idnits 2.17.1 draft-ietf-babel-rfc6126bis-05.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 (May 29, 2018) is 2159 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) ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126) Summary: 1 error (**), 0 flaws (~~), 2 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 Apple Inc. 6 Expires: November 30, 2018 May 29, 2018 8 The Babel Routing Protocol 9 draft-ietf-babel-rfc6126bis-05 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 November 30, 2018. 35 Copyright Notice 37 Copyright (c) 2018 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 . . . . . . . . . . . . . . . . . . . . . . . 33 78 4.4. Sub-TLV Format . . . . . . . . . . . . . . . . . . . . . 34 79 4.5. Parser state . . . . . . . . . . . . . . . . . . . . . . 35 80 4.6. Details of Specific TLVs . . . . . . . . . . . . . . . . 35 81 4.7. Details of specific sub-TLVs . . . . . . . . . . . . . . 46 82 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 47 83 6. Security Considerations . . . . . . . . . . . . . . . . . . . 48 84 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 48 85 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 49 86 8.1. Normative References . . . . . . . . . . . . . . . . . . 49 87 8.2. Informative References . . . . . . . . . . . . . . . . . 49 88 Appendix A. Cost and Metric Computation . . . . . . . . . . . . 50 89 A.1. Maintaining Hello History . . . . . . . . . . . . . . . . 50 90 A.2. Cost Computation . . . . . . . . . . . . . . . . . . . . 51 91 A.3. Metric Computation . . . . . . . . . . . . . . . . . . . 53 92 Appendix B. Constants . . . . . . . . . . . . . . . . . . . . . 53 93 Appendix C. Considerations for protocol extensions . . . . . . . 54 94 Appendix D. Stub Implementations . . . . . . . . . . . . . . . . 56 95 Appendix E. Software Availability . . . . . . . . . . . . . . . 56 96 Appendix F. Changes from previous versions . . . . . . . . . . . 57 97 F.1. Changes since RFC 6126 . . . . . . . . . . . . . . . . . 57 98 F.2. Changes since draft-ietf-babel-rfc6126bis-00 . . . . . . 57 99 F.3. Changes since draft-ietf-babel-rfc6126bis-01 . . . . . . 57 100 F.4. Changes since draft-ietf-babel-rfc6126bis-02 . . . . . . 58 101 F.5. Changes since draft-ietf-babel-rfc6126bis-03 . . . . . . 58 102 F.6. Changes since draft-ietf-babel-rfc6126bis-03 . . . . . . 59 103 F.7. Changes since draft-ietf-babel-rfc6126bis-04 . . . . . . 59 104 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 59 106 1. Introduction 108 Babel is a loop-avoiding distance-vector routing protocol that is 109 designed to be robust and efficient both in networks using prefix- 110 based routing and in networks using flat routing ("mesh networks"), 111 and both in relatively stable wired networks and in highly dynamic 112 wireless networks. 114 1.1. Features 116 The main property that makes Babel suitable for unstable networks is 117 that, unlike naive distance-vector routing protocols [RIP], it 118 strongly limits the frequency and duration of routing pathologies 119 such as routing loops and black-holes during reconvergence. Even 120 after a mobility event is detected, a Babel network usually remains 121 loop-free. Babel then quickly reconverges to a configuration that 122 preserves the loop-freedom and connectedness of the network, but is 123 not necessarily optimal; in many cases, this operation requires no 124 packet exchanges at all. Babel then slowly converges, in a time on 125 the scale of minutes, to an optimal configuration. This is achieved 126 by using sequenced routes, a technique pioneered by Destination- 127 Sequenced Distance-Vector routing [DSDV]. 129 More precisely, Babel has the following properties: 131 o when every prefix is originated by at most one router, Babel never 132 suffers from routing loops; 134 o when a single prefix is originated by multiple routers, Babel may 135 occasionally create a transient routing loop for this particular 136 prefix; this loop disappears in a time proportional to its 137 diameter, and never again (up to an arbitrary garbage-collection 138 (GC) time) will the routers involved participate in a routing loop 139 for the same prefix; 141 o assuming bounded packet loss rates, any routing black-holes that 142 may appear after a mobility event are corrected in a time at most 143 proportional to the network's diameter. 145 Babel has provisions for link quality estimation and for fairly 146 arbitrary metrics. When configured suitably, Babel can implement 147 shortest-path routing, or it may use a metric based, for example, on 148 measured packet loss. 150 Babel nodes will successfully establish an association even when they 151 are configured with different parameters. For example, a mobile node 152 that is low on battery may choose to use larger time constants (hello 153 and update intervals, etc.) than a node that has access to wall 154 power. Conversely, a node that detects high levels of mobility may 155 choose to use smaller time constants. The ability to build such 156 heterogeneous networks makes Babel particularly adapted to the 157 unmanaged and wireless environment. 159 Finally, Babel is a hybrid routing protocol, in the sense that it can 160 carry routes for multiple network-layer protocols (IPv4 and IPv6), 161 whichever protocol the Babel packets are themselves being carried 162 over. 164 1.2. Limitations 166 Babel has two limitations that make it unsuitable for use in some 167 environments. First, Babel relies on periodic routing table updates 168 rather than using a reliable transport; hence, in large, stable 169 networks it generates more traffic than protocols that only send 170 updates when the network topology changes. In such networks, 171 protocols such as OSPF [OSPF], IS-IS [IS-IS], or the Enhanced 172 Interior Gateway Routing Protocol (EIGRP) [EIGRP] might be more 173 suitable. 175 Second, unless the optional algorithm described in Section 3.5.5 is 176 implemented, Babel does impose a hold time when a prefix is 177 retracted. While this hold time does not apply to the exact prefix 178 being retracted, and hence does not prevent fast reconvergence should 179 it become available again, it does apply to any shorter prefix that 180 covers it. This may make those implementations of Babel that do not 181 implement the optional algorithm described in Section 3.5.5 182 unsuitable for use in networks that implement automatic prefix 183 aggregation. 185 1.3. Specification of Requirements 187 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 188 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 189 "OPTIONAL" in this document are to be interpreted as described in BCP 190 14 [RFC2119] [RFC8174] when, and only when, they appear in all 191 capitals, as shown here. 193 2. Conceptual Description of the Protocol 195 Babel is a loop-avoiding distance vector protocol: it is based on the 196 Bellman-Ford protocol, just like the venerable RIP [RIP], but 197 includes a number of refinements that either prevent loop formation 198 altogether, or ensure that a loop disappears in a timely manner and 199 doesn't form again. 201 Conceptually, Bellman-Ford is executed in parallel for every source 202 of routing information (destination of data traffic). In the 203 following discussion, we fix a source S; the reader will recall that 204 the same algorithm is executed for all sources. 206 2.1. Costs, Metrics and Neighbourship 208 For every pair of neighbouring nodes A and B, Babel computes an 209 abstract value known as the cost of the link from A to B., written 210 C(A, B). Given a route between any two (not necessarily 211 neighbouring) nodes, the metric of the route is the sum of the costs 212 of all the edges along the route. The goal of the routing algorithm 213 is to compute, for every source S, the tree of routes of lowest 214 metric to S. 216 Costs and metrics need not be integers. In general, they can be 217 values in any algebra that satisfies two fairly general conditions 218 (Section 3.5.2). 220 A Babel node periodically sends Hello messages to all of its 221 neighbours; it also periodically sends an IHU ("I Heard You") message 222 to every neighbour from which it has recently heard a Hello. From 223 the information derived from Hello and IHU messages received from its 224 neighbour B, a node A computes the cost C(A, B) of the link from A to 225 B. 227 2.2. The Bellman-Ford Algorithm 229 Every node A maintains two pieces of data: its estimated distance to 230 S, written D(A), and its next-hop router to S, written NH(A). 231 Initially, D(S) = 0, D(A) is infinite, and NH(A) is undefined. 233 Periodically, every node B sends to all of its neighbours a route 234 update, a message containing D(B). When a neighbour A of B receives 235 the route update, it checks whether B is its selected next hop; if 236 that is the case, then NH(A) is set to B, and D(A) is set to C(A, B) 237 + D(B). If that is not the case, then A compares C(A, B) + D(B) to 238 its current value of D(A). If that value is smaller, meaning that 239 the received update advertises a route that is better than the 240 currently selected route, then NH(A) is set to B, and D(A) is set to 241 C(A, B) + D(B). 243 A number of refinements to this algorithm are possible, and are used 244 by Babel. In particular, convergence speed may be increased by 245 sending unscheduled "triggered updates" whenever a major change in 246 the topology is detected, in addition to the regular, scheduled 247 updates. Additionally, a node may maintain a number of alternate 248 routes, which are being advertised by neighbours other than its 249 selected neighbour, and which can be used immediately if the selected 250 route were to fail. 252 2.3. Transient Loops in Bellman-Ford 254 It is well known that a naive application of Bellman-Ford to 255 distributed routing can cause transient loops after a topology 256 change. Consider for example the following topology: 258 B 259 1 /| 260 1 / | 261 S --- A |1 262 \ | 263 1 \| 264 C 266 After convergence, D(B) = D(C) = 2, with NH(B) = NH(C) = A. 268 Suppose now that the link between S and A fails: 270 B 271 1 /| 272 / | 273 S A |1 274 \ | 275 1 \| 276 C 278 When it detects the failure of the link, A switches its next hop to B 279 (which is still advertising a route to S with metric 2), and 280 advertises a metric equal to 3, and then advertises a new route with 281 metric 3. This process of nodes changing selected neighbours and 282 increasing their metric continues until the advertised metric reaches 283 "infinity", a value larger than all the metrics that the routing 284 protocol is able to carry. 286 2.4. Feasibility Conditions 288 Bellman-Ford is a very robust algorithm: its convergence properties 289 are preserved when routers delay route acquisition or when they 290 discard some updates. Babel routers discard received route 291 announcements unless they can prove that accepting them cannot 292 possibly cause a routing loop. 294 More formally, we define a condition over route announcements, known 295 as the "feasibility condition", that guarantees the absence of 296 routing loops whenever all routers ignore route updates that do not 297 satisfy the feasibility condition. In effect, this makes Bellman- 298 Ford into a family of routing algorithms, parameterised by the 299 feasibility condition. 301 Many different feasibility conditions are possible. For example, BGP 302 can be modelled as being a distance-vector protocol with a (rather 303 drastic) feasibility condition: a routing update is only accepted 304 when the receiving node's AS number is not included in the update's 305 AS-Path attribute (note that BGP's feasibility condition does not 306 ensure the absence of transient "micro-loops" during reconvergence). 308 Another simple feasibility condition, used in the Destination- 309 Sequenced Distance-Vector (DSDV) routing protocol [DSDV] and in the 310 Ad hoc On-Demand Distance Vector (AODV) protocol, stems from the 311 following observation: a routing loop can only arise after a router 312 has switched to a route with a larger metric than the route that it 313 had previously selected. Hence, one could decide that a route is 314 feasible only when its metric at the local node would be no larger 315 than the metric of the currently selected route, i.e., an 316 announcement carrying a metric D(B) is accepted by A when C(A, B) + 317 D(B) <= D(A). If all routers obey this constraint, then the metric 318 at every router is nonincreasing, and the following invariant is 319 always preserved: if A has selected B as its successor, then D(B) < 320 D(A), which implies that the forwarding graph is loop-free. 322 Babel uses a slightly more refined feasibility condition, derived 323 from EIGRP [DUAL]. Given a router A, define the feasibility distance 324 of A, written FD(A), as the smallest metric that A has ever 325 advertised for S to any of its neighbours. An update sent by a 326 neighbour B of A is feasible when the metric D(B) advertised by B is 327 strictly smaller than A's feasibility distance, i.e., when D(B) < 328 FD(A). 330 It is easy to see that this latter condition is no more restrictive 331 than DSDV-feasibility. Suppose that node A obeys DSDV-feasibility; 332 then D(A) is nonincreasing, hence at all times D(A) <= FD(A). 333 Suppose now that A receives a DSDV-feasible update that advertises a 334 metric D(B). Since the update is DSDV-feasible, C(A, B) + D(B) <= 335 D(A), hence D(B) < D(A), and since D(A) <= FD(A), D(B) < FD(A). 337 To see that it is strictly less restrictive, consider the following 338 diagram, where A has selected the route through B, and D(A) = FD(A) = 339 2. Since D(C) = 1 < FD(A), the alternate route through C is feasible 340 for A, although its metric C(A, C) + D(C) = 5 is larger than that of 341 the currently selected route: 343 B 344 1 / \ 1 345 / \ 346 S A 347 \ / 348 1 \ / 4 349 C 351 To show that this feasibility condition still guarantees loop- 352 freedom, recall that at the time when A accepts an update from B, the 353 metric D(B) announced by B is no smaller than FD(B); since it is 354 smaller than FD(A), at that point in time FD(B) < FD(A). Since this 355 property is preserved when A sends updates, it remains true at all 356 times, which ensures that the forwarding graph has no loops. 