idnits 2.17.1 draft-irtf-panrg-questions-06.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 : ---------------------------------------------------------------------------- ** The document seems to lack a Security Considerations section. ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (13 August 2020) is 1345 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- No issues found here. Summary: 2 errors (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Path Aware Networking RG B. Trammell 3 Internet-Draft Google Switzerland GmbH 4 Intended status: Informational 13 August 2020 5 Expires: 14 February 2021 7 Current Open Questions in Path Aware Networking 8 draft-irtf-panrg-questions-06 10 Abstract 12 In contrast to the present Internet architecture, a path-aware 13 internetworking architecture has two important properties: it exposes 14 the properties of available Internet paths to endpoints, and provides 15 for endpoints and applications to use these properties to select 16 paths through the Internet for their traffic. This document poses 17 questions in path-aware networking open as of 2019, that must be 18 answered in the design, development, and deployment of path-aware 19 internetworks. It was originally written to frame discussions in the 20 Path Aware Networking proposed Research Group (PANRG), and has been 21 published to snapshot current thinking in this space. 23 Status of This Memo 25 This Internet-Draft is submitted in full conformance with the 26 provisions of BCP 78 and BCP 79. 28 Internet-Drafts are working documents of the Internet Engineering 29 Task Force (IETF). Note that other groups may also distribute 30 working documents as Internet-Drafts. The list of current Internet- 31 Drafts is at https://datatracker.ietf.org/drafts/current/. 33 Internet-Drafts are draft documents valid for a maximum of six months 34 and may be updated, replaced, or obsoleted by other documents at any 35 time. It is inappropriate to use Internet-Drafts as reference 36 material or to cite them other than as "work in progress." 38 This Internet-Draft will expire on 14 February 2021. 40 Copyright Notice 42 Copyright (c) 2020 IETF Trust and the persons identified as the 43 document authors. All rights reserved. 45 This document is subject to BCP 78 and the IETF Trust's Legal 46 Provisions Relating to IETF Documents (https://trustee.ietf.org/ 47 license-info) in effect on the date of publication of this document. 48 Please review these documents carefully, as they describe your rights 49 and restrictions with respect to this document. Code Components 50 extracted from this document must include Simplified BSD License text 51 as described in Section 4.e of the Trust Legal Provisions and are 52 provided without warranty as described in the Simplified BSD License. 54 Table of Contents 56 1. Introduction to Path-Aware Networking . . . . . . . . . . . . 2 57 1.1. Definition . . . . . . . . . . . . . . . . . . . . . . . 3 58 2. Questions . . . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2.1. A Vocabulary of Path Properties . . . . . . . . . . . . . 4 60 2.2. Discovery, Distribution, and Trustworthiness of Path 61 Properties . . . . . . . . . . . . . . . . . . . . . . . 4 62 2.3. Supporting Path Selection . . . . . . . . . . . . . . . . 5 63 2.4. Interfaces for Path Awareness . . . . . . . . . . . . . . 5 64 2.5. Implications of Path Awareness for the Data Plane . . . . 6 65 2.6. What is an Endpoint? . . . . . . . . . . . . . . . . . . 6 66 2.7. Operating a Path Aware Network . . . . . . . . . . . . . 7 67 2.8. Deploying a Path Aware Network . . . . . . . . . . . . . 7 68 3. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 8 69 4. Informative References . . . . . . . . . . . . . . . . . . . 8 70 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 9 72 1. Introduction to Path-Aware Networking 74 In the current Internet architecture, the network layer provides an 75 unverifiable, best-effort service: an application can assume that a 76 packet with a given destination address will eventually be forwarded 77 toward that destination, but little else. A transport layer protocol 78 such as TCP can provide reliability over this best-effort service, 79 and a protocol above the network layer such as IPsec AH [RFC4302] or 80 TLS [RFC8446] can authenticate the remote endpoint. However, little, 81 if any, explicit information about the path is available, and 82 assumptions about that path often do not hold, sometimes with serious 83 impacts on the application, as in the case with BGP hijacking 84 attacks. 86 By contrast, in a path-aware internetworking architecture, endpoints 87 have the ability to select or influence the path through the network 88 used by any given packet, and the network and transport layers 89 explicitly expose information about the path or paths available from 90 one endpoint to another, and vice versa, to those endpoints and the 91 applications running on them, so that they can make this selection. 