idnits 2.17.1 draft-ietf-anima-grasp-14.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 : ---------------------------------------------------------------------------- ** There are 2 instances of too long lines in the document, the longest one being 5 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 2, 2017) is 2483 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) -- Looks like a reference, but probably isn't: '7' on line 3313 -- Looks like a reference, but probably isn't: '13767778' on line 3313 -- Looks like a reference, but probably isn't: '34965' on line 3313 == Outdated reference: A later version (-11) exists of draft-greevenbosch-appsawg-cbor-cddl-10 ** Downref: Normative reference to an Informational draft: draft-greevenbosch-appsawg-cbor-cddl (ref. 'I-D.greevenbosch-appsawg-cbor-cddl') == Outdated reference: A later version (-30) exists of draft-ietf-anima-autonomic-control-plane-06 ** Obsolete normative reference: RFC 7049 (Obsoleted by RFC 8949) == Outdated reference: A later version (-09) exists of draft-carpenter-anima-asa-guidelines-01 == Outdated reference: A later version (-45) exists of draft-ietf-anima-bootstrapping-keyinfra-06 == Outdated reference: A later version (-10) exists of draft-ietf-anima-reference-model-03 == Outdated reference: A later version (-10) exists of draft-ietf-anima-stable-connectivity-02 == Outdated reference: A later version (-06) exists of draft-liu-anima-grasp-api-04 -- Obsolete informational reference (is this intentional?): RFC 3315 (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 7564 (Obsoleted by RFC 8264) Summary: 3 errors (**), 0 flaws (~~), 8 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group C. Bormann 3 Internet-Draft Universitaet Bremen TZI 4 Intended status: Standards Track B. Carpenter, Ed. 5 Expires: January 3, 2018 Univ. of Auckland 6 B. Liu, Ed. 7 Huawei Technologies Co., Ltd 8 July 2, 2017 10 A Generic Autonomic Signaling Protocol (GRASP) 11 draft-ietf-anima-grasp-14 13 Abstract 15 This document specifies the GeneRic Autonomic Signaling Protocol 16 (GRASP), which enables autonomic nodes and autonomic service agents 17 to dynamically discover peers, to synchronize state with each other, 18 and to negotiate parameter settings with each other. GRASP depends 19 on an external security environment that is described elsewhere. The 20 technical objectives and parameters for specific application 21 scenarios are to be described in separate documents. Appendices 22 briefly discuss requirements for the protocol and existing protocols 23 with comparable features. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on January 3, 2018. 42 Copyright Notice 44 Copyright (c) 2017 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 60 2. GRASP Protocol Overview . . . . . . . . . . . . . . . . . . . 5 61 2.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 62 2.2. High Level Deployment Model . . . . . . . . . . . . . . . 7 63 2.3. High Level Design . . . . . . . . . . . . . . . . . . . . 8 64 2.4. Quick Operating Overview . . . . . . . . . . . . . . . . 11 65 2.5. GRASP Protocol Basic Properties and Mechanisms . . . . . 12 66 2.5.1. Required External Security Mechanism . . . . . . . . 12 67 2.5.2. Discovery Unsolicited Link-Local (DULL) GRASP . . . . 13 68 2.5.3. Transport Layer Usage . . . . . . . . . . . . . . . . 14 69 2.5.4. Discovery Mechanism and Procedures . . . . . . . . . 14 70 2.5.5. Negotiation Procedures . . . . . . . . . . . . . . . 18 71 2.5.6. Synchronization and Flooding Procedures . . . . . . . 20 72 2.6. GRASP Constants . . . . . . . . . . . . . . . . . . . . . 22 73 2.7. Session Identifier (Session ID) . . . . . . . . . . . . . 23 74 2.8. GRASP Messages . . . . . . . . . . . . . . . . . . . . . 23 75 2.8.1. Message Overview . . . . . . . . . . . . . . . . . . 23 76 2.8.2. GRASP Message Format . . . . . . . . . . . . . . . . 24 77 2.8.3. Message Size . . . . . . . . . . . . . . . . . . . . 25 78 2.8.4. Discovery Message . . . . . . . . . . . . . . . . . . 25 79 2.8.5. Discovery Response Message . . . . . . . . . . . . . 26 80 2.8.6. Request Messages . . . . . . . . . . . . . . . . . . 27 81 2.8.7. Negotiation Message . . . . . . . . . . . . . . . . . 28 82 2.8.8. Negotiation End Message . . . . . . . . . . . . . . . 29 83 2.8.9. Confirm Waiting Message . . . . . . . . . . . . . 29 84 2.8.10. Synchronization Message . . . . . . . . . . . . . . . 29 85 2.8.11. Flood Synchronization Message . . . . . . . . . . . . 30 86 2.8.12. Invalid Message . . . . . . . . . . . . . . . . . . . 31 87 2.8.13. No Operation Message . . . . . . . . . . . . . . . . 31 88 2.9. GRASP Options . . . . . . . . . . . . . . . . . . . . . . 31 89 2.9.1. Format of GRASP Options . . . . . . . . . . . . . . . 31 90 2.9.2. Divert Option . . . . . . . . . . . . . . . . . . . . 32 91 2.9.3. Accept Option . . . . . . . . . . . . . . . . . . . . 32 92 2.9.4. Decline Option . . . . . . . . . . . . . . . . . . . 32 93 2.9.5. Locator Options . . . . . . . . . . . . . . . . . . . 33 94 2.10. Objective Options . . . . . . . . . . . . . . . . . . . . 35 95 2.10.1. Format of Objective Options . . . . . . . . . . . . 35 96 2.10.2. Objective flags . . . . . . . . . . . . . . . . . . 36 97 2.10.3. General Considerations for Objective Options . . . . 36 98 2.10.4. Organizing of Objective Options . . . . . . . . . . 37 99 2.10.5. Experimental and Example Objective Options . . . . . 39 100 3. Implementation Status [RFC Editor: please remove] . . . . . . 39 101 3.1. BUPT C++ Implementation . . . . . . . . . . . . . . . . . 39 102 3.2. Python Implementation . . . . . . . . . . . . . . . . . . 40 103 4. Security Considerations . . . . . . . . . . . . . . . . . . . 41 104 5. CDDL Specification of GRASP . . . . . . . . . . . . . . . . . 43 105 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 45 106 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 47 107 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 47 108 8.1. Normative References . . . . . . . . . . . . . . . . . . 47 109 8.2. Informative References . . . . . . . . . . . . . . . . . 48 110 Appendix A. Open Issues [RFC Editor: This section should be 111 empty. Please remove] . . . . . . . . . . . . . . . 52 112 Appendix B. Closed Issues [RFC Editor: Please remove] . . . . . 52 113 Appendix C. Change log [RFC Editor: Please remove] . . . . . . . 61 114 Appendix D. Example Message Formats . . . . . . . . . . . . . . 68 115 D.1. Discovery Example . . . . . . . . . . . . . . . . . . . . 68 116 D.2. Flood Example . . . . . . . . . . . . . . . . . . . . . . 69 117 D.3. Synchronization Example . . . . . . . . . . . . . . . . . 69 118 D.4. Simple Negotiation Example . . . . . . . . . . . . . . . 69 119 D.5. Complete Negotiation Example . . . . . . . . . . . . . . 70 120 Appendix E. Requirement Analysis of Discovery, Synchronization 121 and Negotiation . . . . . . . . . . . . . . . . . . 71 122 E.1. Requirements for Discovery . . . . . . . . . . . . . . . 71 123 E.2. Requirements for Synchronization and Negotiation 124 Capability . . . . . . . . . . . . . . . . . . . . . . . 73 125 E.3. Specific Technical Requirements . . . . . . . . . . . . . 75 126 Appendix F. Capability Analysis of Current Protocols . . . . . . 76 127 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 78 129 1. Introduction 131 The success of the Internet has made IP-based networks bigger and 132 more complicated. Large-scale ISP and enterprise networks have 133 become more and more problematic for human based management. Also, 134 operational costs are growing quickly. Consequently, there are 135 increased requirements for autonomic behavior in the networks. 136 General aspects of autonomic networks are discussed in [RFC7575] and 137 [RFC7576]. 139 One approach is to largely decentralize the logic of network 140 management by migrating it into network elements. A reference model 141 for autonomic networking on this basis is given in 142 [I-D.ietf-anima-reference-model]. The reader should consult this 143 document to understand how various autonomic components fit together. 144 In order to fulfill autonomy, devices that embody Autonomic Service 145 Agents (ASAs, [RFC7575]) have specific signaling requirements. In 146 particular they need to discover each other, to synchronize state 147 with each other, and to negotiate parameters and resources directly 148 with each other. There is no limitation on the types of parameters 149 and resources concerned, which can include very basic information 150 needed for addressing and routing, as well as anything else that 151 might be configured in a conventional non-autonomic network. The 152 atomic unit of discovery, synchronization or negotiation is referred 153 to as a technical objective, i.e, a configurable parameter or set of 154 parameters (defined more precisely in Section 2.1). 156 Negotiation is an iterative process, requiring multiple message 157 exchanges forming a closed loop between the negotiating entities. In 158 fact, these entities are ASAs, normally but not necessarily in 159 different network devices. State synchronization, when needed, can 160 be regarded as a special case of negotiation, without iteration. 161 Both negotiation and synchronization must logically follow discovery. 162 More details of the requirements are found in Appendix E. 163 Section 2.3 describes a behavior model for a protocol intended to 164 support discovery, synchronization and negotiation. The design of 165 GeneRic Autonomic Signaling Protocol (GRASP) in Section 2 of this 166 document is based on this behavior model. The relevant capabilities 167 of various existing protocols are reviewed in Appendix F. 169 The proposed discovery mechanism is oriented towards synchronization 170 and negotiation objectives. It is based on a neighbor discovery 171 process on the local link, but also supports diversion to peers on 172 other links. There is no assumption of any particular form of 173 network topology. When a device starts up with no pre-configuration, 174 it has no knowledge of the topology. The protocol itself is capable 175 of being used in a small and/or flat network structure such as a 176 small office or home network as well as in a large professionally 177 managed network. Therefore, the discovery mechanism needs to be able 178 to allow a device to bootstrap itself without making any prior 179 assumptions about network structure. 181 Because GRASP can be used as part of a decision process among 182 distributed devices or between networks, it must run in a secure and 183 strongly authenticated environment. 185 In realistic deployments, not all devices will support GRASP. 186 Therefore, some autonomic service agents will directly manage a group 187 of non-autonomic nodes, and other non-autonomic nodes will be managed 188 traditionally. Such mixed scenarios are not discussed in this 189 specification. 191 2. GRASP Protocol Overview 193 2.1. Terminology 195 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 196 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 197 "OPTIONAL" in this document are to be interpreted as described in 198 [RFC2119] when they appear in ALL CAPS. When these words are not in 199 ALL CAPS (such as "should" or "Should"), they have their usual 200 English meanings, and are not to be interpreted as [RFC2119] key 201 words. 203 This document uses terminology defined in [RFC7575]. 205 The following additional terms are used throughout this document: 207 o Discovery: a process by which an ASA discovers peers according to 208 a specific discovery objective. The discovery results may be 209 different according to the different discovery objectives. The 210 discovered peers may later be used as negotiation counterparts or 211 as sources of synchronization data. 213 o Negotiation: a process by which two ASAs interact iteratively to 214 agree on parameter settings that best satisfy the objectives of 215 both ASAs. 217 o State Synchronization: a process by which ASAs interact to receive 218 the current state of parameter values stored in other ASAs. This 219 is a special case of negotiation in which information is sent but 220 the ASAs do not request their peers to change parameter settings. 221 All other definitions apply to both negotiation and 222 synchronization. 224 o Technical Objective (usually abbreviated as Objective): A 225 technical objective is a data structure, whose main contents are a 226 name and a value. The value consists of a single configurable 227 parameter or a set of parameters of some kind. The exact format 228 of an objective is defined in Section 2.10.1. An objective occurs 229 in three contexts: Discovery, Negotiation and Synchronization. 230 Normally, a given objective will not occur in negotiation and 231 synchronization contexts simultaneously. 233 * One ASA may support multiple independent objectives. 235 * The parameter(s) in the value of a given objective apply to a 236 specific service or function or action. They may in principle 237 be anything that can be set to a specific logical, numerical or 238 string value, or a more complex data structure, by a network 239 node. Each node is expected to contain one or more ASAs which 240 may each manage subsidiary non-autonomic nodes. 242 * Discovery Objective: an objective in the process of discovery. 243 Its value may be undefined. 245 * Synchronization Objective: an objective whose specific 246 technical content needs to be synchronized among two or more 247 ASAs. Thus, each ASA will maintain its own copy of the 248 objective. 250 * Negotiation Objective: an objective whose specific technical 251 content needs to be decided in coordination with another ASA. 252 Again, each ASA will maintain its own copy of the objective. 254 A detailed discussion of objectives, including their format, is 255 found in Section 2.10. 257 o Discovery Initiator: an ASA that starts discovery by sending a 258 discovery message referring to a specific discovery objective. 260 o Discovery Responder: a peer that either contains an ASA supporting 261 the discovery objective indicated by the discovery initiator, or 262 caches the locator(s) of the ASA(s) supporting the objective. It 263 sends a Discovery Response, as described later. 265 o Synchronization Initiator: an ASA that starts synchronization by 266 sending a request message referring to a specific synchronization 267 objective. 269 o Synchronization Responder: a peer ASA which responds with the 270 value of a synchronization objective. 272 o Negotiation Initiator: an ASA that starts negotiation by sending a 273 request message referring to a specific negotiation objective. 275 o Negotiation Counterpart: a peer with which the Negotiation 276 Initiator negotiates a specific negotiation objective. 278 o GRASP Instance: This refers to an instantiation of a GRASP 279 protocol engine, likely including multiple threads or processes as 280 well as dynamic data structures such as a discovery cache, running 281 in a given security environment on a single device. 283 o GRASP Core: This refers to the code and shared data structures of 284 a GRASP instance, which will communicate with individual ASAs via 285 a suitable Application Programming Interface (API). 287 o Interface or GRASP Interface: Unless otherwise stated, these refer 288 to a network interface - which might be physical or virtual - that 289 a specific instance of GRASP is currently using. A device might 290 have other interfaces that are not used by GRASP and which are 291 outside the scope of the autonomic network. 293 2.2. High Level Deployment Model 295 A GRASP implementation will be part of the Autonomic Networking 296 Infrastructure (ANI) in an autonomic node, which must also provide an 297 appropriate security environment. In accordance with 298 [I-D.ietf-anima-reference-model], this SHOULD be the Autonomic 299 Control Plane (ACP) [I-D.ietf-anima-autonomic-control-plane]. As a 300 result, all autonomic nodes in the ACP are able to trust each other. 301 It is expected that GRASP will access the ACP by using a typical 302 socket programming interface and the ACP will make available only 303 network interfaces within the autonomic network. If there is no ACP, 304 the considerations described in Section 2.5.1 apply. 306 There will also be one or more Autonomic Service Agents (ASAs). In 307 the minimal case of a single-purpose device, these components might 308 be fully integrated with GRASP and the ACP. A more common model is 309 expected to be a multi-purpose device capable of containing several 310 ASAs, such as a router or large switch. In this case it is expected 311 that the ACP, GRASP and the ASAs will be implemented as separate 312 processes, which are able to support asynchronous and simultaneous 313 operations, for example by multi-threading. 315 In some scenarios, a limited negotiation model might be deployed 316 based on a limited trust relationship such as that between two 317 administrative domains. ASAs might then exchange limited information 318 and negotiate some particular configurations. 320 GRASP is explicitly designed to operate within a single addressing 321 realm. Its discovery and flooding mechanisms do not support 322 autonomic operations that cross any form of address translator or 323 upper layer proxy. 325 A suitable Application Programming Interface (API) will be needed 326 between GRASP and the ASAs. In some implementations, ASAs would run 327 in user space with a GRASP library providing the API, and this 328 library would in turn communicate via system calls with core GRASP 329 functions. Details of the API are out of scope for the present 330 document. For further details of possible deployment models, see 331 [I-D.ietf-anima-reference-model]. 333 An instance of GRASP must be aware of the network interfaces it will 334 use, and of the appropriate global-scope and link-local addresses. 336 In the presence of the ACP, such information will be available from 337 the adjacency table discussed in [I-D.ietf-anima-reference-model]. 338 In other cases, GRASP must determine such information for itself. 339 Details depend on the device and operating system. In the rest of 340 this document, the terms 'interfaces' or 'GRASP interfaces' refers 341 only to the set of network interfaces that a specific instance of 342 GRASP is currently using. 344 Because GRASP needs to work with very high reliability, especially 345 during bootstrapping and during fault conditions, it is essential 346 that every implementation continues to operate in adverse conditions. 347 For example, discovery failures, or any kind of socket exception at 348 any time, must not cause irrecoverable failures in GRASP itself, and 349 must return suitable error codes through the API so that ASAs can 350 also recover. 352 GRASP must not depend upon non-volatile data storage. All run time 353 error conditions, and events such as address renumbering, network 354 interface failures, and CPU sleep/wake cycles, must be handled in 355 such a way that GRASP will still operate correctly and securely 356 (Section 2.5.1) afterwards. 358 An autonomic node will normally run a single instance of GRASP, used 359 by multiple ASAs. Possible exceptions are mentioned below. 361 2.3. High Level Design 363 This section describes the behavior model and general design of 364 GRASP, supporting discovery, synchronization and negotiation, to act 365 as a platform for different technical objectives. 367 o A generic platform: 369 The protocol design is generic and independent of the 370 synchronization or negotiation contents. The technical contents 371 will vary according to the various technical objectives and the 372 different pairs of counterparts. 374 o Normally, a single main instance of the GRASP protocol engine will 375 exist in an autonomic node, and each ASA will run as an 376 independent asynchronous process. However, scenarios where 377 multiple instances of GRASP run in a single node, perhaps with 378 different security properties, are possible (Section 2.5.2). In 379 this case, each instance MUST listen independently for GRASP link- 380 local multicasts, and all instances MUST be woken by each such 381 multicast, in order for discovery and flooding to work correctly. 