idnits 2.17.1 draft-ietf-babel-rfc6126bis-08.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 1 instance of lines with multicast IPv4 addresses in the document. If these are generic example addresses, they should be changed to use the 233.252.0.x range defined in RFC 5771 -- The draft header indicates that this document obsoletes RFC7557, but the abstract doesn't seem to mention this, which it should. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 27, 2019) is 1850 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-10) exists of draft-ietf-babel-dtls-04 == Outdated reference: A later version (-12) exists of draft-ietf-babel-hmac-04 Summary: 0 errors (**), 0 flaws (~~), 4 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group J. Chroboczek 3 Internet-Draft IRIF, University of Paris-Diderot 4 Obsoletes: 6126,7557 (if approved) D. Schinazi 5 Intended status: Standards Track Google LLC 6 Expires: September 28, 2019 March 27, 2019 8 The Babel Routing Protocol 9 draft-ietf-babel-rfc6126bis-08 11 Abstract 13 Babel is a loop-avoiding distance-vector routing protocol that is 14 robust and efficient both in ordinary wired networks and in wireless 15 mesh networks. This document describes the Babel routing protocol, 16 and obsoletes RFCs 6126 and 7557. 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 https://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 September 28, 2019. 35 Copyright Notice 37 Copyright (c) 2019 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 (https://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. Features . . . . . . . . . . . . . . . . . . . . . . . . 3 54 1.2. Limitations . . . . . . . . . . . . . . . . . . . . . . . 4 55 1.3. Specification of Requirements . . . . . . . . . . . . . . 4 56 2. Conceptual Description of the Protocol . . . . . . . . . . . 5 57 2.1. Costs, Metrics and Neighbourship . . . . . . . . . . . . 5 58 2.2. The Bellman-Ford Algorithm . . . . . . . . . . . . . . . 5 59 2.3. Transient Loops in Bellman-Ford . . . . . . . . . . . . . 6 60 2.4. Feasibility Conditions . . . . . . . . . . . . . . . . . 7 61 2.5. Solving Starvation: Sequencing Routes . . . . . . . . . . 8 62 2.6. Requests . . . . . . . . . . . . . . . . . . . . . . . . 10 63 2.7. Multiple Routers . . . . . . . . . . . . . . . . . . . . 10 64 2.8. Overlapping Prefixes . . . . . . . . . . . . . . . . . . 11 65 3. Protocol Operation . . . . . . . . . . . . . . . . . . . . . 12 66 3.1. Message Transmission and Reception . . . . . . . . . . . 12 67 3.2. Data Structures . . . . . . . . . . . . . . . . . . . . . 12 68 3.3. Acknowledgments and acknowledgment requests . . . . . . . 16 69 3.4. Neighbour Acquisition . . . . . . . . . . . . . . . . . . 17 70 3.5. Routing Table Maintenance . . . . . . . . . . . . . . . . 20 71 3.6. Route Selection . . . . . . . . . . . . . . . . . . . . . 24 72 3.7. Sending Updates . . . . . . . . . . . . . . . . . . . . . 25 73 3.8. Explicit Requests . . . . . . . . . . . . . . . . . . . . 27 74 4. Protocol Encoding . . . . . . . . . . . . . . . . . . . . . . 31 75 4.1. Data Types . . . . . . . . . . . . . . . . . . . . . . . 32 76 4.2. Packet Format . . . . . . . . . . . . . . . . . . . . . . 33 77 4.3. TLV Format . . . . . . . . . . . . . . . . . . . . . . . 34 78 4.4. Sub-TLV Format . . . . . . . . . . . . . . . . . . . . . 34 79 4.5. Parser state . . . . . . . . . . . . . . . . . . . . . . 35 80 4.6. Details of Specific TLVs . . . . . . . . . . . . . . . . 36 81 4.7. Details of specific sub-TLVs . . . . . . . . . . . . . . 46 82 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 47 83 6. Security Considerations . . . . . . . . . . . . . . . . . . . 48 84 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 49 85 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 49 86 8.1. Normative References . . . . . . . . . . . . . . . . . . 49 87 8.2. Informative References . . . . . . . . . . . . . . . . . 50 88 Appendix A. Cost and Metric Computation . . . . . . . . . . . . 51 89 A.1. Maintaining Hello History . . . . . . . . . . . . . . . . 51 90 A.2. Cost Computation . . . . . . . . . . . . . . . . . . . . 52 91 A.3. Metric Computation . . . . . . . . . . . . . . . . . . . 53 92 Appendix B. Constants . . . . . . . . . . . . . . . . . . . . . 54 93 Appendix C. Considerations for protocol extensions . . . . . . . 55 94 Appendix D. Stub Implementations . . . . . . . . . . . . . . . . 57 95 Appendix E. Software Availability . . . . . . . . . . . . . . . 57 96 Appendix F. Changes from previous versions . . . . . . . . . . . 58 97 F.1. Changes since RFC 6126 . . . . . . . . . . . . . . . . . 58 98 F.2. Changes since draft-ietf-babel-rfc6126bis-00 . . . . . . 58 99 F.3. Changes since draft-ietf-babel-rfc6126bis-01 . . . . . . 58 100 F.4. Changes since draft-ietf-babel-rfc6126bis-02 . . . . . . 59 101 F.5. Changes since draft-ietf-babel-rfc6126bis-03 . . . . . . 59 102 F.6. Changes since draft-ietf-babel-rfc6126bis-03 . . . . . . 60 103 F.7. Changes since draft-ietf-babel-rfc6126bis-04 . . . . . . 60 104 F.8. Changes since draft-ietf-babel-rfc6126bis-05 . . . . . . 60 105 F.9. Changes since draft-ietf-babel-rfc6126bis-06 . . . . . . 60 106 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 60 108 1. Introduction 110 Babel is a loop-avoiding distance-vector routing protocol that is 111 designed to be robust and efficient both in networks using prefix- 112 based routing and in networks using flat routing ("mesh networks"), 113 and both in relatively stable wired networks and in highly dynamic 114 wireless networks. 116 1.1. Features 118 The main property that makes Babel suitable for unstable networks is 119 that, unlike naive distance-vector routing protocols [RIP], it 120 strongly limits the frequency and duration of routing pathologies 121 such as routing loops and black-holes during reconvergence. Even 122 after a mobility event is detected, a Babel network usually remains 123 loop-free. Babel then quickly reconverges to a configuration that 124 preserves the loop-freedom and connectedness of the network, but is 125 not necessarily optimal; in many cases, this operation requires no 126 packet exchanges at all. Babel then slowly converges, in a time on 127 the scale of minutes, to an optimal configuration. This is achieved 128 by using sequenced routes, a technique pioneered by Destination- 129 Sequenced Distance-Vector routing [DSDV]. 131 More precisely, Babel has the following properties: 133 o when every prefix is originated by at most one router, Babel never 134 suffers from routing loops; 136 o when a single prefix is originated by multiple routers, Babel may 137 occasionally create a transient routing loop for this particular 138 prefix; this loop disappears in a time proportional to its 139 diameter, and never again (up to an arbitrary garbage-collection 140 (GC) time) will the routers involved participate in a routing loop 141 for the same prefix; 143 o assuming bounded packet loss rates, any routing black-holes that 144 may appear after a mobility event are corrected in a time at most 145 proportional to the network's diameter. 147 Babel has provisions for link quality estimation and for fairly 148 arbitrary metrics. When configured suitably, Babel can implement 149 shortest-path routing, or it may use a metric based, for example, on 150 measured packet loss. 152 Babel nodes will successfully establish an association even when they 153 are configured with different parameters. For example, a mobile node 154 that is low on battery may choose to use larger time constants (hello 155 and update intervals, etc.) than a node that has access to wall 156 power. Conversely, a node that detects high levels of mobility may 157 choose to use smaller time constants. The ability to build such 158 heterogeneous networks makes Babel particularly adapted to the 159 unmanaged and wireless environment. 161 Finally, Babel is a hybrid routing protocol, in the sense that it can 162 carry routes for multiple network-layer protocols (IPv4 and IPv6), 163 whichever protocol the Babel packets are themselves being carried 164 over. 166 1.2. Limitations 168 Babel has two limitations that make it unsuitable for use in some 169 environments. First, Babel relies on periodic routing table updates 170 rather than using a reliable transport; hence, in large, stable 171 networks it generates more traffic than protocols that only send 172 updates when the network topology changes. In such networks, 173 protocols such as OSPF [OSPF], IS-IS [IS-IS], or the Enhanced 174 Interior Gateway Routing Protocol (EIGRP) [EIGRP] might be more 175 suitable. 177 Second, unless the optional algorithm described in Section 3.5.5 is 178 implemented, Babel does impose a hold time when a prefix is 179 retracted. While this hold time does not apply to the exact prefix 180 being retracted, and hence does not prevent fast reconvergence should 181 it become available again, it does apply to any shorter prefix that 182 covers it. This may make those implementations of Babel that do not 183 implement the optional algorithm described in Section 3.5.5 184 unsuitable for use in networks that implement automatic prefix 185 aggregation. 187 1.3. Specification of Requirements 189 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 190 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 191 "OPTIONAL" in this document are to be interpreted as described in BCP 192 14 [RFC2119] [RFC8174] when, and only when, they appear in all 193 capitals, as shown here. 195 2. Conceptual Description of the Protocol 197 Babel is a loop-avoiding distance vector protocol: it is based on the 198 Bellman-Ford protocol, just like the venerable RIP [RIP], but 199 includes a number of refinements that either prevent loop formation 200 altogether, or ensure that a loop disappears in a timely manner and 201 doesn't form again. 203 Conceptually, Bellman-Ford is executed in parallel for every source 204 of routing information (destination of data traffic). In the 205 following discussion, we fix a source S; the reader will recall that 206 the same algorithm is executed for all sources. 208 2.1. Costs, Metrics and Neighbourship 210 For every pair of neighbouring nodes A and B, Babel computes an 211 abstract value known as the cost of the link from A to B., written 212 C(A, B). Given a route between any two (not necessarily 213 neighbouring) nodes, the metric of the route is the sum of the costs 214 of all the edges along the route. The goal of the routing algorithm 215 is to compute, for every source S, the tree of routes of lowest 216 metric to S. 218 Costs and metrics need not be integers. In general, they can be 219 values in any algebra that satisfies two fairly general conditions 220 (Section 3.5.2). 222 A Babel node periodically sends Hello messages to all of its 223 neighbours; it also periodically sends an IHU ("I Heard You") message 224 to every neighbour from which it has recently heard a Hello. From 225 the information derived from Hello and IHU messages received from its 226 neighbour B, a node A computes the cost C(A, B) of the link from A to 227 B. 229 2.2. The Bellman-Ford Algorithm 231 Every node A maintains two pieces of data: its estimated distance to 232 S, written D(A), and its next-hop router to S, written NH(A). 233 Initially, D(S) = 0, D(A) is infinite, and NH(A) is undefined. 235 Periodically, every node B sends to all of its neighbours a route 236 update, a message containing D(B). When a neighbour A of B receives 237 the route update, it checks whether B is its selected next hop; if 238 that is the case, then NH(A) is set to B, and D(A) is set to C(A, B) 239 + D(B). If that is not the case, then A compares C(A, B) + D(B) to 240 its current value of D(A). If that value is smaller, meaning that 241 the received update advertises a route that is better than the 242 currently selected route, then NH(A) is set to B, and D(A) is set to 243 C(A, B) + D(B). 245 A number of refinements to this algorithm are possible, and are used 246 by Babel. In particular, convergence speed may be increased by 247 sending unscheduled "triggered updates" whenever a major change in 248 the topology is detected, in addition to the regular, scheduled 249 updates. Additionally, a node may maintain a number of alternate 250 routes, which are being advertised by neighbours other than its 251 selected neighbour, and which can be used immediately if the selected 252 route were to fail. 254 2.3. Transient Loops in Bellman-Ford 256 It is well known that a naive application of Bellman-Ford to 257 distributed routing can cause transient loops after a topology 258 change. Consider for example the following topology: 260 B 261 1 /| 262 1 / | 263 S --- A |1 264 \ | 265 1 \| 266 C 268 After convergence, D(B) = D(C) = 2, with NH(B) = NH(C) = A. 270 Suppose now that the link between S and A fails: 272 B 273 1 /| 274 / | 275 S A |1 276 \ | 277 1 \| 278 C 280 When it detects the failure of the link, A switches its next hop to B 281 (which is still advertising a route to S with metric 2), and 282 advertises a metric equal to 3, and then advertises a new route with 283 metric 3. This process of nodes changing selected neighbours and 284 increasing their metric continues until the advertised metric reaches 285 "infinity", a value larger than all the metrics that the routing 286 protocol is able to carry. 288 2.4. Feasibility Conditions 290 Bellman-Ford is a very robust algorithm: its convergence properties 291 are preserved when routers delay route acquisition or when they 292 discard some updates. Babel routers discard received route 293 announcements unless they can prove that accepting them cannot 294 possibly cause a routing loop. 296 More formally, we define a condition over route announcements, known 297 as the "feasibility condition", that guarantees the absence of 298 routing loops whenever all routers ignore route updates that do not 299 satisfy the feasibility condition. In effect, this makes Bellman- 300 Ford into a family of routing algorithms, parameterised by the 301 feasibility condition. 303 Many different feasibility conditions are possible. For example, BGP 304 can be modelled as being a distance-vector protocol with a (rather 305 drastic) feasibility condition: a routing update is only accepted 306 when the receiving node's AS number is not included in the update's 307 AS-Path attribute (note that BGP's feasibility condition does not 308 ensure the absence of transient "micro-loops" during reconvergence). 310 Another simple feasibility condition, used in the Destination- 311 Sequenced Distance-Vector (DSDV) routing protocol [DSDV] and in the 312 Ad hoc On-Demand Distance Vector (AODV) protocol, stems from the 313 following observation: a routing loop can only arise after a router 314 has switched to a route with a larger metric than the route that it 315 had previously selected. Hence, one could decide that a route is 316 feasible only when its metric at the local node would be no larger 317 than the metric of the currently selected route, i.e., an 318 announcement carrying a metric D(B) is accepted by A when C(A, B) + 319 D(B) <= D(A). If all routers obey this constraint, then the metric 320 at every router is nonincreasing, and the following invariant is 321 always preserved: if A has selected B as its successor, then D(B) < 322 D(A), which implies that the forwarding graph is loop-free. 324 Babel uses a slightly more refined feasibility condition, derived 325 from EIGRP [DUAL]. Given a router A, define the feasibility distance 326 of A, written FD(A), as the smallest metric that A has ever 327 advertised for S to any of its neighbours. An update sent by a 328 neighbour B of A is feasible when the metric D(B) advertised by B is 329 strictly smaller than A's feasibility distance, i.e., when D(B) < 330 FD(A). 