idnits 2.17.1 draft-ietf-homenet-dncp-10.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 204 has weird spacing: '...ntifier an o...' == Line 240 has weird spacing: '...on Time the (...' == Line 262 has weird spacing: '...e trust the ...' == Line 266 has weird spacing: '...y graph the...' == Line 270 has weird spacing: '...ionally a pe...' == (1 more instance...) -- The document date (September 21, 2015) is 3134 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) ** Downref: Normative reference to an Informational RFC: RFC 6234 ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126) -- Obsolete informational reference (is this intentional?): RFC 3315 (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 6347 (Obsoleted by RFC 9147) -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) Summary: 2 errors (**), 0 flaws (~~), 7 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Homenet Working Group M. Stenberg 3 Internet-Draft S. Barth 4 Intended status: Standards Track Independent 5 Expires: March 24, 2016 September 21, 2015 7 Distributed Node Consensus Protocol 8 draft-ietf-homenet-dncp-10 10 Abstract 12 This document describes the Distributed Node Consensus Protocol 13 (DNCP), a generic state synchronization protocol that uses the 14 Trickle algorithm and hash trees. DNCP is an abstract protocol, and 15 must be combined with a specific profile to make a complete 16 implementable protocol. 18 Status of This Memo 20 This Internet-Draft is submitted in full conformance with the 21 provisions of BCP 78 and BCP 79. 23 Internet-Drafts are working documents of the Internet Engineering 24 Task Force (IETF). Note that other groups may also distribute 25 working documents as Internet-Drafts. The list of current Internet- 26 Drafts is at http://datatracker.ietf.org/drafts/current/. 28 Internet-Drafts are draft documents valid for a maximum of six months 29 and may be updated, replaced, or obsoleted by other documents at any 30 time. It is inappropriate to use Internet-Drafts as reference 31 material or to cite them other than as "work in progress." 33 This Internet-Draft will expire on March 24, 2016. 35 Copyright Notice 37 Copyright (c) 2015 IETF Trust and the persons identified as the 38 document authors. All rights reserved. 40 This document is subject to BCP 78 and the IETF Trust's Legal 41 Provisions Relating to IETF Documents 42 (http://trustee.ietf.org/license-info) in effect on the date of 43 publication of this document. Please review these documents 44 carefully, as they describe your rights and restrictions with respect 45 to this document. Code Components extracted from this document must 46 include Simplified BSD License text as described in Section 4.e of 47 the Trust Legal Provisions and are provided without warranty as 48 described in the Simplified BSD License. 50 Table of Contents 52 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 53 1.1. Applicability . . . . . . . . . . . . . . . . . . . . . . 3 54 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 55 2.1. Requirements Language . . . . . . . . . . . . . . . . . . 6 56 3. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 7 57 4. Operation . . . . . . . . . . . . . . . . . . . . . . . . . . 8 58 4.1. Hash Tree . . . . . . . . . . . . . . . . . . . . . . . . 8 59 4.2. Data Transport . . . . . . . . . . . . . . . . . . . . . 8 60 4.3. Trickle-Driven Status Updates . . . . . . . . . . . . . . 9 61 4.4. Processing of Received TLVs . . . . . . . . . . . . . . . 10 62 4.5. Adding and Removing Peers . . . . . . . . . . . . . . . . 13 63 4.6. Data Liveliness Validation . . . . . . . . . . . . . . . 13 64 5. Data Model . . . . . . . . . . . . . . . . . . . . . . . . . 14 65 6. Optional Extensions . . . . . . . . . . . . . . . . . . . . . 16 66 6.1. Keep-Alives . . . . . . . . . . . . . . . . . . . . . . . 16 67 6.1.1. Data Model Additions . . . . . . . . . . . . . . . . 16 68 6.1.2. Per-Endpoint Periodic Keep-Alives . . . . . . . . . . 17 69 6.1.3. Per-Peer Periodic Keep-Alives . . . . . . . . . . . . 17 70 6.1.4. Received TLV Processing Additions . . . . . . . . . . 17 71 6.1.5. Peer Removal . . . . . . . . . . . . . . . . . . . . 17 72 6.2. Support For Dense Multicast-Enabled Links . . . . . . . . 18 73 7. Type-Length-Value Objects . . . . . . . . . . . . . . . . . . 19 74 7.1. Request TLVs . . . . . . . . . . . . . . . . . . . . . . 20 75 7.1.1. Request Network State TLV . . . . . . . . . . . . . . 20 76 7.1.2. Request Node State TLV . . . . . . . . . . . . . . . 20 77 7.2. Data TLVs . . . . . . . . . . . . . . . . . . . . . . . . 20 78 7.2.1. Node Endpoint TLV . . . . . . . . . . . . . . . . . . 20 79 7.2.2. Network State TLV . . . . . . . . . . . . . . . . . . 21 80 7.2.3. Node State TLV . . . . . . . . . . . . . . . . . . . 21 81 7.3. Data TLVs within Node State TLV . . . . . . . . . . . . . 22 82 7.3.1. Peer TLV . . . . . . . . . . . . . . . . . . . . . . 22 83 7.3.2. Keep-Alive Interval TLV . . . . . . . . . . . . . . . 23 84 8. Security and Trust Management . . . . . . . . . . . . . . . . 23 85 8.1. Pre-Shared Key Based Trust Method . . . . . . . . . . . . 23 86 8.2. PKI Based Trust Method . . . . . . . . . . . . . . . . . 24 87 8.3. Certificate Based Trust Consensus Method . . . . . . . . 24 88 8.3.1. Trust Verdicts . . . . . . . . . . . . . . . . . . . 24 89 8.3.2. Trust Cache . . . . . . . . . . . . . . . . . . . . . 25 90 8.3.3. Announcement of Verdicts . . . . . . . . . . . . . . 26 91 8.3.4. Bootstrap Ceremonies . . . . . . . . . . . . . . . . 27 92 9. DNCP Profile-Specific Definitions . . . . . . . . . . . . . . 28 93 10. Security Considerations . . . . . . . . . . . . . . . . . . . 29 94 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 30 95 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 31 96 12.1. Normative references . . . . . . . . . . . . . . . . . . 31 97 12.2. Informative references . . . . . . . . . . . . . . . . . 31 99 Appendix A. Alternative Modes of Operation . . . . . . . . . . . 31 100 A.1. Read-only Operation . . . . . . . . . . . . . . . . . . . 32 101 A.2. Forwarding Operation . . . . . . . . . . . . . . . . . . 32 102 Appendix B. DNCP Profile Additional Guidance . . . . . . . . . . 32 103 B.1. Unicast Transport - UDP or TCP? . . . . . . . . . . . . . 32 104 B.2. (Optional) Multicast Transport . . . . . . . . . . . . . 33 105 B.3. (Optional) Transport Security . . . . . . . . . . . . . . 33 106 Appendix C. Example Profile . . . . . . . . . . . . . . . . . . 33 107 Appendix D. Some Questions and Answers [RFC Editor: please 108 remove] . . . . . . . . . . . . . . . . . . . . . . 35 109 Appendix E. Changelog [RFC Editor: please remove] . . . . . . . 35 110 Appendix F. Draft Source [RFC Editor: please remove] . . . . . . 37 111 Appendix G. Acknowledgements . . . . . . . . . . . . . . . . . . 37 112 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 37 114 1. Introduction 116 DNCP is designed to provide a way for each participating node to 117 publish a small set of TLV (Type-Length-Value) tuples (at most 64 118 KB), and to provide a shared and common view about the data published 119 by every currently or recently bidirectionally reachable DNCP node in 120 a network. 122 For state synchronization a hash tree is used. It is formed by first 123 calculating a hash for the dataset published by each node, called 124 node data, and then calculating another hash over those node data 125 hashes. The single resulting hash, called network state hash, is 126 transmitted using the Trickle algorithm [RFC6206] to ensure that all 127 nodes share the same view of the current state of the published data 128 within the network. The use of Trickle with only short network state 129 hashes sent infrequently (in steady state, once the maximum Trickle 130 interval per link or unicast connection has been reached) makes DNCP 131 very thrifty when updates happen rarely. 133 For maintaining liveliness of the topology and the data within it, a 134 combination of Trickled network state, keep-alives, and "other" means 135 of ensuring reachability are used. The core idea is that if every 136 node ensures its peers are present, transitively, the whole network 137 state also stays up-to-date. 139 1.1. Applicability 141 DNCP is most suitable for data that changes only infrequently to gain 142 the maximum benefit from using Trickle. As the network of nodes 143 grows, or the frequency of data changes per node increases, Trickle 144 is eventually used less and less and the benefit of using DNCP 145 diminishes. In these cases Trickle just provides extra complexity 146 within the specification and little added value. 148 The suitability of DNCP for a particular application can roughly be 149 evaluated by considering the expected average network-wide state 150 change interval A_NC_I; it is computed by dividing the mean interval 151 at which a node originates a new TLV set by the number of 152 participating nodes. If keep-alives are used, A_NC_I is the minimum 153 of the computed A_NC_I and the keep-alive interval. If A_NC_I is 154 less than the (application-specific) Trickle minimum interval, DNCP 155 is most likely unsuitable for the application as Trickle will not be 156 utilized most of the time. 158 If constant rapid state changes are needed, the preferable choice is 159 to use an additional point-to-point channel whose address or locator 160 is published using DNCP. Nevertheless, if doing so does not raise 161 A_NC_I above the (sensibly chosen) Trickle interval parameters for a 162 particular application, using DNCP is probably not suitable for the 163 application. 165 Another consideration is the size of the published TLV set by a node 166 compared to the size of deltas in the TLV set. If the TLV set 167 published by a node is very large, and has frequent small changes, 168 DNCP as currently specified in this specification may be unsuitable 169 as it lacks a delta synchronization scheme to keep implementation 170 simple. 