383 o Security infrastructure: 385 As noted above, the protocol itself has no built-in security 386 functionality, and relies on a separate secure infrastructure. 388 o Discovery, synchronization and negotiation are designed together: 390 The discovery method and the synchronization and negotiation 391 methods are designed in the same way and can be combined when this 392 is useful, allowing a rapid mode of operation described in 393 Section 2.5.4. These processes can also be performed 394 independently when appropriate. 396 * Thus, for some objectives, especially those concerned with 397 application layer services, another discovery mechanism such as 398 the future DNS Service Discovery [RFC7558] MAY be used. The 399 choice is left to the designers of individual ASAs. 401 o A uniform pattern for technical objectives: 403 The synchronization and negotiation objectives are defined 404 according to a uniform pattern. The values that they contain 405 could be carried either in a simple binary format or in a complex 406 object format. The basic protocol design uses the Concise Binary 407 Object Representation (CBOR) [RFC7049], which is readily 408 extensible for unknown future requirements. 410 o A flexible model for synchronization: 412 GRASP supports synchronization between two nodes, which could be 413 used repeatedly to perform synchronization among a small number of 414 nodes. It also supports an unsolicited flooding mode when large 415 groups of nodes, possibly including all autonomic nodes, need data 416 for the same technical objective. 418 * There may be some network parameters for which a more 419 traditional flooding mechanism such as DNCP [RFC7787] is 420 considered more appropriate. GRASP can coexist with DNCP. 422 o A simple initiator/responder model for negotiation: 424 Multi-party negotiations are very complicated to model and cannot 425 readily be guaranteed to converge. GRASP uses a simple bilateral 426 model and can support multi-party negotiations by indirect steps. 428 o Organizing of synchronization or negotiation content: 430 The technical content transmitted by GRASP will be organized 431 according to the relevant function or service. The objectives for 432 different functions or services are kept separate, because they 433 may be negotiated or synchronized with different counterparts or 434 have different response times. Thus a normal arrangement would be 435 a single ASA managing a small set of closely related objectives, 436 with a version of that ASA in each relevant autonomic node. 437 Further discussion of this aspect is out of scope for the current 438 document. 440 o Requests and responses in negotiation procedures: 442 The initiator can negotiate a specific negotiation objective with 443 relevant counterpart ASAs. It can request relevant information 444 from a counterpart so that it can coordinate its local 445 configuration. It can request the counterpart to make a matching 446 configuration. It can request simulation or forecast results by 447 sending some dry run conditions. 449 Beyond the traditional yes/no answer, the responder can reply with 450 a suggested alternative value for the objective concerned. This 451 would start a bi-directional negotiation ending in a compromise 452 between the two ASAs. 454 o Convergence of negotiation procedures: 456 To enable convergence, when a responder suggests a new value or 457 condition in a negotiation step reply, it should be as close as 458 possible to the original request or previous suggestion. The 459 suggested value of later negotiation steps should be chosen 460 between the suggested values from the previous two steps. GRASP 461 provides mechanisms to guarantee convergence (or failure) in a 462 small number of steps, namely a timeout and a maximum number of 463 iterations. 465 o Extensibility: 467 GRASP intentionally does not have a version number, and can be 468 extended by adding new message types and options. The Invalid 469 Message (M_INVALID) will be used to signal that an implementation 470 does not recognize a message or option sent by another 471 implementation. In normal use, new semantics will be added by 472 defining new synchronization or negotiation objectives. 474 2.4. Quick Operating Overview 476 An instance of GRASP is expected to run as a separate core module, 477 providing an API (such as [I-D.liu-anima-grasp-api]) to interface to 478 various ASAs. These ASAs may operate without special privilege, 479 unless they need it for other reasons (such as configuring IP 480 addresses or manipulating routing tables). 482 The GRASP mechanisms used by the ASA are built around GRASP 483 objectives defined as data structures containing administrative 484 information such as the objective's unique name, and its current 485 value. The format and size of the value is not restricted by the 486 protocol, except that it must be possible to serialize it for 487 transmission in CBOR, which is no restriction at all in practice. 489 GRASP provides the following mechanisms: 491 o A discovery mechanism (M_DISCOVERY, M_RESPONSE), by which an ASA 492 can discover other ASAs supporting a given objective. 494 o A negotiation request mechanism (M_REQ_NEG), by which an ASA can 495 start negotiation of an objective with a counterpart ASA. Once a 496 negotiation has started, the process is symmetrical, and there is 497 a negotiation step message (M_NEGOTIATE) for each ASA to use in 498 turn. Two other functions support negotiating steps (M_WAIT, 499 M_END). 501 o A synchronization mechanism (M_REQ_SYN), by which an ASA can 502 request the current value of an objective from a counterpart ASA. 503 With this, there is a corresponding response function (M_SYNCH) 504 for an ASA that wishes to respond to synchronization requests. 506 o A flood mechanism (M_FLOOD), by which an ASA can cause the current 507 value of an objective to be flooded throughout the autonomic 508 network so that any ASA can receive it. One application of this 509 is to act as an announcement, avoiding the need for discovery of a 510 widely applicable objective. 512 Some example messages and simple message flows are provided in 513 Appendix D. 515 2.5. GRASP Protocol Basic Properties and Mechanisms 517 2.5.1. Required External Security Mechanism 519 GRASP does not specify transport security because it is meant to be 520 adapted to different environments. Every solution adopting GRASP 521 MUST specify a security substrate used by GRASP in that solution. 523 The substrate MUST enforce sending and receiving GRASP messages only 524 between members of a mutually trusted group running GRASP. Each 525 group member is an instance of GRASP. The group members are nodes of 526 a connected graph. The group and graph is created by the security 527 substrate and called the GRASP domain. The substrate must support 528 unicast messages between any group members and (link-local) multicast 529 messages between adjacent group members. It must deny messages 530 between group members and non group members. Substrates MUST use 531 cryptographic member authentication and message integrity for GRASP 532 messages. This can be end-to-end or hop-by-hop across the domain. 533 The security substrate MUST provide mechanisms to remove untrusted 534 members from the group. 536 If the substrate does not mandate and enforce GRASP message 537 encryption then any service using GRASP in such a solution MUST 538 provide protection and encryption for message elements whose exposure 539 could constitute an attack vector. 541 The security substrate for GRASP in the ANI is the ACP. Unless 542 otherwise noted, we assume this security substrate in the remainder 543 of this document. The ACP does mandate the use of encryption; 544 therefore GRASP in the ANI can rely on GRASP message being encrypted. 545 The GRASP domain is the ACP: all nodes in an autonomic domain 546 connected by encrypted virtual links formed by the ACP. The ACP uses 547 hop-by-hop security (authentication/encryption) of messages. Removal 548 of nodes relies on standard PKI certificate revocation or expiry of 549 sufficiently short lived certificates. Refer to 550 [I-D.ietf-anima-autonomic-control-plane] for more details. 552 As mentioned in Section 2.3, some GRASP operations might be performed 553 across an administrative domain boundary by mutual agreement, without 554 the benefit of an ACP. Such operations MUST be confined to a 555 separate instance of GRASP with its own copy of all GRASP data 556 structures running across a separate GRASP domain with a security 557 substrate. In the most simple case, each point-to-point interdomain 558 GRASP peering could be a separate domain and the security substrate 559 could be built using transport or network layer security protocols. 560 This is subject to future specifications. 562 An exception to the requirements for the security substrate exists 563 for highly constrained subsets of GRASP meant to support the 564 establishment of a security substrate, described in the following 565 section. 567 2.5.2. Discovery Unsolicited Link-Local (DULL) GRASP 569 Some services may need to use insecure GRASP discovery, response and 570 flood messages without being able to use pre-existing security 571 associations, for example as part of discovery for establishing 572 security associations such as a security substrate for GRASP. 574 Such operations being intrinsically insecure, they need to be 575 confined to link-local use to minimize the risk of malicious actions. 576 Possible examples include discovery of candidate ACP neighbors 577 [I-D.ietf-anima-autonomic-control-plane], discovery of bootstrap 578 proxies [I-D.ietf-anima-bootstrapping-keyinfra] or perhaps 579 initialization services in networks using GRASP without being fully 580 autonomic (e.g., no ACP). Such usage MUST be limited to link-local 581 operations on a single interface and MUST be confined to a separate 582 insecure instance of GRASP with its own copy of all GRASP data 583 structures. This instance is nicknamed DULL - Discovery Unsolicited 584 Link-Local. 586 The detailed rules for the DULL instance of GRASP are as follows: 588 o An initiator MAY send Discovery or Flood Synchronization link- 589 local multicast messages which MUST have a loop count of 1, to 590 prevent off-link operations. Other GRASP message types MUST NOT 591 be sent. 593 o A responder MUST silently discard any message whose loop count is 594 not 1. 596 o A responder MUST silently discard any message referring to a GRASP 597 Objective that is not directly part of a service that requires 598 this insecure mode. 600 o A responder MUST NOT relay any multicast messages. 602 o A Discovery Response MUST indicate a link-local address. 604 o A Discovery Response MUST NOT include a Divert option. 606 o A node MUST silently discard any message whose source address is 607 not link-local. 609 To minimize traffic possibly observed by third parties, GRASP traffic 610 SHOULD be minimized by using only Flood Synchronization to announce 611 objectives and their associated locators, rather than by using 612 Discovery and Response. Further details are out of scope for this 613 document 615 2.5.3. Transport Layer Usage 617 All GRASP messages, after they are serialized as a CBOR byte string, 618 are transmitted as such directly over the transport protocol in use, 619 which itself runs within the security environment discussed in the 620 previous section. 622 GRASP discovery and flooding messages are designed for use over link- 623 local multicast UDP. They MUST NOT be fragmented, and therefore MUST 624 NOT exceed the link MTU size. An implementation should report any 625 attempt to send a longer message as a run-time error. 627 All other GRASP messages are unicast and could in principle run over 628 any transport protocol. An implementation MUST support use of TCP. 629 It MAY support use of another transport protocol but the details are 630 out of scope for this specification. However, GRASP itself does not 631 provide for error detection or retransmission of its unicast 632 messages, so an unreliable transport protocol MUST NOT be used. 634 For considerations when running without an ACP, see Section 2.5.1. 636 For link-local multicast, the GRASP protocol listens to the well- 637 known GRASP Listen Port (Section 2.6). For unicast transport 638 sessions used for discovery responses, synchronization and 639 negotiation, the ASA concerned normally listens on its own 640 dynamically assigned ports, which are communicated to its peers 641 during discovery. However, a minimal implementation MAY use the 642 GRASP Listen Port for this purpose. 644 2.5.4. Discovery Mechanism and Procedures 646 2.5.4.1. Separated discovery and negotiation mechanisms 648 Although discovery and negotiation or synchronization are defined 649 together in GRASP, they are separate mechanisms. The discovery 650 process could run independently from the negotiation or 651 synchronization process. Upon receiving a Discovery (Section 2.8.4) 652 message, the recipient node should return a response message in which 653 it either indicates itself as a discovery responder or diverts the 654 initiator towards another more suitable ASA. However, this response 655 may be delayed if the recipient needs to relay the discovery onwards, 656 as described below. 658 The discovery action (M_DISCOVERY) will normally be followed by a 659 negotiation (M_REQ_NEG) or synchronization (M_REQ_SYN) action. The 660 discovery results could be utilized by the negotiation protocol to 661 decide which ASA the initiator will negotiate with. 663 The initiator of a discovery action for a given objective need not be 664 capable of responding to that objective as a Negotiation Counterpart, 665 as a Synchronization Responder or as source for flooding. For 666 example, an ASA might perform discovery even if it only wishes to act 667 a Synchronization Initiator or Negotiation Initiator. Such an ASA 668 does not itself need to respond to discovery messages. 670 It is also entirely possible to use GRASP discovery without any 671 subsequent negotiation or synchronization action. In this case, the 672 discovered objective is simply used as a name during the discovery 673 process and any subsequent operations between the peers are outside 674 the scope of GRASP. 676 2.5.4.2. Discovery Overview 678 A complete discovery process will start with a multicast (of 679 M_DISCOVERY) on the local link. On-link neighbors supporting the 680 discovery objective will respond directly (with M_RESPONSE). A 681 neighbor with multiple interfaces may respond with a cached discovery 682 response. If it has no cached response, it will relay the discovery 683 on its other GRASP interfaces. If a node receiving the relayed 684 discovery supports the discovery objective, it will respond to the 685 relayed discovery. If it has a cached response, it will respond with 686 that. If not, it will repeat the discovery process, which thereby 687 becomes iterative. The loop count and timeout will ensure that the 688 process ends. Further details are given below. 690 A Discovery message MAY be sent unicast (via UDP or TCP) to a peer 691 node, which SHOULD then proceed exactly as if the message had been 692 multicast, except that when TCP is used, the response will be on the 693 same socket as the query. However, this mode does not guarantee 694 successful discovery in the general case. 696 2.5.4.3. Discovery Procedures 698 Discovery starts as an on-link operation. The Divert option can tell 699 the discovery initiator to contact an off-link ASA for that discovery 700 objective. A Discovery message is sent by a discovery initiator via 701 UDP to the ALL_GRASP_NEIGHBORS link-local multicast address 702 (Section 2.6). Every network device that supports GRASP always 703 listens to a well-known UDP port to capture the discovery messages. 704 Because this port is unique in a device, this is a function of the 705 GRASP instance and not of an individual ASA. As a result, each ASA 706 will need to register the objectives that it supports with the local 707 GRASP instance. 709 If an ASA in a neighbor device supports the requested discovery 710 objective, the device SHOULD respond to the link-local multicast with 711 a unicast Discovery Response message (Section 2.8.5) with locator 712 option(s), unless it is temporarily unavailable. Otherwise, if the 713 neighbor has cached information about an ASA that supports the 714 requested discovery objective (usually because it discovered the same 715 objective before), it SHOULD respond with a Discovery Response 716 message with a Divert option pointing to the appropriate Discovery 717 Responder. However, on a link that supports link-local multicast, it 718 SHOULD NOT respond with a cached response on an interface if it 719 learnt that information from the same interface, because the peer in 720 question will answer directly if still operational. 722 If a device has no information about the requested discovery 723 objective, and is not acting as a discovery relay (see below) it MUST 724 silently discard the Discovery message. 726 The discovery initiator MUST set a reasonable timeout on the 727 discovery process. A suggested value is 100 milliseconds multiplied 728 by the loop count embedded in the objective. 730 If no discovery response is received within the timeout, the 731 Discovery message MAY be repeated, with a newly generated Session ID 732 (Section 2.7). An exponential backoff SHOULD be used for subsequent 733 repetitions, to limit the load during busy periods. The details of 734 the backoff algorithm will depend on the use case for the objective 735 concerned but MUST be consistent with the recommendations in 736 [RFC8085] for low data-volume multicast. Frequent repetition might 737 be symptomatic of a denial of service attack. 739 After a GRASP device successfully discovers a locator for a Discovery 740 Responder supporting a specific objective, it SHOULD cache this 741 information, including the interface index [RFC3493] via which it was 742 discovered. This cache record MAY be used for future negotiation or 743 synchronization, and the locator SHOULD be passed on when appropriate 744 as a Divert option to another Discovery Initiator. 746 The cache mechanism MUST include a lifetime for each entry. The 747 lifetime is derived from a time-to-live (ttl) parameter in each 748 Discovery Response message. Cached entries MUST be ignored or 749 deleted after their lifetime expires. In some environments, 750 unplanned address renumbering might occur. In such cases, the 751 lifetime SHOULD be short compared to the typical address lifetime. 752 The discovery mechanism needs to track the node's current address to 753 ensure that Discovery Responses always indicate the correct address. 755 If multiple Discovery Responders are found for the same objective, 756 they SHOULD all be cached, unless this creates a resource shortage. 757 The method of choosing between multiple responders is an 758 implementation choice. This choice MUST be available to each ASA but 759 the GRASP implementation SHOULD provide a default choice. 761 Because Discovery Responders will be cached in a finite cache, they 762 might be deleted at any time. In this case, discovery will need to 763 be repeated. If an ASA exits for any reason, its locator might still 764 be cached for some time, and attempts to connect to it will fail. 765 ASAs need to be robust in these circumstances. 767 2.5.4.4. Discovery Relaying 769 A GRASP instance with multiple link-layer interfaces (typically 770 running in a router) MUST support discovery on all GRASP interfaces. 771 We refer to this as a 'relaying instance'. 773 Constrained Instances (Section 2.5.2) are always single-interface 774 instances and therefore MUST NOT perform discovery relaying. 776 If a relaying instance receives a Discovery message on a given 777 interface for a specific objective that it does not support and for 778 which it has not previously cached a Discovery Responder, it MUST 779 relay the query by re-issuing a new Discovery message as a link-local 780 multicast on its other GRASP interfaces. 782 The relayed discovery message MUST have the same Session ID as the 783 incoming discovery message and MUST be tagged with the IP address of 784 its original initiator (see Section 2.8.4). Note that this initiator 785 address is only used to allow for disambiguation of the Session ID 786 and is never used to address Response packets, which are sent to the 787 relaying instance, not the original initiator. 