332 It is easy to see that this latter condition is no more restrictive 333 than DSDV-feasibility. Suppose that node A obeys DSDV-feasibility; 334 then D(A) is nonincreasing, hence at all times D(A) <= FD(A). 335 Suppose now that A receives a DSDV-feasible update that advertises a 336 metric D(B). Since the update is DSDV-feasible, C(A, B) + D(B) <= 337 D(A), hence D(B) < D(A), and since D(A) <= FD(A), D(B) < FD(A). 339 To see that it is strictly less restrictive, consider the following 340 diagram, where A has selected the route through B, and D(A) = FD(A) = 341 2. Since D(C) = 1 < FD(A), the alternate route through C is feasible 342 for A, although its metric C(A, C) + D(C) = 5 is larger than that of 343 the currently selected route: 345 B 346 1 / \ 1 347 / \ 348 S A 349 \ / 350 1 \ / 4 351 C 353 To show that this feasibility condition still guarantees loop- 354 freedom, recall that at the time when A accepts an update from B, the 355 metric D(B) announced by B is no smaller than FD(B); since it is 356 smaller than FD(A), at that point in time FD(B) < FD(A). Since this 357 property is preserved when A sends updates, it remains true at all 358 times, which ensures that the forwarding graph has no loops. 360 2.5. Solving Starvation: Sequencing Routes 362 Obviously, the feasibility conditions defined above cause starvation 363 when a router runs out of feasible routes. Consider the following 364 diagram, where both A and B have selected the direct route to S: 366 A 367 1 /| D(A) = 1 368 / | FD(A) = 1 369 S |1 370 \ | D(B) = 2 371 2 \| FD(B) = 2 372 B 374 Suppose now that the link between A and S breaks: 376 A 377 | 378 | FD(A) = 1 379 S |1 380 \ | D(B) = 2 381 2 \| FD(B) = 2 382 B 384 The only route available from A to S, the one that goes through B, is 385 not feasible: A suffers from spurious starvation. At that point, the 386 whole subtree suffering from starvation must be reset, which is 387 essentially what EIGRP does when it performs a global synchronisation 388 of all the routers in the sarving subtree (the "active" phase of 389 EIGRP). 391 Babel reacts to starvation in a less drastic manner, by using 392 sequenced routes, a technique introduced by DSDV and adopted by AODV. 393 In addition to a metric, every route carries a sequence number, a 394 nondecreasing integer that is propagated unchanged through the 395 network and is only ever incremented by the source; a pair (s, m), 396 where s is a sequence number and m a metric, is called a distance. 398 A received update is feasible when either it is more recent than the 399 feasibility distance maintained by the receiving node, or it is 400 equally recent and the metric is strictly smaller. More formally, if 401 FD(A) = (s, m), then an update carrying the distance (s', m') is 402 feasible when either s' > s, or s = s' and m' < m. 404 Assuming the sequence number of S is 137, the diagram above becomes: 406 A 407 | 408 | FD(A) = (137, 1) 409 S |1 410 \ | D(B) = (137, 2) 411 2 \| FD(B) = (137, 2) 412 B 414 After S increases its sequence number, and the new sequence number is 415 propagated to B, we have: 417 A 418 | 419 | FD(A) = (137, 1) 420 S |1 421 \ | D(B) = (138, 2) 422 2 \| FD(B) = (138, 2) 423 B 425 at which point the route through B becomes feasible again. 427 Note that while sequence numbers are used for determining 428 feasibility, they are not necessarily used in route selection: a node 429 will normally ignore the sequence number when selecting the best 430 route to a given destination (Section 3.6). 432 2.6. Requests 434 In DSDV, the sequence number of a source is increased periodically. 435 A route becomes feasible again after the source increases its 436 sequence number, and the new sequence number is propagated through 437 the network, which may, in general, require a significant amount of 438 time. 440 Babel takes a different approach. When a node detects that it is 441 suffering from a potentially spurious starvation, it sends an 442 explicit request to the source for a new sequence number. This 443 request is forwarded hop by hop to the source, with no regard to the 444 feasibility condition. Upon receiving the request, the source 445 increases its sequence number and broadcasts an update, which is 446 forwarded to the requesting node. 448 Note that after a change in network topology not all such requests 449 will, in general, reach the source, as some will be sent over links 450 that are now broken. However, if the network is still connected, 451 then at least one among the nodes suffering from spurious starvation 452 has an (unfeasible) route to the source; hence, in the absence of 453 packet loss, at least one such request will reach the source. 454 (Resending requests a small number of times compensates for packet 455 loss.) 457 Since requests are forwarded with no regard to the feasibility 458 condition, they may, in general, be caught in a forwarding loop; this 459 is avoided by having nodes perform duplicate detection for the 460 requests that they forward. 462 2.7. Multiple Routers 464 The above discussion assumes that every prefix is originated by a 465 single router. In real networks, however, it is often necessary to 466 have a single prefix originated by multiple routers: for example, the 467 default route will be originated by all of the edge routers of a 468 routing domain. 470 Since synchronising sequence numbers between distinct routers is 471 problematic, Babel treats routes for the same prefix as distinct 472 entities when they are originated by different routers: every route 473 announcement carries the router-id of its originating router, and 474 feasibility distances are not maintained per prefix, but per source, 475 where a source is a pair of a router-id and a prefix. In effect, 476 Babel guarantees loop-freedom for the forwarding graph to every 477 source; since the union of multiple acyclic graphs is not in general 478 acyclic, Babel does not in general guarantee loop-freedom when a 479 prefix is originated by multiple routers, but any loops will be 480 broken in a time at most proportional to the diameter of the loop -- 481 as soon as an update has "gone around" the routing loop. 483 Consider for example the following topology, where A has selected the 484 default route through S, and B has selected the one through S': 486 1 1 1 487 ::/0 -- S --- A --- B --- S' -- ::/0 489 Suppose that both default routes fail at the same time; then nothing 490 prevents A from switching to B, and B simultaneously switching to A. 491 However, as soon as A has successfully advertised the new route to B, 492 the route through A will become unfeasible for B. Conversely, as 493 soon as B will have advertised the route through A, the route through 494 B will become unfeasible for A. 496 In effect, the routing loop disappears at the latest when routing 497 information has gone around the loop. Since this process can be 498 delayed by lost packets, Babel makes certain efforts to ensure that 499 updates are sent reliably after a router-id change (Section 3.7.2). 501 Additionally, after the routers have advertised the two routes, both 502 sources will be in their source tables, which will prevent them from 503 ever again participating in a routing loop involving routes from S 504 and S' (up to the source GC time, which, available memory permitting, 505 can be set to arbitrarily large values). 507 2.8. Overlapping Prefixes 509 In the above discussion, we have assumed that all prefixes are 510 disjoint, as is the case in flat ("mesh") routing. In practice, 511 however, prefixes may overlap: for example, the default route 512 overlaps with all of the routes present in the network. 514 After a route fails, it is not correct in general to switch to a 515 route that subsumes the failed route. Consider for example the 516 following configuration: 518 1 1 519 ::/0 -- A --- B --- C 521 Suppose that node C fails. If B forwards packets destined to C by 522 following the default route, a routing loop will form, and persist 523 until A learns of B's retraction of the direct route to C. B avoids 524 this pitfall by installing an "unreachable" route after a route is 525 retracted; this route is maintained until it can be guaranteed that 526 the former route has been retracted by all of B's neighbours 527 (Section 3.5.5). 529 3. Protocol Operation 531 Every Babel speaker is assigned a router-id, which is an arbitrary 532 string of 8 octets that is assumed unique across the routing domain. 533 For example, routers-ids could be assigned randomly, or they could 534 derived from a link-layer address. (The protocol encoding is 535 slightly more compact when router-ids are assigned in the same manner 536 as the IPv6 layer assigns host IDs.) 538 3.1. Message Transmission and Reception 540 Babel protocol packets are sent in the body of a UDP datagram (as 541 described in Section 4 below). Each Babel packet consists of zero or 542 more TLVs. Most TLVs may contain sub-TLVs. 544 The source address of a Babel packet is always a unicast address, 545 link-local in the case of IPv6. Babel packets may be sent to a well- 546 known (link-local) multicast address or to a (link-local) unicast 547 address. In normal operation, a Babel speaker sends both multicast 548 and unicast packets to its neighbours. 550 With the exception of Hello TLVs and acknowledgments, all Babel TLVs 551 can be sent to either unicast or multicast addresses, and their 552 semantics does not depend on whether the destination is a unicast or 553 a multicast address. Hence, a Babel speaker does not need to 554 determine the destination address of a packet that it receives in 555 order to interpret it. 557 A moderate amount of jitter may be applied to packets sent by a Babel 558 speaker: outgoing TLVs are buffered and SHOULD be sent with a small 559 random delay. This is done for two purposes: it avoids 560 synchronisation of multiple Babel speakers across a network [JITTER], 561 and it allows for the aggregation of multiple TLVs into a single 562 packet. 564 The exact delay and amount of jitter applied to a packet depends on 565 whether it contains any urgent TLVs. Acknowledgment TLVs MUST be 566 sent before the deadline specified in the corresponding request. The 567 particular class of updates specified in Section 3.7.2 MUST be sent 568 in a timely manner. The particular class of request and update TLVs 569 specified in Section 3.8.2 SHOULD be sent in a timely manner. 571 3.2. Data Structures 573 In this section, we give a description of the data structures that 574 every Babel speaker maintains. This description is conceptual: a 575 Babel speaker may use different data structures as long as the 576 resulting protocol is the same as the one described in this document. 578 For example, rather than maintaining a single table containing both 579 selected and unselected (fallback) routes, as described in 580 Section 3.2.6 belong, an actual implementation would probably use two 581 tables, one with selected routes and one with fallback routes. 583 3.2.1. Sequence number arithmetic 585 Sequence numbers (seqnos) appear in a number of Babel data 586 structures, and they are interpreted as integers modulo 2^16. For 587 the purposes of this document, arithmetic on sequence numbers is 588 defined as follows. 590 Given a seqno s and an integer n, the sum of s and n is defined by 592 s + n (modulo 2^16) = (s + n) MOD 2^16 594 or, equivalently, 596 s + n (modulo 2^16) = (s + n) AND 65535 598 where MOD is the modulo operation yielding a non-negative integer and 599 AND is the bitwise conjunction operation. 601 Given two sequence numbers s and s', the relation s is less than s' 602 (s < s') is defined by 604 s < s' (modulo 2^16) when 0 < ((s' - s) MOD 2^16) < 32768 606 or equivalently 608 s < s' (modulo 2^16) when s /= s' and ((s' - s) AND 32768) = 0. 610 3.2.2. Node Sequence Number 612 A node's sequence number is a 16-bit integer that is included in 613 route updates sent for routes originated by this node. 615 A node increments its sequence number (modulo 2^16) whenever it 616 receives a request for a new sequence number (Section 3.8.1.2). A 617 node SHOULD NOT increment its sequence number (seqno) spontaneously, 618 since increasing seqnos makes it less likely that other nodes will 619 have feasible alternate routes when their selected routes fail. 621 3.2.3. The Interface Table 623 The interface table contains the list of interfaces on which the node 624 speaks the Babel protocol. Every interface table entry contains the 625 interface's outgoing Multicast Hello seqno, a 16-bit integer that is 626 sent with each Multicast Hello TLV on this interface and is 627 incremented (modulo 2^16) whenever a Multicast Hello is sent. (Note 628 that an interface's Multicast Hello seqno is unrelated to the node's 629 seqno.) 631 There are two timers associated with each interface table entry -- 632 the multicast hello timer, which governs the sending of scheduled 633 Multicast Hello and IHU packets, and the update timer, which governs 634 the sending of periodic route updates. 636 3.2.4. The Neighbour Table 638 The neighbour table contains the list of all neighbouring interfaces 639 from which a Babel packet has been recently received. The neighbour 640 table is indexed by pairs of the form (interface, address), and every 641 neighbour table entry contains the following data: 643 o the local node's interface over which this neighbour is reachable; 645 o the address of the neighbouring interface; 647 o a history of recently received Multicast Hello packets from this 648 neighbour; this can, for example, be a sequence of n bits, for 649 some small value n, indicating which of the n hellos most recently 650 sent by this neighbour have been received by the local node; 652 o a history of recently received Unicast Hello packets from this 653 neighbour; 655 o the "transmission cost" value from the last IHU packet received 656 from this neighbour, or FFFF hexadecimal (infinity) if the IHU 657 hold timer for this neighbour has expired; 659 o the neighbour's expected incoming Multicast Hello sequence number, 660 an integer modulo 2^16. 662 o the neighbour's expected incoming Unicast Hello sequence number, 663 an integer modulo 2^16. 665 o the neighbour's outgoing Unicast Hello sequence number, an integer 666 modulo 2^16 that is sent with each Unicast Hello TLV to this 667 neighbour and is incremented (modulo 2^16) whenever a Unicast 668 Hello is sent. (Note that a neighbour's outgoing Unicast Hello 669 seqno is distinct from the interface's outgoing Multicast Hello 670 seqno.) 