172 DNCP can be used in networks where only unicast transport is 173 available. While DNCP uses the least amount of bandwidth when 174 multicast is utilized, even in pure unicast mode, the use of Trickle 175 (ideally with k < 2) results in a protocol with an exponential 176 backoff timer and fewer transmissions than a simpler protocol not 177 using Trickle. 179 2. Terminology 181 DNCP profile the values for the set of parameters, given in 182 Section 9. They are prefixed with DNCP_ in this 183 document. The profile also specifies the set of 184 optional DNCP extensions to be used. For a simple 185 example DNCP profile, see Appendix C. 187 DNCP-based a protocol which provides a DNCP profile, according 188 protocol to Section 9, and zero or more TLV assignments from 189 the per-DNCP profile TLV registry as well as their 190 processing rules. 192 DNCP node a single node which runs a DNCP-based protocol. 194 Link a link-layer media over which directly connected 195 nodes can communicate. 197 DNCP network a set of DNCP nodes running DNCP-based protocol(s) 198 with matching DNCP profile(s). The set consists of 199 nodes that have discovered each other using the 200 transport method defined in the DNCP profile, via 201 multicast on local links, and / or by using unicast 202 communication. 204 Node identifier an opaque fixed-length identifier consisting of 205 DNCP_NODE_IDENTIFIER_LENGTH bytes which uniquely 206 identifies a DNCP node within a DNCP network. 208 Interface a node's attachment to a particular link. 210 Address an identifier used as source or destination of a 211 DNCP message flow, e.g., a tuple (IPv6 address, UDP 212 port) for an IPv6 UDP transport. 214 Endpoint a locally configured termination point for 215 (potential or established) DNCP message flows. An 216 endpoint is the source and destination for separate 217 unicast message flows to individual nodes and 218 optionally for multicast messages to all thereby 219 reachable nodes (e.g., for node discovery). 220 Endpoints are usually in one of the transport modes 221 specified in Section 4.2. 223 Endpoint a 32-bit opaque and locally unique value, which 224 identifier identifies a particular endpoint of a particular 225 DNCP node. The value 0 is reserved for DNCP and 226 DNCP-based protocol purposes and not used to 227 identify an actual endpoint. This definition is in 228 sync with the interface index definition in 229 [RFC3493], as the non-zero small positive integers 230 should comfortably fit within 32 bits. 232 Peer another DNCP node with which a DNCP node 233 communicates using a particular local and remote 234 endpoint pair. 236 Node data a set of TLVs published and owned by a node in the 237 DNCP network. Other nodes pass it along as-is, even 238 if they cannot fully interpret it. 240 Origination Time the (estimated) time when the node data set with 241 the current sequence number was published. 243 Node state a set of metadata attributes for node data. It 244 includes a sequence number for versioning, a hash 245 value for comparing equality of stored node data, 246 and a timestamp indicating the time passed since 247 its last publication (i.e., since the origination 248 time). The hash function and the length of the hash 249 value are defined in the DNCP profile. 251 Network state a hash value which represents the current state of 252 hash the network. The hash function and the length of 253 the hash value are defined in the DNCP profile. 254 Whenever a node is added, removed or updates its 255 published node data this hash value changes as 256 well. For calculation, please see Section 4.1. 258 Trust verdict a statement about the trustworthiness of a 259 certificate announced by a node participating in 260 the certificate based trust consensus mechanism. 262 Effective trust the trust verdict with the highest priority within 263 verdict the set of trust verdicts announced for the 264 certificate in the DNCP network. 266 Topology graph the undirected graph of DNCP nodes produced by 267 retaining only bidirectional peer relationships 268 between nodes. 270 Bidirectionally a peer is locally unidirectionally reachable if a 271 reachable recent and consistent multicast or any unicast DNCP 272 message has been received by the local node (see 273 Section 4.5). If said peer in return also 274 considers the local node unidirectionally 275 reachable, then bidirectionally reachability is 276 established. As this process is based on 277 publishing peer relationships and evaluating the 278 resulting topology graph as described in Section 279 4.6, this information is available to the whole 280 DNCP network. 282 Trickle Instance a distinct Trickle [RFC6206] algorithm state kept 283 by a node (Section 5) and related to an endpoint or 284 a particular (peer, endpoint) tuple with Trickle 285 variables I, t and c. See Section 4.3. 287 2.1. Requirements Language 289 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 290 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 291 "OPTIONAL" in this document are to be interpreted as described in RFC 292 2119 [RFC2119]. 294 3. Overview 296 DNCP operates primarily using unicast exchanges between nodes, and 297 may use multicast for Trickle-based shared state dissemination and 298 topology discovery. If used in pure unicast mode with unreliable 299 transport, Trickle is also used between peers. 301 DNCP discovers the topology of the nodes in the DNCP network and 302 maintains the liveliness of published node data by ensuring that the 303 publishing node is bidirectionally reachable. New potential peers 304 can be discovered autonomously on multicast-enabled links, their 305 addresses may be manually configured or they may be found by some 306 other means defined in the particular DNCP profile. The DNCP profile 307 may specify, for example, a well-known anycast address or 308 provisioning the remote address to contact via some other protocol 309 such as DHCPv6 [RFC3315]. 311 A hash tree of height 1, rooted in itself, is maintained by each node 312 to represent the state of all currently reachable nodes (see 313 Section 4.1) and the Trickle algorithm is used to trigger 314 synchronization (see Section 4.3). The need to check peer nodes for 315 state changes is thereby determined by comparing the current root of 316 their respective hash trees, i.e., their individually calculated 317 network state hashes. 319 Before joining a DNCP network, a node starts with a hash tree that 320 has only one leaf if the node publishes some TLVs, and no leaves 321 otherwise. It then announces the network state hash calculated from 322 the hash tree by means of the Trickle algorithm on all its configured 323 endpoints. 325 When an update is detected by a node (e.g., by receiving a different 326 network state hash from a peer) the originator of the event is 327 requested to provide a list of the state of all nodes, i.e., all the 328 information it uses to calculate its own hash tree. The node uses 329 the list to determine whether its own information is outdated and - 330 if necessary - requests the actual node data that has changed. 332 Whenever a node's local copy of any node data and its hash tree are 333 updated (e.g., due to its own or another node's node state changing 334 or due to a peer being added or removed) its Trickle instances are 335 reset which eventually causes any update to be propagated to all of 336 its peers. 338 4. Operation 340 4.1. Hash Tree 342 Each DNCP node maintains an arbitrary width hash tree of height 1. 343 Each leaf represents one bidirectionally reachable DNCP node (see 344 Section 4.6), and is represented by a tuple consisting of the node's 345 sequence number in network byte order concatenated with the hash- 346 value of the node's ordered node data published in the Node State TLV 347 (Section 7.2.3). These leaves are ordered in ascending order of the 348 node identifiers of the nodes they represent. The root of the tree - 349 the network state hash - is represented by the hash-value calculated 350 over all such leaf tuples concatenated in order. It is used to 351 determine whether the view of the network of two or more nodes is 352 consistent and shared. 354 The node data hashes in the leaves and the root network state hash 355 are updated on-demand and whenever any locally stored per-node state 356 changes. This includes local unidirectional reachability encoded in 357 the published Peer TLVs (Section 7.3.1) and - when combined with 358 remote data - results in awareness of bidirectional reachability 359 changes. 361 4.2. Data Transport 363 DNCP has few requirements for the underlying transport; it requires 364 some way of transmitting either unicast datagram or stream data to a 365 peer and, if used in multicast mode, a way of sending multicast 366 datagrams. As multicast is used only to identify potential new DNCP 367 nodes and to send status messages which merely notify that a unicast 368 exchange should be triggered, the multicast transport does not have 369 to be secured. If unicast security is desired and one of the built- 370 in security methods is to be used, support for some TLS-derived 371 transport scheme - such as TLS [RFC5246] on top of TCP or DTLS 372 [RFC6347] on top of UDP - is also required. They provide for 373 integrity protection and confidentiality of the node data, as well as 374 authentication and authorization using the schemes defined in 375 Security and Trust Management (Section 8). A specific definition of 376 the transport(s) in use and their parameters MUST be provided by the 377 DNCP profile. 