789 The relaying instance MUST decrement the loop count within the 790 objective, and MUST NOT relay the Discovery message if the result is 791 zero. Also, it MUST limit the total rate at which it relays 792 discovery messages to a reasonable value, in order to mitigate 793 possible denial of service attacks. For example, the rate limit 794 could be set to a small multiple of the observed rate of discovery 795 messages during normal operation. The relaying instance MUST cache 796 the Session ID value and initiator address of each relayed Discovery 797 message until any Discovery Responses have arrived or the discovery 798 process has timed out. To prevent loops, it MUST NOT relay a 799 Discovery message which carries a given cached Session ID and 800 initiator address more than once. These precautions avoid discovery 801 loops and mitigate potential overload. 803 Since the relay device is unaware of the timeout set by the original 804 initiator it SHOULD set a suitable timeout for the relayed discovery. 805 A suggested value is 100 milliseconds multiplied by the remaining 806 loop count. 808 The discovery results received by the relaying instance MUST in turn 809 be sent as a Discovery Response message to the Discovery message that 810 caused the relay action. 812 2.5.4.5. Rapid Mode (Discovery with Negotiation or Synchronization ) 814 A Discovery message MAY include an Objective option. This allows a 815 rapid mode of negotiation (Section 2.5.5.1) or synchronization 816 (Section 2.5.6.3). Rapid mode is currently limited to a single 817 objective for simplicity of design and implementation. A possible 818 future extension is to allow multiple objectives in rapid mode for 819 greater efficiency. 821 2.5.5. Negotiation Procedures 823 A negotiation initiator opens a transport connection to a counterpart 824 ASA using the address, protocol and port obtained during discovery. 825 It then sends a negotiation request (using M_REQ_NEG) to the 826 counterpart, including a specific negotiation objective. It may 827 request the negotiation counterpart to make a specific configuration. 828 Alternatively, it may request a certain simulation or forecast result 829 by sending a dry run configuration. The details, including the 830 distinction between a dry run and a live configuration change, will 831 be defined separately for each type of negotiation objective. Any 832 state associated with a dry run operation, such as temporarily 833 reserving a resource for subsequent use in a live run, is entirely a 834 matter for the designer of the ASA concerned. 836 Each negotiation session as a whole is subject to a timeout (default 837 GRASP_DEF_TIMEOUT milliseconds, Section 2.6), initialised when the 838 request is sent (see Section 2.8.6). If no reply message of any kind 839 is received within the timeout, the negotiation request MAY be 840 repeated, with a newly generated Session ID (Section 2.7). An 841 exponential backoff SHOULD be used for subsequent repetitions. The 842 details of the backoff algorithm will depend on the use case for the 843 objective concerned. 845 If the counterpart can immediately apply the requested configuration, 846 it will give an immediate positive (O_ACCEPT) answer (using M_END). 847 This will end the negotiation phase immediately. Otherwise, it will 848 negotiate (using M_NEGOTIATE). It will reply with a proposed 849 alternative configuration that it can apply (typically, a 850 configuration that uses fewer resources than requested by the 851 negotiation initiator). This will start a bi-directional negotiation 852 (using M_NEGOTIATE) to reach a compromise between the two ASAs. 854 The negotiation procedure is ended when one of the negotiation peers 855 sends a Negotiation Ending (M_END) message, which contains an accept 856 (O_ACCEPT) or decline (O_DECLINE) option and does not need a response 857 from the negotiation peer. Negotiation may also end in failure 858 (equivalent to a decline) if a timeout is exceeded or a loop count is 859 exceeded. When the procedure ends for whatever reason, the transport 860 connection SHOULD be closed. A transport session failure is treated 861 as a negotiation failure. 863 A negotiation procedure concerns one objective and one counterpart. 864 Both the initiator and the counterpart may take part in simultaneous 865 negotiations with various other ASAs, or in simultaneous negotiations 866 about different objectives. Thus, GRASP is expected to be used in a 867 multi-threaded mode or its logical equivalent. Certain negotiation 868 objectives may have restrictions on multi-threading, for example to 869 avoid over-allocating resources. 871 Some configuration actions, for example wavelength switching in 872 optical networks, might take considerable time to execute. The ASA 873 concerned needs to allow for this by design, but GRASP does allow for 874 a peer to insert latency in a negotiation process if necessary 875 (Section 2.8.9, M_WAIT). 877 2.5.5.1. Rapid Mode (Discovery/Negotiation Linkage) 879 A Discovery message MAY include a Negotiation Objective option. In 880 this case it is as if the initiator sent the sequence M_DISCOVERY, 881 immediately followed by M_REQ_NEG. This has implications for the 882 construction of the GRASP core, as it must carefully pass the 883 contents of the Negotiation Objective option to the ASA so that it 884 may evaluate the objective directly. When a Negotiation Objective 885 option is present the ASA replies with an M_NEGOTIATE message (or 886 M_END with O_ACCEPT if it is immediately satisfied with the 887 proposal), rather than with an M_RESPONSE. However, if the recipient 888 node does not support rapid mode, discovery will continue normally. 890 It is possible that a Discovery Response will arrive from a responder 891 that does not support rapid mode, before such a Negotiation message 892 arrives. In this case, rapid mode will not occur. 894 This rapid mode could reduce the interactions between nodes so that a 895 higher efficiency could be achieved. However, a network in which 896 some nodes support rapid mode and others do not will have complex 897 timing-dependent behaviors. Therefore, the rapid negotiation 898 function SHOULD be disabled by default. 900 2.5.6. Synchronization and Flooding Procedures 902 2.5.6.1. Unicast Synchronization 904 A synchronization initiator opens a transport connection to a 905 counterpart ASA using the address, protocol and port obtained during 906 discovery. It then sends a synchronization request (using M_REQ_SYN) 907 to the counterpart, including a specific synchronization objective. 908 The counterpart responds with a Synchronization message (M_SYNCH, 909 Section 2.8.10) containing the current value of the requested 910 synchronization objective. No further messages are needed and the 911 transport connection SHOULD be closed. A transport session failure 912 is treated as a synchronization failure. 914 If no reply message of any kind is received within a given timeout 915 (default GRASP_DEF_TIMEOUT milliseconds, Section 2.6), the 916 synchronization request MAY be repeated, with a newly generated 917 Session ID (Section 2.7). An exponential backoff SHOULD be used for 918 subsequent repetitions. The details of the backoff algorithm will 919 depend on the use case for the objective concerned. 921 2.5.6.2. Flooding 923 In the case just described, the message exchange is unicast and 924 concerns only one synchronization objective. For large groups of 925 nodes requiring the same data, synchronization flooding is available. 926 For this, a flooding initiator MAY send an unsolicited Flood 927 Synchronization message containing one or more Synchronization 928 Objective option(s), if and only if the specification of those 929 objectives permits it. This is sent as a multicast message to the 930 ALL_GRASP_NEIGHBORS multicast address (Section 2.6). 932 Receiving flood multicasts is a function of the GRASP core, as in the 933 case of discovery multicasts (Section 2.5.4.3). 935 To ensure that flooding does not result in a loop, the originator of 936 the Flood Synchronization message MUST set the loop count in the 937 objectives to a suitable value (the default is GRASP_DEF_LOOPCT). 938 Also, a suitable mechanism is needed to avoid excessive multicast 939 traffic. This mechanism MUST be defined as part of the specification 940 of the synchronization objective(s) concerned. It might be a simple 941 rate limit or a more complex mechanism such as the Trickle algorithm 942 [RFC6206]. 944 A GRASP device with multiple link-layer interfaces (typically a 945 router) MUST support synchronization flooding on all GRASP 946 interfaces. If it receives a multicast Flood Synchronization message 947 on a given interface, it MUST relay it by re-issuing a Flood 948 Synchronization message as a link-local multicast on its other GRASP 949 interfaces. The relayed message MUST have the same Session ID as the 950 incoming message and MUST be tagged with the IP address of its 951 original initiator. 953 Link-layer Flooding is supported by GRASP by setting the loop count 954 to 1, and sending with a link-local source address. Floods with 955 link-local source addresses and a loop count other than 1 are 956 invalid, and such messages MUST be discarded. 958 The relaying device MUST decrement the loop count within the first 959 objective, and MUST NOT relay the Flood Synchronization message if 960 the result is zero. Also, it MUST limit the total rate at which it 961 relays Flood Synchronization messages to a reasonable value, in order 962 to mitigate possible denial of service attacks. For example, the 963 rate limit could be set to a small multiple of the observed rate of 964 flood messages during normal operation. The relaying device MUST 965 cache the Session ID value and initiator address of each relayed 966 Flood Synchronization message for a time not less than twice 967 GRASP_DEF_TIMEOUT milliseconds. To prevent loops, it MUST NOT relay 968 a Flood Synchronization message which carries a given cached Session 969 ID and initiator address more than once. These precautions avoid 970 synchronization loops and mitigate potential overload. 972 Note that this mechanism is unreliable in the case of sleeping nodes, 973 or new nodes that join the network, or nodes that rejoin the network 974 after a fault. An ASA that initiates a flood SHOULD repeat the flood 975 at a suitable frequency, which MUST be consistent with the 976 recommendations in [RFC8085] for low data-volume multicast. The ASA 977 SHOULD also act as a synchronization responder for the objective(s) 978 concerned. Thus nodes that require an objective subject to flooding 979 can either wait for the next flood or request unicast synchronization 980 for that objective. 982 The multicast messages for synchronization flooding are subject to 983 the security rules in Section 2.5.1. In practice this means that 984 they MUST NOT be transmitted and MUST be ignored on receipt unless 985 there is an operational ACP or equivalent strong security in place. 986 However, because of the security weakness of link-local multicast 987 (Section 4), synchronization objectives that are flooded SHOULD NOT 988 contain unencrypted private information and SHOULD be validated by 989 the recipient ASA. 991 2.5.6.3. Rapid Mode (Discovery/Synchronization Linkage) 993 A Discovery message MAY include a Synchronization Objective option. 994 In this case the Discovery message also acts as a Request 995 Synchronization message to indicate to the Discovery Responder that 996 it could directly reply to the Discovery Initiator with a 997 Synchronization message Section 2.8.10 with synchronization data for 998 rapid processing, if the discovery target supports the corresponding 999 synchronization objective. The design implications are similar to 1000 those discussed in Section 2.5.5.1. 1002 It is possible that a Discovery Response will arrive from a responder 1003 that does not support rapid mode, before such a Synchronization 1004 message arrives. In this case, rapid mode will not occur. 1006 This rapid mode could reduce the interactions between nodes so that a 1007 higher efficiency could be achieved. However, a network in which 1008 some nodes support rapid mode and others do not will have complex 1009 timing-dependent behaviors. Therefore, the rapid synchronization 1010 function SHOULD be configured off by default and MAY be configured on 1011 or off by Intent. 1013 2.6. GRASP Constants 1015 o ALL_GRASP_NEIGHBORS 1017 A link-local scope multicast address used by a GRASP-enabled 1018 device to discover GRASP-enabled neighbor (i.e., on-link) devices. 1019 All devices that support GRASP are members of this multicast 1020 group. 1022 * IPv6 multicast address: TBD1 1024 * IPv4 multicast address: TBD2 1026 o GRASP_LISTEN_PORT (TBD3) 1028 A well-known UDP user port that every GRASP-enabled network device 1029 MUST always listen to for link-local multicasts. This user port 1030 MAY also be used to listen for TCP or UDP unicast messages in a 1031 simple implementation of GRASP (Section 2.5.3). 1033 o GRASP_DEF_TIMEOUT (60000 milliseconds) 1035 The default timeout used to determine that an operation has failed 1036 to complete. 1038 o GRASP_DEF_LOOPCT (6) 1040 The default loop count used to determine that a negotiation has 1041 failed to complete, and to avoid looping messages. 1043 o GRASP_DEF_MAX_SIZE (2048) 1044 The default maximum message size in bytes. 1046 2.7. Session Identifier (Session ID) 1048 This is an up to 32-bit opaque value used to distinguish multiple 1049 sessions between the same two devices. A new Session ID MUST be 1050 generated by the initiator for every new Discovery, Flood 1051 Synchronization or Request message. All responses and follow-up 1052 messages in the same discovery, synchronization or negotiation 1053 procedure MUST carry the same Session ID. 1055 The Session ID SHOULD have a very low collision rate locally. It 1056 MUST be generated by a pseudo-random number generator (PRNG) using a 1057 locally generated seed which is unlikely to be used by any other 1058 device in the same network. The PRNG SHOULD be cryptographically 1059 strong [RFC4086]. When allocating a new Session ID, GRASP MUST check 1060 that the value is not already in use and SHOULD check that it has not 1061 been used recently, by consulting a cache of current and recent 1062 sessions. In the unlikely event of a clash, GRASP MUST generate a 1063 new value. 1065 However, there is a finite probability that two nodes might generate 1066 the same Session ID value. For that reason, when a Session ID is 1067 communicated via GRASP, the receiving node MUST tag it with the 1068 initiator's IP address to allow disambiguation. In the highly 1069 unlikely event of two peers opening sessions with the same Session ID 1070 value, this tag will allow the two sessions to be distinguished. 1071 Multicast GRASP messages and their responses, which may be relayed 1072 between links, therefore include a field that carries the initiator's 1073 global IP address. 1075 There is a highly unlikely race condition in which two peers start 1076 simultaneous negotiation sessions with each other using the same 1077 Session ID value. Depending on various implementation choices, this 1078 might lead to the two sessions being confused. See Section 2.8.6 for 1079 details of how to avoid this. 1081 2.8. GRASP Messages 1083 2.8.1. Message Overview 1085 This section defines the GRASP message format and message types. 1086 Message types not listed here are reserved for future use. 1088 The messages currently defined are: 1090 Discovery and Discovery Response (M_DISCOVERY, M_RESPONSE). 1092 Request Negotiation, Negotiation, Confirm Waiting and Negotiation 1093 End (M_REQ_NEG, M_NEGOTIATE, M_WAIT, M_END). 1095 Request Synchronization, Synchronization, and Flood 1096 Synchronization (M_REQ_SYN, M_SYNCH, M_FLOOD. 1098 No Operation and Invalid (M_NOOP, M_INVALID). 1100 2.8.2. GRASP Message Format 1102 GRASP messages share an identical header format and a variable format 1103 area for options. GRASP message headers and options are transmitted 1104 in Concise Binary Object Representation (CBOR) [RFC7049]. In this 1105 specification, they are described using CBOR data definition language 1106 (CDDL) [I-D.greevenbosch-appsawg-cbor-cddl]. Fragmentary CDDL is 1107 used to describe each item in this section. A complete and normative 1108 CDDL specification of GRASP is given in Section 5, including 1109 constants such as message types. 1111 Every GRASP message, except the No Operation message, carries a 1112 Session ID (Section 2.7). Options are then presented serially in the 1113 options field. 1115 In fragmentary CDDL, every GRASP message follows the pattern: 1117 grasp-message = (message .within message-structure) / noop-message 1119 message-structure = [MESSAGE_TYPE, session-id, ?initiator, 1120 *grasp-option] 1122 MESSAGE_TYPE = 1..255 1123 session-id = 0..4294967295 ;up to 32 bits 1124 grasp-option = any 1126 The MESSAGE_TYPE indicates the type of the message and thus defines 1127 the expected options. Any options received that are not consistent 1128 with the MESSAGE_TYPE SHOULD be silently discarded. 1130 The No Operation (noop) message is described in Section 2.8.13. 1132 The various MESSAGE_TYPE values are defined in Section 5. 1134 All other message elements are described below and formally defined 1135 in Section 5. 1137 If an unrecognized MESSAGE_TYPE is received in a unicast message, an 1138 Invalid message (Section 2.8.12) MAY be returned. Otherwise the 1139 message MAY be logged and MUST be discarded. If an unrecognized 1140 MESSAGE_TYPE is received in a multicast message, it MAY be logged and 1141 MUST be silently discarded. 1143 2.8.3. Message Size 1145 GRASP nodes MUST be able to receive unicast messages of at least 1146 GRASP_DEF_MAX_SIZE bytes. GRASP nodes MUST NOT send unicast messages 1147 longer than GRASP_DEF_MAX_SIZE bytes unless a longer size is 1148 explicitly allowed for the objective concerned. For example, GRASP 1149 negotiation itself could be used to agree on a longer message size. 1151 The message parser used by GRASP should be configured to know about 1152 the GRASP_DEF_MAX_SIZE, or any larger negotiated message size, so 1153 that it may defend against overly long messages. 1155 The maximum size of multicast messages (M_DISCOVERY and M_FLOOD) 1156 depends on the link layer technology or link adaptation layer in use. 1158 2.8.4. Discovery Message 1160 In fragmentary CDDL, a Discovery message follows the pattern: 1162 discovery-message = [M_DISCOVERY, session-id, initiator, objective] 1164 A discovery initiator sends a Discovery message to initiate a 1165 discovery process for a particular objective option. 1167 The discovery initiator sends all Discovery messages via UDP to port 1168 GRASP_LISTEN_PORT at the link-local ALL_GRASP_NEIGHBORS multicast 1169 address on each link-layer interface in use by GRASP. It then 1170 listens for unicast TCP responses on a given port, and stores the 1171 discovery results (including responding discovery objectives and 1172 corresponding unicast locators). 1174 The listening port used for TCP MUST be the same port as used for 1175 sending the Discovery UDP multicast, on a given interface. In an 1176 implementation with a single GRASP instance in a node this MAY be 1177 GRASP_LISTEN_PORT. To support multiple instances in the same node, 1178 the GRASP discovery mechanism in each instance needs to find, for 1179 each interface, a dynamic port that it can bind to for both sending 1180 UDP link-local multicast and listening for TCP, before initiating any 1181 discovery. 1183 The 'initiator' field in the message is a globally unique IP address 1184 of the initiator, for the sole purpose of disambiguating the Session 1185 ID in other nodes. If for some reason the initiator does not have a 1186 globally unique IP address, it MUST use a link-local address for this 1187 purpose that is highly likely to be unique, for example using 1189 [RFC7217]. Determination of a node's globally unique IP address is 1190 implementation-dependent. 