672 There are three timers associated with each neighbour entry -- the 673 multicast hello timer, which is initialised from the interval value 674 carried by scheduled Multicast Hello TLVs, the unicast hello timer, 675 which is initialised from the interval value carried by scheduled 676 Unicast Hello TLVs, and the IHU timer, which is initialised to a 677 small multiple of the interval carried in IHU TLVs. 679 Note that the neighbour table is indexed by IP addresses, not by 680 router-ids: neighbourship is a relationship between interfaces, not 681 between nodes. Therefore, two nodes with multiple interfaces can 682 participate in multiple neighbourship relationships, a situation that 683 can notably arise when wireless nodes with multiple radios are 684 involved. 686 3.2.5. The Source Table 688 The source table is used to record feasibility distances. It is 689 indexed by triples of the form (prefix, plen, router-id), and every 690 source table entry contains the following data: 692 o the prefix (prefix, plen), where plen is the prefix length, that 693 this entry applies to; 695 o the router-id of a router originating this prefix; 697 o a pair (seqno, metric), this source's feasibility distance. 699 There is one timer associated with each entry in the source table -- 700 the source garbage-collection timer. It is initialised to a time on 701 the order of minutes and reset as specified in Section 3.7.3. 703 3.2.6. The Route Table 705 The route table contains the routes known to this node. It is 706 indexed by triples of the form (prefix, plen, neighbour), and every 707 route table entry contains the following data: 709 o the source (prefix, plen, router-id) for which this route is 710 advertised; 712 o the neighbour that advertised this route; 714 o the metric with which this route was advertised by the neighbour, 715 or FFFF hexadecimal (infinity) for a recently retracted route; 717 o the sequence number with which this route was advertised; 719 o the next-hop address of this route; 720 o a boolean flag indicating whether this route is selected, i.e., 721 whether it is currently being used for forwarding and is being 722 advertised. 724 There is one timer associated with each route table entry -- the 725 route expiry timer. It is initialised and reset as specified in 726 Section 3.5.4. 728 Note that there are two distinct (seqno, metric) pairs associated to 729 each route: the route's distance, which is stored in the route table, 730 and the feasibility distance, stored in the source table and shared 731 between all routes with the same source. 733 3.2.7. The Table of Pending Seqno Requests 735 The table of pending seqno requests contains a list of seqno requests 736 that the local node has sent (either because they have been 737 originated locally, or because they were forwarded) and to which no 738 reply has been received yet. This table is indexed by triples of the 739 form (prefix, plen, router-id), and every entry in this table 740 contains the following data: 742 o the prefix, router-id, and seqno being requested; 744 o the neighbour, if any, on behalf of which we are forwarding this 745 request; 747 o a small integer indicating the number of times that this request 748 will be resent if it remains unsatisfied. 750 There is one timer associated with each pending seqno request; it 751 governs both the resending of requests and their expiry. 753 3.3. Acknowledgments and acknowledgment requests 755 A Babel speaker may request that a neighbour receiving a given packet 756 reply with an explicit acknowledgment within a given time. While the 757 use of acknowledgment requests is optional, every Babel speaker MUST 758 be able to reply to such a request. 760 An acknowledgment MUST be sent to a unicast destination. On the 761 other hand, acknowledgment requests may be sent to either unicast or 762 multicast destinations, in which case they request an acknowledgment 763 from all of the receiving nodes. 765 When to request acknowledgments is a matter of local policy; the 766 simplest strategy is to never request acknowledgments and to rely on 767 periodic updates to ensure that any reachable routes are eventually 768 propagated throughout the routing domain. In order to improve 769 convergence speed and reduce the amount of control traffic, 770 acknowledgment requests MAY be used in order to reliably send urgent 771 updates (Section 3.7.2) and retractions (Section 3.5.5), especially 772 when the number of neighbours on a given interface is small. Since 773 Babel is designed to deal gracefully with packet loss on unreliable 774 media, sending all packets with acknowledgment requests is not 775 necessary, and NOT RECOMMENDED, as the acknowledgments cause 776 additional traffic and may force additional Address Resolution 777 Protocol (ARP) or Neighbour Discovery (ND) exchanges. 779 3.4. Neighbour Acquisition 781 Neighbour acquisition is the process by which a Babel node discovers 782 the set of neighbours heard over each of its interfaces and 783 ascertains bidirectional reachability. On unreliable media, 784 neighbour acquisition additionally provides some statistics that may 785 be useful for link quality computation. 787 Before it can exchange routing information with a neighbour, a Babel 788 node MUST create an entry for that neighbour in the neighbour table. 789 When to do that is implementation-specific; suitable strategies 790 include creating an entry when any Babel packet is received, or 791 creating an entry when a Hello TLV is parsed. Similarly, in order to 792 conserve system resources, an implementation SHOULD discard an entry 793 when it has been unused for long enough; suitable strategies include 794 dropping the neighbour after a timeout, and dropping a neighbour when 795 the associated Hello histories become empty (see Appendix A.2). 797 3.4.1. Reverse Reachability Detection 799 Every Babel node sends Hello TLVs to its neighbours to indicate that 800 it is alive, at regular or irregular intervals. Each Hello TLV 801 carries an increasing (modulo 2^16) sequence number and an upper 802 bound on the time interval until the next Hello of the same type (see 803 below). If the time interval is set to 0, then the Hello TLV does 804 not establish a new promise: the deadline carried by the previous 805 Hello of the same type still applies to the next Hello (if the most 806 recent scheduled Hello of the right kind was received at time t0 and 807 carried interval i, then the previous promise of sending another 808 Hello before time t0 + i still holds). We say that a Hello is 809 "scheduled" if it carries a non-zero interval, and "unscheduled" 810 otherwise. 812 There are two kinds of Hellos: Multicast Hellos, which use a per- 813 interface Hello counter (the Multicast Hello seqno), and Unicast 814 Hellos, which use a per-neighbour counter (the Multicast Hello 815 seqno). A Multicast Hello with a given seqno MUST be sent to all 816 neighbours on a given interface, either by sending it to a multicast 817 address or by sending it to one unicast address per neighbour (hence, 818 the term "Multicast Hello" is a slight misnomer). A Unicast Hello 819 carrying a given seqno should normally be sent to just one neighbour 820 (over unicast), since the sequence numbers of different neighbours 821 are not in general synchronised. 823 Multicast Hellos sent over multicast can be used for neighbour 824 discovery; hence, a node SHOULD send periodic (scheduled) Multicast 825 Hellos unless neighbour discovery is performed by means outside of 826 the Babel protocol. A node MAY send Unicast Hellos or unscheduled 827 Hellos of either kind for any reason, such as reducing the amount of 828 multicast traffic or improving reliability on link technologies with 829 poor support for link-layer multicast. 831 A node MAY send a scheduled Hello ahead of time. A node MAY change 832 its scheduled Hello interval. The Hello interval MAY be decreased at 833 any time; it MAY be increased immediately before sending a Hello TLV, 834 but SHOULD NOT be increased at other times. (Equivalently, a node 835 SHOULD send a scheduled Hello immediately after increasing its Hello 836 interval.) 838 How to deal with received Hello TLVs and what statistics to maintain 839 are considered local implementation matters; typically, a node will 840 maintain some sort of history of recently received Hellos. An 841 example of a suitable algorithm is described in Appendix A.1. 843 After receiving a Hello, or determining that it has missed one, the 844 node recomputes the association's cost (Section 3.4.3) and runs the 845 route selection procedure (Section 3.6). 847 3.4.2. Bidirectional Reachability Detection 849 In order to establish bidirectional reachability, every node sends 850 periodic IHU ("I Heard You") TLVs to each of its neighbours. Since 851 IHUs carry an explicit interval value, they MAY be sent less often 852 than Hellos in order to reduce the amount of routing traffic in dense 853 networks; in particular, they SHOULD be sent less often than Hellos 854 over links with little packet loss. While IHUs are conceptually 855 unicast, they MAY be sent to a multicast address in order to avoid an 856 ARP or Neighbour Discovery exchange and to aggregate multiple IHUs 857 into a single packet. 859 In addition to the periodic IHUs, a node MAY, at any time, send an 860 unscheduled IHU packet. It MAY also, at any time, decrease its IHU 861 interval, and it MAY increase its IHU interval immediately before 862 sending an IHU, but SHOULD NOT increase it at any other time. 864 (Equivalently, a node SHOULD send an extra IHU immediately after 865 increasing its Hello interval.) 867 Every IHU TLV contains two pieces of data: the link's rxcost 868 (reception cost) from the sender's perspective, used by the neighbour 869 for computing link costs (Section 3.4.3), and the interval between 870 periodic IHU packets. A node receiving an IHU sets the value of the 871 txcost (transmission cost) maintained in the neighbour table to the 872 value contained in the IHU, and resets the IHU timer associated to 873 this neighbour to a small multiple of the interval value received in 874 the IHU. When a neighbour's IHU timer expires, the neighbour's 875 txcost is set to infinity. 877 After updating a neighbour's txcost, the receiving node recomputes 878 the neighbour's cost (Section 3.4.3) and runs the route selection 879 procedure (Section 3.6). 881 3.4.3. Cost Computation 883 A neighbourship association's link cost is computed from the values 884 maintained in the neighbour table: the statistics kept in the 885 neighbour table about the reception of Hellos, and the txcost 886 computed from received IHU packets. 888 For every neighbour, a Babel node computes a value known as this 889 neighbour's rxcost. This value is usually derived from the Hello 890 history, which may be combined with other data, such as statistics 891 maintained by the link layer. The rxcost is sent to a neighbour in 892 each IHU. 894 Since nodes do not necessarily send periodic Unicast Hellos but do 895 usually send periodic Multicast Hellos (Section 3.4.1), a node SHOULD 896 use an algorithm that yields a finite rxcost when only Multicast 897 Hellos are received, unless interoperability with nodes that only 898 send Multicast Hellos is not required. 900 How the txcost and rxcost are combined in order to compute a link's 901 cost is a matter of local policy; as far as Babel's correctness is 902 concerned, only the following conditions MUST be satisfied: 904 o the cost is strictly positive; 906 o if no Hello TLVs of either kind were received recently, then the 907 cost is infinite; 909 o if the txcost is infinite, then the cost is infinite. 911 Note that while this document does not constrain cost computation any 912 further, not all cost computation strategies will give good results. 913 See Appendix A.2 for examples of strategies for computing a link's 914 cost that are known to work well in practice. 916 3.5. Routing Table Maintenance 918 Conceptually, a Babel update is a quintuple (prefix, plen, router-id, 919 seqno, metric), where (prefix, plen) is the prefix for which a route 920 is being advertised, router-id is the router-id of the router 921 originating this update, seqno is a nondecreasing (modulo 2^16) 922 integer that carries the originating router seqno, and metric is the 923 announced metric. 925 Before being accepted, an update is checked against the feasibility 926 condition (Section 3.5.1), which ensures that the route does not 927 create a routing loop. If the feasibility condition is not 928 satisfied, the update is either ignored or prevents the route from 929 being selected, as described in Section 3.5.4. If the feasibility 930 condition is satisfied, then the update cannot possibly cause a 931 routing loop. 933 3.5.1. The Feasibility Condition 935 The feasibility condition is applied to all received updates. The 936 feasibility condition compares the metric in the received update with 937 the metrics of the updates previously sent by the receiving node; 938 updates that fail the feasibility condition, and therefore have 939 metrics large enough to cause a routing loop, are either ignored or 940 prevent the resulting route from being selected. 942 A feasibility distance is a pair (seqno, metric), where seqno is an 943 integer modulo 2^16 and metric is a positive integer. Feasibility 944 distances are compared lexicographically, with the first component 945 inverted: we say that a distance (seqno, metric) is strictly better 946 than a distance (seqno', metric'), written 948 (seqno, metric) < (seqno', metric') 950 when 952 seqno > seqno' or (seqno = seqno' and metric < metric') 954 where sequence numbers are compared modulo 2^16. 956 Given a source (prefix, plen, router-id), a node's feasibility 957 distance for this source is the minimum, according to the ordering 958 defined above, of the distances of all the finite updates ever sent 959 by this particular node for the prefix (prefix, plen) and the given 960 router-id. Feasibility distances are maintained in the source table, 961 the exact procedure is given in Section 3.7.3. 963 A received update is feasible when either it is a retraction (its 964 metric is FFFF hexadecimal), or the advertised distance is strictly 965 better, in the sense defined above, than the feasibility distance for 966 the corresponding source. More precisely, a route advertisement 967 carrying the quintuple (prefix, plen, router-id, seqno, metric) is 968 feasible if one of the following conditions holds: 970 o metric is infinite; or 972 o no entry exists in the source table indexed by (prefix, plen, 973 router-id); or 975 o an entry (prefix, plen, router-id, seqno', metric') exists in the 976 source table, and either 978 * seqno' < seqno or 980 * seqno = seqno' and metric < metric'. 982 Note that the feasibility condition considers the metric advertised 983 by the neighbour, not the route's metric; hence, a fluctuation in a 984 neighbour's cost cannot render a selected route unfeasible. Note 985 further that retractions (updates with infinite metric) are always 986 feasible, since they cannot possibly cause a routing loop. 988 3.5.2. Metric Computation 990 A route's metric is computed from the metric advertised by the 991 neighbour and the neighbour's link cost. Just like cost computation, 992 metric computation is considered a local policy matter; as far as 993 Babel is concerned, the function M(c, m) used for computing a metric 994 from a locally computed link cost and the metric advertised by a 995 neighbour MUST only satisfy the following conditions: 997 o if c is infinite, then M(c, m) is infinite; 999 o M is strictly monotonic: M(c, m) > m. 1001 Additionally, the metric SHOULD satisfy the following condition: 1003 o M is left-distributive: if m <= m', then M(c, m) <= M(c, m'). 