358 2.5. Solving Starvation: Sequencing Routes 360 Obviously, the feasibility conditions defined above cause starvation 361 when a router runs out of feasible routes. Consider the following 362 diagram, where both A and B have selected the direct route to S: 364 A 365 1 /| D(A) = 1 366 / | FD(A) = 1 367 S |1 368 \ | D(B) = 2 369 2 \| FD(B) = 2 370 B 372 Suppose now that the link between A and S breaks: 374 A 375 | 376 | FD(A) = 1 377 S |1 378 \ | D(B) = 2 379 2 \| FD(B) = 2 380 B 382 The only route available from A to S, the one that goes through B, is 383 not feasible: A suffers from spurious starvation. At that point, the 384 whole subtree suffering from starvation must be reset, which is 385 essentially what EIGRP does when it performs a global synchronisation 386 of all the routers in the sarving subtree (the "active" phase of 387 EIGRP). 389 Babel reacts to starvation in a less drastic manner, by using 390 sequenced routes, a technique introduced by DSDV and adopted by AODV. 391 In addition to a metric, every route carries a sequence number, a 392 nondecreasing integer that is propagated unchanged through the 393 network and is only ever incremented by the source; a pair (s, m), 394 where s is a sequence number and m a metric, is called a distance. 396 A received update is feasible when either it is more recent than the 397 feasibility distance maintained by the receiving node, or it is 398 equally recent and the metric is strictly smaller. More formally, if 399 FD(A) = (s, m), then an update carrying the distance (s', m') is 400 feasible when either s' > s, or s = s' and m' < m. 402 Assuming the sequence number of S is 137, the diagram above becomes: 404 A 405 | 406 | FD(A) = (137, 1) 407 S |1 408 \ | D(B) = (137, 2) 409 2 \| FD(B) = (137, 2) 410 B 412 After S increases its sequence number, and the new sequence number is 413 propagated to B, we have: 415 A 416 | 417 | FD(A) = (137, 1) 418 S |1 419 \ | D(B) = (138, 2) 420 2 \| FD(B) = (138, 2) 421 B 423 at which point the route through B becomes feasible again. 425 Note that while sequence numbers are used for determining 426 feasibility, they are not necessarily used in route selection: a node 427 will normally ignore the sequence number when selecting the best 428 route to a given destination (Section 3.6). 430 2.6. Requests 432 In DSDV, the sequence number of a source is increased periodically. 433 A route becomes feasible again after the source increases its 434 sequence number, and the new sequence number is propagated through 435 the network, which may, in general, require a significant amount of 436 time. 438 Babel takes a different approach. When a node detects that it is 439 suffering from a potentially spurious starvation, it sends an 440 explicit request to the source for a new sequence number. This 441 request is forwarded hop by hop to the source, with no regard to the 442 feasibility condition. Upon receiving the request, the source 443 increases its sequence number and broadcasts an update, which is 444 forwarded to the requesting node. 446 Note that after a change in network topology not all such requests 447 will, in general, reach the source, as some will be sent over links 448 that are now broken. However, if the network is still connected, 449 then at least one among the nodes suffering from spurious starvation 450 has an (unfeasible) route to the source; hence, in the absence of 451 packet loss, at least one such request will reach the source. 452 (Resending requests a small number of times compensates for packet 453 loss.) 455 Since requests are forwarded with no regard to the feasibility 456 condition, they may, in general, be caught in a forwarding loop; this 457 is avoided by having nodes perform duplicate detection for the 458 requests that they forward. 460 2.7. Multiple Routers 462 The above discussion assumes that every prefix is originated by a 463 single router. In real networks, however, it is often necessary to 464 have a single prefix originated by multiple routers: for example, the 465 default route will be originated by all of the edge routers of a 466 routing domain. 468 Since synchronising sequence numbers between distinct routers is 469 problematic, Babel treats routes for the same prefix as distinct 470 entities when they are originated by different routers: every route 471 announcement carries the router-id of its originating router, and 472 feasibility distances are not maintained per prefix, but per source, 473 where a source is a pair of a router-id and a prefix. In effect, 474 Babel guarantees loop-freedom for the forwarding graph to every 475 source; since the union of multiple acyclic graphs is not in general 476 acyclic, Babel does not in general guarantee loop-freedom when a 477 prefix is originated by multiple routers, but any loops will be 478 broken in a time at most proportional to the diameter of the loop -- 479 as soon as an update has "gone around" the routing loop. 481 Consider for example the following topology, where A has selected the 482 default route through S, and B has selected the one through S': 484 1 1 1 485 ::/0 -- S --- A --- B --- S' -- ::/0 487 Suppose that both default routes fail at the same time; then nothing 488 prevents A from switching to B, and B simultaneously switching to A. 489 However, as soon as A has successfully advertised the new route to B, 490 the route through A will become unfeasible for B. Conversely, as 491 soon as B will have advertised the route through A, the route through 492 B will become unfeasible for A. 494 In effect, the routing loop disappears at the latest when routing 495 information has gone around the loop. Since this process can be 496 delayed by lost packets, Babel makes certain efforts to ensure that 497 updates are sent reliably after a router-id change (Section 3.7.2). 499 Additionally, after the routers have advertised the two routes, both 500 sources will be in their source tables, which will prevent them from 501 ever again participating in a routing loop involving routes from S 502 and S' (up to the source GC time, which, available memory permitting, 503 can be set to arbitrarily large values). 505 2.8. Overlapping Prefixes 507 In the above discussion, we have assumed that all prefixes are 508 disjoint, as is the case in flat ("mesh") routing. In practice, 509 however, prefixes may overlap: for example, the default route 510 overlaps with all of the routes present in the network. 512 After a route fails, it is not correct in general to switch to a 513 route that subsumes the failed route. Consider for example the 514 following configuration: 516 1 1 517 ::/0 -- A --- B --- C 519 Suppose that node C fails. If B forwards packets destined to C by 520 following the default route, a routing loop will form, and persist 521 until A learns of B's retraction of the direct route to C. B avoids 522 this pitfall by installing an "unreachable" route after a route is 523 retracted; this route is maintained until it can be guaranteed that 524 the former route has been retracted by all of B's neighbours 525 (Section 3.5.5). 527 3. Protocol Operation 529 Every Babel speaker is assigned a router-id, which is an arbitrary 530 string of 8 octets that is assumed unique across the routing domain. 531 For example, routers-ids could be assigned randomly, or they could 532 derived from a link-layer address. (The protocol encoding is 533 slightly more compact when router-ids are assigned in the same manner 534 as the IPv6 layer assigns host IDs.) 536 3.1. Message Transmission and Reception 538 Babel protocol packets are sent in the body of a UDP datagram (as 539 described in Section 4 below). Each Babel packet consists of zero or 540 more TLVs. Most TLVs may contain sub-TLVs. 542 The source address of a Babel packet is always a unicast address, 543 link-local in the case of IPv6. Babel packets may be sent to a well- 544 known (link-local) multicast address or to a (link-local) unicast 545 address. In normal operation, a Babel speaker sends both multicast 546 and unicast packets to its neighbours. 548 With the exception of Hello TLVs and acknowledgments, all Babel TLVs 549 can be sent to either unicast or multicast addresses, and their 550 semantics does not depend on whether the destination is a unicast or 551 a multicast address. Hence, a Babel speaker does not need to 552 determine the destination address of a packet that it receives in 553 order to interpret it. 555 A moderate amount of jitter may be applied to packets sent by a Babel 556 speaker: outgoing TLVs are buffered and SHOULD be sent with a small 557 random delay. This is done for two purposes: it avoids 558 synchronisation of multiple Babel speakers across a network [JITTER], 559 and it allows for the aggregation of multiple TLVs into a single 560 packet. 562 The exact delay and amount of jitter applied to a packet depends on 563 whether it contains any urgent TLVs. Acknowledgment TLVs MUST be 564 sent before the deadline specified in the corresponding request. The 565 particular class of updates specified in Section 3.7.2 MUST be sent 566 in a timely manner. The particular class of request and update TLVs 567 specified in Section 3.8.2 SHOULD be sent in a timely manner. 569 3.2. Data Structures 571 In this section, we give a description of the data structures that 572 every Babel speaker maintains. This description is conceptual: a 573 Babel speaker may use different data structures as long as the 574 resulting protocol is the same as the one described in this document. 576 For example, rather than maintaining a single table containing both 577 selected and unselected (fallback) routes, as described in 578 Section 3.2.6 belong, an actual implementation would probably use two 579 tables, one with selected routes and one with fallback routes. 581 3.2.1. Sequence number arithmetic 583 Sequence numbers (seqnos) appear in a number of Babel data 584 structures, and they are interpreted as integers modulo 2^16. For 585 the purposes of this document, arithmetic on sequence numbers is 586 defined as follows. 588 Given a seqno s and an integer n, the sum of s and n is defined by 590 s + n (modulo 2^16) = (s + n) MOD 2^16 592 or, equivalently, 594 s + n (modulo 2^16) = (s + n) AND 65535 596 where MOD is the modulo operation yielding a non-negative integer and 597 AND is the bitwise conjunction operation. 599 Given two sequence numbers s and s', the relation s is less than s' 600 (s < s') is defined by 602 s < s' (modulo 2^16) when 0 < ((s' - s) MOD 2^16) < 32768 604 or equivalently 606 s < s' (modulo 2^16) when s /= s' and ((s' - s) AND 32768) = 0. 608 3.2.2. Node Sequence Number 610 A node's sequence number is a 16-bit integer that is included in 611 route updates sent for routes originated by this node. 613 A node increments its sequence number (modulo 2^16) whenever it 614 receives a request for a new sequence number (Section 3.8.1.2). A 615 node SHOULD NOT increment its sequence number (seqno) spontaneously, 616 since increasing seqnos makes it less likely that other nodes will 617 have feasible alternate routes when their selected routes fail. 619 3.2.3. The Interface Table 621 The interface table contains the list of interfaces on which the node 622 speaks the Babel protocol. Every interface table entry contains the 623 interface's outgoing Multicast Hello seqno, a 16-bit integer that is 624 sent with each Multicast Hello TLV on this interface and is 625 incremented (modulo 2^16) whenever a Multicast Hello is sent. (Note 626 that an interface's Multicast Hello seqno is unrelated to the node's 627 seqno.) 629 There are two timers associated with each interface table entry -- 630 the multicast hello timer, which governs the sending of scheduled 631 Multicast Hello and IHU packets, and the update timer, which governs 632 the sending of periodic route updates. 634 3.2.4. The Neighbour Table 636 The neighbour table contains the list of all neighbouring interfaces 637 from which a Babel packet has been recently received. The neighbour 638 table is indexed by pairs of the form (interface, address), and every 639 neighbour table entry contains the following data: 641 o the local node's interface over which this neighbour is reachable; 643 o the address of the neighbouring interface; 645 o a history of recently received Multicast Hello packets from this 646 neighbour; this can, for example, be a sequence of n bits, for 647 some small value n, indicating which of the n hellos most recently 648 sent by this neighbour have been received by the local node; 650 o a history of recently received Unicast Hello packets from this 651 neighbour; 653 o the "transmission cost" value from the last IHU packet received 654 from this neighbour, or FFFF hexadecimal (infinity) if the IHU 655 hold timer for this neighbour has expired; 657 o the neighbour's expected incoming Multicast Hello sequence number, 658 an integer modulo 2^16. 660 o the neighbour's expected incoming Unicast Hello sequence number, 661 an integer modulo 2^16. 663 o the neighbour's outgoing Unicast Hello sequence number, an integer 664 modulo 2^16 that is sent with each Unicast Hello TLV to this 665 neighbour and is incremented (modulo 2^16) whenever a Unicast 666 Hello is sent. (Note that a neighbour's outgoing Unicast Hello 667 seqno is distinct from the interface's outgoing Multicast Hello 668 seqno.) 670 There are three timers associated with each neighbour entry -- the 671 multicast hello timer, which is initialised from the interval value 672 carried by scheduled Multicast Hello TLVs, the unicast hello timer, 673 which is initialised from the interval value carried by scheduled 674 Unicast Hello TLVs, and the IHU timer, which is initialised to a 675 small multiple of the interval carried in IHU TLVs. 677 Note that the neighbour table is indexed by IP addresses, not by 678 router-ids: neighbourship is a relationship between interfaces, not 679 between nodes. Therefore, two nodes with multiple interfaces can 680 participate in multiple neighbourship relationships, a situation that 681 can notably arise when wireless nodes with multiple radios are 682 involved. 684 3.2.5. The Source Table 686 The source table is used to record feasibility distances. It is 687 indexed by triples of the form (prefix, plen, router-id), and every 688 source table entry contains the following data: 690 o the prefix (prefix, plen), where plen is the prefix length, that 691 this entry applies to; 693 o the router-id of a router originating this prefix; 695 o a pair (seqno, metric), this source's feasibility distance. 697 There is one timer associated with each entry in the source table -- 698 the source garbage-collection timer. It is initialised to a time on 699 the order of minutes and reset as specified in Section 3.7.3. 