise the 197 newly added source/destination prefixes via the BGP session 2. The 198 bi-direction traffic between PF13/PF23 will go through the assigned 199 dedicated physical links as the traffic between PF12/PF22. 201 Such decouple philosophy gives network operator flexible control 202 capability on the network traffic, achieve the determined QoS 203 assurance effect to meet the application's requirement. No complex 204 signaling procedures like MPLS-TE are introduced, the router needs 205 only support native IP and multiple BGP sessions setup via different 206 loopback addresses. 208 +-----+ 209 +----------+ PCE +--------+ 210 | +-----+ | 211 | | 212 | BGP Session 1(lo11/lo21)| 213 +-------------------------+ 214 | | 215 | BGP Session 2(lo12/lo22)| 216 +-------------------------+ 217 PF12 | | PF22 218 PF11 | | PF21 219 +---+ +-----+-----+ +-----+-----+ +---+ 220 |SW1+---------+(lo11/lo12)+-------------+(lo21/lo22)+--------------+SW2| 221 +---+ | R1 +-------------+ R2 | +---+ 222 +-----------+ +-----------+ 224 Figure 1: CCDR framework in simple topology 226 4. CCDR Framework in Large Scale Topology 228 When the assured traffic spans across the large scale network, as 229 that illustrated in Figure 2, the multiple BGP sessions cannot be 230 established hop by hop, especially for the iBGP within one AS. 232 For such scenario, we should consider to use the Route Reflector (RR) 233 [RFC4456] to achieve the similar effect. Every edge router will 234 establish two BGP sessions with the RR via different loopback 235 addresses respectively. The other steps for traffic differentiation 236 are same as that described in the CCDR framework for simple topology. 238 As shown in Figure 2, if we select R3 as the RR, every edge router(R1 239 and R7 in this example) will build two BGP session with the RR. If 240 the PCE selects the dedicated path as R1-R2-R4-R7, then the operator 241 should set the explicit peer routes via PCEP protocol on these 242 routers respectively, pointing to the BGP next hop (loopback 243 addresses of R1 and R7, which are used to send the prefix of the 244 assured traffic) to the selected forwarding address. 246 +-----+ 247 +----------------+ PCE +------------------+ 248 | +--+--+ | 249 | | | 250 | | | 251 | ++-+ | 252 +------------------+R3+-------------------+ 253 PF12 | +--+ | PF22 254 PF11 | | PF21 255 +---+ ++-+ +--+ +--+ +-++ +---+ 256 |SW1+-------+R1+----------+R5+----------+R6+---------+R7+--------+SW2| 257 +---+ ++-+ +--+ +--+ +-++ +---+ 258 | | 259 | | 260 | +--+ +--+ | 261 +------------+R2+----------+R4+-----------+ 262 +--+ +--+ 263 Figure 2: CCDR framework in large scale network 265 5. CCDR Multiple BGP Sessions Strategy 267 In general situation, different applications may require different 268 QoS criteria, which may include: 270 o Traffic that requires low latency and is not sensitive to packet 271 loss. 273 o Traffic that requires low packet loss and can endure higher 274 latency. 276 o Traffic that requires low jitter. 278 These different traffic requirements can be summarized in the 279 following table: 281 +----------------+-------------+---------------+-----------------+ 282 | Prefix Set No. | Latency | Packet Loss | Jitter | 283 +----------------+-------------+---------------+-----------------+ 284 | 1 | Low | Normal | Don't care | 285 +----------------+-------------+---------------+-----------------+ 286 | 2 | Normal | Low | Dont't care | 287 +----------------+-------------+---------------+-----------------+ 288 | 3 | Normal | Normal | Low | 289 +----------------+-------------+---------------+-----------------+ 290 Table 1. Traffic Requirement Criteria 292 For Prefix Set No.1, we can select the shortest distance path to 293 carry the traffic; for Prefix Set No.2, we can select the path that 294 is comprised by under loading links from end to end; For Prefix Set 295 No.3, we can let all assured traffic pass the determined single path, 296 no Equal Cost Multipath (ECMP) distribution on the parallel links is 297 desired. 