idnits 2.17.1 draft-ietf-teas-pcecc-use-cases-07.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 8, 2021) is 1144 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Looks like a reference, but probably isn't: '1' on line 1396 == Outdated reference: A later version (-13) exists of draft-ietf-pce-pcep-flowspec-12 == Outdated reference: A later version (-14) exists of draft-ietf-pce-pcep-extension-for-pce-controller-10 == Outdated reference: A later version (-08) exists of draft-ietf-pce-pcep-extension-pce-controller-sr-00 == Outdated reference: A later version (-16) exists of draft-li-pce-controlled-id-space-07 == Outdated reference: A later version (-04) exists of draft-ietf-pce-stateful-interdomain-00 == Outdated reference: A later version (-25) exists of draft-ietf-pce-segment-routing-ipv6-08 == Outdated reference: A later version (-17) exists of draft-ietf-teas-pce-native-ip-16 == Outdated reference: A later version (-13) exists of draft-ietf-bier-te-arch-07 == Outdated reference: A later version (-13) exists of draft-chen-pce-bier-07 Summary: 0 errors (**), 0 flaws (~~), 10 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 TEAS Working Group Z. Li 3 Internet-Draft D. Dhody 4 Intended status: Informational Huawei Technologies 5 Expires: September 9, 2021 Q. Zhao 6 Etheric Networks 7 K. Ke 8 Tencent Holdings Ltd. 9 B. Khasanov 10 Yandex LLC 11 L. Fang 12 Expedia, Inc. 13 C. Zhou 14 HPE 15 B. Zhang 16 Telus Communications 17 A. Rachitskiy 18 Mobile TeleSystems JLLC 19 A. Gulida 20 LLC "Lifetech" 21 March 8, 2021 23 The Use Cases for Path Computation Element (PCE) as a Central Controller 24 (PCECC). 25 draft-ietf-teas-pcecc-use-cases-07 27 Abstract 29 The Path Computation Element (PCE) is a core component of a Software- 30 Defined Networking (SDN) system. It can compute optimal paths for 31 traffic across a network and can also update the paths to reflect 32 changes in the network or traffic demands. PCE was developed to 33 derive paths for MPLS Label Switched Paths (LSPs), which are supplied 34 to the head end of the LSP using the Path Computation Element 35 Communication Protocol (PCEP). 37 SDN has a broader applicability than signaled MPLS traffic-engineered 38 (TE) networks, and the PCE may be used to determine paths in a range 39 of use cases including static LSPs, segment routing (SR), Service 40 Function Chaining (SFC), and most forms of a routed or switched 41 network. It is, therefore, reasonable to consider PCEP as a control 42 protocol for use in these environments to allow the PCE to be fully 43 enabled as a central controller. 45 This document describes general considerations for PCECC deployment 46 and examines its applicability and benefits, as well as its 47 challenges and limitations, through a number of use cases. PCEP 48 extensions required for stateful PCE usage are covered in separate 49 documents. 51 This is a living document to catalog the use cases for PCECC. There 52 is currently no intention to publish this work as an RFC. [Update: 53 Chairs are evaluating if the document should be published instead.] 55 Requirements Language 57 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 58 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 59 "OPTIONAL" in this document are to be interpreted as described in BCP 60 14 [RFC2119] [RFC8174] when, and only when, they appear in all 61 capitals, as shown here. 63 Status of This Memo 65 This Internet-Draft is submitted in full conformance with the 66 provisions of BCP 78 and BCP 79. 68 Internet-Drafts are working documents of the Internet Engineering 69 Task Force (IETF). Note that other groups may also distribute 70 working documents as Internet-Drafts. The list of current Internet- 71 Drafts is at https://datatracker.ietf.org/drafts/current/. 73 Internet-Drafts are draft documents valid for a maximum of six months 74 and may be updated, replaced, or obsoleted by other documents at any 75 time. It is inappropriate to use Internet-Drafts as reference 76 material or to cite them other than as "work in progress." 78 This Internet-Draft will expire on September 9, 2021. 80 Copyright Notice 82 Copyright (c) 2021 IETF Trust and the persons identified as the 83 document authors. All rights reserved. 85 This document is subject to BCP 78 and the IETF Trust's Legal 86 Provisions Relating to IETF Documents 87 (https://trustee.ietf.org/license-info) in effect on the date of 88 publication of this document. Please review these documents 89 carefully, as they describe your rights and restrictions with respect 90 to this document. Code Components extracted from this document must 91 include Simplified BSD License text as described in Section 4.e of 92 the Trust Legal Provisions and are provided without warranty as 93 described in the Simplified BSD License. 95 Table of Contents 97 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 98 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 99 3. Application Scenarios . . . . . . . . . . . . . . . . . . . . 4 100 3.1. Use Cases of PCECC for Label Management . . . . . . . . . 4 101 3.2. Using PCECC for SR . . . . . . . . . . . . . . . . . . . 6 102 3.2.1. PCECC SID Allocation . . . . . . . . . . . . . . . . 7 103 3.2.2. Use Cases of PCECC for SR Best Effort (BE) Path . . . 8 104 3.2.3. Use Cases of PCECC for SR Traffic Engineering (TE) 105 Path . . . . . . . . . . . . . . . . . . . . . . . . 8 106 3.3. Use Cases of PCECC for TE LSP . . . . . . . . . . . . . . 9 107 3.3.1. PCECC Load Balancing (LB) Use Case . . . . . . . . . 11 108 3.3.2. PCECC and Inter-AS TE . . . . . . . . . . . . . . . . 13 109 3.4. Use Cases of PCECC for Multicast LSPs . . . . . . . . . . 16 110 3.4.1. Using PCECC for P2MP/MP2MP LSPs' Setup . . . . . . . 17 111 3.4.2. Use Cases of PCECC for the Resiliency of P2MP/MP2MP 112 LSPs . . . . . . . . . . . . . . . . . . . . . . . . 18 113 3.5. Use Cases of PCECC for LSP in the Network Migration . . . 20 114 3.6. Use Cases of PCECC for L3VPN and PWE3 . . . . . . . . . . 22 115 3.7. Using PCECC for Traffic Classification Information . . . 23 116 3.8. Use Cases of PCECC for SRv6 . . . . . . . . . . . . . . . 23 117 3.9. Use Cases of PCECC for SFC . . . . . . . . . . . . . . . 25 118 3.10. Use Cases of PCECC for Native IP . . . . . . . . . . . . 25 119 3.11. Use Cases of PCECC for Local Protection (RSVP-TE) . . . . 26 120 3.12. Use Cases of PCECC for BIER . . . . . . . . . . . . . . . 26 121 4. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 26 122 5. Security Considerations . . . . . . . . . . . . . . . . . . . 27 123 6. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 27 124 7. References . . . . . . . . . . . . . . . . . . . . . . . . . 27 125 7.1. Normative References . . . . . . . . . . . . . . . . . . 27 126 7.2. Informative References . . . . . . . . . . . . . . . . . 27 127 Appendix A. Using reliable P2MP TE based multicast delivery for 128 distributed computations (MapReduce-Hadoop) . . . . 31 129 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 34 131 1. Introduction 133 An Architecture for Use of PCE and PCEP [RFC5440] in a Network with 134 Central Control [RFC8283] describes SDN architecture where the Path 135 Computation Element (PCE) determines paths for variety of different 136 usecases, with PCEP as a general southbound communication protocol 137 with all the nodes along the path.. 139 [I-D.ietf-pce-pcep-extension-for-pce-controller] introduces the 140 procedures and extensions for PCEP to support the PCECC architecture 141 [RFC8283]. 143 This draft describes the various usecases for the PCECC architecture. 145 This is a living document to catalog the use cases for PCECC. There 146 is currently no intention to publish this work as an RFC. [Update: 147 Chairs are evaluating if the document should be published instead.] 149 2. Terminology 151 The following terminology is used in this document. 153 IGP: Interior Gateway Protocol. Either of the two routing 154 protocols, Open Shortest Path First (OSPF) or Intermediate System 155 to Intermediate System (IS-IS). 157 PCC: Path Computation Client: any client application requesting a 158 path computation to be performed by a Path Computation Element. 160 PCE: Path Computation Element. An entity (component, application, 161 or network node) that is capable of computing a network path or 162 route based on a network graph and applying computational 163 constraints. 