93 Path selection provides transparency and control to applications and 94 users of the network. Selection may be made at either the 95 application layer or the transport layer. Path control at the packet 96 level enables the design of new transport protocols that can leverage 97 multipath connectivity across disjoint paths through the Internet, 98 even over a single physical interface. When exposed to applications, 99 or to end-users through a system configuration interface, path 100 control allows the specification of constraints on the paths that 101 traffic should traverse, for instance to confound passive 102 surveillance in the network core [RFC7624]. 104 We note that this property of "path awareness" already exists in many 105 Internet-connected networks within single domains. Indeed, much of 106 the practice of network engineering using encapsulation at layer 3 107 can be said to be "path aware", in that it explicitly assigns traffic 108 at tunnel endpoints to a given path within the network. Path-aware 109 internetworking seeks to extend this awareness across domain 110 boundaries without resorting to overlays, except as a transition 111 technology. 113 1.1. Definition 115 For purposes of this document, "path aware networking" describes 116 endpoint discovery of the properties of paths they use for 117 communication, and endpoint reaction to these properties that affects 118 routing and/or transmission; note that this can and already does 119 happen to some extent in the current Internet architecture. 120 Expanding on this definition, a "path aware internetwork" is one in 121 which endpoint discovery of path properties and endpoint selection of 122 paths used by traffic exchanged by the endpoint are explicitly 123 supported, regardless of the specific design of the protocol features 124 which enable this discovery and selection. 126 Research into path aware networking covers any and all aspects of 127 designing, building, and operating path aware internetworks or the 128 networks and endpoints attached to them. This document presents a 129 collection of research questions to address in order to make a path 130 aware Internet a reality. 132 2. Questions 134 Realizing path-aware networking requires answers to a set of open 135 research questions. This document poses these questions, as a 136 starting point for discussions about how to realize path awareness in 137 the Internet, and to direct future research efforts within the Path 138 Aware Networking Research Group. 140 2.1. A Vocabulary of Path Properties 142 In order for information about paths to be exposed to an endpoint, 143 and for the endpoint to make use of that information, it is necessary 144 to define a common vocabulary for paths through an internetwork, and 145 properties of those paths. The elements of this vocabulary could 146 include terminology for components of a path and properties defined 147 for these components, for the entire path, or for subpaths of a path. 148 These properties may be relatively static, such as the presence of a 149 given node or service function on the path; as well as relatively 150 dynamic, such as the current values of metrics such as loss and 151 latency. 153 This vocabulary must be defined carefully, as its design will have 154 impacts on the expressiveness of a given path-aware internetworking 155 architecture. This expressiveness also exhibits tradeoffs. For 156 example, a system that exposes node-level information for the 157 topology through each network would maximize information about the 158 individual components of the path at the endpoints, at the expense of 159 making internal network topology universally public, which may be in 160 conflict with the business goals of each network's operator. 161 Furthermore, properties related to individual components of the path 162 may change frequently and may quickly become outdated. However, 163 aggregating the properties of individual components to distill end- 164 to-end properties for the entire path is not trivial. 166 The first question: how are paths and path properties defined and 167 represented? 169 2.2. Discovery, Distribution, and Trustworthiness of Path Properties 171 Once endpoints and networks have a shared vocabulary for expressing 172 path properties, the network must have some method for distributing 173 those path properties to the endpoint. Regardless of how path 174 property information is distributed to the endpoints, the endpoints 175 require a method to authenticate the properties - to determine that 176 they originated from and pertain to the path that they purport to. 178 Choices in distribution and authentication methods will have impacts 179 on the scalability of a path-aware architecture. Possible dimensions 180 in the space of distribution methods include in-band versus out-of- 181 band, push versus pull versus publish-subscribe, and so on. There 182 are temporal issues with path property dissemination as well, 183 especially with dynamic properties, since the measurement or 184 elicitation of dynamic properties may be outdated by the time that 185 information is available at the endpoints, and interactions between 186 the measurement and dissemination delay may exhibit pathological 187 behavior for unlucky points in the parameter space. 