idnits 2.17.1 draft-ietf-bess-nsh-bgp-control-plane-15.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 15, 2020) is 1409 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'SPI' is mentioned on line 629, but not defined == Missing Reference: 'SI' is mentioned on line 629, but not defined == Outdated reference: A later version (-27) exists of draft-ietf-idr-rfc5575bis-25 == Outdated reference: A later version (-22) exists of draft-ietf-idr-tunnel-encaps-15 ** Obsolete normative reference: RFC 5575 (Obsoleted by RFC 8955) ** Downref: Normative reference to an Informational RFC: RFC 7665 ** Downref: Normative reference to an Informational RFC: RFC 8596 == Outdated reference: A later version (-06) exists of draft-dawra-idr-bgp-ls-sr-service-segments-03 Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 BESS Working Group A. Farrel 3 Internet-Draft Old Dog Consulting 4 Intended status: Standards Track J. Drake 5 Expires: December 17, 2020 E. Rosen 6 Juniper Networks 7 J. Uttaro 8 AT&T 9 L. Jalil 10 Verizon 11 June 15, 2020 13 BGP Control Plane for the Network Service Header in Service Function 14 Chaining 15 draft-ietf-bess-nsh-bgp-control-plane-15 17 Abstract 19 This document describes the use of BGP as a control plane for 20 networks that support Service Function Chaining (SFC). The document 21 introduces a new BGP address family called the SFC Address Family 22 Identifier / Subsequent Address Family Identifier (SFC AFI/SAFI) with 23 two route types. One route type is originated by a node to advertise 24 that it hosts a particular instance of a specified service function. 25 This route type also provides "instructions" on how to send a packet 26 to the hosting node in a way that indicates that the service function 27 has to be applied to the packet. The other route type is used by a 28 Controller to advertise the paths of "chains" of service functions, 29 and to give a unique designator to each such path so that they can be 30 used in conjunction with the Network Service Header defined in RFC 31 8300. 33 This document adopts the SFC architecture described in RFC 7665. 35 Status of This Memo 37 This Internet-Draft is submitted in full conformance with the 38 provisions of BCP 78 and BCP 79. 40 Internet-Drafts are working documents of the Internet Engineering 41 Task Force (IETF). Note that other groups may also distribute 42 working documents as Internet-Drafts. The list of current Internet- 43 Drafts is at https://datatracker.ietf.org/drafts/current/. 45 Internet-Drafts are draft documents valid for a maximum of six months 46 and may be updated, replaced, or obsoleted by other documents at any 47 time. It is inappropriate to use Internet-Drafts as reference 48 material or to cite them other than as "work in progress." 49 This Internet-Draft will expire on December 17, 2020. 51 Copyright Notice 53 Copyright (c) 2020 IETF Trust and the persons identified as the 54 document authors. All rights reserved. 56 This document is subject to BCP 78 and the IETF Trust's Legal 57 Provisions Relating to IETF Documents 58 (https://trustee.ietf.org/license-info) in effect on the date of 59 publication of this document. Please review these documents 60 carefully, as they describe your rights and restrictions with respect 61 to this document. Code Components extracted from this document must 62 include Simplified BSD License text as described in Section 4.e of 63 the Trust Legal Provisions and are provided without warranty as 64 described in the Simplified BSD License. 66 Table of Contents 68 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 69 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 5 70 1.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 71 2. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 6 72 2.1. Overview of Service Function Chaining . . . . . . . . . . 6 73 2.2. Control Plane Overview . . . . . . . . . . . . . . . . . 8 74 3. BGP SFC Routes . . . . . . . . . . . . . . . . . . . . . . . 12 75 3.1. Service Function Instance Route (SFIR) . . . . . . . . . 13 76 3.1.1. SFIR Pool Identifier Extended Community . . . . . . . 14 77 3.1.2. MPLS Mixed Swapping/Stacking Extended Community . . . 15 78 3.2. Service Function Path Route (SFPR) . . . . . . . . . . . 16 79 3.2.1. The SFP Attribute . . . . . . . . . . . . . . . . . . 16 80 3.2.2. General Rules For The SFP Attribute . . . . . . . . . 22 81 4. Mode of Operation . . . . . . . . . . . . . . . . . . . . . . 23 82 4.1. Route Targets . . . . . . . . . . . . . . . . . . . . . . 23 83 4.2. Service Function Instance Routes . . . . . . . . . . . . 24 84 4.3. Service Function Path Routes . . . . . . . . . . . . . . 24 85 4.4. Classifier Operation . . . . . . . . . . . . . . . . . . 26 86 4.5. Service Function Forwarder Operation . . . . . . . . . . 26 87 4.5.1. Processing With 'Gaps' in the SI Sequence . . . . . . 27 88 5. Selection within Service Function Paths . . . . . . . . . . . 29 89 6. Looping, Jumping, and Branching . . . . . . . . . . . . . . . 31 90 6.1. Protocol Control of Looping, Jumping, and Branching . . . 31 91 6.2. Implications for Forwarding State . . . . . . . . . . . . 32 92 7. Advanced Topics . . . . . . . . . . . . . . . . . . . . . . . 33 93 7.1. Correlating Service Function Path Instances . . . . . . . 33 94 7.2. Considerations for Stateful Service Functions . . . . . . 34 95 7.3. VPN Considerations and Private Service Functions . . . . 35 96 7.4. Flow Specification for SFC Classifiers . . . . . . . . . 35 97 7.5. Choice of Data Plane SPI/SI Representation . . . . . . . 37 98 7.5.1. MPLS Representation of the SPI/SI . . . . . . . . . . 38 99 7.6. MPLS Label Swapping/Stacking Operation . . . . . . . . . 38 100 7.7. Support for MPLS-Encapsulated NSH Packets . . . . . . . . 39 101 8. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . 39 102 8.1. Example Explicit SFP With No Choices . . . . . . . . . . 41 103 8.2. Example SFP With Choice of SFIs . . . . . . . . . . . . . 41 104 8.3. Example SFP With Open Choice of SFIs . . . . . . . . . . 42 105 8.4. Example SFP With Choice of SFTs . . . . . . . . . . . . . 42 106 8.5. Example Correlated Bidirectional SFPs . . . . . . . . . . 43 107 8.6. Example Correlated Asymmetrical Bidirectional SFPs . . . 43 108 8.7. Example Looping in an SFP . . . . . . . . . . . . . . . . 44 109 8.8. Example Branching in an SFP . . . . . . . . . . . . . . . 45 110 8.9. Examples of SFPs with Stateful Service Functions . . . . 45 111 8.9.1. Forward and Reverse Choice Made at the SFF . . . . . 46 112 8.9.2. Parallel End-to-End SFPs with Shared SFF . . . . . . 47 113 8.9.3. Parallel End-to-End SFPs with Separate SFFs . . . . . 49 114 8.9.4. Parallel SFPs Downstream of the Choice . . . . . . . 51 115 8.10. Examples Using IPv6 Addressing . . . . . . . . . . . . . 54 116 8.10.1. Example Explicit SFP With No Choices . . . . . . . . 56 117 8.10.2. Example SFP With Choice of SFIs . . . . . . . . . . 56 118 8.10.3. Example SFP With Open Choice of SFIs . . . . . . . . 57 119 8.10.4. Example SFP With Choice of SFTs . . . . . . . . . . 57 120 9. Security Considerations . . . . . . . . . . . . . . . . . . . 58 121 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 60 122 10.1. New BGP AF/SAFI . . . . . . . . . . . . . . . . . . . . 60 123 10.2. New BGP Path Attribute . . . . . . . . . . . . . . . . . 60 124 10.3. New SFP Attribute TLVs Type Registry . . . . . . . . . . 61 125 10.4. New SFP Association Type Registry . . . . . . . . . . . 61 126 10.5. New Service Function Type Registry . . . . . . . . . . . 62 127 10.6. New Generic Transitive Experimental Use Extended 128 Community Sub-Types . . . . . . . . . . . . . . . . . . 63 129 10.7. New BGP Transitive Extended Community Type . . . . . . . 64 130 10.8. New SFC Extended Community Sub-Types Registry . . . . . 64 131 10.9. SPI/SI Representation . . . . . . . . . . . . . . . . . 64 132 10.10. SFC SPI/SI Representation Flags Registry . . . . . . . . 65 133 11. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 65 134 12. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 65 135 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 66 136 13.1. Normative References . . . . . . . . . . . . . . . . . . 66 137 13.2. Informative References . . . . . . . . . . . . . . . . . 67 138 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 68 140 1. Introduction 142 As described in [RFC7498], the delivery of end-to-end services can 143 require a packet to pass through a series of Service Functions (SFs) 144 (e.g., WAN and application accelerators, Deep Packet Inspection (DPI) 145 engines, firewalls, TCP optimizers, and server load balancers) in a 146 specified order: this is termed "Service Function Chaining" (SFC). 147 There are a number of issues associated with deploying and 148 maintaining service function chaining in production networks, which 149 are described below. 151 Historically, if a packet needed to travel through a particular 152 service chain, the nodes hosting the service functions of that chain 153 were placed in the network topology in such a way that the packet 154 could not reach its ultimate destination without first passing 155 through all the service functions in the proper order. This need to 156 place the service functions at particular topological locations 157 limited the ability to adapt a service function chain to changes in 158 network topology (e.g., link or node failures), network utilization, 159 or offered service load. These topological restrictions on where the 160 service functions can be placed raised the following issues: 162 1. The process of configuring or modifying a service function chain 163 is operationally complex and may require changes to the network 164 topology. 166 2. Alternate or redundant service functions may need to be co- 167 located with the primary service functions. 169 3. When there is more than one path between source and destination, 170 forwarding may be asymmetric and it may be difficult to support 171 bidirectional service function chains using simple routing 172 methodologies and protocols without adding mechanisms for traffic 173 steering or traffic engineering. 175 In order to address these issues, the SFC architecture describes 176 Service Function Chains that are built in their own overlay network 177 (the service function overlay network), coexisting with other overlay 178 networks, over a common underlay network [RFC7665]. A Service 179 Function Chain is a sequence of Service Functions through which 180 packet flows that satisfy specified criteria will pass. 182 This document describes the use of BGP as a control plane for 183 networks that support Service Function Chaining (SFC). The document 184 introduces a new BGP address family called the SFC AFI/SAFI with two 185 route types. One route type is originated by a node to advertise 186 that it hosts a particular instance of a specified service function. 187 This route type also provides "instructions" on how to send a packet 188 to the hosting node in a way that indicates that the service function 189 has to be applied to the packet. The other route type is used by a 190 Controller (a centralized network component responsible for planning 191 and coordinating Service Function Chaining within the network) to 192 advertise the paths of "chains" of service functions, and to give a 193 unique designator to each such path so that they can be used in 194 conjunction with the Network Service Header [RFC8300]. 196 This document adopts the SFC architecture described in [RFC7665]. 198 1.1. Requirements Language 200 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 201 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 202 "OPTIONAL" in this document are to be interpreted as described in BCP 203 14 [RFC2119] [RFC8174] when, and only when, they appear in all 204 capitals, as shown here. 206 1.2. Terminology 208 This document uses the following terms from [RFC7665]: 210 o Bidirectional Service Function Chain 212 o Classifier 214 o Service Function (SF) 216 o Service Function Chain (SFC) 218 o Service Function Forwarder (SFF) 220 o Service Function Instance (SFI) 222 o Service Function Path (SFP) 224 o SFC branching 226 Additionally, this document uses the following terms from [RFC8300]: 228 o Network Service Header (NSH) 230 o Service Index (SI) 232 o Service Path Identifier (SPI) 234 This document introduces the following terms: 236 o Service Function Instance Route (SFIR). A new BGP Route Type 237 advertised by the node that hosts an SFI to describe the SFI and 238 to announce the way to forward a packet to the node through the 239 underlay network. 241 o Service Function Overlay Network. The logical network comprised 242 of Classifiers, SFFs, and SFIs that are connected by paths or 243 tunnels through underlay transport networks. 245 o Service Function Path Route (SFPR). A new BGP Route Type 246 originated by Controllers to advertise the details of each SFP. 248 o Service Function Type (SFT). An indication of the function and 249 features of an SFI. 251 2. Overview 253 This section provides an overview of Service Function Chaining in 254 general, and the control plane defined in this document. After 255 reading this section, readers may find it helpful to look through 256 Section 8 for some simple worked examples. 258 2.1. Overview of Service Function Chaining 260 In [RFC8300] a Service Function Chain (SFC) is an ordered list of 261 Service Functions (SFs). A Service Function Path (SFP) is an 262 indication of which instances of SFs are acceptable to be traversed 263 in an instantiation of an SFC in a service function overlay network. 264 The Service Path Identifier (SPI) is a 24-bit number that identifies 265 a specific SFP, and a Service Index (SI) is an 8-bit number that 266 identifies a specific point in that path. In the context of a 267 particular SFP (identified by an SPI), an SI represents a particular 268 Service Function, and indicates the order of that SF in the SFP. 270 Within the context of a specific SFP, an SI references a set of one 271 or more SFs. Each of those SFs may be supported by one or more 272 Service Function Instances (SFIs). Thus an SI may represent a choice 273 of SFIs of one or more Service Function Types. By deploying multiple 274 SFIs for a single SF, one can provide load balancing and redundancy. 276 A special functional element, called a Classifier, is located at each 277 ingress point to a service function overlay network. It assigns the 278 packets of a given packet flow to a specific Service Function Path. 279 This may be done by comparing specific fields in a packet's header 280 with local policy, which may be customer/network/service specific. 281 The Classifier picks an SFP and sets the SPI accordingly, it then 282 sets the SI to the value of the SI for the first hop in the SFP, and 283 then prepends a Network Services Header (NSH) [RFC8300] containing 284 the assigned SPI/SI to that packet. Note that the Classifier and the 285 node that hosts the first Service Function in a Service Function Path 286 need not be located at the same point in the service function overlay 287 network. 289 Note that the presence of the NSH can make it difficult for nodes in 290 the underlay network to locate the fields in the original packet that 291 would normally be used to constrain equal cost multipath (ECMP) 292 forwarding. Therefore, it is recommended that the node prepending 293 the NSH also provide some form of entropy indicator that can be used 294 in the underlay network. How this indicator is generated and 295 supplied, and how an SFF generates a new entropy indicator when it 296 forwards a packet to the next SFF, are out of scope of this document. 298 The Service Function Forwarder (SFF) receives a packet from the 299 previous node in a Service Function Path, removes the packet's link 300 layer or tunnel encapsulation and hands the packet and the NSH to the 301 Service Function Instance for processing. The SFI has no knowledge 302 of the SFP. 304 When the SFF receives the packet and the NSH back from the SFI it 305 must select the next SFI along the path using the SPI and SI in the 306 NSH and potentially choosing between multiple SFIs (possibly of 307 different Service Function Types) as described in Section 5. In the 308 normal case the SPI remains unchanged and the SI will have been 309 decremented to indicate the next SF along the path. But other 310 possibilities exist if the SF makes other changes to the NSH through 311 a process of re-classification: 313 o The SI in the NSH may indicate: 315 * A previous SF in the path: known as "looping" (see Section 6). 317 * An SF further down the path: known as "jumping" (see also 318 Section 6). 320 o The SPI and the SI may point to an SF on a different SFP: known as 321 "branching" (see also Section 6). 323 Such modifications are limited to within the same service function 324 overlay network. That is, an SPI is known within the scope of 325 service function overlay network. Furthermore, the new SI value is 326 interpreted in the context of the SFP identified by the SPI. 328 As described in [RFC8300], an unknown or invalid SPI is treated as an 329 error and the SFF drops the packet: such errors should be logged, and 330 such logs are subject to rate limits. 