idnits 2.17.1 draft-ietf-sfc-architecture-04.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 (January 12, 2015) is 3385 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-13) exists of draft-ietf-sfc-problem-statement-10 Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Halpern, Ed. 3 Internet-Draft Ericsson 4 Intended status: Standards Track C. Pignataro, Ed. 5 Expires: July 16, 2015 Cisco 6 January 12, 2015 8 Service Function Chaining (SFC) Architecture 9 draft-ietf-sfc-architecture-04 11 Abstract 13 This document describes an architecture for the specification, 14 creation, and ongoing maintenance of Service Function Chains (SFC) in 15 a network. It includes architectural concepts, principles, and 16 components used in the construction of composite services through 17 deployment of SFCs, with a focus on those to be standardized in the 18 IETF. This document does not propose solutions, protocols, or 19 extensions to existing protocols. 21 Status of This Memo 23 This Internet-Draft is submitted in full conformance with the 24 provisions of BCP 78 and BCP 79. 26 Internet-Drafts are working documents of the Internet Engineering 27 Task Force (IETF). Note that other groups may also distribute 28 working documents as Internet-Drafts. The list of current Internet- 29 Drafts is at http://datatracker.ietf.org/drafts/current/. 31 Internet-Drafts are draft documents valid for a maximum of six months 32 and may be updated, replaced, or obsoleted by other documents at any 33 time. It is inappropriate to use Internet-Drafts as reference 34 material or to cite them other than as "work in progress." 36 This Internet-Draft will expire on July 16, 2015. 38 Copyright Notice 40 Copyright (c) 2015 IETF Trust and the persons identified as the 41 document authors. All rights reserved. 43 This document is subject to BCP 78 and the IETF Trust's Legal 44 Provisions Relating to IETF Documents 45 (http://trustee.ietf.org/license-info) in effect on the date of 46 publication of this document. Please review these documents 47 carefully, as they describe your rights and restrictions with respect 48 to this document. Code Components extracted from this document must 49 include Simplified BSD License text as described in Section 4.e of 50 the Trust Legal Provisions and are provided without warranty as 51 described in the Simplified BSD License. 53 Table of Contents 55 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 56 1.1. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . 3 57 1.2. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 4 58 1.3. Definition of Terms . . . . . . . . . . . . . . . . . . . 4 59 2. Architectural Concepts . . . . . . . . . . . . . . . . . . . 6 60 2.1. Service Function Chains . . . . . . . . . . . . . . . . . 7 61 2.2. Service Function Chain Symmetry . . . . . . . . . . . . . 8 62 2.3. Service Function Paths . . . . . . . . . . . . . . . . . 9 63 3. Architecture Principles . . . . . . . . . . . . . . . . . . . 10 64 4. Core SFC Architecture Components . . . . . . . . . . . . . . 11 65 4.1. SFC Encapsulation . . . . . . . . . . . . . . . . . . . . 12 66 4.2. Service Function (SF) . . . . . . . . . . . . . . . . . . 13 67 4.3. Service Function Forwarder (SFF) . . . . . . . . . . . . 13 68 4.3.1. Transport Derived SFF . . . . . . . . . . . . . . . . 15 69 4.4. SFC-Enabled Domain . . . . . . . . . . . . . . . . . . . 15 70 4.5. Network Overlay and Network Components . . . . . . . . . 15 71 4.6. SFC Proxy . . . . . . . . . . . . . . . . . . . . . . . . 15 72 4.7. Classification . . . . . . . . . . . . . . . . . . . . . 17 73 4.8. Re-Classification and Branching . . . . . . . . . . . . . 17 74 4.9. Shared Metadata . . . . . . . . . . . . . . . . . . . . . 18 75 5. Additional Architectural Concepts . . . . . . . . . . . . . . 18 76 5.1. The Role of Policy . . . . . . . . . . . . . . . . . . . 18 77 5.2. SFC Control Plane . . . . . . . . . . . . . . . . . . . . 19 78 5.3. Resource Control . . . . . . . . . . . . . . . . . . . . 20 79 5.4. Infinite Loop Detection and Avoidance . . . . . . . . . . 20 80 5.5. Load Balancing Considerations . . . . . . . . . . . . . . 21 81 5.6. MTU and Fragmentation Considerations . . . . . . . . . . 22 82 5.7. SFC OAM . . . . . . . . . . . . . . . . . . . . . . . . . 22 83 5.8. Resilience and Redundancy . . . . . . . . . . . . . . . . 23 84 6. Security Considerations . . . . . . . . . . . . . . . . . . . 24 85 7. Contributors and Acknowledgments . . . . . . . . . . . . . . 24 86 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 26 87 9. Informative References . . . . . . . . . . . . . . . . . . . 26 88 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 26 90 1. Introduction 92 The delivery of end-to-end services often requires various service 93 functions. These include traditional network service functions such 94 as firewalls and traditional IP Network Address Translators (NATs), 95 as well as application-specific functions. The definition and 96 instantiation of an ordered set of service functions and subsequent 97 'steering' of traffic through them is termed Service Function 98 Chaining (SFC). 100 This document describes an architecture used for the creation and 101 ongoing maintenance of Service Function Chains (SFC) in a network. 102 It includes architectural concepts, principles, and components, with 103 a focus on those to be standardized in the IETF. Service function 104 chains enable composite services that are constructed from one or 105 more service functions. 107 An overview of the issues associated with the deployment of end-to- 108 end service function chains, abstract sets of service functions and 109 their ordering constraints that create a composite service and the 110 subsequent "steering" of traffic flows through said service 111 functions, is described in [I-D.ietf-sfc-problem-statement]. 113 The current service function deployment models are relatively static, 114 coupled to network topology and physical resources, greatly reducing 115 or eliminating the ability of an operator to introduce new services 116 or dynamically create service function chains. This architecture 117 presents a model addressing the problematic aspects of existing 118 service deployments, including topological independence and 119 configuration complexity. 121 1.1. Scope 123 This document defines the architecture to realize Service Function 124 Chaining (SFC) with minimum requirements on the physical topology of 125 the network, as standardized in the IETF. In this architecture 126 packets are classifies on ingress for handling by the required set of 127 Service Functions (SFs) in the SFC-enabled domain and are then 128 forwarded through that set of functions for processing by each 129 function in turn. Packets may be reclassified as a result of this 130 processing. 132 The architecture described in this document is independent of the 133 planned usage of the network and deployment context and thus, for 134 example, is applicable to both fixed and mobile networks as well as 135 being useful in many Data Center applications. 