701 3.2.6. The Route Table 703 The route table contains the routes known to this node. It is 704 indexed by triples of the form (prefix, plen, neighbour), and every 705 route table entry contains the following data: 707 o the source (prefix, plen, router-id) for which this route is 708 advertised; 710 o the neighbour that advertised this route; 712 o the metric with which this route was advertised by the neighbour, 713 or FFFF hexadecimal (infinity) for a recently retracted route; 715 o the sequence number with which this route was advertised; 717 o the next-hop address of this route; 718 o a boolean flag indicating whether this route is selected, i.e., 719 whether it is currently being used for forwarding and is being 720 advertised. 722 There is one timer associated with each route table entry -- the 723 route expiry timer. It is initialised and reset as specified in 724 Section 3.5.4. 726 Note that there are two distinct (seqno, metric) pairs associated to 727 each route: the route's distance, which is stored in the route table, 728 and the feasibility distance, stored in the source table and shared 729 between all routes with the same source. 731 3.2.7. The Table of Pending Seqno Requests 733 The table of pending seqno requests contains a list of seqno requests 734 that the local node has sent (either because they have been 735 originated locally, or because they were forwarded) and to which no 736 reply has been received yet. This table is indexed by triples of the 737 form (prefix, plen, router-id), and every entry in this table 738 contains the following data: 740 o the prefix, router-id, and seqno being requested; 742 o the neighbour, if any, on behalf of which we are forwarding this 743 request; 745 o a small integer indicating the number of times that this request 746 will be resent if it remains unsatisfied. 748 There is one timer associated with each pending seqno request; it 749 governs both the resending of requests and their expiry. 751 3.3. Acknowledgments and acknowledgment requests 753 A Babel speaker may request that a neighbour receiving a given packet 754 reply with an explicit acknowledgment within a given time. While the 755 use of acknowledgment requests is optional, every Babel speaker MUST 756 be able to reply to such a request. 758 An acknowledgment MUST be sent to a unicast destination. On the 759 other hand, acknowledgment requests may be sent to either unicast or 760 multicast destinations, in which case they request an acknowledgment 761 from all of the receiving nodes. 763 When to request acknowledgments is a matter of local policy; the 764 simplest strategy is to never request acknowledgments and to rely on 765 periodic updates to ensure that any reachable routes are eventually 766 propagated throughout the routing domain. In order to improve 767 convergence speed and reduce the amount of control traffic, 768 acknowledgment requests MAY be used in order to reliably send urgent 769 updates (Section 3.7.2) and retractions (Section 3.5.5), especially 770 when the number of neighbours on a given interface is small. Since 771 Babel is designed to deal gracefully with packet loss on unreliable 772 media, sending all packets with acknowledgment requests is not 773 necessary, and NOT RECOMMENDED, as the acknowledgments cause 774 additional traffic and may force additional Address Resolution 775 Protocol (ARP) or Neighbour Discovery (ND) exchanges. 777 3.4. Neighbour Acquisition 779 Neighbour acquisition is the process by which a Babel node discovers 780 the set of neighbours heard over each of its interfaces and 781 ascertains bidirectional reachability. On unreliable media, 782 neighbour acquisition additionally provides some statistics that may 783 be useful for link quality computation. 785 Before it can exchange routing information with a neighbour, a Babel 786 node MUST create an entry for that neighbour in the neighbour table. 787 When to do that is implementation-specific; suitable strategies 788 include creating an entry when any Babel packet is received, or 789 creating an entry when a Hello TLV is parsed. Similarly, in order to 790 conserve system resources, an implementation SHOULD discard an entry 791 when it has been unused for long enough; suitable strategies include 792 dropping the neighbour after a timeout, and dropping a neighbour when 793 the associated Hello histories become empty (see Appendix A.2). 795 3.4.1. Reverse Reachability Detection 797 Every Babel node sends Hello TLVs to its neighbours to indicate that 798 it is alive, at regular or irregular intervals. Each Hello TLV 799 carries an increasing (modulo 2^16) sequence number and an upper 800 bound on the time interval until the next Hello of the same type (see 801 below). If the time interval is set to 0, then the Hello TLV does 802 not establish a new promise: the deadline carried by the previous 803 Hello of the same type still applies to the next Hello (if the most 804 recent scheduled Hello of the right kind was received at time t0 and 805 carried interval i, then the previous promise of sending another 806 Hello before time t0 + i still holds). We say that a Hello is 807 "scheduled" if it carries a non-zero interval, and "unscheduled" 808 otherwise. 810 There are two kinds of Hellos: Multicast Hellos, which use a per- 811 interface Hello counter (the Multicast Hello seqno), and Unicast 812 Hellos, which use a per-neighbour counter (the Multicast Hello 813 seqno). A Multicast Hello with a given seqno MUST be sent to all 814 neighbours on a given interface, either by sending it to a multicast 815 address or by sending it to one unicast address per neighbour (hence, 816 the term "Multicast Hello" is a slight misnomer). A Unicast Hello 817 carrying a given seqno should normally be sent to just one neighbour 818 (over unicast), since the sequence numbers of different neighbours 819 are not in general synchronised. 821 Multicast Hellos sent over multicast can be used for neighbour 822 discovery; hence, a node SHOULD send periodic (scheduled) Multicast 823 Hellos unless neighbour discovery is performed by means outside of 824 the Babel protocol. A node MAY send Unicast Hellos or unscheduled 825 Hellos of either kind for any reason, such as reducing the amount of 826 multicast traffic or improving reliability on link technologies with 827 poor support for link-layer multicast. 829 A node MAY send a scheduled Hello ahead of time. A node MAY change 830 its scheduled Hello interval. The Hello interval MAY be decreased at 831 any time; it MAY be increased immediately before sending a Hello TLV, 832 but SHOULD NOT be increased at other times. (Equivalently, a node 833 SHOULD send a scheduled Hello immediately after increasing its Hello 834 interval.) 836 How to deal with received Hello TLVs and what statistics to maintain 837 are considered local implementation matters; typically, a node will 838 maintain some sort of history of recently received Hellos. An 839 example of a suitable algorithm is described in Appendix A.1. 841 After receiving a Hello, or determining that it has missed one, the 842 node recomputes the association's cost (Section 3.4.3) and runs the 843 route selection procedure (Section 3.6). 845 3.4.2. Bidirectional Reachability Detection 847 In order to establish bidirectional reachability, every node sends 848 periodic IHU ("I Heard You") TLVs to each of its neighbours. Since 849 IHUs carry an explicit interval value, they MAY be sent less often 850 than Hellos in order to reduce the amount of routing traffic in dense 851 networks; in particular, they SHOULD be sent less often than Hellos 852 over links with little packet loss. While IHUs are conceptually 853 unicast, they MAY be sent to a multicast address in order to avoid an 854 ARP or Neighbour Discovery exchange and to aggregate multiple IHUs 855 into a single packet. 857 In addition to the periodic IHUs, a node MAY, at any time, send an 858 unscheduled IHU packet. It MAY also, at any time, decrease its IHU 859 interval, and it MAY increase its IHU interval immediately before 860 sending an IHU, but SHOULD NOT increase it at any other time. 862 (Equivalently, a node SHOULD send an extra IHU immediately after 863 increasing its Hello interval.) 865 Every IHU TLV contains two pieces of data: the link's rxcost 866 (reception cost) from the sender's perspective, used by the neighbour 867 for computing link costs (Section 3.4.3), and the interval between 868 periodic IHU packets. A node receiving an IHU sets the value of the 869 txcost (transmission cost) maintained in the neighbour table to the 870 value contained in the IHU, and resets the IHU timer associated to 871 this neighbour to a small multiple of the interval value received in 872 the IHU. When a neighbour's IHU timer expires, the neighbour's 873 txcost is set to infinity. 875 After updating a neighbour's txcost, the receiving node recomputes 876 the neighbour's cost (Section 3.4.3) and runs the route selection 877 procedure (Section 3.6). 879 3.4.3. Cost Computation 881 A neighbourship association's link cost is computed from the values 882 maintained in the neighbour table: the statistics kept in the 883 neighbour table about the reception of Hellos, and the txcost 884 computed from received IHU packets. 886 For every neighbour, a Babel node computes a value known as this 887 neighbour's rxcost. This value is usually derived from the Hello 888 history, which may be combined with other data, such as statistics 889 maintained by the link layer. The rxcost is sent to a neighbour in 890 each IHU. 892 Since nodes do not necessarily send periodic Unicast Hellos but do 893 usually send periodic Multicast Hellos (Section 3.4.1), a node SHOULD 894 use an algorithm that yields a finite rxcost when only Multicast 895 Hellos are received, unless interoperability with nodes that only 896 send Multicast Hellos is not required. 898 How the txcost and rxcost are combined in order to compute a link's 899 cost is a matter of local policy; as far as Babel's correctness is 900 concerned, only the following conditions MUST be satisfied: 902 o the cost is strictly positive; 904 o if no Hello TLVs of either kind were received recently, then the 905 cost is infinite; 907 o if the txcost is infinite, then the cost is infinite. 909 Note that while this document does not constrain cost computation any 910 further, not all cost computation strategies will give good results. 911 See Appendix A.2 for examples of strategies for computing a link's 912 cost that are known to work well in practice. 914 3.5. Routing Table Maintenance 916 Conceptually, a Babel update is a quintuple (prefix, plen, router-id, 917 seqno, metric), where (prefix, plen) is the prefix for which a route 918 is being advertised, router-id is the router-id of the router 919 originating this update, seqno is a nondecreasing (modulo 2^16) 920 integer that carries the originating router seqno, and metric is the 921 announced metric. 923 Before being accepted, an update is checked against the feasibility 924 condition (Section 3.5.1), which ensures that the route does not 925 create a routing loop. If the feasibility condition is not 926 satisfied, the update is either ignored or prevents the route from 927 being selected, as described in Section 3.5.4. If the feasibility 928 condition is satisfied, then the update cannot possibly cause a 929 routing loop. 931 3.5.1. The Feasibility Condition 933 The feasibility condition is applied to all received updates. The 934 feasibility condition compares the metric in the received update with 935 the metrics of the updates previously sent by the receiving node; 936 updates that fail the feasibility condition, and therefore have 937 metrics large enough to cause a routing loop, are either ignored or 938 prevent the resulting route from being selected. 940 A feasibility distance is a pair (seqno, metric), where seqno is an 941 integer modulo 2^16 and metric is a positive integer. Feasibility 942 distances are compared lexicographically, with the first component 943 inverted: we say that a distance (seqno, metric) is strictly better 944 than a distance (seqno', metric'), written 946 (seqno, metric) < (seqno', metric') 948 when 950 seqno > seqno' or (seqno = seqno' and metric < metric') 952 where sequence numbers are compared modulo 2^16. 954 Given a source (prefix, plen, router-id), a node's feasibility 955 distance for this source is the minimum, according to the ordering 956 defined above, of the distances of all the finite updates ever sent 957 by this particular node for the prefix (prefix, plen) and the given 958 router-id. Feasibility distances are maintained in the source table, 959 the exact procedure is given in Section 3.7.3. 961 A received update is feasible when either it is a retraction (its 962 metric is FFFF hexadecimal), or the advertised distance is strictly 963 better, in the sense defined above, than the feasibility distance for 964 the corresponding source. More precisely, a route advertisement 965 carrying the quintuple (prefix, plen, router-id, seqno, metric) is 966 feasible if one of the following conditions holds: 968 o metric is infinite; or 970 o no entry exists in the source table indexed by (prefix, plen, 971 router-id); or 973 o an entry (prefix, plen, router-id, seqno', metric') exists in the 974 source table, and either 976 * seqno' < seqno or 978 * seqno = seqno' and metric < metric'. 980 Note that the feasibility condition considers the metric advertised 981 by the neighbour, not the route's metric; hence, a fluctuation in a 982 neighbour's cost cannot render a selected route unfeasible. Note 983 further that retractions (updates with infinite metric) are always 984 feasible, since they cannot possibly cause a routing loop. 986 3.5.2. Metric Computation 988 A route's metric is computed from the metric advertised by the 989 neighbour and the neighbour's link cost. Just like cost computation, 990 metric computation is considered a local policy matter; as far as 991 Babel is concerned, the function M(c, m) used for computing a metric 992 from a locally computed link cost and the metric advertised by a 993 neighbour MUST only satisfy the following conditions: 995 o if c is infinite, then M(c, m) is infinite; 997 o M is strictly monotonic: M(c, m) > m. 999 Additionally, the metric SHOULD satisfy the following condition: 1001 o M is left-distributive: if m <= m', then M(c, m) <= M(c, m'). 1003 Note that while strict monotonicity is essential to the integrity of 1004 the network (persistent routing loops may arise if it is not 1005 satisfied), left distributivity is not: if it is not satisfied, Babel 1006 will still converge to a loop-free configuration, but might not reach 1007 a global optimum (in fact, a global optimum may not even exist). 1009 As with cost computation, not all strategies for computing route 1010 metrics will give good results. In particular, some metrics are more 1011 likely than others to lead to routing instabilities (route flapping). 