332 Also, as described in [RFC8300], an SFF receiving an SI that is 333 unknown in the context of the SPI can reduce the value to the next 334 meaningful SI value in the SFP indicated by the SPI. If no such 335 value exists or if the SFF does not support reducing the SI, the SFF 336 drops the packet and should log the event: such logs are also subject 337 to rate limits. 339 The SFF then selects an SFI that provides the SF denoted by the SPI/ 340 SI, and forwards the packet to the SFF that supports that SFI. 342 [RFC8300] makes it clear that the intended scope is for use within a 343 single provider's operational domain. 345 This document adopts the SFC architecture described in [RFC7665] and 346 adds a control plane to support the functions as described in 347 Section 2.2. An essential component of this solution is the 348 Controller. This is a network component responsible for planning 349 SFPs within the network. It gathers information about the 350 availability of SFIs and SFFs, instructs the control plane about the 351 SFPs to be programmed, and instructs the Classifiers how to assign 352 traffic flows to individual SFPs. 354 2.2. Control Plane Overview 356 To accomplish the function described in Section 2.1, this document 357 introduces the Service Function Type (SFT) that is the category of SF 358 that is supported by an SFF (such as "firewall"). An IANA registry 359 of Service Function Types is introduced in Section 10.5 and is 360 consistent with types used in other work such as 361 [I-D.dawra-idr-bgp-ls-sr-service-segments]. An SFF may support SFs 362 of multiple different SFTs, and may support multiple SFIs of each SF. 364 The registry of SFT values (see Section 10.5) is split into three 365 ranges with assignment policies per [RFC8126]: 367 o The Special Purpose SFT values range is assigned through Standards 368 Action. Values in that range are used for special SFC operations 369 and do not apply to the types SF that may be placed on the SFC. 371 o The First Come First Served range tracks assignments of STF values 372 made by any party that defines an SF type. Reference through an 373 Internet-Draft is desirable, but not required. 375 o The Private Use range is not tracked by IANA and is primarily 376 intended for use in private networks where the meaning of the SFT 377 values is locally tracked and under the control of a local 378 administrator. 380 It is envisaged that the majority of SFT values used will be assigned 381 from the First Come First Served space in the registry. This will 382 ensure interoperability especially in situations where software and 383 hardware from different vendors is deployed in the same networks, or 384 when networks are merged. However, operators of private networks may 385 choose to develop their own SFs and manage the configuration and 386 operation of their network through their own list of SFT values. 388 This document also introduces a new BGP AFI/SAFI (values to be 389 assigned by IANA) for "SFC Routes". Two SFC Route Types are defined 390 by this document: the Service Function Instance Route (SFIR), and the 391 Service Function Path Route (SFPR). As detailed in Section 3, the 392 route type is indicated by a sub-field in the NLRI. 394 o The SFIR is advertised by the node hosting the service function 395 instance (i.e., the SFF). The SFIR describes a particular 396 instance of a particular Service Function (i.e., an SFI) and the 397 way to forward a packet to it through the underlay network, i.e., 398 IP address and encapsulation information. 400 o The SFPRs are originated by Controllers. One SFPR is originated 401 for each Service Function Path. The SFPR specifies: 403 A. the SPI of the path 405 B. the sequence of SFTs and/or SFIs of which the path consists 407 C. for each such SFT or SFI, the SI that represents it in the 408 identified path. 410 This approach assumes that there is an underlay network that provides 411 connectivity between SFFs and Controllers, and that the SFFs are 412 grouped to form one or more service function overlay networks through 413 which SFPs are built. We assume the the Controllers have BGP 414 connectivity to all SFFs and all Classifiers within each service 415 function overlay network. 417 When choosing the next SFI in a path, the SFF uses the SPI and SI as 418 well as the SFT to choose among the SFIs, applying, for example, a 419 load balancing algorithm or direct knowledge of the underlay network 420 topology as described in Section 4. 422 The SFF then encapsulates the packet using the encapsulation 423 specified by the SFIR of the selected SFI and forwards the packet. 424 See Figure 1. 426 Thus the SFF can be seen as a portal in the underlay network through 427 which a particular SFI is reached. 429 Figure 1 shows a reference model for the SFC architecture. There are 430 four SFFs (SFF-1 through SFF-4) connected by tunnels across the 431 underlay network. Packets arrive at a Classifier and are channeled 432 along SFPs to destinations reachable through SFF-4. 434 SFF-1 and SFF-4 each have one instance of one SF attached (SFa and 435 SFe). SFF-2 has two types of SF attached: there is one instance of 436 one (SFc), and three instances of the other (SFb). SFF-3 has just 437 one instance of an SF (SFd), but it in this case the type of SFd is 438 the same type as SFb (SFTx). 440 This figure demonstrates how load balancing can be achieved by 441 creating several SFPs that satisfy the same SFC. Suppose an SFC 442 needs to include SFa, an SF of type SFTx, and SFc. A number of SFPs 443 can be constructed using any instance of SFb or using SFd. Load 444 balancing may be applied at two places: 446 o The Classifier may distribute different flows onto different SFPs 447 to share the load in the network and across SFIs. 449 o SFF-2 may distribute different flows (on the same SFP) to 450 different instances of SFb to share the processing load. 452 Note that, for convenience and clarity, Figure 1 shows only a few 453 tunnels between SFFs. There could be a full mesh of such tunnels, or 454 more likely, a selection of tunnels connecting key SFFs to enable the 455 construction of SFPs and to balance load and traffic in the network. 456 Further, the figure does not show any controllers: these would each 457 have BGP connectivity to the Classifier and all of the SFFs. 459 Packets 460 | | | 461 ------------ 462 | | 463 | Classifier | 464 | | 465 ------+----- 466 | 467 ---+--- --------- ------- 468 | | Tunnel | | | | 469 | SFF-1 |===============| SFF-2 |=========| SFF-4 | 470 | | | | | | 471 | | -+-----+- | | 472 | | ,,,,,,,,,,,,,,/,, \ | | 473 | | ' .........../. ' ..\...... | | 474 | | ' : SFb / : ' : \ SFc : | | 475 | | ' : ---+- : ' : --+-- : | | 476 | | ' : -| SFI | : ' : | SFI | : | | 477 | | ' : -| ----- : ' : ----- : | | 478 | | ' : | ----- : ' ......... | | 479 | | ' : ----- : ' | | 480 | | ' ............. ' | |--- Dests 481 | | ' ' | |--- Dests 482 | | ' ......... ' | | 483 | | ' : ----- : ' | | 484 | | ' : | SFI | : ' | | 485 | | ' : --+-- : ' | | 486 | | ' :SFd | : ' | | 487 | | ' ....|.... ' | | 488 | | ' | ' | | 489 | | ' SFTx | ' | | 490 | | ',,,,,,,,|,,,,,,,,' | | 491 | | | | | 492 | | ---+--- | | 493 | | | | | | 494 | |======| SFF-3 |====================| | 495 ---+--- | | ---+--- 496 | ------- | 497 ....|.... ....|.... 498 : | SFa: : | SFe: 499 : --+-- : : --+-- : 500 : | SFI | : : | SFI | : 501 : ----- : : ----- : 502 ......... ......... 504 Figure 1: The SFC Architecture Reference Model 506 As previously noted, [RFC8300] makes it clear that the mechanisms it 507 defines are intended for use within a single provider's operational 508 domain. This reduces the requirements on the control plane function. 510 [RFC7665] sets out the functions provided by a control plane for an 511 SFC network in Section 5.2. The functions are broken down into six 512 items the first four of which are completely covered by the 513 mechanisms described in this document: 515 1. Visiblity of all SFs and the SFFs through which they are reached. 517 2. Computation of SFPs and progrmming into the network. 519 3. Selection of SFIs explicitly in the SFP or dynamically within the 520 network. 522 4. Programming of SFFs with forwarding path information. 524 The fifth and six items in the list in RFC 7665 concern the use of 525 metadata. These are more peripheral to the control plane mechanisms 526 defined in this document, but are discussed in Section 4.4. 528 3. BGP SFC Routes 530 This document defines a new AFI/SAFI for BGP, known as "SFC", with an 531 NLRI that is described in this section. 533 The format of the SFC NLRI is shown in Figure 2. 535 +---------------------------------------+ 536 | Route Type (2 octets) | 537 +---------------------------------------+ 538 | Length (2 octets) | 539 +---------------------------------------+ 540 | Route Type specific (variable) | 541 +---------------------------------------+ 543 Figure 2: The Format of the SFC NLRI 545 The Route Type field determines the encoding of the rest of the route 546 type specific SFC NLRI. 548 The Length field indicates the length in octets of the route type 549 specific field of the SFC NLRI. 551 This document defines the following Route Types: 553 1. Service Function Instance Route (SFIR) 555 2. Service Function Path Route (SFPR) 557 A Service Function Instance Route (SFIR) is used to identify an SFI. 558 A Service Function Path Route (SFPR) defines a sequence of Service 559 Functions (each of which has at least one instance advertised in an 560 SFIR) that form an SFP. 562 The detailed encoding and procedures for these Route Types are 563 described in subsequent sections. 565 The SFC NLRI is carried in BGP [RFC4271] using BGP Multiprotocol 566 Extensions [RFC4760] with an Address Family Identifier (AFI) of TBD1 567 and a Subsequent Address Family Identifier (SAFI) of TBD2. The NLRI 568 field in the MP_REACH_NLRI/MP_UNREACH_NLRI attribute contains the SFC 569 NLRI, encoded as specified above. 571 In order for two BGP speakers to exchange SFC NLRIs, they MUST use 572 BGP Capabilities Advertisements to ensure that they both are capable 573 of properly processing such NLRIs. This is done as specified in 574 [RFC4760], by using capability code 1 (Multiprotocol BGP) with an AFI 575 of TBD1 and a SAFI of TBD2. 577 The nexthop field of the MP_REACH_NLRI attribute of the SFC NLRI MUST 578 be set to a loopback address of the advertising SFF. 580 3.1. Service Function Instance Route (SFIR) 582 Figure 3 shows the Route Type specific NLRI of the SFIR. 584 +--------------------------------------------+ 585 | Route Distinguisher (RD) (8 octets) | 586 +--------------------------------------------+ 587 | Service Function Type (2 octets) | 588 +--------------------------------------------+ 590 Figure 3: SFIR Route Type specific NLRI 592 Per [RFC4364] the RD field comprises a two byte Type field and a six 593 byte Value field. If two SFIRs are originated from different 594 administrative domains (within the same provier's operational 595 domain), they MUST have different RDs. In particular, SFIRs from 596 different VPNs (for different service function overlay networks) MUST 597 have different RDs, and those RDs MUST be different from any non-VPN 598 SFIRs. 600 The Service Function Type identifies the functions/features a service 601 function can offer, e.g., Classifier, firewall, load balancer. There 602 may be several SFIs that can perform a given Service Function. Each 603 node hosting an SFI MUST originate an SFIR for each type of SF that 604 it hosts (as indicated by the SFT value), and it MAY advertise an 605 SFIR for each instance of each type of SF. The minimal advertisement 606 allows construction of valid SFPs and leaves the selection of SFIs to 607 the local SFF; the detailed advertisement may have scaling concerns, 608 but allows a Controller that constructs an SFP to make an explicit 609 choice of SFI. 611 Note that a node may advertise all its SFIs of one SFT in one shot 612 using normal BGP Update packing. That is, all of the SFIRs in an 613 Update share a common Tunnel Encapsulation and Route Target (RT) 614 attribute. See also Section 3.2.1. 616 The SFIR representing a given SFI will contain an NLRI with RD field 617 set to an RD as specified above, and with SFT field set to identify 618 that SFI's Service Function Type. The values for the SFT field are 619 taken from a registry administered by IANA (see Section 10). A BGP 620 Update containing one or more SFIRs MUST also include a Tunnel 621 Encapsulation attribute [I-D.ietf-idr-tunnel-encaps]. If a data 622 packet needs to be sent to an SFI identified in one of the SFIRs, it 623 will be encapsulated as specified by the Tunnel Encapsulation 624 attribute, and then transmitted through the underlay network. 626 Note that the Tunnel Encapsulation attribute MUST contain sufficient 627 information to allow the advertising SFF to identify the overlay or 628 VPN network which a received packet is transiting. This is because 629 the [SPI, SI] in a received packet is specific to a particular 630 overlay or VPN network. 632 3.1.1. SFIR Pool Identifier Extended Community 634 This document defines a new transitive extended community [RFC4360] 635 of type TBD6 called the SFC extended community. When used with Sub- 636 Type TBD7, this is called the SFIR Pool Identifier extended 637 community. It MAY be included in SFIR advertisements, and is used to 638 indicate the identity of a pool of SFIRs to which an SFIR belongs. 639 Since an SFIR may be a member of multiple pools, multiple of these 640 extended communities may be present on a single SFIR advertisement. 642 SFIR pools allow SFIRs to be grouped for any purpose. Possible uses 643 include control plane scalability and stability. A pool identifier 644 may be included in an SFPR to indicate a set of SFIs that are 645 acceptable at a specific point on an SFP (see Section 3.2.1.3 and 646 Section 4.3). 648 The SFIR Pool Identifier extended community is encoded in 8 octets as 649 shown in Figure 4. 651 +--------------------------------------------+ 652 | Type = TBD6 (1 octet) | 653 +--------------------------------------------+ 654 | Sub-Type = TBD7 (1 octet) | 655 +--------------------------------------------+ 656 | SFIR Pool Identifier Value (6 octets) | 657 +--------------------------------------------+ 659 Figure 4: The SFIR Pool Identifier Extended Community 661 The SFIR Pool Identifier Value is encoded in a 6 octet field in 662 network byte order, and the value is unique within the scope of an 663 overlay network. This means that pool identifiers need to be 664 centrally managed, which is consistent with the assignment of SFIs to 665 pools. 667 3.1.2. MPLS Mixed Swapping/Stacking Extended Community 669 As noted in Section 3.1.1, this document defines a new transitive 670 extended community of type TBD6 called the SFC extended community. 671 When used with Sub-Type TBD8, this is called the MPLS Mixed Swapping/ 672 Stacking Labels extended community. The community is encoded as 673 shown in Figure 5. It contains a pair of MPLS labels: an SFC Context 674 Label and an SF Label as described in [RFC8595]. Each label is 20 675 bits encoded in a 3-octet (24 bit) field with 4 trailing bits that 676 MUST be set to zero. 678 +--------------------------------------------+ 679 | Type = TBD6 (1 octet) | 680 +--------------------------------------------| 681 | Sub-Type = TBD8 (1 octet) | 682 +--------------------------------------------| 683 | SFC Context Label (3 octets) | 684 +--------------------------------------------| 685 | SF Label (3 octets) | 686 +--------------------------------------------+ 688 Figure 5: The MPLS Mixed Swapping/Stacking Extended Community 690 Note that it is assumed that each SFF has one or more globally unique 691 SFC Context Labels and that the context label space and the SPI 692 address space are disjoint (i.e., a label value cannot be used both 693 to indicate an SFC context and an SPI, and it can be determined from 694 knowledge of the label spaces whether a label indicates an SFC 695 context or an SPI). 697 If an SFF supports SFP Traversal with an MPLS Label Stack it MUST 698 include this extended community with the SFIRs that it advertises. 700 See Section 7.6 for a description of how this extended community is 701 used. 703 3.2. Service Function Path Route (SFPR) 705 Figure 6 shows the Route Type specific NLRI of the SFPR. 707 +-----------------------------------------------+ 708 | Route Distinguisher (RD) (8 octets) | 709 +-----------------------------------------------+ 710 | Service Path Identifier (SPI) (3 octets) | 711 +-----------------------------------------------+ 713 Figure 6: SFPR Route Type Specific NLRI 715 Per [RFC4364] the RD field comprises a two byte Type field and a six 716 byte Value field. All SFPs MUST be associated with an RD. The 717 association of an SFP with an RD is determined by provisioning. If 718 two SFPRs are originated from different Controllers they MUST have 719 different RDs. Additionally, SFPRs from different VPNs (i.e., in 720 different service function overlay networks) MUST have different RDs, 721 and those RDs MUST be different from any non-VPN SFPRs. 