137 The architecture described herein is assumed to be applicable to a 138 single network administrative domain. While it is possible for the 139 architectural principles and components to be applied to inter-domain 140 SFCs, these are left for future study. 142 1.2. Assumptions 144 The following assumptions are made: 146 o There is no standard definition or characterization applicable to 147 all SFs, and thus the architecture considers each SF as an opaque 148 processing element. 150 o There is no global or standard list of SFs enabled in a given 151 administrative domain. The set of SFs enabled in a given domain 152 is a function of the currently active services which may vary with 153 time and according to the networking environment. 155 o There is no global or standard SF chaining logic. The ordered set 156 of SFs that needs to be applied to deliver a given service is 157 specific to each administrative entity. 159 o The chaining of SFs and the criteria to invoke them are specific 160 to each administrative entity that operates an SF-enabled domain. 162 o Several SF chaining policies can be simultaneously applied within 163 an administrative domain to meet various business requirements. 165 o The underlay is assumed to provide the necessary connectivity to 166 interconnect the SFFs, but the architecture places no constraints 167 on how that connectivity is realized other than it have the 168 required bandwidth, latency, and jitter to support the SFC. 170 o No assumption is made on how FIBs and RIBs of involved nodes are 171 populated. 173 o How to bind traffic to a given SF chain is policy-based. 175 1.3. Definition of Terms 177 Network Service: An offering provided by an operator that is 178 delivered using one or more service functions. This may also be 179 referred to as a composite service. The term "service" is used 180 to denote a "network service" in the context of this document. 182 Note: Beyond this document, the term "service" is overloaded 183 with varying definitions. For example, to some a service is an 184 offering composed of several elements within the operator's 185 network, whereas for others a service, or more specifically a 186 network service, is a discrete element such as a "firewall". 187 Traditionally, such services (in the latter sense) host a set of 188 service functions and have a network locator where the service 189 is hosted. 191 Classification: Locally instantiated matching of traffic flows 192 against policy for subsequent application of the required set of 193 network service functions. The policy may be customer/network/ 194 service specific. 196 Classifier: An element that performs Classification. 198 Service Function Chain (SFC): A service function chain defines an 199 ordered set of abstract service functions (SFs) and ordering 200 constraints that must be applied to packets and/or frames and/or 201 flows selected as a result of classification. An example of an 202 abstract service function is "a firewall". The implied order 203 may not be a linear progression as the architecture allows for 204 SFCs that copy to more than one branch, and also allows for 205 cases where there is flexibility in the order in which service 206 functions need to be applied. The term service chain is often 207 used as shorthand for service function chain. 209 Service Function (SF): A function that is responsible for specific 210 treatment of received packets. A Service Function can act at 211 various layers of a protocol stack (e.g., at the network layer 212 or other OSI layers). As a logical component, a Service 213 Function can be realized as a virtual element or be embedded in 214 a physical network element. One or more Service Functions can 215 be embedded in the same network element. Multiple occurrences 216 of the Service Function can exist in the same administrative 217 domain. 219 One or more Service Functions can be involved in the delivery of 220 added-value services. A non-exhaustive list of abstract Service 221 Functions includes: firewalls, WAN and application acceleration, 222 Deep Packet Inspection (DPI), LI (Lawful Intercept), server load 223 balancing, NAT44 [RFC3022], NAT64 [RFC6146], NPTv6 [RFC6296], 224 HOST_ID injection, HTTP Header Enrichment functions, TCP 225 optimizer. 227 An SF may be SFC encapsulation aware, that is it receives and 228 acts on information in the SFC encapsulation, or unaware, in 229 which case data forwarded to the SF does not contain the SFC 230 encapsulation. 232 Service Function Forwarder (SFF): A service function forwarder is 233 responsible for forwarding traffic to one or more connected 234 service functions according to information carried in the SFC 235 encapsulation, as well as handling traffic coming back from the 236 SF. Additionally, a service function forwarder is responsible 237 for delivering traffic to a classifier when needed and 238 supported, transporting traffic to another SFF (in the same or 239 different type of overlay), and terminating the SFP. 241 Metadata: provides the ability to exchange context information 242 between classifiers and SFs and among SFs. 244 Service Function Path (SFP): The SFP provides a level of indirection 245 between the fully abstract notion of service chain as a sequence 246 of abstract service functions to be delivered, and the fully 247 specified notion of exactly which SFF/SFs the packet will visit 248 when it actually traverses the network. By allowing the control 249 components to specify this level of indirection, the operator 250 may control the degree of SFF/SF selection authority that is 251 delegated to the network. 253 SFC Encapsulation: The SFC Encapsulation provides at a minimum SFP 254 identification, and is used by the SFC-aware functions, such as 255 the SFF and SFC-aware SFs. The SFC Encapsulation is not used 256 for network packet forwarding. In addition to SFP 257 identification, the SFC encapsulation carries metadata including 258 data plane context information. 260 Rendered Service Path (RSP): The Service Function Path is a 261 constrained specification of where packets using a certain 262 service chain must go. While it may be so constrained as to 263 identify the exact locations, it can also be less specific. 264 Packets themselves are of course transmitted from and to 265 specific places in the network, visiting a specific sequence of 266 SFFs and SFs. This sequence of actual visits by a packet to 267 specific SFFs and SFs in the network is known as the Rendered 268 Service Path (RSP). This definition is included here for use by 269 later documents, such as when solutions may need to discuss the 270 actual sequence of locations the packets visit. 272 SFC-enabled Domain: A network or region of a network that implements 273 SFC. An SFC-enabled Domain is limited to a single network 274 administrative domain. 276 SFC Proxy: Removes and inserts SFC encapsulation on behalf of an 277 SFC-unaware service function. SFC proxies are logical elements. 279 2. Architectural Concepts 281 The following sections describe the foundational concepts of service 282 function chaining and the SFC architecture. 