idnits 2.17.1 draft-ietf-sfc-architecture-01.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 (September 8, 2014) is 3515 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-13) exists of draft-ietf-sfc-problem-statement-10 Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--).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: March 12, 2015 Cisco 6 September 8, 2014 8 Service Function Chaining (SFC) Architecture 9 draft-ietf-sfc-architecture-01 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. This document does not propose solutions, 18 protocols, or extensions to existing protocols. 20 Status of This Memo 22 This Internet-Draft is submitted in full conformance with the 23 provisions of BCP 78 and BCP 79. 25 Internet-Drafts are working documents of the Internet Engineering 26 Task Force (IETF). Note that other groups may also distribute 27 working documents as Internet-Drafts. The list of current Internet- 28 Drafts is at http://datatracker.ietf.org/drafts/current/. 30 Internet-Drafts are draft documents valid for a maximum of six months 31 and may be updated, replaced, or obsoleted by other documents at any 32 time. It is inappropriate to use Internet-Drafts as reference 33 material or to cite them other than as "work in progress." 35 This Internet-Draft will expire on March 12, 2015. 37 Copyright Notice 39 Copyright (c) 2014 IETF Trust and the persons identified as the 40 document authors. All rights reserved. 42 This document is subject to BCP 78 and the IETF Trust's Legal 43 Provisions Relating to IETF Documents 44 (http://trustee.ietf.org/license-info) in effect on the date of 45 publication of this document. Please review these documents 46 carefully, as they describe your rights and restrictions with respect 47 to this document. Code Components extracted from this document must 48 include Simplified BSD License text as described in Section 4.e of 49 the Trust Legal Provisions and are provided without warranty as 50 described in the Simplified BSD License. 52 Table of Contents 54 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 55 1.1. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . 3 56 1.2. Assumptions . . . . . . . . . . . . . . . . . . . . . . . 3 57 1.3. Definition of Terms . . . . . . . . . . . . . . . . . . . 4 58 2. Architectural Concepts . . . . . . . . . . . . . . . . . . . 6 59 2.1. Service Function Chains . . . . . . . . . . . . . . . . . 6 60 2.2. Service Function Chain Symmetry . . . . . . . . . . . . . 7 61 2.3. Service Function Paths . . . . . . . . . . . . . . . . . 8 62 3. Architecture Principles . . . . . . . . . . . . . . . . . . . 9 63 4. Core SFC Architecture Components . . . . . . . . . . . . . . 10 64 4.1. SFC Encapsulation . . . . . . . . . . . . . . . . . . . . 11 65 4.2. Service Function (SF) . . . . . . . . . . . . . . . . . . 12 66 4.3. Service Function Forwarder (SFF) . . . . . . . . . . . . 12 67 4.3.1. Transport Derived SFF . . . . . . . . . . . . . . . . 14 68 4.4. SFC-Enabled Domain . . . . . . . . . . . . . . . . . . . 14 69 4.5. Network Overlay and Network Components . . . . . . . . . 14 70 4.6. SFC Proxy . . . . . . . . . . . . . . . . . . . . . . . . 14 71 4.7. Classification . . . . . . . . . . . . . . . . . . . . . 16 72 4.8. Re-Classification and Branching . . . . . . . . . . . . . 16 73 4.9. Shared Metadata . . . . . . . . . . . . . . . . . . . . . 17 74 5. Additional Architectural Concepts . . . . . . . . . . . . . . 17 75 5.1. The Role of Policy . . . . . . . . . . . . . . . . . . . 17 76 5.2. SFC Control Plane . . . . . . . . . . . . . . . . . . . . 18 77 5.3. Resource Control . . . . . . . . . . . . . . . . . . . . 19 78 5.4. Infinite Loop Detection and Avoidance . . . . . . . . . . 19 79 5.5. Load Balancing Considerations . . . . . . . . . . . . . . 20 80 5.6. MTU and Fragmentation Considerations . . . . . . . . . . 21 81 5.7. SFC OAM . . . . . . . . . . . . . . . . . . . . . . . . . 21 82 5.8. Resilience and Redundancy . . . . . . . . . . . . . . . . 22 83 6. Security Considerations . . . . . . . . . . . . . . . . . . . 23 84 7. Contributors and Acknowledgments . . . . . . . . . . . . . . 23 85 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 25 86 9. Informative References . . . . . . . . . . . . . . . . . . . 25 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 25 89 1. Introduction 91 This document describes an architecture used for the creation and 92 ongoing maintenance of Service Function Chains (SFC) in a network. 93 It includes architectural concepts, principles, and components. 95 An overview of the issues associated with the deployment of end-to- 96 end service function chains, abstract sets of service functions and 97 their ordering constraints that create a composite service and the 98 subsequent "steering" of traffic flows through said service 99 functions, is described in [I-D.ietf-sfc-problem-statement]. 101 This architecture presents a model addressing the problematic aspects 102 of existing service deployments, including topological independence 103 and configuration complexity. 105 Service function chains enable composite services that are 106 constructed from one or more service functions. 108 1.1. Scope 110 This document defines a framework to realize Service Function 111 Chaining (SFC) with minimum requirements on the physical topology of 112 the network. The proposed solution relies on initial packet 113 classification. Packets initially classified for handling by a set 114 of Service Functions (SFs) in the SFC-enabled domain are then 115 forwarded to that set of SFs for processing. 117 This document does not make any assumption on the deployment context. 118 The proposed framework covers both fixed and mobile networks. 120 The architecture described herein is assumed to be applicable to a 121 single network administrative domain. While it is possible for the 122 architectural principles and components to be applied to inter-domain 123 SFCs, these are left for future study. 125 1.2. Assumptions 127 The following assumptions are made: 129 o Not all SFs can be characterized with a standard definition in 130 terms of technical description, detailed specification, 131 configuration, etc. 133 o There is no global or standard list of SFs enabled in a given 134 administrative domain. The set of SFs varies as a function of the 135 service to be provided and according to the networking 136 environment. 138 o There is no global or standard SF chaining logic. The ordered set 139 of SFs that needs to be enabled to deliver a given service is 140 specific to each administrative entity. 142 o The chaining of SFs and the criteria to invoke them are specific 143 to each administrative entity that operates an SF-enabled domain. 