1012 In Appendix A.3, we give a number of examples of strictly monotonic, 1013 left-distributive routing metrics that are known to work well in 1014 practice. 1016 3.5.3. Encoding of Updates 1018 In a large network, the bulk of Babel traffic consists of route 1019 updates; hence, some care has been given to encoding them 1020 efficiently. An Update TLV itself only contains the prefix, seqno, 1021 and metric, while the next hop is derived either from the network- 1022 layer source address of the packet or from an explicit Next Hop TLV 1023 in the same packet. The router-id is derived from a separate Router- 1024 Id TLV in the same packet, which optimises the case when multiple 1025 updates are sent with the same router-id. 1027 Additionally, a prefix of the advertised prefix can be omitted in an 1028 Update TLV, in which case it is copied from a previous Update TLV in 1029 the same packet -- this is known as address compression 1030 (Section 4.6.9). 1032 Finally, as a special optimisation for the case when a router-id 1033 coincides with the interface-id part of an IPv6 address, the router- 1034 id can optionally be derived from the low-order bits of the 1035 advertised prefix. 1037 The encoding of updates is described in detail in Section 4.6. 1039 3.5.4. Route Acquisition 1041 When a Babel node receives an update (prefix, plen, router-id, seqno, 1042 metric) from a neighbour neigh with a link cost value equal to cost, 1043 it checks whether it already has a route table entry indexed by 1044 (prefix, plen, neigh). 1046 If no such entry exists: 1048 o if the update is unfeasible, it MAY be ignored; 1050 o if the metric is infinite (the update is a retraction of a route 1051 we do not know about), the update is ignored; 1053 o otherwise, a new entry is created in the route table, indexed by 1054 (prefix, plen, neigh), with source equal to (prefix, plen, router- 1055 id), seqno equal to seqno and an advertised metric equal to the 1056 metric carried by the update. 1058 If such an entry exists: 1060 o if the entry is currently selected, the update is unfeasible, and 1061 the router-id of the update is equal to the router-id of the 1062 entry, then the update MAY be ignored; 1064 o otherwise, the entry's sequence number, advertised metric, metric, 1065 and router-id are updated and, if the advertised metric is not 1066 infinite, the route's expiry timer is reset to a small multiple of 1067 the Interval value included in the update. If the update is 1068 unfeasible, then the (now unfeasible) entry MUST be immediately 1069 unselected. If the update caused the router-id of the entry to 1070 change, an update (possibly a retraction) MUST be sent in a timely 1071 manner (see Section 3.7.2). 1073 Note that the route table may contain unfeasible routes, either 1074 because they were created by an unfeasible update or due to a metric 1075 fluctuation. Such routes are never selected, since they are not 1076 known to be loop-free; should all the feasible routes become 1077 unusable, however, the unfeasible routes can be made feasible and 1078 therefore possible to select by sending requests along them (see 1079 Section 3.8.2). 1081 When a route's expiry timer triggers, the behaviour depends on 1082 whether the route's metric is finite. If the metric is finite, it is 1083 set to infinity and the expiry timer is reset. If the metric is 1084 already infinite, the route is flushed from the route table. 1086 After the route table is updated, the route selection procedure 1087 (Section 3.6) is run. 1089 3.5.5. Hold Time 1091 When a prefix P is retracted, because all routes are unfeasible or 1092 have an infinite metric (whether due to the expiry timer or to other 1093 reasons), and a shorter prefix P' that covers P is reachable, P' 1094 cannot in general be used for routing packets destined to P without 1095 running the risk of creating a routing loop (Section 2.8). 1097 To avoid this issue, whenever a prefix P is retracted, a route table 1098 entry with infinite metric is maintained as described in 1099 Section 3.5.4 above. As long as this entry is maintained, packets 1100 destined to an address within P MUST NOT be forwarded by following a 1101 route for a shorter prefix. This entry is removed as soon as a 1102 finite-metric update for prefix P is received and the resulting route 1103 selected. If no such update is forthcoming, the infinite metric 1104 entry SHOULD be maintained at least until it is guaranteed that no 1105 neighbour has selected the current node as next-hop for prefix P. 1106 This can be achieved by either: 1108 o waiting until the route's expiry timer has expired 1109 (Section 3.5.4); 1111 o sending a retraction with an acknowledgment request (Section 3.3) 1112 to every reachable neighbour that has not explicitly retracted 1113 prefix P and waiting for all acknowledgments. 1115 The former option is simpler and ensures that at that point, any 1116 routes for prefix P pointing at the current node have expired. 1117 However, since the expiry time can be as high as a few minutes, doing 1118 that prevents automatic aggregation by creating spurious black-holes 1119 for aggregated routes. The latter option is RECOMMENDED as it 1120 dramatically reduces the time for which a prefix is unreachable in 1121 the presence of aggregated routes. 1123 3.6. Route Selection 1125 Route selection is the process by which a single route for a given 1126 prefix is selected to be used for forwarding packets and to be re- 1127 advertised to a node's neighbours. 1129 Babel is designed to allow flexible route selection policies. As far 1130 as the protocol's correctness is concerned, the route selection 1131 policy MUST only satisfy the following properties: 1133 o a route with infinite metric (a retracted route) is never 1134 selected; 1136 o an unfeasible route is never selected. 1138 Note, however, that Babel does not naturally guarantee the stability 1139 of routing, and configuring conflicting route selection policies on 1140 different routers may lead to persistent route oscillation. 1142 Route selection is a difficult problem, since a good route selection 1143 policy needs to take into account multiple mutually contradictory 1144 criteria; in roughly decreasing order of importance, these are: 1146 o routes with a small metric should be preferred to routes with a 1147 large metric; 1149 o switching router-ids should be avoided; 1151 o routes through stable neighbours should be preferred to routes 1152 through unstable ones; 1154 o stable routes should be preferred to unstable ones; 1156 o switching next hops should be avoided. 1158 A simple but useful strategy is to choose the feasible route with the 1159 smallest metric, with a small amount of hysteresis applied to avoid 1160 switching router-ids too often. 1162 After the route selection procedure is run, triggered updates 1163 (Section 3.7.2) and requests (Section 3.8.2) are sent. 1165 3.7. Sending Updates 1167 A Babel speaker advertises to its neighbours its set of selected 1168 routes. Normally, this is done by sending one or more multicast 1169 packets containing Update TLVs on all of its connected interfaces; 1170 however, on link technologies where multicast is significantly more 1171 expensive than unicast, a node MAY choose to send multiple copies of 1172 updates in unicast packets, especially when the number of neighbours 1173 is small. 1175 Additionally, in order to ensure that any black-holes are reliably 1176 cleared in a timely manner, a Babel node sends retractions (updates 1177 with an infinite metric) for any recently retracted prefixes. 1179 If an update is for a route injected into the Babel domain by the 1180 local node (e.g., it carries the address of a local interface, the 1181 prefix of a directly attached network, or a prefix redistributed from 1182 a different routing protocol), the router-id is set to the local 1183 node's router-id, the metric is set to some arbitrary finite value 1184 (typically 0), and the seqno is set to the local router's sequence 1185 number. 1187 If an update is for a route learned from another Babel speaker, the 1188 router-id and sequence number are copied from the route table entry, 1189 and the metric is computed as specified in Section 3.5.2. 1191 3.7.1. Periodic Updates 1193 Every Babel speaker periodically advertises all of its selected 1194 routes on all of its interfaces, including any recently retracted 1195 routes. Since Babel doesn't suffer from routing loops (there is no 1196 "counting to infinity") and relies heavily on triggered updates 1197 (Section 3.7.2), this full dump only needs to happen infrequently. 1199 3.7.2. Triggered Updates 1201 In addition to periodic routing updates, a Babel speaker sends 1202 unscheduled, or triggered, updates in order to inform its neighbours 1203 of a significant change in the network topology. 1205 A change of router-id for the selected route to a given prefix may be 1206 indicative of a routing loop in formation; hence, a node MUST send a 1207 triggered update in a timely manner whenever it changes the selected 1208 router-id for a given destination. Additionally, it SHOULD make a 1209 reasonable attempt at ensuring that all reachable neighbours receive 1210 this update. 1212 There are two strategies for ensuring that. If the number of 1213 neighbours is small, then it is reasonable to send the update 1214 together with an acknowledgment request; the update is resent until 1215 all neighbours have acknowledged the packet, up to some number of 1216 times. If the number of neighbours is large, however, requesting 1217 acknowledgments from all of them might cause a non-negligible amount 1218 of network traffic; in that case, it may be preferable to simply 1219 repeat the update some reasonable number of times (say, 5 for 1220 wireless and 2 for wired links). 1222 A route retraction is somewhat less worrying: if the route retraction 1223 doesn't reach all neighbours, a black-hole might be created, which, 1224 unlike a routing loop, does not endanger the integrity of the 1225 network. When a route is retracted, a node SHOULD send a triggered 1226 update and SHOULD make a reasonable attempt at ensuring that all 1227 neighbours receive this retraction. 1229 Finally, a node MAY send a triggered update when the metric for a 1230 given prefix changes in a significant manner, due to a received 1231 update, because a link's cost has changed, or because a different 1232 next hop has been selected. A node SHOULD NOT send triggered updates 1233 for other reasons, such as when there is a minor fluctuation in a 1234 route's metric, when the selected next hop changes, or to propagate a 1235 new sequence number (except to satisfy a request, as specified in 1236 Section 3.8). 1238 3.7.3. Maintaining Feasibility Distances 1240 Before sending an update (prefix, plen, router-id, seqno, metric) 1241 with finite metric (i.e., not a route retraction), a Babel node 1242 updates the feasibility distance maintained in the source table. 1243 This is done as follows. 1245 If no entry indexed by (prefix, plen, router-id) exists in the source 1246 table, then one is created with value (prefix, plen, router-id, 1247 seqno, metric). 1249 If an entry (prefix, plen, router-id, seqno', metric') exists, then 1250 it is updated as follows: 1252 o if seqno > seqno', then seqno' := seqno, metric' := metric; 1254 o if seqno = seqno' and metric' > metric, then metric' := metric; 1256 o otherwise, nothing needs to be done. 1258 The garbage-collection timer for the entry is then reset. Note that 1259 the feasibility distance is not updated and the garbage-collection 1260 timer is not reset when a retraction (an update with infinite metric) 1261 is sent. 1263 When the garbage-collection timer expires, the entry is removed from 1264 the source table. 1266 3.7.4. Split Horizon 1268 When running over a transitive, symmetric link technology, e.g., a 1269 point-to-point link or a wired LAN technology such as Ethernet, a 1270 Babel node SHOULD use an optimisation known as split horizon. When 1271 split horizon is used on a given interface, a routing update for 1272 prefix P is not sent on the particular interface over which the 1273 selected route towards prefix P was learnt. 1275 Split horizon SHOULD NOT be applied to an interface unless the 1276 interface is known to be symmetric and transitive; in particular, 1277 split horizon is not applicable to decentralised wireless link 1278 technologies (e.g., IEEE 802.11 in ad hoc mode) when routing updates 1279 are sent over multicast. 1281 3.8. Explicit Requests 1283 In normal operation, a node's route table is populated by the regular 1284 and triggered updates sent by its neighbours. Under some 1285 circumstances, however, a node sends explicit requests in order to 1286 cause a resynchronisation with the source after a mobility event or 1287 to prevent a route from spuriously expiring. 1289 The Babel protocol provides two kinds of explicit requests: route 1290 requests, which simply request an update for a given prefix, and 1291 seqno requests, which request an update for a given prefix with a 1292 specific sequence number. The former are never forwarded; the latter 1293 are forwarded if they cannot be satisfied by the receiver. 1295 3.8.1. Handling Requests 1297 Upon receiving a request, a node either forwards the request or sends 1298 an update in reply to the request, as described in the following 1299 sections. If this causes an update to be sent, the update is either 1300 sent to a multicast address on the interface on which the request was 1301 received, or to the unicast address of the neighbour that sent the 1302 request. 1304 The exact behaviour is different for route requests and seqno 1305 requests. 1307 3.8.1.1. Route Requests 1309 When a node receives a route request for a given prefix, it checks 1310 its route table for a selected route to this exact prefix. If such a 1311 route exists, it MUST send an update (over unicast or over 1312 multicast); if such a route does not exist, it MUST send a retraction 1313 for that prefix. 1315 When a node receives a wildcard route request, it SHOULD send a full 1316 route table dump. Full route dumps MAY be rate-limited, especially 1317 if they are sent over multicast. 