284 Service Function Chaining enables the creation of composite (network) 285 services that consist of an ordered set of Service Functions (SF) 286 that must be applied to packets and/or frames and/or flows selected 287 as a result of classification. Each SF is referenced using an 288 identifier that is unique within an SF-enabled domain. No IANA 289 registry is required to store the identity of SFs. 291 Service Function Chaining is a concept that provides for more than 292 just the application of an ordered set of SFs to selected traffic; 293 rather, it describes a method for deploying SFs in a way that enables 294 dynamic ordering and topological independence of those SFs as well as 295 the exchange of metadata between participating entities. 297 2.1. Service Function Chains 299 In most networks, services are constructed as abstract sequences of 300 SFs that represent SFCs. At a high level, an SFC is an abstracted 301 view of a service that specifies the set of required SFs as well as 302 the order in which they must be executed. Graphs, as illustrated in 303 Figure 1, define an SFC, where each graph node represents the 304 required existence of at least one abstract SF. Such graph nodes 305 (SFs) can be part of zero, one, or many SFCs. A given graph node 306 (SF) can appear one time or multiple times in a given SFC. 308 SFCs can start from the origination point of the service function 309 graph (i.e., node 1 in Figure 1), or from any subsequent node in the 310 graph. SFs may therefore become branching nodes in the graph, with 311 those SFs selecting edges that move traffic to one or more branches. 312 An SFC can have more than one terminus. 314 ,-+-. ,---. ,---. ,---. 315 / \ / \ / \ / \ 316 ( 1 )+--->( 2 )+---->( 6 )+---->( 8 ) 317 \ / \ / \ / \ / 318 `---' `---' `---' `---' 320 ,-+-. ,---. ,---. ,---. ,---. 321 / \ / \ / \ / \ / \ 322 ( 1 )+--->( 2 )+---->( 3 )+---->( 7 )+---->( 9 ) 323 \ / \ / \ / \ / \ / 324 `---' `---' `---' `---' `---' 326 ,-+-. ,---. ,---. ,---. ,---. 327 / \ / \ / \ / \ / \ 328 ( 1 )+--->( 7 )+---->( 8 )+---->( 4 )+---->( 7 ) 329 \ / \ / \ / \ / \ / 330 `---' `---' `---' `---' `---' 332 Figure 1: Service Function Chain Graphs 334 The concepts of classification, re-classification, and branching are 335 covered in subsequent sections of this architecture (see Section 4.7 336 and Section 4.8). 338 2.2. Service Function Chain Symmetry 340 SFCs may be unidirectional or bidirectional. A unidirectional SFC 341 requires that traffic be forwarded through the ordered SFs in one 342 direction (SF1 -> SF2 -> SF3), whereas a bidirectional SFC requires a 343 symmetric path (SF1 -> SF2 -> SF3 and SF3 -> SF2 -> SF1), and in 344 which the SF instances are the same in opposite directions. A hybrid 345 SFC has attributes of both unidirectional and bidirectional SFCs; 346 that is to say some SFs require symmetric traffic, whereas other SFs 347 do not process reverse traffic or are independent of the 348 corresponding forward traffic. 350 SFCs may contain cycles; that is traffic may need to traverse one or 351 more SFs within an SFC more than once. Solutions will need to ensure 352 suitable disambiguation for such situations. 354 The architectural allowance that is made for SFPs that delegate 355 choice to the network for which SFs and/or SFFs a packet will visit 356 creates potential issues here. A solution that allows such 357 delegation needs to also describe how the solution ensures that those 358 service chains that require service function chain symmetry can 359 achieve that. 361 Further, there are state tradeoffs in symmetry. Symmetry may be 362 realized in several ways depending on the SFF and classifier 363 functionality. In some cases, "mirrored" classification (i.e., from 364 Source to Destination and from Destination to Source) policy may be 365 deployed, whereas in others shared state between classifiers may be 366 used to ensure that symmetric flows are correctly identified, then 367 steered along the required SFP. At a high level, there are various 368 common cases. In a non-exhaustive way, there can be for example: 370 o A single classifier (or a small number of classifiers), in which 371 case both incoming and outgoing flows could be recognized at the 372 same classifier, so the synchronization would be feasible by 373 internal mechanisms internal to the classifier. 375 o Stateful classifiers where several classifiers may be clustered 376 and share state. 378 o Fully distributed classifiers, where synchronization needs to be 379 provided through unspecified means. 381 o A classifier that learns state from the egress packets/flows that 382 is then used to provide state for the return packets/flow. 384 o Symmetry may also be provided by stateful forwarding logic in the 385 SFF in some implementations. 387 This is a non-comprehensive list of common cases. 389 2.3. Service Function Paths 391 A service function path (SFP) is a mechanism used by service chaining 392 to express the result of applying more granular policy and 393 operational constraints to the abstract requirements of a service 394 chain (SFC). This architecture does not mandate the degree of 395 specificity of the SFP. Architecturally, within the same SFC-enabled 396 domain, some SFPs may be fully specified, selecting exactly which SFF 397 and which SF are to be visited by packets using that SFP, while other 398 SFPs may be quite vague, deferring to the SFF the decisions about the 399 exact sequence of steps to be used to realize the SFC. The 400 specificity may be anywhere in between these extremes. 402 As an example of such an intermediate specificity, there may be two 403 SFPs associated with a given SFC, where one SFP specifies that any 404 order of SFF and SF may be used as long as it is within data center 405 1, and where the second SFP allows the same latitude, but only within 406 data center 2. 408 Thus, the policies and logic of SFP selection or creation (depending 409 upon the solution) produce what may be thought of as a constrained 410 version of the original SFC. Since multiple policies may apply to 411 different traffic that uses the same SFC, it also follows that there 412 may be multiple SFPs may be associated with a single SFC. 414 The architecture allows for the same SF to be reachable through 415 multiple SFFs. In these cases, some SFPs may constrain which SFF is 416 used to reach which SF, while some SFPs may leave that decision to 417 the SFF itself. 419 Further, the architecture allows for two or more SFs to be attached 420 to the same SFF, and possibly connected via internal means allowing 421 more effective communication. In these cases, some solutions or 422 deployments may choose to use some form of internal inter-process or 423 inter-VM messaging (communication behind the virtual switching 424 element) that is optimized for such an environment. This must be 425 coordinated with the SFF so that the service function forwarding can 426 properly perform its job. Implementation details of such mechanisms 427 are considered out of scope for this document, and can include a 428 spectrum of methods: for example situations including all next-hops 429 explicitly, others where a list of possible next-hops is provided and 430 the selection is local, or cases with just an identifier, where all 431 resolution is local. 