idnits 2.17.1 draft-irtf-panrg-questions-03.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 (October 15, 2019) is 1647 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446) Summary: 3 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 4 Intended status: Informational October 15, 2019 5 Expires: April 17, 2020 7 Current Open Questions in Path Aware Networking 8 draft-irtf-panrg-questions-03 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 intetnetworks. 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 April 17, 2020. 40 Copyright Notice 42 Copyright (c) 2019 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 47 (https://trustee.ietf.org/license-info) in effect on the date of 48 publication of this document. Please review these documents 49 carefully, as they describe your rights and restrictions with respect 50 to this document. Code Components extracted from this document must 51 include Simplified BSD License text as described in Section 4.e of 52 the Trust Legal Provisions and are provided without warranty as 53 described in the Simplified BSD License. 55 Table of Contents 57 1. Introduction to Path-Aware Networking . . . . . . . . . . . . 2 58 2. Questions . . . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2.1. A Vocabulary of Path Properties . . . . . . . . . . . . . 3 60 2.2. Discovery, Distribution, and Trustworthiness of Path 61 Properties . . . . . . . . . . . . . . . . . . . . . . . 4 62 2.3. Supporting Path Selection . . . . . . . . . . . . . . . . 4 63 2.4. Interfaces for Path Awareness . . . . . . . . . . . . . . 4 64 2.5. Implications of Path Awareness for the Data Plane . . . . 5 65 2.6. What is an Endpoint? . . . . . . . . . . . . . . . . . . 5 66 2.7. Operating a Path Aware Network . . . . . . . . . . . . . 6 67 2.8. Deploying a Path Aware Network . . . . . . . . . . . . . 6 68 3. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 7 69 4. References . . . . . . . . . . . . . . . . . . . . . . . . . 7 70 4.1. Normative References . . . . . . . . . . . . . . . . . . 7 71 4.2. Informative References . . . . . . . . . . . . . . . . . 8 72 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 8 74 1. Introduction to Path-Aware Networking 76 In the current Internet architecture, the interdomain network layer 77 provides an unverifiable, best-effort service: an application can 78 assume that a packet with a given destination address will eventually 79 be forwarded toward that destination, but little else. A transport 80 layer protocol such as TCP can provide reliability over this best- 81 effort service, and a protocol above the network layer such as IPsec 82 AH [RFC4302] or TLS [RFC5246] can authenticate the remote endpoint. 83 However, no explicit information about the path is available, and 84 assumptions about that path sometimes do not hold, sometimes with 85 serious impacts on the application, as in the case with BGP hijacking 86 attacks. 88 By contrast, in a path-aware internetworking architecture, endpoints 89 have the ability to select or influence the path through the network 90 used by any given packet, and the network and transport layers 91 explicitly expose information about the path or paths available 92 between two endpoints to those endpoints and the applications running 93 on them, so that they can make this selection. 95 Path selection provides transparency and control to applications and 96 users of the network. Selection may be made at either the 97 application layer or the transport layer. Path control at the packet 98 level enables the design of new transport protocols that can leverage 99 multipath connectivity across maximally-disjoint paths through the 100 Internet, even over a single physical interface. When exposed to 101 applications, or to end-users through a system configuration 102 interface, path control allows the specification of constraints on 103 the paths that traffic should traverse, for instance to confound 104 passive surveillance in the network core. 106 We note that this property of "path awareness" already exists in many 107 Internet-connected networks in an intradomain context. Indeed, much 108 of the practice of network engineering using encapsulation at layer 3 109 can be said to be "path aware", in that it explicitly assigns traffic 110 at tunnel endpoints to a given path within the network. Path-aware 111 internetworking seeks to extend this awareness across domain 112 boundaries without resorting to overlays, except as a transition 113 technology. 115 2. Questions 117 Realizing path-aware networking requires answers to a set of open 118 research questions. This document poses these questions, as a 119 starting point for discussions about how to realize path awareness in 120 the Internet, and to direct future research efforts within the Path 121 Aware Networking Research Group. 123 2.1. A Vocabulary of Path Properties 125 In order for information about paths to be exposed to an endpoint, 126 and for the endpoint to make use of that information, it is necessary 127 to define a common vocabulary for path properties. The elements of 128 this vocabulary could include relatively static properties, such as 129 the presence of a given node or service function on the path; as well 130 as relatively dynamic properties, such as the current values of 131 metrics such as loss and latency. 