idnits 2.17.1 draft-ietf-ipv6-rfc2462bis-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == The document seems to lack the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. (The document does seem to have the reference to RFC 2119 which the ID-Checklist requires). -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (June 17, 2004) is 7250 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 2402 (ref. '1') (Obsoleted by RFC 4302, RFC 4305) ** Obsolete normative reference: RFC 3513 (ref. '4') (Obsoleted by RFC 4291) ** Obsolete normative reference: RFC 2461 (ref. '5') (Obsoleted by RFC 4861) -- Obsolete informational reference (is this intentional?): RFC 3315 (ref. '6') (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 3736 (ref. '7') (Obsoleted by RFC 8415) -- Obsolete informational reference (is this intentional?): RFC 3484 (ref. '8') (Obsoleted by RFC 6724) -- Obsolete informational reference (is this intentional?): RFC 3041 (ref. '9') (Obsoleted by RFC 4941) Summary: 4 errors (**), 0 flaws (~~), 3 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 IETF IPv6 Working Group S. Thomson 2 Internet-Draft Cisco 3 Expires: December 16, 2004 T. Narten 4 IBM 5 T. Jinmei 6 Toshiba 7 June 17, 2004 9 IPv6 Stateless Address Autoconfiguration 10 draft-ietf-ipv6-rfc2462bis-02.txt 12 Status of this Memo 14 This document is an Internet-Draft and is in full conformance with 15 all provisions of Section 10 of RFC2026. 17 Internet-Drafts are working documents of the Internet Engineering 18 Task Force (IETF), its areas, and its working groups. Note that other 19 groups may also distribute working documents as Internet-Drafts. 21 Internet-Drafts are draft documents valid for a maximum of six months 22 and may be updated, replaced, or obsoleted by other documents at any 23 time. It is inappropriate to use Internet-Drafts as reference 24 material or to cite them other than as "work in progress." 26 The list of current Internet-Drafts can be accessed at http:// 27 www.ietf.org/ietf/1id-abstracts.txt. 29 The list of Internet-Draft Shadow Directories can be accessed at 30 http://www.ietf.org/shadow.html. 32 This Internet-Draft will expire on December 16, 2004. 34 Copyright Notice 36 Copyright (C) The Internet Society (2004). All Rights Reserved. 38 Abstract 40 This document specifies the steps a host takes in deciding how to 41 autoconfigure its interfaces in IP version 6. The autoconfiguration 42 process includes creating a link-local address and verifying its 43 uniqueness on a link, determining what information can be 44 autoconfigured (addresses, other information, or both), and in the 45 case of addresses, whether they can be obtained through the stateless 46 mechanism, the stateful mechanism, or both. This document defines the 47 process for generating a link-local address, the process for 48 generating global addresses via stateless address autoconfiguration, 49 and the Duplicate Address Detection procedure. The details of 50 autoconfiguration using the stateful protocol is specified in RFC 51 3315 and RFC 3736. 53 Table of Contents 55 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 56 2. TERMINOLOGY . . . . . . . . . . . . . . . . . . . . . . . . 5 57 2.1 Requirements . . . . . . . . . . . . . . . . . . . . . . . . 7 58 3. DESIGN GOALS . . . . . . . . . . . . . . . . . . . . . . . . 7 59 4. PROTOCOL OVERVIEW . . . . . . . . . . . . . . . . . . . . . 8 60 4.1 Site Renumbering . . . . . . . . . . . . . . . . . . . . . . 10 61 5. PROTOCOL SPECIFICATION . . . . . . . . . . . . . . . . . . . 11 62 5.1 Node Configuration Variables . . . . . . . . . . . . . . . . 11 63 5.2 Autoconfiguration-Related Structures . . . . . . . . . . . . 12 64 5.3 Creation of Link-Local Addresses . . . . . . . . . . . . . . 12 65 5.4 Duplicate Address Detection . . . . . . . . . . . . . . . . 13 66 5.4.1 Message Validation . . . . . . . . . . . . . . . . . . . . . 14 67 5.4.2 Sending Neighbor Solicitation Messages . . . . . . . . . . . 14 68 5.4.3 Receiving Neighbor Solicitation Messages . . . . . . . . . . 15 69 5.4.4 Receiving Neighbor Advertisement Messages . . . . . . . . . 16 70 5.4.5 When Duplicate Address Detection Fails . . . . . . . . . . . 17 71 5.5 Creation of Global Addresses . . . . . . . . . . . . . . . . 17 72 5.5.1 Soliciting Router Advertisements . . . . . . . . . . . . . . 17 73 5.5.2 Absence of Router Advertisements . . . . . . . . . . . . . . 17 74 5.5.3 Router Advertisement Processing . . . . . . . . . . . . . . 18 75 5.5.4 Address Lifetime Expiry . . . . . . . . . . . . . . . . . . 20 76 5.6 Configuration Consistency . . . . . . . . . . . . . . . . . 21 77 5.7 Retaining Configured Addresses for Stability . . . . . . . . 21 78 6. SECURITY CONSIDERATIONS . . . . . . . . . . . . . . . . . . 21 79 7. IANA CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . 22 80 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 22 81 Normative References . . . . . . . . . . . . . . . . . . . . 22 82 Informative References . . . . . . . . . . . . . . . . . . . 23 83 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . 23 84 A. LOOPBACK SUPPRESSION & DUPLICATE ADDRESS DETECTION . . . . . 24 85 B. CHANGES SINCE RFC 1971 . . . . . . . . . . . . . . . . . . . 25 86 C. CHANGE HISTORY . . . . . . . . . . . . . . . . . . . . . . . 26 87 Intellectual Property and Copyright Statements . . . . . . . 29 89 1. Introduction 91 This document specifies the steps a host takes in deciding how to 92 autoconfigure its interfaces in IP version 6. The autoconfiguration 93 process includes creating a link-local address and verifying its 94 uniqueness on a link, determining what information can be 95 autoconfigured (addresses, other information, or both), and in the 96 case of addresses, whether they can be obtained through the stateless 97 mechanism, the stateful mechanism, or both. This document defines the 98 process for generating a link-local address, the process for 99 generating global addresses via stateless address autoconfiguration, 100 and the Duplicate Address Detection procedure. The details of 101 autoconfiguration using the stateful protocol is specified in RFC 102 3315 [6] and RFC 3736 [7]. 