723 The Service Path Identifier is defined in [RFC8300] and is the value 724 to be placed in the Service Path Identifier field of the NSH header 725 of any packet sent on this Service Function Path. It is expected 726 that one or more Controllers will originate these routes in order to 727 configure a service function overlay network. 729 The SFP is described in a new BGP Path attribute, the SFP attribute. 730 Section 3.2.1 shows the format of that attribute. 732 3.2.1. The SFP Attribute 734 [RFC4271] defines BGP Path attributes. This document introduces a 735 new Optional Transitive Path attribute called the SFP attribute with 736 value TBD3 to be assigned by IANA. The first SFP attribute MUST be 737 processed and subsequent instances MUST be ignored. 739 The common fields of the SFP attribute are set as follows: 741 o Optional bit is set to 1 to indicate that this is an optional 742 attribute. 744 o The Transitive bit is set to 1 to indicate that this is a 745 transitive attribute. 747 o The Extended Length bit is set if the length of the SFP attribute 748 is encoded in one octet (set to 0) or two octets (set to 1) as 749 described in [RFC4271]. 751 o The Attribute Type Code is set to TBD3. 753 The content of the SFP attribute is a series of Type-Length-Value 754 (TLV) constructs. Some TLVs may include sub-TLVs. All TLVs and sub- 755 TLVs have a common format that is: 757 o Type: A single octet indicating the type of the SFP attribute TLV. 758 Values are taken from the registry described in Section 10.3. 760 o Length: A two octet field indicating the length of the data 761 following the Length field counted in octets. 763 o Value: The contents of the TLV. 765 The formats of the TLVs defined in this document are shown in the 766 following sections. The presence rules and meanings are as follows. 768 o The SFP attribute contains a sequence of zero or more Association 769 TLVs. That is, the Association TLV is OPTIONAL. Each Association 770 TLV provides an association between this SFPR and another SFPR. 771 Each associated SFPR is indicated using the RD with which it is 772 advertised (we say the SFPR-RD to avoid ambiguity). 774 o The SFP attribute contains a sequence of one or more Hop TLVs. 775 Each Hop TLV contains all of the information about a single hop in 776 the SFP. 778 o Each Hop TLV contains an SI value and a sequence of one or more 779 SFT TLVs. Each SFT TLV contains an SFI reference for each 780 instance of an SF that is allowed at this hop of the SFP for the 781 specific SFT. Each SFI is indicated using the RD with which it is 782 advertised (we say the SFIR-RD to avoid ambiguity). 784 Section 6 of [RFC4271] describes the handling of malformed BGP 785 attributes, or those that are in error in some way. [RFC7606] 786 revises BGP error handling specifically for the UPDATE message, 787 provides guidelines for the authors of documents defining new 788 attributes, and revises the error handling procedures for a number of 789 existing attributes. This document introduces the SFP attribute and 790 so defines error handling as follows: 792 o When parsing a message, an unknown Attribute Type code or a length 793 that suggests that the attribute is longer than the remaining 794 message is treated as a malformed message and the "treat-as- 795 withdraw" approach used as per [RFC7606]. 797 o When parsing a message that contains an SFP attribute, the 798 following cases constitute errors: 800 1. Optional bit is set to 0 in SFP attribute. 802 2. Transitive bit is set to 0 in SFP attribute. 804 3. Unknown TLV type field found in SFP attribute. 806 4. TLV length that suggests the TLV extends beyond the end of the 807 SFP attribute. 809 5. Association TLV contains an unknown SFPR-RD. 811 6. No Hop TLV found in the SFP attribute. 813 7. No sub-TLV found in a Hop TLV. 815 8. Unknown SFIR-RD found in an SFT TLV. 817 o The errors listed above are treated as follows: 819 1., 2., 4., 6., 7.: The attribute MUST be treated as malformed 820 and the "treat-as-withdraw" approach used as per [RFC7606]. 822 3.: Unknown TLVs MUST be ignored, and message processing MUST 823 continue. 825 5., 8.: The absence of an RD with which to correlate is nothing 826 more than a soft error. The receiver SHOULD store the 827 information from the SFP attribute until a corresponding 828 advertisement is received. 830 3.2.1.1. The Association TLV 832 The Association TLV is an optional TLV in the SFP attribute. It MAY 833 be present multiple times. Each occurrence provides an association 834 with another SFP as advertised in another SFPR. The format of the 835 Association TLV is shown in Figure 7 837 +--------------------------------------------+ 838 | Type = 1 (1 octet) | 839 +--------------------------------------------| 840 | Length (2 octets) | 841 +--------------------------------------------| 842 | Association Type (1 octet) | 843 +--------------------------------------------| 844 | Associated SFPR-RD (8 octets) | 845 +--------------------------------------------| 846 | Associated SPI (3 octets) | 847 +--------------------------------------------+ 849 Figure 7: The Format of the Association TLV 851 The fields are as follows: 853 Type is set to 1 to indicate an Association TLV. 855 Length indicates the length in octets of the Association Type and 856 Associated SFPR-RD fields. The value of the Length field is 12. 858 The Association Type field indicate the type of association. The 859 values are tracked in an IANA registry (see Section 10.4). Only 860 one value is defined in this document: type 1 indicates 861 association of two unidirectional SFPs to form a bidirectional 862 SFP. An SFP attribute SHOULD NOT contain more than one 863 Association TLV with Association Type 1: if more than one is 864 present, the first one MUST be processed and subsequent instances 865 MUST be ignored. Note that documents that define new Association 866 Types must also define the presence rules for Association TLVs of 867 the new type. 869 The Associated SFPR-RD contains the RD of the associated SFP as 870 advertised in an SFPR. 872 The Associated SPI contains the SPI of the associated SFP as 873 advertised in an SFPR. 875 Association TLVs with unknown Association Type values SHOULD be 876 ignored. Association TLVs that contain an Associated SFPR-RD value 877 equal to the RD of the SFPR in which they are contained SHOULD be 878 ignored. If the Associated SPI is not equal to the SPI advertised in 879 the SFPR indicated by the Associated SFPR-RD then the Association TLV 880 SHOULD be ignored. In all three of these cases an implementation MAY 881 reject the SFP attribute as malformed and use the "treat-as-withdraw" 882 approach per [RFC7606], however implementers are cautioned that such 883 an approach may make an implementation less flexible in the event of 884 future extensions to this protocol. 886 Note that when two SFPRs reference each other using the Association 887 TLV, one SFPR advertisement will be received before the other. 888 Therefore, processing of an association MUST NOT be rejected simply 889 because the Associated SFPR-RD is unknown. 891 Further discussion of correlation of SFPRs is provided in 892 Section 7.1. 894 3.2.1.2. The Hop TLV 896 There is one Hop TLV in the SFP attribute for each hop in the SFP. 897 The format of the Hop TLV is shown in Figure 8. At least one Hop TLV 898 MUST be present in an SFP attribute. 900 +--------------------------------------------+ 901 | Type = 2 (1 octet) | 902 +--------------------------------------------| 903 | Length (2 octets) | 904 +--------------------------------------------| 905 | Service Index (1 octet) | 906 +--------------------------------------------| 907 | Hop Details (variable) | 908 +--------------------------------------------+ 910 Figure 8: The Format of the Hop TLV 912 The fields are as follows: 914 Type is set to 2 to indicate a Hop TLV. 916 Length indicates the length in octets of the Service Index and Hop 917 Details fields. 919 The Service Index is defined in [RFC8300] and is the value found 920 in the Service Index field of the NSH header that an SFF will use 921 to lookup to which next SFI a packet is to be sent. 923 The Hop Details field consists of a sequence of one or more sub- 924 TLVs. 926 Each hop of the SFP may demand that a specific type of SF is 927 executed, and that type is indicated in sub-TLVs of the Hop TLV. At 928 least one sub-TLV MUST be present. This document defines the SFT 929 Sub-TLV (see Section 3.2.1.3 and the MPLS Swapping/Stacking Sub-TLV 930 (see Section Section 3.2.1.4: other sub-TLVs may be defined in 931 future. This provides a list of which types of SF are acceptable at 932 a specific hop, and for each type it allows a degree of control to be 933 imposed on the choice of SFIs of that particular type. 935 If no Hop TLV is present in an SFP Attribute, it is a malformed 936 attribute 938 3.2.1.3. The SFT Sub-TLV 940 The SFT Sub-TLV MAY be included in the list of sub-TLVs of the Hop 941 TLV. The format of the SFT Sub-TLV is shown in Figure 9. The Sub- 942 TLV contains a list of SFIR-RD values each taken from the 943 advertisement of an SFI. Together they form a list of acceptable 944 SFIs of the indicated type. 946 +--------------------------------------------+ 947 | Type = 3 (1 octet) | 948 +--------------------------------------------| 949 | Length (2 octets) | 950 +--------------------------------------------| 951 | Service Function Type (2 octets) | 952 +--------------------------------------------| 953 | SFIR-RD List (variable) | 954 +--------------------------------------------+ 956 Figure 9: The Format of the SFT Sub-TLV 958 The fields are as follows: 960 Type is set to 3 to indicate an SFT Sub-TLV. 962 Length indicates the length in octets of the Service Function Type 963 and SFIR-RD List fields. 965 The Service Function Type value indicates the category (type) of 966 SF that is to be executed at this hop. The types are as 967 advertised for the SFs supported by the SFFs. SFT values in the 968 range 1-31 are Special Purpose SFT values and have meanings 969 defined by the documents that describe them - the value 'Change 970 Sequence' is defined in Section 6.1 of this document. 972 The hop description is further qualified beyond the specification 973 of the SFTs by listing, for each SFT in each hop, the SFIs that 974 may be used at the hop. The SFIs are identified using the SFIR- 975 RDs from the advertisements of the SFIs in the SFIRs. Note that 976 if the list contains one or more SFIR Pool Identifiers, then for 977 each the SFIR-RD list is effectively expanded to include the SFIR- 978 RD of each SFIR advertised with that SFIR Pool Identifier. An 979 SFIR-RD of value zero has special meaning as described in 980 Section 5. Each entry in the list is eight octets long, and the 981 number of entries in the list can be deduced from the value of the 982 Length field. 984 3.2.1.4. MPLS Swapping/Stacking Sub-TLV 986 The MPLS Swapping/Stacking Sub-TLV (Type value 4) is a zero length 987 sub-TLV that is OPTIONAL in the Hop TLV and is used when the data 988 representation is MPLS (see Section 7.5). When present it indicates 989 to the Classifier imposing an MPLS label stack that the current hop 990 is to use an {SFC Context Label, SF label} rather than an {SPI, SF} 991 label pair. See Section 7.6 for more details. 993 3.2.1.5. SFP Traversal With MPLS Label Stack TLV 995 The SFP Traversal With MPLS Label Stack TLV (Type value 5) is a zero 996 length TLV that can be carried in the SFP Attribute and indicates to 997 the Classifier and the SFFs on the SFP that an MPLS label stack with 998 label swapping/stacking is to be used for packets traversing the SFP. 999 All of the SFF specified at each the SFP's hops MUST have advertised 1000 an MPLS Mixed Swapping/Stacking Extended Community (see 1001 Section 3.1.2) for the SFP to be considered usable. 1003 3.2.2. General Rules For The SFP Attribute 1005 It is possible for the same SFI, as described by an SFIR, to be used 1006 in multiple SFPRs. 1008 When two SFPRs have the same SPI but different SFPR-RDs there can be 1009 three cases: 1011 o Two or more Controllers are originating SFPRs for the same SFP. 1012 In this case the content of the SFPRs is identical and the 1013 duplication is to ensure receipt and to provide Controller 1014 redundancy. 1016 o There is a transition in content of the advertised SFP and the 1017 advertisements may originate from one or more Controllers. In 1018 this case the content of the SFPRs will be different. 1020 o The reuse of an SPI may result from a configuration error. 1022 In all cases, there is no way for the receiving SFF to know which 1023 SFPR to process, and the SFPRs could be received in any order. At 1024 any point in time, when multiple SFPRs have the same SPI but 1025 different SFPR-RDs, the SFF MUST use the SFPR with the numerically 1026 lowest SFPR-RD when interpretting the RDs as 8-octet integers in 1027 network byte order. The SFF SHOULD log this occurrence to assist 1028 with debugging. 1030 Furthermore, a Controller that wants to change the content of an SFP 1031 is RECOMMENDED to use a new SPI and so create a new SFP onto which 1032 the Classifiers can transition packet flows before the SFPR for the 1033 old SFP is withdrawn. This avoids any race conditions with SFPR 1034 advertisements. 1036 Additionally, a Controller SHOULD NOT re-use an SPI after it has 1037 withdrawn the SFPR that used it until at least a configurable amount 1038 of time has passed. This timer SHOULD have a default of one hour. 1040 4. Mode of Operation 1042 This document describes the use of BGP as a control plane to create 1043 and manage a service function overlay network. 1045 4.1. Route Targets 1047 The main feature introduced by this document is the ability to create 1048 multiple service function overlay networks through the use of Route 1049 Targets (RTs) [RFC4364]. 1051 Every BGP UPDATE containing an SFIR or SFPR carries one or more RTs. 1052 The RT carried by a particular SFIR or SFPR is determined by the 1053 provisioning of the route's originator. 1055 Every node in a service function overlay network is configured with 1056 one or more import RTs. Thus, each SFF will import only the SFPRs 1057 with matching RTs allowing the construction of multiple service 1058 function overlay networks or the instantiation of Service Function 1059 Chains within an L3VPN or EVPN instance (see Section 7.3). An SFF 1060 that has a presence in multiple service function overlay networks 1061 (i.e., imports more than one RT) will usually maintain separate 1062 forwarding state for each overlay network. 1064 4.2. Service Function Instance Routes 1066 The SFIR (see Section 3.1) is used to advertise the existence and 1067 location of a specific Service Function Instance and consists of: 1069 o The RT as just described. 1071 o A Service Function Type (SFT) that is the type of service function 1072 that is provided (such as "firewall"). 1074 o A Route Distinguisher (RD) that is unique to a specific overlay. 1076 4.3. Service Function Path Routes 1078 The SFPR (see Section 3.2) describes a specific path of a Service 1079 Function Chain. The SFPR contains the Service Path Identifier (SPI) 1080 used to identify the SFP in the NSH in the data plane. It also 1081 contains a sequence of Service Indexes (SIs). Each SI identifies a 1082 hop in the SFP, and each hop is a choice between one of more SFIs. 1084 As described in this document, each Service Function Path Route is 1085 identified in the service function overlay network by an RD and an 1086 SPI. The SPI is unique within a single VPN instance supported by the 1087 underlay network. 1089 The SFPR advertisement comprises: 1091 o An RT as described in Section 4.1. 1093 o A tuple that identifies the SFPR 1095 * An RD that identifies an advertisement of an SFPR. 1097 * The SPI that uniquely identifies this path within the VPN 1098 instance distinguished by the RD. This SPI also appears in the 1099 NSH. 1101 o A series of Service Indexes. Each SI is used in the context of a 1102 particular SPI and identifies one or more SFs (distinguished by 1103 their SFTs) and for each SF a set of SFIs that instantiate the SF. 1104 The values of the SI indicate the order in which the SFs are to be 1105 executed in the SFP that is represented by the SPI. 1107 o The SI is used in the NSH to identify the entries in the SFP. 1108 Note that the SI values have meaning only relative to a specific 1109 path. They have no semantic other than to indicate the order of 1110 Service Functions within the path and are assumed to be 1111 monotonically decreasing from the start to the end of the path 1112 [RFC8300]. 