1319 3.8.1.2. Seqno Requests 1321 When a node receives a seqno request for a given router-id and 1322 sequence number, it checks whether its route table contains a 1323 selected entry for that prefix. If a selected route for the given 1324 prefix exists, it has finite metric, and either the router-ids are 1325 different or the router-ids are equal and the entry's sequence number 1326 is no smaller (modulo 2^16) than the requested sequence number, the 1327 node MUST send an update for the given prefix. If the router-ids 1328 match but the requested seqno is larger (modulo 2^16) than the route 1329 entry's, the node compares the router-id against its own router-id. 1330 If the router-id is its own, then it increases its sequence number by 1331 1 (modulo 2^16) and sends an update. A node MUST NOT increase its 1332 sequence number by more than 1 in response to a seqno request. 1334 Otherwise, if the requested router-id is not its own, the received 1335 request's hop count is 2 or more, and the node is advertising the 1336 prefix to its neighbours, the node selects a neighbour to forward the 1337 request to as follows: 1339 o if the node has one or more feasible routes toward the requested 1340 prefix with a next hop that is not the requesting node, then the 1341 node MUST forward the request to the next hop of one such route; 1343 o otherwise, if the node has one or more (not necessarily feasible) 1344 routes to the requested prefix with a next hop that is not the 1345 requesting node, then the node SHOULD forward the request to the 1346 next hop of one such route. 1348 In order to actually forward the request, the node decrements the hop 1349 count and sends the request in a unicast packet destined to the 1350 selected neighbour. 1352 A node SHOULD maintain a list of recently forwarded seqno requests 1353 and forward the reply (an update with a seqno sufficiently large to 1354 satisfy the request) in a timely manner. A node SHOULD compare every 1355 incoming seqno request against its list of recently forwarded seqno 1356 requests and avoid forwarding it if it is redundant (i.e., if it has 1357 recently sent a request with the same prefix, router-id and a seqno 1358 that is not smaller modulo 2^16). 1360 Since the request-forwarding mechanism does not necessarily obey the 1361 feasibility condition, it may get caught in routing loops; hence, 1362 requests carry a hop count to limit the time during which they remain 1363 in the network. However, since requests are only ever forwarded as 1364 unicast packets, the initial hop count need not be kept particularly 1365 low, and performing an expanding horizon search is not necessary. A 1366 single request MUST NOT be duplicated: it MUST NOT be forwarded to a 1367 multicast address, and it MUST NOT be forwarded to multiple 1368 neighbours. However, if a seqno request is resent by its originator, 1369 the subsequent copies MAY be forwarded to a different neighbour than 1370 the initial one. 1372 3.8.2. Sending Requests 1374 A Babel node MAY send a route or seqno request at any time, to a 1375 multicast or a unicast address; there is only one case when 1376 originating requests is required (Section 3.8.2.1). 1378 3.8.2.1. Avoiding Starvation 1380 When a route is retracted or expires, a Babel node usually switches 1381 to another feasible route for the same prefix. It may be the case, 1382 however, that no such routes are available. 1384 A node that has lost all feasible routes to a given destination but 1385 still has unexpired unfeasible routes to that destination MUST send a 1386 seqno request; if it doesn't have any such routes, it MAY still send 1387 a seqno request. The router-id of the request is set to the router- 1388 id of the route that it has just lost, and the requested seqno is the 1389 value contained in the source table plus 1. 1391 If the node has any (unfeasible) routes to the requested destination, 1392 then it MUST send the request to at least one of the next-hop 1393 neighbours that advertised these routes, and SHOULD send it to all of 1394 them; in any case, it MAY send the request to any other neighbours, 1395 whether they advertise a route to the requested destination or not. 1396 A simple implementation strategy is therefore to unconditionally 1397 multicast the request over all interfaces. 1399 Similar requests will be sent by other nodes that are affected by the 1400 route's loss. If the network is still connected, and assuming no 1401 packet loss, then at least one of these requests will be forwarded to 1402 the source, resulting in a route being advertised with a new sequence 1403 number. (Due to duplicate suppression, only a small number of such 1404 requests will actually reach the source.) 1406 In order to compensate for packet loss, a node SHOULD repeat such a 1407 request a small number of times if no route becomes feasible within a 1408 short time. In the presence of heavy packet loss, however, all such 1409 requests might be lost; in that case, the mechanism in the next 1410 section will eventually ensure that a new seqno is received. 1412 3.8.2.2. Dealing with Unfeasible Updates 1414 When a route's metric increases, a node might receive an unfeasible 1415 update for a route that it has currently selected. As specified in 1416 Section 3.5.1, the receiving node will either ignore the update or 1417 unselect the route. 1419 In order to keep routes from spuriously expiring because they have 1420 become unfeasible, a node SHOULD send a unicast seqno request when it 1421 receives an unfeasible update for a route that is currently selected. 1422 The requested sequence number is computed from the source table as in 1423 Section 3.8.2.1 above. 1425 Additionally, since metric computation does not necessarily coincide 1426 with the delay in propagating updates, a node might receive an 1427 unfeasible update from a currently unselected neighbour that is 1428 preferable to the currently selected route (e.g., because it has a 1429 much smaller metric); in that case, the node SHOULD send a unicast 1430 seqno request to the neighbour that advertised the preferable update. 1432 3.8.2.3. Preventing Routes from Expiring 1434 In normal operation, a route's expiry timer never triggers: since a 1435 route's hold time is computed from an explicit interval included in 1436 Update TLVs, a new update (possibly a retraction) should arrive in 1437 time to prevent a route from expiring. 1439 In the presence of packet loss, however, it may be the case that no 1440 update is successfully received for an extended period of time, 1441 causing a route to expire. In order to avoid such spurious expiry, 1442 shortly before a selected route expires, a Babel node SHOULD send a 1443 unicast route request to the neighbour that advertised this route; 1444 since nodes always send either updates or retractions in response to 1445 non-wildcard route requests (Section 3.8.1.1), this will usually 1446 result in the route being either refreshed or retracted. 1448 3.8.2.4. Acquiring New Neighbours 1450 In order to speed up convergence after a mobility event, a node MAY 1451 send a unicast wildcard request after acquiring a new neighbour. 1452 Additionally, a node MAY send a small number of multicast wildcard 1453 requests shortly after booting. Note however that doing that 1454 carelessly can cause serious congestion when a whole network is 1455 rebooted, especially on link layers with high per-packet overhead 1456 (e.g., IEEE 802.11). 1458 4. Protocol Encoding 1460 A Babel packet is sent as the body of a UDP datagram, with network- 1461 layer hop count set to 1, destined to a well-known multicast address 1462 or to a unicast address, over IPv4 or IPv6; in the case of IPv6, 1463 these addresses are link-local. Both the source and destination UDP 1464 port are set to a well-known port number. A Babel packet MUST be 1465 silently ignored unless its source address is either a link-local 1466 IPv6 address or an IPv4 address belonging to the local network, and 1467 its source port is the well-known Babel port. It MAY be silently 1468 ignored if its destination address is a global IPv6 address. 1470 In order to minimise the number of packets being sent while avoiding 1471 lower-layer fragmentation, a Babel node SHOULD attempt to maximise 1472 the size of the packets it sends, up to the outgoing interface's MTU 1473 adjusted for lower-layer headers (28 octets for UDP over IPv4, 48 1474 octets for UDP over IPv6). It MUST NOT send packets larger than the 1475 attached interface's MTU adjusted for lower-layer headers or 512 1476 octets, whichever is larger, but not exceeding 2^16 - 1 adjusted for 1477 lower-layer headers. Every Babel speaker MUST be able to receive 1478 packets that are as large as any attached interface's MTU adjusted 1479 for lower-layer headers or 512 octets, whichever is larger. Babel 1480 packets MUST NOT be sent in IPv6 Jumbograms. 1482 In order to avoid global synchronisation of a Babel network and to 1483 aggregate multiple TLVs into large packets, a Babel node SHOULD 1484 buffer every TLV and delay sending a packet by a small, randomly 1485 chosen delay [JITTER]. In order to allow accurate computation of 1486 packet loss rates, this delay MUST NOT be larger than half the 1487 advertised Hello interval. 1489 4.1. Data Types 1491 4.1.1. Interval 1493 Relative times are carried as 16-bit values specifying a number of 1494 centiseconds (hundredths of a second). This allows times up to 1495 roughly 11 minutes with a granularity of 10ms, which should cover all 1496 reasonable applications of Babel. 1498 4.1.2. Router-Id 1500 A router-id is an arbitrary 8-octet value. A router-id MUST NOT 1501 consist of either all zeroes or all ones. 1503 4.1.3. Address 1505 Since the bulk of the protocol is taken by addresses, multiple ways 1506 of encoding addresses are defined. Additionally, a common subnet 1507 prefix may be omitted when multiple addresses are sent in a single 1508 packet -- this is known as address compression (Section 4.6.9). 1510 Address encodings: 1512 o AE 0: wildcard address. The value is 0 octets long. 1514 o AE 1: IPv4 address. Compression is allowed. 4 octets or less. 1516 o AE 2: IPv6 address. Compression is allowed. 16 octets or less. 1518 o AE 3: link-local IPv6 address. Compression is not allowed. The 1519 value is 8 octets long, a prefix of fe80::/64 is implied. 1521 The address family associated to an address encoding is either IPv4 1522 or IPv6; it is undefined for AE 0, IPv4 for AE 1, and IPv6 for AEs 2 1523 and 3. 1525 4.1.4. Prefixes 1527 A network prefix is encoded just like a network address, but it is 1528 stored in the smallest number of octets that are enough to hold the 1529 significant bits (up to the prefix length). 1531 4.2. Packet Format 1533 A Babel packet consists of a 4-octet header, followed by a sequence 1534 of TLVs. 1536 0 1 2 3 1537 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 1538 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1539 | Magic | Version | Body length | 1540 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1541 | Packet Body ... 1542 +-+-+-+-+-+-+-+-+-+-+-+-+- 1544 Fields : 1546 Magic The arbitrary but carefully chosen value 42 (decimal); 1547 packets with a first octet different from 42 MUST be 1548 silently ignored. 1550 Version This document specifies version 2 of the Babel protocol. 1551 Packets with a second octet different from 2 MUST be 1552 silently ignored. 1554 Body length The length in octets of the body following the packet 1555 header (excluding the Magic, Version and Body length 1556 fields). 1558 Body The packet body; a sequence of TLVs. 1560 Any data following the body MUST be silently ignored. 1562 4.3. TLV Format 1564 With the exception of Pad1, all TLVs have the following structure: 1566 0 1 2 3 1567 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 1568 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1569 | Type | Length | Payload... 1570 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1572 Fields : 1574 Type The type of the TLV. 1576 Length The length of the body, exclusive of the Type and Length 1577 fields. If the body is longer than the expected length of 1578 a given type of TLV, any extra data MUST be silently 1579 ignored. 1581 Payload The TLV payload, which consists of a body and, for selected 1582 TLV types, an optional list of sub-TLVs. 1584 TLVs with an unknown type value MUST be silently ignored. 1586 4.4. Sub-TLV Format 1588 Every TLV carries an explicit length in its header; however, most 1589 TLVs are self-terminating, in the sense that it is possible to 1590 determine the length of the body without reference to the explicit 1591 Length field. If a TLV has a self-terminating format, then it MAY 1592 allow a sequence of sub-TLVs to follow the body. 1594 Sub-TLVs have the same structure as TLVs. With the exception of 1595 PAD1, all TLVs have the following structure: 1597 0 1 2 3 1598 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 1599 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1600 | Type | Length | Body... 1601 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1603 Fields : 1605 Type The type of the sub-TLV. 1607 Length The length of the body, in octets, exclusive of the Type 1608 and Length fields. 1610 Body The sub-TLV body, the interpretation of which depends on 1611 both the type of the sub-TLV and the type of the TLV within 1612 which it is embedded. 1614 The most-significant bit of the sub-TLV, called the mandatory bit, 1615 indicates how to handle unknown sub-TLVs. If the mandatory bit is 1616 not set, then an unknown sub-TLV MUST be silently ignored, and the 1617 rest of the TLV processed normally. If the mandatory bit is set, 1618 then the whole enclosing TLV MUST be silently ignored (except for 1619 updating the parser state by a Router-Id, Next-Hop or Update TLV, see 1620 Section 4.6.7, Section 4.6.8, and Section 4.6.9). 1622 4.5. Parser state 1624 Babel uses a stateful parser: a TLV may refer to data from a previous 1625 TLV. The parser state consists of the following pieces of data: 1627 o for each address encoding that allows compression, the current 1628 default prefix; this is undefined at the start of the packet, and 1629 is updated by each Update TLV with the Prefix flag set 1630 (Section 4.6.9); 1632 o for each address family (IPv4 or IPv6), the current next-hop; this 1633 is the source address of the enclosing packet for the matching 1634 address family at the start of a packet, and is updated by each 1635 Next-Hop TLV (Section 4.6.8); 1637 o the current router-id; this is undefined at the start of the 1638 packet, and is updated by each Router-ID TLV (Section 4.6.7) and 1639 by each Update TLV with Router-Id flag set. 1641 Since the parser state is separate from the bulk of Babel's state, 1642 and since for correct parsing it must be identical across 1643 implementations, it is updated before checking for mandatory TLVs: 1644 parsing a TLV MUST update the parser state even if the TLV is 1645 otherwise ignored due to an unknown mandatory sub-TLV. 1647 4.6. Details of Specific TLVs 1649 4.6.1. Pad1 1651 0 1652 0 1 2 3 4 5 6 7 1653 +-+-+-+-+-+-+-+-+ 1654 | Type = 0 | 1655 +-+-+-+-+-+-+-+-+ 1657 Fields : 1659 Type Set to 0 to indicate a Pad1 TLV. 1661 This TLV is silently ignored on reception. 