idnits 2.17.1 draft-ietf-sfc-problem-statement-09.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (August 8, 2014) is 3550 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- No issues found here. Summary: 0 errors (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group P. Quinn, Ed. 3 Internet-Draft Cisco Systems, Inc. 4 Intended status: Informational T. Nadeau, Ed. 5 Expires: February 9, 2015 Brocade 6 August 8, 2014 8 Service Function Chaining Problem Statement 9 draft-ietf-sfc-problem-statement-09.txt 11 Abstract 13 This document provides an overview of the issues associated with the 14 deployment of service functions (such as firewalls, load balancers) 15 in large-scale environments. The term service function chaining is 16 used to describe the definition and instantiation of an ordered set 17 of instances of such service functions, and the subsequent "steering" 18 of traffic flows through those service functions. 20 The set of enabled service function chains reflect operator service 21 offerings and is designed in conjunction with application delivery 22 and service and network policy. 24 Status of this Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on February 9, 2015. 41 Copyright Notice 43 Copyright (c) 2014 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 59 1.1. Definition of Terms . . . . . . . . . . . . . . . . . . . 3 60 2. Problem Space . . . . . . . . . . . . . . . . . . . . . . . . 5 61 2.1. Topological Dependencies . . . . . . . . . . . . . . . . . 5 62 2.2. Configuration complexity . . . . . . . . . . . . . . . . . 5 63 2.3. Constrained High Availability . . . . . . . . . . . . . . 6 64 2.4. Consistent Ordering of Service Functions . . . . . . . . . 6 65 2.5. Application of Service Policy . . . . . . . . . . . . . . 6 66 2.6. Transport Dependence . . . . . . . . . . . . . . . . . . . 7 67 2.7. Elastic Service Delivery . . . . . . . . . . . . . . . . . 7 68 2.8. Traffic Selection Criteria . . . . . . . . . . . . . . . . 7 69 2.9. Limited End-to-End Service Visibility . . . . . . . . . . 7 70 2.10. Per-Service Function (re)Classification . . . . . . . . . 7 71 2.11. Symmetric Traffic Flows . . . . . . . . . . . . . . . . . 8 72 2.12. Multi-vendor Service Functions . . . . . . . . . . . . . . 8 73 3. Service Function Chaining . . . . . . . . . . . . . . . . . . 9 74 3.1. Service Overlay . . . . . . . . . . . . . . . . . . . . . 9 75 3.2. Service Classification . . . . . . . . . . . . . . . . . . 9 76 3.3. Dataplane Metadata . . . . . . . . . . . . . . . . . . . . 9 77 4. Related IETF Work . . . . . . . . . . . . . . . . . . . . . . 11 78 5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 79 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 80 7. Security Considerations . . . . . . . . . . . . . . . . . . . 14 81 8. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 15 82 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 17 83 10. Informative References . . . . . . . . . . . . . . . . . . . . 18 84 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 19 86 1. Introduction 88 The delivery of end-to-end services often require various service 89 functions including traditional network service functions (for 90 example firewalls and server load balancers), as well as application- 91 specific features. Service functions may be delivered within the 92 context of an isolated user group, or shared amongst many users/user 93 groups. 95 Current service function deployment models are relatively static in 96 that they are tightly coupled to network topology and physical 97 resources. The result of that static nature of existing deployments 98 greatly reduces, and in many cases, limits the ability of an operator 99 to introduce new services and/or service functions. Furthermore 100 there is a cascading effect: service changes affect other services. 102 This document outlines the problems encountered with existing service 103 deployment models for Service Function Chaining (SFC) (often referred 104 to simply as service chaining; in this document the terms will be 105 used interchangeably), as well as the problems of service chain 106 creation, deletion, modification/update, policy integration with 107 service chains, and policy enforcement within the network 108 infrastructure. 110 1.1. Definition of Terms 112 Classification: Locally instantiated policy that results in matching 113 of traffic flows for identification of appropriate outbound 114 forwarding actions. 116 Network Overlay: Locally instantiated policy and customer/network/ 117 service profile matching of traffic flows for identification of 118 appropriate outbound forwarding actions. 