104 IPv6 defines both a stateful and stateless address autoconfiguration 105 mechanism. Stateless autoconfiguration requires no manual 106 configuration of hosts, minimal (if any) configuration of routers, 107 and no additional servers. The stateless mechanism allows a host to 108 generate its own addresses using a combination of locally available 109 information and information advertised by routers. Routers advertise 110 prefixes that identify the subnet(s) associated with a link, while 111 hosts generate an "interface identifier" that uniquely identifies an 112 interface on a subnet. An address is formed by combining the two. In 113 the absence of routers, a host can only generate link-local 114 addresses. However, link-local addresses are sufficient for allowing 115 communication among nodes attached to the same link. 117 In the stateful autoconfiguration model, hosts obtain interface 118 addresses and/or configuration information and parameters from a 119 DHCPv6 server. Servers maintain a database that keeps track of which 120 addresses have been assigned to which hosts. The stateful 121 autoconfiguration protocol allows hosts to obtain addresses, other 122 configuration information or both from a server. Stateless and 123 stateful autoconfiguration complement each other. For example, a host 124 can use stateless autoconfiguration to configure its own addresses, 125 but use stateful autoconfiguration to obtain other information. 127 To obtain other configuration information without configuring 128 addresses in the stateful autoconfiguration model, a subset of DHCPv6 129 will be used [7]. While the model is called "stateful" here in order 130 to highlight the contrast to the stateless protocol defined in this 131 document, the intended protocol is also defined to work in a 132 stateless fashion. This is based on a result, through operational 133 experiments, that all known "other" configuration information can be 134 managed by a stateless server, that is, a server that does not 135 maintain state of each client that the server provides with the 136 configuration information. 138 The stateless approach is used when a site is not particularly 139 concerned with the exact addresses hosts use, so long as they are 140 unique and properly routable. The stateful approach is used when a 141 site requires tighter control over exact address assignments. Both 142 stateful and stateless address autoconfiguration may be used 143 simultaneously. The site administrator specifies which type of 144 autoconfiguration is available through the setting of appropriate 145 fields in Router Advertisement messages [5]. 147 IPv6 addresses are leased to an interface for a fixed (possibly 148 infinite) length of time. Each address has an associated lifetime 149 that indicates how long the address is bound to an interface. When a 150 lifetime expires, the binding (and address) become invalid and the 151 address may be reassigned to another interface elsewhere in the 152 Internet. To handle the expiration of address bindings gracefully, an 153 address goes through two distinct phases while assigned to an 154 interface. Initially, an address is "preferred", meaning that its use 155 in arbitrary communication is unrestricted. Later, an address becomes 156 "deprecated" in anticipation that its current interface binding will 157 become invalid. While in a deprecated state, the use of an address is 158 discouraged, but not strictly forbidden. New communication (e.g., 159 the opening of a new TCP connection) should use a preferred address 160 when possible. A deprecated address should be used only by 161 applications that have been using it and would have difficulty 162 switching to another address without a service disruption. 164 To ensure that all configured addresses are likely to be unique on a 165 given link, nodes run a "duplicate address detection" algorithm on 166 addresses before assigning them to an interface. The Duplicate 167 Address Detection algorithm is performed on all addresses, 168 independent of whether they are obtained via stateless or stateful 169 autoconfiguration. This document defines the Duplicate Address 170 Detection algorithm. 172 The autoconfiguration process specified in this document applies only 173 to hosts and not routers. Since host autoconfiguration uses 174 information advertised by routers, routers will need to be configured 175 by some other means. However, it is expected that routers will 176 generate link-local addresses using the mechanism described in this 177 document. In addition, routers are expected to successfully pass the 178 Duplicate Address Detection procedure described in this document on 179 all addresses prior to assigning them to an interface. 181 Section 2 provides definitions for terminology used throughout this 182 document. Section 3 describes the design goals that lead to the 183 current autoconfiguration procedure. Section 4 provides an overview 184 of the protocol, while Section 5 describes the protocol in detail. 186 2. TERMINOLOGY 188 IP - Internet Protocol Version 6. The terms IPv4 and IPv6 are used 189 only in contexts where necessary to avoid ambiguity. 191 node - a device that implements IP. 193 router - a node that forwards IP packets not explicitly addressed to 194 itself. 196 host - any node that is not a router. 198 upper layer - a protocol layer immediately above IP. Examples are 199 transport protocols such as TCP and UDP, control protocols such as 200 ICMP, routing protocols such as OSPF, and internet or lower-layer 201 protocols being "tunneled" over (i.e., encapsulated in) IP such as 202 IPX, AppleTalk, or IP itself. 204 link - a communication facility or medium over which nodes can 205 communicate at the link layer, i.e., the layer immediately below 206 IP. Examples are Ethernets (simple or bridged); PPP links; X.25, 207 Frame Relay, or ATM networks; and internet (or higher) layer 208 "tunnels", such as tunnels over IPv4 or IPv6 itself. 210 interface - a node's attachment to a link. 212 packet - an IP header plus payload. 214 address - an IP-layer identifier for an interface or a set of 215 interfaces. 217 unicast address - an identifier for a single interface. A packet sent 218 to a unicast address is delivered to the interface identified by 219 that address. 221 multicast address - an identifier for a set of interfaces (typically 222 belonging to different nodes). A packet sent to a multicast 223 address is delivered to all interfaces identified by that address. 225 anycast address - an identifier for a set of interfaces (typically 226 belonging to different nodes). A packet sent to an anycast 227 address is delivered to one of the interfaces identified by that 228 address (the "nearest" one, according to the routing protocol's 229 measure of distance). See the IPv6 addressing architecture [4]. 231 solicited-node multicast address - a multicast address to which 232 Neighbor Solicitation messages are sent. The algorithm for 233 computing the address is given in RFC 2461 [5]. 235 link-layer address - a link-layer identifier for an interface. 