133 This vocabulary must be defined carefully, as its design will have 134 impacts on the expressiveness of a given path-aware internetworking 135 architecture. This expressiveness also exhibits tradeoffs. For 136 example, a system that exposes node-level information for the 137 topology through each network would maximize information about the 138 individual components of the path at the endpoints at the expense of 139 making internal network topology universally public, which may be in 140 conflict with the business goals of each network's operator. 142 The first question: how are path properties defined and represented? 144 2.2. Discovery, Distribution, and Trustworthiness of Path Properties 146 Once endpoints and networks have a shared vocabulary for expressing 147 path properties, the network must have some method for distributing 148 those path properties to the endpoint. Regardless of how path 149 property information is distributed to the endpoints, the endpoints 150 require a method to authenticate the properties - to determine that 151 they originated from and pertain to the path that they purport to. 153 Choices in distribution and authentication methods will have impacts 154 on the scalability of a path-aware architecture. Possible dimensions 155 in the space of distribution methods include in-band versus out-of- 156 band, push versus pull versus publish-subscribe, and so on. There 157 are temporal issues with path property dissemination as well, 158 especially with dynamic properties, since the measurement or 159 elicitation of dynamic properties may be outdated by the time that 160 information is available at the endpoints, and interactions between 161 the measurement and dissemination delay may exhibit pathological 162 behavior for unlucky points in the parameter space. 164 The second question: how do endpoints get access to trustworthy path 165 properties? 167 2.3. Supporting Path Selection 169 Access to trustworthy path properties is only half of the challenge 170 in establishing a path-aware architecture. Endpoints must be able to 171 use this information in order to select paths for traffic they send. 172 As with the dissemination of path properties, choices made in path 173 selection methods will also have an impact on the tradeoff between 174 scalability and expressiveness of a path-aware architecture. One key 175 choice here is between in-band and out-of-band control of path 176 selection. Another is granularity of path selection (whether per 177 packet, per flow, or per larger aggregate), which also has a large 178 impact on the scalabilty/expressiveness tradeoff. Path selection 179 must, like path property information, be trustworthy, such that the 180 result of a path selection at an endpoint is predictable. 182 The third question: how can endpoints select paths to use for traffic 183 in a way that can be trusted by both the network and the endpoints? 185 2.4. Interfaces for Path Awareness 187 In order for applications to make effective use of a path-aware 188 networking architecture, the control interfaces presented by the 189 network and transport layers must also expose path properties to the 190 application in a useful way, and provide a useful set of paths among 191 which the application can select. Path selection must be possible 192 based not only on the preferences and policies of the application 193 developer, but of end-users as well. Also, the path selection 194 interfaces presented to applications and end users will need to 195 support multiple levels of granularity. Most applications' 196 requirements can be satisfied with the expression path selection 197 policies in terms of properties of the paths, while some applications 198 may need finer-grained, per-path control. 200 The fourth question: how can interfaces to the transport and 201 application layers support the use of path awareness? 203 2.5. Implications of Path Awareness for the Data Plane 205 In the current Internet, the basic assumption that at a given time t 206 all traffic for a given flow will traverse a single path, for some 207 definition of path, generally holds. In a path aware network, this 208 assumption no longer holds. The absence of this assumption has 209 implications for the design of protocols above any path-aware network 210 layer. 212 For example, one advantage of multipath communication is that a given 213 end-to-end flow can be "sprayed" along multiple paths in order to 214 confound attempts to collect data or metadata from those flows for 215 pervasive surveillance purposes [RFC7624]. However, the benefits of 216 this approach are reduced if the upper-layer protocols use linkable 217 identifiers on packets belonging to the same flow across different 218 paths. Clients may mitigate linkability by opting to not re-use 219 cleartext connection identifiers, such as TLS session IDs or tickets, 220 on separate paths. The privacy-conscious strategies required for 221 effective privacy in a path-aware Internet are only possible if 222 higher-layer protocols such as TLS permit clients to obtain 223 unlinkable identifiers. 225 The fifth question: how should transport-layer and higher layer 226 protocols be redesigned to work most effectively over a path-aware 227 networking layer? 229 2.6. What is an Endpoint? 