165 PCECC: PCE as a central controller. Extension of PCE to support SDN 166 functions as per [RFC8283]. 168 TE: Traffic Engineering. 170 3. Application Scenarios 172 In the following sections, several use cases are described, 173 showcasing scenarios that benefit from the deployment of PCECC. 175 3.1. Use Cases of PCECC for Label Management 177 As per [RFC8283], in some cases, the PCE-based controller can take 178 responsibility for managing some part of the MPLS label space for 179 each of the routers that it controls, and it may taker wider 180 responsibility for partitioning the label space for each router and 181 allocating different parts for different uses, communicating the 182 ranges to the router using PCEP. 184 [I-D.ietf-pce-pcep-extension-for-pce-controller] describe a mode 185 where LSPs are provisioned as explicit label instructions at each hop 186 on the end-to-end path. Each router along the path must be told what 187 label forwarding instructions to program and what resources to 188 reserve. The controller uses PCEP to communicate with each router 189 along the path of the end-to-end LSP. For this to work, the PCE- 190 based controller will take responsibility for managing some part of 191 the MPLS label space for each of the routers that it controls. An 192 extension to PCEP could be done to allow a PCC to inform the PCE of 193 such a label space to control. 195 [RFC8664] specifies extensions to PCEP that allow a stateful PCE to 196 compute, update or initiate SR-TE paths. 197 [I-D.ietf-pce-pcep-extension-pce-controller-sr] describes the 198 mechanism for PCECC to allocate and provision the node/prefix/ 199 adjacency label (SID) via PCEP. To make such allocation PCE needs to 200 be aware of the label space from Segment Routing Global Block (SRGB) 201 or Segment Routing Local Block (SRLB) [RFC8402] of the node that it 202 controls. A mechanism for a PCC to inform the PCE of such a label 203 space to control is needed within PCEP. The full SRGB/SRLB of a node 204 could be learned via existing IGP or BGP-LS mechanism too. 206 [I-D.li-pce-controlled-id-space] defines a PCEP extension to support 207 advertisement of the MPLS label space to the PCE to control. 209 There have been various proposals for Global Labels, the PCECC 210 architecture could be used as means to learn the label space of 211 nodes, and could also be used to determine and provision the global 212 label range. 214 +------------------------------+ +------------------------------+ 215 | PCE DOMAIN 1 | | PCE DOMAIN 2 | 216 | +--------+ | | +--------+ | 217 | | | | | | | | 218 | | PCECC1 | ---------PCEP---------- | PCECC2 | | 219 | | | | | | | | 220 | | | | | | | | 221 | +--------+ | | +--------+ | 222 | ^ ^ | | ^ ^ | 223 | / \ PCEP | | PCEP / \ | 224 | V V | | V V | 225 | +--------+ +--------+ | | +--------+ +--------+ | 226 | |NODE 11 | | NODE 1n| | | |NODE 21 | | NODE 2n| | 227 | | | ...... | | | | | | ...... | | | 228 | | PCECC | | PCECC | | | | PCECC | |PCECC | | 229 | |Enabled | | Enabled| | |Enabled | |Enabled | | 230 | +--------+ +--------+ | | +--------+ +--------+ | 231 | | | | 232 +------------------------------+ +------------------------------+ 234 Figure 1: PCECC for Label Management 236 o PCC would advertise the PCECC capability to the PCE (central 237 controller-PCECC) 238 [I-D.ietf-pce-pcep-extension-for-pce-controller]. 240 o The PCECC could also learn the label range set aside by the PCC 241 ([I-D.li-pce-controlled-id-space]). 243 o Optionally, the PCECC could determine the shared MPLS global label 244 range for the network. 246 o In the case that the shared global label range need to be 247 negotiated across multiple domains, the central controllers of 248 these domains would also need to negotiate a common global 249 label range across domains. 251 o The PCECC would need to set the shared global label range to 252 all PCC nodes in the network. 254 3.2. Using PCECC for SR 256 Segment Routing (SR) leverages the source routing paradigm. Using 257 SR, a source node steers a packet through a path without relying on 258 hop-by-hop signaling protocols such as LDP or RSVP-TE. Each path is 259 specified as an ordered list of instructions called "segments". Each 260 segment is an instruction to route the packet to a specific place in 261 the network, or to perform a specific service on the packet. A 262 database of segments can be distributed through the network using a 263 routing protocol (such as IS-IS or OSPF) or by any other means. PCEP 264 (and PCECC) could be one such means. 266 [RFC8664] specify the SR specific PCEP extensions. PCECC may further 267 use PCEP protocol for SR SID (Segment Identifier) distribution to the 268 SR nodes (PCC) with some benefits. If the PCECC allocates and 269 maintains the SID in the network for the nodes and adjacencies; and 270 further distributes them to the SR nodes directly via the PCEP 271 session has some advantage over the configurations on each SR node 272 and flooding via IGP, especially in a SDN environment. 274 When the PCECC is used for the distribution of the node segment ID 275 and adjacency segment ID, the node segment ID is allocated from the 276 SRGB of the node. For the allocation of adjacency segment ID, the 277 allocation is from the SRLB of the node as described in 278 [I-D.ietf-pce-pcep-extension-pce-controller-sr]. 280 [RFC8355] identifies various protection and resiliency usecases for 281 SR. Path protection lets the ingress node be in charge of the 282 failure recovery (used for SR-TE). Also protection can be performed 283 by the node adjacent to the failed component, commonly referred to as 284 local protection techniques or fast-reroute (FRR) techniques. In 285 case of PCECC, the protection paths can be pre-computed and setup by 286 the PCE. 288 The following example illustrate the use case where the node SID and 289 adjacency SID are allocated by the PCECC. 291 192.0.2.1/32 292 +----------+ 293 | R1(1001) | 294 +----------+ 295 | 296 +----------+ 297 | R2(1002) | 192.0.2.2/32 298 +----------+ 299 * | * * 300 * | * * 301 *link1| * * 302 192.0.2.4/32 * | *link2 * 192.0.2.5/32 303 +-----------+ 9001| * +-----------+ 304 | R4(1004) | | * | R5(1005) | 305 +-----------+ | * +-----------+ 306 * | *9003 * + 307 * | * * + 308 * | * * + 309 +-----------+ +-----------+ 310 192.0.2.3/32 | R3(1003) | |R6(1006) |192.0.2.6/32 311 +-----------+ +-----------+ 312 | 313 +-----------+ 314 | R8(1008) | 192.0.2.8/32 315 +-----------+ 317 3.2.1. PCECC SID Allocation 319 Each node (PCC) is allocated a node-SID by the PCECC. The PCECC 320 needs to update the label map of each node to all the nodes in the 321 domain. On receiving the label map, each node (PCC) uses the local 322 routing information to determine the next-hop and download the label 323 forwarding instructions accordingly. The forwarding behavior and the 324 end result is same as IGP based Node-SID in SR. Thus, from anywhere 325 in the domain, it enforces the ECMP-aware shortest-path forwarding of 326 the packet towards the related node. 328 For each adjacency in the network, PCECC can allocate an Adj-SID. 329 The PCECC sends PCInitiate message to update the label map of each 330 Adj to the corresponding nodes in the domain. Each node (PCC) 331 download the label forwarding instructions accordingly. The 332 forwarding behavior and the end result is similar to IGP based "Adj- 333 SID" in SR. 335 The various mechanism are described in 336 [I-D.ietf-pce-pcep-extension-pce-controller-sr]. 338 3.2.2. Use Cases of PCECC for SR Best Effort (BE) Path 340 In this mode of the solution, the PCECC just need to allocate the 341 node segment ID and adjacency ID (without calculating the explicit 342 path for the SR path). The ingress of the forwarding path just need 343 to encapsulate the destination node segment ID on top of the packet. 344 All the intermediate nodes will forward the packet based on the 345 destination node SID. It is similar to the LDP LSP. 347 R1 may send a packet to R8 simply by pushing an SR header with 348 segment list {1008} (Node SID for R8). The path would be the based 349 on the routing/nexthop calculation on the routers. 