236 Examples include IEEE 802 addresses for Ethernet links and E.164 237 addresses for ISDN links. 239 link-local address - an address having link-only scope that can be 240 used to reach neighboring nodes attached to the same link. All 241 interfaces have a link-local unicast address. 243 global address - an address with unlimited scope. 245 communication - any packet exchange among nodes that requires that 246 the address of each node used in the exchange remain the same for 247 the duration of the packet exchange. Examples are a TCP 248 connection or a UDP request-response. 250 tentative address - an address whose uniqueness on a link is being 251 verified, prior to its assignment to an interface. A tentative 252 address is not considered assigned to an interface in the usual 253 sense. An interface discards received packets addressed to a 254 tentative address, but accepts Neighbor Discovery packets related 255 to Duplicate Address Detection for the tentative address. 257 preferred address - an address assigned to an interface whose use by 258 upper layer protocols is unrestricted. Preferred addresses may be 259 used as the source (or destination) address of packets sent from 260 (or to) the interface. 262 deprecated address - An address assigned to an interface whose use is 263 discouraged, but not forbidden. A deprecated address should no 264 longer be used as a source address in new communications, but 265 packets sent from or to deprecated addresses are delivered as 266 expected. A deprecated address may continue to be used as a 267 source address in communications where switching to a preferred 268 address causes hardship to a specific upper-layer activity (e.g., 269 an existing TCP connection). 271 valid address - a preferred or deprecated address. A valid address 272 may appear as the source or destination address of a packet, and 273 the internet routing system is expected to deliver packets sent to 274 a valid address to their intended recipients. 276 invalid address - an address that is not assigned to any interface. A 277 valid address becomes invalid when its valid lifetime expires. 278 Invalid addresses should not appear as the destination or source 279 address of a packet. In the former case, the internet routing 280 system will be unable to deliver the packet, in the latter case 281 the recipient of the packet will be unable to respond to it. 283 preferred lifetime - the length of time that a valid address is 284 preferred (i.e., the time until deprecation). When the preferred 285 lifetime expires, the address becomes deprecated. 287 valid lifetime - the length of time an address remains in the valid 288 state (i.e., the time until invalidation). The valid lifetime must 289 be greater than or equal to the preferred lifetime. When the 290 valid lifetime expires, the address becomes invalid. 292 interface identifier - a link-dependent identifier for an interface 293 that is (at least) unique per link [4]. Stateless address 294 autoconfiguration combines an interface identifier with a prefix 295 to form an address. From address autoconfiguration's perspective, 296 an interface identifier is a bit string of known length. The exact 297 length of an interface identifier and the way it is created is 298 defined in a separate link-type specific document that covers 299 issues related to the transmission of IP over a particular link 300 type (e.g., IPv6 over Ethernet [2]). Note that the address 301 architecture [4] also defines the length of the interface 302 identifiers for some set of addresses, but the two sets of 303 definitions must be consistent. In many cases, the identifier will 304 be derived from the interface's link-layer address. 306 2.1 Requirements 308 The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, 309 SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this 310 document, are to be interpreted as described in RFC 2119 [3]. 312 3. DESIGN GOALS 314 Stateless autoconfiguration is designed with the following goals in 315 mind: 317 o Manual configuration of individual machines before connecting them 318 to the network should not be required. Consequently, a mechanism 319 is needed that allows a host to obtain or create unique addresses 320 for each of its interfaces. Address autoconfiguration assumes that 321 each interface can provide a unique identifier for that interface 322 (i.e., an "interface identifier"). In the simplest case, an 323 interface identifier consists of the interface's link-layer 324 address. An interface identifier can be combined with a prefix to 325 form an address. 327 o Small sites consisting of a set of machines attached to a single 328 link should not require the presence of a stateful server or 329 router as a prerequisite for communicating. Plug-and-play 330 communication is achieved through the use of link-local addresses. 331 Link-local addresses have a well-known prefix that identifies the 332 (single) shared link to which a set of nodes attach. A host forms 333 a link-local address by appending its interface identifier to the 334 link-local prefix. 336 o A large site with multiple networks and routers should not require 337 the presence of a stateful address configuration server. In order 338 to generate global addresses, hosts must determine the prefixes 339 that identify the subnets to which they attach. Routers generate 340 periodic Router Advertisements that include options listing the 341 set of active prefixes on a link. 343 o Address configuration should facilitate the graceful renumbering 344 of a site's machines. For example, a site may wish to renumber all 345 of its nodes when it switches to a new network service provider. 346 Renumbering is achieved through the leasing of addresses to 347 interfaces and the assignment of multiple addresses to the same 348 interface. Lease lifetimes provide the mechanism through which a 349 site phases out old prefixes. The assignment of multiple 350 addresses to an interface provides for a transition period during 351 which both a new address and the one being phased out work 352 simultaneously. 354 o System administrators need the ability to specify whether 355 stateless autoconfiguration, stateful autoconfiguration, or both 356 are available. Router Advertisements include flags specifying 357 which mechanisms a host can use. 359 4. PROTOCOL OVERVIEW 361 This section provides an overview of the typical steps that take 362 place when an interface autoconfigures itself. Autoconfiguration is 363 performed only on multicast-capable links and begins when a 364 multicast-capable interface is enabled, e.g., during system startup. 