120 Network Service: An offering provided by an operator that is 121 delivered using one or more service functions. This may also be 122 referred to as a composite service. The term "service" is used to 123 denote a "network service" in the context of this document. 125 Service Function: A function that is responsible for specific 126 treatment of received packets. A Service Function can act at the 127 network layer or other OSI layers. A Service Function can be a 128 virtual instance or be embedded in a physical network element. 129 One of multiple Service Functions can be embedded in the same 130 network element. Multiple instances of the Service Function can 131 be enabled in the same administrative domain. 133 A non-exhaustive list of Service Functions includes: firewalls, 134 WAN and application acceleration, Deep Packet Inspection (DPI), 135 server load balancers, NAT44 [RFC3022], NAT64 [RFC6146], HOST_ID 136 injection [RFC6967], HTTP Header Enrichment functions, TCP 137 optimizer, etc. 139 The generic term "L4-L7 services" is often used to describe many 140 service functions. 142 Service Function Chain (SFC): A service Function chain defines an 143 abstract set of service functions and their ordering constraints 144 that must be applied to packets and/or frames selected as a result 145 of classification. The implied order may not be a linear 146 progression as the architecture allows for nodes that copy to more 147 than one branch, and also allows for cases where there is 148 flexibility in the order in which services need to be applied. 149 The term service chain is often used as shorthand for service 150 function chain. 152 Service Overlay: An overlay network created for the purpose of 153 forwarding data to required service functions. 155 Service Topology: The service overlay connectivity forms a service 156 topology. 158 2. Problem Space 160 The following points describe aspects of existing service deployments 161 that are problematic, and that the Service Function Chaining (SFC) 162 working group aims to address. 164 2.1. Topological Dependencies 166 Network service deployments are often coupled to network topology, 167 whether it be physical or virtualized, or a hybrid of the two. Such 168 dependency imposes constraints on the service delivery, potentially 169 inhibiting the network operator from optimally utilizing service 170 resources, and reduces the flexibility. This limits scale, capacity, 171 and redundancy across network resources. 173 These topologies serve only to "insert" the service function (i.e., 174 ensure that traffic traverses a service function); they are not 175 required from a native packet delivery perspective. For example, 176 firewalls often require an "in" and "out" layer-2 segment and adding 177 a new firewall requires changing the topology (i.e., adding new 178 layer-2 segments). 180 As more service functions are required - often with strict ordering - 181 topology changes are needed before and after each service function 182 resulting in complex network changes and device configuration. In 183 such topologies, all traffic, whether a service function needs to be 184 applied or not, often passes through the same strict order. 186 The topological coupling limits placement and selection of service 187 functions: service functions are "fixed" in place by topology and 188 therefore placement and service function selection taking into 189 account network topology information is not viable. Furthermore, 190 altering the services traversed, or their order, based on flow 191 direction is not possible. 193 A common example is web servers using a server load balancer as the 194 default gateway. When the web service responds to non-load balanced 195 traffic (e.g., administrative or backup operations) all traffic from 196 the server must traverse the load balancer forcing network 197 administrators to create complex routing schemes or create additional 198 interfaces to provide an alternate topology. 200 2.2. Configuration complexity 202 A direct consequence of topological dependencies is the complexity of 203 the entire configuration, specifically in deploying service function 204 chains. Simple actions such as changing the order of the service 205 functions in a service function chain require changes to the 206 topology. Changes to the topology are avoided by the network 207 operator once installed, configured and deployed in production 208 environments for fear of misconfiguration and consequent downtime. 