351 3.2.3. Use Cases of PCECC for SR Traffic Engineering (TE) Path 353 SR-TE paths may not follow an IGP SPT. Such paths may be chosen by a 354 PCECC and provisioned on the ingress node of the SR-TE path. The SR 355 header consists of a list of SIDs (or MPLS labels). The header has 356 all necessary information so that, the packets can be guided from the 357 ingress node to the egress node of the path; hence, there is no need 358 for any signaling protocol. For the case where strict traffic 359 engineering path is needed, all the adjacency SID are stacked, 360 otherwise a combination of node-SID or adj-SID can be used for the 361 SR-TE paths. 363 Note that the bandwidth reservations is only guaranteed at controller 364 and through the enforce of the bandwidth admission control. As for 365 the RSVP-TE LSP case, the control plane signaling also does the link 366 bandwidth reservation in each hop of the path. 368 The SR traffic engineering path examples are explained as bellow: 370 Note that the node SID for each node is allocated from the SRGB and 371 adjacency SID for each link are allocated from the SRLB for each 372 node. 374 Example 1: 376 R1 may send a packet P1 to R8 simply by pushing an SR header with 377 segment list {1008}. Based on the best path, it could be: 378 R1-R2-R3-R8. 380 Example 2: 382 R1 may send a packet P2 to R8 by pushing an SR header with segment 383 list {1002, 9001, 1008}. The path should be: R1-R2-link1-R3-R8. 385 Example 3: 387 R1 may send a packet P3 to R8 via R4 by pushing an SR header with 388 segment list {1004, 1008}. The path could be : R1-R2-R4-R3-R8 390 The local protection examples for SR TE path are explained below: 392 Example 4: local link protection: 394 o R1 may send a packet P4 to R8 by pushing an SR header with segment 395 list {1002, 9001, 1008}. The path should be: R1-R2-link1-R3-R8. 397 o When node R2 receives the packet from R1 which has the header of 398 link1-R3-R8, and also find out there is a link failure of link1, 399 then the R2 can enforce the traffic over the bypass to send out 400 the packet with header of R3-R8 through link2. 402 Example 5: local node protection: 404 o R1 may send a packet P5 to R8 by pushing an SR header with segment 405 list {1004, 1008}. The path could be : R1-R2-R4-R3-R8. 407 o When node R2 receives the packet from R1 which has the header of 408 {1004, 1008}, and also finds out there is a node failure for 409 node4, then it can enforce the traffic over the bypass and send 410 out the packet with header of {1005, 1008} to node5 instead of 411 node4. 413 3.3. Use Cases of PCECC for TE LSP 415 In the Section 3.2 the case of SR path via PCECC is discussed. 416 Although those cases give the simplicity and scalability, but there 417 are existing functionalities for the traffic engineering path such as 418 the bandwidth guarantee, monitoring where SR based solution are 419 complex. Also there are cases where the depth of the label stack is 420 an issue for existing deployment and certain vendors. 422 So to address these issues, PCECC architecture also support the TE 423 LSP functionalities. To achieve this, the existing PCEP can be used 424 to communicate between the PCECC and nodes along the path. This is 425 similar to static LSPs, where LSPs can be provisioned as explicit 426 label instructions at each hop on the end-to-end path. Each router 427 along the path must be told what label- forwarding instructions to 428 program and what resources to reserve. The PCE-based controller 429 keeps a view of the network and determines the paths of the end-to- 430 end LSPs, and the controller uses PCEP to communicate with each 431 router along the path of the end-to-end LSP. 433 192.0.2.1/32 434 +----------+ 435 | R1 | 436 +----------+ 437 | | 438 |link1 | 439 | |link2 440 +----------+ 441 | R2 | 192.0.2.2/32 442 +----------+ 443 link3 * | * * link4 444 * | * * 445 *link5| * * 446 192.0.2.4/32 * | *link6 * 192.0.2.5/32 447 +-----------+ | * +-----------+ 448 | R4 | | * | R5 | 449 +-----------+ | * +-----------+ 450 * | * * + 451 link10 * | * *link7 + 452 * | * * + 453 +-----------+ +-----------+ 454 192.0.2.3/32 | R3 | |R6 |192.0.2.6/32 455 +-----------+ +-----------+ 456 | | 457 |link8 | 458 | |link9 459 +-----------+ 460 | R8 | 192.0.2.8/32 461 +-----------+ 463 Figure 2: PCECC TE LSP Setup Example 465 o Based on path computation request / delegation or PCE initiation, 466 the PCECC receives the PCECC request with constraints and 467 optimization criteria. 469 o PCECC would calculate the optimal path according to given 470 constrains (e.g. bandwidth). 472 o PCECC would provision each node along the path and assign incoming 473 and outgoing labels from R1 to R8 with the path: {R1, link1, 474 1001}, {1001, R2, link3, 2003], {2003, R4, link10, 4010}, {4010, 475 R3, link8, 3008}, {3008, R8}. 477 o For the end to end protection, PCECC program each node along the 478 path from R1 to R8 with the secondary path: {R1, link2, 1002}, 479 {1002, R2, link4, 2004], {2004, R5, link7, 5007}, {5007, R3, 480 link9, 3009}, {3009, R8}. 482 o It is also possible to have a bypass path for the local protection 483 setup by the PCECC. For example, the primary path as above, then 484 to protect the node R4 locally, PCECC can program the bypass path 485 like this: {R2, link5, 2005}, {2005, R3}. By doing this, the node 486 R4 is locally protected at R2. 488 3.3.1. PCECC Load Balancing (LB) Use Case 490 Very often many service providers use TE tunnels for solving issues 491 with non-deterministic paths in their networks. One example of such 492 applications is usage of TEs in the mobile backhaul (MBH). Consider 493 the following topology - 495 TE1 --------------> 496 +---------+ +--------+ +--------+ +--------+ +------+ +---+ 497 | Access |----| Access |----| AGG 1 |----| AGG N-1|----|Core 1|--|SR1| 498 | SubNode1| | Node 1 | +--------+ +--------+ +------+ +---+ 499 +---------+ +--------+ | | | ^ | 500 | Access | Access | AGG Ring 1 | | | 501 | SubRing 1 | Ring 1 | | | | | 502 +---------+ +--------+ +--------+ | | | 503 | Access | | Access | | AGG 2 | | | | 504 | SubNode2| | Node 2 | +--------+ | | | 505 +---------+ +--------+ | | | | | 506 | | | | | | | 507 | | | +----TE2----|-+ | 508 +---------+ +--------+ +--------+ +--------+ +------+ +---+ 509 | Access | | Access |----| AGG 3 |----| AGG N |----|Core N|--|SRn| 510 | SubNodeN|----| Node N | +--------+ +--------+ +------+ +---+ 511 +---------+ +--------+ 513 This MBH architecture uses L2 access rings and sub-rings. L3 starts 514 at the aggregation layer. For the sake of simplicity, the figure 515 shows only one access sub-ring, access ring and aggregation ring 516 (AGG1...AGGN), connected by Nx10GE interfaces. Aggregation domain 517 runs its own IGP. There are two Egress routers (AGG N-1,AGG N) that 518 are connected to the Core domain via L2 interfaces. Core also have 519 connections to service routers, RSVP-TEs are used for MPLS transport 520 inside the ring. There could be at least 2 tunnels (one way) from 521 each AGG router to egress AGG routers. There are also many L2 access 522 rings connected to AGG routers. 524 Service deployment made by means of either L2VPNs (VPLS) or L3VPNs. 525 Those services use MPLS TE as transport towards egress AGG routers. 526 TE tunnels could be also used as transport towards service routers in 527 case of seamless MPLS based architecture in the future. 529 There is a need to solve the following tasks: 531 o Perform automatic load-balance amongst TE tunnels according to 532 current traffic load. 534 o TE bandwidth (BW) management: Provide guaranteed BW for specific 535 service: HSI, IPTV, etc., provide time-based BW reservation (BoD) 536 for other services. 538 o Simplify development of TE tunnels by automation without any 539 manual intervention. 541 o Provide flexibility for Service Router placement (anywhere in the 542 network by creation of transport LSPs to them). 544 Since other tasks are already considered by other PCECC use cases, in 545 this section, the focus is on load balancing (LB) task. LB task 546 could be solved by means of PCECC in the following way: 548 o After application or network service or operator can ask SDN 549 controller (PCECC) for LSP based LB between AGG X and AGG N/AGG 550 N-1 (egress AGG routers which have connections to core) via North 551 Bound Interface (NBI). Each of these would have associated 552 constrains (i.e. Path Setup Type (PST), bandwidth, inclusion or 553 exclusion specific links or nodes, number of paths, objective 554 function (OF), need for disjoint LSP paths etc.). 