1192 A Discovery message MUST include exactly one of the following: 1194 o a discovery objective option (Section 2.10.1). Its loop count 1195 MUST be set to a suitable value to prevent discovery loops 1196 (default value is GRASP_DEF_LOOPCT). If the discovery initiator 1197 requires only on-link responses, the loop count MUST be set to 1. 1199 o a negotiation objective option (Section 2.10.1). This is used 1200 both for the purpose of discovery and to indicate to the discovery 1201 target that it MAY directly reply to the discovery initiatior with 1202 a Negotiation message for rapid processing, if it could act as the 1203 corresponding negotiation counterpart. The sender of such a 1204 Discovery message MUST initialize a negotiation timer and loop 1205 count in the same way as a Request Negotiation message 1206 (Section 2.8.6). 1208 o a synchronization objective option (Section 2.10.1). This is used 1209 both for the purpose of discovery and to indicate to the discovery 1210 target that it MAY directly reply to the discovery initiator with 1211 a Synchronization message for rapid processing, if it could act as 1212 the corresponding synchronization counterpart. Its loop count 1213 MUST be set to a suitable value to prevent discovery loops 1214 (default value is GRASP_DEF_LOOPCT). 1216 As mentioned in Section 2.5.4.2, a Discovery message MAY be sent 1217 unicast to a peer node, which SHOULD then proceed exactly as if the 1218 message had been multicast. 1220 2.8.5. Discovery Response Message 1222 In fragmentary CDDL, a Discovery Response message follows the 1223 pattern: 1225 response-message = [M_RESPONSE, session-id, initiator, ttl, 1226 (+locator-option // divert-option), ?objective)] 1228 ttl = 0..4294967295 ; in milliseconds 1230 A node which receives a Discovery message SHOULD send a Discovery 1231 Response message if and only if it can respond to the discovery. 1233 It MUST contain the same Session ID and initiator as the Discovery 1234 message. 1236 It MUST contain a time-to-live (ttl) for the validity of the 1237 response, given as a positive integer value in milliseconds. Zero 1238 implies a value significantly greater than GRASP_DEF_TIMEOUT 1239 milliseconds (Section 2.6). A suggested value is ten times that 1240 amount. 1242 It MAY include a copy of the discovery objective from the 1243 Discovery message. 1245 It is sent to the sender of the Discovery message via TCP at the port 1246 used to send the Discovery message (as explained in Section 2.8.4). 1247 In the case of a relayed Discovery message, the Discovery Response is 1248 thus sent to the relay, not the original initiator. 1250 In all cases, the transport session SHOULD be closed after sending 1251 the Discovery Response. A transport session failure is treated as no 1252 response. 1254 If the responding node supports the discovery objective of the 1255 discovery, it MUST include at least one kind of locator option 1256 (Section 2.9.5) to indicate its own location. A sequence of multiple 1257 kinds of locator options (e.g. IP address option and FQDN option) is 1258 also valid. 1260 If the responding node itself does not support the discovery 1261 objective, but it knows the locator of the discovery objective, then 1262 it SHOULD respond to the discovery message with a divert option 1263 (Section 2.9.2) embedding a locator option or a combination of 1264 multiple kinds of locator options which indicate the locator(s) of 1265 the discovery objective. 1267 More details on the processing of Discovery Responses are given in 1268 Section 2.5.4. 1270 2.8.6. Request Messages 1272 In fragmentary CDDL, Request Negotiation and Request Synchronization 1273 messages follow the patterns: 1275 request-negotiation-message = [M_REQ_NEG, session-id, objective] 1277 request-synchronization-message = [M_REQ_SYN, session-id, objective] 1279 A negotiation or synchronization requesting node sends the 1280 appropriate Request message to the unicast address of the negotiation 1281 or synchronization counterpart, using the appropriate protocol and 1282 port numbers (selected from the discovery result). If the discovery 1283 result is an FQDN, it will be resolved first. 1285 A Request message MUST include the relevant objective option. In the 1286 case of Request Negotiation, the objective option MUST include the 1287 requested value. 1289 When an initiator sends a Request Negotiation message, it MUST 1290 initialize a negotiation timer for the new negotiation thread. The 1291 default is GRASP_DEF_TIMEOUT milliseconds. Unless this timeout is 1292 modified by a Confirm Waiting message (Section 2.8.9), the initiator 1293 will consider that the negotiation has failed when the timer expires. 1295 Similarly, when an initiator sends a Request Synchronization, it 1296 SHOULD initialize a synchronization timer. The default is 1297 GRASP_DEF_TIMEOUT milliseconds. The initiator will consider that 1298 synchronization has failed if there is no response before the timer 1299 expires. 1301 When an initiator sends a Request message, it MUST initialize the 1302 loop count of the objective option with a value defined in the 1303 specification of the option or, if no such value is specified, with 1304 GRASP_DEF_LOOPCT. 1306 If a node receives a Request message for an objective for which no 1307 ASA is currently listening, it MUST immediately close the relevant 1308 socket to indicate this to the initiator. This is to avoid 1309 unnecessary timeouts if, for example, an ASA exits prematurely but 1310 the GRASP core is listening on its behalf. 1312 To avoid the highly unlikely race condition in which two nodes 1313 simultaneously request sessions with each other using the same 1314 Session ID (Section 2.7), when a node receives a Request message, it 1315 MUST verify that the received Session ID is not already locally 1316 active. In case of a clash, it MUST discard the Request message, in 1317 which case the initiator will detect a timeout. 1319 2.8.7. Negotiation Message 1321 In fragmentary CDDL, a Negotiation message follows the pattern: 1323 negotiate-message = [M_NEGOTIATE, session-id, objective] 1325 A negotiation counterpart sends a Negotiation message in response to 1326 a Request Negotiation message, a Negotiation message, or a Discovery 1327 message in Rapid Mode. A negotiation process MAY include multiple 1328 steps. 1330 The Negotiation message MUST include the relevant Negotiation 1331 Objective option, with its value updated according to progress in the 1332 negotiation. The sender MUST decrement the loop count by 1. If the 1333 loop count becomes zero the message MUST NOT be sent. In this case 1334 the negotiation session has failed and will time out. 1336 2.8.8. Negotiation End Message 1338 In fragmentary CDDL, a Negotiation End message follows the pattern: 1340 end-message = [M_END, session-id, accept-option / decline-option] 1342 A negotiation counterpart sends an Negotiation End message to close 1343 the negotiation. It MUST contain either an accept or a decline 1344 option, defined in Section 2.9.3 and Section 2.9.4. It could be sent 1345 either by the requesting node or the responding node. 1347 2.8.9. Confirm Waiting Message 1349 In fragmentary CDDL, a Confirm Waiting message follows the pattern: 1351 wait-message = [M_WAIT, session-id, waiting-time] 1352 waiting-time = 0..4294967295 ; in milliseconds 1354 A responding node sends a Confirm Waiting message to ask the 1355 requesting node to wait for a further negotiation response. It might 1356 be that the local process needs more time or that the negotiation 1357 depends on another triggered negotiation. This message MUST NOT 1358 include any other options. When received, the waiting time value 1359 overwrites and restarts the current negotiation timer 1360 (Section 2.8.6). 1362 The responding node SHOULD send a Negotiation, Negotiation End or 1363 another Confirm Waiting message before the negotiation timer expires. 1364 If not, when the initiator's timer expires, the initiator MUST treat 1365 the negotiation procedure as failed. 1367 2.8.10. Synchronization Message 1369 In fragmentary CDDL, a Synchronization message follows the pattern: 1371 synch-message = [M_SYNCH, session-id, objective] 1373 A node which receives a Request Synchronization, or a Discovery 1374 message in Rapid Mode, sends back a unicast Synchronization message 1375 with the synchronization data, in the form of a GRASP Option for the 1376 specific synchronization objective present in the Request 1377 Synchronization. 1379 2.8.11. Flood Synchronization Message 1381 In fragmentary CDDL, a Flood Synchronization message follows the 1382 pattern: 1384 flood-message = [M_FLOOD, session-id, initiator, ttl, 1385 +[objective, (locator-option / [])]] 1387 ttl = 0..4294967295 ; in milliseconds 1389 A node MAY initiate flooding by sending an unsolicited Flood 1390 Synchronization Message with synchronization data. This MAY be sent 1391 to port GRASP_LISTEN_PORT at the link-local ALL_GRASP_NEIGHBORS 1392 multicast address, in accordance with the rules in Section 2.5.6. 1394 The initiator address is provided, as described for Discovery 1395 messages (Section 2.8.4), only to disambiguate the Session ID. 1397 The message MUST contain a time-to-live (ttl) for the validity of 1398 the contents, given as a positive integer value in milliseconds. 1399 There is no default; zero indicates an indefinite lifetime. 1401 The synchronization data are in the form of GRASP Option(s) for 1402 specific synchronization objective(s). The loop count(s) MUST be 1403 set to a suitable value to prevent flood loops (default value is 1404 GRASP_DEF_LOOPCT). 1406 Each objective option MAY be followed by a locator option 1407 associated with the flooded objective. In its absence, an empty 1408 option MUST be included to indicate a null locator. 1410 A node that receives a Flood Synchronization message MUST cache the 1411 received objectives for use by local ASAs. Each cached objective 1412 MUST be tagged with the locator option sent with it, or with a null 1413 tag if an empty locator option was sent. If a subsequent Flood 1414 Synchronization message carrying an objective with same name and the 1415 same tag, the corresponding cached copy of the objective MUST be 1416 overwritten. If a subsequent Flood Synchronization message carrying 1417 an objective with same name arrives with a different tag, a new 1418 cached entry MUST be created. 1420 Note: the purpose of this mechanism is to allow the recipient of 1421 flooded values to distinguish between different senders of the same 1422 objective, and if necessary communicate with them using the locator, 1423 protocol and port included in the locator option. Many objectives 1424 will not need this mechanism, so they will be flooded with a null 1425 locator. 1427 Cached entries MUST be ignored or deleted after their lifetime 1428 expires. 1430 2.8.12. Invalid Message 1432 In fragmentary CDDL, an Invalid message follows the pattern: 1434 invalid-message = [M_INVALID, session-id, ?any] 1436 This message MAY be sent by an implementation in response to an 1437 incoming unicast message that it considers invalid. The session-id 1438 MUST be copied from the incoming message. The content SHOULD be 1439 diagnostic information such as a partial copy of the invalid message 1440 up to the maximum message size. An M_INVALID message MAY be silently 1441 ignored by a recipient. However, it could be used in support of 1442 extensibility, since it indicates that the remote node does not 1443 support a new or obsolete message or option. 1445 An M_INVALID message MUST NOT be sent in response to an M_INVALID 1446 message. 1448 2.8.13. No Operation Message 1450 In fragmentary CDDL, a No Operation message follows the pattern: 1452 noop-message = [M_NOOP] 1454 This message MAY be sent by an implementation that for practical 1455 reasons needs to initialize a socket. It MUST be silently ignored by 1456 a recipient. 1458 2.9. GRASP Options 1460 This section defines the GRASP options for the negotiation and 1461 synchronization protocol signaling. Additional options may be 1462 defined in the future. 1464 2.9.1. Format of GRASP Options 1466 GRASP options are CBOR objects that MUST start with an unsigned 1467 integer identifying the specific option type carried in this option. 1468 These option types are formally defined in Section 5. Apart from 1469 that the only format requirement is that each option MUST be a well- 1470 formed CBOR object. In general a CBOR array format is RECOMMENDED to 1471 limit overhead. 1473 GRASP options may be defined to include encapsulated GRASP options. 1475 2.9.2. Divert Option 1477 The Divert option is used to redirect a GRASP request to another 1478 node, which may be more appropriate for the intended negotiation or 1479 synchronization. It may redirect to an entity that is known as a 1480 specific negotiation or synchronization counterpart (on-link or off- 1481 link) or a default gateway. The divert option MUST only be 1482 encapsulated in Discovery Response messages. If found elsewhere, it 1483 SHOULD be silently ignored. 1485 A discovery initiator MAY ignore a Divert option if it only requires 1486 direct discovery responses. 1488 In fragmentary CDDL, the Divert option follows the pattern: 1490 divert-option = [O_DIVERT, +locator-option] 1492 The embedded Locator Option(s) (Section 2.9.5) point to diverted 1493 destination target(s) in response to a Discovery message. 1495 2.9.3. Accept Option 1497 The accept option is used to indicate to the negotiation counterpart 1498 that the proposed negotiation content is accepted. 1500 The accept option MUST only be encapsulated in Negotiation End 1501 messages. If found elsewhere, it SHOULD be silently ignored. 1503 In fragmentary CDDL, the Accept option follows the pattern: 1505 accept-option = [O_ACCEPT] 1507 2.9.4. Decline Option 1509 The decline option is used to indicate to the negotiation counterpart 1510 the proposed negotiation content is declined and end the negotiation 1511 process. 1513 The decline option MUST only be encapsulated in Negotiation End 1514 messages. If found elsewhere, it SHOULD be silently ignored. 1516 In fragmentary CDDL, the Decline option follows the pattern: 1518 decline-option = [O_DECLINE, ?reason] 1519 reason = text ;optional UTF-8 error message 1521 Note: there might be scenarios where an ASA wants to decline the 1522 proposed value and restart the negotiation process. In this case it 1523 is an implementation choice whether to send a Decline option or to 1524 continue with a Negotiate message, with an objective option that 1525 contains a null value, or one that contains a new value that might 1526 achieve convergence. 1528 2.9.5. Locator Options 1530 These locator options are used to present reachability information 1531 for an ASA, a device or an interface. They are Locator IPv6 Address 1532 Option, Locator IPv4 Address Option, Locator FQDN (Fully Qualified 1533 Domain Name) Option and URI (Uniform Resource Identifier) Option. 1535 Since ASAs will normally run as independent user programs, locator 1536 options need to indicate the network layer locator plus the transport 1537 protocol and port number for reaching the target. For this reason, 1538 the Locator Options for IP addresses and FQDNs include this 1539 information explicitly. In the case of the URI Option, this 1540 information can be encoded in the URI itself. 1542 Note: It is assumed that all locators used in locator options are in 1543 scope throughout the GRASP domain. As stated in Section 2.2, GRASP 1544 is not intended to work across disjoint addressing or naming realms. 1546 2.9.5.1. Locator IPv6 address option 1548 In fragmentary CDDL, the IPv6 address option follows the pattern: 1550 ipv6-locator-option = [O_IPv6_LOCATOR, ipv6-address, 1551 transport-proto, port-number] 1552 ipv6-address = bytes .size 16 1554 transport-proto = IPPROTO_TCP / IPPROTO_UDP 1555 IPPROTO_TCP = 6 1556 IPPROTO_UDP = 17 1557 port-number = 0..65535 1559 The content of this option is a binary IPv6 address followed by the 1560 protocol number and port number to be used. 1562 Note 1: The IPv6 address MUST normally have global scope. However, 1563 during initialization, a link-local address MAY be used for specific 1564 objectives only (Section 2.5.2). In this case the corresponding 1565 Discovery Response message MUST be sent via the interface to which 1566 the link-local address applies. 1568 Note 2: A link-local IPv6 address MUST NOT be used when this option 1569 is included in a Divert option. 1571 Note 3: The IPPROTO values are taken from the existing IANA Protocol 1572 Numbers registry in order to specify TCP or UDP. If GRASP requires 1573 future values that are not in that registry, a new registry for 1574 values outside the range 0..255 will be needed. 1576 2.9.5.2. Locator IPv4 address option 1578 In fragmentary CDDL, the IPv4 address option follows the pattern: 1580 ipv4-locator-option = [O_IPv4_LOCATOR, ipv4-address, 1581 transport-proto, port-number] 1582 ipv4-address = bytes .size 4 1584 The content of this option is a binary IPv4 address followed by the 1585 protocol number and port number to be used. 1587 Note: If an operator has internal network address translation for 1588 IPv4, this option MUST NOT be used within the Divert option. 1590 2.9.5.3. Locator FQDN option 1592 In fragmentary CDDL, the FQDN option follows the pattern: 1594 fqdn-locator-option = [O_FQDN_LOCATOR, text, 1595 transport-proto, port-number] 1597 The content of this option is the Fully Qualified Domain Name of the 1598 target followed by the protocol number and port number to be used. 1600 Note 1: Any FQDN which might not be valid throughout the network in 1601 question, such as a Multicast DNS name [RFC6762], MUST NOT be used 1602 when this option is used within the Divert option. 1604 Note 2: Normal GRASP operations are not expected to use this option. 1605 It is intended for special purposes such as discovering external 1606 services. 1608 2.9.5.4. Locator URI option 1610 In fragmentary CDDL, the URI option follows the pattern: 1612 uri-locator = [O_URI_LOCATOR, text, 1613 transport-proto / null, port-number / null] 1615 The content of this option is the Uniform Resource Identifier of the 1616 target followed by the protocol number and port number to be used (or 1617 by null values if not required) [RFC3986]. 1619 Note 1: Any URI which might not be valid throughout the network in 1620 question, such as one based on a Multicast DNS name [RFC6762], MUST 1621 NOT be used when this option is used within the Divert option. 1623 Note 2: Normal GRASP operations are not expected to use this option. 1624 It is intended for special purposes such as discovering external 1625 services. Therefore its use is not further described in this 1626 specification. 1628 2.10. Objective Options 1630 2.10.1. Format of Objective Options 1632 An objective option is used to identify objectives for the purposes 1633 of discovery, negotiation or synchronization. All objectives MUST be 1634 in the following format, described in fragmentary CDDL: 1636 objective = [objective-name, objective-flags, loop-count, ?objective-value] 1638 objective-name = text 1639 objective-value = any 1640 loop-count = 0..255 1642 All objectives are identified by a unique name which is a UTF-8 1643 string [RFC3629], to be compared byte by byte. 1645 The names of generic objectives MUST NOT include a colon (":") and 1646 MUST be registered with IANA (Section 6). 1648 The names of privately defined objectives MUST include at least one 1649 colon (":"). The string preceding the last colon in the name MUST be 1650 globally unique and in some way identify the entity or person 1651 defining the objective. The following three methods MAY be used to 1652 create such a globally unique string: 1654 1. The unique string is a decimal number representing a registered 1655 32 bit Private Enterprise Number (PEN) [RFC5612] that uniquely 1656 identifies the enterprise defining the objective. 1658 2. The unique string is a fully qualified domain name that uniquely 1659 identifies the entity or person defining the objective. 1661 3. The unique string is an email address that uniquely identifies 1662 the entity or person defining the objective. 