1114 o Each Service Index is associated with a set of one or more Service 1115 Function Instances that can be used to provide the indexed Service 1116 Function within the path. Each member of the set comprises: 1118 * The RD used in an SFIR advertisement of the SFI. 1120 * The SFT that indicates the type of function as used in the same 1121 SFIR advertisement of the SFI. 1123 This may be summarized as follows where the notations "SFPR-RD" and 1124 "SFIR-RD" are used to distinguish the two different RDs, and where 1125 "*" indicates a multiplier: 1127 RT, {SFPR-RD, SPI}, m * {SI, {n * {SFT, p * SFIR-RD} } } 1129 Where: 1131 RT: Route Target 1133 SFPR-RD: The Route Descriptor of the Service Function Path Route 1134 advertisement 1136 SPI: Service Path Identifier used in the NSH 1138 m: The number of hops in the Service Function Path 1140 n: The number of choices of Service Function Type for a specific 1141 hop 1143 p: The number of choices of Service Function Instance for given 1144 Service Function Type in a specific hop 1146 SI: Service Index used in the NSH to indicate a specific hop 1148 SFT: The Service Function Type used in the same advertisement of 1149 the Service Function Instance Route 1151 SFIR-RD: The Route Descriptor used in an advertisement of the 1152 Service Function Instance Route 1154 That is, there can be multiple SFTs at a given hop as described in 1155 Section 5. 1157 Note that the values of SI are from the set {255, ..., 1} and are 1158 monotonically decreasing within the SFP. SIs MUST appear in order 1159 within the SFPR (i.e., monotonically decreasing) and MUST NOT appear 1160 more than once. Gaps MAY appear in the sequence as described in 1161 Section 4.5.1. Malformed SFPRs MUST be discarded and MUST cause any 1162 previous instance of the SFPR (same SFPR-RD and SPI) to be discarded. 1164 Note that if the SFIR-RD list in an SFT TLV contains one or more SFIR 1165 Pool identifiers, then in the above expression, 'p' is the sum of the 1166 number of individual SFIR-RD values and the sum for each SFIR Pool 1167 Identifier of the number of SFIRs advertised with that SFIR Pool 1168 Identifier. I.e., the list of SFIR-RD values is effectively expanded 1169 to include the SFIR-RD of each SFIR advertised with each SFIR Pool 1170 Identifier in the SFIR-RD list. 1172 The choice of SFI is explained further in Section 5. Note that an 1173 SFIR-RD value of zero has special meaning as described in that 1174 Section. 1176 4.4. Classifier Operation 1178 As shown in Figure 1, the Classifier is a component that is used to 1179 assign packets to an SFP. 1181 The Classifier is responsible for determining to which packet flow a 1182 packet belongs. The mechanism it uses to achieve that classification 1183 is out of scope of this document, but might include inspection of the 1184 packet header. The Classifier has been instructed (by the Controller 1185 or through some other configuration mechanism - see Section 7.4) 1186 which flows are to be assigned to which SFPs, and so it can impose an 1187 NSH on each packet and initialize the NSH with the SPI of the 1188 selected SFP and the SI of its first hop. 1190 Note that instructions delivered to the Classifier may include 1191 information about the metadata to encode (and the format for that 1192 encoding) on packets that are classified by the Classifier to a 1193 particular SFP. As mentioned in Section 2.2, this corresponds to the 1194 fifth element of control plane functionality described in [RFC7665]. 1195 Such instructions fall outside the scope of this specification 1196 (although, see Section 7.4), as do instructions to other SFC elements 1197 on how to interpret metadata (as described in the sixth element of 1198 control plane functionality described in [RFC7665]. 1200 4.5. Service Function Forwarder Operation 1202 Each packet sent to an SFF is transmitted encapsulated in an NSH. 1203 The NSH includes an SPI and SI: the SPI indicates the SFPR 1204 advertisement that announced the Service Function Path; the tuple 1205 SPI/SI indicates a specific hop in a specific path and maps to the 1206 RD/SFT of a particular SFIR advertisement. 1208 When an SFF gets an SFPR advertisement it will first determine 1209 whether to import the route by examining the RT. If the SFPR is 1210 imported the SFF then determines whether it is on the SFP by looking 1211 for its own SFIR-RDs or any SFIR-RD with value zero in the SFPR. For 1212 each occurrence in the SFP, the SFF creates forwarding state for 1213 incoming packets and forwarding state for outgoing packets that have 1214 been processed by the specified SFI. 1216 The SFF creates local forwarding state for packets that it receives 1217 from other SFFs. This state makes the association between the SPI/SI 1218 in the NSH of the received packet and one or more specific local SFIs 1219 as identified by the SFIR-RD/SFT. If there are multiple local SFIs 1220 that match this is because a single advertisement was made for a set 1221 of equivalent SFIs and the SFF may use local policy (such as load 1222 balancing) to determine to which SFI to forward a received packet. 1224 The SFF also creates next hop forwarding state for packets received 1225 back from the local SFI that need to be forwarded to the next hop in 1226 the SFP. There may be a choice of next hops as described in 1227 Section 4.3. The SFF could install forwarding state for all 1228 potential next hops, or it could choose to only install forwarding 1229 state to a subset of the potential next hops. If a choice is made 1230 then it will be as described in Section 5. 1232 The installed forwarding state may change over time reacting to 1233 changes in the underlay network and the availability of particular 1234 SFIs. Note that the forwarding state describes how one SFF send 1235 packets to another SFF, but not how those packets are routed through 1236 the underlay network. SFFs may be connected by tunnels across the 1237 underlay, or packets may be sent addressed to the next SFF and routed 1238 through the underlay. In any case, transmission across the underlay 1239 requires encapsulation of packets with a header for transport in the 1240 underlay network. 1242 Note that SFFs only create and store forwarding state for the SFPs on 1243 which they are included. They do not retain state for all SFPs 1244 advertised. 1246 An SFF may also install forwarding state to support looping, jumping, 1247 and branching. The protocol mechanism for explicit control of 1248 looping, jumping, and branching uses a specific reserved SFT value at 1249 a given hop of an SFPR and is described in Section 6.1. 1251 4.5.1. Processing With 'Gaps' in the SI Sequence 1253 The behavior of an SF as described in [RFC8300] is to decrement the 1254 value of the SI field in the NSH by one before returning a packet to 1255 the local SFF for further processing. This means that there is a 1256 good reason to assume that the SFP is composed of a series of SFs 1257 each indicated by an SI value one less than the previous. 1259 However, there is an advantage to having non-successive SIs in an 1260 SPI. Consider the case where an SPI needs to be modified by the 1261 insertion or removal of an SF. In the latter case this would lead to 1262 a "gap" in the sequence of SIs, and in the former case, this could 1263 only be achieved if a gap already existed into which the new SF with 1264 its new SI value could be inserted. Otherwise, all "downstream" SFs 1265 would need to be renumbered. 1267 Now, of course, such renumbering could be performed, but would lead 1268 to a significant disruption to the SFC as all the SFFs along the SFP 1269 were "reprogrammed". Thus, to achieve dynamic modification of an SFP 1270 (and even, in-service modification) it is desirable to be able to 1271 make these modifications without changing the SIs of the elements 1272 that were present before the modification. This will produce much 1273 more consistent/predictable behavior during the convergence period 1274 where otherwise the change would need to be fully propagated. 1276 Another approach says that any change to an SFP simply creates a new 1277 SFP that can be assigned a new SPI. All that would be needed would 1278 be to give a new instruction to the Classifier and traffic would be 1279 switched to the new SFP that contains the new set of SFs. This 1280 approach is practical, but neglects to consider that the SFP may be 1281 referenced by other SFPs (through "branch" instructions) and used by 1282 many Classifiers. In those cases the corresponding configuration 1283 resulting from a change in SPI may have wide ripples and give scope 1284 for errors that are hard to trace. 1286 Therefore, while this document requires that the SI values in an SFP 1287 are monotonic decreasing, it makes no assumption that the SI values 1288 are sequential. Configuration tools may apply that rule, but they 1289 are not required to. To support this, an SFF SHOULD process as 1290 follows when it receives a packet: 1292 o If the SI indicates a known entry in the SFP, the SFF MUST process 1293 the packet as normal, looking up the SI and determining to which 1294 local SFI to deliver the packet. 1296 o If the SI does not match an entry in the SFP, the SFF MUST reduce 1297 the SI value to the next (smaller) value present in the SFP and 1298 process the packet using that SI. 1300 o If there is no smaller SI (i.e., if the end of the SFP has been 1301 reached) the SFF MUST treat the SI value as invalid as described 1302 in [RFC8300]. 1304 This makes the behavior described in this document a superset of the 1305 function in [RFC8300]. That is, an implementation that strictly 1306 follows RFC 8300 in performing SI decrements in units of one, is 1307 perfectly in line with the mechanisms defined in this document. 1309 SFF implementations MAY choose to only support contiguous SI values 1310 in an SFP. Such an implementation will not support receiving an SI 1311 value that is not present in the SFP and will discard the packets as 1312 described in [RFC8300]. 1314 5. Selection within Service Function Paths 1316 As described in Section 2 the SPI/SI in the NSH passed back from an 1317 SFI to the SFF may leave the SFF with a choice of next hop SFTs, and 1318 a choice of SFIs for each SFT. That is, the SPI indicates an SFPR, 1319 and the SI indicates an entry in that SFPR. Each entry in an SFPR is 1320 a set of one or more SFT/SFIR-RD pairs. The SFF MUST choose one of 1321 these, identify the SFF that supports the chosen SFI, and send the 1322 packet to that next hop SFF. 1324 The choice be may offered for load balancing across multiple SFIs, or 1325 for discrimination between different actions necessary at a specific 1326 hop in the SFP. Different SFT values may exist at a given hop in an 1327 SFP to support several cases: 1329 o There may be multiple instances of similar service functions that 1330 are distinguished by different SFT values. For example, firewalls 1331 made by vendor A and vendor B may need to be identified by 1332 different SFT values because, while they have similar 1333 functionality, their behavior is not identical. Then, some SFPs 1334 may limit the choice of SF at a given hop by specifying the SFT 1335 for vendor A, but other SFPs might not need to control which 1336 vendor's SF is used and so can indicate that either SFT can be 1337 used. 1339 o There may be an obvious branch needed in an SFP such as the 1340 processing after a firewall where admitted packets continue along 1341 the SFP, but suspect packets are diverted to a "penalty box". In 1342 this case, the next hop in the SFP will be indicated with two 1343 different SFT values. 1345 In the typical case, the SFF chooses a next hop SFF by looking at the 1346 set of all SFFs that support the SFs identified by the SI (that set 1347 having been advertised in individual SFIR advertisements), finding 1348 the one or more that are "nearest" in the underlay network, and 1349 choosing between next hop SFFs using its own load-balancing 1350 algorithm. 1352 An SFI may influence this choice process by passing additional 1353 information back along with the packet and NSH. This information may 1354 influence local policy at the SFF to cause it to favor a next hop SFF 1355 (perhaps selecting one that is not nearest in the underlay), or to 1356 influence the load-balancing algorithm. 1358 This selection applies to the normal case, but also applies in the 1359 case of looping, jumping, and branching (see Section 6). 1361 Suppose an SFF in a particular service overlay network (identified by 1362 a particular import RT, RT-z) needs to forward an NSH-encapsulated 1363 packet whose SPI is SPI-x and whose SI is SI-y. It does the 1364 following: 1366 1. It looks for an installed SFPR that carries RT-z and that has 1367 SPI-x in its NLRI. If there is none, then such packets cannot be 1368 forwarded. 1370 2. From the SFP attribute of that SFPR, it finds the Hop TLV with SI 1371 value set to SI-y. If there is no such Hop TLV, then such 1372 packets cannot be forwarded. 1374 3. It then finds the "relevant" set of SFIRs by going through the 1375 list of SFT TLVs contained in the Hop TLV as follows: 1377 A. An SFIR is relevant if it carries RT-z, the SFT in its NLRI 1378 matches the SFT value in one of the SFT TLVs, and the RD 1379 value in its NLRI matches an entry in the list of SFIR-RDs in 1380 that SFT TLV. 1382 B. If an entry in the SFIR-RD list of an SFT TLV contains the 1383 value zero, then an SFIR is relevant if it carries RT-z and 1384 the SFT in its NLRI matches the SFT value in that SFT TLV. 1385 I.e., any SFIR in the service function overlay network 1386 defined by RT-z and with the correct SFT is relevant. 1388 C. If a pool identifier is in use then an SFIR is relevant if it 1389 is a member of the pool. 1391 Each of the relevant SFIRs identifies a single SFI, and contains a 1392 Tunnel Encapsulation attribute that specifies how to send a packet to 1393 that SFI. For a particular packet, the SFF chooses a particular SFI 1394 from the set of relevant SFIRs. This choice is made according to 1395 local policy. 1397 A typical policy might be to figure out the set of SFIs that are 1398 closest, and to load balance among them. But this is not the only 1399 possible policy. 1401 Thus, at any point in time when an SFF selects its next hop, it 1402 chooses from the intersection of the set of next hop RDs contained in 1403 the SFPR and the RDs contained in the SFF's local set of SFIRs (i.e., 1404 according to the determination of "relevance", above). If the 1405 intersection is null, the SFPR is unusable. Similarly, when this 1406 condition applies on the controller that originated the SFPR, it 1407 SHOULD either withdraw the SFPR or re-advertise it with a new set of 1408 RDs for the affected hop. 1410 6. Looping, Jumping, and Branching 1412 As described in Section 2 an SFI or an SFF may cause a packet to 1413 "loop back" to a previous SF on a path in order that a sequence of 1414 functions may be re-executed. This is simply achieved by replacing 1415 the SI in the NSH with a higher value instead of decreasing it as 1416 would normally be the case to determine the next hop in the path. 1418 Section 2 also describes how an SFI or an SFF may cause a packets to 1419 "jump forward" to an SF on a path that is not the immediate next SF 1420 in the SFP. This is simply achieved by replacing the SI in the NSH 1421 with a lower value than would be achieved by decreasing it by the 1422 normal amount. 1424 A more complex option to move packets from one SFP to another is 1425 described in [RFC8300] and Section 2 where it is termed "branching". 1426 This mechanism allows an SFI or SFF to make a choice of downstream 1427 treatments for packets based on local policy and output of the local 1428 SF. Branching is achieved by changing the SPI in the NSH to indicate 1429 the new path and setting the SI to indicate the point in the path at 1430 which the packets enter. 1432 Note that the NSH does not include a marker to indicate whether a 1433 specific packet has been around a loop before. Therefore, the use of 1434 NSH metadata ([RFC8300]) may be required in order to prevent infinite 1435 loops. 1437 6.1. Protocol Control of Looping, Jumping, and Branching 1439 If the SFT value in an SFT TLV in an SFPR has the Special Purpose SFT 1440 value "Change Sequence" (see Section 10) then this is an indication 1441 that the SFF may make a loop, jump, or branch according to local 1442 policy and information returned by the local SFI. 1444 In this case, the SPI and SI of the next hop are encoded in the eight 1445 bytes of an entry in the SFIR-RD list as follows: 1447 3 bytes SPI 1448 1 bytes SI 1450 4 bytes Reserved (SHOULD be set to zero and ignored) 1452 If the SI in this encoding is not part of the SFPR indicated by the 1453 SPI in this encoding, then this is an explicit error that SHOULD be 1454 detected by the SFF when it parses the SFPR. The SFPR SHOULD NOT 1455 cause any forwarding state to be installed in the SFF and packets 1456 received with the SPI that indicates this SFPR SHOULD be silently 1457 discarded. 