556 o PCECC could calculate multiple (Say N) LSPs according to given 557 constrains, calculation is based on results of Objective Function 558 (OF) [RFC5541], constraints, endpoints, same or different 559 bandwidth (BW) , different links (in case of disjoint paths) and 560 other constrains. 562 o Depending on given LSP Path setup type (PST), PCECC would use 563 download instructions to the PCC. At this stage it is assumed the 564 PCECC is aware of the label space it controls and in case of SR 565 the SID allocation and distribution is already done. 567 o PCECC would send PCInitiate PCEP message [RFC8281] towards ingress 568 AGG X router(PCC) for each of N LSPs and receives PCRpt PCEP 569 message [RFC8231] back from PCCs. If the PST is PCECC-SR, the 570 PCECC would include the SID stack as per [RFC8664]. If the PST is 571 PCECC (basic), then the PCECC would assigns labels along the 572 calculated path; and set up the path by sending central controller 573 instructions in PCEP message to each node along the path of the 574 LSP as per [I-D.ietf-pce-pcep-extension-for-pce-controller] and 575 then send PCUpd message to the ingress AGG X router with 576 information about new LSP and AGG X(PCC) would respond with PCRpt 577 with LSP status. 579 o AGG X as ingress router now have N LSPs towards AGG N and AGG N-1 580 which are available for installing to router's forwarding and LB 581 of traffic between them. Traffic distribution between those LSPs 582 depends on particular realization of hash-function on that router. 584 o Since PCECC is aware of TEDB (TE state) and LSP-DB, it can manage 585 and prevent possible over-subscriptions and limit number of 586 available LB states. Via PCECC mechanism the control can take 587 quick actions into the network by directly provisioning the 588 central control instructions. 590 3.3.2. PCECC and Inter-AS TE 592 There are various signaling options for establishing Inter-AS TE LSP: 593 contiguous TE LSP [RFC5151], stitched TE LSP [RFC5150], nested TE LSP 594 [RFC4206]. 596 Requirements for PCE-based Inter-AS setup [RFC5376] describe the 597 approach and PCEP functionality that are needed for establishing 598 Inter-AS TE LSPs. 600 [RFC5376] also gives Inter- and Intra-AS PCE Reference Model that is 601 provided below in shorten form for the sake of simplicity. 603 Inter-AS Inter-AS 604 PCC <-->PCE1<--------->PCE2 605 :: :: :: 606 :: :: :: 607 R1----ASBR1====ASBR3---R3---ASBR5 608 | AS1 | | PCC | 609 | | | AS2 | 610 R2----ASBR2====ASBR4---R4---ASBR6 611 :: :: 612 :: :: 613 Intra-AS Intra-AS 614 PCE3 PCE4 616 Figure 3: Shorten form of Inter- and Intra-AS PCE Reference Model 617 [RFC5376] 619 The PCECC belonging to different domain can co-operate to setup 620 inter-AS TE LSP. The stateful H-PCE [RFC8751] mechanism could also 621 be used to first establish a per-domain PCECC LSP. These could be 622 stitched together to form inter-AS TE LSP as described in 623 [I-D.ietf-pce-stateful-interdomain]. 625 For the sake of simplicity, here after the focus is on a simplified 626 Inter-AS case when both AS1 and AS2 belong to the same service 627 provider administration. In that case Inter and Intra-AS PCEs could 628 be combined in one single PCE if such combined PCE performance is 629 enough for handling all path computation request and setup. There is 630 a potential to use a single PCE for both ASes if the scalability and 631 performance are enough. The PCE would require interfaces (PCEP and 632 BGP-LS) to both domains. PCECC redundancy mechanisms are described 633 in [RFC8283]. Thus routers in AS1 and AS2 (PCCs) can send PCEP 634 messages towards same PCECC. 636 +----BGP-LS------+ +------BGP-LS-----+ 637 | | | | 638 +-PCEP-|----++-+-------PCECC-----PCEP--++-+-|-------+ 639 +-:------|----::-:-+ +--::-:-|-------:---+ 640 | : | :: : | | :: : | : | 641 | : RR1 :: : | | :: : RR2 : | 642 | v v: : | LSP1 | :: v v | 643 | R1---------ASBR1=======================ASBR3--------R3 | 644 | | v : | | :v | | 645 | +----------ASBR2=======================ASBR4---------+ | 646 | | Region 1 : | | : Region 1 | | 647 |----------------:-| |--:-------------|--| 648 | | v | LSP2 | v | | 649 | +----------ASBR5=======================ASBR6---------+ | 650 | Region 2 | | Region 2 | 651 +------------------+ <--------------> +-------------------+ 652 MPLS Domain 1 Inter-AS MPLS Domain 2 653 <=======AS1=======> <========AS2=======> 655 Figure 4: Particular case of Inter-AS PCE 657 In a case of PCECC Inter-AS TE scenario where service provider 658 controls both domains (AS1 and AS2), each of them have own IGP and 659 MPLS transport. There is a need is to setup Inter-AS LSPs for 660 transporting different services on top of them (Voice, L3VPN etc.). 661 Inter-AS links with different capacity exist in several regions. The 662 task is not only to provision those Inter-AS LSPs with given 663 constrains but also calculate the path and pre-setup the backup 664 Inter-AS LSPs that will be used if primary LSP fails. 666 As per the Figure 4, LSP1 from R1 to R3 goes via ASBR1 and ASBR3, and 667 it is the primary Inter-AS LSP. R1-R3 LSP2 that go via ASBR5 and 668 ASBR6 is the backup one. In addition there could also be a bypass 669 LSP setup to protect against ASBR or inter-AS link failure. 671 After the addition of PCECC functionality to PCE (SDN controller), 672 PCECC based Inter-AS TE model SHOULD follow as PCECC usecase for TE 673 LSP as requirements of [RFC5376] with the following details: 675 o Since PCECC needs to know the topology of both domains AS1 and 676 AS2, PCECC could use BGP-LS peering with routers (or RRs) in both 677 domains. 679 o PCECC needs to PCEP connectivity towards all routers in both 680 domains (see also section 4 in [RFC5376]) in a similar manner as a 681 SDN controller. 683 o After operator's application or service orchestrator will create 684 request for tunnel creation of specific service, PCECC should 685 receive that request via NBI (NBI type is implementation 686 dependent, could be NETCONF/Yang, REST etc.). Then PCECC would 687 calculate the optimal path based on Objective Function (OF) and 688 given constraints (i.e. path setup type, bandwidth etc.), 689 including those from [RFC5376]: priority, AS sequence, preferred 690 ASBR, disjoint paths, protection. On this step we would have two 691 paths: R1-ASBR1-ASBR3-R3, R1-ASBR5-ASBR6-R3 693 o Depending on given LSP PST (PCECC or PCECC-SR), PCECC would use 694 central control download instructions to the PCC. At this stage 695 it is assumed the PCECC is aware of the label space it controls 696 and in case of SR the SID allocation and distribution is already 697 done. 699 o PCECC would send PCInitiate PCEP message [RFC8281] towards ingress 700 router R1 (PCC) in AS1 and receives PCRpt PCEP message [RFC8231] 701 back from PCC. If the PST is PCECC-SR, the PCECC would include 702 the SID stack as per [RFC8664]. It may also include binding SID 703 based on AS boundary. The backup SID stack could also be 704 installed at ingress but more importantly each node along the SR 705 path could also do local protection just based on the top segment. 706 If the PST is PCECC (basic), then the PCECC would assigns labels 707 along the calculated paths (R1-ASBR1-ASBR3-R3, R1-ASBR5-ASBR6-R3); 708 and set up the path by sending central controller instructions in 709 PCEP message to each node along the path of the LSPs as per 710 [I-D.ietf-pce-pcep-extension-for-pce-controller] and then send 711 PCUpd message to the ingress R1 router with information about new 712 LSPs and R1 would respond with PCRpt with LSP(s) status. 714 o After that step R1 now have primary and backup TEs (LSP1 and LSP2) 715 towards R3. It is up to router implementation how to make 716 switchover to backup LSP2 if LSP1 fails. 718 3.4. Use Cases of PCECC for Multicast LSPs 720 The current multicast LSPs are setup either using the RSVP-TE P2MP or 721 mLDP protocols. The setup of these LSPs may require manual 722 configurations and complex signaling when the protection is 723 considered. By using the PCECC solution, the multicast LSP can be 724 computed and setup through centralized controller which has the full 725 picture of the topology and bandwidth usage for each link. It not 726 only reduces the complex configurations comparing the distributed 727 RSVP-TE P2MP or mLDP signaling, but also it can compute the disjoint 728 primary path and secondary P2MP path efficiently. 