145 o Several SF chaining policies can be simultaneously applied within 146 an administrative domain to meet various business requirements. 148 o No assumption is made on how FIBs and RIBs of involved nodes are 149 populated. 151 o How to bind traffic to a given SF chain is policy-based. 153 1.3. Definition of Terms 155 Network Service: An offering provided by an operator that is 156 delivered using one or more service functions. This may also be 157 referred to as a composite service. The term "service" is used 158 to denote a "network service" in the context of this document. 160 Note: Beyond this document, the term "service" is overloaded 161 with varying definitions. For example, to some a service is an 162 offering composed of several elements within the operator's 163 network, whereas for others a service, or more specifically a 164 network service, is a discrete element such as a firewall. 165 Traditionally, such services (in the latter sense) host a set of 166 service functions and have a network locator where the service 167 is hosted. 169 SFC Encapsulation: The SFC Encapsulation provides at a minimum SFP 170 identification, and is used by the SFC-aware functions, such as 171 the SFF and SFC-aware SFs. The SFC Encapsulation is not used 172 for network packet forwarding. In addition to SFP 173 identification, the SFC encapsulation carries dataplane context 174 information, also referred to as metadata. 176 Classification: Locally instantiated policy and customer/network/ 177 service profile matching of traffic flows for identification of 178 appropriate outbound forwarding actions. 180 Classifier: An element that performs Classification. 182 Service Function (SF): A function that is responsible for specific 183 treatment of received packets. A Service Function can act at 184 various layers of a protocol stack (e.g., at the network layer 185 or other OSI layers). As a logical component, a Service 186 Function can be realized as a virtual element or be embedded in 187 a physical network element. One of multiple Service Functions 188 can be embedded in the same network element. Multiple 189 occurrences of the Service Function can exist in the same 190 administrative domain. 192 One or more Service Functions can be involved in the delivery of 193 added-value services. A non-exhaustive list of abstract Service 194 Functions includes: firewalls, WAN and application acceleration, 195 Deep Packet Inspection (DPI), LI (Lawful Intercept), server load 196 balancing, NAT44 [RFC3022], NAT64 [RFC6146], NPTv6 [RFC6296], 197 HOST_ID injection, HTTP Header Enrichment functions, TCP 198 optimizer. 200 An SF may be SFC encapsulation aware, that is it receives and 201 acts on information in the SFC encapsulation, or unaware, in 202 which case data forwarded to the SF does not contain the SFC 203 encapsulation. 205 Service Function Forwarder (SFF): A service function forwarder is 206 responsible for delivering traffic received from the network to 207 one or more connected service functions according to information 208 carried in the SFC encapsulation, as well as handling traffic 209 coming back from the SF. Additionally, a service function 210 forwarder is responsible for delivering traffic to a classifier 211 when needed and supported, mapping out traffic to another SFF 212 (in the same or different type of overlay), and terminating the 213 SFP. 215 Service Function Chain (SFC): A service Function chain defines a set 216 of abstract service functions and ordering constraints that must 217 be applied to packets and/or frames selected as a result of 218 classification. The implied order may not be a linear 219 progression as the architecture allows for SFCs that copy to 220 more than one branch, and also allows for cases where there is 221 flexibility in the order in which service functions need to be 222 applied. The term service chain is often used as shorthand for 223 service function chain. 225 Service Function Path (SFP): The SFP provides a level of indirection 226 between the fully abstract notion of service chain as a sequence 227 of abstract service functions to be delivered, and the fully 228 specified notion of exactly which SFF/SFs the packet will visit 229 when it actually traverses the network. By allowing the control 230 components to specify this level of indirection, the operator 231 may control the degree of SFF/SF selection authority that is 232 delegated to the network. 234 Rendered Service Path (RSP): The Service Function Path is a 235 constrained specification of where packets using a certain 236 service chain must go. While it may be so constrained as to 237 identify the exact locations, it can also be less specific. 238 Packets themselves are of course transmitted from and to 239 specific places in the network, visiting a specific sequence of 240 SFFs and SFs. This sequence of actual visits by a packet to 241 specific SFFs and SFs in the network is known as the Rendered 242 Service Path (RSP). This definition is included here for use by 243 later documents, such as when solutions may need to discuss the 244 actual sequence of locations the packets visit. 246 SFC-enabled Domain: A network or region of a network that implements 247 SFC. An SFC-enabled Domain is limited to a single network 248 administrative domain. 250 SFC Proxy: Removes and inserts SFC encapsulation on behalf of an 251 SFC-unaware service function. SFC proxies are logical elements. 253 2. Architectural Concepts 255 The following sections describe the foundational concepts of service 256 function chaining and the SFC architecture. 258 Service Function Chaining enables the creation of composite (network) 259 services that consist of an ordered set of Service Functions (SF) 260 that must be applied to packets and/or frames selected as a result of 261 classification. Each SF is referenced using an identifier that is 262 unique within an SF-enabled domain. No IANA registry is required to 263 store the identity of SFs. 265 Service Function Chaining is a concept that provides for more than 266 just the application of an ordered set of SFs to selected traffic; 267 rather, it describes a method for deploying SFs in a way that enables 268 dynamic ordering and topological independence of those SFs as well as 269 the exchange of metadata between participating entities. 271 2.1. Service Function Chains 273 In most networks, services are constructed as abstract sequences of 274 SFs that represent SFCs. At a high level, an SFC is an abstracted 275 view of a service that specifies the set of required SFs as well as 276 the order in which they must be executed. Graphs, as illustrated in 277 Figure 1, define an SFC, where each graph node represents the 278 required existence of at least one abstract SF. Such graph nodes 279 (SFs) can be part of zero, one, or many SFCs. A given graph node 280 (SF) can appear one time or multiple times in a given SFC. 282 SFCs can start from the origination point of the service function 283 graph (i.e.: node 1 in Figure 1), or from any subsequent node in the 284 graph. SFs may therefore become branching nodes in the graph, with 285 those SFs selecting edges that move traffic to one or more branches. 