433 This architecture also allows the same SF to be part of multiple 434 SFPs. 436 3. Architecture Principles 438 Service function chaining is predicated on several key architectural 439 principles: 441 1. Topological independence: no changes to the underlay network 442 forwarding topology - implicit, or explicit - are needed to 443 deploy and invoke SFs or SFCs. 445 2. Plane separation: dynamic realization of SFPs is separated from 446 packet handling operations (e.g., packet forwarding). 448 3. Classification: traffic that satisfies classification rules is 449 forwarded according to a specific SFP. For example, 450 classification can be as simple as an explicit forwarding entry 451 that forwards all traffic from one address into the SFP. 452 Multiple classification points are possible within an SFC (i.e., 453 forming a service graph) thus enabling changes/updates to the SFC 454 by SFs. 456 Classification can occur at varying degrees of granularity; for 457 example, classification can use a 5-tuple, a transport port or 458 set of ports, part of the packet payload, it can be the result of 459 high-level inspections, or it can come from external systems. 461 4. Shared Metadata: Metadata/context data can be shared amongst SFs 462 and classifiers, between SFs, and between external systems and 463 SFs (e.g., orchestration). 465 One use of metadata is to provide and share the result of 466 classification (that occurs within the SFC-enabled domain, or 467 external to it) along an SFP. For example, an external 468 repository might provide user/subscriber information to a service 469 chain classifier. This classifier could in turn impose that 470 information in the SFC encapsulation for delivery to the 471 requisite SFs. The SFs could in turn utilize the user/subscriber 472 information for local policy decisions. Metadata can also share 473 SF output along the SFP. 475 5. Service definition independence: The SFC architecture does not 476 depend on the details of SFs themselves. Additionally, no IANA 477 registry is required to store the list of SFs. 479 6. Service function chain independence: The creation, modification, 480 or deletion of an SFC has no impact on other SFCs. The same is 481 true for SFPs. 483 7. Heterogeneous control/policy points: The architecture allows SFs 484 to use independent mechanisms (out of scope for this document) to 485 populate and resolve local policy and (if needed) local 486 classification criteria. 488 4. Core SFC Architecture Components 490 At a very high level, the logical architecture of an SFC-enabled 491 Domain comprises: 493 o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494 . +--------------+ +------------------~~~ 495 . | Service | SFC | Service +---+ +---+ 496 . |Classification| Encapsulation | Function |sf1|...|sfn| 497 +---->| Function |+---------------->| Path +---+ +---+ 498 . +--------------+ +------------------~~~ 499 . SFC-enabled Domain 500 o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502 Figure 2: Service Function Chain Architecture 504 The following sub-sections provide details on each logical component 505 that form the basis of the SFC architecture. A detailed overview of 506 how each of these architectural components interact is provided in 507 Figure 3: 509 +----------------+ +----------------+ 510 | SFC-aware | | SFC-unaware | 511 |Service Function| |Service Function| 512 +-------+--------+ +-------+--------+ 513 | | 514 SFC Encapsulation No SFC Encapsulation 515 | SFC | 516 +---------+ +----------------+ Encapsulation +---------+ 517 |SFC-Aware|-----------------+ \ +------------|SFC Proxy| 518 | SF | ... ----------+ \ \ / +---------+ 519 +---------+ \ \ \ / 520 +-------+--------+ 521 | SF Forwarder | 522 | (SFF) | 523 +-------+--------+ 524 | 525 SFC Encapsulation 526 | 527 ... SFC-enabled Domain ... 528 | 529 Network Overlay Transport 530 | 531 _,....._ 532 ,-' `-. 533 / `. 534 | Network | 535 `. / 536 `.__ __,-' 537 `'''' 539 Figure 3: Service Function Chain Architecture Components 541 4.1. SFC Encapsulation 543 The SFC encapsulation enables service function path selection. It 544 also enables the sharing of metadata/context information when such 545 metadata exchange is required. 547 The SFC encapsulation carries explicit information used to identify 548 the SFP. However, the SFC encapsulation is not a transport 549 encapsulation itself: it is not used to forward packets within the 550 network fabric. If packets need to flow between separate physical 551 platforms, the SFC encapsulation therefore relies on an outer network 552 transport. Transit forwarders -- such as router and switches -- 553 forward SFC encapsulated packets based on the outer (non-SFC) 554 encapsulation. 556 One of the key architecture principles of SFC is that the SFC 557 encapsulation remain transport independent. As such any network 558 transport protocol may be used to carry the SFC encapsulated traffic. 560 4.2. Service Function (SF) 562 The concept of an SF evolves; rather than being viewed as a bump in 563 the wire, an SF becomes a resource within a specified administrative 564 domain that is available for consumption as part of a composite 565 service. SFs send/receive data to/from one or more SFFs. SFC-aware 566 SFs receive this traffic with the SFC encapsulation. 568 While the SFC architecture defines the concept and specifies some 569 characteristics of a new encapsulation - the SFC encapsulation - and 570 several logical components for the construction of SFCs, existing SF 571 implementations may not have the capabilities to act upon or fully 572 integrate with the new SFC encapsulation. In order to provide a 573 mechanism for such SFs to participate in the architecture, an SFC 574 proxy function is defined (see Section 4.6). The SFC proxy acts as a 575 gateway between the SFC encapsulation and SFC-unaware SFs. The 576 integration of SFC-unaware service functions is discussed in more 577 detail in the SFC proxy section. 579 This architecture allows an SF to be part of multiple SFPs and SFCs. 581 4.3. Service Function Forwarder (SFF) 583 The SFF is responsible for forwarding packets and/or frames received 584 from the network to one or more SFs associated with a given SFF using 585 information conveyed in the SFC encapsulation. Traffic from SFs 586 eventually returns to the same SFF, which is responsible for 587 injecting traffic back onto the network. Some SFs, such as 588 firewalls, could also consume a packet. 590 The collection of SFFs and associated SFs creates a service plane 591 overlay in which SFC-aware SFs, as well as SFC-unaware SFs reside. 