730 3.4.1. Using PCECC for P2MP/MP2MP LSPs' Setup 732 It is assumed the PCECC is aware of the label space it controls for 733 all nodes and make allocations accordingly. 735 +----------+ 736 | R1 | Root node of the multicast LSP 737 +----------+ 738 |6000 739 +----------+ 740 Transit Node | R2 | 741 branch +----------+ 742 * | * * 743 9001* | * *9002 744 * | * * 745 +-----------+ | * +-----------+ 746 | R4 | | * | R5 | Transit Nodes 747 +-----------+ | * +-----------+ 748 * | * * + 749 9003* | * * +9004 750 * | * * + 751 +-----------+ +-----------+ 752 | R3 | | R6 | Leaf Node 753 +-----------+ +-----------+ 754 9005| 755 +-----------+ 756 | R8 | Leaf Node 757 +-----------+ 759 The P2MP examples are explained here, where R1 is root and R8 and R6 760 are the leaves. 762 o Based on the P2MP path computation request / delegation or PCE 763 initiation, the PCECC receives the PCECC request with constraints 764 and optimization criteria. 766 o PCECC would calculate the optimal P2MP path according to given 767 constrains (i.e.bandwidth). 769 o PCECC would provision each node along the path and assign incoming 770 and outgoing labels from R1 to {R6, R8} with the path: {R1, 6000}, 771 {6000, R2, {9001,9002}}, {9001, R4, 9003}, {9002, R5, 9004} {9003, 772 R3, 9005}, {9004, R6}, {9005, R8}. The main difference is in the 773 branch node instruction at R2 where two copies of packet are sent 774 towards R4 and R5 with 9001 and 9002 labels respectively. 776 The packet forwarding involves - 777 Step1: R1 may send a packet P1 to R2 simply by pushing an label of 778 6000 to the packet. 780 Step2: After R2 receives the packet with label 6000, it will 781 forwarding to R4 by swapping label to 9001 and by swapping label 782 of 9002 towards R5. 784 Step3: After R4 receives the packet with label 9001, it will 785 forwarding to R3 by swapping to 9003. After R5 receives the 786 packet with label 9002, it will forwarding to R6 by swapping to 787 9004. 789 Step4: After R3 receives the packet with label 9003, it will 790 forwarding to R8 by swapping to 9005 and when R5 receives the 791 packet with label 9004, it will swap to 9004 and send to R6. 793 Step5: Packet received at R8 and 9005 is popped; packet receives 794 at R6 and 9004 is popped. 796 3.4.2. Use Cases of PCECC for the Resiliency of P2MP/MP2MP LSPs 798 3.4.2.1. PCECC for the End-to-End Protection of the P2MP/MP2MP LSPs 800 In this section we describe the end-to-end managed path protection 801 service as well as the local protection with the operation management 802 in the PCECC network for the P2MP/MP2MP LSP. 804 An end-to-end protection principle can be applied for computing 805 backup P2MP or MP2MP LSPs. During computation of the primary 806 multicast trees, PCECC server may also take the computation of a 807 secondary tree into consideration. A PCE may compute the primary and 808 backup P2MP (or MP2MP) LSP together or sequentially. 810 +----+ +----+ 811 Root node of LSP | R1 |--| R11| 812 +----+ +----+ 813 / + 814 10/ +20 815 / + 816 +----------+ +-----------+ 817 Transit Node | R2 | | R3 | 818 +----------+ +-----------+ 819 | \ + + 820 | \ + + 821 10| 10\ +20 20+ 822 | \ + + 823 | \ + 824 | + \ + 825 +-----------+ +-----------+ Leaf Nodes 826 | R4 | | R5 | (Downstream LSR) 827 +-----------+ +-----------+ 829 In the example above, when the PCECC setup the primary multicast tree 830 from the root node R1 to the leaves, which is R1->R2->{R4, R5}, at 831 same time, it can setup the backup tree, which is R1->R11->R3->{R4, 832 R5}. Both the these two primary forwarding tree and secondary 833 forwarding tree will be downloaded to each routers along the primary 834 path and the secondary path. The traffic will be forwarded through 835 the R1->R2->{R4, R5} path normally, and when there is a node in the 836 primary tree fails (say R2), then the root node R1 will switch the 837 flow to the backup tree, which is R1->R11->R3->{R4, R5}. By using 838 the PCECC, the path computation and forwarding path downloading can 839 all be done without the complex signaling used in the P2MP RSVP-TE or 840 mLDP. 842 3.4.2.2. PCECC for the Local Protection of the P2MP/MP2MP LSPs 844 In this section we describe the local protection service in the PCECC 845 network for the P2MP/MP2MP LSP. 847 While the PCECC sets up the primary multicast tree, it can also build 848 the back LSP among PLR, the protected node, and MPs (the downstream 849 nodes of the protected node). In the cases where the amount of 850 downstream nodes are huge, this mechanism can avoid unnecessary 851 packet duplication on PLR and protect the network from traffic 852 congestion risk. 854 +------------+ 855 | R1 | Root Node 856 +------------+ 857 . 858 . 859 . 860 +------------+ Point of Local Repair/ 861 | R10 | Switchover Point 862 +------------+ (Upstream LSR) 863 / + 864 10/ +20 865 / + 866 +----------+ +-----------+ 867 Protected Node | R20 | | R30 | 868 +----------+ +-----------+ 869 | \ + + 870 | \ + + 871 10| 10\ +20 20+ 872 | \ + + 873 | \ + 874 | + \ + 875 +-----------+ +-----------+ Merge Point 876 | R40 | | R50 | (Downstream LSR) 877 +-----------+ +-----------+ 878 . . 879 . . 881 In the example above, when the PCECC setup the primary multicast path 882 around the PLR node R10 to protect node R20, which is R10->R20->{R40, 883 R50}, at same time, it can setup the backup path R10->R30->{R40, 884 R50}. Both the these two primary forwarding path and secondary 885 bypass forwarding path will be downloaded to each routers along the 886 primary path and the secondary bypass path. The traffic will be 887 forwarded through the R10->R20->{R40, R50} path normally, and when 888 there is a node failure for node R20, then the PLR node R10 will 889 switch the flow to the backup path, which is R10->R30->{R40, R50}. 890 By using the PCECC, the path computation and forwarding path 891 downloading can all be done without the complex signaling used in the 892 P2MP RSVP-TE or mLDP. 894 3.5. Use Cases of PCECC for LSP in the Network Migration 896 One of the main advantages for PCECC solution is that it has backward 897 compatibility naturally since the PCE server itself can function as a 898 proxy node of MPLS network for all the new nodes which may no longer 899 support the signaling protocols. 901 As it is illustrated in the following example, the current network 902 could migrate to a total PCECC controlled network gradually by 903 replacing the legacy nodes. During the migration, the legacy nodes 904 still need to signal using the existing MPLS protocol such as LDP and 905 RSVP-TE, and the new nodes setup their portion of the forwarding path 906 through PCECC directly. With the PCECC function as the proxy of 907 these new nodes, MPLS signaling can populate through network as 908 normal. 910 Example described in this section is based on network configurations 911 illustrated using the following figure: 913 +------------------------------------------------------------------+ 914 | PCE DOMAIN | 915 | +-----------------------------------------------------+ | 916 | | PCECC | | 917 | +-----------------------------------------------------+ | 918 | ^ ^ ^ ^ | 919 | | PCEP | | PCEP | | 920 | V V V V | 921 | +--------+ +--------+ +--------+ +--------+ +--------+ | 922 | | NODE 1 | | NODE 2 | | NODE 3 | | NODE 4 | | NODE 5 | | 923 | | |...| |...| |...| |...| | | 924 | | Legacy |if1| Legacy |if2|Legacy |if3| PCECC |if4| PCECC | | 925 | | Node | | Node | |Enabled | |Enabled | | Enabled| | 926 | +--------+ +--------+ +--------+ +--------+ +--------+ | 927 | | 928 +------------------------------------------------------------------+ 930 Example: PCECC Initiated LSP Setup In the Network Migration 932 In this example, there are five nodes for the TE LSP from head end 933 (Node1) to the tail end (Node5). Where the Node4 and Node5 are 934 centrally controlled and other nodes are legacy nodes. 936 o Node1 sends a path request message for the setup of LSP 937 destinating to Node5. 939 o PCECC sends to node1 a reply message for LSP setup with the path: 940 (Node1, if1),(Node2, if2), (Node3, if3), (Node4, if4), Node5. 942 o Node1, Node2, Node3 will setup the LSP to Node5 using the local 943 labels as usual. Node 3 with help of PCECC could proxy the 944 signaling. 946 o Then the PCECC will program the out-segment of Node3, the in- 947 segment/ out-segment of Node4, and the in-segment for Node5. 