379 TLVs are sent across the transport as is, and they SHOULD be sent 380 together where, e.g., MTU considerations do not recommend sending 381 them in multiple batches. TLVs in general are handled individually 382 and statelessly, with one exception: To form bidirectional peer 383 relationships DNCP requires identification of the endpoints used for 384 communication. As bidirectional peer relationships are required for 385 validating liveliness of published node data as described in 386 Section 4.6, a DNCP node MUST send a Node Endpoint TLV 387 (Section 7.2.1). When it is sent varies, depending on the underlying 388 transport, but conceptually it should be available whenever 389 processing a Network State TLV: 391 o If using a stream transport, the TLV MUST be sent at least once 392 per connection, but SHOULD NOT be sent more than once. 394 o If using a datagram transport, it MUST be included in every 395 datagram that also contains a Network State TLV (Section 7.2.2) 396 and MUST be located before any such TLV. It SHOULD also be 397 included in any other datagram, to speed up initial peer 398 detection. 400 Given the assorted transport options as well as potential endpoint 401 configuration, a DNCP endpoint may be used in various transport 402 modes: 404 Unicast: 406 * If only reliable unicast transport is used, Trickle is not used 407 at all. Where Trickle reset has been specified, a single 408 Network State TLV (Section 7.2.2) is sent instead to every 409 unicast peer. Additionally, recently changed Node State TLVs 410 (Section 7.2.3) MAY be included. 412 * If only unreliable unicast transport is used, Trickle state is 413 kept per peer and it is used to send Network State TLVs 414 intermittently, as specified in Section 4.3. 416 Multicast+Unicast: If multicast datagram transport is available on 417 an endpoint, Trickle state is only maintained for the endpoint as 418 a whole. It is used to send Network State TLVs periodically, as 419 specified in Section 4.3. Additionally, per-endpoint keep-alives 420 MAY be defined in the DNCP profile, as specified in Section 6.1.2. 422 MulticastListen+Unicast: Just like Unicast, except multicast 423 transmissions are listened to in order to detect changes of the 424 highest node identifier. This mode is used only if the DNCP 425 profile supports dense multicast-enabled link optimization 426 (Section 6.2). 428 4.3. Trickle-Driven Status Updates 430 The Trickle algorithm [RFC6206] has 3 parameters: Imin, Imax and k. 431 Imin and Imax represent the minimum value for I and the maximum 432 number of doublings of Imin, where I is the time interval during 433 which at least k Trickle updates must be seen on an endpoint to 434 prevent local state transmission. The actual suggested Trickle 435 algorithm parameters are DNCP profile specific, as described in 436 Section 9. 438 The Trickle state for all Trickle instances defined in Section 5 is 439 considered inconsistent and reset if and only if the locally 440 calculated network state hash changes. This occurs either due to a 441 change in the local node's own node data, or due to receipt of more 442 recent data from another node. A node MUST NOT reset its Trickle 443 state merely based on receiving a Network State TLV (Section 7.2.2) 444 with a network state hash which is different from its locally 445 calculated one. 447 Every time a particular Trickle instance indicates that an update 448 should be sent, the node MUST send a Network State TLV 449 (Section 7.2.2) if and only if: 451 o the endpoint is in Multicast+Unicast transport mode, in which case 452 the TLV MUST be sent over multicast. 454 o the endpoint is NOT in Multicast+Unicast transport mode, and the 455 unicast transport is unreliable, in which case the TLV MUST be 456 sent over unicast. 458 A (sub)set of all Node State TLVs (Section 7.2.3) MAY also be 459 included, unless it is defined as undesirable for some reason by the 460 DNCP profile, or to avoid exposure of the node state TLVs by 461 transmitting them within insecure multicast when using secure 462 unicast. 464 4.4. Processing of Received TLVs 466 This section describes how received TLVs are processed. The DNCP 467 profile may specify when to ignore particular TLVs, e.g., to modify 468 security properties - see Section 9 for what may be safely defined to 469 be ignored in a profile. Any 'reply' mentioned in the steps below 470 denotes sending of the specified TLV(s) over unicast to the 471 originator of the TLV being processed. If the TLV being replied to 472 was received via multicast and it was sent to a multiple access link, 473 the reply MUST be delayed by a random timespan in [0, Imin/2], to 474 avoid potential simultaneous replies that may cause problems on some 475 links, unless specified differently in the DNCP profile. Sending of 476 replies MAY also be rate-limited or omitted for a short period of 477 time by an implementation. However, if the TLV is not forbidden by 478 the DNCP profile, an implementation MUST reply to retransmissions of 479 the TLV with a non-zero probability to avoid starvation which would 480 break the state synchronization. 482 A DNCP node MUST process TLVs received from any valid (e.g., 483 correctly scoped) address, as specified by the DNCP profile and the 484 configuration of a particular endpoint, whether this address is known 485 to be the address of a peer or not. This provision satisfies the 486 needs of monitoring or other host software that needs to discover the 487 DNCP topology without adding to the state in the network. 489 Upon receipt of: 491 o Request Network State TLV (Section 7.1.1): The receiver MUST reply 492 with a Network State TLV (Section 7.2.2) and a Node State TLV 493 (Section 7.2.3) for each node data used to calculate the network 494 state hash. The Node State TLVs SHOULD NOT contain the optional 495 node data part to avoid redundant transmission of node data, 496 unless explicitly specified in the DNCP profile. 498 o Request Node State TLV (Section 7.1.2): If the receiver has node 499 data for the corresponding node, it MUST reply with a Node State 500 TLV (Section 7.2.3) for the corresponding node. The optional node 501 data part MUST be included in the TLV. 503 o Network State TLV (Section 7.2.2): If the network state hash 504 differs from the locally calculated network state hash, and the 505 receiver is unaware of any particular node state differences with 506 the sender, the receiver MUST reply with a Request Network State 507 TLV (Section 7.1.1). These replies MUST be rate limited to only 508 at most one reply per link per unique network state hash within 509 Imin. The simplest way to ensure this rate limit is a timestamp 510 indicating requests, and sending at most one Request Network State 511 TLV (Section 7.1.1) per Imin. To facilitate faster state 512 synchronization, if a Request Network State TLV is sent in a 513 reply, a local, current Network State TLV MAY also be sent. 515 o Node State TLV (Section 7.2.3): 517 * If the node identifier matches the local node identifier and 518 the TLV has a greater sequence number than its current local 519 value, or the same sequence number and a different hash, the 520 node SHOULD re-publish its own node data with a sequence number 521 significantly (e.g., 1000) greater than the received one, to 522 reclaim the node identifier. This difference is needed in 523 order to ensure that it is higher than any potentially 524 lingering copies of the node state in the network. This may 525 occur normally once due to the local node restarting and not 526 storing the most recently used sequence number. If this occurs 527 more than once or for nodes not re-publishing their own node 528 data, the DNCP profile MUST provide guidance on how to handle 529 these situations as it indicates the existence of another 530 active node with the same node identifier. 532 * If the node identifier does not match the local node 533 identifier, and one or more of the following conditions are 534 true: 536 + The local information is outdated for the corresponding node 537 (local sequence number is less than that within the TLV). 539 + The local information is potentially incorrect (local 540 sequence number matches but the node data hash differs). 542 + There is no data for that node altogether. 544 Then: 546 + If the TLV contains the Node Data field, it SHOULD also be 547 verified by ensuring that the locally calculated hash of the 548 Node Data matches the content of the H(Node Data) field 549 within the TLV. If they differ, the TLV SHOULD be ignored 550 and not processed further. 552 + If the TLV does not contain the Node Data field, and the 553 H(Node Data) field within the TLV differs from the local 554 node data hash for that node (or there is none), the 555 receiver MUST reply with a Request Node State TLV 556 (Section 7.1.2) for the corresponding node. 558 + Otherwise the receiver MUST update its locally stored state 559 for that node (node data based on Node Data field if 560 present, sequence number and relative time) to match the 561 received TLV. 563 For comparison purposes of the sequence number, a looping 564 comparison function MUST be used to avoid problems in case of 565 overflow. The comparison function a < b <=> ((a - b) % (2^32)) & 566 (2^31) != 0 where (a % b) represents the remainder of a modulo b 567 and (a & b) represents bitwise conjunction of a and b is 568 RECOMMENDED unless the DNCP profile defines another. 570 o Any other TLV: TLVs not recognized by the receiver MUST be 571 silently ignored unless they are sent within another TLV (for 572 example, TLVs within the Node Data field of a Node State TLV). 574 If secure unicast transport is configured for an endpoint, any Node 575 State TLVs received over insecure multicast MUST be silently ignored. 577 4.5. Adding and Removing Peers 579 When receiving a Node Endpoint TLV (Section 7.2.1) on an endpoint 580 from an unknown peer: 582 o If received over unicast, the remote node MUST be added as a peer 583 on the endpoint and a Peer TLV (Section 7.3.1) MUST be created for 584 it. 586 o If received over multicast, the node MAY be sent a (possibly rate- 587 limited) unicast Request Network State TLV (Section 7.1.1). 589 If keep-alives specified in Section 6.1 are NOT sent by the peer 590 (either the DNCP profile does not specify the use of keep-alives or 591 the particular peer chooses not to send keep-alives), some other 592 existing local transport-specific means (such as Ethernet carrier- 593 detection or TCP keep-alive) MUST be used to ensure its presence. If 594 the peer does not send keep-alives, and no means to verify presence 595 of the peer are available, the peer MUST be considered no longer 596 present and it SHOULD NOT be added back as a peer until it starts 597 sending keep-alives again. When the peer is no longer present, the 598 Peer TLV and the local DNCP peer state MUST be removed. 