299 It is almost impossible to provide an End-to-End (E2E) path 300 efficiently with latency, jitter, packet loss constraints to meet the 301 above requirements in large scale IP-based network via the 302 distributed routing protocol, but these requirements can be solved 303 with the assistance of PCE, as that described in [RFC4655] and 304 [RFC8283] because the PCE has the overall network view, can collect 305 real network topology and network performance information about the 306 underlying network, select the appropriate path to meet various 307 network performance requirements of different traffics. 309 The framework to implement the CCDR Multiple BGP sessions strategy 310 are the followings. Here PCE is the main component of the Software 311 Definition Network (SDN) controller and is responsible for optimal 312 path computation for priority traffic. 314 o SDN controller gets topology via BGP-LS[RFC7752] and link 315 utilization information via existing Network Monitor System (NMS) 316 from the underlying network. 318 o PCE calculates the appropriate path upon application's 319 requirements, sends the key parameters to edge/RR routers(R1, R7 320 and R3 in Fig.3) to establish multiple BGP sessions and advertises 321 different prefixes via them. The loopback addresses used for BGP 322 sessions should be planned in advance and distributed in the 323 domain. 325 o PCE sends the route information to the routers (R1,R2,R4,R7 in 326 Fig.3) on forwarding path via PCEP 327 [I-D.ietf-pce-pcep-extension-native-ip], to build the path to the 328 BGP next-hop of the advertised prefixes. 330 o If the assured traffic prefixes were changed but the total volume 331 of assured traffic does not exceed the physical capacity of the 332 previous E2E path, PCE needs only change the prefixed advertised 333 via the edge routers (R1,R7 in Fig.3). 335 o If the volume of assured traffic exceeds the capacity of previous 336 calculated path, PCE can recalculate and add the appropriate paths 337 to accommodate the exceeding traffic. After that, PCE needs to 338 update on-path routers to build the forwarding path hop by hop. 340 +------------+ 341 | Application| 342 +------+-----+ 343 | 344 +--------+---------+ 345 +----------+SDN Controller/PCE+-----------+ 346 | +--------^---------+ | 347 | | | 348 | | | 349 PCEP | BGP-LS|PCEP | PCEP 350 | | | 351 | +v-+ | 352 +------------------+R3+-------------------+ 353 PF12 | +--+ | PF22 354 PF11 | | PF21 355 +---+ +v-+ +--+ +--+ +-v+ +---+ 356 |SW1+-------+R1+----------+R5+----------+R6+---------+R7+--------+SW2| 357 +---+ ++-+ +--+ +--+ +-++ +---+ 358 | | 359 | | 360 | +--+ +--+ | 361 +------------+R2+----------+R4+-----------+ 363 Figure 3: CCDR framework for Multi-BGP deployment 365 6. PCEP Extension for Key Parameters Delivery 367 The PCEP protocol needs to be extended to transfer the following key 368 parameters: 370 o Peer addresses pair that is used to build the BGP session 372 o Advertised prefixes and their associated BGP session. 374 o Explicit route information to BGP next hop of advertised prefixes. 376 Once the router receives such information, it should establish the 377 BGP session with the peer appointed in the PCEP message, advertise 378 the prefixes that contained in the corresponding PCEP message, and 379 build the end to end dedicated path hop by hop. 381 The dedicated path is preferred by making sure that the explicit 382 route created by PCE has the higher priority (lower route preference) 383 than the route information created by any other protocols (including 384 the route manually configured). 