1005 Note that while strict monotonicity is essential to the integrity of 1006 the network (persistent routing loops may arise if it is not 1007 satisfied), left distributivity is not: if it is not satisfied, Babel 1008 will still converge to a loop-free configuration, but might not reach 1009 a global optimum (in fact, a global optimum may not even exist). 1011 As with cost computation, not all strategies for computing route 1012 metrics will give good results. In particular, some metrics are more 1013 likely than others to lead to routing instabilities (route flapping). 1014 In Appendix A.3, we give a number of examples of strictly monotonic, 1015 left-distributive routing metrics that are known to work well in 1016 practice. 1018 3.5.3. Encoding of Updates 1020 In a large network, the bulk of Babel traffic consists of route 1021 updates; hence, some care has been given to encoding them 1022 efficiently. An Update TLV itself only contains the prefix, seqno, 1023 and metric, while the next hop is derived either from the network- 1024 layer source address of the packet or from an explicit Next Hop TLV 1025 in the same packet. The router-id is derived from a separate Router- 1026 Id TLV in the same packet, which optimises the case when multiple 1027 updates are sent with the same router-id. 1029 Additionally, a prefix of the advertised prefix can be omitted in an 1030 Update TLV, in which case it is copied from a previous Update TLV in 1031 the same packet -- this is known as address compression 1032 (Section 4.6.9). 1034 Finally, as a special optimisation for the case when a router-id 1035 coincides with the interface-id part of an IPv6 address, the router- 1036 id can optionally be derived from the low-order bits of the 1037 advertised prefix. 1039 The encoding of updates is described in detail in Section 4.6. 1041 3.5.4. Route Acquisition 1043 When a Babel node receives an update (prefix, plen, router-id, seqno, 1044 metric) from a neighbour neigh with a link cost value equal to cost, 1045 it checks whether it already has a route table entry indexed by 1046 (prefix, plen, neigh). 1048 If no such entry exists: 1050 o if the update is unfeasible, it MAY be ignored; 1052 o if the metric is infinite (the update is a retraction of a route 1053 we do not know about), the update is ignored; 1055 o otherwise, a new entry is created in the route table, indexed by 1056 (prefix, plen, neigh), with source equal to (prefix, plen, router- 1057 id), seqno equal to seqno and an advertised metric equal to the 1058 metric carried by the update. 1060 If such an entry exists: 1062 o if the entry is currently selected, the update is unfeasible, and 1063 the router-id of the update is equal to the router-id of the 1064 entry, then the update MAY be ignored; 1066 o otherwise, the entry's sequence number, advertised metric, metric, 1067 and router-id are updated and, if the advertised metric is not 1068 infinite, the route's expiry timer is reset to a small multiple of 1069 the Interval value included in the update. If the update is 1070 unfeasible, then the (now unfeasible) entry MUST be immediately 1071 unselected. If the update caused the router-id of the entry to 1072 change, an update (possibly a retraction) MUST be sent in a timely 1073 manner (see Section 3.7.2). 1075 Note that the route table may contain unfeasible routes, either 1076 because they were created by an unfeasible update or due to a metric 1077 fluctuation. Such routes are never selected, since they are not 1078 known to be loop-free; should all the feasible routes become 1079 unusable, however, the unfeasible routes can be made feasible and 1080 therefore possible to select by sending requests along them (see 1081 Section 3.8.2). 1083 When a route's expiry timer triggers, the behaviour depends on 1084 whether the route's metric is finite. If the metric is finite, it is 1085 set to infinity and the expiry timer is reset. If the metric is 1086 already infinite, the route is flushed from the route table. 1088 After the route table is updated, the route selection procedure 1089 (Section 3.6) is run. 1091 3.5.5. Hold Time 1093 When a prefix P is retracted, because all routes are unfeasible or 1094 have an infinite metric (whether due to the expiry timer or to other 1095 reasons), and a shorter prefix P' that covers P is reachable, P' 1096 cannot in general be used for routing packets destined to P without 1097 running the risk of creating a routing loop (Section 2.8). 1099 To avoid this issue, whenever a prefix P is retracted, a route table 1100 entry with infinite metric is maintained as described in 1101 Section 3.5.4 above. As long as this entry is maintained, packets 1102 destined to an address within P MUST NOT be forwarded by following a 1103 route for a shorter prefix. This entry is removed as soon as a 1104 finite-metric update for prefix P is received and the resulting route 1105 selected. If no such update is forthcoming, the infinite metric 1106 entry SHOULD be maintained at least until it is guaranteed that no 1107 neighbour has selected the current node as next-hop for prefix P. 1108 This can be achieved by either: 1110 o waiting until the route's expiry timer has expired 1111 (Section 3.5.4); 1113 o sending a retraction with an acknowledgment request (Section 3.3) 1114 to every reachable neighbour that has not explicitly retracted 1115 prefix P and waiting for all acknowledgments. 1117 The former option is simpler and ensures that at that point, any 1118 routes for prefix P pointing at the current node have expired. 1119 However, since the expiry time can be as high as a few minutes, doing 1120 that prevents automatic aggregation by creating spurious black-holes 1121 for aggregated routes. The latter option is RECOMMENDED as it 1122 dramatically reduces the time for which a prefix is unreachable in 1123 the presence of aggregated routes. 1125 3.6. Route Selection 1127 Route selection is the process by which a single route for a given 1128 prefix is selected to be used for forwarding packets and to be re- 1129 advertised to a node's neighbours. 1131 Babel is designed to allow flexible route selection policies. As far 1132 as the protocol's correctness is concerned, the route selection 1133 policy MUST only satisfy the following properties: 1135 o a route with infinite metric (a retracted route) is never 1136 selected; 1138 o an unfeasible route is never selected. 1140 Note, however, that Babel does not naturally guarantee the stability 1141 of routing, and configuring conflicting route selection policies on 1142 different routers may lead to persistent route oscillation. 1144 Route selection is a difficult problem, since a good route selection 1145 policy needs to take into account multiple mutually contradictory 1146 criteria; in roughly decreasing order of importance, these are: 1148 o routes with a small metric should be preferred to routes with a 1149 large metric; 1151 o switching router-ids should be avoided; 1153 o routes through stable neighbours should be preferred to routes 1154 through unstable ones; 1156 o stable routes should be preferred to unstable ones; 1158 o switching next hops should be avoided. 1160 A simple but useful strategy is to choose the feasible route with the 1161 smallest metric, with a small amount of hysteresis applied to avoid 1162 switching router-ids too often. 1164 After the route selection procedure is run, triggered updates 1165 (Section 3.7.2) and requests (Section 3.8.2) are sent. 1167 3.7. Sending Updates 1169 A Babel speaker advertises to its neighbours its set of selected 1170 routes. Normally, this is done by sending one or more multicast 1171 packets containing Update TLVs on all of its connected interfaces; 1172 however, on link technologies where multicast is significantly more 1173 expensive than unicast, a node MAY choose to send multiple copies of 1174 updates in unicast packets, especially when the number of neighbours 1175 is small. 1177 Additionally, in order to ensure that any black-holes are reliably 1178 cleared in a timely manner, a Babel node sends retractions (updates 1179 with an infinite metric) for any recently retracted prefixes. 1181 If an update is for a route injected into the Babel domain by the 1182 local node (e.g., it carries the address of a local interface, the 1183 prefix of a directly attached network, or a prefix redistributed from 1184 a different routing protocol), the router-id is set to the local 1185 node's router-id, the metric is set to some arbitrary finite value 1186 (typically 0), and the seqno is set to the local router's sequence 1187 number. 1189 If an update is for a route learned from another Babel speaker, the 1190 router-id and sequence number are copied from the route table entry, 1191 and the metric is computed as specified in Section 3.5.2. 1193 3.7.1. Periodic Updates 1195 Every Babel speaker periodically advertises all of its selected 1196 routes on all of its interfaces, including any recently retracted 1197 routes. Since Babel doesn't suffer from routing loops (there is no 1198 "counting to infinity") and relies heavily on triggered updates 1199 (Section 3.7.2), this full dump only needs to happen infrequently. 1201 3.7.2. Triggered Updates 1203 In addition to periodic routing updates, a Babel speaker sends 1204 unscheduled, or triggered, updates in order to inform its neighbours 1205 of a significant change in the network topology. 1207 A change of router-id for the selected route to a given prefix may be 1208 indicative of a routing loop in formation; hence, a node MUST send a 1209 triggered update in a timely manner whenever it changes the selected 1210 router-id for a given destination. Additionally, it SHOULD make a 1211 reasonable attempt at ensuring that all reachable neighbours receive 1212 this update. 1214 There are two strategies for ensuring that. If the number of 1215 neighbours is small, then it is reasonable to send the update 1216 together with an acknowledgment request; the update is resent until 1217 all neighbours have acknowledged the packet, up to some number of 1218 times. If the number of neighbours is large, however, requesting 1219 acknowledgments from all of them might cause a non-negligible amount 1220 of network traffic; in that case, it may be preferable to simply 1221 repeat the update some reasonable number of times (say, 5 for 1222 wireless and 2 for wired links). 1224 A route retraction is somewhat less worrying: if the route retraction 1225 doesn't reach all neighbours, a black-hole might be created, which, 1226 unlike a routing loop, does not endanger the integrity of the 1227 network. When a route is retracted, a node SHOULD send a triggered 1228 update and SHOULD make a reasonable attempt at ensuring that all 1229 neighbours receive this retraction. 1231 Finally, a node MAY send a triggered update when the metric for a 1232 given prefix changes in a significant manner, due to a received 1233 update, because a link's cost has changed, or because a different 1234 next hop has been selected. A node SHOULD NOT send triggered updates 1235 for other reasons, such as when there is a minor fluctuation in a 1236 route's metric, when the selected next hop changes, or to propagate a 1237 new sequence number (except to satisfy a request, as specified in 1238 Section 3.8). 1240 3.7.3. Maintaining Feasibility Distances 1242 Before sending an update (prefix, plen, router-id, seqno, metric) 1243 with finite metric (i.e., not a route retraction), a Babel node 1244 updates the feasibility distance maintained in the source table. 1245 This is done as follows. 1247 If no entry indexed by (prefix, plen, router-id) exists in the source 1248 table, then one is created with value (prefix, plen, router-id, 1249 seqno, metric). 1251 If an entry (prefix, plen, router-id, seqno', metric') exists, then 1252 it is updated as follows: 1254 o if seqno > seqno', then seqno' := seqno, metric' := metric; 1256 o if seqno = seqno' and metric' > metric, then metric' := metric; 1258 o otherwise, nothing needs to be done. 1260 The garbage-collection timer for the entry is then reset. Note that 1261 the feasibility distance is not updated and the garbage-collection 1262 timer is not reset when a retraction (an update with infinite metric) 1263 is sent. 1265 When the garbage-collection timer expires, the entry is removed from 1266 the source table. 1268 3.7.4. Split Horizon 1270 When running over a transitive, symmetric link technology, e.g., a 1271 point-to-point link or a wired LAN technology such as Ethernet, a 1272 Babel node SHOULD use an optimisation known as split horizon. When 1273 split horizon is used on a given interface, a routing update for 1274 prefix P is not sent on the particular interface over which the 1275 selected route towards prefix P was learnt. 1277 Split horizon SHOULD NOT be applied to an interface unless the 1278 interface is known to be symmetric and transitive; in particular, 1279 split horizon is not applicable to decentralised wireless link 1280 technologies (e.g., IEEE 802.11 in ad hoc mode) when routing updates 1281 are sent over multicast. 1283 3.8. Explicit Requests 1285 In normal operation, a node's route table is populated by the regular 1286 and triggered updates sent by its neighbours. Under some 1287 circumstances, however, a node sends explicit requests in order to 1288 cause a resynchronisation with the source after a mobility event or 1289 to prevent a route from spuriously expiring. 1291 The Babel protocol provides two kinds of explicit requests: route 1292 requests, which simply request an update for a given prefix, and 1293 seqno requests, which request an update for a given prefix with a 1294 specific sequence number. The former are never forwarded; the latter 1295 are forwarded if they cannot be satisfied by the receiver. 1297 3.8.1. Handling Requests 1299 Upon receiving a request, a node either forwards the request or sends 1300 an update in reply to the request, as described in the following 1301 sections. If this causes an update to be sent, the update is either 1302 sent to a multicast address on the interface on which the request was 1303 received, or to the unicast address of the neighbour that sent the 1304 request. 1306 The exact behaviour is different for route requests and seqno 1307 requests. 1309 3.8.1.1. Route Requests 1311 When a node receives a route request for a given prefix, it checks 1312 its route table for a selected route to this exact prefix. If such a 1313 route exists, it MUST send an update (over unicast or over 1314 multicast); if such a route does not exist, it MUST send a retraction 1315 for that prefix. 1317 When a node receives a wildcard route request, it SHOULD send a full 1318 route table dump. Full route dumps MAY be rate-limited, especially 1319 if they are sent over multicast. 