365 Nodes (both hosts and routers) begin the autoconfiguration process by 366 generating a link-local address for the interface. A link-local 367 address is formed by appending the interface's identifier to the 368 well-known link-local prefix. 370 Before the link-local address can be assigned to an interface and 371 used, however, a node must attempt to verify that this "tentative" 372 address is not already in use by another node on the link. 373 Specifically, it sends a Neighbor Solicitation message containing the 374 tentative address as the target. If another node is already using 375 that address, it will return a Neighbor Advertisement saying so. If 376 another node is also attempting to use the same address, it will send 377 a Neighbor Solicitation for the target as well. The exact number of 378 times the Neighbor Solicitation is (re)transmitted and the delay time 379 between consecutive solicitations is link-specific and may be set by 380 system management. 382 If a node determines that its tentative link-local address is not 383 unique, autoconfiguration stops and manual configuration of the 384 interface is required. To simplify recovery in this case, it should 385 be possible for an administrator to supply an alternate interface 386 identifier that overrides the default identifier in such a way that 387 the autoconfiguration mechanism can then be applied using the new 388 (presumably unique) interface identifier. Alternatively, link-local 389 and other addresses will need to be configured manually. 391 Once a node ascertains that its tentative link-local address is 392 unique, it assigns the address to the interface. At this point, the 393 node has IP-level connectivity with neighboring nodes. The remaining 394 autoconfiguration steps are performed only by hosts; the 395 (auto)configuration of routers is beyond the scope of this document. 397 The next phase of autoconfiguration involves obtaining a Router 398 Advertisement or determining that no routers are present. If routers 399 are present, they will send Router Advertisements that specify what 400 sort of autoconfiguration a host can do. Note that stateful 401 autoconfiguration may still be available even if no routers are 402 present. 404 Routers send Router Advertisements periodically, but the delay 405 between successive advertisements will generally be longer than a 406 host performing autoconfiguration will want to wait [5]. To obtain an 407 advertisement quickly, a host sends one or more Router Solicitations 408 to the all-routers multicast group. Router Advertisements contain two 409 flags indicating what type of stateful autoconfiguration (if any) is 410 available. A "managed address configuration (M)" flag indicates 411 whether hosts can use stateful autoconfiguration [6] to obtain 412 addresses. An "other stateful configuration (O)" flag indicates 413 whether hosts can use stateful autoconfiguration [7] to obtain 414 additional information (excluding addresses). 416 The details of how a host may use the M flags, including any use of 417 the "on" and "off" transitions for this flag, to control the use of 418 the stateful protocol for address assignment will be described in a 419 separate document. Similarly, the details of how a host may use the O 420 flags, including any use of the "on" and "off" transitions for this 421 flag, to control the use of the stateful protocol for getting other 422 configuration information will be described in a separate document. 424 Router Advertisements also contain zero or more Prefix Information 425 options that contain information used by stateless address 426 autoconfiguration to generate global addresses. It should be noted 427 that the stateless and stateful address autoconfiguration fields in 428 Router Advertisements are processed independently of one another, and 429 a host may use both stateful and stateless address autoconfiguration 430 simultaneously. One Prefix Information option field, the "autonomous 431 address-configuration flag", indicates whether or not the option even 432 applies to stateless autoconfiguration. If it does, additional 433 option fields contain a subnet prefix together with lifetime values 434 indicating how long addresses created from the prefix remain 435 preferred and valid. 437 Because routers generate Router Advertisements periodically, hosts 438 will continually receive new advertisements. Hosts process the 439 information contained in each advertisement as described above, 440 adding to and refreshing information received in previous 441 advertisements. 443 For safety, all addresses must be tested for uniqueness prior to 444 their assignment to an interface. The test should individually be 445 performed on all addresses obtained manually, via stateless address 446 autoconfiguration, or via stateful address autoconfiguration. To 447 accommodate sites that believe the overhead of performing Duplicate 448 Address Detection outweighs its benefits, the use of Duplicate 449 Address Detection can be disabled through the administrative setting 450 of a per-interface configuration flag. 452 To speed the autoconfiguration process, a host may generate its 453 link-local address (and verify its uniqueness) in parallel with 454 waiting for a Router Advertisement. Because a router may delay 455 responding to a Router Solicitation for a few seconds, the total time 456 needed to complete autoconfiguration can be significantly longer if 457 the two steps are done serially. 459 4.1 Site Renumbering 461 Address leasing facilitates site renumbering by providing a mechanism 462 to time-out addresses assigned to interfaces in hosts. At present, 463 upper layer protocols such as TCP provide no support for changing 464 end-point addresses while a connection is open. If an end-point 465 address becomes invalid, existing connections break and all 466 communication to the invalid address fails. Even when applications 467 use UDP as a transport protocol, addresses must generally remain the 468 same during a packet exchange. 470 Dividing valid addresses into preferred and deprecated categories 471 provides a way of indicating to upper layers that a valid address may 472 become invalid shortly and that future communication using the 473 address will fail, should the address's valid lifetime expire before 474 communication ends. To avoid this scenario, higher layers should use 475 a preferred address (assuming one of sufficient scope exists) to 476 increase the likelihood that an address will remain valid for the 477 duration of the communication. It is up to system administrators to 478 set appropriate prefix lifetimes in order to minimize the impact of 479 failed communication when renumbering takes place. The deprecation 480 period should be long enough that most, if not all, communications 481 are using the new address at the time an address becomes invalid. 