600 If the local endpoint is in the Multicast-Listen+Unicast transport 601 mode, a Peer TLV (Section 7.3.1) MUST NOT be published for the peers 602 not having the highest node identifier. 604 4.6. Data Liveliness Validation 606 The topology graph MUST be traversed either immediately or with a 607 small delay shorter than the DNCP profile-defined Trickle Imin, 608 whenever: 610 o A Peer TLV or a whole node is added or removed, or 612 o the origination time (in milliseconds) of some node's node data is 613 less than current time - 2^32 + 2^15. 615 The topology graph traversal starts with the local node marked as 616 reachable. Other nodes are then iteratively marked as reachable 617 using the following algorithm: A candidate not-yet-reachable node N 618 with an endpoint NE is marked as reachable if there is a reachable 619 node R with an endpoint RE that meet all of the following criteria: 621 o The origination time (in milliseconds) of R's node data is greater 622 than current time - 2^32 + 2^15. 624 o R publishes a Peer TLV with: 626 * Peer Node Identifier = N's node identifier 628 * Peer Endpoint Identifier = NE's endpoint identifier 630 * Endpoint Identifier = RE's endpoint identifier 632 o N publishes a Peer TLV with: 634 * Peer Node Identifier = R's node identifier 636 * Peer Endpoint Identifier = RE's endpoint identifier 638 * Endpoint Identifier = NE's endpoint identifier 640 The algorithm terminates, when no more candidate nodes fulfilling 641 these criteria can be found. 643 DNCP nodes that have not been reachable in the most recent topology 644 graph traversal MUST NOT be used for calculation of the network state 645 hash, be provided to any applications that need to use the whole TLV 646 graph, or be provided to remote nodes. They MAY be forgotten 647 immediately after the topology graph traversal, however it is 648 RECOMMENDED to keep them at least briefly to improve the speed of 649 DNCP network state convergence. This reduces the number of queries 650 needed to reconverge during both initial network convergence and when 651 a part of the network loses and regains bidirectional connectivity 652 within that time period. 654 5. Data Model 656 This section describes the local data structures a minimal 657 implementation might use. This section is provided only as a 658 convenience for the implementor. Some of the optional extensions 659 (Section 6) describe additional data requirements, and some optional 660 parts of the core protocol may also require more. 662 A DNCP node has: 664 o A data structure containing data about the most recently sent 665 Request Network State TLVs (Section 7.1.1). The simplest option 666 is keeping a timestamp of the most recent request (required to 667 fulfill reply rate limiting specified in Section 4.4). 669 A DNCP node has for every DNCP node in the DNCP network: 671 o Node identifier: the unique identifier of the node. The length, 672 how it is produced, and how collisions are handled, is up to the 673 DNCP profile. 675 o Node data: the set of TLV tuples published by that particular 676 node. As they are transmitted ordered (see Node State TLV 677 (Section 7.2.3) for details), maintaining the order within the 678 data structure here may be reasonable. 680 o Latest sequence number: the 32-bit sequence number that is 681 incremented any time the TLV set is published. The comparison 682 function used to compare them is described in Section 4.4. 684 o Origination time: the (estimated) time when the current TLV set 685 with the current sequence number was published. It is used to 686 populate the Milliseconds Since Origination field in a Node State 687 TLV (Section 7.2.3). Ideally it also has millisecond accuracy. 689 Additionally, a DNCP node has a set of endpoints for which DNCP is 690 configured to be used. For each such endpoint, a node has: 692 o Endpoint identifier: the 32-bit opaque locally unique value 693 identifying the endpoint within a node. It SHOULD NOT be reused 694 immediately after an endpoint is disabled. 696 o Trickle instance: the endpoint's Trickle instance with parameters 697 I, T, and c (only on an endpoint in Multicast+Unicast transport 698 mode). 700 and one (or more) of the following: 702 o Interface: the assigned local network interface. 704 o Unicast address: the DNCP node it should connect with. 706 o Set of addresses: the DNCP nodes from which connections are 707 accepted. 709 For each remote (peer, endpoint) pair detected on a local endpoint, a 710 DNCP node has: 712 o Node identifier: the unique identifier of the peer. 714 o Endpoint identifier: the unique endpoint identifier used by the 715 peer. 717 o Peer address: the most recently used address of the peer 718 (authenticated and authorized, if security is enabled). 720 o Trickle instance: the particular peer's Trickle instance with 721 parameters I, T, and c (only on an endpoint in Unicast mode, when 722 using an unreliable unicast transport) . 724 6. Optional Extensions 726 This section specifies extensions to the core protocol that a DNCP 727 profile may specify to be used. 729 6.1. Keep-Alives 731 Trickle-driven status updates (Section 4.3) provide a mechanism for 732 handling of new peer detection on an endpoint, as well as state 733 change notifications. Another mechanism may be needed to get rid of 734 old, no longer valid peers if the transport or lower layers do not 735 provide one. 737 If keep-alives are not specified in the DNCP profile, the rest of 738 this subsection MUST be ignored. 740 A DNCP profile MAY specify either per-endpoint (sent using multicast 741 to all DNCP nodes connected to a multicast-enabled link) or per-peer 742 (sent using unicast to each peer individually) keep-alive support. 744 For every endpoint that a keep-alive is specified for in the DNCP 745 profile, the endpoint-specific keep-alive interval MUST be 746 maintained. By default, it is DNCP_KEEPALIVE_INTERVAL. If there is 747 a local value that is preferred for that for any reason 748 (configuration, energy conservation, media type, ..), it can be 749 substituted instead. If a non-default keep-alive interval is used on 750 any endpoint, a DNCP node MUST publish appropriate Keep-Alive 751 Interval TLV(s) (Section 7.3.2) within its node data. 753 6.1.1. Data Model Additions 755 The following additions to the Data Model (Section 5) are needed to 756 support keep-alives: 758 For each configured endpoint that has per-endpoint keep-alives 759 enabled: 761 o Last sent: If a timestamp which indicates the last time a Network 762 State TLV (Section 7.2.2) was sent over that interface. 764 For each remote (peer, endpoint) pair detected on a local endpoint, a 765 DNCP node has: 767 o Last contact timestamp: a timestamp which indicates the last time 768 a consistent Network State TLV (Section 7.2.2) was received from 769 the peer over multicast, or anything was received over unicast. 770 When adding a new peer, it is initialized to the current time. 772 o Last sent: If per-peer keep-alives are enabled, a timestamp which 773 indicates the last time a Network State TLV (Section 7.2.2) was 774 sent to to that point-to-point peer. When adding a new peer, it 775 is initialized to the current time. 777 6.1.2. Per-Endpoint Periodic Keep-Alives 779 If per-endpoint keep-alives are enabled on an endpoint in 780 Multicast+Unicast transport mode, and if no traffic containing a 781 Network State TLV (Section 7.2.2) has been sent to a particular 782 endpoint within the endpoint-specific keep-alive interval, a Network 783 State TLV (Section 7.2.2) MUST be sent on that endpoint, and a new 784 Trickle interval started, as specified in the step 2 of Section 4.2 785 of [RFC6206]. The actual sending time SHOULD be further delayed by a 786 random timespan in [0, Imin/2]. 788 6.1.3. Per-Peer Periodic Keep-Alives 790 If per-peer keep-alives are enabled on a unicast-only endpoint, and 791 if no traffic containing a Network State TLV (Section 7.2.2) has been 792 sent to a particular peer within the endpoint-specific keep-alive 793 interval, a Network State TLV (Section 7.2.2) MUST be sent to the 794 peer, and a new Trickle interval started, as specified in the step 2 795 of Section 4.2 of [RFC6206]. 797 6.1.4. Received TLV Processing Additions 799 If a TLV is received over unicast from the peer, the Last contact 800 timestamp for the peer MUST be updated. 802 On receipt of a Network State TLV (Section 7.2.2) which is consistent 803 with the locally calculated network state hash, the Last contact 804 timestamp for the peer MUST be updated. 806 6.1.5. Peer Removal 808 For every peer on every endpoint, the endpoint-specific keep-alive 809 interval must be calculated by looking for Keep-Alive Interval TLVs 810 (Section 7.3.2) published by the node, and if none exist, using the 811 default value of DNCP_KEEPALIVE_INTERVAL. If the peer's last contact 812 timestamp has not been updated for at least locally chosen 813 potentially endpoint-specific keep-alive multiplier (defaults to 814 DNCP_KEEPALIVE_MULTIPLIER) times the peer's endpoint-specific keep- 815 alive interval, the Peer TLV for that peer and the local DNCP peer 816 state MUST be removed. 818 6.2. Support For Dense Multicast-Enabled Links 820 This optimization is needed to avoid a state space explosion. Given 821 a large set of DNCP nodes publishing data on an endpoint that uses 822 multicast on a link, every node will add a Peer TLV (Section 7.3.1) 823 for each peer. While Trickle limits the amount of traffic on the 824 link in stable state to some extent, the total amount of data that is 825 added to and maintained in the DNCP network given N nodes on a 826 multicast-enabled link is O(N^2). Additionally if per-peer keep- 827 alives are used, there will be O(N^2) keep-alives running on the link 828 if liveliness of peers is not ensured using some other way (e.g., TCP 829 connection lifetime, layer 2 notification, per-endpoint keep-alive). 