231 The vision of path-aware networking articulated so far makes an 232 assumption that path properties will be disseminated to endpoints on 233 which applications are running (terminals with user agents, servers, 234 and so on). However, incremental deployment may require that a path- 235 aware network "core" be used to interconnect islands of legacy 236 protocol networks. In these cases, it is the gateways, not the 237 application endpoints, that receive path properties and make path 238 selections for that traffic. The interfaces provided by this gateway 239 are necessarily different than those a path-aware networking layer 240 provides to its transport and application layers, and the path 241 property information the gateway needs and makes available over those 242 interfaces may also be different. 244 The sixth question: how is path awareness (in terms of vocabulary and 245 interfaces) different when applied to tunnel and overlay endpoints? 247 2.7. Operating a Path Aware Network 249 The network operations model in the current Internet architecture 250 assumes that traffic flows are controlled by the decisions and 251 policies made by network operators, as expressed in interdomain 252 routing protocols. In a network providing path selection to the 253 endpoints, however, this assumption no longer holds, as endpoints may 254 react to path properties by selecting alternate paths. Competing 255 control inputs from path-aware endpoints and the interdomain routing 256 control plane may lead to more difficult traffic engineering or 257 nonconvergent forwarding, especially if the endpoints' and operators' 258 notion of the "best" path for given traffic diverges significantly. 260 A concept for path aware network operations will need to have clear 261 methods for the resolution of apparent (if not actual) conflicts of 262 intent between the network's operator and the path selection at an 263 endpoint. It will also need set of safety principles to ensure that 264 increasing path control does not lead to decreasing connectivity; one 265 such safety principle could be "the existence of at least one path 266 between two endpoints guarantees the selection of at least one path 267 between those endpoints." 269 The seventh question: how can a path aware network in a path aware 270 internetwork be effectively operated, given control inputs from the 271 network administrator as well as from the endpoints? 273 2.8. Deploying a Path Aware Network 275 The vision presented in the introduction discusses path aware 276 networking from the point of view of the benefits accruing at the 277 endpoints, to designers of transport protocols and applications as 278 well as to the end users of those applications. However, this vision 279 requires action not only at the endpoints but within the 280 interconnected networks offering path aware connectivity. While the 281 specific actions required are a matter of the design and 282 implementation of a specific realization of a path aware protocol 283 stack, it is clear than any path aware architecture will require 284 network operators to give up some control of their networks over to 285 endpoint-driven control inputs. 287 Here the question of apparent versus actual conflicts of intent 288 arises again: certain network operations requirements may appear 289 essential, but are merely accidents of the interfaces provided by 290 current routing and management protocols. Incentives for deployment 291 must show how existing network operations requirements are met 292 through new path selection and property dissemination mechanisms. 294 The incentives for network operators and equipment vendors need to be 295 made clear, in terms of a plan to transition [RFC8170] an 296 internetwork to path-aware operation, one network and facility at a 297 time. This plan to transition must also take into account that the 298 dynamics of path aware networking early in this transition (when few 299 endpoints and flows in the Internet use path selection) may be 300 different than those later in the transition. 302 The eighth question: how can the incentives of network operators and 303 end-users be aligned to realize the vision of path aware networking, 304 and how can the transition from current ("path-oblivious") to path- 305 aware networking be managed? 307 3. Acknowledgments 309 Many thanks to Adrian Perrig, Jean-Pierre Smith, Mirja Kuehlewind, 310 Olivier Bonaventure, Martin Thomson, Shwetha Bhandari, Chris Wood, 311 Lee Howard, Mohamed Boucadair, and Thorben Krueger for discussions 312 leading to questions in this document, and for feedback on the 313 document itself. 315 This work is partially supported by the European Commission under 316 Horizon 2020 grant agreement no. 688421 Measurement and Architecture 317 for a Middleboxed Internet (MAMI), and by the Swiss State Secretariat 318 for Education, Research, and Innovation under contract no. 15.0268. 319 This support does not imply endorsement. 321 4. References 323 4.1. Normative References 325 [RFC4302] Kent, S., "IP Authentication Header", RFC 4302, 326 DOI 10.17487/RFC4302, December 2005, 327 . 329 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 330 (TLS) Protocol Version 1.2", RFC 5246, 331 DOI 10.17487/RFC5246, August 2008, 332 . 334 4.2. Informative References 336 [RFC7624] Barnes, R., Schneier, B., Jennings, C., Hardie, T., 337 Trammell, B., Huitema, C., and D. Borkmann, 338 "Confidentiality in the Face of Pervasive Surveillance: A 339 Threat Model and Problem Statement", RFC 7624, 340 DOI 10.17487/RFC7624, August 2015, 341 . 343 [RFC8170] Thaler, D., Ed., "Planning for Protocol Adoption and 344 Subsequent Transitions", RFC 8170, DOI 10.17487/RFC8170, 345 May 2017, . 347 Author's Address 349 Brian Trammell 350 Google 351 Gustav-Gull-Platz 1 352 8004 Zurich 353 Switzerland 355 Email: ietf@trammell.ch