286 An SFC can have more than one terminus. 288 ,-+-. ,---. ,---. ,---. 289 / \ / \ / \ / \ 290 ( 1 )+--->( 2 )+---->( 6 )+---->( 8 ) 291 \ / \ / \ / \ / 292 `---' `---' `---' `---' 294 ,-+-. ,---. ,---. ,---. ,---. 295 / \ / \ / \ / \ / \ 296 ( 1 )+--->( 2 )+---->( 3 )+---->( 7 )+---->( 9 ) 297 \ / \ / \ / \ / \ / 298 `---' `---' `---' `---' `---' 300 ,-+-. ,---. ,---. ,---. ,---. 301 / \ / \ / \ / \ / \ 302 ( 1 )+--->( 7 )+---->( 8 )+---->( 4 )+---->( 7 ) 303 \ / \ / \ / \ / \ / 304 `---' `---' `---' `---' `---' 306 Figure 1: Service Function Chain Graphs 308 2.2. Service Function Chain Symmetry 310 SFCs may be unidirectional or bidirectional. A unidirectional SFC 311 requires that traffic be forwarded through the ordered SFs in one 312 direction (SF1 -> SF2 -> SF3), whereas a bidirectional SFC requires a 313 symmetric path (SF1 -> SF2 -> SF3 and SF3 -> SF2 -> SF1), and in 314 which the SF instances are the same in opposite directions. A hybrid 315 SFC has attributes of both unidirectional and bidirectional SFCs; 316 that is to say some SFs require symmetric traffic, whereas other SFs 317 do not process reverse traffic or are independent of the 318 corresponding forward traffic. 320 SFCs may contain cycles; that is traffic may need to traverse one or 321 more SFs within an SFC more than once. Solutions will need to ensure 322 suitable disambiguation for such situations. 324 The architectural allowance that is made for SFPs that delegate 325 choice to the network for which SFs and/or SFFs a packet will visit 326 creates potential issues here. A solution that allows such 327 delegation needs to also describe how the solution ensures that those 328 service chains that require service function chain symmetry can 329 achieve that. 331 Further, there are state tradeoffs in symmetry. Symmetry may be 332 realized in several ways depending on the SFF and classifier 333 functionality. In some cases, "mirrored" classification (S -> D and 334 D -> S) policy may be deployed, whereas in others shared state 335 between classifiers may be used to ensure that symmetric flows are 336 correctly identified, then steered along the required SFP. At a high 337 level, there are various common cases. In a non-exhaustive way, 338 there can be for example: a single classifier (or a small number of 339 classifiers), in which case both incoming and outgoing flows could be 340 recognized at the same classifier, so the synchronization would be 341 feasible by internal mechanisms internal to the classifier. Another 342 case is the one of stateful classifiers where several classifiers may 343 be clustered and share state. Lastly, another case entails fully 344 distributed classifiers, where synchronization needs to be provided 345 through unspecified means. This is a non-comprehensive list of 346 common cases. 348 2.3. Service Function Paths 350 A service function path (SFP) is a mechanism used by service chaining 351 to express the result of applying more granular policy and 352 operational constraints to the abstract requirements of a service 353 chain (SFC). This architecture does not mandate the degree of 354 specificity of the SFP. Architecturally, within the same SFC-enabled 355 domain, some SFPs may be fully specified, selecting exactly which SFF 356 and which SF are to be visited by packets using that SFP, while other 357 SFPs may be quite vague, deferring to the SFF the decisions about the 358 exact sequence of steps to be used to realize the SFC. The 359 specificity may be anywhere in between these extremes. 361 As an example of such an intermediate specificity, there may be two 362 SFPs associated with a given SFC, where one SFP says essentially that 363 any order of SFF and SF may be used as long as it is within data 364 center 1, and where the second SFP allows the same latitude, but only 365 within data center 2. 367 Thus, the policies and logic of SFP selection or creation (depending 368 upon the solution) produce what may be thought of as a constrained 369 version of the original SFC. Since multiple policies may apply to 370 different traffic that uses the same SFC, it also follows that there 371 may be multiple SFPs may be associated with a single SFC. 373 The architecture allows for the same SF to be reachable through 374 multiple SFFs. In these cases, some SFPs may constrain which SFF is 375 used to reach which SF, while some SFPs may leave that decision to 376 the SFF itself. 378 Further, the architecture allows for two or more SFs to be attached 379 to the same SFF, and possibly connected via internal means allowing 380 more effective communication. In these cases, some solutions or 381 deployments may choose to use some form of internal inter-process or 382 inter-VM messaging (communication behind the virtual switching 383 element) that is optimized for such an environment. This must be 384 coordinated with the SFF so that the service function forwarding can 385 properly perform its job. Implementation details of such mechanisms 386 are considered out of scope for this document, and can include a 387 spectrum of methods: for example situations including all next-hops 388 explicitly, others where a list of possible next-hops is provided and 389 the selection is local, or cases with just an identifier, where all 390 resolution is local. 392 This architecture also allows the same SF to be part of multiple 393 SFPs. 395 3. Architecture Principles 397 Service function chaining is predicated on several key architectural 398 principles: 400 1. Topological independence: no changes to the underlay network 401 forwarding topology - implicit, or explicit - are needed to 402 deploy and invoke SFs or SFCs. 404 2. Plane separation: dynamic realization of SFPs is separated from 405 packet handling operations (e.g., packet forwarding). 407 3. Classification: traffic that satisfies classification rules is 408 forwarded according to a specific SFP. For example, 409 classification can be as simple as an explicit forwarding entry 410 that forwards all traffic from one address into the SFP. 411 Multiple classification points are possible within an SFC (i.e. 412 forming a service graph) thus enabling changes/updates to the SFC 413 by SFs. 415 Classification can occur at varying degrees of granularity; for 416 example, classification can use a 5-tuple, a transport port or 417 set of ports, part of the packet payload, or it can come from 418 external systems. 