831 An upper bound for the number of peers that are allowed for a 832 particular type of link that an endpoint in Multicast+Unicast 833 transport mode is used on SHOULD be provided by a DNCP profile, but 834 MAY also be chosen at runtime. The main consideration when selecting 835 a bound (if any) for a particular type of link should be whether it 836 supports multicast traffic, and whether a too large number of peers 837 case is likely to happen during the use of that DNCP profile on that 838 particular type of link. If neither is likely, there is little point 839 specifying support for this for that particular link type. 841 If a DNCP profile does not support this extension at all, the rest of 842 this subsection MUST be ignored. This is because when this extension 843 is used, the state within the DNCP network only contains a subset of 844 the full topology of the network. Therefore every node must be aware 845 of the potential of it being used in a particular DNCP profile. 847 If the specified upper bound is exceeded for some endpoint in 848 Multicast+Unicast transport mode and if the node does not have the 849 highest node identifier on the link, it SHOULD treat the endpoint as 850 a unicast endpoint connected to the node that has the highest node 851 identifier detected on the link, therefore transitioning to 852 Multicast-listen+Unicast transport mode. See Section 4.2 for 853 implications on the specific endpoint behavior. The nodes in 854 Multicast-listen+Unicast transport mode MUST keep listening to 855 multicast traffic to both receive messages from the node(s) still in 856 Multicast+Unicast mode, and as well to react to nodes with a greater 857 node identifier appearing. If the highest node identifier present on 858 the link changes, the remote unicast address of the endpoints in 859 Multicast-Listen+Unicast transport mode MUST be changed. If the node 860 identifier of the local node is the highest one, the node MUST switch 861 back to, or stay in Multicast+Unicast mode, and normally form peer 862 relationships with all peers. 864 7. Type-Length-Value Objects 866 0 1 2 3 867 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 868 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 869 | Type | Length | 870 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 871 | Value (if any) (+padding (if any)) | 872 .. 873 | (variable # of bytes) | 874 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 875 | (Optional nested TLVs) | 876 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 878 Each TLV is encoded as: 880 o a 2 byte Type field 882 o a 2 byte Length field which contains the length of the Value field 883 in bytes; 0 means no Value 885 o the Value itself (if any) 887 o padding bytes with value of zero up to the next 4 byte boundary if 888 the Length is not divisible by 4. 890 While padding bytes MUST NOT be included in the number stored in the 891 Length field of the TLV, if the TLV is enclosed within another TLV, 892 then the padding is included in the enclosing TLV's Length value. 894 Each TLV which does not define optional fields or variable-length 895 content MAY be sent with additional sub-TLVs appended after the TLV 896 to allow for extensibility. When handling such TLV types, each node 897 MUST accept received TLVs that are longer than the fixed fields 898 specified for the particular type, and ignore the sub-TLVs with 899 either unknown types, or not supported within that particular TLV 900 type. If any sub-TLVs are present, the Length field of the TLV 901 describes the number of bytes from the first byte of the TLV's own 902 Value (if any) to the last (padding) byte of the last sub-TLV. 904 For example, type=123 (0x7b) TLV with value 'x' (120 = 0x78) is 905 encoded as: 007B 0001 7800 0000. If it were to have sub-TLV of 906 type=124 (0x7c) with value 'y', it would be encoded as 007B 000C 7800 907 0000 007C 0001 7900 0000. 909 In this section, the following special notation is used: 911 .. = octet string concatenation operation. 913 H(x) = non-cryptographic hash function specified by DNCP profile. 915 7.1. Request TLVs 917 7.1.1. Request Network State TLV 919 0 1 2 3 920 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 921 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 922 | Type: REQ-NETWORK-STATE (1) | Length: >= 0 | 923 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 925 This TLV is used to request response with a Network State TLV 926 (Section 7.2.2) and all Node State TLVs (Section 7.2.3) (without node 927 data). 929 7.1.2. Request Node State TLV 931 0 1 2 3 932 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 933 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 934 | Type: REQ-NODE-STATE (2) | Length: > 0 | 935 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 936 | Node Identifier | 937 | (length fixed in DNCP profile) | 938 ... 939 | | 940 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 942 This TLV is used to request a Node State TLV (Section 7.2.3) 943 (including node data) for the node with the matching node identifier. 945 7.2. Data TLVs 947 7.2.1. Node Endpoint TLV 949 0 1 2 3 950 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 951 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 952 | Type: NODE-ENDPOINT (3) | Length: > 4 | 953 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 954 | Node Identifier | 955 | (length fixed in DNCP profile) | 956 ... 957 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 958 | Endpoint Identifier | 959 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 960 This TLV identifies both the local node's node identifier, as well as 961 the particular endpoint's endpoint identifier. Section 4.2 specifies 962 when it is sent. 964 7.2.2. Network State TLV 966 0 1 2 3 967 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 968 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 969 | Type: NETWORK-STATE (4) | Length: > 0 | 970 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 971 | H(sequence number of node 1 .. H(node data of node 1) .. | 972 | .. sequence number of node N .. H(node data of node N)) | 973 | (length fixed in DNCP profile) | 974 ... 975 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 977 This TLV contains the current locally calculated network state hash, 978 see Section 4.1 for how it is calculated. 980 7.2.3. Node State TLV 982 0 1 2 3 983 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 984 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 985 | Type: NODE-STATE (5) | Length: > 8 | 986 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 987 | Node Identifier | 988 | (length fixed in DNCP profile) | 989 ... 990 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 991 | Sequence Number | 992 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 993 | Milliseconds Since Origination | 994 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 995 | H(Node Data) | 996 | (length fixed in DNCP profile) | 997 ... 998 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 999 | (optionally) Node Data (a set of nested TLVs) | 1000 ... 1001 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1003 This TLV represents the local node's knowledge about the published 1004 state of a node in the DNCP network identified by the Node Identifier 1005 field in the TLV. 1007 Every node, including the node publishing the node data, MUST update 1008 the Milliseconds Since Origination whenever it sends a Node State TLV 1009 based on when the node estimates the data was originally published. 1010 This is, e.g., to ensure that any relative timestamps contained 1011 within the published node data can be correctly offset and 1012 interpreted. Ultimately, what is provided is just an approximation, 1013 as transmission delays are not accounted for. 1015 Absent any changes, if the originating node notices that the 32-bit 1016 milliseconds since origination value would be close to overflow 1017 (greater than 2^32-2^16), the node MUST re-publish its TLVs even if 1018 there is no change. In other words, absent any other changes, the 1019 TLV set MUST be re-published roughly every 48 days. 1021 The actual node data of the node may be included within the TLV as 1022 well in the optional Node Data field. The set of TLVs MUST be 1023 strictly ordered based on ascending binary content (including TLV 1024 type and length). This enables, e.g., efficient state delta 1025 processing and no-copy indexing by TLV type by the recipient. The 1026 Node Data content MUST be passed along exactly as it was received. 1027 It SHOULD be also verified on receipt that the locally calculated 1028 H(Node Data) matches the content of the field within the TLV, and if 1029 the hash differs, the TLV SHOULD be ignored. 1031 7.3. Data TLVs within Node State TLV 1033 These TLVs are published by the DNCP nodes, and therefore only 1034 encoded in the Node Data field of Node State TLVs. If encountered 1035 outside Node State TLV, they MUST be silently ignored. 1037 7.3.1. Peer TLV 1039 0 1 2 3 1040 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1041 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1042 | Type: PEER (8) | Length: > 8 | 1043 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1044 | Peer Node Identifier | 1045 | (length fixed in DNCP profile) | 1046 ... 1047 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1048 | Peer Endpoint Identifier | 1049 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1050 | (Local) Endpoint Identifier | 1051 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1053 This TLV indicates that the node in question vouches that the 1054 specified peer is reachable by it on the specified local endpoint. 1056 The presence of this TLV at least guarantees that the node publishing 1057 it has received traffic from the peer recently. For guaranteed up- 1058 to-date bidirectional reachability, the existence of both nodes' 1059 matching Peer TLVs needs to be checked. 