189 The second question: how do endpoints and applications get access to 190 trustworthy path properties? 192 2.3. Supporting Path Selection 194 Access to trustworthy path properties is only half of the challenge 195 in establishing a path-aware architecture. Endpoints must be able to 196 use this information in order to select paths for specific traffic 197 they send. As with the dissemination of path properties, choices 198 made in path selection methods will also have an impact on the 199 tradeoff between scalability and expressiveness of a path-aware 200 architecture. One key choice here is between in-band and out-of-band 201 control of path selection. Another is granularity of path selection 202 (whether per packet, per flow, or per larger aggregate), which also 203 has a large impact on the scalabilty/expressiveness tradeoff. Path 204 selection must, like path property information, be trustworthy, such 205 that the result of a path selection at an endpoint is predictable. 206 Moreover, any path selection mechanism should aim to provide an 207 outcome that is not worse than using a single path, or selecting 208 paths at random. 210 The third question: how can endpoints select paths to use for traffic 211 in a way that can be trusted by the network, the endpoints, and the 212 applications using them? 214 2.4. Interfaces for Path Awareness 216 In order for applications to make effective use of a path-aware 217 networking architecture, the control interfaces presented by the 218 network and transport layers must also expose path properties to the 219 application in a useful way, and provide a useful set of paths among 220 which the application can select. Path selection must be possible 221 based not only on the preferences and policies of the application 222 developer, but of end-users as well. Also, the path selection 223 interfaces presented to applications and end users will need to 224 support multiple levels of granularity. Most applications' 225 requirements can be satisfied with the expression path selection 226 policies in terms of properties of the paths, while some applications 227 may need finer-grained, per-path control. These interfaces will need 228 to support incremental development and deployment of applications, 229 and provide sensible defaults, to avoid hindering their adoption. 231 The fourth question: how can interfaces to the transport and 232 application layers support the use of path awareness? 234 2.5. Implications of Path Awareness for the Data Plane 236 In the current Internet, the basic assumption that at a given time 237 all traffic for a given flow will receive the same network treatment 238 and traverse the same path or equivalend paths often holds. In a 239 path aware network, this assumption is more easily violated holds. 240 The weakening of this assumption has implications for the design of 241 protocols above any path-aware network layer. 243 For example, one advantage of multipath communication is that a given 244 end-to-end flow can be "sprayed" along multiple paths in order to 245 confound attempts to collect data or metadata from those flows for 246 pervasive surveillance purposes [RFC7624]. However, the benefits of 247 this approach are reduced if the upper-layer protocols use linkable 248 identifiers on packets belonging to the same flow across different 249 paths. Clients may mitigate linkability by opting to not re-use 250 cleartext connection identifiers, such as TLS session IDs or tickets, 251 on separate paths. The privacy-conscious strategies required for 252 effective privacy in a path-aware Internet are only possible if 253 higher-layer protocols such as TLS permit clients to obtain 254 unlinkable identifiers. 256 The fifth question: how should transport-layer and higher layer 257 protocols be redesigned to work most effectively over a path-aware 258 networking layer? 260 2.6. What is an Endpoint? 262 The vision of path-aware networking articulated so far makes an 263 assumption that path properties will be disseminated to endpoints on 264 which applications are running (terminals with user agents, servers, 265 and so on). However, incremental deployment may require that a path- 266 aware network "core" be used to interconnect islands of legacy 267 protocol networks. In these cases, it is the gateways, not the 268 application endpoints, that receive path properties and make path 269 selections for that traffic. The interfaces provided by this gateway 270 are necessarily different than those a path-aware networking layer 271 provides to its transport and application layers, and the path 272 property information the gateway needs and makes available over those 273 interfaces may also be different. 275 The sixth question: how is path awareness (in terms of vocabulary and 276 interfaces) different when applied to tunnel and overlay endpoints? 