483 The IP layer is expected to provide a means for upper layers 484 (including applications) to select the most appropriate source 485 address given a particular destination and possibly other 486 constraints. An application may choose to select the source address 487 itself before starting a new communication or may leave the address 488 unspecified, in which case the upper networking layers will use the 489 mechanism provided by the IP layer to choose a suitable address on 490 the application's behalf. 492 Detailed address selection rules are beyond the scope of this 493 document. 495 5. PROTOCOL SPECIFICATION 497 Autoconfiguration is performed on a per-interface basis on 498 multicast-capable interfaces. For multihomed hosts, 499 autoconfiguration is performed independently on each interface. 500 Autoconfiguration applies primarily to hosts, with two exceptions. 501 Routers are expected to generate a link-local address using the 502 procedure outlined below. In addition, routers perform Duplicate 503 Address Detection on all addresses prior to assigning them to an 504 interface. 506 5.1 Node Configuration Variables 508 A node MUST allow the following autoconfiguration-related variable to 509 be configured by system management for each multicast interface: 511 DupAddrDetectTransmits 513 The number of consecutive Neighbor Solicitation messages sent 514 while performing Duplicate Address Detection on a tentative 515 address. A value of zero indicates that Duplicate Address 516 Detection is not performed on tentative addresses. A value of one 517 indicates a single transmission with no follow up retransmissions. 519 Default: 1, but may be overridden by a link-type specific value in 520 the document that covers issues related to the transmission of IP 521 over a particular link type (e.g., IPv6 over Ethernet [2]). 523 Autoconfiguration also assumes the presence of the variable 524 RetransTimer as defined in RFC 2461 [5]. For autoconfiguration 525 purposes, RetransTimer specifies the delay between consecutive 526 Neighbor Solicitation transmissions performed during Duplicate 527 Address Detection (if DupAddrDetectTransmits is greater than 1), 528 as well as the time a node waits after sending the last Neighbor 529 Solicitation before ending the Duplicate Address Detection 530 process. 532 5.2 Autoconfiguration-Related Structures 534 Beyond the formation of a link-local address and using Duplicate 535 Address Detection, how routers (auto)configure their interfaces is 536 beyond the scope of this document. 538 A host maintains a list of addresses together with their 539 corresponding lifetimes. The address list contains both 540 autoconfigured addresses and those configured manually. 542 5.3 Creation of Link-Local Addresses 544 A node forms a link-local address whenever an interface becomes 545 enabled. An interface may become enabled after any of the following 546 events: 548 - The interface is initialized at system startup time. 550 - The interface is reinitialized after a temporary interface failure 551 or after being temporarily disabled by system management. 553 - The interface attaches to a link for the first time. 555 - The interface becomes enabled by system management after having 556 been administratively disabled. 558 A link-local address is formed by prepending the well-known link- 559 local prefix FE80::0 [4] (of appropriate length) to the interface 560 identifier. If the interface identifier has a length of N bits, the 561 interface identifier replaces the right-most N zero bits of the 562 link-local prefix. If the interface identifier is more than 118 bits 563 in length, autoconfiguration fails and manual configuration is 564 required. The length of the interface identifier is defined in a 565 separate link-type specific document, which should also be consistent 566 with the address architecture [4] (see Section 2). These documents 567 will carefully define the length so that link-local addresses can be 568 autoconfigured on the link. 570 A link-local address has an infinite preferred and valid lifetime; it 571 is never timed out. 573 5.4 Duplicate Address Detection 575 Duplicate Address Detection is performed on unicast addresses prior 576 to assigning them to an interface whose DupAddrDetectTransmits 577 variable is greater than zero. Duplicate Address Detection MUST take 578 place on all unicast addresses, regardless of whether they are 579 obtained through stateful, stateless or manual configuration, with 580 the exception of the following cases: 582 IP - Duplicate Address Detection MUST NOT be performed on anycast 583 addresses. 585 IP - Each individual unicast address SHOULD be tested for uniqueness. 586 Note that there are implementations deployed that only perform 587 Duplicate Address Detection for the link-local address and skip 588 the test for the global address using the same interface 589 identifier as that of the link-local address. Whereas this 590 document does not invalidate such implementations, this kind of 591 "optimization" is NOT RECOMMENDED, and new implementations MUST 592 NOT do that optimization. This optimization came from the 593 assumption that all of an interface's addresses are generated from 594 the same identifier. However, the assumption does actually not 595 stand; new types of addresses have been introduced where the 596 interface identifiers are not necessarily the same for all unicast 597 addresses on a single interface [9] [10]. Requiring to perform 598 Duplicate Address Detection for all unicast addresses will make 599 the algorithm robust for the current and future such special 600 interface identifiers. 602 The procedure for detecting duplicate addresses uses Neighbor 603 Solicitation and Advertisement messages as described below. If a 604 duplicate address is discovered during the procedure, the address 605 cannot be assigned to the interface. If the address is derived from 606 an interface identifier, a new identifier will need to be assigned to 607 the interface, or all IP addresses for the interface will need to be 608 manually configured. Note that the method for detecting duplicates 609 is not completely reliable, and it is possible that duplicate 610 addresses will still exist (e.g., if the link was partitioned while 611 Duplicate Address Detection was performed). 