1663 4.6.2. PadN 1665 0 1 2 3 1666 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 1667 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1668 | Type = 1 | Length | MBZ... 1669 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 1670 Fields : 1672 Type Set to 1 to indicate a PadN TLV. 1674 Length The length of the body, exclusive of the Type and Length 1675 fields. 1677 MBZ Set to 0 on transmission. 1679 This TLV is silently ignored on reception. 1681 4.6.3. Acknowledgment Request 1683 0 1 2 3 1684 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 1685 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1686 | Type = 2 | Length | Reserved | 1687 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1688 | Nonce | Interval | 1689 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1691 This TLV requests that the receiver send an Acknowledgment TLV within 1692 the number of centiseconds specified by the Interval field. 1694 Fields : 1696 Type Set to 2 to indicate an Acknowledgment Request TLV. 1698 Length The length of the body, exclusive of the Type and Length 1699 fields. 1701 Reserved Sent as 0 and MUST be ignored on reception. 1703 Nonce An arbitrary value that will be echoed in the receiver's 1704 Acknowledgment TLV. 1706 Interval A time interval in centiseconds after which the sender will 1707 assume that this packet has been lost. This MUST NOT be 0. 1708 The receiver MUST send an Acknowledgment TLV before this 1709 time has elapsed (with a margin allowing for propagation 1710 time). 1712 This TLV is self-terminating, and allows sub-TLVs. 1714 4.6.4. Acknowledgment 1716 0 1 2 3 1717 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 1718 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1719 | Type = 3 | Length | Nonce | 1720 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1722 This TLV is sent by a node upon receiving an Acknowledgment Request. 1724 Fields : 1726 Type Set to 3 to indicate an Acknowledgment TLV. 1728 Length The length of the body, exclusive of the Type and Length 1729 fields. 1731 Nonce Set to the Nonce value of the Acknowledgment Request that 1732 prompted this Acknowledgment. 1734 Since nonce values are not globally unique, this TLV MUST be sent to 1735 a unicast address. 1737 This TLV is self-terminating, and allows sub-TLVs. 1739 4.6.5. Hello 1741 0 1 2 3 1742 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 1743 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1744 | Type = 4 | Length | Flags | 1745 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1746 | Seqno | Interval | 1747 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1749 This TLV is used for neighbour discovery and for determining a 1750 neighbour's reception cost. 1752 Fields : 1754 Type Set to 4 to indicate a Hello TLV. 1756 Length The length of the body, exclusive of the Type and Length 1757 fields. 1759 Flags The individual bits of this field specify special handling 1760 of this TLV (see below). 1762 Seqno If the Unicast flag is set, this is the value of the 1763 sending node's outgoing Unicast Hello seqno for this 1764 neighbour. Otherwise, it is the sending node's outgoing 1765 Multicast Hello seqno for this interface. 1767 Interval If non-zero, this is an upper bound, expressed in 1768 centiseconds, on the time after which the sending node will 1769 send a new scheduled Hello TLV with the same setting of the 1770 Unicast flag. If this is 0, then this Hello represents an 1771 unscheduled Hello, and doesn't carry any new information 1772 about times at which Hellos are sent. 1774 The Flags field is interpreted as follows: 1776 0 1 1777 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 1778 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1779 |U|X|X|X|X|X|X|X|X|X|X|X|X|X|X|X| 1780 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1782 o U (Unicast) flag (8000 hexadecimal): if set, then this Hello 1783 represents a Unicast Hello, otherwise it represents a Multicast 1784 Hello; 1786 o X: all other bits MUST be sent as 0 and silently ignored on 1787 reception. 1789 Every time a Hello is sent, the corresponding seqno counter MUST be 1790 incremented. Since there is a single seqno counter for all the 1791 Multicast Hellos sent by a given node over a given interface, if the 1792 Unicast flag is not set, this TLV MUST be sent to all neighbors on 1793 this link, which can be achieved by sending to a multicast 1794 destination, or by sending multiple packets to the unicast addresses 1795 of all reachable neighbours. Conversely, if the Unicast flag is set, 1796 this TLV MUST be sent to a single neighbour, which can achieved by 1797 sending to a unicast destination. In order to avoid large 1798 discontinuities in link quality, multiple Hello TLVs SHOULD NOT be 1799 sent in the same packet. 1801 This TLV is self-terminating, and allows sub-TLVs. 1803 4.6.6. IHU 1804 0 1 2 3 1805 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 1806 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1807 | Type = 5 | Length | AE | Reserved | 1808 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1809 | Rxcost | Interval | 1810 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1811 | Address... 1812 +-+-+-+-+-+-+-+-+-+-+-+- 1814 An IHU ("I Heard You") TLV is used for confirming bidirectional 1815 reachability and carrying a link's transmission cost. 1817 Fields : 1819 Type Set to 5 to indicate an IHU TLV. 1821 Length The length of the body, exclusive of the Type and Length 1822 fields. 1824 AE The encoding of the Address field. This should be 1 or 3 1825 in most cases. As an optimisation, it MAY be 0 if the TLV 1826 is sent to a unicast address, if the association is over a 1827 point-to-point link, or when bidirectional reachability is 1828 ascertained by means outside of the Babel protocol. 1830 Reserved Sent as 0 and MUST be ignored on reception. 1832 Rxcost The rxcost according to the sending node of the interface 1833 whose address is specified in the Address field. The value 1834 FFFF hexadecimal (infinity) indicates that this interface 1835 is unreachable. 1837 Interval An upper bound, expressed in centiseconds, on the time 1838 after which the sending node will send a new IHU; this MUST 1839 NOT be 0. The receiving node will use this value in order 1840 to compute a hold time for this symmetric association. 1842 Address The address of the destination node, in the format 1843 specified by the AE field. Address compression is not 1844 allowed. 1846 Conceptually, an IHU is destined to a single neighbour. However, IHU 1847 TLVs contain an explicit destination address, and MAY be sent to a 1848 multicast address, as this allows aggregation of IHUs destined to 1849 distinct neighbours into a single packet and avoids the need for an 1850 ARP or Neighbour Discovery exchange when a neighbour is not being 1851 used for data traffic. 1853 IHU TLVs with an unknown value in the AE field MUST be silently 1854 ignored. 1856 This TLV is self-terminating, and allows sub-TLVs. 1858 4.6.7. Router-Id 1860 0 1 2 3 1861 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 1862 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1863 | Type = 6 | Length | Reserved | 1864 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1865 | | 1866 + Router-Id + 1867 | | 1868 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1870 A Router-Id TLV establishes a router-id that is implied by subsequent 1871 Update TLVs. This TLV sets the router-id even if it is otherwise 1872 ignored due to an unknown mandatory sub-TLV. 1874 Fields : 1876 Type Set to 6 to indicate a Router-Id TLV. 1878 Length The length of the body, exclusive of the Type and Length 1879 fields. 1881 Reserved Sent as 0 and MUST be ignored on reception. 1883 Router-Id The router-id for routes advertised in subsequent Update 1884 TLVs. This MUST NOT consist of all zeroes or all ones. 1886 This TLV is self-terminating, and allows sub-TLVs. 1888 4.6.8. Next Hop 1890 0 1 2 3 1891 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 1892 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1893 | Type = 7 | Length | AE | Reserved | 1894 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1895 | Next hop... 1896 +-+-+-+-+-+-+-+-+-+-+-+- 1898 A Next Hop TLV establishes a next-hop address for a given address 1899 family (IPv4 or IPv6) that is implied in subsequent Update TLVs. 1901 This TLV sets up the next-hop for subsequent Update TLVs even if it 1902 is otherwise ignored due to an unknown mandatory sub-TLV. 1904 Fields : 1906 Type Set to 7 to indicate a Next Hop TLV. 1908 Length The length of the body, exclusive of the Type and Length 1909 fields. 1911 AE The encoding of the Address field. This SHOULD be 1 (IPv4) 1912 or 3 (link-local IPv6), and MUST NOT be 0. 1914 Reserved Sent as 0 and MUST be ignored on reception. 1916 Next hop The next-hop address advertised by subsequent Update TLVs, 1917 for this address family. 1919 When the address family matches the network-layer protocol that this 1920 packet is transported over, a Next Hop TLV is not needed: in the 1921 absence of a Next Hop TLV in a given address family, the next hop 1922 address is taken to be the source address of the packet. 1924 Next Hop TLVs with an unknown value for the AE field MUST be silently 1925 ignored. 1927 This TLV is self-terminating, and allows sub-TLVs. 1929 4.6.9. Update 1931 0 1 2 3 1932 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 1933 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1934 | Type = 8 | Length | AE | Flags | 1935 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1936 | Plen | Omitted | Interval | 1937 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1938 | Seqno | Metric | 1939 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1940 | Prefix... 1941 +-+-+-+-+-+-+-+-+-+-+-+- 1943 An Update TLV advertises or retracts a route. As an optimisation, it 1944 can optionally have the side effect of establishing a new implied 1945 router-id and a new default prefix. 1947 Fields : 1949 Type Set to 8 to indicate an Update TLV. 1951 Length The length of the body, exclusive of the Type and Length 1952 fields. 1954 AE The encoding of the Prefix field. 1956 Flags The individual bits of this field specify special handling 1957 of this TLV (see below). 1959 Plen The length of the advertised prefix. 1961 Omitted The number of octets that have been omitted at the 1962 beginning of the advertised prefix and that should be taken 1963 from a preceding Update TLV in the same address family with 1964 the Prefix flag set. 1966 Interval An upper bound, expressed in centiseconds, on the time 1967 after which the sending node will send a new update for 1968 this prefix. This MUST NOT be 0. The receiving node will 1969 use this value to compute a hold time for the route table 1970 entry. The value FFFF hexadecimal (infinity) expresses 1971 that this announcement will not be repeated unless a 1972 request is received (Section 3.8.2.3). 1974 Seqno The originator's sequence number for this update. 1976 Metric The sender's metric for this route. The value FFFF 1977 hexadecimal (infinity) means that this is a route 1978 retraction. 1980 Prefix The prefix being advertised. This field's size is 1981 (Plen/8 - Omitted) rounded upwards. 1983 The Flags field is interpreted as follows: 1985 0 1 2 3 4 5 6 7 1986 +-+-+-+-+-+-+-+-+ 1987 |P|R|X|X|X|X|X|X| 1988 +-+-+-+-+-+-+-+-+ 1990 o P (Prefix) flag (80 hexadecimal): if set, then this Update 1991 establishes a new default prefix for subsequent Update TLVs with a 1992 matching address encoding within the same packet, even if this TLV 1993 is otherwise ignored due to an unknown mandatory sub-TLV; 1995 o R (Router-Id) flag (40 hexadecimal): if set, then this TLV 1996 establishes a new default router-id for this TLV and subsequent 1997 Update TLVs in the same packet, even if this TLV is otherwise 1998 ignored due to an unknown mandatory sub-TLV. This router-id is 1999 computed from the first address of the advertised prefix as 2000 follows: 2002 * if the length of the address is 8 octets or more, then the new 2003 router-id is taken from the 8 last octets of the address; 2005 * if the length of the address is smaller than 8 octets, then the 2006 new router-id consists of the required number of zero octets 2007 followed by the address, i.e., the address is stored on the 2008 right of the router-id. For example, for an IPv4 address, the 2009 router-id consists of 4 octets of zeroes followed by the IPv4 2010 address. 2012 o X: all other bits MUST be sent as 0 and silently ignored on 2013 reception. 2015 The prefix being advertised by an Update TLV is computed as follows: 2017 o the first Omitted octets of the prefix are taken from the previous 2018 Update TLV with the Prefix flag set and the same address encoding, 2019 even if it was ignored due to an unknown mandatory sub-TLV; 2021 o the next (Plen/8 - Omitted) rounded upwards octets are taken from 2022 the Prefix field; 2024 o the remaining octets are set to 0. If AE is 3 (link-local IPv6), 2025 Omitted MUST be 0) 2027 If the Metric field is finite, the router-id of the originating node 2028 for this announcement is taken from the prefix advertised by this 2029 Update if the Router-Id flag is set, computed as described above. 2030 Otherwise, it is taken either from the preceding Router-Id packet, or 2031 the preceding Update packet with the Router-Id flag set, whichever 2032 comes last, even if that TLV is otherwise ignored due to an unknown 2033 mandatory sub-TLV. 2035 The next-hop address for this update is taken from the last preceding 2036 Next Hop TLV with a matching address family (IPv4 or IPv6) in the 2037 same packet even if it was otherwise ignored due to an unknown 2038 mandatory sub-TLV; if no such TLV exists, it is taken from the 2039 network-layer source address of this packet. 2041 If the metric field is FFFF hexadecimal, this TLV specifies a 2042 retraction. In that case, the router-id, next-hop and seqno are not 2043 used. AE MAY then be 0, in which case this Update retracts all of 2044 the routes previously advertised by the sending interface. If the 2045 metric is finite, AE MUST NOT be 0. If the metric is infinite and AE 2046 is 0, Plen and Omitted MUST both be 0. 2048 Update TLVs with an unknown value in the AE field MUST be silently 2049 ignored. 