1459 If the SPI in this encoding is unknown, the SFF SHOULD NOT install 1460 any forwarding state for this SFPR, but MAY hold the SFPR pending 1461 receipt of another SFPR that does use the encoded SPI. 1463 If the SPI matches the current SPI for the path, this is a loop or 1464 jump. In this case, if the SI is greater than to the current SI it 1465 is a loop. If the SPI matches and the SI is less than the next SI, 1466 it is a jump. 1468 If the SPI indicates another path, this is a branch and the SI 1469 indicates the point at which to enter that path. 1471 The Change Sequence SFT is just another SFT that may appear in a set 1472 of SFI/SFT tuples within an SI and is selected as described in 1473 Section 5. 1475 Note that Special Purpose SFTs MUST NOT be advertised in SFIRs. If 1476 such an SFIR is received it SHOULD be ignored. 1478 6.2. Implications for Forwarding State 1480 Support for looping and jumping requires that the SFF has forwarding 1481 state established to an SFF that provides access to an instance of 1482 the appropriate SF. This means that the SFF must have seen the 1483 relevant SFIR advertisements and known that it needed to create the 1484 forwarding state. This is a matter of local configuration and 1485 implementation: for example, an implementation could be configured to 1486 install forwarding state for specific looping/jumping. 1488 Support for branching requires that the SFF has forwarding state 1489 established to an SFF that provides access to an instance of the 1490 appropriate entry SF on the other SFP. This means that the SFF must 1491 have seen the relevant SFIR and SFPR advertisements and known that it 1492 needed to create the forwarding state. This is a matter of local 1493 configuration and implementation: for example, an implementation 1494 could be configured to install forwarding state for specific 1495 branching (identified by SPI and SI). 1497 7. Advanced Topics 1499 This section highlights several advanced topics introduced elsewhere 1500 in this document. 1502 7.1. Correlating Service Function Path Instances 1504 It is often useful to create bidirectional SFPs to enable packet 1505 flows to traverse the same set of SFs, but in the reverse order. 1506 However, packets on SFPs in the data plane (per [RFC8300]) do not 1507 contain a direction indicator, so each direction must use a different 1508 SPI. 1510 As described in Section 3.2.1.1 an SFPR can contain one or more 1511 correlators encoded in Association TLVs. If the Association Type 1512 indicates "Bidirectional SFP" then the SFP advertised in the SFPR is 1513 one direction of a bidirectional pair of SFPs where the other in the 1514 pair is advertised in the SFPR with RD as carried in the Associated 1515 SFPR-RD field of the Association TLV. The SPI carried in the 1516 Associated SPI field of the Association TLV provides a cross-check 1517 against the SPI advertised in the SFPR with RD as carried in the 1518 Associated SFPR-RD field of the Association TLV. 1520 As noted in Section 3.2.1.1, when SFPRs reference each other, one 1521 SFPR advertisement will be received before the other. Therefore, 1522 processing of an association will require that the first SFPR is not 1523 rejected simply because the Associated SFPR-RD it carries is unknown. 1524 However, the SFP defined by the first SFPR is valid and SHOULD be 1525 available for use as a unidirectional SFP even in the absence of an 1526 advertisement of its partner. 1528 Furthermore, in error cases where SFPR-a associates with SFPR-b, but 1529 SFPR-b associates with SFPR-c such that a bidirectional pair of SFPs 1530 cannot be formed, the individual SFPs are still valid and SHOULD be 1531 available for use as unidirectional SFPs. An implementation SHOULD 1532 log this situation because it represents a Controller error. 1534 Usage of a bidirectional SFP may be programmed into the Classifiers 1535 by the Controller. Alternatively, a Classifier may look at incoming 1536 packets on a bidirectional packet flow, extract the SPI from the 1537 received NSH, and look up the SFPR to find the reverse direction SFP 1538 to use when it sends packets. 1540 See Section 8 for an example of how this works. 1542 7.2. Considerations for Stateful Service Functions 1544 Some service functions are stateful. That means that they build and 1545 maintain state derived from configuration or from the packet flows 1546 that they handle. In such cases it can be important or necessary 1547 that all packets from a flow continue to traverse the same instance 1548 of a service function so that the state can be leveraged and does not 1549 need to be regenerated. 1551 In the case of bidirectional SFPs, it may be necessary to traverse 1552 the same instances of a stateful service function in both directions. 1553 A firewall is a good example of such a service function. 1555 This issue becomes a concern where there are multiple parallel 1556 instances of a service function and a determination of which one to 1557 use could normally be left to the SFF as a load-balancing or local 1558 policy choice. 1560 For the forward direction SFP, the concern is that the same choice of 1561 service function is made for all packets of a flow under normal 1562 network conditions. It may be possible to guarantee that the load 1563 balancing functions applied in the SFFs are stable and repeatable, 1564 but a controller that constructs SFPs might not want to trust to 1565 this. The controller can, in these cases, build a number of more 1566 specific SFPs each traversing a specific instance of the stateful 1567 SFs. In this case, the load balancing choice can be left up to the 1568 Classifier. Thus the Classifier selects which instance of a stateful 1569 SF is used by a particular flow by selecting the SFP that the flow 1570 uses. 1572 For bidirectional SFPs where the same instance of a stateful SF must 1573 be traversed in both directions, it is not enough to leave the choice 1574 of service function instance as a local choice even if the load 1575 balancing is stable because coordination would be required between 1576 the decision points in the forward and reverse directions and this 1577 may be hard to achieve in all cases except where it is the same SFF 1578 that makes the choice in both directions. 1580 Note that this approach necessarily increases the amount of SFP state 1581 in the network (i.e., there are more SFPs). It is possible to 1582 mitigate this effect by careful construction of SFPs built from a 1583 concatenation of other SFPs. 1585 Section 8.9 includes some simple examples of SFPs for stateful 1586 service functions. 1588 7.3. VPN Considerations and Private Service Functions 1590 Likely deployments include reserving specific instances of Service 1591 Functions for specific customers or allowing customers to deploy 1592 their own Service Functions within the network. Building Service 1593 Functions in such environments requires that suitable identifiers are 1594 used to ensure that SFFs distinguish which SFIs can be used and which 1595 cannot. 1597 This problem is similar to how VPNs are supported and is solved in a 1598 similar way. The RT field is used to indicate a set of Service 1599 Functions from which all choices must be made. 1601 7.4. Flow Specification for SFC Classifiers 1603 [RFC5575] and [I-D.ietf-idr-rfc5575bis] define a set of BGP routes 1604 that can be used to identify the packets in a given flow using fields 1605 in the header of each packet, and a set of actions, encoded as 1606 extended communities, that can be used to disposition those packets. 1607 This document enables the use of these mechanisms by SFC Classifiers 1608 by defining a new action extended community called "Flow 1609 Specification for SFC Classifiers" identified by the value TBD4. 1610 Note that implementation of this specification MUST NOT include other 1611 action extended communities at the same time as an SFC Classifier: 1612 the inclusion of the "Flow Specification for SFC Classifiers" action 1613 extended community along with any other action MUST be treated by 1614 implementation of this specification as an error which SHOULD result 1615 in the Flow Specification UPDATE message being handled as Treat-as- 1616 withdraw according to [RFC7606] Section 2. 1618 To put the Flow Specification into context when multiple SFC overlays 1619 are present in one network, each FlowSpec update MUST be tagged with 1620 the route target of the overlay or VPN network for which it is 1621 intended. 1623 This extended community is encoded as an 8-octet value, as shown in 1624 Figure 10. 1626 1 2 3 1627 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1628 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1629 | Type=0x80 | Sub-Type=TBD4 | SPI | 1630 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1631 | SPI (cont.) | SI | SFT | 1632 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1634 Figure 10: The Format of the Flow Specification for SFC Classifiers 1635 Extended Community 1637 The extended community contains the Service Path Identifier (SPI), 1638 Service Index (SI), and Service Function Type (SFT) as defined 1639 elsewhere in this document. Thus, each action extended community 1640 defines the entry point (not necessarily the first hop) into a 1641 specific service function path. This allows, for example, different 1642 flows to enter the same service function path at different points. 1644 Note that a given Flow Specification update according to [RFC5575] 1645 and [I-D.ietf-idr-rfc5575bis] may include multiple of these action 1646 extended communities, and that if a given action extended community 1647 does not contain an installed SFPR with the specified {SPI, SI, SFT} 1648 it MUST NOT be used for dispositioning the packets of the specified 1649 flow. 1651 The normal case of packet classification for SFC will see a packet 1652 enter the SFP at its first hop. In this case the SI in the extended 1653 community is superfluous and the SFT may also be unnecessary. To 1654 allow these cases to be handled, a special meaning is assigned to a 1655 Service Index of zero (not a valid value) and an SFT of zero (a 1656 reserved value in the registry - see Section 10.5). 1658 o If an SFC Classifiers Extended Community is received with SI = 0 1659 then it means that the first hop of the SFP indicated by the SPI 1660 MUST be used. 1662 o If an SFC Classifiers Extended Community is received with SFT = 0 1663 then there are two sub-cases: 1665 * If there is a choice of SFT in the hop indicated by the value 1666 of the SI (including SI = 0) then SFT = 0 means there is a free 1667 choice according to local policy of which SFT to use). 1669 * If there is no choice of SFT in the hop indicated by the value 1670 of SI, then SFT = 0 means that the value of the SFT at that hop 1671 as indicated in the SFPR for the indicated SPI MUST be used. 1673 One of the filters that the Flow Specification may describe is the 1674 VPN to which the traffic belongs. Additionally, as noted above, to 1675 put the indicated SPI into context when multiple SFC overlays are 1676 present in one network, each FlowSpec update MUST be tagged with the 1677 route target of the overlay or VPN network for which it is intended. 1679 Note that future extensions might be made to the Flow Specification 1680 for SFC Classifiers Extended Community to provide instruction to the 1681 Classifier about what metadata to add to packets that it classifies 1682 for forwarding on a specific SFP, but that is outside the scope of 1683 this document. 1685 7.5. Choice of Data Plane SPI/SI Representation 1687 This document ties together the control and data planes of an SFC 1688 overlay network through the use of the SPI/SI which is nominally 1689 carried in the NSH of a given packet. However, in order to handle 1690 situations in which the NSH is not ubiquitously deployed, it is also 1691 possible to use alternative data plane representations of the SPI/SI 1692 by carrying the identical semantics in other protocol fields such as 1693 MPLS labels [RFC8595]. 1695 This document defines a new sub-TLV for the Tunnel Encapsulation 1696 attribute [I-D.ietf-idr-tunnel-encaps], the SPI/SI Representation 1697 sub-TLV of type TBD5. This sub-TLV MAY be present in each Tunnel TLV 1698 contained in a Tunnel Encapsulation attribute when the attribute is 1699 carried by an SFIR. The value field of this sub-TLV is a two octet 1700 field of flags numbered counting from the the most significant bit, 1701 each of which describes how the originating SFF expects to see the 1702 SPI/SI represented in the data plane for packets carried in the 1703 tunnels described by the Tunnel TLV. 1705 The following bits are defined by this document and are tracked in an 1706 IANA registry desribed in Section 10.10: 1708 Bit TBD9: If this bit is set the NSH is to be used to carry the SPI/ 1709 SI in the data plane. 1711 Bit TBD10: If this bit is set two labels in an MPLS label stack are 1712 to be used as described in Section 7.5.1. 1714 If a given Tunnel TLV does not contain an SPI/SI Representation sub- 1715 TLV then it MUST be processed as if such a sub-TLV is present with 1716 Bit TBD9 set and no other bits set. That is, the absence of the sub- 1717 TLV SHALL be interpreted to mean that the NSH is to be used. 1719 If a given Tunnel TLV contains an SPI/SI Representation sub-TLV with 1720 value field that has no flag set then the tunnel indicated by the 1721 Tunnel TLV MUST NOT be used for forwarding SFC packets. If a given 1722 Tunnel TLV contains an SPI/SI Representation sub-TLV with both bit 1723 TBD9 and bit TBD10 set then the tunnel indicated by the Tunnel TLV 1724 MUST NOT be used for forwarding SFC packets. The meaning and rules 1725 for presence of other bits is to be defined in future documents, but 1726 implementations of this specification MUST set other bits to zero and 1727 ignore them on receipt. 1729 If a given Tunnel TLV contains more than one SPI/SI Representation 1730 sub-TLV then the first one MUST be considered and subsequent 1731 instances MUST be ignored. 1733 Note that the MPLS representation of the logical NSH may be used even 1734 if the tunnel is not an MPLS tunnel. Conversely, MPLS tunnels may be 1735 used to carry other encodings of the logical NSH (specifically, the 1736 NSH itself). It is a requirement that both ends of a tunnel over the 1737 underlay network know that the tunnel is used for SFC and know what 1738 form of NSH representation is used. The signaling mechanism 1739 described here allows coordination of this information. 1741 7.5.1. MPLS Representation of the SPI/SI 1743 If bit TBD10 is set in the in the SPI/SI Representation sub-TLV then 1744 labels in the MPLS label stack are used to indicate SFC forwarding 1745 and processing instructions to achieve the semantics of a logical 1746 NSH. The label stack is encoded as shown in [RFC8595]. 1748 7.6. MPLS Label Swapping/Stacking Operation 1750 When a Classifier constructs an MPLS label stack for an SFP it starts 1751 with that SFP's last hop. If the last hop requires an {SPI, SI} 1752 label pair for label swapping, it pushes the SI (set to the SI value 1753 of the last hop) and the SFP's SPI onto the MPLS label stack. If the 1754 last hop requires a {context label, SFI label} label pair for label 1755 stacking it selects a specific SFIR and pushes that SFIR's SFI label 1756 and context label onto the MPLS label stack. 1758 The Classifier then moves sequentially back through the SFP one hop 1759 at a time. For each hop, if the hop requires an {SPI, SI]} and there 1760 is an {SPI, SI} at the top of the MPLS label stack, the SI is set to 1761 the SI value of the current hop. If there is not an {SPI, SI} at the 1762 top of the MPLS label stack, it pushes the SI (set to the SI value of 1763 the current hop) and the SFP's SPI onto the MPLS label stack. 1765 If the hop requires a {context label, SFI label}, it selects a 1766 specific SFIR and pushes that SFIR's SFI label and context label onto 1767 the MPLS label stack. 1769 7.7. Support for MPLS-Encapsulated NSH Packets 1771 [RFC8596] describes how to transport SFC packets using the NSH over 1772 an MPLS transport network. Signaling MPLS encapsulation of SFC 1773 packets using the NSH is also supported by this document by using the 1774 "BGP Tunnel Encapsulation Attribute Sub-TLV" with the codepoint 10 1775 (representing "MPLS Label Stack") from the "BGP Tunnel Encapsulation 1776 Attribute Sub-TLVs" registry defined in [I-D.ietf-idr-tunnel-encaps], 1777 and also using the "SFP Traversal With MPLS Label Stack TLV" and the 1778 "SPI/SI Representation sub-TLV" with bit TBD9 set and bit TBD10 1779 cleared. 1781 In this case the MPLS label stack constructed by the SFF to forward a 1782 packet to the next SFF on the SFP will consist of the labels needed 1783 to reach that SFF, and if label stacking is used it will also include 1784 the labels advertised in the MPLS Label Stack sub-TLV and the labels 1785 remaining in the stack needed to traverse the remainder of the SFP. 1787 8. Examples 1789 Most of the examples in this section use IPv4 addressing. But there 1790 is nothing special about IPv4 in the mechanisms described in this 1791 document, and they are equally applicable to IPv6. A few examples 1792 using IPv6 addressing are provided in Section 8.10. 