1321 3.8.1.2. Seqno Requests 1323 When a node receives a seqno request for a given router-id and 1324 sequence number, it checks whether its route table contains a 1325 selected entry for that prefix. If a selected route for the given 1326 prefix exists, it has finite metric, and either the router-ids are 1327 different or the router-ids are equal and the entry's sequence number 1328 is no smaller (modulo 2^16) than the requested sequence number, the 1329 node MUST send an update for the given prefix. If the router-ids 1330 match but the requested seqno is larger (modulo 2^16) than the route 1331 entry's, the node compares the router-id against its own router-id. 1332 If the router-id is its own, then it increases its sequence number by 1333 1 (modulo 2^16) and sends an update. A node MUST NOT increase its 1334 sequence number by more than 1 in response to a seqno request. 1336 Otherwise, if the requested router-id is not its own, the received 1337 request's hop count is 2 or more, and the node is advertising the 1338 prefix to its neighbours, the node selects a neighbour to forward the 1339 request to as follows: 1341 o if the node has one or more feasible routes toward the requested 1342 prefix with a next hop that is not the requesting node, then the 1343 node MUST forward the request to the next hop of one such route; 1345 o otherwise, if the node has one or more (not necessarily feasible) 1346 routes to the requested prefix with a next hop that is not the 1347 requesting node, then the node SHOULD forward the request to the 1348 next hop of one such route. 1350 In order to actually forward the request, the node decrements the hop 1351 count and sends the request in a unicast packet destined to the 1352 selected neighbour. 1354 A node SHOULD maintain a list of recently forwarded seqno requests 1355 and forward the reply (an update with a seqno sufficiently large to 1356 satisfy the request) in a timely manner. A node SHOULD compare every 1357 incoming seqno request against its list of recently forwarded seqno 1358 requests and avoid forwarding it if it is redundant (i.e., if it has 1359 recently sent a request with the same prefix, router-id and a seqno 1360 that is not smaller modulo 2^16). 1362 Since the request-forwarding mechanism does not necessarily obey the 1363 feasibility condition, it may get caught in routing loops; hence, 1364 requests carry a hop count to limit the time during which they remain 1365 in the network. However, since requests are only ever forwarded as 1366 unicast packets, the initial hop count need not be kept particularly 1367 low, and performing an expanding horizon search is not necessary. A 1368 single request MUST NOT be duplicated: it MUST NOT be forwarded to a 1369 multicast address, and it MUST NOT be forwarded to multiple 1370 neighbours. However, if a seqno request is resent by its originator, 1371 the subsequent copies MAY be forwarded to a different neighbour than 1372 the initial one. 1374 3.8.2. Sending Requests 1376 A Babel node MAY send a route or seqno request at any time, to a 1377 multicast or a unicast address; there is only one case when 1378 originating requests is required (Section 3.8.2.1). 1380 3.8.2.1. Avoiding Starvation 1382 When a route is retracted or expires, a Babel node usually switches 1383 to another feasible route for the same prefix. It may be the case, 1384 however, that no such routes are available. 1386 A node that has lost all feasible routes to a given destination but 1387 still has unexpired unfeasible routes to that destination MUST send a 1388 seqno request; if it doesn't have any such routes, it MAY still send 1389 a seqno request. The router-id of the request is set to the router- 1390 id of the route that it has just lost, and the requested seqno is the 1391 value contained in the source table plus 1. 1393 If the node has any (unfeasible) routes to the requested destination, 1394 then it MUST send the request to at least one of the next-hop 1395 neighbours that advertised these routes, and SHOULD send it to all of 1396 them; in any case, it MAY send the request to any other neighbours, 1397 whether they advertise a route to the requested destination or not. 1398 A simple implementation strategy is therefore to unconditionally 1399 multicast the request over all interfaces. 1401 Similar requests will be sent by other nodes that are affected by the 1402 route's loss. If the network is still connected, and assuming no 1403 packet loss, then at least one of these requests will be forwarded to 1404 the source, resulting in a route being advertised with a new sequence 1405 number. (Due to duplicate suppression, only a small number of such 1406 requests will actually reach the source.) 1408 In order to compensate for packet loss, a node SHOULD repeat such a 1409 request a small number of times if no route becomes feasible within a 1410 short time. In the presence of heavy packet loss, however, all such 1411 requests might be lost; in that case, the mechanism in the next 1412 section will eventually ensure that a new seqno is received. 1414 3.8.2.2. Dealing with Unfeasible Updates 1416 When a route's metric increases, a node might receive an unfeasible 1417 update for a route that it has currently selected. As specified in 1418 Section 3.5.1, the receiving node will either ignore the update or 1419 unselect the route. 1421 In order to keep routes from spuriously expiring because they have 1422 become unfeasible, a node SHOULD send a unicast seqno request when it 1423 receives an unfeasible update for a route that is currently selected. 1424 The requested sequence number is computed from the source table as in 1425 Section 3.8.2.1 above. 1427 Additionally, since metric computation does not necessarily coincide 1428 with the delay in propagating updates, a node might receive an 1429 unfeasible update from a currently unselected neighbour that is 1430 preferable to the currently selected route (e.g., because it has a 1431 much smaller metric); in that case, the node SHOULD send a unicast 1432 seqno request to the neighbour that advertised the preferable update. 1434 3.8.2.3. Preventing Routes from Expiring 1436 In normal operation, a route's expiry timer never triggers: since a 1437 route's hold time is computed from an explicit interval included in 1438 Update TLVs, a new update (possibly a retraction) should arrive in 1439 time to prevent a route from expiring. 1441 In the presence of packet loss, however, it may be the case that no 1442 update is successfully received for an extended period of time, 1443 causing a route to expire. In order to avoid such spurious expiry, 1444 shortly before a selected route expires, a Babel node SHOULD send a 1445 unicast route request to the neighbour that advertised this route; 1446 since nodes always send either updates or retractions in response to 1447 non-wildcard route requests (Section 3.8.1.1), this will usually 1448 result in the route being either refreshed or retracted. 1450 3.8.2.4. Acquiring New Neighbours 1452 In order to speed up convergence after a mobility event, a node MAY 1453 send a unicast wildcard request after acquiring a new neighbour. 1454 Additionally, a node MAY send a small number of multicast wildcard 1455 requests shortly after booting. Note however that doing that 1456 carelessly can cause serious congestion when a whole network is 1457 rebooted, especially on link layers with high per-packet overhead 1458 (e.g., IEEE 802.11). 1460 4. Protocol Encoding 1462 A Babel packet is sent as the body of a UDP datagram, with network- 1463 layer hop count set to 1, destined to a well-known multicast address 1464 or to a unicast address, over IPv4 or IPv6; in the case of IPv6, 1465 these addresses are link-local. Both the source and destination UDP 1466 port are set to a well-known port number. A Babel packet MUST be 1467 silently ignored unless its source address is either a link-local 1468 IPv6 address or an IPv4 address belonging to the local network, and 1469 its source port is the well-known Babel port. It MAY be silently 1470 ignored if its destination address is a global IPv6 address. 1472 In order to minimise the number of packets being sent while avoiding 1473 lower-layer fragmentation, a Babel node SHOULD attempt to maximise 1474 the size of the packets it sends, up to the outgoing interface's MTU 1475 adjusted for lower-layer headers (28 octets for UDP over IPv4, 48 1476 octets for UDP over IPv6). It MUST NOT send packets larger than the 1477 attached interface's MTU adjusted for lower-layer headers or 512 1478 octets, whichever is larger, but not exceeding 2^16 - 1 adjusted for 1479 lower-layer headers. Every Babel speaker MUST be able to receive 1480 packets that are as large as any attached interface's MTU adjusted 1481 for lower-layer headers or 512 octets, whichever is larger. Babel 1482 packets MUST NOT be sent in IPv6 Jumbograms. 1484 In order to avoid global synchronisation of a Babel network and to 1485 aggregate multiple TLVs into large packets, a Babel node SHOULD 1486 buffer every TLV and delay sending a packet by a small, randomly 1487 chosen delay [JITTER]. In order to allow accurate computation of 1488 packet loss rates, this delay MUST NOT be larger than half the 1489 advertised Hello interval. 1491 4.1. Data Types 1493 4.1.1. Interval 1495 Relative times are carried as 16-bit values specifying a number of 1496 centiseconds (hundredths of a second). This allows times up to 1497 roughly 11 minutes with a granularity of 10ms, which should cover all 1498 reasonable applications of Babel. 1500 4.1.2. Router-Id 1502 A router-id is an arbitrary 8-octet value. A router-id MUST NOT 1503 consist of either all zeroes or all ones. 1505 4.1.3. Address 1507 Since the bulk of the protocol is taken by addresses, multiple ways 1508 of encoding addresses are defined. Additionally, a common subnet 1509 prefix may be omitted when multiple addresses are sent in a single 1510 packet -- this is known as address compression (Section 4.6.9). 1512 Address encodings: 1514 o AE 0: wildcard address. The value is 0 octets long. 1516 o AE 1: IPv4 address. Compression is allowed. 4 octets or less. 1518 o AE 2: IPv6 address. Compression is allowed. 16 octets or less. 1520 o AE 3: link-local IPv6 address. Compression is not allowed. The 1521 value is 8 octets long, a prefix of fe80::/64 is implied. 1523 The address family associated to an address encoding is either IPv4 1524 or IPv6; it is undefined for AE 0, IPv4 for AE 1, and IPv6 for AEs 2 1525 and 3. 1527 4.1.4. Prefixes 1529 A network prefix is encoded just like a network address, but it is 1530 stored in the smallest number of octets that are enough to hold the 1531 significant bits (up to the prefix length). 1533 4.2. Packet Format 1535 A Babel packet consists of a 4-octet header, followed by a sequence 1536 of TLVs (the packet body), optionally followed by a second sequence 1537 of TLVs (the packet trailer). 1539 0 1 2 3 1540 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 1541 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1542 | Magic | Version | Body length | 1543 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1544 | Packet Body ... 1545 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1546 | Packet Trailer... 1547 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1549 Fields : 1551 Magic The arbitrary but carefully chosen value 42 (decimal); 1552 packets with a first octet different from 42 MUST be 1553 silently ignored. 1555 Version This document specifies version 2 of the Babel protocol. 1556 Packets with a second octet different from 2 MUST be 1557 silently ignored. 1559 Body length The length in octets of the body following the packet 1560 header (excluding the Magic, Version and Body length 1561 fields, and excluding the packet trailer). 1563 Packet Body The packet body; a sequence of TLVs. 1565 Packet Trailer The packet trailer; another sequence of TLVs. 1567 The packet body and trailer are both sequences of TLVs. The packet 1568 body is the normal place to store TLVs; the packet trailer only 1569 contains specialised TLVs that do not need to be protected by 1570 cryptographic security mechanisms. When parsing the trailer, the 1571 receiver MUST ignore any TLV unless its definition explicitly states 1572 that it is allowed to appear there. Among the TLVs defined in this 1573 document, only Pad1 and PadN are allowed in the trailer; since these 1574 TLVs are ignored in any case, an implementation MAY silently ignore 1575 the packet trailer without even parsing it, unless it implements at 1576 least one extension that defines TLVs that are allowed to appear in 1577 the trailer. 1579 4.3. TLV Format 1581 With the exception of Pad1, all TLVs have the following structure: 1583 0 1 2 3 1584 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 1585 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1586 | Type | Length | Payload... 1587 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1589 Fields : 1591 Type The type of the TLV. 1593 Length The length of the body, exclusive of the Type and Length 1594 fields. If the body is longer than the expected length of 1595 a given type of TLV, any extra data MUST be silently 1596 ignored. 1598 Payload The TLV payload, which consists of a body and, for selected 1599 TLV types, an optional list of sub-TLVs. 1601 TLVs with an unknown type value MUST be silently ignored. 1603 4.4. Sub-TLV Format 1605 Every TLV carries an explicit length in its header; however, most 1606 TLVs are self-terminating, in the sense that it is possible to 1607 determine the length of the body without reference to the explicit 1608 Length field. If a TLV has a self-terminating format, then it MAY 1609 allow a sequence of sub-TLVs to follow the body. 1611 Sub-TLVs have the same structure as TLVs. With the exception of 1612 PAD1, all TLVs have the following structure: 1614 0 1 2 3 1615 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 1616 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1617 | Type | Length | Body... 1618 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1620 Fields : 1622 Type The type of the sub-TLV. 1624 Length The length of the body, in octets, exclusive of the Type 1625 and Length fields. 1627 Body The sub-TLV body, the interpretation of which depends on 1628 both the type of the sub-TLV and the type of the TLV within 1629 which it is embedded. 1631 The most-significant bit of the sub-TLV, called the mandatory bit, 1632 indicates how to handle unknown sub-TLVs. If the mandatory bit is 1633 not set, then an unknown sub-TLV MUST be silently ignored, and the 1634 rest of the TLV processed normally. If the mandatory bit is set, 1635 then the whole enclosing TLV MUST be silently ignored (except for 1636 updating the parser state by a Router-Id, Next-Hop or Update TLV, see 1637 Section 4.6.7, Section 4.6.8, and Section 4.6.9). 