2051 This TLV is self-terminating, and allows sub-TLVs. 2053 4.6.10. Route Request 2055 0 1 2 3 2056 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 2057 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2058 | Type = 9 | Length | AE | Plen | 2059 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2060 | Prefix... 2061 +-+-+-+-+-+-+-+-+-+-+-+- 2063 A Route Request TLV prompts the receiver to send an update for a 2064 given prefix, or a full route table dump. 2066 Fields : 2068 Type Set to 9 to indicate a Route Request TLV. 2070 Length The length of the body, exclusive of the Type and Length 2071 fields. 2073 AE The encoding of the Prefix field. The value 0 specifies 2074 that this is a request for a full route table dump (a 2075 wildcard request). 2077 Plen The length of the requested prefix. 2079 Prefix The prefix being requested. This field's size is Plen/8 2080 rounded upwards. 2082 A Request TLV prompts the receiver to send an update message 2083 (possibly a retraction) for the prefix specified by the AE, Plen, and 2084 Prefix fields, or a full dump of its route table if AE is 0 (in which 2085 case Plen MUST be 0 and Prefix is of length 0). 2087 This TLV is self-terminating, and allows sub-TLVs. 2089 4.6.11. Seqno Request 2091 0 1 2 3 2092 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 2093 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2094 | Type = 10 | Length | AE | Plen | 2095 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2096 | Seqno | Hop Count | Reserved | 2097 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2098 | | 2099 + Router-Id + 2100 | | 2101 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2102 | Prefix... 2103 +-+-+-+-+-+-+-+-+-+-+ 2105 A Seqno Request TLV prompts the receiver to send an Update for a 2106 given prefix with a given sequence number, or to forward the request 2107 further if it cannot be satisfied locally. 2109 Fields : 2111 Type Set to 10 to indicate a Seqno Request message. 2113 Length The length of the body, exclusive of the Type and Length 2114 fields. 2116 AE The encoding of the Prefix field. This MUST NOT be 0. 2118 Plen The length of the requested prefix. 2120 Seqno The sequence number that is being requested. 2122 Hop Count The maximum number of times that this TLV may be forwarded, 2123 plus 1. This MUST NOT be 0. 2125 Reserved Sent as 0 and MUST be ignored on reception. 2127 Router Id The Router-Id that is being requested. This MUST NOT 2128 consist of all zeroes or all ones. 2130 Prefix The prefix being requested. This field's size is Plen/8 2131 rounded upwards. 2133 A Seqno Request TLV prompts the receiving node to send a finite- 2134 metric Update for the prefix specified by the AE, Plen, and Prefix 2135 fields, with either a router-id different from what is specified by 2136 the Router-Id field, or a Seqno no less (modulo 2^16) than what is 2137 specified by the Seqno field. If this request cannot be satisfied 2138 locally, then it is forwarded according to the rules set out in 2139 Section 3.8.1.2. 2141 While a Seqno Request MAY be sent to a multicast address, it MUST NOT 2142 be forwarded to a multicast address and MUST NOT be forwarded to more 2143 than one neighbour. A request MUST NOT be forwarded if its Hop Count 2144 field is 1. 2146 This TLV is self-terminating, and allows sub-TLVs. 2148 4.7. Details of specific sub-TLVs 2150 4.7.1. Pad1 2152 0 1 2 3 4 5 6 7 2153 +-+-+-+-+-+-+-+-+ 2154 | Type = 0 | 2155 +-+-+-+-+-+-+-+-+ 2157 Fields : 2159 Type Set to 0 to indicate a Pad1 sub-TLV. 2161 This sub-TLV is silently ignored on reception. It is allowed within 2162 any TLV that allows sub-TLVs. 2164 4.7.2. PadN 2166 0 1 2 3 2167 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 2168 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2169 | Type = 1 | Length | MBZ... 2170 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- 2172 Fields : 2174 Type Set to 1 to indicate a PadN sub-TLV. 2176 Length The length of the body, in octets, exclusive of the Type 2177 and Length fields. 2179 MBZ Set to 0 on transmission. 2181 This sub-TLV is silently ignored on reception. It is allowed within 2182 any TLV that allows sub-TLVs. 2184 5. IANA Considerations 2186 IANA has registered the UDP port number 6696, called "babel", for use 2187 by the Babel protocol. 2189 IANA has registered the IPv6 multicast group ff02::1:6 and the IPv4 2190 multicast group 224.0.0.111 for use by the Babel protocol. 2192 IANA has created a registry called "Babel TLV Types". The values in 2193 this registry are not changed by this specification. 2195 IANA has created a registry called "Babel sub-TLV Types". Due to the 2196 addition of a Mandatory bit to the Babel protocol, the values in the 2197 "Babel sub-TLV Types" registry are amended as follows: 2199 +---------+-----------------------------------------+---------------+ 2200 | Type | Name | Reference | 2201 +---------+-----------------------------------------+---------------+ 2202 | 0 | Pad1 | this document | 2203 | | | | 2204 | 1 | PadN | this document | 2205 | | | | 2206 | 112-126 | Reserved for Experimental Use | this document | 2207 | | | | 2208 | 127 | Reserved for expansion of the type | this document | 2209 | | space | | 2210 | | | | 2211 | 240-254 | Reserved for Experimental Use | this document | 2212 | | | | 2213 | 255 | Reserved for expansion of the type | this document | 2214 | | space | | 2215 +---------+-----------------------------------------+---------------+ 2217 Existing assignments in the "Babel sub-TLV Types" registry in the 2218 range 2 to 111 are not changed by this specification. The values 224 2219 through 239, previously reserved for Experimental Use, are now 2220 unassigned. 2222 IANA has created a registry called "Babel Flags Values". IANA is 2223 instructed to rename this registry to "Babel Update Flags Values", 2224 with its contents unchanged. 2226 IANA is instructed to create a new registry called "Babel Hello Flags 2227 Values". The allocation policy for this registry is Specification 2228 Required [RFC5226]. The initial values in this registry are as 2229 follows: 2231 +------+------------+---------------+ 2232 | Bit | Name | Reference | 2233 +------+------------+---------------+ 2234 | 0 | Unicast | this document | 2235 | | | | 2236 | 1-15 | Unassigned | | 2237 +------+------------+---------------+ 2239 IANA is instructed to replace all references to RFCs 6126 and 7557 in 2240 all of the registries mentioned above by references to this document. 2242 6. Security Considerations 2244 As defined in this document, Babel is a completely insecure protocol. 2245 Any attacker can misdirect data traffic by advertising routes with a 2246 low metric or a high seqno. This issue can be solved either by a 2247 lower-layer security mechanism (e.g. link-layer security), or by 2248 deploying a suitable authentication mechanism within Babel itself. 2249 With the exception of Hello TLVs used for discovery, Babel control 2250 traffic can be carried over unicast, which makes it possible to 2251 protect Babel traffic with a protocol that can only protect unicast 2252 data, for example IPsec with IKEv2, or DTLS. 2254 The information that a Babel node announces to the whole routing 2255 domain is often sufficient to determine a mobile node's physical 2256 location with reasonable precision. The privacy issues that this 2257 causes can be mitigated somewhat by using randomly chosen router-ids 2258 and randomly chosen IP addresses, and changing them periodically. 2260 When carried over IPv6, Babel packets are ignored unless they are 2261 sent from a link-local IPv6 address; since routers don't forward 2262 link-local IPv6 packets, this provides protection against spoofed 2263 Babel packets being sent from the global Internet. No such natural 2264 protection exists when Babel packets are carried over IPv4. 2266 7. Acknowledgments 2268 A number of people have contributed text and ideas to this 2269 specification. The authors are particularly indebted to Matthieu 2270 Boutier, Gwendoline Chouasne, Margaret Cullen, Donald Eastlake and 2271 Toke Hoiland-Jorgensen. Earlier versions of this document greatly 2272 benefited from the input of Joel Halpern. The address compression 2273 technique was inspired by [PACKETBB]. 2275 8. References 2277 8.1. Normative References 2279 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2280 Requirement Levels", BCP 14, RFC 2119, 2281 DOI 10.17487/RFC2119, March 1997. 2283 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 2284 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 2285 May 2008. 2287 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2288 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 2289 May 2017. 2291 8.2. Informative References 2293 [DSDV] Perkins, C. and P. Bhagwat, "Highly Dynamic Destination- 2294 Sequenced Distance-Vector Routing (DSDV) for Mobile 2295 Computers", ACM SIGCOMM'94 Conference on Communications 2296 Architectures, Protocols and Applications 234-244, 1994. 2298 [DUAL] Garcia Luna Aceves, J., "Loop-Free Routing Using Diffusing 2299 Computations", IEEE/ACM Transactions on Networking 1:1, 2300 February 1993. 2302 [EIGRP] Albrightson, B., Garcia Luna Aceves, J., and J. Boyle, 2303 "EIGRP -- a Fast Routing Protocol Based on Distance 2304 Vectors", Proc. Interop 94, 1994. 2306 [ETX] De Couto, D., Aguayo, D., Bicket, J., and R. Morris, "A 2307 high-throughput path metric for multi-hop wireless 2308 networks", Proc. MobiCom 2003, 2003. 2310 [IS-IS] "Information technology -- Telecommunications and 2311 information exchange between systems -- Intermediate 2312 System to Intermediate System intra-domain routeing 2313 information exchange protocol for use in conjunction with 2314 the protocol for providing the connectionless-mode network 2315 service (ISO 8473)", ISO/IEC 10589:2002, 2002. 2317 [JITTER] Floyd, S. and V. Jacobson, "The synchronization of 2318 periodic routing messages", IEEE/ACM Transactions on 2319 Networking 2, 2, 122-136, April 1994. 2321 [OSPF] Moy, J., "OSPF Version 2", RFC 2328, April 1998. 2323 [PACKETBB] 2324 Clausen, T., Dearlove, C., Dean, J., and C. Adjih, 2325 "Generalized Mobile Ad Hoc Network (MANET) Packet/Message 2326 Format", RFC 5444, February 2009. 2328 [RIP] Malkin, G., "RIP Version 2", RFC 2453, November 1998. 2330 Appendix A. Cost and Metric Computation 2332 The strategy for computing link costs and route metrics is a local 2333 matter; Babel itself only requires that it comply with the conditions 2334 given in Section 3.4.3 and Section 3.5.2. Different nodes may use 2335 different strategies in a single network and may use different 2336 strategies on different interface types. This section describes the 2337 strategies used by the sample implementation of Babel. 2339 The sample implementation of Babel sends periodic Multicast Hellos, 2340 and never sends Unicast Hellos. It maintains statistics about the 2341 last 16 received Hello TLVs of each kind (Appendix A.1), computes 2342 costs by using the 2-out-of-3 strategy (Appendix A.2.1) on wired 2343 links, and ETX (Appendix A.2.2) on wireless links. It uses an 2344 additive algebra for metric computation (Appendix A.3.1). 2346 A.1. Maintaining Hello History 2348 For each neighbour, the sample implementation of Babel maintains two 2349 sets of Hello history, one for each kind of Hello, and an expected 2350 sequence number, one for Multicast and one for Unicast Hellos. Each 2351 Hello history is a vector of 16 bits, where a 1 value represents a 2352 received Hello, and a 0 value a missed Hello. For each kind of 2353 Hello, the expected sequence number, written ne, is the sequence 2354 number that is expected to be carried by the next received Hello from 2355 this neighbour. 2357 Whenever it receives a Hello packet of a given kind from a neighbour, 2358 a node compares the received sequence number nr for that kind of 2359 Hello with its expected sequence number ne. Depending on the outcome 2360 of this comparison, one of the following actions is taken: 2362 o if the two differ by more than 16 (modulo 2^16), then the sending 2363 node has probably rebooted and lost its sequence number; the whole 2364 associated neighbour table entry is flushed and a new one is 2365 created; 2367 o otherwise, if the received nr is smaller (modulo 2^16) than the 2368 expected sequence number ne, then the sending node has increased 2369 its Hello interval without us noticing; the receiving node removes 2370 the last (ne - nr) entries from this neighbour's Hello history (we 2371 "undo history"); 2373 o otherwise, if nr is larger (modulo 2^16) than ne, then the sending 2374 node has decreased its Hello interval, and some Hellos were lost; 2375 the receiving node adds (nr - ne) 0 bits to the Hello history (we 2376 "fast-forward"). 2378 The receiving node then appends a 1 bit to the Hello history and sets 2379 ne to (nr + 1). If the Interval field of the received Hello is not 2380 zero, it resets the neighbour's hello timer to 1.5 times the 2381 advertised Interval (the extra margin allows for delay due to 2382 jitter). 2384 Whenever either Hello timer associated to a neighbour expires, the 2385 local node adds a 0 bit to this neighbour's Hello history, and 2386 increments the expected Hello number. If both Hello histories are 2387 empty (they contain 0 bits only), the neighbour entry is flushed; 2388 otherwise, the relevant hello timer is reset to the value advertised 2389 in the last Hello of that kind received from this neighbour (no extra 2390 margin is necessary in this case, since jitter was already taken into 2391 account when computing the timeout that has just expired). 2393 A.2. Cost Computation 2395 This section discusses how to compute costs based on Hello history. 2397 A.2.1. k-out-of-j 2399 K-out-of-j link sensing is suitable for wired links that are either 2400 up, in which case they only occasionally drop a packet, or down, in 2401 which case they drop all packets. 2403 The k-out-of-j strategy is parameterised by two small integers k and 2404 j, such that 0 < k <= j, and the nominal link cost, a constant K >= 2405 1. A node keeps a history of the last j hellos; if k or more of 2406 those have been correctly received, the link is assumed to be up, and 2407 the rxcost is set to K; otherwise, the link is assumed to be down, 2408 and the rxcost is set to infinity. 2410 Since Babel supports two kinds of Hellos, a Babel node performs k- 2411 out-of-j twice for each neighbour, once on the Unicast and once on 2412 the Multicast Hello history. If either of the instances of k-out- 2413 of-j indicates that the link is up, then the link is assumed to be 2414 up, and the rxcost is set to K; if both instances indicate that the 2415 link is down, then the link is assumed to be down, and the rxcost is 2416 set to infinity. In other words, the resulting rxcost is the minimum 2417 of the rxcosts yielded by the two instances of k-out-of-j link 2418 sensing. 