949 3.6. Use Cases of PCECC for L3VPN and PWE3 951 As described in [RFC8283], various network services may be offered 952 over a network. These include protection services (including Virtual 953 Private Network (VPN) services (such as Layer 3 VPNs [RFC4364] or 954 Ethernet VPNs [RFC7432]); or Pseudowires [RFC3985]. Delivering 955 services over a network in an optimal way requires coordination in 956 the way that network resources are allocated to support the services. 957 A PCE-based central controller can consider the whole network and all 958 components of a service at once when planning how to deliver the 959 service. It can then use PCEP to manage the network resources and to 960 install the necessary associations between those resources. 962 In the case of L3VPN, VPN labels can be assigned and distributed 963 through the PCECC PCEP among the PE router instead of using the BGP 964 protocols. 966 Example described in this section is based on network configurations 967 illustrated using the following figure: 969 +-------------------------------------------+ 970 | PCE DOMAIN | 971 | +-----------------------------------+ | 972 | | PCECC | | 973 | +-----------------------------------+ | 974 | ^ ^ ^ | 975 |PWE3/L3VPN | PCEP PCEP|LSP PWE3/L3VPN|PCEP | 976 | V V V | 977 +--------+ | +--------+ +--------+ +--------+ | +--------+ 978 | CE | | | PE1 | | NODE x | | PE2 | | | CE | 979 | |...... | |...| |...| |.....| | 980 | Legacy | |if1 | PCECC |if2|PCCEC |if3| PCECC |if4 | Legacy | 981 | Node | | | Enabled| |Enabled | |Enabled | | | Node | 982 +--------+ | +--------+ +--------+ +--------+ | +--------+ 983 | | 984 +-------------------------------------------+ 986 Example: Using PCECC for L3VPN and PWE3 988 In the case PWE3, instead of using the LDP signaling protocols, the 989 label and port pairs assigned to each pseudowire can be assigned 990 through PCECC among the PE routers and the corresponding forwarding 991 entries will be distributed into each PE routers through the extended 992 PCEP protocols and PCECC mechanism. 994 3.7. Using PCECC for Traffic Classification Information 996 As described in [RFC8283], traffic classification is an important 997 part of traffic engineering. It is the process of looking at a 998 packet to determine how it should be treated as it is forwarded 999 through the network. It applies in many scenarios including MPLS 1000 traffic engineering (where it determines what traffic is forwarded 1001 onto which LSPs); segment routing (where it is used to select which 1002 set of forwarding instructions to add to a packet); and SFC (where it 1003 indicates along which service function path a packet should be 1004 forwarded). In conjunction with traffic engineering, traffic 1005 classification is an important enabler for load balancing. Traffic 1006 classification is closely linked to the computational elements of 1007 planning for the network functions just listed because it determines 1008 how traffic load is balanced and distributed through the network. 1009 Therefore, selecting what traffic classification should be performed 1010 by a router is an important part of the work done by a PCECC. 1012 Instructions can be passed from the controller to the routers using 1013 PCEP. These instructions tell the routers how to map traffic to 1014 paths or connections. Refer [I-D.ietf-pce-pcep-flowspec]. 1016 Along with traffic classification, there are few more question that 1017 needs to be considered once the path is setup - 1019 o how to use it 1021 o Whether it is a virtual link 1023 o Whether to advertise it in the IGP as a virtual link 1025 o What bits of this information to signal to the tail end 1027 These are out of scope of this document. 1029 3.8. Use Cases of PCECC for SRv6 1031 As per [RFC8402], with Segment Routing (SR), a node steers a packet 1032 through an ordered list of instructions, called segments. Segment 1033 Routing can be applied to the IPv6 architecture with the Segment 1034 Routing Header (SRH) [RFC8754]. A segment is encoded as an IPv6 1035 address. An ordered list of segments is encoded as an ordered list 1036 of IPv6 addresses in the routing header. The active segment is 1037 indicated by the Destination Address of the packet. Upon completion 1038 of a segment, a pointer in the new routing header is incremented and 1039 indicates the next segment. 1041 As per [RFC8754], an SRv6 Segment is a 128-bit value. "SRv6 SID" or 1042 simply "SID" are often used as a shorter reference for "SRv6 1043 Segment". Further details are in An illustration is provided in 1044 [RFC8986] where SRv6 SID is represented as LOC:FUNCT. 1046 [I-D.ietf-pce-segment-routing-ipv6] extends [RFC8664] to support SR 1047 for IPv6 data plane. Further a PCECC could be extended to support 1048 SRv6 SID allocation and distribution. 1050 2001:db8::1 1051 +----------+ 1052 | R1 | 1053 +----------+ 1054 | 1055 +----------+ 1056 | R2 | 2001:db8::2 1057 +----------+ 1058 * | * * 1059 * | * * 1060 *link1| * * 1061 2001:db8::4 * | *link2 * 2001:db8::5 1062 +-----------+ | * +-----------+ 1063 | R4 | | * | R5 | 1064 +-----------+ | * +-----------+ 1065 * | * * + 1066 * | * * + 1067 * | * * + 1068 +-----------+ +-----------+ 1069 2001:db8::3 | R3 | |R6 |2001:db8::6 1070 +-----------+ +-----------+ 1071 | 1072 +-----------+ 1073 | R8 | 2001:db8::8 1074 +-----------+ 1076 In this case, PCECC could assign the SRv6 SID (in form of a IPv6 1077 address) to be used for node and adjacency. Later SRv6 path in form 1078 of list of SRv6 SID could be used at the ingress. Some examples - 1080 o SRv6 SID-List={2001:db8::8} - The best path towards R8 1082 o SRv6 SID-List={2001:db8::5, 2001:db8::8} - The path towards R8 via 1083 R5 1085 3.9. Use Cases of PCECC for SFC 1087 Service Function Chaining (SFC) is described in [RFC7665]. It is the 1088 process of directing traffic in a network such that it passes through 1089 specific hardware devices or virtual machines (known as service 1090 function nodes) that can perform particular desired functions on the 1091 traffic. The set of functions to be performed and the order in which 1092 they are to be performed is known as a service function chain. The 1093 chain is enhanced with the locations at which the service functions 1094 are to be performed to derive a Service Function Path (SFP). Each 1095 packet is marked as belonging to a specific SFP, and that marking 1096 lets each successive service function node know which functions to 1097 perform and to which service function node to send the packet next. 1098 To operate an SFC network, the service function nodes must be 1099 configured to understand the packet markings, and the edge nodes must 1100 be told how to mark packets entering the network. Additionally, it 1101 may be necessary to establish tunnels between service function nodes 1102 to carry the traffic. Planning an SFC network requires load 1103 balancing between service function nodes and traffic engineering 1104 across the network that connects them. As per [RFC8283], these are 1105 operations that can be performed by a PCE-based controller, and that 1106 controller can use PCEP to program the network and install the 1107 service function chains and any required tunnels. 1109 PCECC can play the role for setting the traffic classification rules 1110 at the classifier as well as downloading the forwarding instructions 1111 to the SFFs so that they could process the NSH and forward 1112 accordingly. 1114 [Editor's Note - more details to be added] 1116 3.10. Use Cases of PCECC for Native IP 1118 [RFC8735] describes the scenarios, and suggestions for the "Centrally 1119 Control Dynamic Routing (CCDR)" architecture, which integrates the 1120 merit of traditional distributed protocols (IGP/BGP), and the power 1121 of centrally control technologies (PCE/SDN) to provide one feasible 1122 traffic engineering solution in various complex scenarios for the 1123 service provider. [I-D.ietf-teas-pce-native-ip] defines the 1124 framework for CCDR traffic engineering within Native IP network, 1125 using Dual/Multi-BGP session strategy and CCDR architecture. PCEP 1126 protocol can be used to transfer the key parameters between PCE and 1127 the underlying network devices (PCC) using PCECC technique. The 1128 central control instructions from PCECC to identify which prefix 1129 should be advertised on which BGP session. 