1639 4.5. Parser state 1641 Babel uses a stateful parser: a TLV may refer to data from a previous 1642 TLV. The parser state consists of the following pieces of data: 1644 o for each address encoding that allows compression, the current 1645 default prefix; this is undefined at the start of the packet, and 1646 is updated by each Update TLV with the Prefix flag set 1647 (Section 4.6.9); 1649 o for each address family (IPv4 or IPv6), the current next-hop; this 1650 is the source address of the enclosing packet for the matching 1651 address family at the start of a packet, and is updated by each 1652 Next-Hop TLV (Section 4.6.8); 1654 o the current router-id; this is undefined at the start of the 1655 packet, and is updated by each Router-ID TLV (Section 4.6.7) and 1656 by each Update TLV with Router-Id flag set. 1658 Since the parser state is separate from the bulk of Babel's state, 1659 and since for correct parsing it must be identical across 1660 implementations, it is updated before checking for mandatory TLVs: 1661 parsing a TLV MUST update the parser state even if the TLV is 1662 otherwise ignored due to an unknown mandatory sub-TLV. 1664 None of the TLVs that modify the parser state are allowed in the 1665 packet trailer; hence, an implementation may choose to use a 1666 dedicated stateless parser to parse the packet trailer. 1668 4.6. Details of Specific TLVs 1670 4.6.1. Pad1 1672 0 1673 0 1 2 3 4 5 6 7 1674 +-+-+-+-+-+-+-+-+ 1675 | Type = 0 | 1676 +-+-+-+-+-+-+-+-+ 1678 Fields : 1680 Type Set to 0 to indicate a Pad1 TLV. 1682 This TLV is silently ignored on reception. It is allowed in the 1683 packet trailer. 1685 4.6.2. PadN 1687 0 1 2 3 1688 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 1689 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1690 | Type = 1 | Length | MBZ... 1691 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1693 Fields : 1695 Type Set to 1 to indicate a PadN TLV. 1697 Length The length of the body, exclusive of the Type and Length 1698 fields. 1700 MBZ Set to 0 on transmission. 1702 This TLV is silently ignored on reception. It is allowed in the 1703 packet trailer. 1705 4.6.3. Acknowledgment Request 1707 0 1 2 3 1708 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 1709 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1710 | Type = 2 | Length | Reserved | 1711 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1712 | Nonce | Interval | 1713 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1714 This TLV requests that the receiver send an Acknowledgment TLV within 1715 the number of centiseconds specified by the Interval field. 1717 Fields : 1719 Type Set to 2 to indicate an Acknowledgment Request TLV. 1721 Length The length of the body, exclusive of the Type and Length 1722 fields. 1724 Reserved Sent as 0 and MUST be ignored on reception. 1726 Nonce An arbitrary value that will be echoed in the receiver's 1727 Acknowledgment TLV. 1729 Interval A time interval in centiseconds after which the sender will 1730 assume that this packet has been lost. This MUST NOT be 0. 1731 The receiver MUST send an Acknowledgment TLV before this 1732 time has elapsed (with a margin allowing for propagation 1733 time). 1735 This TLV is self-terminating, and allows sub-TLVs. 1737 4.6.4. Acknowledgment 1739 0 1 2 3 1740 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 1741 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1742 | Type = 3 | Length | Nonce | 1743 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1745 This TLV is sent by a node upon receiving an Acknowledgment Request. 1747 Fields : 1749 Type Set to 3 to indicate an Acknowledgment TLV. 1751 Length The length of the body, exclusive of the Type and Length 1752 fields. 1754 Nonce Set to the Nonce value of the Acknowledgment Request that 1755 prompted this Acknowledgment. 1757 Since nonce values are not globally unique, this TLV MUST be sent to 1758 a unicast address. 1760 This TLV is self-terminating, and allows sub-TLVs. 1762 4.6.5. Hello 1764 0 1 2 3 1765 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 1766 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1767 | Type = 4 | Length | Flags | 1768 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1769 | Seqno | Interval | 1770 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1772 This TLV is used for neighbour discovery and for determining a 1773 neighbour's reception cost. 1775 Fields : 1777 Type Set to 4 to indicate a Hello TLV. 1779 Length The length of the body, exclusive of the Type and Length 1780 fields. 1782 Flags The individual bits of this field specify special handling 1783 of this TLV (see below). 1785 Seqno If the Unicast flag is set, this is the value of the 1786 sending node's outgoing Unicast Hello seqno for this 1787 neighbour. Otherwise, it is the sending node's outgoing 1788 Multicast Hello seqno for this interface. 1790 Interval If non-zero, this is an upper bound, expressed in 1791 centiseconds, on the time after which the sending node will 1792 send a new scheduled Hello TLV with the same setting of the 1793 Unicast flag. If this is 0, then this Hello represents an 1794 unscheduled Hello, and doesn't carry any new information 1795 about times at which Hellos are sent. 1797 The Flags field is interpreted as follows: 1799 0 1 1800 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 1801 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1802 |U|X|X|X|X|X|X|X|X|X|X|X|X|X|X|X| 1803 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1805 o U (Unicast) flag (8000 hexadecimal): if set, then this Hello 1806 represents a Unicast Hello, otherwise it represents a Multicast 1807 Hello; 1809 o X: all other bits MUST be sent as 0 and silently ignored on 1810 reception. 1812 Every time a Hello is sent, the corresponding seqno counter MUST be 1813 incremented. Since there is a single seqno counter for all the 1814 Multicast Hellos sent by a given node over a given interface, if the 1815 Unicast flag is not set, this TLV MUST be sent to all neighbors on 1816 this link, which can be achieved by sending to a multicast 1817 destination, or by sending multiple packets to the unicast addresses 1818 of all reachable neighbours. Conversely, if the Unicast flag is set, 1819 this TLV MUST be sent to a single neighbour, which can achieved by 1820 sending to a unicast destination. In order to avoid large 1821 discontinuities in link quality, multiple Hello TLVs SHOULD NOT be 1822 sent in the same packet. 1824 This TLV is self-terminating, and allows sub-TLVs. 1826 4.6.6. IHU 1828 0 1 2 3 1829 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 1830 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1831 | Type = 5 | Length | AE | Reserved | 1832 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1833 | Rxcost | Interval | 1834 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1835 | Address... 1836 +-+-+-+-+-+-+-+-+-+-+-+- 1838 An IHU ("I Heard You") TLV is used for confirming bidirectional 1839 reachability and carrying a link's transmission cost. 1841 Fields : 1843 Type Set to 5 to indicate an IHU TLV. 1845 Length The length of the body, exclusive of the Type and Length 1846 fields. 1848 AE The encoding of the Address field. This should be 1 or 3 1849 in most cases. As an optimisation, it MAY be 0 if the TLV 1850 is sent to a unicast address, if the association is over a 1851 point-to-point link, or when bidirectional reachability is 1852 ascertained by means outside of the Babel protocol. 1854 Reserved Sent as 0 and MUST be ignored on reception. 1856 Rxcost The rxcost according to the sending node of the interface 1857 whose address is specified in the Address field. The value 1858 FFFF hexadecimal (infinity) indicates that this interface 1859 is unreachable. 1861 Interval An upper bound, expressed in centiseconds, on the time 1862 after which the sending node will send a new IHU; this MUST 1863 NOT be 0. The receiving node will use this value in order 1864 to compute a hold time for this symmetric association. 1866 Address The address of the destination node, in the format 1867 specified by the AE field. Address compression is not 1868 allowed. 1870 Conceptually, an IHU is destined to a single neighbour. However, IHU 1871 TLVs contain an explicit destination address, and MAY be sent to a 1872 multicast address, as this allows aggregation of IHUs destined to 1873 distinct neighbours into a single packet and avoids the need for an 1874 ARP or Neighbour Discovery exchange when a neighbour is not being 1875 used for data traffic. 1877 IHU TLVs with an unknown value in the AE field MUST be silently 1878 ignored. 1880 This TLV is self-terminating, and allows sub-TLVs. 1882 4.6.7. Router-Id 1884 0 1 2 3 1885 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 1886 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1887 | Type = 6 | Length | Reserved | 1888 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1889 | | 1890 + Router-Id + 1891 | | 1892 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1894 A Router-Id TLV establishes a router-id that is implied by subsequent 1895 Update TLVs. This TLV sets the router-id even if it is otherwise 1896 ignored due to an unknown mandatory sub-TLV. 1898 Fields : 1900 Type Set to 6 to indicate a Router-Id TLV. 1902 Length The length of the body, exclusive of the Type and Length 1903 fields. 1905 Reserved Sent as 0 and MUST be ignored on reception. 1907 Router-Id The router-id for routes advertised in subsequent Update 1908 TLVs. This MUST NOT consist of all zeroes or all ones. 1910 This TLV is self-terminating, and allows sub-TLVs. 1912 4.6.8. Next Hop 1914 0 1 2 3 1915 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 1916 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1917 | Type = 7 | Length | AE | Reserved | 1918 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1919 | Next hop... 1920 +-+-+-+-+-+-+-+-+-+-+-+- 1922 A Next Hop TLV establishes a next-hop address for a given address 1923 family (IPv4 or IPv6) that is implied in subsequent Update TLVs. 1924 This TLV sets up the next-hop for subsequent Update TLVs even if it 1925 is otherwise ignored due to an unknown mandatory sub-TLV. 1927 Fields : 1929 Type Set to 7 to indicate a Next Hop TLV. 1931 Length The length of the body, exclusive of the Type and Length 1932 fields. 1934 AE The encoding of the Address field. This SHOULD be 1 (IPv4) 1935 or 3 (link-local IPv6), and MUST NOT be 0. 1937 Reserved Sent as 0 and MUST be ignored on reception. 1939 Next hop The next-hop address advertised by subsequent Update TLVs, 1940 for this address family. 1942 When the address family matches the network-layer protocol that this 1943 packet is transported over, a Next Hop TLV is not needed: in the 1944 absence of a Next Hop TLV in a given address family, the next hop 1945 address is taken to be the source address of the packet. 1947 Next Hop TLVs with an unknown value for the AE field MUST be silently 1948 ignored. 1950 This TLV is self-terminating, and allows sub-TLVs. 1952 4.6.9. Update 1954 0 1 2 3 1955 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 1956 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1957 | Type = 8 | Length | AE | Flags | 1958 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1959 | Plen | Omitted | Interval | 1960 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1961 | Seqno | Metric | 1962 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1963 | Prefix... 1964 +-+-+-+-+-+-+-+-+-+-+-+- 1966 An Update TLV advertises or retracts a route. As an optimisation, it 1967 can optionally have the side effect of establishing a new implied 1968 router-id and a new default prefix. 1970 Fields : 1972 Type Set to 8 to indicate an Update TLV. 1974 Length The length of the body, exclusive of the Type and Length 1975 fields. 1977 AE The encoding of the Prefix field. 1979 Flags The individual bits of this field specify special handling 1980 of this TLV (see below). 1982 Plen The length of the advertised prefix. 1984 Omitted The number of octets that have been omitted at the 1985 beginning of the advertised prefix and that should be taken 1986 from a preceding Update TLV in the same address family with 1987 the Prefix flag set. 1989 Interval An upper bound, expressed in centiseconds, on the time 1990 after which the sending node will send a new update for 1991 this prefix. This MUST NOT be 0. The receiving node will 1992 use this value to compute a hold time for the route table 1993 entry. The value FFFF hexadecimal (infinity) expresses 1994 that this announcement will not be repeated unless a 1995 request is received (Section 3.8.2.3). 1997 Seqno The originator's sequence number for this update. 1999 Metric The sender's metric for this route. The value FFFF 2000 hexadecimal (infinity) means that this is a route 2001 retraction. 2003 Prefix The prefix being advertised. This field's size is 2004 (Plen/8 - Omitted) rounded upwards. 2006 The Flags field is interpreted as follows: 2008 0 1 2 3 4 5 6 7 2009 +-+-+-+-+-+-+-+-+ 2010 |P|R|X|X|X|X|X|X| 2011 +-+-+-+-+-+-+-+-+ 2013 o P (Prefix) flag (80 hexadecimal): if set, then this Update 2014 establishes a new default prefix for subsequent Update TLVs with a 2015 matching address encoding within the same packet, even if this TLV 2016 is otherwise ignored due to an unknown mandatory sub-TLV; 2018 o R (Router-Id) flag (40 hexadecimal): if set, then this TLV 2019 establishes a new default router-id for this TLV and subsequent 2020 Update TLVs in the same packet, even if this TLV is otherwise 2021 ignored due to an unknown mandatory sub-TLV. This router-id is 2022 computed from the first address of the advertised prefix as 2023 follows: 2025 * if the length of the address is 8 octets or more, then the new 2026 router-id is taken from the 8 last octets of the address; 2028 * if the length of the address is smaller than 8 octets, then the 2029 new router-id consists of the required number of zero octets 2030 followed by the address, i.e., the address is stored on the 2031 right of the router-id. For example, for an IPv4 address, the 2032 router-id consists of 4 octets of zeroes followed by the IPv4 2033 address. 2035 o X: all other bits MUST be sent as 0 and silently ignored on 2036 reception. 2038 The prefix being advertised by an Update TLV is computed as follows: 2040 o the first Omitted octets of the prefix are taken from the previous 2041 Update TLV with the Prefix flag set and the same address encoding, 2042 even if it was ignored due to an unknown mandatory sub-TLV; 2044 o the next (Plen/8 - Omitted) rounded upwards octets are taken from 2045 the Prefix field; 2047 o the remaining octets are set to 0. If AE is 3 (link-local IPv6), 2048 Omitted MUST be 0) 2050 If the Metric field is finite, the router-id of the originating node 2051 for this announcement is taken from the prefix advertised by this 2052 Update if the Router-Id flag is set, computed as described above. 2053 Otherwise, it is taken either from the preceding Router-Id packet, or 2054 the preceding Update packet with the Router-Id flag set, whichever 2055 comes last, even if that TLV is otherwise ignored due to an unknown 2056 mandatory sub-TLV. 2058 The next-hop address for this update is taken from the last preceding 2059 Next Hop TLV with a matching address family (IPv4 or IPv6) in the 2060 same packet even if it was otherwise ignored due to an unknown 2061 mandatory sub-TLV; if no such TLV exists, it is taken from the 2062 network-layer source address of this packet. 2064 If the metric field is FFFF hexadecimal, this TLV specifies a 2065 retraction. In that case, the router-id, next-hop and seqno are not 2066 used. AE MAY then be 0, in which case this Update retracts all of 2067 the routes previously advertised by the sending interface. If the 2068 metric is finite, AE MUST NOT be 0. If the metric is infinite and AE 2069 is 0, Plen and Omitted MUST both be 0. 