1794 Assume we have a service function overlay network with four SFFs 1795 (SFF1, SFF3, SFF3, and SFF4). The SFFs have addresses in the 1796 underlay network as follows: 1798 SFF1 192.0.2.1 1799 SFF2 192.0.2.2 1800 SFF3 192.0.2.3 1801 SFF4 192.0.2.4 1803 Each SFF provides access to some SFIs from the four Service Function 1804 Types SFT=41, SFT=42, SFT=43, and SFT=44 as follows: 1806 SFF1 SFT=41 and SFT=42 1807 SFF2 SFT=41 and SFT=43 1808 SFF3 SFT=42 and SFT=44 1809 SFF4 SFT=43 and SFT=44 1811 The service function network also contains a Controller with address 1812 198.51.100.1. 1814 This example service function overlay network is shown in Figure 11. 1816 -------------- 1817 | Controller | 1818 | 198.51.100.1 | ------ ------ ------ ------ 1819 -------------- | SFI | | SFI | | SFI | | SFI | 1820 |SFT=41| |SFT=42| |SFT=41| |SFT=43| 1821 ------ ------ ------ ------ 1822 \ / \ / 1823 --------- --------- 1824 ---------- | SFF1 | | SFF2 | 1825 Packet --> | | |192.0.2.1| |192.0.2.2| 1826 Flows --> |Classifier| --------- --------- -->Dest 1827 | | --> 1828 ---------- --------- --------- 1829 | SFF3 | | SFF4 | 1830 |192.0.2.3| |192.0.2.4| 1831 --------- --------- 1832 / \ / \ 1833 ------ ------ ------ ------ 1834 | SFI | | SFI | | SFI | | SFI | 1835 |SFT=42| |SFT=44| |SFT=43| |SFT=44| 1836 ------ ------ ------ ------ 1838 Figure 11: Example Service Function Overlay Network 1840 The SFFs advertise routes to the SFIs they support. So we see the 1841 following SFIRs: 1843 RD = 192.0.2.1/1, SFT = 41 1844 RD = 192.0.2.1/2, SFT = 42 1845 RD = 192.0.2.2/1, SFT = 41 1846 RD = 192.0.2.2/2, SFT = 43 1847 RD = 192.0.2.3/7, SFT = 42 1848 RD = 192.0.2.3/8, SFT = 44 1849 RD = 192.0.2.4/5, SFT = 43 1850 RD = 192.0.2.4/6, SFT = 44 1852 Note that the addressing used for communicating between SFFs is taken 1853 from the Tunnel Encapsulation attribute of the SFIR and not from the 1854 SFIR-RD. 1856 8.1. Example Explicit SFP With No Choices 1858 Consider the following SFPR. 1860 SFP1: RD = 198.51.100.1/101, SPI = 15, 1861 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1862 [SI = 250, SFT = 43, RD = 192.0.2.2/2] 1864 The Service Function Path consists of an SF of type 41 located at 1865 SFF1 followed by an SF of type 43 located at SFF2. This path is 1866 fully explicit and each SFF is offered no choice in forwarding 1867 packets along the path. 1869 SFF1 will receive packets on the path from the Classifier and will 1870 identify the path from the SPI (15). The initial SI will be 255 and 1871 so SFF1 will deliver the packets to the SFI for SFT 41. 1873 When the packets are returned to SFF1 by the SFI the SI will be 1874 decreased to 250 for the next hop. SFF1 has no flexibility in the 1875 choice of SFF to support the next hop SFI and will forward the packet 1876 to SFF2 which will send the packets to the SFI that supports SFT 43 1877 before forwarding the packets to their destinations. 1879 8.2. Example SFP With Choice of SFIs 1881 SFP2: RD = 198.51.100.1/102, SPI = 16, 1882 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1883 [SI = 250, SFT = 43, {RD = 192.0.2.2/2, 1884 RD = 192.0.2.4/5 } ] 1886 In this example the path also consists of an SF of type 41 located at 1887 SFF1 and this is followed by an SF of type 43, but in this case the 1888 SI = 250 contains a choice between the SFI located at SFF2 and the 1889 SFI located at SFF4. 1891 SFF1 will receive packets on the path from the Classifier and will 1892 identify the path from the SPI (16). The initial SI will be 255 and 1893 so SFF1 will deliver the packets to the SFI for SFT 41. 1895 When the packets are returned to SFF1 by the SFI the SI will be 1896 decreased to 250 for the next hop. SFF1 now has a choice of next hop 1897 SFF to execute the next hop in the path. It can either forward 1898 packets to SFF2 or SFF4 to execute a function of type 43. It uses 1899 its local load balancing algorithm to make this choice. The chosen 1900 SFF will send the packets to the SFI that supports SFT 43 before 1901 forwarding the packets to their destinations. 1903 8.3. Example SFP With Open Choice of SFIs 1905 SFP3: RD = 198.51.100.1/103, SPI = 17, 1906 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1907 [SI = 250, SFT = 44, RD = 0] 1909 In this example the path also consists of an SF of type 41 located at 1910 SFF1 and this is followed by an SI with an RD of zero and SF of type 1911 44. This means that a choice can be made between any SFF that 1912 supports an SFI of type 44. 1914 SFF1 will receive packets on the path from the Classifier and will 1915 identify the path from the SPI (17). The initial SI will be 255 and 1916 so SFF1 will deliver the packets to the SFI for SFT 41. 1918 When the packets are returned to SFF1 by the SFI the SI will be 1919 decreased to 250 for the next hop. SFF1 now has a free choice of 1920 next hop SFF to execute the next hop in the path selecting between 1921 all SFFs that support SFs of type 44. Looking at the SFIRs it has 1922 received, SFF1 knows that SF type 44 is supported by SFF3 and SFF4. 1923 SFF1 uses its local load balancing algorithm to make this choice. 1924 The chosen SFF will send the packets to the SFI that supports SFT 44 1925 before forwarding the packets to their destinations. 1927 8.4. Example SFP With Choice of SFTs 1929 SFP4: RD = 198.51.100.1/104, SPI = 18, 1930 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1931 [SI = 250, {SFT = 43, RD = 192.0.2.2/2, 1932 SFT = 44, RD = 192.0.2.3/8 } ] 1934 This example provides a choice of SF type in the second hop in the 1935 path. The SI of 250 indicates a choice between SF type 43 located at 1936 SF2 and SF type 44 located at SF3. 1938 SFF1 will receive packets on the path from the Classifier and will 1939 identify the path from the SPI (18). The initial SI will be 255 and 1940 so SFF1 will deliver the packets to the SFI for SFT 41. 1942 When the packets are returned to SFF1 by the SFI the SI will be 1943 decreased to 250 for the next hop. SFF1 now has a free choice of 1944 next hop SFF to execute the next hop in the path selecting between 1945 all SFFs that support an SF of type 43 and SFF3 that supports an SF 1946 of type 44. These may be completely different functions that are to 1947 be executed dependent on specific conditions, or may be similar 1948 functions identified with different type identifiers (such as 1949 firewalls from different vendors). SFF1 uses its local policy and 1950 load balancing algorithm to make this choice, and may use additional 1951 information passed back from the local SFI to help inform its 1952 selection. The chosen SFF will send the packets to the SFI that 1953 supports the chose SFT before forwarding the packets to their 1954 destinations. 1956 8.5. Example Correlated Bidirectional SFPs 1958 SFP5: RD = 198.51.100.1/105, SPI = 19, 1959 Assoc-Type = 1, Assoc-RD = 198.51.100.1/106, Assoc-SPI = 20, 1960 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1961 [SI = 250, SFT = 43, RD = 192.0.2.2/2] 1963 SFP6: RD = 198.51.100.1/106, SPI = 20, 1964 Assoc-Type = 1, Assoc-RD = 198.51.100.1/105, Assoc-SPI = 19, 1965 [SI = 254, SFT = 43, RD = 192.0.2.2/2], 1966 [SI = 249, SFT = 41, RD = 192.0.2.1/1] 1968 This example demonstrates correlation of two SFPs to form a 1969 bidirectional SFP as described in Section 7.1. 1971 Two SFPRs are advertised by the Controller. They have different SPIs 1972 (19 and 20) so they are known to be separate SFPs, but they both have 1973 Association TLVs with Association Type set to 1 indicating 1974 bidirectional SFPs. Each has an Associated SFPR-RD field containing 1975 the value of the other SFPR-RD to correlated the two SFPs as a 1976 bidirectional pair. 1978 As can be seen from the SFPRs in this example, the paths are 1979 symmetric: the hops in SFP5 appear in the reverse order in SFP6. 1981 8.6. Example Correlated Asymmetrical Bidirectional SFPs 1982 SFP7: RD = 198.51.100.1/107, SPI = 21, 1983 Assoc-Type = 1, Assoc-RD = 198.51.100.1/108, Assoc-SPI = 22, 1984 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 1985 [SI = 250, SFT = 43, RD = 192.0.2.2/2] 1987 SFP8: RD = 198.51.100.1/108, SPI = 22, 1988 Assoc-Type = 1, Assoc-RD = 198.51.100.1/107, Assoc-SPI = 21, 1989 [SI = 254, SFT = 44, RD = 192.0.2.4/6], 1990 [SI = 249, SFT = 41, RD = 192.0.2.1/1] 1992 Asymmetric bidirectional SFPs can also be created. This example 1993 shows a pair of SFPs with distinct SPIs (21 and 22) that are 1994 correlated in the same way as in the example in Section 8.5. 1996 However, unlike in that example, the SFPs are different in each 1997 direction. Both paths include a hop of SF type 41, but SFP7 includes 1998 a hop of SF type 43 supported at SFF2 while SFP8 includes a hop of SF 1999 type 44 supported at SFF4. 2001 8.7. Example Looping in an SFP 2003 SFP9: RD = 198.51.100.1/109, SPI = 23, 2004 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 2005 [SI = 250, SFT = 44, RD = 192.0.2.4/5], 2006 [SI = 245, {SFT = 1, RD = {SPI=23, SI=255, Rsv=0}, 2007 SFT = 42, RD = 192.0.2.3/7 } ] 2009 Looping and jumping are described in Section 6. This example shows 2010 an SFP that contains an explicit loop-back instruction that is 2011 presented as a choice within an SFP hop. 2013 The first two hops in the path (SI = 255 and SI = 250) are normal. 2014 That is, the packets will be delivered to SFF1 and SFF4 in turn for 2015 execution of SFs of type 41 and 44 respectively. 2017 The third hop (SI = 245) presents SFF4 with a choice of next hop. It 2018 can either forward the packets to SFF3 for an SF of type 42 (the 2019 second choice), or it can loop back. 2021 The loop-back entry in the SFPR for SI = 245 is indicated by the 2022 special purpose SFT value 1 ("Change Sequence"). Within this hop, 2023 the RD is interpreted as encoding the SPI and SI of the next hop (see 2024 Section 6.1. In this case the SPI is 23 which indicates that this is 2025 loop or branch: i.e., the next hop is on the same SFP. The SI is set 2026 to 255: this is a higher number than the current SI (245) indicating 2027 a loop. 2029 SFF4 must make a choice between these two next hops. Either the 2030 packets will be forwarded to SFF3 with the NSH SI decreased to 245 or 2031 looped back to SFF1 with the NSH SI reset to 255. This choice will 2032 be made according to local policy, information passed back by the 2033 local SFI, and details in the packets' metadata that are used to 2034 prevent infinite looping. 2036 8.8. Example Branching in an SFP 2038 SFP10: RD = 198.51.100.1/110, SPI = 24, 2039 [SI = 254, SFT = 42, RD = 192.0.2.3/7], 2040 [SI = 249, SFT = 43, RD = 192.0.2.2/2] 2042 SFP11: RD = 198.51.100.1/111, SPI = 25, 2043 [SI = 255, SFT = 41, RD = 192.0.2.1/1], 2044 [SI = 250, SFT = 1, RD = {SPI=24, SI=254, Rsv=0}] 2046 Branching follows a similar procedure to that for looping (and 2047 jumping) as shown in Section 8.7 however there are two SFPs involved. 2049 SFP10 shows a normal path with packets forwarded to SFF3 and SFF2 for 2050 execution of service functions of type 42 and 43 respectively. 2052 SFP11 starts as normal (SFF1 for an SF of type 41), but then SFF1 2053 processes the next hop in the path and finds a "Change Sequence" 2054 Special Purpose SFT. The SFIR-RD field includes an SPI of 24 which 2055 indicates SFP10, not the current SFP. The SI in the SFIR-RD is 254, 2056 so SFF1 knows that it must set the SPI/SI in the NSH to 24/254 and 2057 send the packets to the appropriate SFF as advertised in the SFPR for 2058 SFP10 (that is, SFF3). 2060 8.9. Examples of SFPs with Stateful Service Functions 2062 This section provides some examples to demonstrate establishing SFPs 2063 when there is a choice of service functions at a particular hop, and 2064 where consistency of choice is required in both directions. The 2065 scenarios that give rise to this requirement are discussed in 2066 Section 7.2. 2068 8.9.1. Forward and Reverse Choice Made at the SFF 2070 Consider the topology shown in Figure 12. There are three SFFs 2071 arranged neatly in a line, and the middle one (SFF2) supports three 2072 SFIs all of SFT 42. These three instances can be used by SFF2 to 2073 load balance so that no one instance is swamped. 2075 ------ ------ ------ ------ ------ 2076 | SFI | | SFIa | | SFIb | | SFIc | | SFI | 2077 |SFT=41| |SFT=42| |SFT=42| |SFT=42| |SFT=43| 2078 ------ ------\ ------ /------ ------ 2079 \ \ | / / 2080 --------- --------- --------- 2081 ---------- | SFF1 | | SFF2 | | SFF3 | 2082 --> | |..|192.0.2.1|...|192.0.2.2|...|192.0.2.3|--> 2083 --> |Classifier| --------- --------- --------- 2084 | | 2085 ---------- 2087 Figure 12: Example Where Choice is Made at the SFF 2089 This leads to the following SFIRs being advertised. 2091 RD = 192.0.2.1/11, SFT = 41 2092 RD = 192.0.2.2/11, SFT = 42 (for SFIa) 2093 RD = 192.0.2.2/12, SFT = 42 (for SFIb) 2094 RD = 192.0.2.2/13, SFT = 42 (for SFIc) 2095 RD = 192.0.2.3/11, SFT = 43 2097 The controller can create a single forward SFP (SFP12) giving SFF2 2098 the choice of which SFI to use to provide function of SFT 42 as 2099 follows. The load-balancing choice between the three available SFIs 2100 is assumed to be within the capabilities of the SFF and if the SFs 2101 are stateful it is assumed that the SFF knows this and arranges load 2102 balancing in a stable, flow-dependent way. 2104 SFP12: RD = 198.51.100.1/112, SPI = 26, 2105 Assoc-Type = 1, Assoc-RD = 198.51.100.1/113, Assoc-SPI = 27, 2106 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2107 [SI = 254, SFT = 42, {RD = 192.0.2.2/11, 2108 192.0.2.2/12, 2109 192.0.2.2/13 }], 2110 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2112 The reverse SFP (SFP13) in this case may also be created as shown 2113 below using association with the forward SFP and giving the load- 2114 balancing choice to SFF2. This is safe, even in the case that the 2115 SFs of type 42 are stateful because SFF2 is doing the load balancing 2116 in both directions and can apply the same algorithm to ensure that 2117 packets associated with the same flow use the same SFI regardless of 2118 the direction of travel. 2120 SFP13: RD = 198.51.100.1/113, SPI = 27, 2121 Assoc-Type = 1, Assoc-RD = 198.51.100.1/112, Assoc-SPI = 26, 2122 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2123 [SI = 254, SFT = 42, {RD = 192.0.2.2/11, 2124 192.0.2.2/12, 2125 192.0.2.2/13 }], 2126 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2128 How an SFF knows that an attached SFI is stateful is out of scope of 2129 this document. It is assumed that this will form part of the process 2130 by which SFIs are registered as local to SFFs. Section 7.2 provides 2131 additional observations about the coordination of the use of stateful 2132 SFIs in the case of bidirectional SFPs. 2134 In general, the problems of load balancing and the selection of the 2135 same SFIs in both directions of a bidirectional SFP can be addressed 2136 by using sufficiently precisely specified SFPs (specifying the exact 2137 SFIs to use) and suitable programming of the Classifiers at each end 2138 of the SFPs to make sure that the matching pair of SFPs are used. 2140 8.9.2. Parallel End-to-End SFPs with Shared SFF 2142 The mechanism described in Section 8.9.1 might not be desirable 2143 because of the functional assumptions it places on SFF2 to be able to 2144 load balance with suitable flow identification, stability, and 2145 equality in both directions. Instead, it may be desirable to place 2146 the responsibility for flow classification in the Classifier and let 2147 it determine load balancing with the implied choice of SFIs. 2149 Consider the network graph as shown in Figure 12 and with the same 2150 set of SFIRs as listed in Section 8.9.1. In this case the controller 2151 could specify three forward SFPs with their corresponding associated 2152 reverse SFPs. Each bidirectional pair of SFPs uses a different SFI 2153 for the SF of type 42. The controller can instruct the Classifier 2154 how to place traffic on the three bidirectional SFPs, or can treat 2155 them as a group leaving the Classifier responsible for balancing the 2156 load. 2158 SFP14: RD = 198.51.100.1/114, SPI = 28, 2159 Assoc-Type = 1, Assoc-RD = 198.51.100.1/117, Assoc-SPI = 31, 2160 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2161 [SI = 254, SFT = 42, RD = 192.0.2.2/11], 2162 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2164 SFP15: RD = 198.51.100.1/115, SPI = 29, 2165 Assoc-Type = 1, Assoc-RD = 198.51.100.1/118, Assoc-SPI = 32, 2166 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2167 [SI = 254, SFT = 42, RD = 192.0.2.2/12], 2168 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2170 SFP16: RD = 198.51.100.1/116, SPI = 30, 2171 Assoc-Type = 1, Assoc-RD = 198.51.100.1/119, Assoc-SPI = 33, 2172 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2173 [SI = 254, SFT = 42, RD = 192.0.2.2/13], 2174 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2176 SFP17: RD = 198.51.100.1/117, SPI = 31, 2177 Assoc-Type = 1, Assoc-RD = 198.51.100.1/114, Assoc-SPI = 28, 2178 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2179 [SI = 254, SFT = 42, RD = 192.0.2.2/11], 2180 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2182 SFP18: RD = 198.51.100.1/118, SPI = 32, 2183 Assoc-Type = 1, Assoc-RD = 198.51.100.1/115, Assoc-SPI = 29, 2184 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2185 [SI = 254, SFT = 42, RD = 192.0.2.2/12], 2186 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2188 SFP19: RD = 198.51.100.1/119, SPI = 33, 2189 Assoc-Type = 1, Assoc-RD = 198.51.100.1/116, Assoc-SPI = 30, 2190 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2191 [SI = 254, SFT = 42, RD = 192.0.2.2/13], 2192 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2194 8.9.3. Parallel End-to-End SFPs with Separate SFFs 2196 While the examples in Section 8.9.1 and Section 8.9.2 place the 2197 choice of SFI as subtended from the same SFF, it is also possible 2198 that the SFIs are each subtended from a different SFF as shown in 2199 Figure 13. In this case it is harder to coordinate the choices for 2200 forward and reverse paths without some form of coordination between 2201 SFF1 and SFF3. Therefore it would be normal to consider end-to-end 2202 parallel SFPs as described in Section 8.9.2. 