386 All above dynamically created states (BGP sessions, Prefix advertised 387 prefix, Explicit route) will be cleared on the expiration of state 388 timeout interval which is based on the existing Stateful PCE 389 [RFC8231] and PCECC [RFC8283] mechanism. 391 Details of communications between PCEP and BGP subsystems in router's 392 control plane are out of scope of this draft and will be described in 393 separate draft [I-D.ietf-pce-pcep-extension-native-ip] . 395 The reason that we select PCEP as the southbound protocol instead of 396 OpenFlow, is that PCEP is suitable for the changes in control plane 397 of the network devices, while OpenFlow dramatically changes the 398 forwarding plane. We also think that the level of centralization 399 that required by OpenFlow is hardly achievable in SP networks so 400 hybrid BGP+PCEP approach looks much more interesting. 402 7. Deployment Consideration 404 7.1. Scalability 406 In CCDR framework, PCE needs only influence the edge routers for the 407 prefixes advertisement via the multiple BGP sessions deployment. The 408 route information for these prefixes within the on-path routers were 409 distributed via the BGP protocol. 411 For multiple domains deployment, the PCE or the pool of PCEs that 412 reponsible for these domains need only control the edge router to 413 build multiple eBGP sessions, all other procedures are the same that 414 in one domain. 416 Unlike the solution from BGP Flowspec, the on-path router need only 417 keep the specific policy routes to the BGP next-hop of the 418 differentiate prefixes, not the specific routes to the prefixes 419 themselves. This can lessen the burden from the table size of policy 420 based routes for the on-path routers, and has more expandability when 421 comparing with the solution from BGP flowspec or Openflow. For 422 example, if we want to differentiate 1000 prefixes from the normal 423 traffic, CCDR needs only one explicit peer route in every on-path 424 router, but the BGP flowspec or Openflow needs 1000 policy routes on 425 them. 427 7.2. High Availability 429 The CCDR framework is based on the distributed IP protocol. If the 430 PCE failed, the forwarding plane will not be impacted, as the BGP 431 session between all devices will not flap, and the forwarding table 432 will remain unchanged. 434 If one node on the optimal path is failed, the priority traffic will 435 fall over to the best-effort forwarding path. One can even design 436 several assurance paths to load balance/hot-standby the priority 437 traffic to meet the path failure situation. 439 For high availability of PCE/SDN-controller, operator should rely on 440 existing HA solutions for SDN controller, such as clustering 441 technology and deployment. 443 7.3. Incremental deployment 445 Not every router within the network will support the PCEP extension 446 that defined in [I-D.ietf-pce-pcep-extension-native-ip] 447 simultaneously. 449 For such situations, router on the edge of domain can be upgraded 450 first, and then the traffic can be assured between different domains. 451 Within each domain, the traffic will be forwarded along the best- 452 effort path. Service provider can selectively upgrade the routers on 453 each domain in sequence. 455 8. Security Considerations 457 A PCE needs to assure calculation of E2E path based on the status of 458 network and the service requirements in real-time. 460 The PCE need consider the explicit route deployment order (for 461 example, from tail router to head router) to eliminate the possible 462 transient traffic loop. 464 The setup of BGP session, prefix advertisement and explicit peer 465 route establishment are all controlled by the PCE. To prevent the 466 bogus PCE to send harmful messages to the network nodes, the network 467 devices should authenticate the validity of PCE and keep secures 468 communication channel between them. Mechanism described in [RFC8253] 469 should be used to avoid such situation. 