420 4. Shared Metadata: Metadata/context data can be shared amongst SFs 421 and classifiers, between SFs, and between external systems and 422 SFs (e.g., orchestration). 424 Generally speaking, metadata can be thought of as providing and 425 sharing the result of classification (that occurs within the SFC- 426 enabled domain, or external to it) along an SFP. For example, an 427 external repository might provide user/subscriber information to 428 a service chain classifier. This classifier could in turn impose 429 that information in the SFC encapsulation for delivery to the 430 requisite SFs. The SFs could in turn utilize the user/subscriber 431 information for local policy decisions. 433 5. Service definition independence: The SFC architecture does not 434 depend on the details of SFs themselves. Additionally, no IANA 435 registry is required to store the list of SFs. 437 6. Service function chain independence: The creation, modification, 438 or deletion of an SFC has no impact on other SFCs. The same is 439 true for SFPs. 441 7. Heterogeneous control/policy points: The architecture allows SFs 442 to use independent mechanisms (out of scope for this document) to 443 populate and resolve local policy and (if needed) local 444 classification criteria. 446 4. Core SFC Architecture Components 448 At a very high level, the logical architecture of an SFC-enabled 449 Domain comprises: 451 o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452 . +--------------+ +------------------~~~ 453 . | Service | SFC | Service +---+ +---+ 454 . |Classification| Encapsulation | Function |sf1|...|sfn| 455 +---->| Function |+---------------->| Path +---+ +---+ 456 . +--------------+ +------------------~~~ 457 . SFC-enabled Domain 458 o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460 Figure 2: Service Function Chain Architecture 462 The following sub-sections provide details on each logical component 463 that form the basis of the SFC architecture. A detailed overview of 464 how each of these architectural components interact is provided in 465 Figure 3: 467 +----------------+ +----------------+ 468 | SFC-aware | | SFC-unaware | 469 |Service Function| |Service Function| 470 +-------+--------+ +-------+--------+ 471 | | 472 SFC Encapsulation No SFC Encapsulation 473 | SFC | 474 +---------+ +----------------+ Encapsulation +---------+ 475 |SFC-Aware|-----------------+ \ +------------|SFC Proxy| 476 | SF | ... ----------+ \ \ / +---------+ 477 +---------+ \ \ \ / 478 +-------+--------+ 479 | SF Forwarder | 480 | (SFF) | 481 +-------+--------+ 482 | 483 SFC Encapsulation 484 | 485 ... SFC-enabled Domain ... 486 | 487 Network Overlay Transport 488 | 489 _,....._ 490 ,-' `-. 491 / `. 492 | Network | 493 `. / 494 `.__ __,-' 495 `'''' 497 Figure 3: Service Function Chain Architecture Components 499 4.1. SFC Encapsulation 501 The SFC encapsulation enables service function path selection. It 502 also enables the sharing of metadata/context information when such 503 metadata exchange is required. 505 The SFC encapsulation provides explicit information used to identify 506 the SFP. However, the SFC encapsulation is not a transport 507 encapsulation itself: it is not used to forward packets within the 508 network fabric. If packets need to flow between separate physical 509 platforms, the SFC encapsulation therefore relies on an outer network 510 transport. Transit forwarders -- such as router and switches -- 511 simply forward SFC encapsulated packets based on the outer (non-SFC) 512 encapsulation. 514 One of the key architecture principles of SFC is that the SFC 515 encapsulation remain transport independent. As such any network 516 transport protocol may be used to carry the SFC encapsulated traffic. 518 4.2. Service Function (SF) 520 The concept of an SF evolves; rather than being viewed as a bump in 521 the wire, an SF becomes a resource within a specified administrative 522 domain that is available for consumption as part of a composite 523 service. SFs send/receive data to/from one or more SFFs. SFC-aware 524 SFs receive this traffic with the SFC encapsulation. 526 While the SFC architecture defines a new encapsulation - the SFC 527 encapsulation - and several logical components for the construction 528 of SFCs, existing SF implementations may not have the capabilities to 529 act upon or fully integrate with the new SFC encapsulation. In order 530 to provide a mechanism for such SFs to participate in the 531 architecture, an SFC proxy function is defined. The SFC proxy acts 532 as a gateway between the SFC encapsulation and SFC-unaware SFs. The 533 integration of SFC-unaware service functions is discussed in more 534 detail in the SFC proxy section. 536 This architecture allows an SF to be part of multiple SFPs and SFCs. 538 4.3. Service Function Forwarder (SFF) 540 The SFF is responsible for forwarding packets and/or frames received 541 from the network to one or more SFs associated with a given SFF using 542 information conveyed in the SFC encapsulation. Traffic from SFs 543 eventually returns to the same SFF, which is responsible for putting 544 it back onto the network. 546 The collection of SFFs and associated SFs creates a service plane 547 overlay in which SFC-aware SFs, as well as SFC-unaware SFs reside. 548 Within this service plane, the SFF component connects different SFs 549 that form a service function path. 551 SFFs maintain the requisite SFP forwarding information. SFP 552 forwarding information is associated with a service path identifier 553 that is used to uniquely identify an SFP. The service forwarding 554 state enables an SFF to identify which SFs of a given SFP should be 555 applied, and in what order, as traffic flows through the associated 556 SFP. While there may appear to the SFF to be only one available way 557 to deliver the given SF, there may also be multiple choices allowed 558 by the constraints of the SFP. 560 If there are multiple choices, the SFF needs to preserve the property 561 that all packets of a given flow are handled the same way, since the 562 SF may well be stateful. Additionally, the SFF may preserve the 563 handling of packets based on other properties on top of a flow, such 564 as a subscriber, session, or application instance identification. 566 The SFF also has the information to allow it to forward packets to 567 the next SFF after applying local service functions. Again, while 568 there may be only a single choice available, the architecture allows 569 for multiple choices for the next SFF. As with SFs, the solution 570 needs to operate such that the behavior with regard to specific flows 571 (see the Rendered Service Path) is stable. It should be noted that 572 the selection of available SFs and next SFFs may be interwoven when 573 an SFF supports multiple distinct service functions and the same 574 service function is available at multiple SFFs. Solutions need to be 575 clear about what is allowed in these cases. 