613 An address on which the Duplicate Address Detection Procedure is 614 applied is said to be tentative until the procedure has completed 615 successfully. A tentative address is not considered "assigned to an 616 interface" in the traditional sense. That is, the interface must 617 accept Neighbor Solicitation and Advertisement messages containing 618 the tentative address in the Target Address field, but processes such 619 packets differently from those whose Target Address matches an 620 address assigned to the interface. Other packets addressed to the 621 tentative address should be silently discarded. Note that the "other 622 packets" include Neighbor Solicitation and Advertisement messages to 623 the tentative address containing the tentative address in the Target 624 Address field. Such a case should not happen in normal operation, 625 though, since these messages are multicasted in the Duplicate Address 626 Detection Procedure. 628 It should also be noted that Duplicate Address Detection must be 629 performed prior to assigning an address to an interface in order to 630 prevent multiple nodes from using the same address simultaneously. If 631 a node begins using an address in parallel with Duplicate Address 632 Detection, and another node is already using the address, the node 633 performing Duplicate Address Detection will erroneously process 634 traffic intended for the other node, resulting in such possible 635 negative consequences as the resetting of open TCP connections. 637 The following subsections describe specific tests a node performs to 638 verify an address's uniqueness. An address is considered unique if 639 none of the tests indicate the presence of a duplicate address within 640 RetransTimer milliseconds after having sent DupAddrDetectTransmits 641 Neighbor Solicitations. Once an address is determined to be unique, 642 it may be assigned to an interface. 644 5.4.1 Message Validation 646 A node MUST silently discard any Neighbor Solicitation or 647 Advertisement message that does not pass the validity checks 648 specified in RFC 2461 [5]. A Neighbor Solicitation or Advertisement 649 message that passes these validity checks is called a valid 650 solicitation or valid advertisement, respectively. 652 5.4.2 Sending Neighbor Solicitation Messages 654 Before sending a Neighbor Solicitation, an interface MUST join the 655 all-nodes multicast address and the solicited-node multicast address 656 of the tentative address. The former ensures that the node receives 657 Neighbor Advertisements from other nodes already using the address; 658 the latter ensures that two nodes attempting to use the same address 659 simultaneously detect each other's presence. 661 To check an address, a node sends DupAddrDetectTransmits Neighbor 662 Solicitations, each separated by RetransTimer milliseconds. The 663 solicitation's Target Address is set to the address being checked, 664 the IP source is set to the unspecified address and the IP 665 destination is set to the solicited-node multicast address of the 666 target address. 668 If the Neighbor Solicitation is going to be the first message to be 669 sent from an interface after interface (re)initialization, the node 670 SHOULD delay joining the solicited-node multicast address by a random 671 delay between 0 and MAX_RTR_SOLICITATION_DELAY as specified in RFC 672 2461 [5]. This serves to alleviate congestion when many nodes start 673 up on the link at the same time, such as after a power failure, and 674 may help to avoid race conditions when more than one node is trying 675 to solicit for the same address at the same time. 677 Even if the Neighbor Solicitation is not going to be the first 678 message to be sent, the node SHOULD delay joining the solicited-node 679 multicast address by a random delay between 0 and 680 MAX_RTR_SOLICITATION_DELAY if the address being checked is configured 681 by a router advertisement message sent to a multicast address. The 682 delay will avoid similar congestion when multiple nodes are going to 683 configure addresses by receiving a same single multicast router 684 advertisement. 686 Note that the delay for joining the multicast address implicitly 687 means delaying transmission of the corresponding MLD report message 688 [11]. Since RFC 2710 [11] does not request a random delay to avoid 689 race conditions, just delaying Neighbor Solicitation would cause 690 congestion by the MLD report messages. The congestion would then 691 prevent MLD-snooping switches from working correctly, and, as a 692 result, prevent Duplicate Address Detection from working. The 693 requirement to include the delay for the MLD report in this case 694 avoids this scenario. 696 In order to improve the robustness of the Duplicate Address Detection 697 algorithm, an interface MUST receive and process datagrams sent to 698 the all-nodes multicast address or solicited-node multicast address 699 of the tentative address while the delaying period. This does not 700 necessarily conflict with the requirement that joining the multicast 701 group be delayed. In fact, in some cases it is possible for a node to 702 start listening to the group during the delay period before MLD 703 report transmission. It should be noted, however, that in some 704 link-layer environments, particularly with MLD-snooping switches, no 705 multicast reception will be available until the MLD report is sent. 707 5.4.3 Receiving Neighbor Solicitation Messages 709 On receipt of a valid Neighbor Solicitation message on an interface, 710 node behavior depends on whether the target address is tentative or 711 not. If the target address is not tentative (i.e., it is assigned to 712 the receiving interface), the solicitation is processed as described 713 in RFC 2461 [5]. If the target address is tentative, and the source 714 address is a unicast address, the solicitation's sender is performing 715 address resolution on the target; the solicitation should be silently 716 ignored. Otherwise, processing takes place as described below. In 717 all cases, a node MUST NOT respond to a Neighbor Solicitation for a 718 tentative address. 720 If the source address of the Neighbor Solicitation is the unspecified 721 address, the solicitation is from a node performing Duplicate Address 722 Detection. If the solicitation is from another node, the tentative 723 address is a duplicate and should not be used (by either node). If 724 the solicitation is from the node itself (because the node loops back 725 multicast packets), the solicitation does not indicate the presence 726 of a duplicate address. 728 Implementor's Note: many interfaces provide a way for upper layers to 729 selectively enable and disable the looping back of multicast packets. 