592 Within this service plane, the SFF component connects different SFs 593 that form a service function path. 595 SFFs maintain the requisite SFP forwarding information. SFP 596 forwarding information is associated with a service path identifier 597 that is used to uniquely identify an SFP. The service forwarding 598 state enables an SFF to identify which SFs of a given SFP should be 599 applied, and in what order, as traffic flows through the associated 600 SFP. While there may appear to the SFF to be only one available way 601 to deliver the given SF, there may also be multiple choices allowed 602 by the constraints of the SFP. 604 If there are multiple choices, the SFF needs to preserve the property 605 that all packets of a given flow are handled the same way, since the 606 SF may well be stateful. Additionally, the SFF may preserve the 607 handling of packets based on other properties on top of a flow, such 608 as a subscriber, session, or application instance identification. 610 The SFF also has the information to allow it to forward packets to 611 the next SFF after applying local service functions. Again, while 612 there may be only a single choice available, the architecture allows 613 for multiple choices for the next SFF. As with SFs, the solution 614 needs to operate such that the behavior with regard to specific flows 615 (see the Rendered Service Path) is stable. The selection of 616 available SFs and next SFFs may be interwoven when an SFF supports 617 multiple distinct service functions and the same service function is 618 available at multiple SFFs. Solutions need to be clear about what is 619 allowed in these cases. 621 Even when the SFF supports and utilizes multiple choices, the 622 decision as to whether to use flow-specific mechanisms or coarser 623 grained means to ensure that the behavior of specific flows is stable 624 is a matter for specific solutions and specific implementations. 626 The SFF component has the following primary responsibilities: 628 1. SFP forwarding : Traffic arrives at an SFF from the network. The 629 SFF determines the appropriate SF the traffic should be forwarded 630 to via information contained in the SFC encapsulation. Post-SF, 631 the traffic is returned to the SFF, and if needed forwarded to 632 another SF associated with that SFF. If there is another non- 633 local (i.e., different SFF) hop in the SFP, the SFF further 634 encapsulates the traffic in the appropriate network transport 635 protocol and delivers it to the network for delivery to the next 636 SFF along the path. Related to this forwarding responsibility, 637 an SFF should be able to interact with metadata. 639 2. Terminating SFPs : An SFC is completely executed when traffic has 640 traversed all required SFs in a chain. When traffic arrives at 641 the SFF after the last SF has finished processing it, the final 642 SFF knows from the service forwarding state that the SFC is 643 complete. The SFF removes the SFC encapsulation and delivers the 644 packet back to the network for forwarding. 646 3. Maintaining flow state: In some cases, the SFF may be stateful. 647 It creates flows and stores flow-centric information. This state 648 information may be used for a range of SFP-related tasks such as 649 ensuring consistent treatment of all packets in a given flow, 650 ensuring symmetry or for state-aware SFC Proxy functionality (see 651 Section 4.8). 653 4.3.1. Transport Derived SFF 655 Service function forwarding, as described above, directly depends 656 upon the use of the service path information contained in the SFC 657 encapsulation. However, existing implementations may not be able to 658 act on the SFC encapsulation. These platforms may opt to use 659 existing transport information if it can be arranged to provide 660 explicit service path information. 662 This results in the same architectural behavior and meaning for 663 service function forwarding and service function paths. It is the 664 responsibility of the control components to ensure that the transport 665 path executed in such a case is fully aligned with the path 666 identified by the information in the service chaining encapsulation. 668 4.4. SFC-Enabled Domain 670 Specific features may need to be enforced at the boundaries of an 671 SFC-enabled domain, for example to avoid leaking SFC information. 672 Using the term node to refer generically to an entity that is 673 performing a set of functions, in this context, an SFC Boundary Node 674 denotes a node that connects one SFC-enabled domain to a node either 675 located in another SFC-enabled domain or in a domain that is SFC- 676 unaware. 678 An SFC Boundary node can act as egress or ingress. An SFC Egress 679 Node denotes a SFC Boundary Node that handles traffic leaving the 680 SFC-enabled domain the Egress Node belongs to. Such a node is 681 required to remove any information specific to the SFC Domain, 682 typically the SFC Encapsulation. An SFC Ingress Node denotes an SFC 683 Boundary Node that handles traffic entering the SFC-enabled domain. 684 In most solutions and deployments this will need to include a 685 classifier, and will be responsible for adding the SFC encapsulation 686 to the packet. 688 4.5. Network Overlay and Network Components 690 Underneath the SFF there are components responsible for performing 691 the transport (overlay) forwarding. They do not consult the SFC 692 encapsulation or inner payload for performing this forwarding. They 693 only consult the outer-transport encapsulation for the transport 694 (overlay) forwarding. 696 4.6. SFC Proxy 698 In order for the SFC architecture to support SFC-unaware SFs (e.g., 699 legacy service functions) a logical SFC proxy function may be used. 701 This function sits between an SFF and one or more SFs to which the 702 SFF is directing traffic. 704 The proxy accepts packets from the SFF on behalf of the SF. It 705 removes the SFC encapsulation, and then uses a local attachment 706 circuit to deliver packets to SFC unaware SFs. It also receives 707 packets back from the SF, reapplies the SFC encapsulation, and 708 returns them to the SFF for processing along the service function 709 path. 711 Thus, from the point of view of the SFF, the SFC proxy appears to be 712 part of an SFC aware SF. 714 Communication details between the SFF and the SFC Proxy are the same 715 as those between the SFF and an SFC aware SF. The details of that 716 are not part of this architecture. The details of the communication 717 methods over the local attachment circuit between the SFC proxy and 718 the SFC-unaware SF are dependent upon the specific behaviors and 719 capabilities of that SFC-unaware SF, and thus are also out of scope 720 for this architecture. 722 Specifically, for traffic received from the SFF intended for the SF 723 the proxy is representing, the SFC proxy: 725 o Removes the SFC encapsulation from SFC encapsulated packets. 727 o Identifies the required SF to be applied based on available 728 information including that carried in the SFC encapsulation. 730 o Selects the appropriate outbound local attachment circuit through 731 which the next SF for this SFP is reachable. This is derived from 732 the identification of the SF carried in the SFC encapsulation, and 733 may include local techniques. Examples of a local attachment 734 circuit include, but are not limited to, VLAN, IP-in-IP, L2TPv3, 735 GRE, VXLAN. 