577 It should be noted that even when the SFF supports and utilizes 578 multiple choices, the decision as to whether to use flow-specific 579 mechanisms or coarser grained means to ensure that the behavior of 580 specific flows is stable is a matter for specific solutions and 581 specific implementations. 583 The SFF component has the following primary responsibilities: 585 1. SFP forwarding : Traffic arrives at an SFF from the network. The 586 SFF determines the appropriate SF the traffic should be forwarded 587 to via information contained in the SFC encapsulation. Post-SF, 588 the traffic is returned to the SFF, and if needed forwarded to 589 another SF associated with that SFF. If there is another non- 590 local (i.e., different SFF) hop in the SFP, the SFF further 591 encapsulates the traffic in the appropriate network transport 592 protocol and delivers it to the network for delivery to the next 593 SFF along the path. Related to this forwarding responsibility, 594 an SFF should be able to interact with metadata. 596 2. Terminating SFPs : An SFC is completely executed when traffic has 597 traversed all required SFs in a chain. When traffic arrives at 598 the SFF after the last SF has finished processing it, the final 599 SFF knows from the service forwarding state that the SFC is 600 complete. The SFF removes the SFC encapsulation and delivers the 601 packet back to the network for forwarding. 603 3. Maintaining flow state: In some cases, the SFF may be stateful. 604 It creates flows and stores flow-centric information. This state 605 information may be used for a range of SFP-related tasks such as 606 ensuring consistent treatment of all packets in a given flow, 607 ensuring symmetry or for state-aware SFC Proxy functionality (see 608 Section 4.8). 610 4.3.1. Transport Derived SFF 612 Service function forwarding, as described above, directly depends 613 upon the use of the service path information contained in the SFC 614 encapsulation. However, existing implementations may not be able to 615 act on the SFC encapsulation. These platforms may opt to use 616 existing transport information if it can be arranged to provide 617 explicit service path information. 619 This results in the same architectural behavior and meaning for 620 service function forwarding and service function paths. It is the 621 responsibility of the control components to ensure that the transport 622 path executed in such a case is fully aligned with the path 623 identified by the information in the service chaining encapsulation. 625 4.4. SFC-Enabled Domain 627 Specific features may need to be enforced at the boundaries of an 628 SFC-enabled domain, for example to avoid leaking SFC information. 629 Using the term node to refer generically to an entity that is 630 performing a set of functions, in this context, an SFC Boundary Node 631 denotes a node that connects one SFC-enabled domain to a node either 632 located in another SFC-enabled domain or in a domain that is SFC- 633 unaware. 635 An SFC Boundary node can act as egress or ingress. An SFC Egress 636 Node denotes a SFC Boundary Node that handles traffic leaving the 637 SFC-enabled domain the Egress Node belongs to. Such a node is 638 required to remove any information specific to the SFC Domain, 639 typically the SFC Encapsulation. An SFC Ingress Node denotes an SFC 640 Boundary Node that handles traffic entering the SFC-enabled domain. 641 In most solutions and deployments this will need to include a 642 classifier, and will be responsible for adding the SFC encapsulation 643 to the packet. 645 4.5. Network Overlay and Network Components 647 Underneath the SFF there are components responsible for performing 648 the transport (overlay) forwarding. They do not consult the SFC 649 encapsulation or inner payload for performing this forwarding. They 650 only consult the outer-transport encapsulation for the transport 651 (overlay) forwarding. 653 4.6. SFC Proxy 655 In order for the SFC architecture to support SFC-unaware SFs (.e.g 656 legacy service functions) a logical SFC proxy function may be used. 658 This function sits between an SFF and one or more SFs to which the 659 SFF is directing traffic. 661 The proxy accepts packets from the SFF on behalf of the SF. It 662 removes the SFC encapsulation, and then uses a local attachment 663 circuit to deliver packets to SFC unaware SFs. It also receives 664 packets back from the SF, reapplies the SFC encapsulation, and 665 returns them to the SFF for processing along the service function 666 path. 668 Thus, from the point of view of the SFF, the SFC proxy appears to be 669 part of an SFC aware SF. 671 Communication details between the SFF and the SFC Proxy are the same 672 as those between the SFF and an SFC aware SF. The details of that 673 are not part of this architecture. The details of the communication 674 methods over the local attachment circuit between the SFC proxy and 675 the SFC-unaware SF are dependent upon the specific behaviors and 676 capabilities of that SFC-unaware SF, and thus are also out of scope 677 for this architecture. 679 Specifically, for traffic received from the SFF intended for the SF 680 the proxy is representing, the SFC proxy: 682 o Removes the SFC encapsulation from SFC encapsulated packets. 684 o Identifies the required SF to be applied based on available 685 information including that carried in the SFC encapsulation. 687 o Selects the appropriate outbound local attachment circuit through 688 which the next SF for this SFP is reachable. This is derived from 689 the identification of the SF carried in the SFC encapsulation, and 690 may include local techniques. Examples of a local attachment 691 circuit include, but are not limited to, VLAN, IP-in-IP, L2TPv3, 692 GRE, VXLAN. 694 o Forwards the original payload via the selected local attachment 695 circuit to the appropriate SF. 697 When traffic is returned from the SF: 699 o Applies the required SFC encapsulation. The determination of the 700 encapsulation details may be inferred by the local attachment 701 circuit through which the packet and/or frame was received, or via 702 packet classification, or other local policy. In some cases, 703 packet ordering or modification by the SF may necessitate 704 additional classification in order to re-apply the correct SFC 705 encapsulation. 707 o Delivers the packet with the SFC Encapsulation to the SFF, as 708 would happen with packets returned from an SFC-aware SF. 710 Alternatively, a service provider may decide to exclude legacy 711 service functions from an SFC domain. 