2204 ------ 2205 | SFIa | 2206 |SFT=42| 2207 ------ 2208 ------ | 2209 | SFI | --------- 2210 |SFT=41| | SFF5 | 2211 ------ ..|192.0.2.5|.. 2212 | ..: --------- :.. 2213 ---------.: :.--------- 2214 ---------- | SFF1 | --------- | SFF3 | 2215 --> | |..|192.0.2.1|....| SFF6 |....|192.0.2.3| --> 2216 --> |Classifier| ---------: |192.0.2.6| :--------- 2217 | | : --------- : | 2218 ---------- : | : ------ 2219 : ------ : | SFI | 2220 :.. | SFIb | ..: |SFT=43| 2221 :.. |SFT=42| ..: ------ 2222 : ------ : 2223 :.---------.: 2224 | SFF7 | 2225 |192.0.2.7| 2226 --------- 2227 | 2228 ------ 2229 | SFIc | 2230 |SFT=42| 2231 ------ 2233 Figure 13: Second Example With Parallel End-to-End SFPs 2235 In this case, five SFIRs are advertised as follows: 2237 RD = 192.0.2.1/11, SFT = 41 2238 RD = 192.0.2.5/11, SFT = 42 (for SFIa) 2239 RD = 192.0.2.6/11, SFT = 42 (for SFIb) 2240 RD = 192.0.2.7/11, SFT = 42 (for SFIc) 2241 RD = 192.0.2.3/11, SFT = 43 2243 In this case the controller could specify three forward SFPs with 2244 their corresponding associated reverse SFPs. Each bidirectional pair 2245 of SFPs uses a different SFF and SFI for middle hop (for an SF of 2246 type 42). The controller can instruct the Classifier how to place 2247 traffic on the three bidirectional SFPs, or can treat them as a group 2248 leaving the Classifier responsible for balancing the load. 2250 SFP20: RD = 198.51.100.1/120, SPI = 34, 2251 Assoc-Type = 1, Assoc-RD = 198.51.100.1/123, Assoc-SPI = 37, 2252 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2253 [SI = 254, SFT = 42, RD = 192.0.2.5/11], 2254 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2256 SFP21: RD = 198.51.100.1/121, SPI = 35, 2257 Assoc-Type = 1, Assoc-RD = 198.51.100.1/124, Assoc-SPI = 38, 2258 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2259 [SI = 254, SFT = 42, RD = 192.0.2.6/11], 2260 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2262 SFP22: RD = 198.51.100.1/122, SPI = 36, 2263 Assoc-Type = 1, Assoc-RD = 198.51.100.1/125, Assoc-SPI = 39, 2264 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2265 [SI = 254, SFT = 42, RD = 192.0.2.7/11], 2266 [SI = 253, SFT = 43, RD = 192.0.2.3/11] 2268 SFP23: RD = 198.51.100.1/123, SPI = 37, 2269 Assoc-Type = 1, Assoc-RD = 198.51.100.1/120, Assoc-SPI = 34, 2270 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2271 [SI = 254, SFT = 42, RD = 192.0.2.5/11], 2272 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2274 SFP24: RD = 198.51.100.1/124, SPI = 38, 2275 Assoc-Type = 1, Assoc-RD = 198.51.100.1/121, Assoc-SPI = 35, 2276 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2277 [SI = 254, SFT = 42, RD = 192.0.2.6/11], 2278 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2280 SFP25: RD = 198.51.100.1/125, SPI = 39, 2281 Assoc-Type = 1, Assoc-RD = 198.51.100.1/122, Assoc-SPI = 36, 2282 [SI = 255, SFT = 43, RD = 192.0.2.3/11], 2283 [SI = 254, SFT = 42, RD = 192.0.2.7/11], 2284 [SI = 253, SFT = 41, RD = 192.0.2.1/11] 2286 8.9.4. Parallel SFPs Downstream of the Choice 2288 The mechanism of parallel SFPs demonstrated in Section 8.9.3 is 2289 perfectly functional and may be practical in many environments. 2290 However, there may be scaling concerns because of the large amount of 2291 state (knowledge of SFPs, i.e., SFPR advertisements retained) if 2292 there is a very large amount of choice of SFIs (for example, tens of 2293 instances of the same stateful SF), or if there are multiple choices 2294 of stateful SF along a path. This situation may be mitigated using 2295 SFP fragments that are combined to form the end to end SFPs. 2297 The example presented here is necessarily simplistic, but should 2298 convey the basic principle. The example presented in Figure 14 is 2299 similar to that in Section 8.9.3 but with an additional first hop. 2301 ------ 2302 | SFIa | 2303 |SFT=43| 2304 ------ 2305 ------ ------ | 2306 | SFI | | SFI | --------- 2307 |SFT=41| |SFT=42| | SFF5 | 2308 ------ ------ ..|192.0.2.5|.. 2309 | | ..: --------- :.. 2310 --------- ---------.: :.--------- 2311 ------ | SFF1 | | SFF2 | --------- | SFF3 | 2312 -->|Class-|.|192.0.2.1|.|192.0.2.2|....| SFF6 |....|192.0.2.3|--> 2313 -->| ifier| --------- ---------: |192.0.2.6| :--------- 2314 ------ : --------- : | 2315 : | : ------ 2316 : ------ : | SFI | 2317 :.. | SFIb | ..: |SFT=44| 2318 :.. |SFT=43| ..: ------ 2319 : ------ : 2320 :.---------.: 2321 | SFF7 | 2322 |192.0.2.7| 2323 --------- 2324 | 2325 ------ 2326 | SFIc | 2327 |SFT=43| 2328 ------ 2330 Figure 14: Example With Parallel SFPs Downstream of Choice 2332 The six SFIs are advertised as follows: 2334 RD = 192.0.2.1/11, SFT = 41 2335 RD = 192.0.2.2/11, SFT = 42 2336 RD = 192.0.2.5/11, SFT = 43 (for SFIa) 2337 RD = 192.0.2.6/11, SFT = 43 (for SFIb) 2338 RD = 192.0.2.7/11, SFT = 43 (for SFIc) 2339 RD = 192.0.2.3/11, SFT = 44 2341 SFF2 is the point at which a load balancing choice must be made. So 2342 "tail-end" SFPs are constructed as follows. Each takes in a 2343 different SFF that provides access to an SF of type 43. 2345 SFP26: RD = 198.51.100.1/126, SPI = 40, 2346 Assoc-Type = 1, Assoc-RD = 198.51.100.1/130, Assoc-SPI = 44, 2347 [SI = 255, SFT = 43, RD = 192.0.2.5/11], 2348 [SI = 254, SFT = 44, RD = 192.0.2.3/11] 2350 SFP27: RD = 198.51.100.1/127, SPI = 41, 2351 Assoc-Type = 1, Assoc-RD = 198.51.100.1/131, Assoc-SPI = 45, 2352 [SI = 255, SFT = 43, RD = 192.0.2.6/11], 2353 [SI = 254, SFT = 44, RD = 192.0.2.3/11] 2355 SFP28: RD = 198.51.100.1/128, SPI = 42, 2356 Assoc-Type = 1, Assoc-RD = 198.51.100.1/132, Assoc-SPI = 46, 2357 [SI = 255, SFT = 43, RD = 192.0.2.7/11], 2358 [SI = 254, SFT = 44, RD = 192.0.2.3/11] 2360 Now an end-to-end SFP with load balancing choice can be constructed 2361 as follows. The choice made by SFF2 is expressed in terms of 2362 entering one of the three "tail end" SFPs. 2364 SFP29: RD = 198.51.100.1/129, SPI = 43, 2365 [SI = 255, SFT = 41, RD = 192.0.2.1/11], 2366 [SI = 254, SFT = 42, RD = 192.0.2.2/11], 2367 [SI = 253, {SFT = 1, RD = {SPI=40, SI=255, Rsv=0}, 2368 RD = {SPI=41, SI=255, Rsv=0}, 2369 RD = {SPI=42, SI=255, Rsv=0} } ] 2371 Now, despite the load balancing choice being made other than at the 2372 initial Classifier, it is possible for the reverse SFPs to be well- 2373 constructed without any ambiguity. The three reverse paths appear as 2374 follows. 2376 SFP30: RD = 198.51.100.1/130, SPI = 44, 2377 Assoc-Type = 1, Assoc-RD = 198.51.100.1/126, Assoc-SPI = 40, 2378 [SI = 255, SFT = 44, RD = 192.0.2.4/11], 2379 [SI = 254, SFT = 43, RD = 192.0.2.5/11], 2380 [SI = 253, SFT = 42, RD = 192.0.2.2/11], 2381 [SI = 252, SFT = 41, RD = 192.0.2.1/11] 2383 SFP31: RD = 198.51.100.1/131, SPI = 45, 2384 Assoc-Type = 1, Assoc-RD = 198.51.100.1/127, Assoc-SPI = 41, 2385 [SI = 255, SFT = 44, RD = 192.0.2.4/11], 2386 [SI = 254, SFT = 43, RD = 192.0.2.6/11], 2387 [SI = 253, SFT = 42, RD = 192.0.2.2/11], 2388 [SI = 252, SFT = 41, RD = 192.0.2.1/11] 2390 SFP32: RD = 198.51.100.1/132, SPI = 46, 2391 Assoc-Type = 1, Assoc-RD = 198.51.100.1/128, Assoc-SPI = 42, 2392 [SI = 255, SFT = 44, RD = 192.0.2.4/11], 2393 [SI = 254, SFT = 43, RD = 192.0.2.7/11], 2394 [SI = 253, SFT = 42, RD = 192.0.2.2/11], 2395 [SI = 252, SFT = 41, RD = 192.0.2.1/11] 2397 8.10. Examples Using IPv6 Addressing 2399 This section provides several examples using IPv6 addressing. As 2400 will be seen from the examples, there is nothing special or clever 2401 about using IPv6 addressing rather than IPv4 addressing. 2403 The reference network for these IPv6 examples is based on that 2404 described at the top of Section 8 and shown in Figure 11. 2406 Assume we have a service function overlay network with four SFFs 2407 (SFF1, SFF3, SFF3, and SFF4). The SFFs have addresses in the 2408 underlay network as follows: 2410 SFF1 2001:db8::192:0:2:1 2411 SFF2 2001:db8::192:0:2:2 2412 SFF3 2001:db8::192:0:2:3 2413 SFF4 2001:db8::192:0:2:4 2415 Each SFF provides access to some SFIs from the four Service Function 2416 Types SFT=41, SFT=42, SFT=43, and SFT=44 just as before: 2418 SFF1 SFT=41 and SFT=42 2419 SFF2 SFT=41 and SFT=43 2420 SFF3 SFT=42 and SFT=44 2421 SFF4 SFT=43 and SFT=44 2423 The service function network also contains a Controller with address 2424 2001:db8::198:51:100:1. 2426 This example service function overlay network is shown in Figure 15. 2428 ------------------------ 2429 | Controller | 2430 | 2001:db8::198:51:100:1 | 2431 ------------------------ 2432 ------ ------ ------ ------ 2433 | SFI | | SFI | | SFI | | SFI | 2434 |SFT=41| |SFT=42| |SFT=41| |SFT=43| 2435 ------ ------ ------ ------ 2436 \ / \ / 2437 ------------------- ------------------- 2438 | SFF1 | | SFF2 | 2439 |2001:db8::192:0:2:1| |2001:db8::192:0:2:2| 2440 ------------------- ------------------- 2441 ---------- 2442 Packet --> | | --> 2443 Flows --> |Classifier| -->Dest 2444 | | --> 2445 ---------- 2446 ------------------- ------------------- 2447 | SFF3 | | SFF4 | 2448 |2001:db8::192:0:2:3| |2001:db8::192:0:2:4| 2449 ------------------- ------------------- 2450 / \ / \ 2451 ------ ------ ------ ------ 2452 | SFI | | SFI | | SFI | | SFI | 2453 |SFT=42| |SFT=44| |SFT=43| |SFT=44| 2454 ------ ------ ------ ------ 2456 Figure 15: Example Service Function Overlay Network 2458 The SFFs advertise routes to the SFIs they support. So we see the 2459 following SFIRs: 2461 RD = 2001:db8::192:0:2:1/1, SFT = 41 2462 RD = 2001:db8::192:0:2:1/2, SFT = 42 2463 RD = 2001:db8::192:0:2:2/1, SFT = 41 2464 RD = 2001:db8::192:0:2:2/2, SFT = 43 2465 RD = 2001:db8::192:0:2:3/7, SFT = 42 2466 RD = 2001:db8::192:0:2:3/8, SFT = 44 2467 RD = 2001:db8::192:0:2:4/5, SFT = 43 2468 RD = 2001:db8::192:0:2:4/6, SFT = 44 2470 Note that the addressing used for communicating between SFFs is taken 2471 from the Tunnel Encapsulation attribute of the SFIR and not from the 2472 SFIR-RD. 2474 8.10.1. Example Explicit SFP With No Choices 2476 Consider the following SFPR similar to that in Section 8.1. 2478 SFP1: RD = 2001:db8::198:51:100:1/101, SPI = 15, 2479 [SI = 255, SFT = 41, RD = 2001:db8::192:0:2:1/1], 2480 [SI = 250, SFT = 43, RD = 2001:db8::192:0:2:2/2] 2482 The Service Function Path consists of an SF of type 41 located at 2483 SFF1 followed by an SF of type 43 located at SFF2. This path is 2484 fully explicit and each SFF is offered no choice in forwarding packet 2485 along the path. 2487 SFF1 will receive packets on the path from the Classifier and will 2488 identify the path from the SPI (15). The initial SI will be 255 and 2489 so SFF1 will deliver the packets to the SFI for SFT 41. 2491 When the packets are returned to SFF1 by the SFI the SI will be 2492 decreased to 250 for the next hop. SFF1 has no flexibility in the 2493 choice of SFF to support the next hop SFI and will forward the packet 2494 to SFF2 which will send the packets to the SFI that supports SFT 43 2495 before forwarding the packets to their destinations. 2497 8.10.2. Example SFP With Choice of SFIs 2499 SFP2: RD = 2001:db8::198:51:100:1/102, SPI = 16, 2500 [SI = 255, SFT = 41, RD = 2001:db8::192:0:2:1/1], 2501 [SI = 250, SFT = 43, {RD = 2001:db8::192:0:2:2/2, 2502 RD = 2001:db8::192:0:2:4/5 } ] 2504 In this example, like that in Section 8.2, the path also consists of 2505 an SF of type 41 located at SFF1 and this is followed by an SF of 2506 type 43, but in this case the SI = 250 contains a choice between the 2507 SFI located at SFF2 and the SFI located at SFF4. 2509 SFF1 will receive packets on the path from the Classifier and will 2510 identify the path from the SPI (16). The initial SI will be 255 and 2511 so SFF1 will deliver the packets to the SFI for SFT 41. 2513 When the packets are returned to SFF1 by the SFI the SI will be 2514 decreased to 250 for the next hop. SFF1 now has a choice of next hop 2515 SFF to execute the next hop in the path. It can either forward 2516 packets to SFF2 or SFF4 to execute a function of type 43. It uses 2517 its local load balancing algorithm to make this choice. The chosen 2518 SFF will send the packets to the SFI that supports SFT 43 before 2519 forwarding the packets to their destinations. 2521 8.10.3. Example SFP With Open Choice of SFIs 2523 SFP3: RD = 2001:db8::198:51:100:1/103, SPI = 17, 2524 [SI = 255, SFT = 41, RD = 2001:db8::192:0:2:1/1], 2525 [SI = 250, SFT = 44, RD = 0] 2527 In this example, like that in Section 8.3 the path also consists of 2528 an SF of type 41 located at SFF1 and this is followed by an SI with 2529 an RD of zero and SF of type 44. This means that a choice can be 2530 made between any SFF that supports an SFI of type 44. 2532 SFF1 will receive packets on the path from the Classifier and will 2533 identify the path from the SPI (17). The initial SI will be 255 and 2534 so SFF1 will deliver the packets to the SFI for SFT 41. 2536 When the packets are returned to SFF1 by the SFI the SI will be 2537 decreased to 250 for the next hop. SFF1 now has a free choice of 2538 next hop SFF to execute the next hop in the path selecting between 2539 all SFFs that support SFs of type 44. Looking at the SFIRs it has 2540 received, SFF1 knows that SF type 44 is supported by SFF3 and SFF4. 2541 SFF1 uses its local load balancing algorithm to make this choice. 2542 The chosen SFF will send the packets to the SFI that supports SFT 44 2543 before forwarding the packets to their destinations. 2545 8.10.4. Example SFP With Choice of SFTs 2546 SFP4: RD = 2001:db8::198:51:100:1/104, SPI = 18, 2547 [SI = 255, SFT = 41, RD = 2001:db8::192:0:2:1/1], 2548 [SI = 250, {SFT = 43, RD = 2001:db8::192:0:2:2/2, 2549 SFT = 44, RD = 2001:db8::192:0:2:3/8 } ] 2551 This example, similar to that in Section 8.4 provides a choice of SF 2552 type in the second hop in the path. The SI of 250 indicates a choice 2553 between SF type 43 located through SF2 and SF type 44 located at SF3. 2555 SFF1 will receive packets on the path from the Classifier and will 2556 identify the path from the SPI (18). The initial SI will be 255 and 2557 so SFF1 will deliver the packets to the SFI for SFT 41. 2559 When the packets are returned to SFF1 by the SFI the SI will be 2560 decreased to 250 for the next hop. SFF1 now has a free choice of 2561 next hop SFF to execute the next hop in the path selecting between 2562 all SFF2 that support an SF of type 43 and SFF3 that supports an SF 2563 of type 44. These may be completely different functions that are to 2564 be executed dependent on specific conditions, or may be similar 2565 functions identified with different type identifiers (such as 2566 firewalls from different vendors). SFF1 uses its local policy and 2567 load balancing algorithm to make this choice, and may use additional 2568 information passed back from the local SFI to help inform its 2569 selection. The chosen SFF will send the packets to the SFI that 2570 supports the chose SFT before forwarding the packets to their 2571 destinations. 2573 9. Security Considerations 2575 The mechanisms in this document use BGP for the control plane. 2576 Hence, techniques such as those discussed in [RFC5925]] can be used 2577 to help authenticate BGP sessions and thus the messages between BGP 2578 peers, making it harder to spoof updates (which could be used to 2579 install bogus SFPs or to advertise false SIs) or withdrawals. 2581 Further discussion of security considerations for BGP may be found in 2582 the BGP specification itself [RFC4271] and in the security analysis 2583 for BGP [RFC4272]. The original discussion of the use of the TCP MD5 2584 signature option to protect BGP sessions is found in [RFC5925], while 2585 [RFC6952] includes an analysis of BGP keying and authentication 2586 issues. 2588 Additionally, this document depends on other documents that specify 2589 BGP Multiprotocol Extensions and the documents that define the 2590 attributes that are carried by BGP UPDATEs of the SFC AFI/SAFI. 2591 Relevant additional security measures are considered in [RFC4760] and 2592 [I-D.ietf-idr-tunnel-encaps]. 