737 o Forwards the original payload via the selected local attachment 738 circuit to the appropriate SF. 740 When traffic is returned from the SF: 742 o Applies the required SFC encapsulation. The determination of the 743 encapsulation details may be inferred by the local attachment 744 circuit through which the packet and/or frame was received, or via 745 packet classification, or other local policy. In some cases, 746 packet ordering or modification by the SF may necessitate 747 additional classification in order to re-apply the correct SFC 748 encapsulation. 750 o Delivers the packet with the SFC Encapsulation to the SFF, as 751 would happen with packets returned from an SFC-aware SF. 753 Alternatively, a service provider may decide to exclude legacy 754 service functions from an SFC domain. 756 4.7. Classification 758 Traffic from the network that satisfies classification criteria is 759 directed into an SFP and forwarded to the requisite service 760 function(s). Classification is handled by a service classification 761 function; initial classification occurs at the ingress to the SFC 762 domain. The granularity of the initial classification is determined 763 by the capabilities of the classifier and the requirements of the SFC 764 policy. For instance, classification might be relatively coarse: all 765 packets from this port are subject to SFC policy X and directed into 766 SFP A, or quite granular: all packets matching this 5-tuple are 767 subject to SFC policy Y and directed into SFP B. 769 As a consequence of the classification decision, the appropriate SFC 770 encapsulation is imposed on the data, and a suitable SFP is selected 771 or created. Classification results in attaching the traffic to a 772 specific SFP. 774 4.8. Re-Classification and Branching 776 The SFC architecture supports re-classification (or non-initial 777 classification) as well. As packets traverse an SFP, re- 778 classification may occur - typically performed by a classification 779 function co-resident with a service function. Reclassification may 780 result in the selection of a new SFP, an update of the associated 781 metadata, or both. This is referred to as "branching". 783 For example, an initial classification results in the selection of 784 SFP A: DPI_1 --> SLB_8. However, when the DPI service function is 785 executed, attack traffic is detected at the application layer. DPI_1 786 re-classifies the traffic as attack and alters the service path to 787 SFP B, to include a firewall for policy enforcement: dropping the 788 traffic: DPI_1 --> FW_4. Subsequent to FW_4, surviving traffic would 789 be returned to the original SFF. In this simple example, the DPI 790 service function re-classifies the traffic based on local application 791 layer classification capabilities (that were not available during the 792 initial classification step). 794 When traffic arrives after being steered through an SFC-unaware SF, 795 the SFC Proxy must perform re-classification of traffic to determine 796 the SFP. The SFC Proxy is concerned with re-attaching information 797 for SFC-unaware SFs, and a stateful SFC Proxy simplifies such 798 classification to a flow lookup. 800 4.9. Shared Metadata 802 Sharing metadata allows the network to provide network-derived 803 information to the SFs, SF-to-SF information exchange and the sharing 804 of service-derived information to the network. Some SFCs may not 805 require metadata exchange. SFC infrastructure enables the exchange 806 of this shared data along the SFP. The shared metadata serves 807 several possible roles within the SFC architecture: 809 o Allows elements that typically operate as ships in the night to 810 exchange information. 812 o Encodes information about the network and/or data for post- 813 service forwarding. 815 o Creates an identifier used for policy binding by SFs. 817 Context information can be derived in several ways: 819 o External sources 821 o Network node classification 823 o Service function classification 825 5. Additional Architectural Concepts 827 There are a number of issues which solutions need to address, and 828 which the architecture informs but does not determine. This section 829 lays out some of those concepts. 831 5.1. The Role of Policy 833 Much of the behavior of service chains is driven by operator and per- 834 customer policy. This architecture is structured to isolate the 835 policy interactions from the data plane and control logic. 837 Specifically, it is assumed that the service chaining control plane 838 creates the service paths. The service chaining data plane is used 839 to deliver the classified packets along the service chains to the 840 intended service functions. 842 Policy, in contrast, interacts with the system in other places. 843 Policies and policy engines may monitor service functions to decide 844 if additional (or fewer) instances of services are needed. When 845 applicable, those decisions may in turn result in interactions that 846 direct the control logic to change the SFP placement or packet 847 classification rules. 849 Similarly, operator service policy, often managed by operational or 850 business support systems (OSS or BSS), will frequently determine what 851 service functions are available. Operator service policies also 852 determine which sequences of functions are valid and are to be used 853 or made available. 855 The offering of service chains to customers, and the selection of 856 which service chain a customer wishes to use, are driven by a 857 combination of operator and customer policies using appropriate 858 portals in conjunction with the OSS and BSS tools. These selections 859 then drive the service chaining control logic, which in turn 860 establishes the appropriate packet classification rules. 862 5.2. SFC Control Plane 864 This is part of the overall architecture but outside the scope of 865 this document. 867 The SFC control plane is responsible for constructing SFPs, 868 translating SFCs to forwarding paths and propagating path information 869 to participating nodes to achieve requisite forwarding behavior to 870 construct the service overlay. For instance, an SFC construction may 871 be static; selecting exactly which SFFs and which SFs from those SFFs 872 are to be used, or it may be dynamic, allowing the network to perform 873 some or all of the choices of SFF or SF to use to deliver the 874 selected service chain within the constraints represented by the 875 service path. 877 In the SFC architecture, SFs are resources; the control plane manages 878 and communicates their capabilities, availability and location in 879 fashions suitable for the transport and SFC operations in use. The 880 control plane is also responsible for the creation of the context 881 (see below). The control plane may be distributed (using new or 882 existing control plane protocols), or be centralized, or a 883 combination of the two. 885 The SFC control plane provides the following functionality: 887 1. An SFC-enabled domain wide view of all available service function 888 resources as well as the network locators through which they are 889 reachable. 891 2. Uses SFC policy to construct service function chains, and 892 associated service function paths. 