1131 3.11. Use Cases of PCECC for Local Protection (RSVP-TE) 1133 [I-D.cbrt-pce-stateful-local-protection] describes the need for the 1134 PCE to maintain and associate the local protection paths for the 1135 RSVP-TE LSP. Local protection requires the setup of a bypass at the 1136 PLR. This bypass can be PCC-initiated and delegated, or PCE- 1137 initiated. In either case, the PLR MUST maintain a PCEP session to 1138 the PCE. The Bypass LSPs need to mapped to the primary LSP. This 1139 could be done locally at the PLR based on a local policy but there is 1140 a need for a PCE to do the mapping as well to exert greater control. 1142 This mapping can be done via PCECC procedures where the PCE could 1143 instruct the PLR to the mapping and identify the primary LSP for 1144 which bypass should be used. 1146 3.12. Use Cases of PCECC for BIER 1148 Bit Index Explicit Replication (BIER) [RFC8279] defines an 1149 architecture where all intended multicast receivers are encoded as a 1150 bitmask in the multicast packet header within different 1151 encapsulations. A router that receives such a packet will forward 1152 the packet based on the bit position in the packet header towards the 1153 receiver(s) following a precomputed tree for each of the bits in the 1154 packet. Each receiver is represented by a unique bit in the bitmask. 1156 BIER-TE [I-D.ietf-bier-te-arch] shares architecture and packet 1157 formats with BIER. BIER-TE forwards and replicates packets based on 1158 a BitString in the packet header, but every BitPosition of the 1159 BitString of a BIER-TE packet indicates one or more adjacencies. 1160 BIER-TE Path can be derived from a PCE and used at the ingress as 1161 described in [I-D.chen-pce-bier]. 1163 Further, PCECC mechanism could be used for the allocation of bits for 1164 the BIER router for BIER as well as for the adjacencies for BIER-TE. 1165 PCECC based controller can use PCEP to instruct the BIER capable 1166 routers the meaning of the bits as well as other fields needed for 1167 BIER encapsulation. 1169 [Editor's Note - more details to be added] 1171 4. IANA Considerations 1173 This document does not require any action from IANA. 1175 5. Security Considerations 1177 TBD. 1179 6. Acknowledgments 1181 We would like to thank Adrain Farrel, Aijun Wang, Robert Tao, 1182 Changjiang Yan, Tieying Huang, Sergio Belotti, Dieter Beller, Andrey 1183 Elperin and Evgeniy Brodskiy for their useful comments and 1184 suggestions. 1186 7. References 1188 7.1. Normative References 1190 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1191 Requirement Levels", BCP 14, RFC 2119, 1192 DOI 10.17487/RFC2119, March 1997, 1193 . 1195 [RFC5440] Vasseur, JP., Ed. and JL. Le Roux, Ed., "Path Computation 1196 Element (PCE) Communication Protocol (PCEP)", RFC 5440, 1197 DOI 10.17487/RFC5440, March 2009, 1198 . 1200 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1201 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1202 May 2017, . 1204 [RFC8283] Farrel, A., Ed., Zhao, Q., Ed., Li, Z., and C. Zhou, "An 1205 Architecture for Use of PCE and the PCE Communication 1206 Protocol (PCEP) in a Network with Central Control", 1207 RFC 8283, DOI 10.17487/RFC8283, December 2017, 1208 . 1210 7.2. Informative References 1212 [RFC3985] Bryant, S., Ed. and P. Pate, Ed., "Pseudo Wire Emulation 1213 Edge-to-Edge (PWE3) Architecture", RFC 3985, 1214 DOI 10.17487/RFC3985, March 2005, 1215 . 1217 [RFC4206] Kompella, K. and Y. Rekhter, "Label Switched Paths (LSP) 1218 Hierarchy with Generalized Multi-Protocol Label Switching 1219 (GMPLS) Traffic Engineering (TE)", RFC 4206, 1220 DOI 10.17487/RFC4206, October 2005, 1221 . 1223 [RFC4364] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private 1224 Networks (VPNs)", RFC 4364, DOI 10.17487/RFC4364, February 1225 2006, . 1227 [RFC5150] Ayyangar, A., Kompella, K., Vasseur, JP., and A. Farrel, 1228 "Label Switched Path Stitching with Generalized 1229 Multiprotocol Label Switching Traffic Engineering (GMPLS 1230 TE)", RFC 5150, DOI 10.17487/RFC5150, February 2008, 1231 . 1233 [RFC5151] Farrel, A., Ed., Ayyangar, A., and JP. Vasseur, "Inter- 1234 Domain MPLS and GMPLS Traffic Engineering -- Resource 1235 Reservation Protocol-Traffic Engineering (RSVP-TE) 1236 Extensions", RFC 5151, DOI 10.17487/RFC5151, February 1237 2008, . 1239 [RFC5541] Le Roux, JL., Vasseur, JP., and Y. Lee, "Encoding of 1240 Objective Functions in the Path Computation Element 1241 Communication Protocol (PCEP)", RFC 5541, 1242 DOI 10.17487/RFC5541, June 2009, 1243 . 1245 [RFC5376] Bitar, N., Zhang, R., and K. Kumaki, "Inter-AS 1246 Requirements for the Path Computation Element 1247 Communication Protocol (PCECP)", RFC 5376, 1248 DOI 10.17487/RFC5376, November 2008, 1249 . 1251 [RFC7432] Sajassi, A., Ed., Aggarwal, R., Bitar, N., Isaac, A., 1252 Uttaro, J., Drake, J., and W. Henderickx, "BGP MPLS-Based 1253 Ethernet VPN", RFC 7432, DOI 10.17487/RFC7432, February 1254 2015, . 1256 [RFC7665] Halpern, J., Ed. and C. Pignataro, Ed., "Service Function 1257 Chaining (SFC) Architecture", RFC 7665, 1258 DOI 10.17487/RFC7665, October 2015, 1259 . 1261 [RFC8231] Crabbe, E., Minei, I., Medved, J., and R. Varga, "Path 1262 Computation Element Communication Protocol (PCEP) 1263 Extensions for Stateful PCE", RFC 8231, 1264 DOI 10.17487/RFC8231, September 2017, 1265 . 1267 [RFC8279] Wijnands, IJ., Ed., Rosen, E., Ed., Dolganow, A., 1268 Przygienda, T., and S. Aldrin, "Multicast Using Bit Index 1269 Explicit Replication (BIER)", RFC 8279, 1270 DOI 10.17487/RFC8279, November 2017, 1271 . 1273 [RFC8281] Crabbe, E., Minei, I., Sivabalan, S., and R. Varga, "Path 1274 Computation Element Communication Protocol (PCEP) 1275 Extensions for PCE-Initiated LSP Setup in a Stateful PCE 1276 Model", RFC 8281, DOI 10.17487/RFC8281, December 2017, 1277 . 1279 [RFC8355] Filsfils, C., Ed., Previdi, S., Ed., Decraene, B., and R. 1280 Shakir, "Resiliency Use Cases in Source Packet Routing in 1281 Networking (SPRING) Networks", RFC 8355, 1282 DOI 10.17487/RFC8355, March 2018, 1283 . 1285 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 1286 Decraene, B., Litkowski, S., and R. Shakir, "Segment 1287 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 1288 July 2018, . 1290 [RFC8664] Sivabalan, S., Filsfils, C., Tantsura, J., Henderickx, W., 1291 and J. Hardwick, "Path Computation Element Communication 1292 Protocol (PCEP) Extensions for Segment Routing", RFC 8664, 1293 DOI 10.17487/RFC8664, December 2019, 1294 . 1296 [RFC8735] Wang, A., Huang, X., Kou, C., Li, Z., and P. Mi, 1297 "Scenarios and Simulation Results of PCE in a Native IP 1298 Network", RFC 8735, DOI 10.17487/RFC8735, February 2020, 1299 . 1301 [RFC8751] Dhody, D., Lee, Y., Ceccarelli, D., Shin, J., and D. King, 1302 "Hierarchical Stateful Path Computation Element (PCE)", 1303 RFC 8751, DOI 10.17487/RFC8751, March 2020, 1304 . 1306 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 1307 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 1308 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 1309 . 1311 [RFC8986] Filsfils, C., Ed., Camarillo, P., Ed., Leddy, J., Voyer, 1312 D., Matsushima, S., and Z. Li, "Segment Routing over IPv6 1313 (SRv6) Network Programming", RFC 8986, 1314 DOI 10.17487/RFC8986, February 2021, 1315 . 1317 [I-D.ietf-pce-pcep-flowspec] 1318 Dhody, D., Farrel, A., and Z. Li, "PCEP Extension for Flow 1319 Specification", draft-ietf-pce-pcep-flowspec-12 (work in 1320 progress), October 2020. 1322 [I-D.ietf-pce-pcep-extension-for-pce-controller] 1323 Li, Z., Peng, S., Negi, M., Zhao, Q., and C. Zhou, "PCEP 1324 Procedures and Protocol Extensions for Using PCE as a 1325 Central Controller (PCECC) of LSPs", draft-ietf-pce-pcep- 1326 extension-for-pce-controller-10 (work in progress), 1327 January 2021. 1329 [I-D.ietf-pce-pcep-extension-pce-controller-sr] 1330 Li, Z., Peng, S., Negi, M., Zhao, Q., and C. Zhou, "PCEP 1331 Procedures and Protocol Extensions for Using PCE as a 1332 Central Controller (PCECC) for Segment Routing (SR) MPLS 1333 Segment Identifier (SID) Allocation and Distribution.", 1334 draft-ietf-pce-pcep-extension-pce-controller-sr-00 (work 1335 in progress), December 2020. 1337 [I-D.li-pce-controlled-id-space] 1338 Li, C., Chen, M., Wang, A., Cheng, W., and C. Zhou, "PCE 1339 Controlled ID Space", draft-li-pce-controlled-id-space-07 1340 (work in progress), October 2020. 