471 CCDR framework does not require the change of forward behavior on the 472 underlay devices, then there will no additional security impact on 473 the devices. 475 9. IANA Considerations 477 This document does not require any IANA actions. 479 10. Acknowledgement 481 The author would like to thank Deborah Brungard, Adrian Farrel, 482 Vishnu Beeram, Lou Berger, Dhruv Dhody, Raghavendra Mallya , Mike 483 Koldychev, Haomian Zheng, Penghui Mi, Shaofu Peng and Jessica Chen 484 for their supports and comments on this draft. 486 11. References 488 11.1. Normative References 490 [RFC4456] Bates, T., Chen, E., and R. Chandra, "BGP Route 491 Reflection: An Alternative to Full Mesh Internal BGP 492 (IBGP)", RFC 4456, DOI 10.17487/RFC4456, April 2006, 493 . 495 [RFC4655] Farrel, A., Vasseur, J., and J. Ash, "A Path Computation 496 Element (PCE)-Based Architecture", RFC 4655, 497 DOI 10.17487/RFC4655, August 2006, 498 . 500 [RFC4657] Ash, J., Ed. and J. Le Roux, Ed., "Path Computation 501 Element (PCE) Communication Protocol Generic 502 Requirements", RFC 4657, DOI 10.17487/RFC4657, September 503 2006, . 505 [RFC5440] Vasseur, JP., Ed. and JL. Le Roux, Ed., "Path Computation 506 Element (PCE) Communication Protocol (PCEP)", RFC 5440, 507 DOI 10.17487/RFC5440, March 2009, 508 . 510 [RFC7752] Gredler, H., Ed., Medved, J., Previdi, S., Farrel, A., and 511 S. Ray, "North-Bound Distribution of Link-State and 512 Traffic Engineering (TE) Information Using BGP", RFC 7752, 513 DOI 10.17487/RFC7752, March 2016, 514 . 516 [RFC8231] Crabbe, E., Minei, I., Medved, J., and R. Varga, "Path 517 Computation Element Communication Protocol (PCEP) 518 Extensions for Stateful PCE", RFC 8231, 519 DOI 10.17487/RFC8231, September 2017, 520 . 522 [RFC8253] Lopez, D., Gonzalez de Dios, O., Wu, Q., and D. Dhody, 523 "PCEPS: Usage of TLS to Provide a Secure Transport for the 524 Path Computation Element Communication Protocol (PCEP)", 525 RFC 8253, DOI 10.17487/RFC8253, October 2017, 526 . 528 [RFC8281] Crabbe, E., Minei, I., Sivabalan, S., and R. Varga, "Path 529 Computation Element Communication Protocol (PCEP) 530 Extensions for PCE-Initiated LSP Setup in a Stateful PCE 531 Model", RFC 8281, DOI 10.17487/RFC8281, December 2017, 532 . 534 [RFC8283] Farrel, A., Ed., Zhao, Q., Ed., Li, Z., and C. Zhou, "An 535 Architecture for Use of PCE and the PCE Communication 536 Protocol (PCEP) in a Network with Central Control", 537 RFC 8283, DOI 10.17487/RFC8283, December 2017, 538 . 540 [RFC8735] Wang, A., Huang, X., Kou, C., Li, Z., and P. Mi, 541 "Scenarios and Simulation Results of PCE in a Native IP 542 Network", RFC 8735, DOI 10.17487/RFC8735, February 2020, 543 . 545 11.2. Informative References 547 [I-D.ietf-pce-pcep-extension-native-ip] 548 Wang, A., Khasanov, B., Fang, S., and C. Zhu, "PCEP 549 Extension for Native IP Network", draft-ietf-pce-pcep- 550 extension-native-ip-05 (work in progress), February 2020. 552 Authors' Addresses 554 Aijun Wang 555 China Telecom 556 Beiqijia Town, Changping District 557 Beijing 102209 558 China 560 Email: wangaj3@chinatelecom.cn 562 Boris Khasanov 563 Huawei Technologies 564 Moskovskiy Prospekt 97A 565 St.Petersburg 196084 566 Russia 568 Email: khasanov.boris@huawei.com 569 Quintin Zhao 570 Etheric Networks 571 1009 S CLAREMONT ST 572 SAN MATEO, CA 94402 573 USA 575 Email: qzhao@ethericnetworks.com 577 Huaimo Chen 578 Futurewei 579 Boston, MA 580 USA 582 Email: huaimo.chen@futurewei.com