1664 The GRASP protocol treats the objective name as an opaque string. 1665 For example, "EX1", "32473:EX1", "example.com:EX1", "example.org:EX1 1666 and "user@example.org:EX1" would be five different objectives. 1668 The 'objective-flags' field is described below. 1670 The 'loop-count' field is used for terminating negotiation as 1671 described in Section 2.8.7. It is also used for terminating 1672 discovery as described in Section 2.5.4, and for terminating flooding 1673 as described in Section 2.5.6.2. It is placed in the objective 1674 rather than in the GRASP message format because, as far as the ASA is 1675 concerned, it is a property of the objective itself. 1677 The 'objective-value' field is to express the actual value of a 1678 negotiation or synchronization objective. Its format is defined in 1679 the specification of the objective and may be a simple value or a 1680 data structure of any kind, as long as it can be represented in CBOR. 1681 It is optional because it is optional in a Discovery or Discovery 1682 Response message. 1684 2.10.2. Objective flags 1686 An objective may be relevant for discovery only, for discovery and 1687 negotiation, or for discovery and synchronization. This is expressed 1688 in the objective by logical flag bits: 1690 objective-flags = uint .bits objective-flag 1691 objective-flag = &( 1692 F_DISC: 0 ; valid for discovery 1693 F_NEG: 1 ; valid for negotiation 1694 F_SYNCH: 2 ; valid for synchronization 1695 F_NEG_DRY: 3 ; negotiation is dry-run 1696 ) 1698 These bits are independent and may be combined appropriately, e.g. 1699 (F_DISC and F_SYNCH) or (F_DISC and F_NEG) or (F_DISC and F_NEG and 1700 F_NEG_DRY). 1702 Note that for a given negotiation session, an objective must be 1703 either used for negotiation, or for dry-run negotiation. Mixing the 1704 two modes in a single negotiation is not possible. 1706 2.10.3. General Considerations for Objective Options 1708 As mentioned above, Objective Options MUST be assigned a unique name. 1709 As long as privately defined Objective Options obey the rules above, 1710 this document does not restrict their choice of name, but the entity 1711 or person concerned SHOULD publish the names in use. 1713 Names are expressed as UTF-8 strings for convenience in designing 1714 Objective Options for localized use. For generic usage, names 1715 expressed in the ASCII subset of UTF-8 are RECOMMENDED. Designers 1716 planning to use non-ASCII names are strongly advised to consult 1717 [RFC7564] or its successor to understand the complexities involved. 1718 Since the GRASP protocol compares names byte by byte, all issues of 1719 Unicode profiling and canonicalization MUST be specified in the 1720 design of the Objective Option. 1722 All Objective Options MUST respect the CBOR patterns defined above as 1723 "objective" and MUST replace the "any" field with a valid CBOR data 1724 definition for the relevant use case and application. 1726 An Objective Option that contains no additional fields beyond its 1727 "loop-count" can only be a discovery objective and MUST only be used 1728 in Discovery and Discovery Response messages. 1730 The Negotiation Objective Options contain negotiation objectives, 1731 which vary according to different functions/services. They MUST be 1732 carried by Discovery, Request Negotiation or Negotiation messages 1733 only. The negotiation initiator MUST set the initial "loop-count" to 1734 a value specified in the specification of the objective or, if no 1735 such value is specified, to GRASP_DEF_LOOPCT. 1737 For most scenarios, there should be initial values in the negotiation 1738 requests. Consequently, the Negotiation Objective options MUST 1739 always be completely presented in a Request Negotiation message, or 1740 in a Discovery message in rapid mode. If there is no initial value, 1741 the value field SHOULD be set to the 'null' value defined by CBOR. 1743 Synchronization Objective Options are similar, but MUST be carried by 1744 Discovery, Discovery Response, Request Synchronization, or Flood 1745 Synchronization messages only. They include value fields only in 1746 Synchronization or Flood Synchronization messages. 1748 The design of an objective interacts in various ways with the design 1749 of the ASAs that will use it. ASA design considerations are 1750 discussed in [I-D.carpenter-anima-asa-guidelines]. 1752 2.10.4. Organizing of Objective Options 1754 Generic objective options MUST be specified in documents available to 1755 the public and SHOULD be designed to use either the negotiation or 1756 the synchronization mechanism described above. 1758 As noted earlier, one negotiation objective is handled by each GRASP 1759 negotiation thread. Therefore, a negotiation objective, which is 1760 based on a specific function or action, SHOULD be organized as a 1761 single GRASP option. It is NOT RECOMMENDED to organize multiple 1762 negotiation objectives into a single option, nor to split a single 1763 function or action into multiple negotiation objectives. 1765 It is important to understand that GRASP negotiation does not support 1766 transactional integrity. If transactional integrity is needed for a 1767 specific objective, this must be ensured by the ASA. For example, an 1768 ASA might need to ensure that it only participates in one negotiation 1769 thread at the same time. Such an ASA would need to stop listening 1770 for incoming negotiation requests before generating an outgoing 1771 negotiation request. 1773 A synchronization objective SHOULD be organized as a single GRASP 1774 option. 1776 Some objectives will support more than one operational mode. An 1777 example is a negotiation objective with both a "dry run" mode (where 1778 the negotiation is to find out whether the other end can in fact make 1779 the requested change without problems) and a "live" mode, as 1780 explained in Section 2.5.5. The semantics of such modes will be 1781 defined in the specification of the objectives. These objectives 1782 SHOULD include flags indicating the applicable mode(s). 1784 An issue requiring particular attention is that GRASP itself is not a 1785 transactionally safe protocol. Any state associated with a dry run 1786 operation, such as temporarily reserving a resource for subsequent 1787 use in a live run, is entirely a matter for the designer of the ASA 1788 concerned. 1790 As indicated in Section 2.1, an objective's value may include 1791 multiple parameters. Parameters might be categorized into two 1792 classes: the obligatory ones presented as fixed fields; and the 1793 optional ones presented in some other form of data structure embedded 1794 in CBOR. The format might be inherited from an existing management 1795 or configuration protocol, with the objective option acting as a 1796 carrier for that format. The data structure might be defined in a 1797 formal language, but that is a matter for the specifications of 1798 individual objectives. There are many candidates, according to the 1799 context, such as ABNF, RBNF, XML Schema, YANG, etc. The GRASP 1800 protocol itself is agnostic on these questions. The only restriction 1801 is that the format can be mapped into CBOR. 1803 It is NOT RECOMMENDED to mix parameters that have significantly 1804 different response time characteristics in a single objective. 1805 Separate objectives are more suitable for such a scenario. 1807 All objectives MUST support GRASP discovery. However, as mentioned 1808 in Section 2.3, it is acceptable for an ASA to use an alternative 1809 method of discovery. 1811 Normally, a GRASP objective will refer to specific technical 1812 parameters as explained in Section 2.1. However, it is acceptable to 1813 define an abstract objective for the purpose of managing or 1814 coordinating ASAs. It is also acceptable to define a special-purpose 1815 objective for purposes such as trust bootstrapping or formation of 1816 the ACP. 1818 To guarantee convergence, a limited number of rounds or a timeout is 1819 needed for each negotiation objective. Therefore, the definition of 1820 each negotiation objective SHOULD clearly specify this, for example a 1821 default loop count and timeout, so that the negotiation can always be 1822 terminated properly. If not, the GRASP defaults will apply. 1824 There must be a well-defined procedure for concluding that a 1825 negotiation cannot succeed, and if so deciding what happens next 1826 (e.g., deadlock resolution, tie-breaking, or revert to best-effort 1827 service). This MUST be specified for individual negotiation 1828 objectives. 1830 2.10.5. Experimental and Example Objective Options 1832 The names "EX0" through "EX9" have been reserved for experimental 1833 options. Multiple names have been assigned because a single 1834 experiment may use multiple options simultaneously. These 1835 experimental options are highly likely to have different meanings 1836 when used for different experiments. Therefore, they SHOULD NOT be 1837 used without an explicit human decision and MUST NOT be used in 1838 unmanaged networks such as home networks. 1840 These names are also RECOMMENDED for use in documentation examples. 1842 3. Implementation Status [RFC Editor: please remove] 1844 Two prototype implementations of GRASP have been made. 1846 3.1. BUPT C++ Implementation 1848 o Name: BaseNegotiator.cpp, msg.cpp, Client.cpp, Server.cpp 1850 o Description: C++ implementation of GRASP core and API 1852 o Maturity: Prototype code, interoperable between Ubuntu. 1854 o Coverage: Corresponds to draft-carpenter-anima-gdn-protocol-03. 1855 Since it was implemented based on the old version draft, the most 1856 significant limitations comparing to current protocol design 1857 include: 1859 * Not support CBOR 1860 * Not support Flooding 1862 * Not support loop avoidance 1864 * only coded for IPv6, any IPv4 is accidental 1866 o Licensing: Huawei License. 1868 o Experience: https://github.com/liubingpang/IETF-Anima-Signaling- 1869 Protocol/blob/master/README.md 1871 o Contact: https://github.com/liubingpang/IETF-Anima-Signaling- 1872 Protocol 1874 3.2. Python Implementation 1876 o Name: graspy 1878 o Description: Python 3 implementation of GRASP core and API. 1880 o Maturity: Prototype code, interoperable between Windows 7 and 1881 Linux. 1883 o Coverage: Corresponds to draft-ietf-anima-grasp-13. Limitations 1884 include: 1886 * insecure: uses a dummy ACP module 1888 * only coded for IPv6, any IPv4 is accidental 1890 * FQDN and URI locators incompletely supported 1892 * no code for rapid mode 1894 * relay code is lazy (no rate control) 1896 * all unicast transactions use TCP (no unicast UDP). 1897 Experimental code for unicast UDP proved to be complex and 1898 brittle. 1900 * optional Objective option in Response messages not implemented 1902 * workarounds for defects in Python socket module and Windows 1903 socket peculiarities 1905 o Licensing: Simplified BSD 1906 o Experience: Tested on Windows, Linux and MacOS. 1907 https://www.cs.auckland.ac.nz/~brian/graspy/graspy.pdf 1909 o Contact: https://www.cs.auckland.ac.nz/~brian/graspy/ 1911 4. Security Considerations 1913 A successful attack on negotiation-enabled nodes would be extremely 1914 harmful, as such nodes might end up with a completely undesirable 1915 configuration that would also adversely affect their peers. GRASP 1916 nodes and messages therefore require full protection. As explained 1917 in Section 2.5.1, GRASP MUST run within a secure environment such as 1918 the Autonomic Control Plane [I-D.ietf-anima-autonomic-control-plane], 1919 except for the constrained instances described in Section 2.5.2. 1921 - Authentication 1923 A cryptographically authenticated identity for each device is 1924 needed in an autonomic network. It is not safe to assume that a 1925 large network is physically secured against interference or that 1926 all personnel are trustworthy. Each autonomic node MUST be 1927 capable of proving its identity and authenticating its messages. 1928 GRASP relies on a separate external certificate-based security 1929 mechanism to support authentication, data integrity protection, 1930 and anti-replay protection. 1932 Since GRASP must be deployed in an existing secure environment, 1933 the protocol itself specifies nothing concerning the trust anchor 1934 and certification authority. For example, in the Autonomic 1935 Control Plane [I-D.ietf-anima-autonomic-control-plane], all nodes 1936 can trust each other and the ASAs installed in them. 1938 If GRASP is used temporarily without an external security 1939 mechanism, for example during system bootstrap (Section 2.5.1), 1940 the Session ID (Section 2.7) will act as a nonce to provide 1941 limited protection against third parties injecting responses. A 1942 full analysis of the secure bootstrap process is in 1943 [I-D.ietf-anima-bootstrapping-keyinfra]. 1945 - Authorization and Roles 1947 The GRASP protocol is agnostic about the roles and capabilities of 1948 individual ASAs and about which objectives a particular ASA is 1949 authorized to support. An implementation might support 1950 precautions such as allowing only one ASA in a given node to 1951 modify a given objective, but this may not be appropriate in all 1952 cases. For example, it might be operationally useful to allow an 1953 old and a new version of the same ASA to run simultaneously during 1954 an overlap period. These questions are out of scope for the 1955 present specification. 1957 - Privacy and confidentiality 1959 GRASP is intended for network management purposes involving 1960 network elements, not end hosts. Therefore, no personal 1961 information is expected to be involved in the signaling protocol, 1962 so there should be no direct impact on personal privacy. 1963 Nevertheless, applications that do convey personal information 1964 cannot be excluded. Also, traffic flow paths, VPNs, etc. could be 1965 negotiated, which could be of interest for traffic analysis. 1966 Operators generally want to conceal details of their network 1967 topology and traffic density from outsiders. Therefore, since 1968 insider attacks cannot be excluded in a large network, the 1969 security mechanism for the protocol MUST provide message 1970 confidentiality. This is why Section 2.5.1 requires either an ACP 1971 or an alternative security mechanism. 1973 - Link-local multicast security 1975 GRASP has no reasonable alternative to using link-local multicast 1976 for Discovery or Flood Synchronization messages and these messages 1977 are sent in clear and with no authentication. They are only sent 1978 on interfaces within the autonomic network (see Section 2.1 and 1979 Section 2.5.1). They are however available to on-link 1980 eavesdroppers, and could be forged by on-link attackers. In the 1981 case of Discovery, the Discovery Responses are unicast and will 1982 therefore be protected (Section 2.5.1), and an untrusted forger 1983 will not be able to receive responses. In the case of Flood 1984 Synchronization, an on-link eavesdropper will be able to receive 1985 the flooded objectives but there is no response message to 1986 consider. Some precautions for Flood Synchronization messages are 1987 suggested in Section 2.5.6.2. 1989 - DoS Attack Protection 1991 GRASP discovery partly relies on insecure link-local multicast. 1992 Since routers participating in GRASP sometimes relay discovery 1993 messages from one link to another, this could be a vector for 1994 denial of service attacks. Some mitigations are specified in 1995 Section 2.5.4. However, malicious code installed inside the 1996 Autonomic Control Plane could always launch DoS attacks consisting 1997 of spurious discovery messages, or of spurious discovery 1998 responses. It is important that firewalls prevent any GRASP 1999 messages from entering the domain from an unknown source. 2001 - Security during bootstrap and discovery 2002 A node cannot trust GRASP traffic from other nodes until the 2003 security environment (such as the ACP) has identified the trust 2004 anchor and can authenticate traffic by validating certificates for 2005 other nodes. Also, until it has succesfully enrolled 2006 [I-D.ietf-anima-bootstrapping-keyinfra] a node cannot assume that 2007 other nodes are able to authenticate its own traffic. Therefore, 2008 GRASP discovery during the bootstrap phase for a new device will 2009 inevitably be insecure. Secure synchronization and negotiation 2010 will be impossible until enrollment is complete. Further details 2011 are given in Section 2.5.2. 2013 - Security of discovered locators 2015 When GRASP discovery returns an IP address, it MUST be that of a 2016 node within the secure environment (Section 2.5.1). If it returns 2017 an FQDN or a URI, the ASA that receives it MUST NOT assume that 2018 the target of the locator is within the secure environment. 2020 5. CDDL Specification of GRASP 2022 2023 grasp-message = (message .within message-structure) / noop-message 2025 message-structure = [MESSAGE_TYPE, session-id, ?initiator, 2026 *grasp-option] 2028 MESSAGE_TYPE = 0..255 2029 session-id = 0..4294967295 ;up to 32 bits 2030 grasp-option = any 2032 message /= discovery-message 2033 discovery-message = [M_DISCOVERY, session-id, initiator, objective] 2035 message /= response-message ;response to Discovery 2036 response-message = [M_RESPONSE, session-id, initiator, ttl, 2037 (+locator-option // divert-option), ?objective] 2039 message /= synch-message ;response to Synchronization request 2040 synch-message = [M_SYNCH, session-id, objective] 2042 message /= flood-message 2043 flood-message = [M_FLOOD, session-id, initiator, ttl, 2044 +[objective, (locator-option / [])]] 2046 message /= request-negotiation-message 2047 request-negotiation-message = [M_REQ_NEG, session-id, objective] 2049 message /= request-synchronization-message 2050 request-synchronization-message = [M_REQ_SYN, session-id, objective] 2052 message /= negotiation-message 2053 negotiation-message = [M_NEGOTIATE, session-id, objective] 2055 message /= end-message 2056 end-message = [M_END, session-id, accept-option / decline-option ] 2058 message /= wait-message 2059 wait-message = [M_WAIT, session-id, waiting-time] 2061 message /= invalid-message 2062 invalid-message = [M_INVALID, session-id, ?any] 2064 noop-message = [M_NOOP] 2066 divert-option = [O_DIVERT, +locator-option] 2068 accept-option = [O_ACCEPT] 2070 decline-option = [O_DECLINE, ?reason] 2071 reason = text ;optional UTF-8 error message 2073 waiting-time = 0..4294967295 ; in milliseconds 2074 ttl = 0..4294967295 ; in milliseconds 2076 locator-option /= [O_IPv4_LOCATOR, ipv4-address, 2077 transport-proto, port-number] 2078 ipv4-address = bytes .size 4 2080 locator-option /= [O_IPv6_LOCATOR, ipv6-address, 2081 transport-proto, port-number] 2082 ipv6-address = bytes .size 16 2084 locator-option /= [O_FQDN_LOCATOR, text, transport-proto, port-number] 2086 locator-option /= [O_URI_LOCATOR, text, 2087 transport-proto / null, port-number / null] 2089 transport-proto = IPPROTO_TCP / IPPROTO_UDP 2090 IPPROTO_TCP = 6 2091 IPPROTO_UDP = 17 2092 port-number = 0..65535 2094 initiator = ipv4-address / ipv6-address 2096 objective-flags = uint .bits objective-flag 2097 objective-flag = &( 2098 F_DISC: 0 ; valid for discovery 2099 F_NEG: 1 ; valid for negotiation 2100 F_SYNCH: 2 ; valid for synchronization 2101 F_NEG_DRY: 3 ; negotiation is dry-run 2102 ) 2104 objective = [objective-name, objective-flags, loop-count, ?objective-value] 2106 objective-name = text ;see section "Format of Objective Options" 2108 objective-value = any 2110 loop-count = 0..255 2112 ; Constants for message types and option types 2114 M_NOOP = 0 2115 M_DISCOVERY = 1 2116 M_RESPONSE = 2 2117 M_REQ_NEG = 3 2118 M_REQ_SYN = 4 2119 M_NEGOTIATE = 5 2120 M_END = 6 2121 M_WAIT = 7 2122 M_SYNCH = 8 2123 M_FLOOD = 9 2124 M_INVALID = 99 2126 O_DIVERT = 100 2127 O_ACCEPT = 101 2128 O_DECLINE = 102 2129 O_IPv6_LOCATOR = 103 2130 O_IPv4_LOCATOR = 104 2131 O_FQDN_LOCATOR = 105 2132 O_URI_LOCATOR = 106 2133 2135 6. IANA Considerations 2137 This document defines the GeneRic Autonomic Signaling Protocol 2138 (GRASP). 2140 Section 2.6 explains the following link-local multicast addresses, 2141 which IANA is requested to assign for use by GRASP: 2143 ALL_GRASP_NEIGHBORS multicast address (IPv6): (TBD1). Assigned in 2144 the IPv6 Link-Local Scope Multicast Addresses registry. 