894 3. Selection of specific SFs for a requested SFC, either statically 895 (using specific SFs) or dynamically (using service explicit SFs 896 at the time of delivering traffic to them). 898 4. Provides requisite SFC data plane information to the SFC 899 architecture components, most notably the SFF. 901 5. Specification of the metadata associated with a given SFP and 902 propagation of said metadata to the relevant SFs and/or SFC 903 encapsulation-proxies or their respective policy planes. 905 5.3. Resource Control 907 The SFC system may be responsible for managing all resources 908 necessary for the SFC components to function. This includes network 909 constraints used to plan and choose network path(s) between service 910 function forwarders, network communication paths between service 911 function forwarders and their attached service functions, 912 characteristics of the nodes themselves such as memory, number of 913 virtual interfaces, routes, and instantiation, configuration, and 914 deletion of SFs. 916 The SFC system will also be required to reflect policy decisions 917 about resource control, as expressed by other components in the 918 system. 920 While all of these aspects are part of the overall system, they are 921 beyond the scope of this architecture. 923 5.4. Infinite Loop Detection and Avoidance 925 This SFC architecture is predicated on topological independence from 926 the underlying forwarding topology. Consequently, a service topology 927 is created by Service Function Paths or by the local decisions of the 928 Service Function Forwarders based on the constraints expressed in the 929 SFP. Due to the overlay constraints, the packet-forwarding path may 930 need to visit the same SFF multiple times, and in some less common 931 cases may even need to visit the same SF more than once. The Service 932 Chaining solution needs to permit these limited and policy-compliant 933 loops. At the same time, the solutions must ensure that indefinite 934 and unbounded loops cannot be formed, as such would consume unbounded 935 resources without delivering any value. 937 In other words, this architecture requires the solution to prevent 938 infinite Service Function Loops, even when Service Functions may be 939 invoked multiple times in the same SFP. 941 5.5. Load Balancing Considerations 943 Supporting function elasticity and high-availability should not 944 overly complicate SFC or lead to unnecessary scalability problems. 946 In the simplest case, where there is only a single function in the 947 SFP (the next hop is either the destination address of the flow or 948 the appropriate next hop to that destination), one could argue that 949 there may be no need for SFC. 951 In the cases where the classifier is separate from the single 952 function or a function at the terminal address may need sub-prefix or 953 per-subscriber metadata, a single SFP exists (the metadata changes 954 but the SFP does not), regardless of the number of potential terminal 955 addresses for the flow. This is the case of the simple load 956 balancer. See Figure 4. 958 +---+ +---++--->web server 959 source+-->|sff|+-->|sf1|+--->web server 960 +---+ +---++--->web server 962 Figure 4: Simple Load Balancing 964 By extrapolation, in the case where intermediary functions within a 965 chain had similar "elastic" behaviors, we do not need separate chains 966 to account for this behavior - as long as the traffic coalesces to a 967 common next-hop after the point of elasticity. 969 In Figure 5, we have a chain of five service functions between the 970 traffic source and its destination. 972 +---+ +---+ +---+ +---+ +---+ +---+ 973 |sf2| |sf2| |sf3| |sf3| |sf4| |sf4| 974 +---+ +---+ +---+ +---+ +---+ +---+ 975 | | | | | | 976 +-----+-----+ +-----+-----+ 977 | | 978 + + 979 +---+ +---+ +---+ +---+ +---+ 980 source+-->|sff|+-->|sff|+--->|sff|+--->|sff|+-->|sff|+-->destination 981 +---+ +---+ +---+ +---+ +---+ 982 + + + 983 | | | 984 +---+ +---+ +---+ 985 |sf1| |sf3| |sf5| 986 +---+ +---+ +---+ 988 Figure 5: Load Balancing 990 This would be represented as one service function path: 991 sf1->sf2->sf3->sf4->sf5. The SFF is a logical element, which may be 992 made up of one or multiple components. In this architecture, the SFF 993 may handle load distribution based on policy. 995 It can also be seen in the above that the same service function may 996 be reachable through multiple SFFs, as discussed earlier. The 997 selection of which SFF to use to reach SF3 may be made by the control 998 logic in defining the SFP, or may be left to the SFFs themselves, 999 depending upon policy, solution, and deployment constraints. In the 1000 latter case, it needs to be assured that exactly one SFF takes 1001 responsibility to steer traffic through SF3. 1003 5.6. MTU and Fragmentation Considerations 1005 This architecture prescribes additional information being added to 1006 packets to identify service function paths and often to represent 1007 metadata. It also envisions adding transport information to carry 1008 packets along service function paths, at least between service 1009 function forwarders. This added information increases the size of 1010 the packet to be carried by service chaining. Such additions could 1011 potentially increase the packet size beyond the MTU supported on some 1012 or all of the media used in the service chaining domain. 1014 Such packet size increases can thus cause operational MTU problems. 1015 Requiring fragmentation and reassembly in an SFF would be a major 1016 processing increase, and might be impossible with some transports. 1017 Expecting service functions to deal with packets fragmented by the 1018 SFC function might be onerous even when such fragmentation was 1019 possible. Thus, at the very least, solutions need to pay attention 1020 to the size cost of their approach. There may be alternative or 1021 additional means available, although any solution needs to consider 1022 the tradeoffs. 1024 These considerations apply to any generic architecture that increases 1025 the header size. There are also more specific MTU considerations: 1026 Effects on Path MTU Discovery (PMTUD) as well as deployment 1027 considerations. Deployments within a single administrative control 1028 or even a single Data Center complex can afford more flexibility in 1029 dealing with larger packets, and deploying existing mitigations that 1030 decrease the likelihood of fragmentation or discard. 1032 5.7. SFC OAM 1034 Operations, Administration, and Maintenance (OAM) tools are an 1035 integral part of the architecture. These serve various purposes, 1036 including fault detection and isolation, and performance management. 1037 For example, there are many advantages of SFP liveness detection, 1038 including status reporting, support for resiliency operations and 1039 policies, and an enhanced ability to balance load. 1041 Service Function Paths create a services topology, and OAM performs 1042 various functions within this service layer. Furthermore, SFC OAM 1043 follows the same architectural principles of SFC in general. For 1044 example, topological independence (including the ability to run OAM 1045 over various overlay technologies) and classification-based policy. 