2071 Update TLVs with an unknown value in the AE field MUST be silently 2072 ignored. 2074 This TLV is self-terminating, and allows sub-TLVs. 2076 4.6.10. Route Request 2078 0 1 2 3 2079 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 2080 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2081 | Type = 9 | Length | AE | Plen | 2082 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2083 | Prefix... 2084 +-+-+-+-+-+-+-+-+-+-+-+- 2086 A Route Request TLV prompts the receiver to send an update for a 2087 given prefix, or a full route table dump. 2089 Fields : 2091 Type Set to 9 to indicate a Route Request TLV. 2093 Length The length of the body, exclusive of the Type and Length 2094 fields. 2096 AE The encoding of the Prefix field. The value 0 specifies 2097 that this is a request for a full route table dump (a 2098 wildcard request). 2100 Plen The length of the requested prefix. 2102 Prefix The prefix being requested. This field's size is Plen/8 2103 rounded upwards. 2105 A Request TLV prompts the receiver to send an update message 2106 (possibly a retraction) for the prefix specified by the AE, Plen, and 2107 Prefix fields, or a full dump of its route table if AE is 0 (in which 2108 case Plen MUST be 0 and Prefix is of length 0). 2110 This TLV is self-terminating, and allows sub-TLVs. 2112 4.6.11. Seqno Request 2114 0 1 2 3 2115 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 2116 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2117 | Type = 10 | Length | AE | Plen | 2118 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2119 | Seqno | Hop Count | Reserved | 2120 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2121 | | 2122 + Router-Id + 2123 | | 2124 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2125 | Prefix... 2126 +-+-+-+-+-+-+-+-+-+-+ 2128 A Seqno Request TLV prompts the receiver to send an Update for a 2129 given prefix with a given sequence number, or to forward the request 2130 further if it cannot be satisfied locally. 2132 Fields : 2134 Type Set to 10 to indicate a Seqno Request message. 2136 Length The length of the body, exclusive of the Type and Length 2137 fields. 2139 AE The encoding of the Prefix field. This MUST NOT be 0. 2141 Plen The length of the requested prefix. 2143 Seqno The sequence number that is being requested. 2145 Hop Count The maximum number of times that this TLV may be forwarded, 2146 plus 1. This MUST NOT be 0. 2148 Reserved Sent as 0 and MUST be ignored on reception. 2150 Router Id The Router-Id that is being requested. This MUST NOT 2151 consist of all zeroes or all ones. 2153 Prefix The prefix being requested. This field's size is Plen/8 2154 rounded upwards. 2156 A Seqno Request TLV prompts the receiving node to send a finite- 2157 metric Update for the prefix specified by the AE, Plen, and Prefix 2158 fields, with either a router-id different from what is specified by 2159 the Router-Id field, or a Seqno no less (modulo 2^16) than what is 2160 specified by the Seqno field. If this request cannot be satisfied 2161 locally, then it is forwarded according to the rules set out in 2162 Section 3.8.1.2. 2164 While a Seqno Request MAY be sent to a multicast address, it MUST NOT 2165 be forwarded to a multicast address and MUST NOT be forwarded to more 2166 than one neighbour. A request MUST NOT be forwarded if its Hop Count 2167 field is 1. 2169 This TLV is self-terminating, and allows sub-TLVs. 2171 4.7. Details of specific sub-TLVs 2173 4.7.1. Pad1 2175 0 1 2 3 4 5 6 7 2176 +-+-+-+-+-+-+-+-+ 2177 | Type = 0 | 2178 +-+-+-+-+-+-+-+-+ 2180 Fields : 2182 Type Set to 0 to indicate a Pad1 sub-TLV. 2184 This sub-TLV is silently ignored on reception. It is allowed within 2185 any TLV that allows sub-TLVs. 2187 4.7.2. PadN 2188 0 1 2 3 2189 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 2190 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2191 | Type = 1 | Length | MBZ... 2192 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 2194 Fields : 2196 Type Set to 1 to indicate a PadN sub-TLV. 2198 Length The length of the body, in octets, exclusive of the Type 2199 and Length fields. 2201 MBZ Set to 0 on transmission. 2203 This sub-TLV is silently ignored on reception. It is allowed within 2204 any TLV that allows sub-TLVs. 2206 5. IANA Considerations 2208 IANA has registered the UDP port number 6696, called "babel", for use 2209 by the Babel protocol. 2211 IANA has registered the IPv6 multicast group ff02::1:6 and the IPv4 2212 multicast group 224.0.0.111 for use by the Babel protocol. 2214 IANA has created a registry called "Babel TLV Types". The values in 2215 this registry are not changed by this specification. 2217 IANA has created a registry called "Babel sub-TLV Types". Due to the 2218 addition of a Mandatory bit to the Babel protocol, the values in the 2219 "Babel sub-TLV Types" registry are amended as follows: 2221 +---------+-----------------------------------------+---------------+ 2222 | Type | Name | Reference | 2223 +---------+-----------------------------------------+---------------+ 2224 | 0 | Pad1 | this document | 2225 | | | | 2226 | 1 | PadN | this document | 2227 | | | | 2228 | 112-126 | Reserved for Experimental Use | this document | 2229 | | | | 2230 | 127 | Reserved for expansion of the type | this document | 2231 | | space | | 2232 | | | | 2233 | 240-254 | Reserved for Experimental Use | this document | 2234 | | | | 2235 | 255 | Reserved for expansion of the type | this document | 2236 | | space | | 2237 +---------+-----------------------------------------+---------------+ 2239 Existing assignments in the "Babel sub-TLV Types" registry in the 2240 range 2 to 111 are not changed by this specification. The values 224 2241 through 239, previously reserved for Experimental Use, are now 2242 unassigned. 2244 IANA has created a registry called "Babel Flags Values". IANA is 2245 instructed to rename this registry to "Babel Update Flags Values", 2246 with its contents unchanged. 2248 IANA is instructed to create a new registry called "Babel Hello Flags 2249 Values". The allocation policy for this registry is Specification 2250 Required [RFC8126]. The initial values in this registry are as 2251 follows: 2253 +------+------------+---------------+ 2254 | Bit | Name | Reference | 2255 +------+------------+---------------+ 2256 | 0 | Unicast | this document | 2257 | | | | 2258 | 1-15 | Unassigned | | 2259 +------+------------+---------------+ 2261 IANA is instructed to replace all references to RFCs 6126 and 7557 in 2262 all of the registries mentioned above by references to this document. 2264 6. Security Considerations 2266 As defined in this document, Babel is a completely insecure protocol. 2267 Any attacker can misdirect data traffic by advertising routes with a 2268 low metric or a high seqno. This issue can be solved either by a 2269 lower-layer security mechanism (e.g., link-layer security or IPsec), 2270 or by deploying a suitable authentication mechanism within Babel 2271 itself. There are currently two such mechanisms: Babel over DTLS 2272 [BABEL-DTLS] and HMAC-based authentication [BABEL-HMAC]. Both 2273 mechanisms ensure integrity of messages and prevent message replay, 2274 but only DTLS supports asymmetric keying and message confidentiality. 2275 HMAC is simpler and does not depend on DTLS, and therefore its use is 2276 RECOMMENDED whenever both mechanisms are applicable. 2278 The information that a Babel node announces to the whole routing 2279 domain is often sufficient to determine a mobile node's physical 2280 location with reasonable precision. The privacy issues that this 2281 causes can be mitigated somewhat by using randomly chosen router-ids 2282 and randomly chosen IP addresses, and changing them periodically. 2284 When carried over IPv6, Babel packets are ignored unless they are 2285 sent from a link-local IPv6 address; since routers don't forward 2286 link-local IPv6 packets, this provides protection against spoofed 2287 Babel packets being sent from the global Internet. No such natural 2288 protection exists when Babel packets are carried over IPv4. 2290 7. Acknowledgments 2292 A number of people have contributed text and ideas to this 2293 specification. The authors are particularly indebted to Matthieu 2294 Boutier, Gwendoline Chouasne, Margaret Cullen, Donald Eastlake and 2295 Toke Hoiland-Jorgensen. Earlier versions of this document greatly 2296 benefited from the input of Joel Halpern. The address compression 2297 technique was inspired by [PACKETBB]. 2299 8. References 2301 8.1. Normative References 2303 [BABEL-DTLS] 2304 Decimo, A., Schinazi, D., and J. Chroboczek, "Babel 2305 Routing Protocol over Datagram Transport Layer Security", 2306 Internet Draft draft-ietf-babel-dtls-04, February 2019. 2308 [BABEL-HMAC] 2309 Do, C., Kolodziejak, W., and J. Chroboczek, "HMAC 2310 authentication for the Babel routing protocol", Internet 2311 Draft draft-ietf-babel-hmac-04, March 2019. 2313 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2314 Requirement Levels", BCP 14, RFC 2119, 2315 DOI 10.17487/RFC2119, March 1997. 2317 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2318 Writing an IANA Considerations Section in RFCs", BCP 26, 2319 RFC 8126, June 2017. 2321 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2322 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 2323 May 2017. 2325 8.2. Informative References 2327 [DSDV] Perkins, C. and P. Bhagwat, "Highly Dynamic Destination- 2328 Sequenced Distance-Vector Routing (DSDV) for Mobile 2329 Computers", ACM SIGCOMM'94 Conference on Communications 2330 Architectures, Protocols and Applications 234-244, 1994. 2332 [DUAL] Garcia Luna Aceves, J., "Loop-Free Routing Using Diffusing 2333 Computations", IEEE/ACM Transactions on Networking 1:1, 2334 February 1993. 2336 [EIGRP] Albrightson, B., Garcia Luna Aceves, J., and J. Boyle, 2337 "EIGRP -- a Fast Routing Protocol Based on Distance 2338 Vectors", Proc. Interop 94, 1994. 2340 [ETX] De Couto, D., Aguayo, D., Bicket, J., and R. Morris, "A 2341 high-throughput path metric for multi-hop wireless 2342 networks", Proc. MobiCom 2003, 2003. 2344 [IS-IS] Standardization, I. O. F., "Information technology -- 2345 Telecommunications and information exchange between 2346 systems -- Intermediate System to Intermediate System 2347 intra-domain routeing information exchange protocol for 2348 use in conjunction with the protocol for providing the 2349 connectionless-mode network service (ISO 8473)", ISO/ 2350 IEC 10589:2002, 2002. 2352 [JITTER] Floyd, S. and V. Jacobson, "The synchronization of 2353 periodic routing messages", IEEE/ACM Transactions on 2354 Networking 2, 2, 122-136, April 1994. 2356 [OSPF] Moy, J., "OSPF Version 2", RFC 2328, April 1998. 2358 [PACKETBB] 2359 Clausen, T., Dearlove, C., Dean, J., and C. Adjih, 2360 "Generalized Mobile Ad Hoc Network (MANET) Packet/Message 2361 Format", RFC 5444, February 2009. 2363 [RIP] Malkin, G., "RIP Version 2", RFC 2453, November 1998. 2365 Appendix A. Cost and Metric Computation 2367 The strategy for computing link costs and route metrics is a local 2368 matter; Babel itself only requires that it comply with the conditions 2369 given in Section 3.4.3 and Section 3.5.2. Different nodes may use 2370 different strategies in a single network and may use different 2371 strategies on different interface types. This section describes the 2372 strategies used by the sample implementation of Babel. 2374 The sample implementation of Babel sends periodic Multicast Hellos, 2375 and never sends Unicast Hellos. It maintains statistics about the 2376 last 16 received Hello TLVs of each kind (Appendix A.1), computes 2377 costs by using the 2-out-of-3 strategy (Appendix A.2.1) on wired 2378 links, and ETX (Appendix A.2.2) on wireless links. It uses an 2379 additive algebra for metric computation (Appendix A.3.1). 2381 A.1. Maintaining Hello History 2383 For each neighbour, the sample implementation of Babel maintains two 2384 sets of Hello history, one for each kind of Hello, and an expected 2385 sequence number, one for Multicast and one for Unicast Hellos. Each 2386 Hello history is a vector of 16 bits, where a 1 value represents a 2387 received Hello, and a 0 value a missed Hello. For each kind of 2388 Hello, the expected sequence number, written ne, is the sequence 2389 number that is expected to be carried by the next received Hello from 2390 this neighbour. 2392 Whenever it receives a Hello packet of a given kind from a neighbour, 2393 a node compares the received sequence number nr for that kind of 2394 Hello with its expected sequence number ne. Depending on the outcome 2395 of this comparison, one of the following actions is taken: 2397 o if the two differ by more than 16 (modulo 2^16), then the sending 2398 node has probably rebooted and lost its sequence number; the whole 2399 associated neighbour table entry is flushed and a new one is 2400 created; 2402 o otherwise, if the received nr is smaller (modulo 2^16) than the 2403 expected sequence number ne, then the sending node has increased 2404 its Hello interval without us noticing; the receiving node removes 2405 the last (ne - nr) entries from this neighbour's Hello history (we 2406 "undo history"); 2408 o otherwise, if nr is larger (modulo 2^16) than ne, then the sending 2409 node has decreased its Hello interval, and some Hellos were lost; 2410 the receiving node adds (nr - ne) 0 bits to the Hello history (we 2411 "fast-forward"). 2413 The receiving node then appends a 1 bit to the Hello history and sets 2414 ne to (nr + 1). If the Interval field of the received Hello is not 2415 zero, it resets the neighbour's hello timer to 1.5 times the 2416 advertised Interval (the extra margin allows for delay due to 2417 jitter). 2419 Whenever either Hello timer associated to a neighbour expires, the 2420 local node adds a 0 bit to this neighbour's Hello history, and 2421 increments the expected Hello number. If both Hello histories are 2422 empty (they contain 0 bits only), the neighbour entry is flushed; 2423 otherwise, the relevant hello timer is reset to the value advertised 2424 in the last Hello of that kind received from this neighbour (no extra 2425 margin is necessary in this case, since jitter was already taken into 2426 account when computing the timeout that has just expired). 2428 A.2. Cost Computation 2430 This section discusses how to compute costs based on Hello history. 2432 A.2.1. k-out-of-j 2434 K-out-of-j link sensing is suitable for wired links that are either 2435 up, in which case they only occasionally drop a packet, or down, in 2436 which case they drop all packets. 2438 The k-out-of-j strategy is parameterised by two small integers k and 2439 j, such that 0 < k <= j, and the nominal link cost, a constant K >= 2440 1. A node keeps a history of the last j hellos; if k or more of 2441 those have been correctly received, the link is assumed to be up, and 2442 the rxcost is set to K; otherwise, the link is assumed to be down, 2443 and the rxcost is set to infinity. 2445 Since Babel supports two kinds of Hellos, a Babel node performs k- 2446 out-of-j twice for each neighbour, once on the Unicast and once on 2447 the Multicast Hello history. If either of the instances of k-out- 2448 of-j indicates that the link is up, then the link is assumed to be 2449 up, and the rxcost is set to K; if both instances indicate that the 2450 link is down, then the link is assumed to be down, and the rxcost is 2451 set to infinity. In other words, the resulting rxcost is the minimum 2452 of the rxcosts yielded by the two instances of k-out-of-j link 2453 sensing. 