2146 ALL_GRASP_NEIGHBORS multicast address (IPv4): (TBD2). Assigned in 2147 the IPv4 Multicast Local Network Control Block. 2149 Section 2.6 explains the following User Port, which IANA is requested 2150 to assign for use by GRASP for both UDP and TCP: 2152 GRASP_LISTEN_PORT: (TBD3) 2153 Service Name: Generic Autonomic Signaling Protocol (GRASP) 2154 Transport Protocols: UDP, TCP 2155 Assignee: iesg@ietf.org 2156 Contact: chair@ietf.org 2157 Description: See Section 2.6 2158 Reference: RFC XXXX (this document) 2160 The IANA is requested to create a GRASP Parameter Registry including 2161 two registry tables. These are the GRASP Messages and Options 2162 Table and the GRASP Objective Names Table. 2164 GRASP Messages and Options Table. The values in this table are names 2165 paired with decimal integers. Future values MUST be assigned using 2166 the Standards Action policy defined by [RFC8126]. The following 2167 initial values are assigned by this document: 2169 M_NOOP = 0 2170 M_DISCOVERY = 1 2171 M_RESPONSE = 2 2172 M_REQ_NEG = 3 2173 M_REQ_SYN = 4 2174 M_NEGOTIATE = 5 2175 M_END = 6 2176 M_WAIT = 7 2177 M_SYNCH = 8 2178 M_FLOOD = 9 2179 M_INVALID = 99 2181 O_DIVERT = 100 2182 O_ACCEPT = 101 2183 O_DECLINE = 102 2184 O_IPv6_LOCATOR = 103 2185 O_IPv4_LOCATOR = 104 2186 O_FQDN_LOCATOR = 105 2187 O_URI_LOCATOR = 106 2189 GRASP Objective Names Table. The values in this table are UTF-8 2190 strings which MUST NOT include a colon (":"), according to 2191 Section 2.10.1. Future values MUST be assigned using the 2192 Specification Required policy defined by [RFC8126]. 2194 To assist expert review of a new objective, the specification should 2195 include a precise description of the format of the new objective, 2196 with sufficient explanation of its semantics to allow independent 2197 implementations. See Section 2.10.3 for more details. If the new 2198 objective is similar in name or purpose to a previously registered 2199 objective, the specification should explain why a new objective is 2200 justified. 2202 The following initial values are assigned by this document: 2204 EX0 2205 EX1 2206 EX2 2207 EX3 2208 EX4 2209 EX5 2210 EX6 2211 EX7 2212 EX8 2213 EX9 2215 7. Acknowledgements 2217 A major contribution to the original version of this document was 2218 made by Sheng Jiang and significant contributions were made by 2219 Toerless Eckert. Significant early review inputs were received from 2220 Joel Halpern, Barry Leiba, Charles E. Perkins, and Michael 2221 Richardson. William Atwood provided important assistance in 2222 debugging a prototype implementation. 2224 Valuable comments were received from Michael Behringer, Jeferson 2225 Campos Nobre, Laurent Ciavaglia, Zongpeng Du, Yu Fu, Joel Jaeggli, 2226 Zhenbin Li, Dimitri Papadimitriou, Pierre Peloso, Reshad Rahman, 2227 Markus Stenberg, Martin Stiemerling, Rene Struik, Martin Thomson, 2228 Dacheng Zhang, and participants in the NMRG research group, the ANIMA 2229 working group, and the IESG. 2231 8. References 2233 8.1. Normative References 2235 [I-D.greevenbosch-appsawg-cbor-cddl] 2236 Birkholz, H., Vigano, C., and C. Bormann, "CBOR data 2237 definition language (CDDL): a notational convention to 2238 express CBOR data structures", draft-greevenbosch-appsawg- 2239 cbor-cddl-10 (work in progress), March 2017. 2241 [I-D.ietf-anima-autonomic-control-plane] 2242 Behringer, M., Eckert, T., and S. Bjarnason, "An Autonomic 2243 Control Plane", draft-ietf-anima-autonomic-control- 2244 plane-06 (work in progress), March 2017. 2246 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2247 Requirement Levels", BCP 14, RFC 2119, 2248 DOI 10.17487/RFC2119, March 1997, 2249 . 2251 [RFC3629] Yergeau, F., "UTF-8, a transformation format of ISO 2252 10646", STD 63, RFC 3629, DOI 10.17487/RFC3629, November 2253 2003, . 2255 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 2256 Resource Identifier (URI): Generic Syntax", STD 66, 2257 RFC 3986, DOI 10.17487/RFC3986, January 2005, 2258 . 2260 [RFC4086] Eastlake 3rd, D., Schiller, J., and S. Crocker, 2261 "Randomness Requirements for Security", BCP 106, RFC 4086, 2262 DOI 10.17487/RFC4086, June 2005, 2263 . 2265 [RFC7049] Bormann, C. and P. Hoffman, "Concise Binary Object 2266 Representation (CBOR)", RFC 7049, DOI 10.17487/RFC7049, 2267 October 2013, . 2269 [RFC7217] Gont, F., "A Method for Generating Semantically Opaque 2270 Interface Identifiers with IPv6 Stateless Address 2271 Autoconfiguration (SLAAC)", RFC 7217, 2272 DOI 10.17487/RFC7217, April 2014, 2273 . 2275 [RFC8085] Eggert, L., Fairhurst, G., and G. Shepherd, "UDP Usage 2276 Guidelines", BCP 145, RFC 8085, DOI 10.17487/RFC8085, 2277 March 2017, . 2279 8.2. Informative References 2281 [I-D.carpenter-anima-asa-guidelines] 2282 Carpenter, B. and S. Jiang, "Guidelines for Autonomic 2283 Service Agents", draft-carpenter-anima-asa-guidelines-01 2284 (work in progress), January 2017. 2286 [I-D.chaparadza-intarea-igcp] 2287 Behringer, M., Chaparadza, R., Petre, R., Li, X., and H. 2288 Mahkonen, "IP based Generic Control Protocol (IGCP)", 2289 draft-chaparadza-intarea-igcp-00 (work in progress), July 2290 2011. 2292 [I-D.ietf-anima-bootstrapping-keyinfra] 2293 Pritikin, M., Richardson, M., Behringer, M., Bjarnason, 2294 S., and K. Watsen, "Bootstrapping Remote Secure Key 2295 Infrastructures (BRSKI)", draft-ietf-anima-bootstrapping- 2296 keyinfra-06 (work in progress), May 2017. 2298 [I-D.ietf-anima-reference-model] 2299 Behringer, M., Carpenter, B., Eckert, T., Ciavaglia, L., 2300 Pierre, P., Liu, B., Nobre, J., and J. Strassner, "A 2301 Reference Model for Autonomic Networking", draft-ietf- 2302 anima-reference-model-03 (work in progress), March 2017. 2304 [I-D.ietf-anima-stable-connectivity] 2305 Eckert, T. and M. Behringer, "Using Autonomic Control 2306 Plane for Stable Connectivity of Network OAM", draft-ietf- 2307 anima-stable-connectivity-02 (work in progress), February 2308 2017. 2310 [I-D.liu-anima-grasp-api] 2311 Carpenter, B., Liu, B., Wang, W., and X. Gong, "Generic 2312 Autonomic Signaling Protocol Application Program Interface 2313 (GRASP API)", draft-liu-anima-grasp-api-04 (work in 2314 progress), June 2017. 2316 [I-D.stenberg-anima-adncp] 2317 Stenberg, M., "Autonomic Distributed Node Consensus 2318 Protocol", draft-stenberg-anima-adncp-00 (work in 2319 progress), March 2015. 2321 [RFC2205] Braden, R., Ed., Zhang, L., Berson, S., Herzog, S., and S. 2322 Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1 2323 Functional Specification", RFC 2205, DOI 10.17487/RFC2205, 2324 September 1997, . 2326 [RFC2334] Luciani, J., Armitage, G., Halpern, J., and N. Doraswamy, 2327 "Server Cache Synchronization Protocol (SCSP)", RFC 2334, 2328 DOI 10.17487/RFC2334, April 1998, 2329 . 2331 [RFC2608] Guttman, E., Perkins, C., Veizades, J., and M. Day, 2332 "Service Location Protocol, Version 2", RFC 2608, 2333 DOI 10.17487/RFC2608, June 1999, 2334 . 2336 [RFC2865] Rigney, C., Willens, S., Rubens, A., and W. Simpson, 2337 "Remote Authentication Dial In User Service (RADIUS)", 2338 RFC 2865, DOI 10.17487/RFC2865, June 2000, 2339 . 2341 [RFC3315] Droms, R., Ed., Bound, J., Volz, B., Lemon, T., Perkins, 2342 C., and M. Carney, "Dynamic Host Configuration Protocol 2343 for IPv6 (DHCPv6)", RFC 3315, DOI 10.17487/RFC3315, July 2344 2003, . 2346 [RFC3416] Presuhn, R., Ed., "Version 2 of the Protocol Operations 2347 for the Simple Network Management Protocol (SNMP)", 2348 STD 62, RFC 3416, DOI 10.17487/RFC3416, December 2002, 2349 . 2351 [RFC3493] Gilligan, R., Thomson, S., Bound, J., McCann, J., and W. 2352 Stevens, "Basic Socket Interface Extensions for IPv6", 2353 RFC 3493, DOI 10.17487/RFC3493, February 2003, 2354 . 2356 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 2357 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 2358 DOI 10.17487/RFC4861, September 2007, 2359 . 2361 [RFC5612] Eronen, P. and D. Harrington, "Enterprise Number for 2362 Documentation Use", RFC 5612, DOI 10.17487/RFC5612, August 2363 2009, . 2365 [RFC5971] Schulzrinne, H. and R. Hancock, "GIST: General Internet 2366 Signalling Transport", RFC 5971, DOI 10.17487/RFC5971, 2367 October 2010, . 2369 [RFC6206] Levis, P., Clausen, T., Hui, J., Gnawali, O., and J. Ko, 2370 "The Trickle Algorithm", RFC 6206, DOI 10.17487/RFC6206, 2371 March 2011, . 2373 [RFC6241] Enns, R., Ed., Bjorklund, M., Ed., Schoenwaelder, J., Ed., 2374 and A. Bierman, Ed., "Network Configuration Protocol 2375 (NETCONF)", RFC 6241, DOI 10.17487/RFC6241, June 2011, 2376 . 2378 [RFC6733] Fajardo, V., Ed., Arkko, J., Loughney, J., and G. Zorn, 2379 Ed., "Diameter Base Protocol", RFC 6733, 2380 DOI 10.17487/RFC6733, October 2012, 2381 . 2383 [RFC6762] Cheshire, S. and M. Krochmal, "Multicast DNS", RFC 6762, 2384 DOI 10.17487/RFC6762, February 2013, 2385 . 2387 [RFC6763] Cheshire, S. and M. Krochmal, "DNS-Based Service 2388 Discovery", RFC 6763, DOI 10.17487/RFC6763, February 2013, 2389 . 2391 [RFC6887] Wing, D., Ed., Cheshire, S., Boucadair, M., Penno, R., and 2392 P. Selkirk, "Port Control Protocol (PCP)", RFC 6887, 2393 DOI 10.17487/RFC6887, April 2013, 2394 . 2396 [RFC7558] Lynn, K., Cheshire, S., Blanchet, M., and D. Migault, 2397 "Requirements for Scalable DNS-Based Service Discovery 2398 (DNS-SD) / Multicast DNS (mDNS) Extensions", RFC 7558, 2399 DOI 10.17487/RFC7558, July 2015, 2400 . 2402 [RFC7564] Saint-Andre, P. and M. Blanchet, "PRECIS Framework: 2403 Preparation, Enforcement, and Comparison of 2404 Internationalized Strings in Application Protocols", 2405 RFC 7564, DOI 10.17487/RFC7564, May 2015, 2406 . 2408 [RFC7575] Behringer, M., Pritikin, M., Bjarnason, S., Clemm, A., 2409 Carpenter, B., Jiang, S., and L. Ciavaglia, "Autonomic 2410 Networking: Definitions and Design Goals", RFC 7575, 2411 DOI 10.17487/RFC7575, June 2015, 2412 . 2414 [RFC7576] Jiang, S., Carpenter, B., and M. Behringer, "General Gap 2415 Analysis for Autonomic Networking", RFC 7576, 2416 DOI 10.17487/RFC7576, June 2015, 2417 . 2419 [RFC7787] Stenberg, M. and S. Barth, "Distributed Node Consensus 2420 Protocol", RFC 7787, DOI 10.17487/RFC7787, April 2016, 2421 . 2423 [RFC7788] Stenberg, M., Barth, S., and P. Pfister, "Home Networking 2424 Control Protocol", RFC 7788, DOI 10.17487/RFC7788, April 2425 2016, . 2427 [RFC8040] Bierman, A., Bjorklund, M., and K. Watsen, "RESTCONF 2428 Protocol", RFC 8040, DOI 10.17487/RFC8040, January 2017, 2429 . 2431 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2432 Writing an IANA Considerations Section in RFCs", BCP 26, 2433 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2434 . 2436 Appendix A. Open Issues [RFC Editor: This section should be empty. 2437 Please remove] 2439 o 68. (Placeholder) 2441 Appendix B. Closed Issues [RFC Editor: Please remove] 2443 o 1. UDP vs TCP: For now, this specification suggests UDP and TCP 2444 as message transport mechanisms. This is not clarified yet. UDP 2445 is good for short conversations, is necessary for multicast 2446 discovery, and generally fits the discovery and divert scenarios 2447 well. However, it will cause problems with large messages. TCP 2448 is good for stable and long sessions, with a little bit of time 2449 consumption during the session establishment stage. If messages 2450 exceed a reasonable MTU, a TCP mode will be required in any case. 2451 This question may be affected by the security discussion. 2453 RESOLVED by specifying UDP for short message and TCP for longer 2454 one. 2456 o 2. DTLS or TLS vs built-in security mechanism. For now, this 2457 specification has chosen a PKI based built-in security mechanism 2458 based on asymmetric cryptography. However, (D)TLS might be chosen 2459 as security solution to avoid duplication of effort. It also 2460 allows essentially similar security for short messages over UDP 2461 and longer ones over TCP. The implementation trade-offs are 2462 different. The current approach requires expensive asymmetric 2463 cryptographic calculations for every message. (D)TLS has startup 2464 overheads but cheaper crypto per message. DTLS is less mature 2465 than TLS. 2467 RESOLVED by specifying external security (ACP or (D)TLS). 2469 o The following open issues applied only if the original security 2470 model was retained: 2472 * 2.1. For replay protection, GRASP currently requires every 2473 participant to have an NTP-synchronized clock. Is this OK for 2474 low-end devices, and how does it work during device 2475 bootstrapping? We could take the Timestamp out of signature 2476 option, to become an independent and OPTIONAL (or RECOMMENDED) 2477 option. 2479 * 2.2. The Signature Option states that this option could be any 2480 place in a message. Wouldn't it be better to specify a 2481 position (such as the end)? That would be much simpler to 2482 implement. 2484 RESOLVED by changing security model. 2486 o 3. DoS Attack Protection needs work. 2488 RESOLVED by adding text. 2490 o 4. Should we consider preferring a text-based approach to 2491 discovery (after the initial discovery needed for bootstrapping)? 2492 This could be a complementary mechanism for multicast based 2493 discovery, especially for a very large autonomic network. 2494 Centralized registration could be automatically deployed 2495 incrementally. At the very first stage, the repository could be 2496 empty; then it could be filled in by the objectives discovered by 2497 different devices (for example using Dynamic DNS Update). The 2498 more records are stored in the repository, the less the multicast- 2499 based discovery is needed. However, if we adopt such a mechanism, 2500 there would be challenges: stateful solution, and security. 2502 RESOLVED for now by adding optional use of DNS-SD by ASAs. 2503 Subsequently removed by editors as irrelevant to GRASP istelf. 2505 o 5. Need to expand description of the minimum requirements for the 2506 specification of an individual discovery, synchronization or 2507 negotiation objective. 2509 RESOLVED for now by extra wording. 2511 o 6. Use case and protocol walkthrough. A description of how a 2512 node starts up, performs discovery, and conducts negotiation and 2513 synchronisation for a sample use case would help readers to 2514 understand the applicability of this specification. Maybe it 2515 should be an artificial use case or maybe a simple real one, based 2516 on a conceptual API. However, the authors have not yet decided 2517 whether to have a separate document or have it in the protocol 2518 document. 2520 RESOLVED: recommend a separate document. 2522 o 7. Cross-check against other ANIMA WG documents for consistency 2523 and gaps. 2525 RESOLVED: Satisfied by WGLC. 2527 o 8. Consideration of ADNCP proposal. 2529 RESOLVED by adding optional use of DNCP for flooding-type 2530 synchronization. 2532 o 9. Clarify how a GDNP instance knows whether it is running inside 2533 the ACP. (Sheng) 2535 RESOLVED by improved text. 2537 o 10. Clarify how a non-ACP GDNP instance initiates (D)TLS. 2538 (Sheng) 2540 RESOLVED by improved text and declaring DTLS out of scope for this 2541 draft. 2543 o 11. Clarify how UDP/TCP choice is made. (Sheng) [Like DNS? - 2544 Brian] 2546 RESOLVED by improved text. 2548 o 12. Justify that IP address within ACP or (D)TLS environment is 2549 sufficient to prove AN identity; or explain how Device Identity 2550 Option is used. (Sheng) 2552 RESOLVED for now: we assume that all ASAs in a device are trusted 2553 as soon as the device is trusted, so they share credentials. In 2554 that case the Device Identity Option is useless. This needs to be 2555 reviewed later. 2557 o 13. Emphasise that negotiation/synchronization are independent 2558 from discovery, although the rapid discovery mode includes the 2559 first step of a negotiation/synchronization. (Sheng) 2561 RESOLVED by improved text. 2563 o 14. Do we need an unsolicited flooding mechanism for discovery 2564 (for discovery results that everyone needs), to reduce scaling 2565 impact of flooding discovery messages? (Toerless) 2567 RESOLVED: Yes, added to requirements and solution. 2569 o 15. Do we need flag bits in Objective Options to distinguish 2570 distinguish Synchronization and Negotiation "Request" or rapid 2571 mode "Discovery" messages? (Bing) 2573 RESOLVED: yes, work on the API showed that these flags are 2574 essential. 2576 o 16. (Related to issue 14). Should we revive the "unsolicited 2577 Response" for flooding synchronisation data? This has to be done 2578 carefully due to the well-known issues with flooding, but it could 2579 be useful, e.g. for Intent distribution, where DNCP doesn't seem 2580 applicable. 2582 RESOLVED: Yes, see #14. 2584 o 17. Ensure that the discovery mechanism is completely proof 2585 against loops and protected against duplicate responses. 2587 RESOLVED: Added loop count mechanism. 2589 o 18. Discuss the handling of multiple valid discovery responses. 2591 RESOLVED: Stated that the choice must be available to the ASA but 2592 GRASP implementation should pick a default. 2594 o 19. Should we use a text-oriented format such as JSON/CBOR 2595 instead of native binary TLV format? 2597 RESOLVED: Yes, changed to CBOR. 2599 o 20. Is the Divert option needed? If a discovery response 2600 provides a valid IP address or FQDN, the recipient doesn't gain 2601 any extra knowledge from the Divert. On the other hand, the 2602 presence of Divert informs the receiver that the target is off- 2603 link, which might be useful sometimes. 2605 RESOLVED: Decided to keep Divert option. 2607 o 21. Rename the protocol as GRASP (GeneRic Autonomic Signaling 2608 Protocol)? 2610 RESOLVED: Yes, name changed. 2612 o 22. Does discovery mechanism scale robustly as needed? Need hop 2613 limit on relaying? 2615 RESOLVED: Added hop limit. 2617 o 23. Need more details on TTL for caching discovery responses. 2619 RESOLVED: Done. 2621 o 24. Do we need "fast withdrawal" of discovery responses? 2623 RESOLVED: This doesn't seem necessary. If an ASA exits or stops 2624 supporting a given objective, peers will fail to start future 2625 sessions and will simply repeat discovery. 2627 o 25. Does GDNP discovery meet the needs of multi-hop DNS-SD? 2629 RESOLVED: Decided not to consider this further as a GRASP protocol 2630 issue. GRASP objectives could embed DNS-SD formats if needed. 2632 o 26. Add a URL type to the locator options (for security bootstrap 2633 etc.) 2635 RESOLVED: Done, later renamed as URI. 2637 o 27. Security of Flood multicasts (Section 2.5.6.2). 2639 RESOLVED: added text. 2641 o 28. Does ACP support secure link-local multicast? 2643 RESOLVED by new text in the Security Considerations. 2645 o 29. PEN is used to distinguish vendor options. Would it be 2646 better to use a domain name? Anything unique will do. 2648 RESOLVED: Simplified this by removing PEN field and changing 2649 naming rules for objectives. 2651 o 30. Does response to discovery require randomized delays to 2652 mitigate amplification attacks? 2654 RESOLVED: WG feedback is that it's unnecessary. 2656 o 31. We have specified repeats for failed discovery etc. Is that 2657 sufficient to deal with sleeping nodes? 2659 RESOLVED: WG feedback is that it's unnecessary to say more. 2661 o 32. We have one-to-one synchronization and flooding 2662 synchronization. Do we also need selective flooding to a subset 2663 of nodes? 2664 RESOLVED: This will be discussed as a protocol extension in a 2665 separate draft (draft-liu-anima-grasp-distribution). 2667 o 33. Clarify if/when discovery needs to be repeated. 2669 RESOLVED: Done. 2671 o 34. Clarify what is mandatory for running in ACP, expand 2672 discussion of security boundary when running with no ACP - might 2673 rely on the local PKI infrastructure. 2675 RESOLVED: Done. 2677 o 35. State that role-based authorization of ASAs is out of scope 2678 for GRASP. GRASP doesn't recognize/handle any "roles". 2680 RESOLVED: Done. 2682 o 36. Reconsider CBOR definition for PEN syntax. ( objective-name 2683 = text / [pen, text] ; pen = uint ) 2685 RESOLVED: See issue 29. 2687 o 37. Are URI locators really needed? 2689 RESOLVED: Yes, e.g. for security bootstrap discovery, but added 2690 note that addresses are the normal case (same for FQDN locators). 2692 o 38. Is Session ID sufficient to identify relayed responses? 