209 All of this leads to very static service delivery deployments. 210 Furthermore, the speed at which these topological changes can be made 211 is not rapid or dynamic enough as it often requires manual 212 intervention, or use of slow provisioning systems. 214 2.3. Constrained High Availability 216 An effect of topological dependency is constrained service function 217 high availability. Worse, when modified, inadvertent non-high 218 availability or downtime can result. 220 Since traffic reaches many service functions based on network 221 topology, alternate, or redundant service functions must be placed in 222 the same topology as the primary service. 224 2.4. Consistent Ordering of Service Functions 226 Service functions are typically independent; service function_1 227 (SF1)...service function_n (SFn) are unrelated and there is no notion 228 at the service layer that SF1 occurs before SF2. However, to an 229 administrator many service functions have a strict ordering that must 230 be in place, yet the administrator has no consistent way to impose 231 and verify the ordering of the service functions that are used to 232 deliver a given service. 234 Service function chains today are most typically built through manual 235 configuration processes. These are slow and error prone. With the 236 advent of newer service deployment models the control and policy 237 planes provide not only connectivity state, but will also be 238 increasingly utilized for the creation of network services. Such 239 control/management planes could be centralized, or be distributed. 241 2.5. Application of Service Policy 243 Service functions rely on topology information such as VLANs or 244 packet (re) classification to determine service policy selection, 245 i.e. the service function specific action taken. Topology 246 information is increasingly less viable due to scaling, tenancy and 247 complexity reasons. The topological information is often stale, 248 providing the operator with inaccurate placement that can result in 249 suboptimal resource utilization. Furthermore topology-centric 250 information often does not convey adequate information to the service 251 functions, forcing functions to individually perform more granular 252 classification. 254 2.6. Transport Dependence 256 Service functions can and will be deployed in networks with a range 257 of transports, including under and overlays. The coupling of service 258 functions to topology requires service functions to support many 259 transport encapsulations or for a transport gateway function to be 260 present. 262 2.7. Elastic Service Delivery 264 Given that the current state of the art for adding/removing service 265 functions largely centers around VLANs and routing changes, rapid 266 changes to the service deployment can be hard to realize due to the 267 risk and complexity of such changes. 269 2.8. Traffic Selection Criteria 271 Traffic selection is coarse, that is, all traffic on a particular 272 segment traverse service functions whether the traffic requires 273 service enforcement or not. This lack of traffic selection is 274 largely due to the topological nature of service deployment since the 275 forwarding topology dictates how (and what) data traverses service 276 function(s). In some deployments, more granular traffic selection is 277 achieved using policy routing or access control filtering. This 278 results in operationally complex configurations and is still 279 relatively inflexible. 281 2.9. Limited End-to-End Service Visibility 283 Troubleshooting service related issues is a complex process that 284 involve both network-specific and service-specific expertise. This 285 is especially the case when service function chains span multiple 286 DCs, or across administrative boundaries. Furthermore, the physical 287 and virtual environments (network and service), can be highly 288 divergent in terms of topology and that topological variance adds to 289 these challenges. 291 2.10. Per-Service Function (re)Classification 293 Classification occurs at each service function independent from 294 previously applied service functions. More importantly, the 295 classification functionality often differs per service function and 296 service functions may not leverage the results from other service 297 functions. 299 2.11. Symmetric Traffic Flows 301 Service function chains may be unidirectional or bidirectional 302 depending on the state requirements of the service functions. In a 303 unidirectional chain traffic is passed through a set of service 304 functions in one forwarding direction only. Bidirectional chains 305 require traffic to be passed through a set of service functions in 306 both forwarding directions. Many common service functions such as 307 DPI and firewall often require bidirectional chaining in order to 308 ensure flow state is consistent. 