2455 The cost of a link performing k-out-of-j link sensing is defined as 2456 follows: 2458 o cost = FFFF hexadecimal if rxcost = FFFF hexadecimal; 2460 o cost = txcost otherwise. 2462 A.2.2. ETX 2464 Unlike wired links, which are bimodal (either up or down), wireless 2465 links exhibit continuous variation of the link quality. Naive 2466 application of hop-count routing in networks that use wireless links 2467 for transit tends to select long, lossy links in preference to 2468 shorter, lossless links, which can dramatically reduce throughput. 2469 For that reason, a routing protocol designed to support wireless 2470 links must perform some form of link-quality estimation. 2472 ETX [ETX] is a simple link-quality estimation algorithm that is 2473 designed to work well with the IEEE 802.11 MAC. By default, the 2474 IEEE 802.11 MAC performs ARQ and rate adaptation on unicast frames, 2475 but not on multicast frames, which are sent at a fixed rate with no 2476 ARQ; therefore, measuring the loss rate of multicast frames yields a 2477 useful estimate of a link's quality. 2479 A node performing ETX link quality estimation uses a neighbour's 2480 Multicast Hello history to compute an estimate, written beta, of the 2481 probability that a Hello TLV is successfully received. Beta can be 2482 computed as the fraction of 1 bits within a small number (say, 6) of 2483 the most recent entries in the Multicast Hello history, or it can be 2484 an exponential average, or some combination of both approaches. 2486 Let alpha be MIN(1, 256/txcost), an estimate of the probability of 2487 successfully sending a Hello TLV. The cost is then computed by 2489 cost = 256/(alpha * beta) 2491 or, equivalently, 2493 cost = (MAX(txcost, 256) * rxcost) / 256. 2495 Since the IEEE 802.11 MAC performs ARQ on unicast frames, unicast 2496 frames do not provide a useful measure of link quality, and therefore 2497 ETX ignores the Unicast Hello history. Thus, a node performing ETX 2498 link-quality estimation will not route through neighbouring nodes 2499 unless they send periodic Multicast Hellos (possibly in addition to 2500 Unicast Hellos). 2502 A.3. Metric Computation 2504 As described in Section 3.5.2, the metric advertised by a neighbour 2505 is combined with the link cost to yield a metric. 2507 A.3.1. Additive Metrics 2509 The simplest approach for obtaining a monotonic, left-distributive 2510 metric is to define the metric of a route as the sum of the costs of 2511 the component links. More formally, if a neighbour advertises a 2512 route with metric m over a link with cost c, then the resulting route 2513 has metric M(c, m) = c + m. 2515 A multiplicative metric can be converted into an additive one by 2516 taking the logarithm (in some suitable base) of the link costs. 2518 A.3.2. External Sources of Willingness 2520 A node may want to vary its willingness to forward packets by taking 2521 into account information that is external to the Babel protocol, such 2522 as the monetary cost of a link, the node's battery status, CPU load, 2523 etc. This can be done by adding to every route's metric a value k 2524 that depends on the external data. For example, if a battery-powered 2525 node receives an update with metric m over a link with cost c, it 2526 might compute a metric M(c, m) = k + c + m, where k depends on the 2527 battery status. 2529 In order to preserve strict monotonicity (Section 3.5.2), the value k 2530 must be greater than -c. 2532 Appendix B. Constants 2534 The choice of time constants is a trade-off between fast detection of 2535 mobility events and protocol overhead. Two implementations of Babel 2536 with different time constants will interoperate, although the 2537 resulting convergence time will most likely be dictated by the slower 2538 of the two. 2540 Experience with the sample implementation of Babel indicates that the 2541 Hello interval is the most important time constant: a mobility event 2542 is detected within 1.5 to 3 Hello intervals. Due to Babel's reliance 2543 on triggered updates and explicit requests, the Update interval only 2544 has an effect on the time it takes for accurate metrics to be 2545 propagated after variations in link costs too small to trigger an 2546 unscheduled update or in the presence of packet loss. 2548 At the time of writing, the sample implementation of Babel uses the 2549 following default values: 2551 Multicast Hello Interval: 4 seconds. 2553 IHU Interval: the advertised IHU interval is always 3 times the 2554 Multicast Hello interval. IHUs are actually sent with each Hello 2555 on lossy links (as determined from the Hello history), but only 2556 with every third Multicast Hello on lossless links. 2558 Unicast Hello Interval: the sample implementation never sends 2559 scheduled Unicast Hellos; 2561 Update Interval: 4 times the Multicast Hello interval. 2563 IHU Hold Time: 3.5 times the advertised IHU interval. 2565 Route Expiry Time: 3.5 times the advertised update interval. 2567 Source GC time: 3 minutes. 2569 Request timeout: initially 2 seconds, doubled every time a request 2570 is resent, up to a maximum of three times. 2572 The amount of jitter applied to a packet depends on whether it 2573 contains any urgent TLVs or not (Section 3.1). Urgent triggered 2574 updates and urgent requests are delayed by no more than 200ms; 2575 acknowledgments, by no more than the associated deadline; and other 2576 TLVs by no more than one-half the Multicast Hello interval. 2578 Appendix C. Considerations for protocol extensions 2580 Babel is an extensible protocol, and this document defines a number 2581 of mechanisms that can be used to extend the protocol in a backwards 2582 compatible manner: 2584 o increasing the version number in the packet header; 2586 o defining new TLVs; 2588 o defining new sub-TLVs (with or without the mandatory bit set); 2590 o defining new AEs; 2592 o using the packet trailer. 2594 This appendix is intended to guide designers of protocol extensions 2595 in chosing a particular encoding. 2597 The version number in the Babel header should only be increased if 2598 the new version is not backwards compatible with the original 2599 protocol. 2601 In many cases, an extension could be implemented either by defining a 2602 new TLV, or by adding a new sub-TLV to an existing TLV. For example, 2603 an extension whose purpose is to attach additional data to route 2604 updates can be implemented either by creating a new "enriched" Update 2605 TLV, by adding a non-mandatory sub-TLV to the Update TLV, or by 2606 adding a mandatory sub-TLV. 2608 The various encodings are treated differently by implementations that 2609 do not understand the extension. In the case of a new TLV or of a 2610 sub-TLV with the mandatory bit set, the whole TLV is ignored by 2611 implementations that do not implement the extension, while in the 2612 case of a non-mandatory sub-TLV, the TLV is parsed and acted upon, 2613 and only the unknown sub-TLV is silently ignored. Therefore, a non- 2614 mandatory sub-TLV should be used by extensions that extend the Update 2615 in a compatible manner (the extension data may be silently ignored), 2616 while a mandatory sub-TLV or a new TLV must be used by extensions 2617 that make incompatible extensions to the meaning of the TLV (the 2618 whole TLV must be thrown away if the extension data is not 2619 understood). 2621 Experience shows that the need for additional data tends to crop up 2622 in the most unexpected places. Hence, it is recommended that 2623 extensions that define new TLVs should make them self-terminating, 2624 and allow attaching sub-TLVs to them. 2626 Adding a new AE is essentially equivalent to adding a new TLV: Update 2627 TLVs with an unknown AE are ignored, just like unknown TLVs. 2628 However, adding a new AE is more involved than adding a new TLV, 2629 since it creates a new set of compression state. Additionally, since 2630 the Next Hop TLV creates state specific to a given address family, as 2631 opposed to a given AE, a new AE for a previously defined address 2632 family must not be used in the Next Hop TLV if backwards 2633 compatibility is required. A similar issue arises with Update TLVs 2634 with unknown AEs establishing a new router-id (due to the Router-Id 2635 flag being set). Therefore, defining new AEs must be done with care 2636 if compatibility with unextended implementations is required. 2638 The packet trailer is intended to carry cryptographic signatures that 2639 only cover the packet body; storing the cryptographic signatures in 2640 the packet trailer avoids clearing the signature before computing a 2641 hash of the packet body, and makes it possible to check a 2642 cryptographic signature before running the full, stateful TLV parser. 2643 Hence, only TLVs that don't need to be protected by cryptographic 2644 security protocols should be allowed in the packet trailer. Any such 2645 TLVs should be easy to parse, and in particular should not require 2646 stateful parsing. 2648 Appendix D. Stub Implementations 2650 Babel is a fairly economic protocol. Updates take between 12 and 40 2651 octets per destination, depending on the address family and how 2652 successful compression is; in a double-stack flat network, an average 2653 of less than 24 octets per update is typical. The route table 2654 occupies about 35 octets per IPv6 entry. To put these values into 2655 perspective, a single full-size Ethernet frame can carry some 65 2656 route updates, and a megabyte of memory can contain a 20000-entry 2657 route table and the associated source table. 2659 Babel is also a reasonably simple protocol. The sample 2660 implementation consists of less than 12 000 lines of C code, and it 2661 compiles to less than 120 kB of text on a 32-bit CISC architecture; 2662 about half of this figure is due to protocol extensions and user- 2663 interface code. 2665 Nonetheless, in some very constrained environments, such as PDAs, 2666 microwave ovens, or abacuses, it may be desirable to have subset 2667 implementations of the protocol. 2669 There are many different definitions of a stub router, but for the 2670 needs of this section a stub implementation of Babel is one that 2671 announces one or more directly attached prefixes into a Babel network 2672 but doesn't reannounce any routes that it has learnt from its 2673 neighbours. It may either maintain a full routing table, or simply 2674 select a default gateway amongst any one of its neighbours that 2675 announces a default route. Since a stub implementation never 2676 forwards packets except from or to directly attached links, it cannot 2677 possibly participate in a routing loop, and hence it need not 2678 evaluate the feasibility condition or maintain a source table. 2680 No matter how primitive, a stub implementation MUST parse sub-TLVs 2681 attached to any TLVs that it understands and check the mandatory bit. 2682 It MUST answer acknowledgment requests and MUST participate in the 2683 Hello/IHU protocol. It MUST also be able to reply to seqno requests 2684 for routes that it announces and SHOULD be able to reply to route 2685 requests. 2687 Experience shows that an IPv6-only stub implementation of Babel can 2688 be written in less than 1000 lines of C code and compile to 13 kB of 2689 text on 32-bit CISC architecture. 2691 Appendix E. Software Availability 2693 The sample implementation of Babel is available from 2694 . 2696 Appendix F. Changes from previous versions 2698 F.1. Changes since RFC 6126 2700 o Changed UDP port number to 6696. 2702 o Consistently use router-id rather than id. 2704 o Clarified that the source garbage collection timer is reset after 2705 sending an update even if the entry was not modified. 2707 o In section "Seqno Requests", fixed an erroneous "route request". 2709 o In the description of the Seqno Request TLV, added the description 2710 of the Router-Id field. 2712 o Made router-ids all-0 and all-1 forbidden. 2714 F.2. Changes since draft-ietf-babel-rfc6126bis-00 2716 o Added security considerations. 2718 F.3. Changes since draft-ietf-babel-rfc6126bis-01 2720 o Integrated the format of sub-TLVs. 2722 o Mentioned for each TLV whether it supports sub-TLVs. 2724 o Added Appendix C. 2726 o Added a mandatory bit in sub-TLVs. 2728 o Changed compression state to be per-AF rather than per-AE. 2730 o Added implementation hint for the routing table. 2732 o Clarified how router-ids are computed when bit 0x40 is set in 2733 Updates. 2735 o Relaxed the conditions for sending requests, and tightened the 2736 conditions for forwarding requests. 2738 o Clarified that neighbours should be acquired at some point, but it 2739 doesn't matter when. 2741 F.4. Changes since draft-ietf-babel-rfc6126bis-02 2743 o Added Unicast Hellos. 2745 o Added unscheduled (interval-less) Hellos. 2747 o Changed Appendix A to consider Unicast and unscheduled Hellos. 2749 o Changed Appendix B to agree with the reference implementation. 2751 o Added optional algorithm to avoid the hold time. 2753 o Changed the table of pending seqno requests to be indexed by 2754 router-id in addition to prefixes. 2756 o Relaxed the route acquisition algorithm. 2758 o Replaced minimal implementations by stub implementations. 2760 o Added acknowledgments section. 2762 F.5. Changes since draft-ietf-babel-rfc6126bis-03 2764 o Clarified that all the data structures are conceptual. 2766 o Made sending and receiving Multicast Hellos a SHOULD, avoids 2767 expressing any opinion about Unicast Hellos. 2769 o Removed opinion about Multicast vs. Unicast Hellos (Appendix A.4). 2771 o Made hold-time into a SHOULD rather than MUST. 2773 o Clarified that Seqno Requests are for a finite-metric Update. 2775 o Clarified that sub-TLVs Pad1 and PadN are allowed within any TLV 2776 that allows sub-TLVs. 2778 o Updated IANA Considerations. 2780 o Updated Security Considerations. 2782 o Renamed routing table back to route table. 2784 o Made buffering outgoing updates a SHOULD. 2786 o Weakened advice to use modified EUI-64 in router-ids. 2788 o Added information about sending requests to Appendix B. 2790 o A number of minor wording changes and clarifications. 2792 F.6. Changes since draft-ietf-babel-rfc6126bis-03 2794 Minor editorial changes. 2796 F.7. Changes since draft-ietf-babel-rfc6126bis-04 2798 o Renamed isotonicity to left-distributivity. 2800 o Minor clarifications to unicast hellos. 2802 o Updated requirements boilerplate to RFC 8174. 2804 o Minor editorial changes. 2806 F.8. Changes since draft-ietf-babel-rfc6126bis-05 2808 o Added information about the packet trailer, now that it is used by 2809 draft-ietf-babel-hmac. 2811 F.9. Changes since draft-ietf-babel-rfc6126bis-06 2813 o Added references to security documents. 2815 Authors' Addresses 2817 Juliusz Chroboczek 2818 IRIF, University of Paris-Diderot 2819 Case 7014 2820 75205 Paris Cedex 13 2821 France 2823 Email: jch@irif.fr 2825 David Schinazi 2826 Google LLC 2827 1600 Amphitheatre Parkway 2828 Mountain View, California 94043 2829 USA 2831 Email: dschinazi.ietf@gmail.com