713 4.7. Classification 715 Traffic from the network that satisfies classification criteria is 716 directed into an SFP and forwarded to the requisite service 717 function(s). Classification is handled by a service classification 718 function; initial classification occurs at the ingress to the SFC 719 domain. The granularity of the initial classification is determined 720 by the capabilities of the classifier and the requirements of the SFC 721 policy. For instance, classification might be relatively coarse: all 722 packets from this port are subject to SFC policy X and directed into 723 SFP A, or quite granular: all packets matching this 5-tuple are 724 subject to SFC policy Y and directed into SFP B. 726 As a consequence of the classification decision, the appropriate SFC 727 encapsulation is imposed on the data, and a suitable SFP is selected 728 or created. Classification results in attaching the traffic to a 729 specific SFP. 731 4.8. Re-Classification and Branching 733 The SFC architecture supports re-classification (or non-initial 734 classification) as well. As packets traverse an SFP, re- 735 classification may occur - typically performed by a classification 736 function co-resident with a service function. Reclassification may 737 result in the selection of a new SFP, an update of the associated 738 metadata, or both. This is referred to as "branching". 740 For example, an initial classification results in the selection of 741 SFP A: DPI_1 --> SLB_8. However, when the DPI service function is 742 executed, attack traffic is detected at the application layer. DPI_1 743 re-classifies the traffic as attack and alters the service path to 744 SFP B, to include a firewall for policy enforcement: dropping the 745 traffic: DPI_1 --> FW_4. Subsequent to FW_4, surviving traffic would 746 be returned to the original SFF. In this simple example, the DPI 747 service function re-classifies the traffic based on local application 748 layer classification capabilities (that were not available during the 749 initial classification step). 751 When traffic arrives after being steered through an SFC-unaware SF, 752 the SFC Proxy must perform re-classification of traffic to determine 753 the SFP. The SFC Proxy is concerned with re-attaching information 754 for SFC-unaware SFs, and a stateful SFC Proxy simplifies such 755 classification to a flow lookup. 757 4.9. Shared Metadata 759 Sharing metadata allows the network to provide network-derived 760 information to the SFs, SF-to-SF information exchange and the sharing 761 of service-derived information to the network. Some SFCs may not 762 require metadata exchange. SFC infrastructure enables the exchange 763 of this shared data along the SFP. The shared metadata serves 764 several possible roles within the SFC architecture: 766 o Allows elements that typically operate as ships in the night to 767 exchange information. 769 o Encodes information about the network and/or data for post- 770 service forwarding. 772 o Creates an identifier used for policy binding by SFs. 774 Context information can be derived in several ways: 776 o External sources 778 o Network node classification 780 o Service function classification 782 5. Additional Architectural Concepts 784 There are a number of issues which solutions need to address, and 785 which the architecture informs but does not determine. This section 786 lays out some of those concepts. 788 5.1. The Role of Policy 790 Much of the behavior of service chains is driven by operator and per- 791 customer policy. This architecture is structured to isolate the 792 policy interactions from the data plane and control logic. 794 Specifically, it is assumed that the service chaining control plane 795 creates the service paths. The service chaining data plane is used 796 to deliver the classified packets along the service chains to the 797 intended service functions. 799 Policy, in contrast, interacts with the system in other places. 800 Policies and policy engines may monitor service functions to decide 801 if additional (or fewer) instances of services are needed. When 802 applicable, those decisions may in turn result in interactions that 803 direct the control logic to change the SFP placement or packet 804 classification rules. 806 Similarly, operator service policy, often managed by operational or 807 business support systems (OSS or BSS), will frequently determine what 808 service functions are available. Operator service policies also 809 determine which sequences of functions are valid and are to be used 810 or made available. 812 The offering of service chains to customers, and the selection of 813 which service chain a customer wishes to use, are driven by a 814 combination of operator and customer policies using appropriate 815 portals in conjunction with the OSS and BSS tools. These selections 816 then drive the service chaining control logic, which in turn 817 establishes the appropriate packet classification rules. 819 5.2. SFC Control Plane 821 This is part of the overall architecture but outside the scope of 822 this document. 824 The SFC control plane is responsible for constructing SFPs, 825 translating SFCs to forwarding paths and propagating path information 826 to participating nodes to achieve requisite forwarding behavior to 827 construct the service overlay. For instance, an SFC construction may 828 be static; selecting exactly which SFFs and which SFs from those SFFs 829 are to be used, or it may be dynamic, allowing the network to perform 830 some or all of the choices of SFF or SF to use to deliver the 831 selected service chain within the constraints represented by the 832 service path. 834 In the SFC architecture, SFs are resources; the control plane manages 835 and communicates their capabilities, availability and location in 836 fashions suitable for the transport and SFC operations in use. The 837 control plane is also responsible for the creation of the context 838 (see below). The control plane may be distributed (using new or 839 existing control plane protocols), or be centralized, or a 840 combination of the two. 842 The SFC control plane provides the following functionality: 844 1. An SFC-enabled domain wide view of all available service function 845 resources as well as the network locators through which they are 846 reachable. 848 2. Uses SFC policy to construct service function chains, and 849 associated service function paths. 851 3. Selection of specific SFs for a requested SFC, either statically 852 (using specific SFs) or dynamically (using service explicit SFs 853 at the time of delivering traffic to them). 855 4. Provides requisite SFC data plane information to the SFC 856 architecture components, most notably the SFF. 858 5. Allocation of metadata associated with a given SFP and 859 propagation of the metadata to relevant SFs and/or SFC 860 encapsulation-proxies or their respective policy planes. 