2594 This document does not fundamentally change the security behavior of 2595 BGP deployments which depend considerably on the network operator's 2596 perception of risk in their network. It may be observed that the 2597 application of the mechanisms described in this document are scoped 2598 to a single domain as implied by [RFC8300] noted in Section 2.1. 2599 Applicability of BGP within a single domain may enable a network 2600 operator to make easier and more consistent decisions about what 2601 security measures to apply, and the domain boundary, which BGP 2602 enforces by definition, provides a safeguard that prevents leakage of 2603 SFC programming in either direction at the boundary. 2605 Service Function Chaining provides a significant attack opportunity: 2606 packets can be diverted from their normal paths through the network, 2607 packets can be made to execute unexpected functions, and the 2608 functions that are instantiated in software can be subverted. 2609 However, this specification does not change the existence of Service 2610 Function Chaining and security issues specific to Service Function 2611 Chaining are covered in [RFC7665] and [RFC8300]. 2613 This document defines a control plane for Service Function Chaining. 2614 Clearly, this provides an attack vector for a Service Function 2615 Chaining system as an attack on this control plane could be used to 2616 make the system misbehave. Thus, the security of the BGP system is 2617 critically important to the security of the whole Service Function 2618 Chaining system. The control plane mechanisms are very similar to 2619 those used for BGP/MPLS IP VPNs as described in [RFC4364], and so the 2620 security considerations in that document (Section 13) provide good 2621 guidance for securing SFC systems reliant on this specification. Of 2622 particular relevance is the need to securely distinguish between 2623 messages intended for the control of different SFC overlays which is 2624 similar to the need to distinguish between different VPNs. 2625 Section 19 of [RFC7432] also provides useful guidance on the use of 2626 BGP in a similar environment. 2628 Note that a component of an SFC system that uses the procedures 2629 described in this document also requires communications between a 2630 controller and the SFC network elements. This communication covers 2631 instructing the Classifiers using BGP mechanisms (see Section 7.4), 2632 thus the use of BGP security is strongly recommended.. But it also 2633 covers other mechanisms for programming the Classifier and 2634 instructing the SFFs and SFs (for example, to bind SFs to an SFF, and 2635 to cause the establishment of tunnels between SFFs). This document 2636 does not cover these latter mechanisms and so their security is out 2637 of scope, but it should be noted that these communications provide an 2638 attack vector on the SFC system and so attention must be paid to 2639 ensuring that they are secure. 2641 There is an intrinsic assumption in SFC systems that nodes that 2642 announce support for specific SFs actually offer those functions, and 2643 that SFs are not, themselves, attacked or subverted. This is 2644 particularly important when the SFs are implemented as software that 2645 can be updated. Protection against this sort of concern forms part 2646 of the security of any SFC system and so is outside the scope of the 2647 control plane mechanisms described in this document. 2649 Similarly, there is a vulnerablity if a rogue or subverted controller 2650 announces SFPs especially if that controller "takes over" an existing 2651 SFP and changes its contents. This is corresponds to a rogue BGP 2652 speaker entering a routing system, or even to a Route Reflector 2653 becoming subverted. Protection mechanisms, as above, include 2654 securing BGP sessions and protecting software loads on the 2655 controllers. 2657 Lastly, note that Section 3.2.2 makes two operational suggestions 2658 that have implications for the stability and security of the 2659 mechanisms described in this document: 2661 o That modifications to active SFPs not be made. 2663 o That SPIs not be immediately re-used. 2665 10. IANA Considerations 2667 10.1. New BGP AF/SAFI 2669 IANA maintains a registry of "Address Family Numbers". IANA is 2670 requested to assign a new Address Family Number from the "Standards 2671 Action" range called "BGP SFC" (TBD1 in this document) with this 2672 document as a reference. 2674 IANA maintains a registry of "Subsequent Address Family Identifiers 2675 (SAFI) Parameters". IANA is requested to assign a new SAFI value 2676 from the "Standards Action" range called "BGP SFC" (TBD2 in this 2677 document) with this document as a reference. 2679 10.2. New BGP Path Attribute 2681 IANA maintains a registry of "Border Gateway Protocol (BGP) 2682 Parameters" with a subregistry of "BGP Path Attributes". IANA is 2683 requested to assign a new Path attribute called "SFP attribute" (TBD3 2684 in this document) with this document as a reference. 2686 10.3. New SFP Attribute TLVs Type Registry 2688 IANA maintains a registry of "Border Gateway Protocol (BGP) 2689 Parameters". IANA is request to create a new subregistry called the 2690 "SFP Attribute TLVs" registry. 2692 Valid values are in the range 0 to 65535. 2694 o Values 0 and 65535 are to be marked "Reserved, not to be 2695 allocated". 2697 o Values 1 through 65534 are to be assigned according to the "First 2698 Come First Served" policy [RFC8126]. 2700 This document should be given as a reference for this registry. 2702 The new registry should track: 2704 o Type 2706 o Name 2708 o Reference Document or Contact 2710 o Registration Date 2712 The registry should initially be populated as follows: 2714 Type | Name | Reference | Date 2715 ------+-------------------------+---------------+--------------- 2716 1 | Association TLV | [This.I-D] | Date-to-be-set 2717 2 | Hop TLV | [This.I-D] | Date-to-be-set 2718 3 | SFT TLV | [This.I-D] | Date-to-be-set 2719 4 | MPLS Swapping/Stacking | [This.I-D] | Date-to-be-set 2720 5 | SFP Traversal With MPLS | [This.I-D] | Date-to-be-set 2722 10.4. New SFP Association Type Registry 2724 IANA maintains a registry of "Border Gateway Protocol (BGP) 2725 Parameters". IANA is request to create a new subregistry called the 2726 "SFP Association Type" registry. 2728 Valid values are in the range 0 to 65535. 2730 o Values 0 and 65535 are to be marked "Reserved, not to be 2731 allocated". 2733 o Values 1 through 65534 are to be assigned according to the "First 2734 Come First Served" policy [RFC8126]. 2736 This document should be given as a reference for this registry. 2738 The new registry should track: 2740 o Association Type 2742 o Name 2744 o Reference Document or Contact 2746 o Registration Date 2748 The registry should initially be populated as follows: 2750 Association Type | Name | Reference | Date 2751 -----------------+--------------------+------------+--------------- 2752 1 | Bidirectional SFP | [This.I-D] | Date-to-be-set 2754 10.5. New Service Function Type Registry 2756 IANA is request to create a new top-level registry called "Service 2757 Function Chaining Service Function Types". 2759 Valid values are in the range 0 to 65535. 2761 o Values 0 and 65535 are to be marked "Reserved, not to be 2762 allocated". 2764 o Values 1 through 31 are to be assigned by "Standards Action" 2765 [RFC8126] and are referred to as the Special Purpose SFT values. 2767 o Values 32 through 64495 are to be assigned according to the "First 2768 Come First Served" policy [RFC8126]. 2770 o Values 64496 through 65534 are for Private Use and are not to be 2771 recorded by IANA. 2773 This document should be given as a reference for this registry. 2775 The new registry should track: 2777 o Value 2778 o Name 2780 o Reference Document or Contact 2782 o Registration Date 2784 The registry should initially be populated as follows: 2786 Value | Name | Reference | Date 2787 ------+-------------------------------+------------+--------------- 2788 0 | Reserved, not to be allocated | [This.I-D] | Date-to-be-set 2789 1 | Change Sequence | [This.I-D] | Date-to-be-set 2790 2-31 | Unassigned | | 2791 32 | Classifier | [This.I-D] | Date-to-be-set 2792 33 | Firewall | [This.I-D] | Date-to-be-set 2793 34 | Load balancer | [This.I-D] | Date-to-be-set 2794 35 | Deep packet inspection engine | [This.I-D] | Date-to-be-set 2795 36 | Penalty box | [This.I-D] | Date-to-be-set 2796 37 | WAN accelerator | [This.I-D] | Date-to-be-set 2797 38 | Application accelerator | [This.I-D] | Date-to-be-set 2798 39 | TCP optimizer | [This.I-D] | Date-to-be-set 2799 40 | Network Address Translator | [This.I-D] | Date-to-be-set 2800 41 | NAT44 | [This.I-D] | Date-to-be-set 2801 42 | NAT64 | [This.I-D] | Date-to-be-set 2802 43 | NPTv6 | [This.I-D] | Date-to-be-set 2803 44 | Lawful intercept | [This.I-D] | Date-to-be-set 2804 45 | HOST_ID injection | [This.I-D] | Date-to-be-set 2805 46 | HTTP header enrichment | [This.I-D] | Date-to-be-set 2806 47 | Caching engine | [This.I-D] | Date-to-be-set 2807 48- | | | 2808 -65534|Unassigned | | 2809 65535 | Reserved, not to be allocated | [This.I-D] | Date-to-be-set 2811 10.6. New Generic Transitive Experimental Use Extended Community Sub- 2812 Types 2814 IANA maintains a registry of "Border Gateway Protocol (BGP) 2815 Parameters" with a subregistry of "Generic Transitive Experimental 2816 Use Extended Community Sub-Type". IANA is requested to assign a new 2817 sub-type as follows: 2819 "Flow Specification for SFC Classifiers" (TBD4 in this document) 2820 with this document as the reference. 2822 10.7. New BGP Transitive Extended Community Type 2824 IANA maintains a registry of "Border Gateway Protocol (BGP) 2825 Parameters" with a subregistry of "BGP Transitive Extended Community 2826 Types". IANA is requested to assign a new type as follows: 2828 o SFC (Sub-Types are defined in the "SFC Extended Community Sub- 2829 Types" registry) (TBD6 in this document) with this document as the 2830 reference. 2832 10.8. New SFC Extended Community Sub-Types Registry 2834 IANA maintains a registry of "Border Gateway Protocol (BGP) 2835 Parameters". IANA is requested to create a new sub-registry called 2836 the "SFC Extended Community Sub-Types Registry". 2838 IANA should include the following note replacing the string "TBD6" 2839 with the value assigned for Section 10.7: 2841 This registry contains values of the second octet (the "Sub-Type" 2842 field) of an extended community when the value of the first octet 2843 (the "Type" field) is set to TBD6. 2845 The allocation policy for this registry should be First Come First 2846 Served. 2848 IANA is requested to populate this registry with the following 2849 entries: 2851 Sub-Type | | | 2852 Value | Name | Reference | Date 2853 ---------+----------------------+-------------+--------------- 2854 TBD7 | SFIR Pool Identifier | [This.I-D] | Date-to-be-set 2855 TBD8 | MPLS Label Stack | [This.I-D] | Date-to-be-set 2856 | Mixed Swapping/ | | 2857 | Stacking Labels | | 2859 All other values should be marked "Unassigned". 2861 10.9. SPI/SI Representation 2863 IANA is requested to assign a codepoint from the "BGP Tunnel 2864 Encapsulation Attribute Sub-TLVs" registry for the "SPI/SI 2865 Representation Sub-TLV" (TBD5 in this document) with this document 2866 being the reference. 2868 10.10. SFC SPI/SI Representation Flags Registry 2870 IANA maintains the "BGP Tunnel Encapsulation Attribute Sub-TLVs" 2871 registry and is requested to create an associated registry called the 2872 "SFC SPI/SI Representation Flags" registry. 2874 Bits are to be assigned by Standards Action. The field is 16 bits 2875 long, and bits are counted from the the most significant bit as bit 2876 zero. 2878 IANA is requested to populate the registry as follows: 2880 Bit number | Name | Reference 2881 -----------+----------------------+----------- 2882 TBD9 | NSH data plane | [This.I-D] 2883 TBD10 | MPLS data plane | [This.I-D] 2885 11. Contributors 2887 Stuart Mackie 2888 Juniper Networks 2890 Email: wsmackie@juinper.net 2892 Keyur Patel 2893 Arrcus, Inc. 2895 Email: keyur@arrcus.com 2897 Avinash Lingala 2898 AT&T 2900 Email: ar977m@att.com 2902 12. Acknowledgements 2904 Thanks to Tony Przygienda, Jeff Haas, and Andy Malis for helpful 2905 comments, and to Joel Halpern for discussions that improved this 2906 document. Yuanlong Jiang provided a useful review and caught some 2907 important issues. Stephane Litkowski did an exceptionally good and 2908 detailed document shepherd review. 2910 Andy Malis contributed text that formed the basis of Section 7.7. 2912 Brian Carpenter and Martin Vigoureux provided useful reviews during 2913 IETF last call. Thanks also to Sheng Jiang, Ravi Singh, Benjamin 2914 Kaduk, Roman Danyliw, Adam Roach, Alvaro Retana, and Barry Leiba for 2915 review comments. Ketan Talaulikar provided helpful discussion of the 2916 SFT code point registry. 2918 13. References 2920 13.1. Normative References 2922 [I-D.ietf-idr-rfc5575bis] 2923 Loibl, C., Hares, S., Raszuk, R., McPherson, D., and M. 2924 Bacher, "Dissemination of Flow Specification Rules", 2925 draft-ietf-idr-rfc5575bis-25 (work in progress), May 2020. 2927 [I-D.ietf-idr-tunnel-encaps] 2928 Patel, K., Velde, G., and S. Ramachandra, "The BGP Tunnel 2929 Encapsulation Attribute", draft-ietf-idr-tunnel-encaps-15 2930 (work in progress), December 2019. 2932 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2933 Requirement Levels", BCP 14, RFC 2119, 2934 DOI 10.17487/RFC2119, March 1997, 2935 . 2937 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 2938 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 2939 DOI 10.17487/RFC4271, January 2006, 2940 . 2942 [RFC4360] Sangli, S., Tappan, D., and Y. Rekhter, "BGP Extended 2943 Communities Attribute", RFC 4360, DOI 10.17487/RFC4360, 2944 February 2006, . 2946 [RFC4364] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private 2947 Networks (VPNs)", RFC 4364, DOI 10.17487/RFC4364, February 2948 2006, . 2950 [RFC4760] Bates, T., Chandra, R., Katz, D., and Y. Rekhter, 2951 "Multiprotocol Extensions for BGP-4", RFC 4760, 2952 DOI 10.17487/RFC4760, January 2007, 2953 . 2955 [RFC5575] Marques, P., Sheth, N., Raszuk, R., Greene, B., Mauch, J., 2956 and D. McPherson, "Dissemination of Flow Specification 2957 Rules", RFC 5575, DOI 10.17487/RFC5575, August 2009, 2958 . 2960 [RFC7432] Sajassi, A., Ed., Aggarwal, R., Bitar, N., Isaac, A., 2961 Uttaro, J., Drake, J., and W. Henderickx, "BGP MPLS-Based 2962 Ethernet VPN", RFC 7432, DOI 10.17487/RFC7432, February 2963 2015, . 2965 [RFC7606] Chen, E., Ed., Scudder, J., Ed., Mohapatra, P., and K. 2966 Patel, "Revised Error Handling for BGP UPDATE Messages", 2967 RFC 7606, DOI 10.17487/RFC7606, August 2015, 2968 . 2970 [RFC7665] Halpern, J., Ed. and C. Pignataro, Ed., "Service Function 2971 Chaining (SFC) Architecture", RFC 7665, 2972 DOI 10.17487/RFC7665, October 2015, 2973 . 2975 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2976 Writing an IANA Considerations Section in RFCs", BCP 26, 2977 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2978 . 2980 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2981 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 2982 May 2017, . 2984 [RFC8300] Quinn, P., Ed., Elzur, U., Ed., and C. Pignataro, Ed., 2985 "Network Service Header (NSH)", RFC 8300, 2986 DOI 10.17487/RFC8300, January 2018, 2987 . 2989 [RFC8595] Farrel, A., Bryant, S., and J. Drake, "An MPLS-Based 2990 Forwarding Plane for Service Function Chaining", RFC 8595, 2991 DOI 10.17487/RFC8595, June 2019, 2992 . 2994 [RFC8596] Malis, A., Bryant, S., Halpern, J., and W. Henderickx, 2995 "MPLS Transport Encapsulation for the Service Function 2996 Chaining (SFC) Network Service Header (NSH)", RFC 8596, 2997 DOI 10.17487/RFC8596, June 2019, 2998 . 3000 13.2. Informative References 3002 [I-D.dawra-idr-bgp-ls-sr-service-segments] 3003 Dawra, G., Filsfils, C., Talaulikar, K., Clad, F., 3004 daniel.bernier@bell.ca, d., Uttaro, J., Decraene, B., 3005 Elmalky, H., Xu, X., Guichard, J., and C. Li, "BGP-LS 3006 Advertisement of Segment Routing Service Segments", draft- 3007 dawra-idr-bgp-ls-sr-service-segments-03 (work in 3008 progress), January 2020. 3010 [RFC4272] Murphy, S., "BGP Security Vulnerabilities Analysis", 3011 RFC 4272, DOI 10.17487/RFC4272, January 2006, 3012 . 3014 [RFC5925] Touch, J., Mankin, A., and R. Bonica, "The TCP 3015 Authentication Option", RFC 5925, DOI 10.17487/RFC5925, 3016 June 2010, . 3018 [RFC6952] Jethanandani, M., Patel, K., and L. Zheng, "Analysis of 3019 BGP, LDP, PCEP, and MSDP Issues According to the Keying 3020 and Authentication for Routing Protocols (KARP) Design 3021 Guide", RFC 6952, DOI 10.17487/RFC6952, May 2013, 3022 . 3024 [RFC7498] Quinn, P., Ed. and T. Nadeau, Ed., "Problem Statement for 3025 Service Function Chaining", RFC 7498, 3026 DOI 10.17487/RFC7498, April 2015, 3027 . 3029 Authors' Addresses 3031 Adrian Farrel 3032 Old Dog Consulting 3034 Email: adrian@olddog.co.uk 3036 John Drake 3037 Juniper Networks 3039 Email: jdrake@juniper.net 3041 Eric Rosen 3042 Juniper Networks 3044 Email: erosen52@gmail.com 3045 Jim Uttaro 3046 AT&T 3048 Email: ju1738@att.com 3050 Luay Jalil 3051 Verizon 3053 Email: luay.jalil@verizon.com