1342 [I-D.ietf-pce-stateful-interdomain] 1343 Dugeon, O., Meuric, J., Lee, Y., and D. Ceccarelli, "PCEP 1344 Extension for Stateful Inter-Domain Tunnels", draft-ietf- 1345 pce-stateful-interdomain-00 (work in progress), January 1346 2021. 1348 [I-D.cbrt-pce-stateful-local-protection] 1349 Barth, C. and R. Torvi, "PCEP Extensions for RSVP-TE 1350 Local-Protection with PCE-Stateful", draft-cbrt-pce- 1351 stateful-local-protection-01 (work in progress), June 1352 2018. 1354 [I-D.ietf-pce-segment-routing-ipv6] 1355 Li, C., Negi, M., Sivabalan, S., Koldychev, M., 1356 Kaladharan, P., and Y. Zhu, "PCEP Extensions for Segment 1357 Routing leveraging the IPv6 data plane", draft-ietf-pce- 1358 segment-routing-ipv6-08 (work in progress), November 2020. 1360 [I-D.ietf-teas-pce-native-ip] 1361 Wang, A., Khasanov, B., Zhao, Q., and H. Chen, "Path 1362 Computation Element (PCE) based Traffic Engineering (TE) 1363 in Native IP Networks", draft-ietf-teas-pce-native-ip-16 1364 (work in progress), January 2021. 1366 [I-D.ietf-bier-te-arch] 1367 Eckert, T., Cauchie, G., and M. Menth, "Tree Engineering 1368 for Bit Index Explicit Replication (BIER-TE)", draft-ietf- 1369 bier-te-arch-07 (work in progress), March 2020. 1371 [I-D.chen-pce-bier] 1372 Chen, R., Zhang, Z., Dhanaraj, S., and F. Qin, "PCEP 1373 Extensions for BIER-TE", draft-chen-pce-bier-07 (work in 1374 progress), May 2020. 1376 [MAP-REDUCE] 1377 Lee, K., Choi, T., Ganguly, A., Wolinsky, D., Boykin, P., 1378 and R. Figueiredo, "Parallel Processing Framework on a P2P 1379 System Using Map and Reduce Primitives", , may 2011, 1380 . 1382 [MPLS-DC] Afanasiev, D. and D. Ginsburg, "MPLS in DC and inter-DC 1383 networks: the unified forwarding mechanism for network 1384 programmability at scale", , march 2014, 1385 . 1388 7.3. URIs 1390 [1] https://hadoop.apache.org/ 1392 Appendix A. Using reliable P2MP TE based multicast delivery for 1393 distributed computations (MapReduce-Hadoop) 1395 MapReduce model of distributed computations in computing clusters is 1396 widely deployed. In Hadoop [1] 1.0 architecture MapReduce operations 1397 on big data in the Hadoop Distributed File System (HDFS), where 1398 NameNode has the knowledge about resources of the cluster and where 1399 actual data (chunks) for particular task are located (which 1400 DataNode). Each chunk of data (64MB or more) should have 3 saved 1401 copies in different DataNodes based on their proximity. 1403 Proximity level currently has semi-manual allocation and based on 1404 Rack IDs (Assumption is that closer data are better because of access 1405 speed/smaller latency). 1407 JobTracker node is responsible for computation tasks, scheduling 1408 across DataNodes and also have Rack-awareness. Currently transport 1409 protocols between NameNode/JobTracker and DataNodes are based on IP 1410 unicast. It has simplicity as pros but has numerous drawbacks 1411 related with its flat approach. 1413 It is clear that we should go beyond of one DC for Hadoop cluster 1414 creation and move towards distributed clusters. In that case we need 1415 to handle performance and latency issues. Latency depends on speed 1416 of light in fiber links and also latency introduced by intermediate 1417 devices in between. The last one is closely correlated with network 1418 device architecture and performance. Current performance of NPU 1419 based routers should be enough for creating distribute Hadoop 1420 clusters with predicted latency. Performance of SW based routers 1421 (mainly as VNF) together with additional HW features such as DPDK are 1422 promising but require additional research and testing. 1424 Main question is how can we create simple but effective architecture 1425 for distributed Hadoop cluster? 1427 There is research [MAP-REDUCE] which show how usage of multicast tree 1428 could improve speed of resource or cluster members discovery inside 1429 the cluster as well as increase redundancy in communications between 1430 cluster nodes. 1432 Is traditional IP based multicast enough for that? We doubt it 1433 because it requires additional control plane (IGMP, PIM) and a lot of 1434 signaling, that is not suitable for high performance computations, 1435 that are very sensitive to latency. 1437 P2MP TE tunnels looks much more suitable as potential solution for 1438 creation of multicast based communications between NameNode as root 1439 and DataNodes as leaves inside the cluster. Obviously these P2MP 1440 tunnels should be dynamically created and turned down (no manual 1441 intervention). Here, the PCECC comes to play with main objective to 1442 create optimal topology of each particular request for MapReduce 1443 computation and also create P2MP tunnels with needed parameters such 1444 as bandwidth and delay. 1446 This solution would require to use MPLS label based forwarding inside 1447 the cluster. Usage of label based forwarding inside DC was proposed 1448 by Yandex [MPLS-DC]. Technically it is already possible because MPLS 1449 on switches is already supported by some vendors, MPLS also exists on 1450 Linux and OVS. 1452 The following framework can make this task: 1454 +--------+ 1455 | APP | 1456 +--------+ 1457 | NBI (REST API,...) 1458 | 1459 PCEP +----------+ REST API 1460 +---------+ +---| PCECC |----------+ 1461 | Client |---|---| | | 1462 +---------+ | +----------+ | 1463 | | | | | | 1464 +-----|---+ |PCEP| | 1465 +--------+ | | | | | 1466 | | | | | | 1467 | REST API | | | | | 1468 | | | | | | 1469 +-------------+ | | | | +----------+ 1470 | Job Tracker | | | | | | NameNode | 1471 | | | | | | | | 1472 +-------------+ | | | | +----------+ 1473 +------------------+ | +-----------+ 1474 | | | | 1475 |---+-----P2MP TE--+-----|-----------| | 1476 +----------+ +----------+ +----------+ 1477 | DataNode1| | DataNode2| | DataNodeN| 1478 |TaskTraker| |TaskTraker| .... |TaskTraker| 1479 +----------+ +----------+ +----------+ 1481 Communication between JobTracker, NameNode and PCECC can be done via 1482 REST API directly or via cluster manager such as Mesos. 1484 Phase 1: Distributed cluster resources discovery During this phase 1485 JobTracker and NameNode SHOULD identify and find available DataNodes 1486 according to computing request from application (APP). NameNode 1487 SHOULD query PCECC about available DataNodes, NameNode MAY provide 1488 additional constrains to PCECC such as topological proximity, 1489 redundancy level. 1491 PCECC SHOULD analyze the topology of distributed cluster and perform 1492 constrain based path calculation from client towards most suitable 1493 NameNodes. PCECC SHOULD reply to NameNode the list of most suitable 1494 DataNodes and their resource capabilities. Topology discovery 1495 mechanism for PCECC will be added later to that framework. 1497 Phase 2: PCECC SHOULD create P2MP LSP from client towards those 1498 DataNodes by means of PCEP messages following previously calculated 1499 path. 1501 Phase 3. NameNode SHOULD send this information to client, PCECC 1502 informs client about optimal P2MP path towards DataNodes via PCEP 1503 message. 1505 Phase 4. Client sends data blocks to those DataNodes for writing via 1506 created P2MP tunnel. 1508 When this task will be finished, P2MP tunnel could be turned down. 1510 Authors' Addresses 1512 Zhenbin (Robin) Li 1513 Huawei Technologies 1514 Huawei Bld., No.156 Beiqing Rd. 1515 Beijing 100095 1516 China 1518 Email: lizhenbin@huawei.com 1520 Dhruv Dhody 1521 Huawei Technologies 1522 Divyashree Techno Park, Whitefield 1523 Bangalore, Karnataka 560066 1524 India 1526 Email: dhruv.ietf@gmail.com 1528 Quintin Zhao 1529 Etheric Networks 1530 1009 S CLAREMONT ST 1531 SAN MATEO, CA 94402 1532 USA 1534 Email: qzhao@ethericnetworks.com 1536 King Ke 1537 Tencent Holdings Ltd. 1538 Shenzhen 1539 China 1541 Email: kinghe@tencent.com 1542 Boris Khasanov 1543 Yandex LLC 1544 Ulitsa Lva Tolstogo 16 1545 Moscow 1546 Russia 1548 Email: bhassanov@yahoo.com 1550 Luyuan Fang 1551 Expedia, Inc. 1552 USA 1554 Email: luyuanf@gmail.com 1556 Chao Zhou 1557 HPE 1559 Email: chaozhou_us@yahoo.com 1561 Boris Zhang 1562 Telus Communications 1564 Email: Boris.zhang@telus.com 1566 Artem Rachitskiy 1567 Mobile TeleSystems JLLC 1568 Nezavisimosti ave., 95 1569 Minsk 220043 1570 Belarus 1572 Email: arachitskiy@mts.by 1574 Anton Gulida 1575 LLC "Lifetech" 1576 Krasnoarmeyskaya str., 24 1577 Minsk 220030 1578 Belarus 1580 Email: anton.gulida@life.com.by