278 2.7. Operating a Path Aware Network 280 The network operations model in the current Internet architecture 281 assumes that traffic flows are controlled by the decisions and 282 policies made by network operators, as expressed in interdomain and 283 intradomain routing protocols. In a network providing path selection 284 to the endpoints, however, this assumption no longer holds, as 285 endpoints may react to path properties by selecting alternate paths. 286 Competing control inputs from path-aware endpoints and the routing 287 control plane may lead to more difficult traffic engineering or 288 nonconvergent forwarding, especially if the endpoints' and operators' 289 notion of the "best" path for given traffic diverges significantly. 290 The degree of difficulty may depend on the fidelity of information 291 made available to path selection algorithms at the endpoints. 292 Explicit path selection can also specify outbound paths, while BGP 293 policies are expressed in terms of inbound traffic. 295 A concept for path aware network operations will need to have clear 296 methods for the resolution of apparent (if not actual) conflicts of 297 intent between the network's operator and the path selection at an 298 endpoint. It will also need set of safety principles to ensure that 299 increasing path control does not lead to decreasing connectivity; one 300 such safety principle could be "the existence of at least one path 301 between two endpoints guarantees the selection of at least one path 302 between those endpoints." 304 The seventh question: how can a path aware network in a path aware 305 internetwork be effectively operated, given control inputs from the 306 network administrator as well as from the endpoints? 308 2.8. Deploying a Path Aware Network 310 The vision presented in the introduction discusses path aware 311 networking from the point of view of the benefits accruing at the 312 endpoints, to designers of transport protocols and applications as 313 well as to the end users of those applications. However, this vision 314 requires action not only at the endpoints but within the 315 interconnected networks offering path aware connectivity. While the 316 specific actions required are a matter of the design and 317 implementation of a specific realization of a path aware protocol 318 stack, it is clear than any path aware architecture will require 319 network operators to give up some control of their networks over to 320 endpoint-driven control inputs. 322 Here the question of apparent versus actual conflicts of intent 323 arises again: certain network operations requirements may appear 324 essential, but are merely accidents of the interfaces provided by 325 current routing and management protocols. Incentives for deployment 326 must show how existing network operations requirements are met 327 through new path selection and property dissemination mechanisms. 329 The incentives for network operators and equipment vendors need to be 330 made clear, in terms of a plan to transition [RFC8170] an 331 internetwork to path-aware operation, one network and facility at a 332 time. This plan to transition must also take into account that the 333 dynamics of path aware networking early in this transition (when few 334 endpoints and flows in the Internet use path selection) may be 335 different than those later in the transition. 337 The eighth question: how can the incentives of network operators and 338 end-users be aligned to realize the vision of path aware networking, 339 and how can the transition from current ("path-oblivious") to path- 340 aware networking be managed? 342 3. Acknowledgments 344 Many thanks to Adrian Perrig, Jean-Pierre Smith, Mirja Kuehlewind, 345 Olivier Bonaventure, Martin Thomson, Shwetha Bhandari, Chris Wood, 346 Lee Howard, Mohamed Boucadair, Thorben Krueger, Gorry Fairhurst, 347 Spencer Dawkins, and Theresa Enghardt for discussions leading to 348 questions in this document, and for feedback on the document itself. 350 This work is partially supported by the European Commission under 351 Horizon 2020 grant agreement no. 688421 Measurement and Architecture 352 for a Middleboxed Internet (MAMI), and by the Swiss State Secretariat 353 for Education, Research, and Innovation under contract no. 15.0268. 354 This support does not imply endorsement. 356 4. Informative References 358 [RFC4302] Kent, S., "IP Authentication Header", RFC 4302, 359 DOI 10.17487/RFC4302, December 2005, 360 . 362 [RFC7624] Barnes, R., Schneier, B., Jennings, C., Hardie, T., 363 Trammell, B., Huitema, C., and D. Borkmann, 364 "Confidentiality in the Face of Pervasive Surveillance: A 365 Threat Model and Problem Statement", RFC 7624, 366 DOI 10.17487/RFC7624, August 2015, 367 . 369 [RFC8170] Thaler, D., Ed., "Planning for Protocol Adoption and 370 Subsequent Transitions", RFC 8170, DOI 10.17487/RFC8170, 371 May 2017, . 373 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 374 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 375 . 377 Author's Address 379 Brian Trammell 380 Google Switzerland GmbH 381 Gustav-Gull-Platz 1 382 CH- 8004 Zurich 383 Switzerland 385 Email: ietf@trammell.ch