2420 The cost of a link performing k-out-of-j link sensing is defined as 2421 follows: 2423 o cost = FFFF hexadecimal if rxcost = FFFF hexadecimal; 2425 o cost = txcost otherwise. 2427 A.2.2. ETX 2429 Unlike wired links, which are bimodal (either up or down), wireless 2430 links exhibit continuous variation of the link quality. Naive 2431 application of hop-count routing in networks that use wireless links 2432 for transit tends to select long, lossy links in preference to 2433 shorter, lossless links, which can dramatically reduce throughput. 2434 For that reason, a routing protocol designed to support wireless 2435 links must perform some form of link-quality estimation. 2437 ETX [ETX] is a simple link-quality estimation algorithm that is 2438 designed to work well with the IEEE 802.11 MAC. By default, the 2439 IEEE 802.11 MAC performs ARQ and rate adaptation on unicast frames, 2440 but not on multicast frames, which are sent at a fixed rate with no 2441 ARQ; therefore, measuring the loss rate of multicast frames yields a 2442 useful estimate of a link's quality. 2444 A node performing ETX link quality estimation uses a neighbour's 2445 Multicast Hello history to compute an estimate, written beta, of the 2446 probability that a Hello TLV is successfully received. Beta can be 2447 computed as the fraction of 1 bits within a small number (say, 6) of 2448 the most recent entries in the Multicast Hello history, or it can be 2449 an exponential average, or some combination of both approaches. 2451 Let alpha be MIN(1, 256/txcost), an estimate of the probability of 2452 successfully sending a Hello TLV. The cost is then computed by 2454 cost = 256/(alpha * beta) 2456 or, equivalently, 2458 cost = (MAX(txcost, 256) * rxcost) / 256. 2460 Since the IEEE 802.11 MAC performs ARQ on unicast frames, unicast 2461 frames do not provide a useful measure of link quality, and therefore 2462 ETX ignores the Unicast Hello history. Thus, a node performing ETX 2463 link-quality estimation will not route through neighbouring nodes 2464 unless they send periodic Multicast Hellos (possibly in addition to 2465 Unicast Hellos). 2467 A.3. Metric Computation 2469 As described in Section 3.5.2, the metric advertised by a neighbour 2470 is combined with the link cost to yield a metric. 2472 A.3.1. Additive Metrics 2474 The simplest approach for obtaining a monotonic, left-distributive 2475 metric is to define the metric of a route as the sum of the costs of 2476 the component links. More formally, if a neighbour advertises a 2477 route with metric m over a link with cost c, then the resulting route 2478 has metric M(c, m) = c + m. 2480 A multiplicative metric can be converted into an additive one by 2481 taking the logarithm (in some suitable base) of the link costs. 2483 A.3.2. External Sources of Willingness 2485 A node may want to vary its willingness to forward packets by taking 2486 into account information that is external to the Babel protocol, such 2487 as the monetary cost of a link, the node's battery status, CPU load, 2488 etc. This can be done by adding to every route's metric a value k 2489 that depends on the external data. For example, if a battery-powered 2490 node receives an update with metric m over a link with cost c, it 2491 might compute a metric M(c, m) = k + c + m, where k depends on the 2492 battery status. 2494 In order to preserve strict monotonicity (Section 3.5.2), the value k 2495 must be greater than -c. 2497 Appendix B. Constants 2499 The choice of time constants is a trade-off between fast detection of 2500 mobility events and protocol overhead. Two implementations of Babel 2501 with different time constants will interoperate, although the 2502 resulting convergence time will most likely be dictated by the slower 2503 of the two. 2505 Experience with the sample implementation of Babel indicates that the 2506 Hello interval is the most important time constant: a mobility event 2507 is detected within 1.5 to 3 Hello intervals. Due to Babel's reliance 2508 on triggered updates and explicit requests, the Update interval only 2509 has an effect on the time it takes for accurate metrics to be 2510 propagated after variations in link costs too small to trigger an 2511 unscheduled update or in the presence of packet loss. 2513 At the time of writing, the sample implementation of Babel uses the 2514 following default values: 2516 Multicast Hello Interval: 4 seconds. 2518 IHU Interval: the advertised IHU interval is always 3 times the 2519 Multicast Hello interval. IHUs are actually sent with each Hello 2520 on lossy links (as determined from the Hello history), but only 2521 with every third Multicast Hello on lossless links. 2523 Unicast Hello Interval: the sample implementation never sends 2524 scheduled Unicast Hellos; 2526 Update Interval: 4 times the Multicast Hello interval. 2528 IHU Hold Time: 3.5 times the advertised IHU interval. 2530 Route Expiry Time: 3.5 times the advertised update interval. 2532 Source GC time: 3 minutes. 2534 Request timeout: initially 2 seconds, doubled every time a request 2535 is resent, up to a maximum of three times. 2537 The amount of jitter applied to a packet depends on whether it 2538 contains any urgent TLVs or not (Section 3.1). Urgent triggered 2539 updates and urgent requests are delayed by no more than 200ms; 2540 acknowledgments, by no more than the associated deadline; and other 2541 TLVs by no more than one-half the Multicast Hello interval. 2543 Appendix C. Considerations for protocol extensions 2545 Babel is an extensible protocol, and this document defines a number 2546 of mechanisms that can be used to extend the protocol in a backwards 2547 compatible manner: 2549 o increasing the version number in the packet header; 2551 o defining new TLVs; 2553 o defining new sub-TLVs (with or without the mandatory bit set); 2555 o defining new AEs; 2557 o using the packet trailer. 2559 This appendix is intended to guide designers of protocol extensions 2560 in chosing a particular encoding. 2562 The version number in the Babel header should only be increased if 2563 the new version is not backwards compatible with the original 2564 protocol. 2566 In many cases, an extension could be implemented either by defining a 2567 new TLV, or by adding a new sub-TLV to an existing TLV. For example, 2568 an extension whose purpose is to attach additional data to route 2569 updates can be implemented either by creating a new "enriched" Update 2570 TLV, by adding a non-mandatory sub-TLV to the Update TLV, or by 2571 adding a mandatory sub-TLV. 2573 The various encodings are treated differently by implementations that 2574 do not understand the extension. In the case of a new TLV or of a 2575 sub-TLV with the mandatory bit set, the whole TLV is ignored by 2576 implementations that do not implement the extension, while in the 2577 case of a non-mandatory sub-TLV, the TLV is parsed and acted upon, 2578 and only the unknown sub-TLV is silently ignored. Therefore, a non- 2579 mandatory sub-TLV should be used by extensions that extend the Update 2580 in a compatible manner (the extension data may be silently ignored), 2581 while a mandatory sub-TLV or a new TLV must be used by extensions 2582 that make incompatible extensions to the meaning of the TLV (the 2583 whole TLV must be thrown away if the extension data is not 2584 understood). 2586 Experience shows that the need for additional data tends to crop up 2587 in the most unexpected places. Hence, it is recommended that 2588 extensions that define new TLVs should make them self-terminating, 2589 and allow attaching sub-TLVs to them. 2591 Adding a new AE is essentially equivalent to adding a new TLV: Update 2592 TLVs with an unknown AE are ignored, just like unknown TLVs. 2593 However, adding a new AE is more involved than adding a new TLV, 2594 since it creates a new set of compression state. Additionally, since 2595 the Next Hop TLV creates state specific to a given address family, as 2596 opposed to a given AE, a new AE for a previously defined address 2597 family must not be used in the Next Hop TLV if backwards 2598 compatibility is required. A similar issue arises with Update TLVs 2599 with unknown AEs establishing a new router-id (due to the Router-Id 2600 flag being set). Therefore, defining new AEs must be done with care 2601 if compatibility with unextended implementations is required. 2603 The packet trailer (the space after the declared length of the packet 2604 but within the payload of the UDP datagram) was originally intended 2605 to carry a cryptographic signature. However, no extension has used 2606 it to date, and therefore we refrain from making any recommendations 2607 about its use due to the lack of implementation experience. 2609 Appendix D. Stub Implementations 2611 Babel is a fairly economic protocol. Updates take between 12 and 40 2612 octets per destination, depending on the address family and how 2613 successful compression is; in a double-stack flat network, an average 2614 of less than 24 octets per update is typical. The route table 2615 occupies about 35 octets per IPv6 entry. To put these values into 2616 perspective, a single full-size Ethernet frame can carry some 65 2617 route updates, and a megabyte of memory can contain a 20000-entry 2618 route table and the associated source table. 2620 Babel is also a reasonably simple protocol. The sample 2621 implementation consists of less than 12 000 lines of C code, and it 2622 compiles to less than 120 kB of text on a 32-bit CISC architecture; 2623 about half of this figure is due to protocol extensions and user- 2624 interface code. 2626 Nonetheless, in some very constrained environments, such as PDAs, 2627 microwave ovens, or abacuses, it may be desirable to have subset 2628 implementations of the protocol. 2630 There are many different definitions of a stub router, but for the 2631 needs of this section a stub implementation of Babel is one that 2632 announces one or more directly attached prefixes into a Babel network 2633 but doesn't reannounce any routes that it has learnt from its 2634 neighbours. It may either maintain a full routing table, or simply 2635 select a default gateway amongst any one of its neighbours that 2636 announces a default route. Since a stub implementation never 2637 forwards packets except from or to directly attached links, it cannot 2638 possibly participate in a routing loop, and hence it need not 2639 evaluate the feasibility condition or maintain a source table. 2641 No matter how primitive, a stub implementation MUST parse sub-TLVs 2642 attached to any TLVs that it understands and check the mandatory bit. 2643 It MUST answer acknowledgment requests and MUST participate in the 2644 Hello/IHU protocol. It MUST also be able to reply to seqno requests 2645 for routes that it announces and SHOULD be able to reply to route 2646 requests. 2648 Experience shows that an IPv6-only stub implementation of Babel can 2649 be written in less than 1000 lines of C code and compile to 13 kB of 2650 text on 32-bit CISC architecture. 2652 Appendix E. Software Availability 2654 The sample implementation of Babel is available from 2655 . 2657 Appendix F. Changes from previous versions 2659 F.1. Changes since RFC 6126 2661 o Changed UDP port number to 6696. 2663 o Consistently use router-id rather than id. 2665 o Clarified that the source garbage collection timer is reset after 2666 sending an update even if the entry was not modified. 2668 o In section "Seqno Requests", fixed an erroneous "route request". 2670 o In the description of the Seqno Request TLV, added the description 2671 of the Router-Id field. 2673 o Made router-ids all-0 and all-1 forbidden. 2675 F.2. Changes since draft-ietf-babel-rfc6126bis-00 2677 o Added security considerations. 2679 F.3. Changes since draft-ietf-babel-rfc6126bis-01 2681 o Integrated the format of sub-TLVs. 2683 o Mentioned for each TLV whether it supports sub-TLVs. 2685 o Added Appendix C. 2687 o Added a mandatory bit in sub-TLVs. 2689 o Changed compression state to be per-AF rather than per-AE. 2691 o Added implementation hint for the routing table. 2693 o Clarified how router-ids are computed when bit 0x40 is set in 2694 Updates. 2696 o Relaxed the conditions for sending requests, and tightened the 2697 conditions for forwarding requests. 2699 o Clarified that neighbours should be acquired at some point, but it 2700 doesn't matter when. 2702 F.4. Changes since draft-ietf-babel-rfc6126bis-02 2704 o Added Unicast Hellos. 2706 o Added unscheduled (interval-less) Hellos. 2708 o Changed Appendix A to consider Unicast and unscheduled Hellos. 2710 o Changed Appendix B to agree with the reference implementation. 2712 o Added optional algorithm to avoid the hold time. 2714 o Changed the table of pending seqno requests to be indexed by 2715 router-id in addition to prefixes. 2717 o Relaxed the route acquisition algorithm. 2719 o Replaced minimal implementations by stub implementations. 2721 o Added acknowledgments section. 2723 F.5. Changes since draft-ietf-babel-rfc6126bis-03 2725 o Clarified that all the data structures are conceptual. 2727 o Made sending and receiving Multicast Hellos a SHOULD, avoids 2728 expressing any opinion about Unicast Hellos. 2730 o Removed opinion about Multicast vs. Unicast Hellos (Appendix A.4). 2732 o Made hold-time into a SHOULD rather than MUST. 2734 o Clarified that Seqno Requests are for a finite-metric Update. 2736 o Clarified that sub-TLVs Pad1 and PadN are allowed within any TLV 2737 that allows sub-TLVs. 2739 o Updated IANA Considerations. 2741 o Updated Security Considerations. 2743 o Renamed routing table back to route table. 2745 o Made buffering outgoing updates a SHOULD. 2747 o Weakened advice to use modified EUI-64 in router-ids. 2749 o Added information about sending requests to Appendix B. 2751 o A number of minor wording changes and clarifications. 2753 F.6. Changes since draft-ietf-babel-rfc6126bis-03 2755 Minor editorial changes. 2757 F.7. Changes since draft-ietf-babel-rfc6126bis-04 2759 o Renamed isotonicity to left-distributivity. 2761 o Minor clarifications to unicast hellos. 2763 o Updated requirements boilerplate to RFC 8174. 2765 o Minor editorial changes. 2767 Authors' Addresses 2769 Juliusz Chroboczek 2770 IRIF, University of Paris-Diderot 2771 Case 7014 2772 75205 Paris Cedex 13 2773 France 2775 Email: jch@irif.fr 2777 David Schinazi 2778 Apple Inc. 2779 1 Infinite Loop 2780 Cupertino, California 95014 2781 US 2783 Email: dschinazi@apple.com