2693 Isn't the originator's address needed too? 2695 RESOLVED: Yes, this is needed for multicast messages and their 2696 responses. 2698 o 39. Clarify that a node will contain one GRASP instance 2699 supporting multiple ASAs. 2701 RESOLVED: Done. 2703 o 40. Add a "reason" code to the DECLINE option? 2705 RESOLVED: Done. 2707 o 41. What happens if an ASA cannot conveniently use one of the 2708 GRASP mechanisms? Do we (a) add a message type to GRASP, or (b) 2709 simply pass the discovery results to the ASA so that it can open 2710 its own socket? 2711 RESOLVED: Both would be possible, but (b) is preferred. 2713 o 42. Do we need a feature whereby an ASA can bypass the ACP and 2714 use the data plane for efficiency/throughput? This would require 2715 discovery to return non-ACP addresses and would evade ACP 2716 security. 2718 RESOLVED: This is considered out of scope for GRASP, but a comment 2719 has been added in security considerations. 2721 o 43. Rapid mode synchronization and negotiation is currently 2722 limited to a single objective for simplicity of design and 2723 implementation. A future consideration is to allow multiple 2724 objectives in rapid mode for greater efficiency. 2726 RESOLVED: This is considered out of scope for this version. 2728 o 44. In requirement T9, the words that encryption "may not be 2729 required in all deployments" were removed. Is that OK?. 2731 RESOLVED: No objections. 2733 o 45. Device Identity Option is unused. Can we remove it 2734 completely?. 2736 RESOLVED: No objections. Done. 2738 o 46. The 'initiator' field in DISCOVER, RESPONSE and FLOOD 2739 messages is intended to assist in loop prevention. However, we 2740 also have the loop count for that. Also, if we create a new 2741 Session ID each time a DISCOVER or FLOOD is relayed, that ID can 2742 be disambiguated by recipients. It would be simpler to remove the 2743 initiator from the messages, making parsing more uniform. Is that 2744 OK? 2746 RESOLVED: Yes. Done. 2748 o 47. REQUEST is a dual purpose message (request negotiation or 2749 request synchronization). Would it be better to split this into 2750 two different messages (and adjust various message names 2751 accordingly)? 2753 RESOLVED: Yes. Done. 2755 o 48. Should the Appendix "Capability Analysis of Current 2756 Protocols" be deleted before RFC publication? 2758 RESOLVED: No (per WG meeting at IETF 96). 2760 o 49. Section 2.5.1 Should say more about signaling between two 2761 autonomic networks/domains. 2763 RESOLVED: Description of separate GRASP instance added. 2765 o 50. Is Rapid mode limited to on-link only? What happens if first 2766 discovery responder does not support Rapid Mode? Section 2.5.5, 2767 Section 2.5.6) 2769 RESOLVED: Not limited to on-link. First responder wins. 2771 o 51. Should flooded objectives have a time-to-live before they are 2772 deleted from the flood cache? And should they be tagged in the 2773 cache with their source locator? 2775 RESOLVED: TTL added to Flood (and Discovery Response) messages. 2776 Cached flooded objectives must be tagged with their originating 2777 ASA locator, and multiple copies must be kept if necessary. 2779 o 52. Describe in detail what is allowed and disallowed in an 2780 insecure instance of GRASP. 2782 RESOLVED: Done. 2784 o 53. Tune IANA Considerations to support early assignment request. 2786 o 54. Is there a highly unlikely race condition if two peers 2787 simultaneously choose the same Session ID and send each other 2788 simultaneous M_REQ_NEG messages? 2790 RESOLVED: Yes. Enhanced text on Session ID generation, and added 2791 precaution when receiving a Request message. 2793 o 55. Could discovery be performed over TCP? 2795 RESOLVED: Unicast discovery added as an option. 2797 o 56. Change Session-ID to 32 bits? 2799 RESOLVED: Done. 2801 o 57. Add M_INVALID message? 2803 RESOLVED: Done. 2805 o 58. Maximum message size? 2806 RESOLVED by specifying default maximum message size (2048 bytes). 2808 o 59. Add F_NEG_DRY flag to specify a "dry run" objective?. 2810 RESOLVED: Done. 2812 o 60. Change M_FLOOD syntax to associate a locator with each 2813 objective? 2815 RESOLVED: Done. 2817 o 61. Is the SONN constrained instance really needed? 2819 RESOLVED: Retained but only as an option. 2821 o 62. Is it helpful to tag descriptive text with message names 2822 (M_DISCOVER etc.)? 2824 RESOLVED: Yes, done in various parts of the text. 2826 o 63. Should encryption be MUST instead of SHOULD in Section 2.5.1 2827 and Section 2.5.1? 2829 RESOLVED: Yes, MUST implement in both cases. 2831 o 64. Should more security text be moved from the main text into 2832 the Security Considerations? 2834 RESOLVED: No, on AD advice. 2836 o 65. Do we need to formally restrict Unicode characters allowed in 2837 objective names? 2839 RESOLVED: No, but need to point to guidance from PRECIS WG. 2841 o 66. Split requirements into separate document? 2843 RESOLVED: No, on AD advice. 2845 o 67. Remove normative dependency on draft-greevenbosch-appsawg- 2846 cbor-cddl? 2848 RESOLVED: No, on AD advice. In worst case, fix at AUTH48. 2850 Appendix C. Change log [RFC Editor: Please remove] 2852 draft-ietf-anima-grasp-14, 2017-07-02: 2854 Updates following additional IESG comments: 2856 Updated 2.5.1 and 2.5.2 based on IESG security feedback (specify 2857 dependency against security substrate). 2859 Strengthened requirement for reliable transport protocol. 2861 draft-ietf-anima-grasp-13, 2017-06-06: 2863 Updates following additional IESG comments: 2865 Removed all mention of TLS, including SONN, since it was under- 2866 specified. 2868 Clarified other text about trust and security model. 2870 Banned Rapid Mode when multicast is insecure. 2872 Explained use of M_INVALID to support extensibility 2874 Corrected details on discovery cache TTL and discovery timeout. 2876 Improved description of multicast UDP w.r.t. RFC8085. 2878 Clarified when transport connections are opened or closed. 2880 Noted that IPPROTO values come from the Protocol Numbers registry 2882 Protocol change: Added protocol and port numbers to URI locator. 2884 Removed inaccurate text about routing protocols 2886 Moved Requirements section to an Appendix. 2888 Other editorial and technical clarifications. 2890 draft-ietf-anima-grasp-12, 2017-05-19: 2892 Updates following IESG comments: 2894 Clarified that GRASP runs in a single addressing realm 2896 Improved wording about FQDN resolution, clarified that URI usage is 2897 out of scope. 2899 Clarified description of negotiation timeout. 2901 Noted that 'dry run' semantics are ASA-dependent 2903 Made the ACP a normative reference 2905 Clarified that LL multicasts are limited to GRASP interfaces 2907 Unicast UDP moved out of scope 2909 Editorial clarifications 2911 draft-ietf-anima-grasp-11, 2017-03-30: 2913 Updates following IETF 98 discussion: 2915 Encryption changed to a MUST implement. 2917 Pointed to guidance on UTF-8 names. 2919 draft-ietf-anima-grasp-10, 2017-03-10: 2921 Updates following IETF Last call: 2923 Protocol change: Specify that an objective with no initial value 2924 should have its value field set to CBOR 'null'. 2926 Protocol change: Specify behavior on receiving unrecognized message 2927 type. 2929 Noted that UTF-8 names are matched byte-for-byte. 2931 Added brief guidance for Expert Reviewer of new generic objectives. 2933 Numerous editorial improvements and clarifications and minor text 2934 rearrangements, none intended to change the meaning. 2936 draft-ietf-anima-grasp-09, 2016-12-15: 2938 Protocol change: Add F_NEG_DRY flag to specify a "dry run" objective. 2940 Protocol change: Change M_FLOOD syntax to associate a locator with 2941 each objective. 2943 Concentrated mentions of TLS in one section, with all details out of 2944 scope. 2946 Clarified text around constrained instances of GRASP. 2948 Strengthened text restricting LL addresses in locator options. 2950 Clarified description of rapid mode processsing. 2952 Specified that cached discovery results should not be returned on the 2953 same interface where they were learned. 2955 Shortened text in "High Level Design Choices" 2957 Dropped the word 'kernel' to avoid confusion with o/s kernel mode. 2959 Editorial improvements and clarifications. 2961 draft-ietf-anima-grasp-08, 2016-10-30: 2963 Protocol change: Added M_INVALID message. 2965 Protocol change: Increased Session ID space to 32 bits. 2967 Enhanced rules to avoid Session ID clashes. 2969 Corrected and completed description of timeouts for Request messages. 2971 Improved wording about exponential backoff and DoS. 2973 Clarified that discovery relaying is not done by limited security 2974 instances. 2976 Corrected and expanded explanation of port used for Discovery 2977 Response. 2979 Noted that Discovery message could be sent unicast in special cases. 2981 Added paragraph on extensibility. 2983 Specified default maximum message size. 2985 Added Appendix for sample messages. 2987 Added short protocol overview. 2989 Editorial fixes, including minor re-ordering for readability. 2991 draft-ietf-anima-grasp-07, 2016-09-13: 2993 Protocol change: Added TTL field to Flood message (issue 51). 2995 Protocol change: Added Locator option to Flood message (issue 51). 2997 Protocol change: Added TTL field to Discovery Response message 2998 (corrollary to issue 51). 3000 Clarified details of rapid mode (issues 43 and 50). 3002 Description of inter-domain GRASP instance added (issue 49). 3004 Description of limited security GRASP instances added (issue 52). 3006 Strengthened advice to use TCP rather than UDP. 3008 Updated IANA considerations and text about well-known port usage 3009 (issue 53). 3011 Amended text about ASA authorization and roles to allow for 3012 overlapping ASAs. 3014 Added text recommending that Flood should be repeated periodically. 3016 Editorial fixes. 3018 draft-ietf-anima-grasp-06, 2016-06-27: 3020 Added text on discovery cache timeouts. 3022 Noted that ASAs that are only initiators do not need to respond to 3023 discovery message. 3025 Added text on unexpected address changes. 3027 Added text on robust implementation. 3029 Clarifications and editorial fixes for numerous review comments 3031 Added open issues for some review comments. 3033 draft-ietf-anima-grasp-05, 2016-05-13: 3035 Noted in requirement T1 that it should be possible to implement ASAs 3036 independently as user space programs. 3038 Protocol change: Added protocol number and port to discovery 3039 response. Updated protocol description, CDDL and IANA considerations 3040 accordingly. 3042 Clarified that discovery and flood multicasts are handled by the 3043 GRASP core, not directly by ASAs. 3045 Clarified that a node may discover an objective without supporting it 3046 for synchronization or negotiation. 3048 Added Implementation Status section. 3050 Added reference to SCSP. 3052 Editorial fixes. 3054 draft-ietf-anima-grasp-04, 2016-03-11: 3056 Protocol change: Restored initiator field in certain messages and 3057 adjusted relaying rules to provide complete loop detection. 3059 Updated IANA Considerations. 3061 draft-ietf-anima-grasp-03, 2016-02-24: 3063 Protocol change: Removed initiator field from certain messages and 3064 adjusted relaying requirement to simplify loop detection. Also 3065 clarified narrative explanation of discovery relaying. 3067 Protocol change: Split Request message into two (Request Negotiation 3068 and Request Synchronization) and updated other message names for 3069 clarity. 3071 Protocol change: Dropped unused Device ID option. 3073 Further clarified text on transport layer usage. 3075 New text about multicast insecurity in Security Considerations. 3077 Various other clarifications and editorial fixes, including moving 3078 some material to Appendix. 3080 draft-ietf-anima-grasp-02, 2016-01-13: 3082 Resolved numerous issues according to WG discussions. 3084 Renumbered requirements, added D9. 3086 Protocol change: only allow one objective in rapid mode. 3088 Protocol change: added optional error string to DECLINE option. 3090 Protocol change: removed statement that seemed to say that a Request 3091 not preceded by a Discovery should cause a Discovery response. That 3092 made no sense, because there is no way the initiator would know where 3093 to send the Request. 3095 Protocol change: Removed PEN option from vendor objectives, changed 3096 naming rule accordingly. 3098 Protocol change: Added FLOOD message to simplify coding. 3100 Protocol change: Added SYNCH message to simplify coding. 3102 Protocol change: Added initiator id to DISCOVER, RESPONSE and FLOOD 3103 messages. But also allowed the relay process for DISCOVER and FLOOD 3104 to regenerate a Session ID. 3106 Protocol change: Require that discovered addresses must be global 3107 (except during bootstrap). 3109 Protocol change: Receiver of REQUEST message must close socket if no 3110 ASA is listening for the objective. 3112 Protocol change: Simplified Waiting message. 3114 Protocol change: Added No Operation message. 3116 Renamed URL locator type as URI locator type. 3118 Updated CDDL definition. 3120 Various other clarifications and editorial fixes. 3122 draft-ietf-anima-grasp-01, 2015-10-09: 3124 Updated requirements after list discussion. 3126 Changed from TLV to CBOR format - many detailed changes, added co- 3127 author. 3129 Tightened up loop count and timeouts for various cases. 3131 Noted that GRASP does not provide transactional integrity. 3133 Various other clarifications and editorial fixes. 3135 draft-ietf-anima-grasp-00, 2015-08-14: 3137 File name and protocol name changed following WG adoption. 3139 Added URL locator type. 3141 draft-carpenter-anima-gdn-protocol-04, 2015-06-21: 3143 Tuned wording around hierarchical structure. 3145 Changed "device" to "ASA" in many places. 3147 Reformulated requirements to be clear that the ASA is the main 3148 customer for signaling. 3150 Added requirement for flooding unsolicited synch, and added it to 3151 protocol spec. Recognized DNCP as alternative for flooding synch 3152 data. 3154 Requirements clarified, expanded and rearranged following design team 3155 discussion. 3157 Clarified that GDNP discovery must not be a prerequisite for GDNP 3158 negotiation or synchronization (resolved issue 13). 3160 Specified flag bits for objective options (resolved issue 15). 3162 Clarified usage of ACP vs TLS/DTLS and TCP vs UDP (resolved issues 3163 9,10,11). 3165 Updated DNCP description from latest DNCP draft. 3167 Editorial improvements. 3169 draft-carpenter-anima-gdn-protocol-03, 2015-04-20: 3171 Removed intrinsic security, required external security 3173 Format changes to allow DNCP co-existence 3175 Recognized DNS-SD as alternative discovery method. 3177 Editorial improvements 3179 draft-carpenter-anima-gdn-protocol-02, 2015-02-19: 3181 Tuned requirements to clarify scope, 3183 Clarified relationship between types of objective, 3185 Clarified that objectives may be simple values or complex data 3186 structures, 3188 Improved description of objective options, 3189 Added loop-avoidance mechanisms (loop count and default timeout, 3190 limitations on discovery relaying and on unsolicited responses), 3192 Allow multiple discovery objectives in one response, 3194 Provided for missing or multiple discovery responses, 3196 Indicated how modes such as "dry run" should be supported, 3198 Minor editorial and technical corrections and clarifications, 3200 Reorganized future work list. 3202 draft-carpenter-anima-gdn-protocol-01, restructured the logical flow 3203 of the document, updated to describe synchronization completely, add 3204 unsolicited responses, numerous corrections and clarifications, 3205 expanded future work list, 2015-01-06. 3207 draft-carpenter-anima-gdn-protocol-00, combination of draft-jiang- 3208 config-negotiation-ps-03 and draft-jiang-config-negotiation-protocol- 3209 02, 2014-10-08. 3211 Appendix D. Example Message Formats 3213 For readers unfamiliar with CBOR, this appendix shows a number of 3214 example GRASP messages conforming to the CDDL syntax given in 3215 Section 5. Each message is shown three times in the following 3216 formats: 3218 1. CBOR diagnostic notation. 3220 2. Similar, but showing the names of the constants. (Details of the 3221 flag bit encoding are omitted.) 3223 3. Hexadecimal version of the CBOR wire format. 3225 Long lines are split for display purposes only. 3227 D.1. Discovery Example 3229 The initiator (2001:db8:f000:baaa:28cc:dc4c:9703:6781) multicasts a 3230 discovery message looking for objective EX1: 3232 [1, 13948744, h'20010db8f000baaa28ccdc4c97036781', ["EX1", 5, 2, 0]] 3233 [M_DISCOVERY, 13948744, h'20010db8f000baaa28ccdc4c97036781', 3234 ["EX1", F_SYNCH_bits, 2, 0]] 3235 h'84011a00d4d7485020010db8f000baaa28ccdc4c970367818463455831050200' 3236 A peer (2001:0db8:f000:baaa:f000:baaa:f000:baaa) responds with a 3237 locator: 3239 [2, 13948744, h'20010db8f000baaa28ccdc4c97036781', 60000, 3240 [103, h'20010db8f000baaaf000baaaf000baaa', 6, 49443]] 3241 [M_RESPONSE, 13948744, h'20010db8f000baaa28ccdc4c97036781', 60000, 3242 [O_IPv6_LOCATOR, h'20010db8f000baaaf000baaaf000baaa', 3243 IPPROTO_TCP, 49443]] 3244 h'85021a00d4d7485020010db8f000baaa28ccdc4c9703678119ea6084186750 3245 20010db8f000baaaf000baaaf000baaa0619c123' 3247 D.2. Flood Example 3249 The initiator multicasts a flood message. The single objective has a 3250 null locator. There is no response: 3252 [9, 3504974, h'20010db8f000baaa28ccdc4c97036781', 10000, 3253 [["EX1", 5, 2, ["Example 1 value=", 100]],[] ] ] 3254 [M_FLOOD, 3504974, h'20010db8f000baaa28ccdc4c97036781', 10000, 3255 [["EX1", F_SYNCH_bits, 2, ["Example 1 value=", 100]],[] ] ] 3256 h'86091a00357b4e5020010db8f000baaa28ccdc4c97036781192710 3257 828463455831050282704578616d706c6520312076616c75653d186480' 3259 D.3. Synchronization Example 3261 Following successful discovery of objective EX2, the initiator 3262 unicasts a request: 3264 [4, 4038926, ["EX2", 5, 5, 0]] 3265 [M_REQ_SYN, 4038926, ["EX2", F_SYNCH_bits, 5, 0]] 3266 h'83041a003da10e8463455832050500' 3268 The peer responds with a value: 3270 [8, 4038926, ["EX2", 5, 5, ["Example 2 value=", 200]]] 3271 [M_SYNCH, 4038926, ["EX2", F_SYNCH_bits, 5, ["Example 2 value=", 200]]] 3272 h'83081a003da10e8463455832050582704578616d706c6520322076616c75653d18c8' 3274 D.4. Simple Negotiation Example 3276 Following successful discovery of objective EX3, the initiator 3277 unicasts a request: 3279 [3, 802813, ["EX3", 3, 6, ["NZD", 47]]] 3280 [M_REQ_NEG, 802813, ["EX3", F_NEG_bits, 6, ["NZD", 47]]] 3281 h'83031a000c3ffd8463455833030682634e5a44182f' 3282 The peer responds with immediate acceptance. Note that no objective 3283 is needed, because the initiator's request was accepted without 3284 change: 3286 [6, 802813, [101]] 3287 [M_END , 802813, [O_ACCEPT]] 3288 h'83061a000c3ffd811865' 3290 D.5. Complete Negotiation Example 3292 Again the initiator unicasts a request: 3294 [3, 13767778, ["EX3", 3, 6, ["NZD", 410]]] 3295 [M_REQ_NEG, 13767778, ["EX3", F_NEG_bits, 6, ["NZD", 410]]] 3296 h'83031a00d214628463455833030682634e5a4419019a' 3298 The responder starts to negotiate (making an offer): 3300 [5, 13767778, ["EX3", 3, 6, ["NZD", 80]]] 3301 [M_NEGOTIATE, 13767778, ["EX3", F_NEG_bits, 6, ["NZD", 80]]] 3302 h'83051a00d214628463455833030682634e5a441850' 3304 The initiator continues to negotiate (reducing its request, and note 3305 that the loop count is decremented): 3307 [5, 13767778, ["EX3", 3, 5, ["NZD", 307]]] 3308 [M_NEGOTIATE, 13767778, ["EX3", F_NEG_bits, 5, ["NZD", 307]]] 3309 h'83051a00d214628463455833030582634e5a44190133' 3311 The responder asks for more time: 3313 [7, 13767778, 34965] 3314 [M_WAIT, 13767778, 34965] 3315 h'83071a00d21462198895' 3317 The responder continues to negotiate (increasing its offer): 3319 [5, 13767778, ["EX3", 3, 4, ["NZD", 120]]] 3320 [M_NEGOTIATE, 13767778, ["EX3", F_NEG_bits, 4, ["NZD", 120]]] 3321 h'83051a00d214628463455833030482634e5a441878' 3323 The initiator continues to negotiate (reducing its request): 3325 [5, 13767778, ["EX3", 3, 3, ["NZD", 246]]] 3326 [M_NEGOTIATE, 13767778, ["EX3", F_NEG_bits, 3, ["NZD", 246]]] 3327 h'83051a00d214628463455833030382634e5a4418f6' 3329 The responder refuses to negotiate further: 3331 [6, 13767778, [102, "Insufficient funds"]] 3332 [M_END , 13767778, [O_DECLINE, "Insufficient funds"]] 3333 h'83061a00d2146282186672496e73756666696369656e742066756e6473' 3335 This negotiation has failed. If either side had sent [M_END, 3336 13767778, [O_ACCEPT]] it would have succeeded, converging on the 3337 objective value in the preceding M_NEGOTIATE. Note that apart from 3338 the initial M_REQ_NEG, the process is symmetrical. 