730 The details of how such a facility is implemented may prevent 731 Duplicate Address Detection from working correctly. See the Appendix 732 A for further discussion. 734 The following tests identify conditions under which a tentative 735 address is not unique: 737 - If a Neighbor Solicitation for a tentative address is received 738 prior to having sent one, the tentative address is a duplicate. 739 This condition occurs when two nodes run Duplicate Address 740 Detection simultaneously, but transmit initial solicitations at 741 different times (e.g., by selecting different random delay values 742 before joining the solicited-node multicast address and 743 transmitting an initial solicitation). 745 - If the actual number of Neighbor Solicitations received exceeds 746 the number expected based on the loopback semantics (e.g., the 747 interface does not loopback packet, yet one or more solicitations 748 was received), the tentative address is a duplicate. This 749 condition occurs when two nodes run Duplicate Address Detection 750 simultaneously and transmit solicitations at roughly the same 751 time. 753 5.4.4 Receiving Neighbor Advertisement Messages 755 On receipt of a valid Neighbor Advertisement message on an interface, 756 node behavior depends on whether the target address is tentative or 757 matches a unicast or anycast address assigned to the interface. If 758 the target address is assigned to the receiving interface, the 759 solicitation is processed as described in RFC 2461 [5]. If the target 760 address is tentative, the tentative address is not unique. 762 5.4.5 When Duplicate Address Detection Fails 764 A tentative address that is determined to be a duplicate as described 765 above MUST NOT be assigned to an interface and the node SHOULD log a 766 system management error. If the address is a link-local address 767 formed from an interface identifier based on the hardware address 768 (e.g., EUI-64), the interface SHOULD be disabled. In this case, the 769 IP address duplication probably means duplicate hardware addresses 770 are in use, and trying to recover from it by configuring another IP 771 address will not result in a usable network. In fact, it probably 772 makes things worse by creating problems that are harder to diagnose 773 than just shutting down the interface; the user will see a partially 774 working network where some things work, and other things will not. On 775 the other hand, if the duplicated link-local address is not formed 776 from an interface identifier based on the hardware address, the 777 interface MAY continue to be used. 779 5.5 Creation of Global Addresses 781 Global addresses are formed by appending an interface identifier to a 782 prefix of appropriate length. Prefixes are obtained from Prefix 783 Information options contained in Router Advertisements. Creation of 784 global addresses and configuration of other parameters as described 785 in this section SHOULD be locally configurable. However, the 786 processing described below MUST be enabled by default. 788 5.5.1 Soliciting Router Advertisements 790 Router Advertisements are sent periodically to the all-nodes 791 multicast address. To obtain an advertisement quickly, a host sends 792 out Router Solicitations as described in RFC 2461 [5]. 794 5.5.2 Absence of Router Advertisements 796 Even if a link has no routers, stateful autoconfiguration to obtain 797 addresses and other configuration information may still be available, 798 and hosts may want to use the mechanism. From the perspective of 799 autoconfiguration, a link has no routers if no Router Advertisements 800 are received after having sent a small number of Router Solicitations 801 as described in RFC 2461 [5]. 803 Note that it is possible that there is no router on the link in this 804 sense but there is a node that has the ability to forward packets. In 805 this case, the forwarding node's address must be manually configured 806 in hosts to be able to send packets off-link, since the only 807 mechanism to configure the default router's address automatically is 808 the one using router advertisements. 810 5.5.3 Router Advertisement Processing 812 For each Prefix-Information option in the Router Advertisement: 814 a) If the Autonomous flag is not set, silently ignore the Prefix 815 Information option. 817 b) If the prefix is the link-local prefix, silently ignore the 818 Prefix Information option. 820 c) If the preferred lifetime is greater than the valid lifetime, 821 silently ignore the Prefix Information option. A node MAY wish to 822 log a system management error in this case. 824 d) If the prefix advertised is not equal to the prefix of an address 825 configured by stateless autoconfiguration already in the list of 826 addresses associated with the interface (where "equal" means the 827 two prefix lengths are the same and the first prefix-length bits 828 of the prefixes are identical), and the Valid Lifetime is not 0, 829 form an address (and add it to the list) by combining the 830 advertised prefix with the link's interface identifier as follows: 832 | 128 - N bits | N bits | 833 +---------------------------------------+------------------------+ 834 | link prefix | interface identifier | 835 +----------------------------------------------------------------+ 837 If the sum of the prefix length and interface identifier length 838 does not equal 128 bits, the Prefix Information option MUST be 839 ignored. An implementation MAY wish to log a system management 840 error in this case. The length of the interface identifier is 841 defined in a separate link-type specific document, which should 842 also be consistent with the address architecture [4] (see Section 843 2). 845 It is the responsibility of the system administrator to insure 846 that the lengths of prefixes contained in Router Advertisements 847 are consistent with the length of interface identifiers for that 848 link type. It should be noted, however, that this does not mean 849 the advertised prefix length is meaningless. In fact, the 850 advertised length has non trivial meaning for on-link 851 determination in RFC 2461 [5] where the sum of the prefix length 852 and the interface identifier length may not be equal to 128. Thus, 853 it should be safe to validate the advertised prefix length here, 854 in order to detect and avoid a configuration error specifying an 855 invalid prefix length in the context of address autoconfiguration. 857 Note that a future revision of the address architecture [4] and a 858 future link-type specific document, which will still be consistent 859 with each other, could potentially allow for an interface 860 identifier of length other than the value defined in the current 861 documents. Thus, an implementation should not assume a particular 862 constant. Rather, it should expect any lengths of interface 863 identifiers. 