310 Existing service deployment models provide a static approach to 311 realizing forward and reverse service function chain association most 312 often requiring complex configuration of each network device 313 throughout the SFC. 315 2.12. Multi-vendor Service Functions 317 Deploying service functions from multiple vendors often require per- 318 vendor expertise: insertion models differ, there are limited common 319 attributes and inter-vendor service functions do not share 320 information. 322 3. Service Function Chaining 324 Service Function Chaining aims to address the aforementioned problems 325 associated with service deployment. Concretely, the SFC working 326 group will investigate solutions that address the following elements: 328 3.1. Service Overlay 330 Service function chaining utilizes a service specific overlay that 331 creates the service topology. The service overlay provides service 332 function connectivity and is built "on top" of the existing network 333 topology and allows operators to use whatever overlay or underlay 334 they prefer to create a path between service functions, and to locate 335 service functions in the network as needed. 337 Within the service topology, service functions can be viewed as 338 resources for consumption and an arbitrary topology constructed to 339 connect those resources in a required order. Adding new service 340 functions to the topology is easily accomplished, and no underlying 341 network changes are required. 343 Lastly, the service overlay can provide service specific information 344 needed for troubleshooting service-related issues. 346 3.2. Service Classification 348 Classification is used to select which traffic enters a service 349 overlay. The granularity of the classification varies based on 350 device capabilities, customer requirements, and service offered. 351 Initial classification determines the service function chain required 352 to process the traffic. Subsequent classification can be used within 353 a given service function chain to alter the sequence of service 354 functions applied. Symmetric classification ensures that forward and 355 reverse chains are in place. Similarly, asymmetric -- relative to 356 required service function -- chains can be achieved via service 357 classification. 359 3.3. Dataplane Metadata 361 Data plane metadata provides the ability to exchange information 362 between logical classification points and service functions (and vice 363 versa) and between service functions. As such metadata is not used 364 as forwarding information to deliver packets along the service 365 overlay. 367 Metadata can include the result of antecedent classification and/or 368 information from external sources. Service functions utilize 369 metadata, as required, for localized policy decisions. 371 In addition to sharing of information, the use of metadata addresses 372 several of the issues raised in section 2, most notably the 373 decoupling of policy from the network topology, and the need for per- 374 service function classification (and re-classification). 376 A common approach to service metadata creates a common foundation for 377 interoperability between service functions, regardless of vendor. 379 4. Related IETF Work 381 The following subsections discuss related IETF work and are provided 382 for reference. This section is not exhaustive, rather it provides an 383 overview of the various initiatives and how they relate to network 384 service chaining. 386 1. [L3VPN]: The L3VPN working group is responsible for defining, 387 specifying and extending BGP/MPLS IP VPNs solutions. Although 388 BGP/MPLS IP VPNs can be used as transport for service chaining 389 deployments, the SFC WG focuses on the service specific 390 protocols, not the general case of VPNs. Furthermore, BGP/MPLS 391 IP VPNs do not address the requirements for service chaining. 393 2. [LISP]: LISP provides locator and ID separation. LISP can be 394 used as an L3 overlay to transport service chaining data but does 395 not address the specific service chaining problems highlighted in 396 this document. 398 3. [NVO3]: The NVO3 working group is chartered with creation of 399 problem statement and requirements documents for multi-tenant 400 network overlays. NVO3 WG does not address service chaining 401 protocols. 