1047 We can subdivide the SFC OAM architecture in two parts: 1049 o In-band: OAM packets follow the same path and share fate with user 1050 packets, within the service topology. For this, they also follow 1051 the architectural principle of consistent policy identifiers, and 1052 use the same path IDs as the service chain data packets. Load 1053 balancing and SFC encapsulation with packet forwarding are 1054 particularly important here. 1056 o Out-of-band: reporting beyond the actual data plane. An 1057 additional layer beyond the data-plane OAM allows for additional 1058 alerting and measurements. 1060 This architecture prescribes end-to-end SFP OAM functions, which 1061 implies SFF understanding of whether an in-band packet is an OAM or 1062 user packet. However, service function validation is outside of the 1063 scope of this architecture, and application-level OAM is not what 1064 this architecture prescribes. 1066 Some of the detailed functions performed by SFC OAM include fault 1067 detection and isolation in a Service Function Path or a Service 1068 Function, verification that connectivity using SFPs is both effective 1069 and directing packets to the intended service functions, service path 1070 tracing, diagnostic and fault isolation, alarm reporting, performance 1071 measurement, locking and testing of service functions, validation 1072 with the control plane (see Section 5.2), and also allow for vendor- 1073 specific as well as experimental functions. SFC should leverage, and 1074 if needed extend relevant existing OAM mechanisms. 1076 5.8. Resilience and Redundancy 1078 As a practical operational requirement, any service chaining solution 1079 needs to be able to respond effectively, and usually very quickly, to 1080 failure conditions. These may be failures of connectivity in the 1081 network between SFFs, failures of SFFs, or failures of SFs. Per-SF 1082 state, as for example stateful-firewall state, is the responsibility 1083 of the SF, and not addressed by this architecture. 1085 Multiple techniques are available to address this issue. Solutions 1086 can describe both what they require and what they allow to address 1087 failure. Solutions can make use of flexible specificity of service 1088 function paths, if the SFF can be given enough information in a 1089 timely fashion to do this. Solutions can also make use of MAC or IP 1090 level redundancy mechanisms such as VRRP. Also, particularly for SF 1091 failures, load balancers co-located with the SFF or as part of the 1092 service function delivery mechanism can provide such robustness. 1094 Similarly, operational requirements imply resilience in the face of 1095 load changes. While mechanisms for managing (e.g., monitoring, 1096 instantiating, loading images, providing configuration to service 1097 function chaining control, deleting, etc.) virtual machines are out 1098 of scope for this architecture, solutions can and are aided by 1099 describing how they can make use of scaling mechanisms. 1101 6. Security Considerations 1103 Security considerations apply to the realization of this 1104 architecture, in particular to the documents that will define 1105 protocols. Such realization ought to provide means to protect the 1106 SFC-enabled domain and its borders against various forms of attacks, 1107 including DDoS attacks. Further, SFC OAM Functions need to not 1108 negatively affect the security considerations of an SFC-enabled 1109 domain. 1111 An operator may consider the SFC Metadata as sensitive. Therefore, 1112 solutions should consider whether there is a risk of sensitive 1113 information slipping out of the operators control. 1115 Additionally, all entities (software or hardware) interacting with 1116 the service chaining mechanisms need to provide means of security 1117 against malformed, poorly configured (deliberate or not) protocol 1118 constructs and loops. These considerations are largely the same as 1119 those in any network, particularly an overlay network. 1121 7. Contributors and Acknowledgments 1123 The editors would like to thank Sam Aldrin, Nicolas Bouthors, Stewart 1124 Bryant, Linda Dunbar, Alla Goldner, Ken Gray, Barry Greene, Anil 1125 Gunturu, David Harrington, Shunsuke Homma, Dave Hood, Nagendra Kumar, 1126 Hongyu Li, Andrew Malis, Guy Meador III, Kengo Naito, Ron Parker, 1127 Reinaldo Penno, Naiming Shen, Xiaohu Xu, and Lucy Yong for a thorough 1128 review and useful comments. 1130 The initial version of this "Service Function Chaining (SFC) 1131 Architecture" document is the result of merging two previous 1132 documents, and this section lists the aggregate of authors, editors, 1133 contributors and acknowledged participants, all who provided 1134 important ideas and text that fed into this architecture. 1136 [I-D.boucadair-sfc-framework]: 1138 Authors: 1140 Mohamed Boucadair 1141 Christian Jacquenet 1142 Ron Parker 1143 Diego R. Lopez 1144 Jim Guichard 1145 Carlos Pignataro 1147 Contributors: 1149 Parviz Yegani 1150 Paul Quinn 1151 Linda Dunbar 1153 Acknowledgements: 1155 Many thanks to D. Abgrall, D. Minodier, Y. Le Goff, D. 1156 Cheng, R. White, and B. Chatras for their review and 1157 comments. 1159 [I-D.quinn-sfc-arch]: 1161 Authors: 1163 Paul Quinn (editor) 1164 Joel Halpern (editor) 1166 Contributors: 1168 Puneet Agarwal 1169 Andre Beliveau 1170 Kevin Glavin 1171 Ken Gray 1172 Jim Guichard 1173 Surendra Kumar 1174 Darrel Lewis 1175 Nic Leymann 1176 Rajeev Manur 1177 Thomas Nadeau 1178 Carlos Pignataro 1179 Michael Smith 1180 Navindra Yadav 1182 Acknowledgements: 1184 The authors would like to thank David Ward, Abhijit Patra, 1185 Nagaraj Bagepalli, Darrel Lewis, Ron Parker, Lucy Yong and 1186 Christian Jacquenet for their review and comments. 1188 8. IANA Considerations 1190 [RFC Editor: please remove this section prior to publication.] 1192 This document has no IANA actions. 1194 9. Informative References 1196 [I-D.boucadair-sfc-framework] 1197 Boucadair, M., Jacquenet, C., Parker, R., Lopez, D., 1198 Guichard, J., and C. Pignataro, "Service Function 1199 Chaining: Framework & Architecture", draft-boucadair-sfc- 1200 framework-02 (work in progress), February 2014. 1202 [I-D.ietf-sfc-problem-statement] 1203 Quinn, P. and T. Nadeau, "Service Function Chaining 1204 Problem Statement", draft-ietf-sfc-problem-statement-10 1205 (work in progress), August 2014. 1207 [I-D.quinn-sfc-arch] 1208 Quinn, P. and J. Halpern, "Service Function Chaining (SFC) 1209 Architecture", draft-quinn-sfc-arch-05 (work in progress), 1210 May 2014. 1212 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 1213 Address Translator (Traditional NAT)", RFC 3022, January 1214 2001. 1216 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 1217 NAT64: Network Address and Protocol Translation from IPv6 1218 Clients to IPv4 Servers", RFC 6146, April 2011. 1220 [RFC6296] Wasserman, M. and F. Baker, "IPv6-to-IPv6 Network Prefix 1221 Translation", RFC 6296, June 2011. 1223 Authors' Addresses 1225 Joel Halpern (editor) 1226 Ericsson 1228 Email: jmh@joelhalpern.com 1229 Carlos Pignataro (editor) 1230 Cisco Systems, Inc. 1232 Email: cpignata@cisco.com