862 5.3. Resource Control 864 The SFC system may be responsible for managing all resources 865 necessary for the SFC components to function. This includes network 866 constraints used to plan and choose network path(s) between service 867 function forwarders, network communication paths between service 868 function forwarders and their attached service functions, 869 characteristics of the nodes themselves such as memory, number of 870 virtual interfaces, routes, and instantiation, configuration, and 871 deletion of SFs. 873 The SFC system will also be required to reflect policy decisions 874 about resource control, as expressed by other components in the 875 system. 877 While all of these aspects are part of the overall system, they are 878 beyond the scope of this architecture. 880 5.4. Infinite Loop Detection and Avoidance 882 This SFC architecture is predicated on topological independence from 883 the underlying forwarding topology. Consequently, a service topology 884 is created by Service Function Paths or by the local decisions of the 885 Service Function Forwarders based on the constraints expressed in the 886 SFP. Due to the overlay constraints, the packet-forwarding path may 887 need to visit the same SFF multiple times, and in some less common 888 cases may even need to visit the same SF more than once. The Service 889 Chaining solution needs to permit these limited and policy-compliant 890 loops. At the same time, the solutions must ensure that indefinite 891 and unbounded loops cannot be formed, as such would consume unbounded 892 resources without delivering any value. 894 In other words, this architecture prevents infinite Service Function 895 Loops, even when Service Functions may be invoked multiple times in 896 the same SFP. 898 5.5. Load Balancing Considerations 900 Supporting function elasticity and high-availability should not 901 overly complicate SFC or lead to unnecessary scalability problems. 903 In the simplest case, where there is only a single function in the 904 SFP (the next hop is either the destination address of the flow or 905 the appropriate next hop to that destination), one could argue that 906 there may be no need for SFC. 908 In the cases where the classifier is separate from the single 909 function or a function at the terminal address may need sub-prefix or 910 per-subscriber metadata, a single SFP exists (the metadata changes 911 but the SFP does not), regardless of the number of potential terminal 912 addresses for the flow. This is the case of the simple load 913 balancer. See Figure 4. 915 +---+ +---++--->web server 916 source+-->|sff|+-->|sf1|+--->web server 917 +---+ +---++--->web server 919 Figure 4: Simple Load Balancing 921 By extrapolation, in the case where intermediary functions within a 922 chain had similar "elastic" behaviors, we do not need separate chains 923 to account for this behavior - as long as the traffic coalesces to a 924 common next-hop after the point of elasticity. 926 In Figure 5, we have a chain of five service functions between the 927 traffic source and its destination. 929 +---+ +---+ +---+ +---+ +---+ +---+ 930 |sf2| |sf2| |sf3| |sf3| |sf4| |sf4| 931 +---+ +---+ +---+ +---+ +---+ +---+ 932 | | | | | | 933 +-----+-----+ +-----+-----+ 934 | | 935 + + 936 +---+ +---+ +---+ +---+ +---+ 937 source+-->|sff|+-->|sff|+--->|sff|+--->|sff|+-->|sff|+-->destination 938 +---+ +---+ +---+ +---+ +---+ 939 + + + 940 | | | 941 +---+ +---+ +---+ 942 |sf1| |sf3| |sf5| 943 +---+ +---+ +---+ 945 Figure 5: Load Balancing 947 This would be represented as one service function path: 948 sf1->sf2->sf3->sf4->sf5. The SFF is a logical element, which may be 949 made up of one or multiple components. In this architecture, the SFF 950 may handle load distribution based on policy. 952 It can also be seen in the above that the same service function may 953 be reachable through multiple SFFs, as discussed earlier. The 954 selection of which SFF to use to reach SF3 may be made by the control 955 logic in defining the SFP, or may be left to the SFFs themselves, 956 depending upon policy, solution, and deployment constraints. In the 957 latter case, it needs to be assured that exactly one SFF takes 958 responsibility to steer traffic through SF3. 960 5.6. MTU and Fragmentation Considerations 962 This architecture prescribes additional information being added to 963 packets to identify service function paths and often to represent 964 metadata. It also envisions adding transport information to carry 965 packets along service function paths, at least between service 966 function forwarders. This added information increases the size of 967 the packet to be carried by service chaining. Such additions could 968 potentially increase the packet size beyond the MTU supported on some 969 or all of the media used in the service chaining domain. 971 Such packet size increases can thus cause operational MTU problems. 972 Requiring fragmentation and reassembly in an SFF would be a major 973 processing increase, and might be impossible with some transports. 974 Expecting service functions to deal with packets fragmented by the 975 SFC function might be onerous even when such fragmentation was 976 possible. Thus, at the very least, solutions need to pay attention 977 to the size cost of their approach. There may be alternative or 978 additional means available, although any solution needs to consider 979 the tradeoffs. 981 These considerations apply to any generic architecture that increases 982 the header size. There are also more specific MTU considerations: 983 Effects on Path MTU Discovery (PMTUD) as well as deployment 984 considerations. Deployments within a single administrateive control 985 or even a single Data Center complex can afford more flexibility in 986 dealing with larger packets, and deploying existing mitigations that 987 decrease the likelihood of fragmentation or discard. 989 5.7. SFC OAM 991 Operations, Administration, and Maintenance (OAM) tools are an 992 integral part of the architecture. These serve various purposes, 993 including fault detection and isolation, and performance management. 994 For example, there are many advantages of SFP liveness detection, 995 including status reporting, support for resiliency operations and 996 policies, and an enhanced ability to balance load. 998 Service Function Paths create a services topology, and OAM performs 999 various functions within this service layer. Furthermore, SFC OAM 1000 follows the same architectural principles of SFC in general. For 1001 example, topological independence (including the ability to run OAM 1002 over various overlay technologies) and classification-based policy. 