1061 7.3.2. Keep-Alive Interval TLV 1063 0 1 2 3 1064 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1065 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1066 | Type: KEEP-ALIVE-INTERVAL (9) | Length: >= 8 | 1067 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1068 | Endpoint Identifier | 1069 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1070 | Interval | 1071 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1073 This TLV indicates a non-default interval being used to send keep- 1074 alives specified in Section 6.1. 1076 Endpoint identifier is used to identify the particular endpoint for 1077 which the interval applies. If 0, it applies for ALL endpoints for 1078 which no specific TLV exists. 1080 Interval specifies the interval in milliseconds at which the node 1081 sends keep-alives. A value of zero means no keep-alives are sent at 1082 all; in that case, some lower layer mechanism that ensures presence 1083 of nodes MUST be available and used. 1085 8. Security and Trust Management 1087 If specified in the DNCP profile, either DTLS [RFC6347] or TLS 1088 [RFC5246] may be used to authenticate and encrypt either some (if 1089 specified optional in the profile), or all unicast traffic. The 1090 following methods for establishing trust are defined, but it is up to 1091 the DNCP profile to specify which ones may, should or must be 1092 supported. 1094 8.1. Pre-Shared Key Based Trust Method 1096 A PSK-based trust model is a simple security management mechanism 1097 that allows an administrator to deploy devices to an existing network 1098 by configuring them with a pre-defined key, similar to the 1099 configuration of an administrator password or WPA-key. Although 1100 limited in nature it is useful to provide a user-friendly security 1101 mechanism for smaller networks. 1103 8.2. PKI Based Trust Method 1105 A PKI-based trust-model enables more advanced management capabilities 1106 at the cost of increased complexity and bootstrapping effort. It 1107 however allows trust to be managed in a centralized manner and is 1108 therefore useful for larger networks with a need for an authoritative 1109 trust management. 1111 8.3. Certificate Based Trust Consensus Method 1113 For some scenarios - such as bootstrapping a mostly unmanaged network 1114 - the methods described above may not provide a desirable tradeoff 1115 between security and user experience. This section includes guidance 1116 for implementing an opportunistic security [RFC7435] method which 1117 DNCP profiles can build upon and adapt for their specific 1118 requirements. 1120 The certificate-based consensus model is designed to be a compromise 1121 between trust management effort and flexibility. It is based on 1122 X.509-certificates and allows each DNCP node to provide a trust 1123 verdict on any other certificate and a consensus is found to 1124 determine whether a node using this certificate or any certificate 1125 signed by it is to be trusted. 1127 A DNCP node not using this security method MUST ignore all announced 1128 trust verdicts and MUST NOT announce any such verdicts by itself, 1129 i.e., any other normative language in this subsection does not apply 1130 to it. 1132 The current effective trust verdict for any certificate is defined as 1133 the one with the highest priority from all trust verdicts announced 1134 for said certificate at the time. 1136 8.3.1. Trust Verdicts 1138 Trust verdicts are statements of DNCP nodes about the trustworthiness 1139 of X.509-certificates. There are 5 possible trust verdicts in order 1140 of ascending priority: 1142 0 (Neutral): no trust verdict exists but the DNCP network should 1143 determine one. 1145 1 (Cached Trust): the last known effective trust verdict was 1146 Configured or Cached Trust. 1148 2 (Cached Distrust): the last known effective trust verdict was 1149 Configured or Cached Distrust. 1151 3 (Configured Trust): trustworthy based upon an external ceremony 1152 or configuration. 1154 4 (Configured Distrust): not trustworthy based upon an external 1155 ceremony or configuration. 1157 Trust verdicts are differentiated in 3 groups: 1159 o Configured verdicts are used to announce explicit trust verdicts a 1160 node has based on any external trust bootstrap or predefined 1161 relation a node has formed with a given certificate. 1163 o Cached verdicts are used to retain the last known trust state in 1164 case all nodes with configured verdicts about a given certificate 1165 have been disconnected or turned off. 1167 o The Neutral verdict is used to announce a new node intending to 1168 join the network so a final verdict for it can be found. 1170 The current effective trust verdict for any certificate is defined as 1171 the one with the highest priority within the set of trust verdicts 1172 announced for the certificate in the DNCP network. A node MUST be 1173 trusted for participating in the DNCP network if and only if the 1174 current effective trust verdict for its own certificate or any one in 1175 its certificate hierarchy is (Cached or Configured) Trust and none of 1176 the certificates in its hierarchy have an effective trust verdict of 1177 (Cached or Configured) Distrust. In case a node has a configured 1178 verdict, which is different from the current effective trust verdict 1179 for a certificate, the current effective trust verdict takes 1180 precedence in deciding trustworthiness. Despite that, the node still 1181 retains and announces its configured verdict. 1183 8.3.2. Trust Cache 1185 Each node SHOULD maintain a trust cache containing the current 1186 effective trust verdicts for all certificates currently announced in 1187 the DNCP network. This cache is used as a backup of the last known 1188 state in case there is no node announcing a configured verdict for a 1189 known certificate. It SHOULD be saved to a non-volatile memory at 1190 reasonable time intervals to survive a reboot or power outage. 1192 Every time a node (re)joins the network or detects the change of an 1193 effective trust verdict for any certificate, it will synchronize its 1194 cache, i.e., store new effective trust verdicts overwriting any 1195 previously cached verdicts. Configured verdicts are stored in the 1196 cache as their respective cached counterparts. Neutral verdicts are 1197 never stored and do not override existing cached verdicts. 1199 8.3.3. Announcement of Verdicts 1201 A node SHOULD always announce any configured trust verdicts it has 1202 established by itself, and it MUST do so if announcing the configured 1203 trust verdict leads to a change in the current effective trust 1204 verdict for the respective certificate. In absence of configured 1205 verdicts, it MUST announce cached trust verdicts it has stored in its 1206 trust cache, if one of the following conditions applies: 1208 o The stored trust verdict is Cached Trust and the current effective 1209 trust verdict for the certificate is Neutral or does not exist. 1211 o The stored trust verdict is Cached Distrust and the current 1212 effective trust verdict for the certificate is Cached Trust. 1214 A node rechecks these conditions whenever it detects changes of 1215 announced trust verdicts anywhere in the network. 1217 Upon encountering a node with a hierarchy of certificates for which 1218 there is no effective trust verdict, a node adds a Neutral Trust- 1219 Verdict-TLV to its node data for all certificates found in the 1220 hierarchy, and publishes it until an effective trust verdict 1221 different from Neutral can be found for any of the certificates, or a 1222 reasonable amount of time (10 minutes is suggested) with no reaction 1223 and no further authentication attempts has passed. Such trust 1224 verdicts SHOULD also be limited in rate and number to prevent denial- 1225 of-service attacks. 1227 Trust verdicts are announced using Trust-Verdict TLVs: 1229 0 1 2 3 1230 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1231 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1232 | Type: Trust-Verdict (10) | Length: > 36 | 1233 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1234 | Verdict | (reserved) | 1235 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1236 | | 1237 | | 1238 | | 1239 | SHA-256 Fingerprint | 1240 | | 1241 | | 1242 | | 1243 | | 1244 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1245 | Common Name | 1246 Verdict represents the numerical index of the trust verdict. 1248 (reserved) is reserved for future additions and MUST be set to 0 1249 when creating TLVs and ignored when parsing them. 1251 SHA-256 Fingerprint contains the SHA-256 [RFC6234] hash value of 1252 the certificate in DER-format. 1254 Common Name contains the variable-length (1-64 bytes) common name 1255 of the certificate. 1257 8.3.4. Bootstrap Ceremonies 1259 The following non-exhaustive list of methods describes possible ways 1260 to establish trust relationships between DNCP nodes and node 1261 certificates. Trust establishment is a two-way process in which the 1262 existing network must trust the newly added node and the newly added 1263 node must trust at least one of its peer nodes. It is therefore 1264 necessary that both the newly added node and an already trusted node 1265 perform such a ceremony to successfully introduce a node into the 1266 DNCP network. In all cases an administrator MUST be provided with 1267 external means to identify the node belonging to a certificate based 1268 on its fingerprint and a meaningful common name. 1270 8.3.4.1. Trust by Identification 1272 A node implementing certificate-based trust MUST provide an interface 1273 to retrieve the current set of effective trust verdicts, fingerprints 1274 and names of all certificates currently known and set configured 1275 trust verdicts to be announced. Alternatively it MAY provide a 1276 companion DNCP node or application with these capabilities with which 1277 it has a pre-established trust relationship. 