3340 Appendix E. Requirement Analysis of Discovery, Synchronization and 3341 Negotiation 3343 This section discusses the requirements for discovery, negotiation 3344 and synchronization capabilities. The primary user of the protocol 3345 is an autonomic service agent (ASA), so the requirements are mainly 3346 expressed as the features needed by an ASA. A single physical device 3347 might contain several ASAs, and a single ASA might manage several 3348 technical objectives. If a technical objective is managed by several 3349 ASAs, any necessary coordination is outside the scope of the GRASP 3350 signaling protocol. Furthermore, requirements for ASAs themselves, 3351 such as the processing of Intent [RFC7575], are out of scope for the 3352 present document. 3354 E.1. Requirements for Discovery 3356 D1. ASAs may be designed to manage any type of configurable device 3357 or software, as required in Appendix E.2. A basic requirement is 3358 therefore that the protocol can represent and discover any kind of 3359 technical objective (as defined in Section 2.1) among arbitrary 3360 subsets of participating nodes. 3362 In an autonomic network we must assume that when a device starts up 3363 it has no information about any peer devices, the network structure, 3364 or what specific role it must play. The ASA(s) inside the device are 3365 in the same situation. In some cases, when a new application session 3366 starts up within a device, the device or ASA may again lack 3367 information about relevant peers. For example, it might be necessary 3368 to set up resources on multiple other devices, coordinated and 3369 matched to each other so that there is no wasted resource. Security 3370 settings might also need updating to allow for the new device or 3371 user. The relevant peers may be different for different technical 3372 objectives. Therefore discovery needs to be repeated as often as 3373 necessary to find peers capable of acting as counterparts for each 3374 objective that a discovery initiator needs to handle. From this 3375 background we derive the next three requirements: 3377 D2. When an ASA first starts up, it may have no knowledge of the 3378 specific network to which it is attached. Therefore the discovery 3379 process must be able to support any network scenario, assuming only 3380 that the device concerned is bootstrapped from factory condition. 3382 D3. When an ASA starts up, it must require no configured location 3383 information about any peers in order to discover them. 3385 D4. If an ASA supports multiple technical objectives, relevant peers 3386 may be different for different discovery objectives, so discovery 3387 needs to be performed separately to find counterparts for each 3388 objective. Thus, there must be a mechanism by which an ASA can 3389 separately discover peer ASAs for each of the technical objectives 3390 that it needs to manage, whenever necessary. 3392 D5. Following discovery, an ASA will normally perform negotiation or 3393 synchronization for the corresponding objectives. The design should 3394 allow for this by conveniently linking discovery to negotiation and 3395 synchronization. It may provide an optional mechanism to combine 3396 discovery and negotiation/synchronization in a single protocol 3397 exchange. 3399 D6. Some objectives may only be significant on the local link, but 3400 others may be significant across the routed network and require off- 3401 link operations. Thus, the relevant peers might be immediate 3402 neighbors on the same layer 2 link, or they might be more distant and 3403 only accessible via layer 3. The mechanism must therefore provide 3404 both on-link and off-link discovery of ASAs supporting specific 3405 technical objectives. 3407 D7. The discovery process should be flexible enough to allow for 3408 special cases, such as the following: 3410 o During initialization, a device must be able to establish mutual 3411 trust with autonomic nodes elsewhere in the network and 3412 participate in an authentication mechanism. Although this will 3413 inevitably start with a discovery action, it is a special case 3414 precisely because trust is not yet established. This topic is the 3415 subject of [I-D.ietf-anima-bootstrapping-keyinfra]. We require 3416 that once trust has been established for a device, all ASAs within 3417 the device inherit the device's credentials and are also trusted. 3418 This does not preclude the device having multiple credentials. 3420 o Depending on the type of network involved, discovery of other 3421 central functions might be needed, such as the Network Operations 3422 Center (NOC) [I-D.ietf-anima-stable-connectivity]. The protocol 3423 must be capable of supporting such discovery during 3424 initialization, as well as discovery during ongoing operation. 3426 D8. The discovery process must not generate excessive traffic and 3427 must take account of sleeping nodes. 3429 D9. There must be a mechanism for handling stale discovery results. 3431 E.2. Requirements for Synchronization and Negotiation Capability 3433 Autonomic networks need to be able to manage many different types of 3434 parameter and consider many dimensions, such as latency, load, unused 3435 or limited resources, conflicting resource requests, security 3436 settings, power saving, load balancing, etc. Status information and 3437 resource metrics need to be shared between nodes for dynamic 3438 adjustment of resources and for monitoring purposes. While this 3439 might be achieved by existing protocols when they are available, the 3440 new protocol needs to be able to support parameter exchange, 3441 including mutual synchronization, even when no negotiation as such is 3442 required. In general, these parameters do not apply to all 3443 participating nodes, but only to a subset. 3445 SN1. A basic requirement for the protocol is therefore the ability 3446 to represent, discover, synchronize and negotiate almost any kind of 3447 network parameter among selected subsets of participating nodes. 3449 SN2. Negotiation is an iterative request/response process that must 3450 be guaranteed to terminate (with success or failure). While tie- 3451 breaking rules must be defined specifically for each use case, the 3452 protocol should have some general mechanisms in support of loop and 3453 deadlock prevention, such as hop count limits or timeouts. 3455 SN3. Synchronization must be possible for groups of nodes ranging 3456 from small to very large. 3458 SN4. To avoid "reinventing the wheel", the protocol should be able 3459 to encapsulate the data formats used by existing configuration 3460 protocols (such as NETCONF/YANG) in cases where that is convenient. 3462 SN5. Human intervention in complex situations is costly and error- 3463 prone. Therefore, synchronization or negotiation of parameters 3464 without human intervention is desirable whenever the coordination of 3465 multiple devices can improve overall network performance. It follows 3466 that the protocol's resource requirements must be small enough to fit 3467 in any device that would otherwise need human intervention. The 3468 issue of running in constrained nodes is discussed in 3469 [I-D.ietf-anima-reference-model]. 3471 SN6. Human intervention in large networks is often replaced by use 3472 of a top-down network management system (NMS). It therefore follows 3473 that the protocol, as part of the Autonomic Networking 3474 Infrastructure, should be capable of running in any device that would 3475 otherwise be managed by an NMS, and that it can co-exist with an NMS, 3476 and with protocols such as SNMP and NETCONF. 3478 SN7. Specific autonomic features are expected to be implemented by 3479 individual ASAs, but the protocol must be general enough to allow 3480 them. Some examples follow: 3482 o Dependencies and conflicts: In order to decide upon a 3483 configuration for a given device, the device may need information 3484 from neighbors. This can be established through the negotiation 3485 procedure, or through synchronization if that is sufficient. 3486 However, a given item in a neighbor may depend on other 3487 information from its own neighbors, which may need another 3488 negotiation or synchronization procedure to obtain or decide. 3489 Therefore, there are potential dependencies and conflicts among 3490 negotiation or synchronization procedures. Resolving dependencies 3491 and conflicts is a matter for the individual ASAs involved. To 3492 allow this, there need to be clear boundaries and convergence 3493 mechanisms for negotiations. Also some mechanisms are needed to 3494 avoid loop dependencies or uncontrolled growth in a tree of 3495 dependencies. It is the ASA designer's responsibility to avoid or 3496 detect looping dependencies or excessive growth of dependency 3497 trees. The protocol's role is limited to bilateral signaling 3498 between ASAs, and the avoidance of loops during bilateral 3499 signaling. 3501 o Recovery from faults and identification of faulty devices should 3502 be as automatic as possible. The protocol's role is limited to 3503 discovery, synchronization and negotiation. These processes can 3504 occur at any time, and an ASA may need to repeat any of these 3505 steps when the ASA detects an event such as a negotiation 3506 counterpart failing. 3508 o Since a major goal is to minimize human intervention, it is 3509 necessary that the network can in effect "think ahead" before 3510 changing its parameters. One aspect of this is an ASA that relies 3511 on a knowledge base to predict network behavior. This is out of 3512 scope for the signaling protocol. However, another aspect is 3513 forecasting the effect of a change by a "dry run" negotiation 3514 before actually installing the change. Signaling a dry run is 3515 therefore a desirable feature of the protocol. 3517 Note that management logging, monitoring, alerts and tools for 3518 intervention are required. However, these can only be features of 3519 individual ASAs, not of the protocol itself. Another document 3520 [I-D.ietf-anima-stable-connectivity] discusses how such agents may be 3521 linked into conventional OAM systems via an Autonomic Control Plane 3522 [I-D.ietf-anima-autonomic-control-plane]. 3524 SN8. The protocol will be able to deal with a wide variety of 3525 technical objectives, covering any type of network parameter. 3526 Therefore the protocol will need a flexible and easily extensible 3527 format for describing objectives. At a later stage it may be 3528 desirable to adopt an explicit information model. One consideration 3529 is whether to adopt an existing information model or to design a new 3530 one. 3532 E.3. Specific Technical Requirements 3534 T1. It should be convenient for ASA designers to define new 3535 technical objectives and for programmers to express them, without 3536 excessive impact on run-time efficiency and footprint. In 3537 particular, it should be convenient for ASAs to be implemented 3538 independently of each other as user space programs rather than as 3539 kernel code, where such a programming model is possible. The classes 3540 of device in which the protocol might run is discussed in 3541 [I-D.ietf-anima-reference-model]. 3543 T2. The protocol should be easily extensible in case the initially 3544 defined discovery, synchronization and negotiation mechanisms prove 3545 to be insufficient. 3547 T3. To be a generic platform, the protocol payload format should be 3548 independent of the transport protocol or IP version. In particular, 3549 it should be able to run over IPv6 or IPv4. However, some functions, 3550 such as multicasting on a link, might need to be IP version 3551 dependent. By default, IPv6 should be preferred. 3553 T4. The protocol must be able to access off-link counterparts via 3554 routable addresses, i.e., must not be restricted to link-local 3555 operation. 3557 T5. It must also be possible for an external discovery mechanism to 3558 be used, if appropriate for a given technical objective. In other 3559 words, GRASP discovery must not be a prerequisite for GRASP 3560 negotiation or synchronization. 3562 T6. The protocol must be capable of distinguishing multiple 3563 simultaneous operations with one or more peers, especially when wait 3564 states occur. 3566 T7. Intent: Although the distribution of Intent is out of scope for 3567 this document, the protocol must not by design exclude its use for 3568 Intent distribution. 3570 T8. Management monitoring, alerts and intervention: Devices should 3571 be able to report to a monitoring system. Some events must be able 3572 to generate operator alerts and some provision for emergency 3573 intervention must be possible (e.g. to freeze synchronization or 3574 negotiation in a mis-behaving device). These features might not use 3575 the signaling protocol itself, but its design should not exclude such 3576 use. 3578 T9. Because this protocol may directly cause changes to device 3579 configurations and have significant impacts on a running network, all 3580 protocol exchanges need to be fully secured against forged messages 3581 and man-in-the middle attacks, and secured as much as reasonably 3582 possible against denial of service attacks. There must also be an 3583 encryption mechanism to resist unwanted monitoring. However, it is 3584 not required that the protocol itself provides these security 3585 features; it may depend on an existing secure environment. 3587 Appendix F. Capability Analysis of Current Protocols 3589 This appendix discusses various existing protocols with properties 3590 related to the requirements described in Appendix E. The purpose is 3591 to evaluate whether any existing protocol, or a simple combination of 3592 existing protocols, can meet those requirements. 3594 Numerous protocols include some form of discovery, but these all 3595 appear to be very specific in their applicability. Service Location 3596 Protocol (SLP) [RFC2608] provides service discovery for managed 3597 networks, but requires configuration of its own servers. DNS-SD 3598 [RFC6763] combined with mDNS [RFC6762] provides service discovery for 3599 small networks with a single link layer. [RFC7558] aims to extend 3600 this to larger autonomous networks but this is not yet standardized. 3601 However, both SLP and DNS-SD appear to target primarily application 3602 layer services, not the layer 2 and 3 objectives relevant to basic 3603 network configuration. Both SLP and DNS-SD are text-based protocols. 3605 Simple Network Management Protocol (SNMP) [RFC3416] uses a command/ 3606 response model not well suited for peer negotiation. Network 3607 Configuration Protocol (NETCONF) [RFC6241] uses an RPC model that 3608 does allow positive or negative responses from the target system, but 3609 this is still not adequate for negotiation. 3611 There are various existing protocols that have elementary negotiation 3612 abilities, such as Dynamic Host Configuration Protocol for IPv6 3613 (DHCPv6) [RFC3315], Neighbor Discovery (ND) [RFC4861], Port Control 3614 Protocol (PCP) [RFC6887], Remote Authentication Dial In User Service 3615 (RADIUS) [RFC2865], Diameter [RFC6733], etc. Most of them are 3616 configuration or management protocols. However, they either provide 3617 only a simple request/response model in a master/slave context or 3618 very limited negotiation abilities. 3620 There are some signaling protocols with an element of negotiation. 3621 For example Resource ReSerVation Protocol (RSVP) [RFC2205] was 3622 designed for negotiating quality of service parameters along the path 3623 of a unicast or multicast flow. RSVP is a very specialised protocol 3624 aimed at end-to-end flows. A more generic design is General Internet 3625 Signalling Transport (GIST) [RFC5971], but it is complex, tries to 3626 solve many problems, and is also aimed at per-flow signaling across 3627 many hops rather than at device-to-device signaling. However, we 3628 cannot completely exclude extended RSVP or GIST as a synchronization 3629 and negotiation protocol. They do not appear to be directly useable 3630 for peer discovery. 3632 RESTCONF [RFC8040] is a protocol intended to convey NETCONF 3633 information expressed in the YANG language via HTTP, including the 3634 ability to transit HTML intermediaries. While this is a powerful 3635 approach in the context of centralised configuration of a complex 3636 network, it is not well adapted to efficient interactive negotiation 3637 between peer devices, especially simple ones that might not include 3638 YANG processing already. 3640 The Distributed Node Consensus Protocol (DNCP) [RFC7787] is defined 3641 as a generic form of state synchronization protocol, with a proposed 3642 usage profile being the Home Networking Control Protocol (HNCP) 3643 [RFC7788] for configuring Homenet routers. A specific application of 3644 DNCP for autonomic networking was proposed in 3645 [I-D.stenberg-anima-adncp]. 3647 DNCP "is designed to provide a way for each participating node to 3648 publish a set of TLV (Type-Length-Value) tuples, and to provide a 3649 shared and common view about the data published... DNCP is most 3650 suitable for data that changes only infrequently... If constant rapid 3651 state changes are needed, the preferable choice is to use an 3652 additional point-to-point channel..." 3654 Specific features of DNCP include: 3656 o Every participating node has a unique node identifier. 3658 o DNCP messages are encoded as a sequence of TLV objects, sent over 3659 unicast UDP or TCP, with or without (D)TLS security. 3661 o Multicast is used only for discovery of DNCP neighbors when lower 3662 security is acceptable. 3664 o Synchronization of state is maintained by a flooding process using 3665 the Trickle algorithm. There is no bilateral synchronization or 3666 negotiation capability. 3668 o The HNCP profile of DNCP is designed to operate between directly 3669 connected neighbors on a shared link using UDP and link-local IPv6 3670 addresses. 3672 DNCP does not meet the needs of a general negotiation protocol, 3673 because it is designed specifically for flooding synchronization. 3674 Also, in its HNCP profile it is limited to link-local messages and to 3675 IPv6. However, at the minimum it is a very interesting test case for 3676 this style of interaction between devices without needing a central 3677 authority, and it is a proven method of network-wide state 3678 synchronization by flooding. 3680 The Server Cache Synchronization Protocol (SCSP) [RFC2334] also 3681 describes a method for cache synchronization and cache replication 3682 among a group of nodes. 3684 A proposal was made some years ago for an IP based Generic Control 3685 Protocol (IGCP) [I-D.chaparadza-intarea-igcp]. This was aimed at 3686 information exchange and negotiation but not directly at peer 3687 discovery. However, it has many points in common with the present 3688 work. 3690 None of the above solutions appears to completely meet the needs of 3691 generic discovery, state synchronization and negotiation in a single 3692 solution. Many of the protocols assume that they are working in a 3693 traditional top-down or north-south scenario, rather than a fluid 3694 peer-to-peer scenario. Most of them are specialized in one way or 3695 another. As a result, we have not identified a combination of 3696 existing protocols that meets the requirements in Appendix E. Also, 3697 we have not identified a path by which one of the existing protocols 3698 could be extended to meet the requirements. 3700 Authors' Addresses 3702 Carsten Bormann 3703 Universitaet Bremen TZI 3704 Postfach 330440 3705 D-28359 Bremen 3706 Germany 3708 Email: cabo@tzi.org 3709 Brian Carpenter (editor) 3710 Department of Computer Science 3711 University of Auckland 3712 PB 92019 3713 Auckland 1142 3714 New Zealand 3716 Email: brian.e.carpenter@gmail.com 3718 Bing Liu (editor) 3719 Huawei Technologies Co., Ltd 3720 Q14, Huawei Campus 3721 No.156 Beiqing Road 3722 Hai-Dian District, Beijing 100095 3723 P.R. China 3725 Email: leo.liubing@huawei.com