1004 We can subdivide the SFC OAM architecture in two parts: 1006 o In-band: OAM packets follow the same path and share fate with user 1007 packets, within the service topology. For this, they also follow 1008 the architectural principle of consistent policy identifiers, and 1009 use the same path IDs as the service chain data packets. Load 1010 balancing and SFC encapsulation with packet forwarding are 1011 particularly important here. 1013 o Out-of-band: reporting beyond the actual data plane. An 1014 additional layer beyond the data-plane OAM allows for additional 1015 alerting and measurements. 1017 This architecture prescribes end-to-end SFP OAM functions, which 1018 implies SFF understanding of whether an in-band packet is an OAM or 1019 user packet. However, service function validation is outside of the 1020 scope of this architecture, and application-level OAM is not what 1021 this architecture prescribes. 1023 Some of the detailed functions performed by SFC OAM include fault 1024 detection and isolation in a Service Function Path or a Service 1025 Function, verification that connectivity using SFPs is both effective 1026 and directing packets to the intended service functions, service path 1027 tracing, diagnostic and fault isolation, alarm reporting, performance 1028 measurement, locking and testing of service functions, validation 1029 with the control plane (see Section 5.2), and also allow for vendor- 1030 specific as well as experimental functions. SFC should leverage, and 1031 if needed extend relevant existing OAM mechanisms. 1033 5.8. Resilience and Redundancy 1035 As a practical operational requirement, any service chaining solution 1036 needs to be able to respond effectively, and usually very quickly, to 1037 failure conditions. These may be failures of connectivity in the 1038 network between SFFs, failures of SFFs, or failures of SFs. Per-SF 1039 state, as for example stateful-firewall state, is the responsibility 1040 of the SF, and not addressed by this architecture. 1042 Multiple techniques are available to address this issue. Solutions 1043 can describe both what they require and what they allow to address 1044 failure. Solutions can make use of flexible specificity of service 1045 function paths, if the SFF can be given enough information in a 1046 timely fashion to do this. Solutions can also make use of MAC or IP 1047 level redundancy mechanisms such as VRRP. Also, particularly for SF 1048 failures, load balancers co-located with the SFF or as part of the 1049 service function delivery mechanism can provide such robustness. 1051 Similarly, operational requirements imply resilience in the face of 1052 load changes. While mechanisms for managing (e.g., monitoring, 1053 instantiating, loading images, providing configuration to service 1054 function chaining control, deleting, etc.) virtual machines are out 1055 of scope for this architecture, solutions can and are aided by 1056 describing how they can make use of scaling mechanisms. 1058 6. Security Considerations 1060 This document does not define a new protocol and therefore creates no 1061 new security issues. 1063 Security considerations apply to the realization of this 1064 architecture. Such realization ought to provide means to protect the 1065 SFC-enabled domain and its borders against various forms of attacks, 1066 including DDoS attacks. Further, SFC OAM Functions need to not 1067 negatively affect the security considerations of an SFC-enabled 1068 domain. Additionally, all entities (software or hardware) 1069 interacting with the service chaining mechanisms need to provide 1070 means of security against malformed, poorly configured (deliberate or 1071 not) protocol constructs and loops. These considerations are largely 1072 the same as those in any network, particularly an overlay network. 1074 7. Contributors and Acknowledgments 1076 The editors would like to thank Sam Aldrin, Linda Dunbar, Alla 1077 Goldner, Ken Gray, Anil Gunturu, Shunsuke Homma, Dave Hood, Nagendra 1078 Kumar, Hongyu Li, Andrew Malis, Guy Meador III, Kengo Naito, Ron 1079 Parker, Reinaldo Penno, Naiming Shen, Xiaohu Xu, and Lucy Yong for a 1080 thorough review and useful comments. 1082 The initial version of this "Service Function Chaining (SFC) 1083 Architecture" document is the result of merging two previous 1084 documents, and this section lists the aggregate of authors, editors, 1085 contributors and acknowledged participants, all who provided 1086 important ideas and text that fed into this architecture. 1088 [I-D.boucadair-sfc-framework]: 1090 Authors: 1092 Mohamed Boucadair 1093 Christian Jacquenet 1094 Ron Parker 1095 Diego R. Lopez 1096 Jim Guichard 1097 Carlos Pignataro 1099 Contributors: 1101 Parviz Yegani 1102 Paul Quinn 1103 Linda Dunbar 1105 Acknowledgements: 1107 Many thanks to D. Abgrall, D. Minodier, Y. Le Goff, D. 1108 Cheng, R. White, and B. Chatras for their review and 1109 comments. 1111 [I-D.quinn-sfc-arch]: 1113 Authors: 1115 Paul Quinn (editor) 1116 Joel Halpern (editor) 1118 Contributors: 1120 Puneet Agarwal 1121 Andre Beliveau 1122 Kevin Glavin 1123 Ken Gray 1124 Jim Guichard 1125 Surendra Kumar 1126 Darrel Lewis 1127 Nic Leymann 1128 Rajeev Manur 1129 Thomas Nadeau 1130 Carlos Pignataro 1131 Michael Smith 1132 Navindra Yadav 1134 Acknowledgements: 1136 The authors would like to thank David Ward, Abhijit Patra, 1137 Nagaraj Bagepalli, Darrel Lewis, Ron Parker, Lucy Yong and 1138 Christian Jacquenet for their review and comments. 1140 8. IANA Considerations 1142 [RFC Editor: please remove this section prior to publication.] 1144 This document has no IANA actions. 1146 9. Informative References 1148 [I-D.boucadair-sfc-framework] 1149 Boucadair, M., Jacquenet, C., Parker, R., Lopez, D., 1150 Guichard, J., and C. Pignataro, "Service Function 1151 Chaining: Framework & Architecture", draft-boucadair-sfc- 1152 framework-02 (work in progress), February 2014. 1154 [I-D.ietf-sfc-problem-statement] 1155 Quinn, P. and T. Nadeau, "Service Function Chaining 1156 Problem Statement", draft-ietf-sfc-problem-statement-10 1157 (work in progress), August 2014. 1159 [I-D.quinn-sfc-arch] 1160 Quinn, P. and J. Halpern, "Service Function Chaining (SFC) 1161 Architecture", draft-quinn-sfc-arch-05 (work in progress), 1162 May 2014. 1164 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 1165 Address Translator (Traditional NAT)", RFC 3022, January 1166 2001. 1168 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 1169 NAT64: Network Address and Protocol Translation from IPv6 1170 Clients to IPv4 Servers", RFC 6146, April 2011. 1172 [RFC6296] Wasserman, M. and F. Baker, "IPv6-to-IPv6 Network Prefix 1173 Translation", RFC 6296, June 2011. 1175 Authors' Addresses 1177 Joel Halpern (editor) 1178 Ericsson 1180 Email: jmh@joelhalpern.com 1181 Carlos Pignataro (editor) 1182 Cisco Systems, Inc. 1184 Email: cpignata@cisco.com