1279 8.3.4.2. Preconfigured Trust 1281 A node MAY be preconfigured to trust a certain set of node or CA 1282 certificates. However such trust relationships MUST NOT result in 1283 unwanted or unrelated trust for nodes not intended to be run inside 1284 the same network (e.g., all other devices by the same manufacturer). 1286 8.3.4.3. Trust on Button Press 1288 A node MAY provide a physical or virtual interface to put one or more 1289 of its internal network interfaces temporarily into a mode in which 1290 it trusts the certificate of the first DNCP node it can successfully 1291 establish a connection with. 1293 8.3.4.4. Trust on First Use 1295 A node which is not associated with any other DNCP node MAY trust the 1296 certificate of the first DNCP node it can successfully establish a 1297 connection with. This method MUST NOT be used when the node has 1298 already associated with any other DNCP node. 1300 9. DNCP Profile-Specific Definitions 1302 Each DNCP profile MUST specify the following aspects: 1304 o Unicast and optionally multicast transport protocol(s) to be used. 1305 If multicast-based node and status discovery is desired, a 1306 datagram-based transport supporting multicast has to be available. 1308 o How the chosen transport(s) are secured: Not at all, optionally or 1309 always with the TLS scheme defined here using one or more of the 1310 methods, or with something else. If the links with DNCP nodes can 1311 be sufficiently secured or isolated, it is possible to run DNCP in 1312 a secure manner without using any form of authentication or 1313 encryption. 1315 o Transport protocols' parameters such as port numbers to be used, 1316 or multicast address to be used. Unicast, multicast, and secure 1317 unicast may each require different parameters, if applicable. 1319 o When receiving TLVs, what sort of TLVs are ignored in addition - 1320 as specified in Section 4.4 - e.g., for security reasons. While 1321 the security of the node data published within the Node State TLVs 1322 is already ensured by the base specification (if secure mode is 1323 enabled, Node State TLVs are sent only via unicast as multicast 1324 ones are ignored on receipt), if a profile adds TLVs that are sent 1325 outside the node data, a profile should indicate whether or not 1326 those TLVs should be ignored if they are received via multicast or 1327 non-secured unicast. A DNCP profile may define the following DNCP 1328 TLVs to be safely ignored: 1330 * Anything received over multicast, except Node Endpoint TLV 1331 (Section 7.2.1) and Network State TLV (Section 7.2.2). 1333 * Any TLVs received over unreliable unicast or multicast at too 1334 high rate; Trickle will ensure eventual convergence given the 1335 rate slows down at some point. 1337 o How to deal with node identifier collision as described in 1338 Section 4.4. Main options are either for one or both nodes to 1339 assign new node identifiers to themselves, or to notify someone 1340 about a fatal error condition in the DNCP network. 1342 o Imin, Imax and k ranges to be suggested for implementations to be 1343 used in the Trickle algorithm. The Trickle algorithm does not 1344 require these to be the same across all implementations for it to 1345 work, but similar orders of magnitude helps implementations of a 1346 DNCP profile to behave more consistently and to facilitate 1347 estimation of lower and upper bounds for convergence behavior of 1348 the network. 1350 o Hash function H(x) to be used, and how many bits of the output are 1351 actually used. The chosen hash function is used to handle both 1352 hashing of node specific data, and network state hash, which is a 1353 hash of node specific data hashes. SHA-256 defined in [RFC6234] 1354 is the recommended default choice, but a non-cryptographic hash 1355 function could be used as well. 1357 o DNCP_NODE_IDENTIFIER_LENGTH: The fixed length of a node identifier 1358 (in bytes). 1360 o Whether to send keep-alives, and if so, whether per-endpoint 1361 (requires multicast transport), or per-peer. Keep-alive has also 1362 associated parameters: 1364 * DNCP_KEEPALIVE_INTERVAL: How often keep-alives are to be sent 1365 by default (if enabled). 1367 * DNCP_KEEPALIVE_MULTIPLIER: How many times the 1368 DNCP_KEEPALIVE_INTERVAL (or peer-supplied keep-alive interval 1369 value) a node may not be heard from to be considered still 1370 valid. This is just a default used in absence of any other 1371 configuration information, or particular per-endpoint 1372 configuration. 1374 o Whether to support dense multicast-enabled link optimization 1375 (Section 6.2) or not. 1377 For some guidance on choosing transport and security options, please 1378 see Appendix B. 1380 10. Security Considerations 1382 DNCP-based protocols may use multicast to indicate DNCP state changes 1383 and for keep-alive purposes. However, no actual published data TLVs 1384 will be sent across that channel. Therefore an attacker may only 1385 learn hash values of the state within DNCP and may be able to trigger 1386 unicast synchronization attempts between nodes on a local link this 1387 way. A DNCP node MUST therefore rate-limit its reactions to 1388 multicast packets. 1390 When using DNCP to bootstrap a network, PKI based solutions may have 1391 issues when validating certificates due to potentially unavailable 1392 accurate time, or due to inability to use the network to either check 1393 Certificate Revocation Lists or perform on-line validation. 1395 The Certificate-based trust consensus mechanism defined in this 1396 document allows for a consenting revocation, however in case of a 1397 compromised device the trust cache may be poisoned before the actual 1398 revocation happens allowing the distrusted device to rejoin the 1399 network using a different identity. Stopping such an attack might 1400 require physical intervention and flushing of the trust caches. 1402 11. IANA Considerations 1404 IANA should set up a registry for the (decimal 16-bit) "DNCP TLV 1405 Types" under "Distributed Node Consensus Protocol (DNCP)", with the 1406 following initial contents: ([RFC Editor: please remove] ideally as 1407 http://www.iana.org/assignments/dncp-registry) 1409 0: Reserved 1411 1: Request network state 1413 2: Request node state 1415 3: Node endpoint 1417 4: Network state 1419 5: Node state 1421 6: Reserved (was: Custom) 1423 7: Reserved (was: Fragment count) 1425 8: Peer 1427 9: Keep-alive interval 1429 10: Trust-Verdict 1431 11-31: Free - policy of standards action [RFC5226] should be used 1433 32-511: Reserved for per-DNCP profile use 1435 512-767: Free - policy of standards action [RFC5226] should be 1436 used 1437 768-1023: Private use [RFC5226] 1439 1024-65535: Reserved for future protocol evolution (for example, 1440 DNCP version 2) 1442 12. References 1444 12.1. Normative references 1446 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1447 Requirement Levels", BCP 14, RFC 2119, March 1997. 1449 [RFC6206] Levis, P., Clausen, T., Hui, J., Gnawali, O., and J. Ko, 1450 "The Trickle Algorithm", RFC 6206, March 2011. 1452 [RFC6234] Eastlake, D. and T. Hansen, "US Secure Hash Algorithms 1453 (SHA and SHA-based HMAC and HKDF)", RFC 6234, May 2011. 1455 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 1456 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 1457 DOI 10.17487/RFC5226, May 2008, 1458 . 1460 12.2. Informative references 1462 [RFC3493] Gilligan, R., Thomson, S., Bound, J., McCann, J., and W. 1463 Stevens, "Basic Socket Interface Extensions for IPv6", RFC 1464 3493, February 2003. 1466 [RFC3315] Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C., 1467 and M. Carney, "Dynamic Host Configuration Protocol for 1468 IPv6 (DHCPv6)", RFC 3315, July 2003. 1470 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 1471 Security Version 1.2", RFC 6347, January 2012. 1473 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1474 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 1476 [RFC7435] Dukhovni, V., "Opportunistic Security: Some Protection 1477 Most of the Time", RFC 7435, DOI 10.17487/RFC7435, 1478 December 2014, . 1480 Appendix A. Alternative Modes of Operation 1482 Beyond what is described in the main text, the protocol allows for 1483 other uses. These are provided as examples. 1485 A.1. Read-only Operation 1487 If a node uses just a single endpoint and does not need to publish 1488 any TLVs, full DNCP node functionality is not required. Such limited 1489 node can acquire and maintain view of the TLV space by implementing 1490 the processing logic as specified in Section 4.4. Such node would 1491 not need Trickle, peer-maintenance or even keep-alives at all, as the 1492 DNCP nodes' use of it would guarantee eventual receipt of network 1493 state hashes, and synchronization of node data, even in presence of 1494 unreliable transport. 1496 A.2. Forwarding Operation 1498 If a node with a pair of endpoints does not need to publish any TLVs, 1499 it can detect (for example) nodes with the highest node identifier on 1500 each of the endpoints (if any). Any TLVs received from one of them 1501 would be forwarded verbatim as unicast to the other node with highest 1502 node identifier. 1504 Any tinkering with the TLVs would remove guarantees of this scheme 1505 working; however passive monitoring would obviously be fine. This 1506 type of simple forwarding cannot be chained, as it does not send 1507 anything proactively. 1509 Appendix B. DNCP Profile Additional Guidance 1511 This appendix explains implications of design choices made when 1512 specifying DNCP profile to use particular transport or security 1513 options. 1515 B.1. Unicast Transport - UDP or TCP? 1517 The node data published by a DNCP node is limited to 64KB due to the 1518 16-bit size of the length field of the TLV it is published within. 