403 4. [ALTO]: The Application Layer Traffic Optimization Working Group 404 is chartered to provide topological information at a higher 405 abstraction layer, which can be based upon network policy, and 406 with application-relevant service functions located in it. The 407 mechanism for ALTO obtaining the topology can vary and policy can 408 apply to what is provided or abstracted. This work could be 409 leveraged and extended to address the need for services 410 discovery. 412 5. [I2RS]: The Interface to the Routing System Working Group is 413 chartered to investigate the rapid programming of a device's 414 routing system, as well as the service of a generalized, multi- 415 layered network topology. This work could be leveraged and 416 extended to address some of the needs for service chaining in the 417 topology and device programming areas. 419 6. [ForCES]: The ForCES working group has created a framework, 420 requirements, a solution protocol, a logical function block 421 library, and other associated documents in support of Forwarding 422 and Control Element Separation. The work done by ForCES may 423 provide a basis for both the separation of SFC elements, as well 424 as provide protocol and design guidance for those elements. 426 5. Summary 428 This document highlights problems associated with network service 429 deployment today and identifies several key areas that will be 430 addressed by the SFC working group. Furthermore, this document 431 identifies three components that are the basis for service function 432 chaining. These components will form the areas of focus for the 433 working group. 435 6. IANA Considerations 437 This document makes no request to IANA. 439 7. Security Considerations 441 Security considerations are not addressed in this problem statement 442 only document. Given the scope of service chaining, and the 443 implications on data and control planes, security considerations are 444 clearly important and will be addressed in the specific protocol and 445 deployment documents created by the SFC WG. 447 8. Contributors 449 The following people are active contributors to this document and 450 have provided review, content and concepts (listed alphabetically by 451 surname): 453 Puneet Agarwal 454 Broadcom 455 Email: pagarwal@broadcom.com 457 Mohamed Boucadair 458 France Telecom 459 Email: mohamed.boucadair@orange.com 461 Abhishek Chauhan 462 Citrix 463 Email: Abhishek.Chauhan@citrix.com 465 Uri Elzur 466 Intel 467 Email: uri.elzur@intel.com 469 Kevin Glavin 470 Riverbed 471 Email: Kevin.Glavin@riverbed.com 473 Ken Gray 474 Cisco Systems 475 Email: kegray@cisco.com 477 Jim Guichard 478 Cisco Systems 479 Email:jguichar@cisco.com 481 Christian Jacquenet 482 France Telecom 483 Email: christian.jacquenet@orange.com 485 Surendra Kumar 486 Cisco Systems 487 Email: smkumar@cisco.com 489 Nic Leymann 490 Deutsche Telekom 491 Email: n.leymann@telekom.de 493 Darrel Lewis 494 Cisco Systems 495 Email: darlewis@cisco.com 497 Rajeev Manur 498 Broadcom 499 Email:rmanur@broadcom.com 501 Brad McConnell 502 Rackspace 503 Email: bmcconne@rackspace.com 505 Carlos Pignataro 506 Cisco Systems 507 Email: cpignata@cisco.com 509 Michael Smith 510 Cisco Systems 511 Email: michsmit@cisco.com 513 Navindra Yadav 514 Cisco Systems 515 Email: nyadav@cisco.com 517 9. Acknowledgments 519 The authors would like to thank David Ward, Rex Fernando, David 520 Mcdysan, Jamal Hadi Salim, Charles Perkins, Andre Beliveau, Joel 521 Halpern and Jim French for their reviews and comments. 523 10. Informative References 525 [ALTO] "Application-Layer Traffic Optimization (alto)", 526 . 528 [ForCES] "Forwarding and Control Element Separation (forces)", 529 . 531 [I2RS] "Interface to the Routing System (i2rs)", 532 . 534 [L3VPN] "Layer 3 Virtual Private Networks (l3vpn)", 535 . 537 [LISP] "Locator/ID Separation Protocol (lisp)", 538 . 540 [NVO3] "Network Virtualization Overlays (nvo3)", 541 . 543 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 544 Address Translator (Traditional NAT)", RFC 3022, 545 January 2001. 547 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 548 NAT64: Network Address and Protocol Translation from IPv6 549 Clients to IPv4 Servers", RFC 6146, April 2011. 551 [RFC6967] Boucadair, M., Touch, J., Levis, P., and R. Penno, 552 "Analysis of Potential Solutions for Revealing a Host 553 Identifier (HOST_ID) in Shared Address Deployments", 554 RFC 6967, June 2013. 556 Authors' Addresses 558 Paul Quinn (editor) 559 Cisco Systems, Inc. 561 Email: paulq@cisco.com 563 Thomas Nadeau (editor) 564 Brocade 566 Email: tnadeau@lucidvision.com