1519 Some transport choices may decrease this limit; if using e.g. UDP 1520 datagrams for unicast transport the upper bound of node data size is 1521 whatever the nodes and the underlying network can pass to each other 1522 as DNCP does not define its own fragmentation scheme. A profile 1523 which chooses UDP has to be limited to small node data (e.g. somewhat 1524 smaller than IPv6 default MTU if using IPv6), or specify a minimum 1525 which all nodes have to support. Even then, if using non-link-local 1526 communications, there is some concern about what middleboxes do to 1527 fragmented packets. Therefore, the use of stream transport such as 1528 TCP is probably a good idea if either non-link-local communication is 1529 desired, or fragmentation is expected to cause problems. 1531 TCP also provides some other facilities, such as a relatively long 1532 built-in keep-alive which in conjunction with connection closes 1533 occurring from eventual failed retransmissions may be sufficient to 1534 avoid the use of in-protocol keep-alive defined in Section 6.1. 1535 Additionally it is reliable, so there is no need for Trickle on such 1536 unicast connections. 1538 The major downside of using TCP instead of UDP with DNCP-based 1539 profiles lies in the loss of control over the time at which TLVs are 1540 received; while unreliable UDP datagrams also have some delay, TLVs 1541 within reliable stream transport may be delayed significantly due to 1542 retransmissions. This is not a problem if no relative time dependent 1543 information is stored within the TLVs in the DNCP-based protocol; for 1544 such a protocol, TCP is a reasonable choice for unicast transport if 1545 it is available. 1547 B.2. (Optional) Multicast Transport 1549 Multicast is needed for dynamic peer discovery and to trigger unicast 1550 exchanges; for that, unreliable datagram transport (=typically UDP) 1551 is the only transport option defined within this specification, 1552 although DNCP-based protocols may themselves define some other 1553 transport or peer discovery mechanism (e.g. based on mDNS or DNS). 1555 If multicast is used, a well-known address should be specified, and 1556 for e.g. IPv6 respectively the desired address scopes. In most 1557 cases link-local and possibly site-local are useful scopes. 1559 B.3. (Optional) Transport Security 1561 In terms of provided security, DTLS and TLS are equivalent; they also 1562 consume similar amount of state on the devices. While TLS is on top 1563 of a stream protocol, using DTLS also requires relatively long 1564 session caching within the DTLS layer to avoid expensive re- 1565 authentication/authorization steps if and when any state within the 1566 DNCP network changes or per-peer keep-alive (if enabled) is sent. 1568 TLS implementations (at the time of the writing of the specification) 1569 seem more mature and available (as open source) than DTLS ones. This 1570 may be due to a long history of use with HTTPS. 1572 Some libraries seem not to support multiplexing between insecure and 1573 secure communication on the same port, so specifying distinct ports 1574 for secured and unsecured communication may be beneficial. 1576 Appendix C. Example Profile 1578 This is the DNCP profile of SHSP, an experimental (and for the 1579 purposes of this document fictional) home automation protocol. The 1580 protocol itself is used to make key-value store published by each of 1581 the nodes available to all other nodes for distributed monitoring and 1582 control of a home infrastructure. It defines only one additional TLV 1583 type: a key=value TLV which contains a single key=value assignment 1584 for publication. 1586 o Unicast transport: IPv6 TCP on port EXAMPLE-P1 since only absolute 1587 timestamps are used within the key=value data and since it focuses 1588 primarily on Linux-based nodes which support both protocols well. 1589 Connections from and to non-link-local addresses are ignored to 1590 avoid exposing this protocol outside the secure links. 1592 o Multicast transport: IPv6 UDP on port EXAMPLE-P2 to link-local 1593 scoped multicast address ff02:EXAMPLE. At least one node per link 1594 in the home is assumed to facilitate node discovery without 1595 depending on any other infrastructure. 1597 o Security: None. It is to be used only on trusted links (WPA2-x 1598 wireless, physically secure wired links). 1600 o Additional TLVs to be ignored: None. No DNCP security is 1601 specified, and no new TLVs are defined outside of node data. 1603 o Node identifier length (DNCP_NODE_IDENTIFIER_LENGTH): 32 bits that 1604 are randomly generated. 1606 o Node identifier collision handling: Pick new random node 1607 identifier. 1609 o Trickle parameters: Imin = 200ms, Imax = 7, k = 1. It means at 1610 least one multicast per link in 25 seconds in stable state (0.2 * 1611 2^7). 1613 o Hash function H(x) + length: SHA-256, only 128 bits used. 1614 Relatively fast, and 128 bits should be plenty to prevent random 1615 conflicts (64 bits would most likely be sufficient, too). 1617 o No in-protocol keep-alives (Section 6.1); TCP keep-alive is to be 1618 used. In practice TCP keep-alive is seldom encountered anyway as 1619 changes in network state cause packets to be sent on the unicast 1620 connections, and those that fail sufficiently many retransmissions 1621 are dropped much before keep-alive actually would fire. 1623 o No support for dense multicast-enabled link optimization 1624 (Section 6.2); SHSP is a simple protocol for few nodes (network- 1625 wide, not even to mention on a single link), and therefore would 1626 not provide any benefit. 1628 Appendix D. Some Questions and Answers [RFC Editor: please remove] 1630 Q: 32-bit endpoint id? 1632 A: Here, it would save 32 bits per peer if it was 16 bits (and less 1633 is not realistic). However, TLVs defined elsewhere would not seem to 1634 even gain that much on average. 32 bits is also used for ifindex in 1635 various operating systems, making for simpler implementation. 1637 Q: Why have topology information at all? 1639 A: It is an alternative to the more traditional seq#/TTL-based 1640 flooding schemes. In steady state, there is no need to, e.g., re- 1641 publish every now and then. 1643 Appendix E. Changelog [RFC Editor: please remove] 1645 draft-ietf-homenet-dncp-10: 1647 o Added profile guidance section, as well as example profile. 1649 draft-ietf-homenet-dncp-09: 1651 o Reserved 1024+ TLV types for future versions (=versioning 1652 mechanism); private use section moved from 192-255 to 512-767. 1654 o Added applicability statement and clarified some text based on 1655 reviews. 1657 draft-ietf-homenet-dncp-08: 1659 o Removed fragmentation as it is somewhat underspecified and 1660 unimplemented. It may be specified in some future extension draft 1661 or new version of DNCP. 1663 o Added generic sub-TLV extensibility mechanism. 1665 draft-ietf-homenet-dncp-06: 1667 o Removed custom TLV. 1669 o Made keep-alive multipliers local implementation choice, profiles 1670 just provide guidance on sane default value. 1672 o Removed the DNCP_GRACE_INTERVAL as it is really implementation 1673 choice. 1675 o Simplified the suggested structures in data model. 1677 o Reorganized the document and provided an overview section. 1679 draft-ietf-homenet-dncp-04: 1681 o Added mandatory rate limiting for network state requests, and 1682 optional slightly faster convergence mechanism by including 1683 current local network state in the remote network state requests. 1685 draft-ietf-homenet-dncp-03: 1687 o Renamed connection -> endpoint. 1689 o !!! Backwards incompatible change: Renumbered TLVs, and got rid of 1690 node data TLV; instead, node data TLV's contents are optionally 1691 within node state TLV. 1693 draft-ietf-homenet-dncp-02: 1695 o Changed DNCP "messages" into series of TLV streams, allowing 1696 optimized round-trip saving synchronization. 1698 o Added fragmentation support for bigger node data and for chunking 1699 in absence of reliable L2 and L3 fragmentation. 1701 draft-ietf-homenet-dncp-01: 1703 o Fixed keep-alive semantics to consider unicast requests also 1704 updates of most recently consistent, and added proactive unicast 1705 request to ensure even inconsistent keep-alive messages eventually 1706 triggering consistency timestamp update. 1708 o Facilitated (simple) read-only clients by making Node Connection 1709 TLV optional if just using DNCP for read-only purposes. 1711 o Added text describing how to deal with "dense" networks, but left 1712 actual numbers and mechanics up to DNCP profiles and (local) 1713 configurations. 1715 draft-ietf-homenet-dncp-00: Split from pre-version of draft-ietf- 1716 homenet-hncp-03 generic parts. Changes that affect implementations: 1718 o TLVs were renumbered. 1720 o TLV length does not include header (=-4). This facilitates, e.g., 1721 use of DHCPv6 option parsing libraries (same encoding), and 1722 reduces complexity (no need to handle error values of length less 1723 than 4). 1725 o Trickle is reset only when locally calculated network state hash 1726 is changes, not as remote different network state hash is seen. 1727 This prevents, e.g., attacks by multicast with one multicast 1728 packet to force Trickle reset on every interface of every node on 1729 a link. 1731 o Instead of 'ping', use 'keep-alive' (optional) for dead peer 1732 detection. Different message used! 1734 Appendix F. Draft Source [RFC Editor: please remove] 1736 As usual, this draft is available at https://github.com/fingon/ietf- 1737 drafts/ in source format (with nice Makefile too). Feel free to send 1738 comments and/or pull requests if and when you have changes to it! 1740 Appendix G. Acknowledgements 1742 Thanks to Ole Troan, Pierre Pfister, Mark Baugher, Mark Townsley, 1743 Juliusz Chroboczek, Jiazi Yi, Mikael Abrahamsson, Brian Carpenter, 1744 Thomas Clausen, DENG Hui and Margaret Cullen for their contributions 1745 to the draft. 1747 Thanks to Kaiwen Jin and Xavier Bonnetain for their related research 1748 work. 1750 Authors' Addresses 1752 Markus Stenberg 1753 Independent 1754 Helsinki 00930 1755 Finland 1757 Email: markus.stenberg@iki.fi 1759 Steven Barth 1760 Independent 1761 Halle 06114 1762 Germany 1764 Email: cyrus@openwrt.org