865 If an address is formed successfully, the host adds it to the list 866 of addresses assigned to the interface, initializing its preferred 867 and valid lifetime values from the Prefix Information option. 869 e) If the advertised prefix is equal to the prefix of an address 870 configured by stateless autoconfiguration in the list, the 871 preferred lifetime of the address is reset to the Preferred 872 Lifetime in the received advertisement. The specific action to 873 perform for the valid lifetime of the address depends on the Valid 874 Lifetime in the received advertisement and the remaining time to 875 the valid lifetime expiration of the previously autoconfigured 876 address. We call the remaining time "RemainingLifetime" in the 877 following discussion: 879 1. If the received Valid Lifetime is greater than 2 hours or 880 greater than RemainingLifetime, set the valid lifetime of the 881 corresponding address to the advertised Valid Lifetime. 883 2. If RemainingLifetime is less than or equal to 2 hours, ignore 884 the Prefix Information option with regards to the valid 885 lifetime, unless the Router Advertisement from which this 886 option was obtained has been authenticated (e.g., via IP 887 security [1]). If the Router Advertisement was authenticated, 888 the valid lifetime of the corresponding address should be set 889 to the Valid Lifetime in the received option. 891 3. Otherwise, reset the valid lifetime of the corresponding 892 address to two hours. 894 The above rules address a specific denial of service attack in 895 which a bogus advertisement could contain prefixes with very small 896 Valid Lifetimes. Without the above rules, a single unauthenticated 897 advertisement containing bogus Prefix Information options with 898 short Valid Lifetimes could cause all of a node's addresses to 899 expire prematurely. The above rules ensure that legitimate 900 advertisements (which are sent periodically) will "cancel" the 901 short Valid Lifetimes before they actually take effect. 903 Note that the preferred lifetime of the corresponding address is 904 always reset to the Preferred Lifetime in the received Prefix 905 Information option, regardless of whether the valid lifetime is 906 also reset or ignored. The difference comes from the fact that the 907 possible attack for the preferred lifetime is relatively minor. 908 Additionally, it is even undesirable to ignore the preferred 909 lifetime when a valid administrator wants to deprecate a 910 particular address by sending a short preferred lifetime (and the 911 valid lifetime is ignored by accident). 913 5.5.4 Address Lifetime Expiry 915 A preferred address becomes deprecated when its preferred lifetime 916 expires. A deprecated address SHOULD continue to be used as a source 917 address in existing communications, but SHOULD NOT be used to 918 initiate new communications if an alternate (non-deprecated) address 919 of sufficient scope can easily be used instead. 921 Note that the feasibility of initiating new communication using a 922 non-deprecated address may be an application-specific decision, as 923 only the application may have knowledge about whether the (now) 924 deprecated address was (or still is) in use by the application. For 925 example, if an application explicitly specifies the protocol stack to 926 use a deprecated address as a source address, the protocol stack must 927 accept that; the application might request it because that IP address 928 is used for in higher-level communication and there might be a 929 requirement that the multiple connections in such a grouping use the 930 same pair of IP addresses. 932 IP and higher layers (e.g., TCP, UDP) MUST continue to accept and 933 process datagrams destined to a deprecated address as normal since a 934 deprecated address is still a valid address for the interface. In the 935 case of TCP, this means TCP SYN segments sent to a deprecated address 936 are responded to using the deprecated address as a source address in 937 the corresponding SYN-ACK (if the connection would otherwise be 938 allowed). 940 An implementation MAY prevent any new communication from using a 941 deprecated address, but system management MUST have the ability to 942 disable such a facility, and the facility MUST be disabled by 943 default. 945 Other subtle cases should also be noted about source address 946 selection. For example, the above description does not clarify which 947 address should be used between a deprecated, smaller-scope address 948 and a non-deprecated, enough scope address. The details of the 949 address selection including this case are described in RFC 3484 [8] 950 and beyond the scope of this document. 952 An address (and its association with an interface) becomes invalid 953 when its valid lifetime expires. An invalid address MUST NOT be used 954 as a source address in outgoing communications and MUST NOT be 955 recognized as a destination on a receiving interface. 957 5.6 Configuration Consistency 959 It is possible for hosts to obtain address information using both 960 stateless and stateful protocols since both may be enabled at the 961 same time. It is also possible that the values of other 962 configuration parameters such as MTU size and hop limit will be 963 learned from both Router Advertisements and the stateful 964 autoconfiguration protocol. If the same configuration information is 965 provided by multiple sources, the value of this information should be 966 consistent. However, it is not considered a fatal error if 967 information received from multiple sources is inconsistent. Hosts 968 accept the union of all information received via the stateless and 969 stateful protocols. If inconsistent information is learned different 970 sources, the most recently obtained values always have precedence 971 over information learned earlier. 973 5.7 Retaining Configured Addresses for Stability 975 An implementation that has stable storage may want to retain 976 addresses in the storage when the addresses were acquired using 977 stateless address autoconfiguration. Assuming the lifetimes used are 978 reasonable, this technique implies that a temporary outage (less than 979 the valid lifetime) of a router will never result in the node losing 980 its global address even if the node were to reboot. When this 981 technique is used, it should also be noted that the expiration times 982 of the preferred and valid lifetimes must be retained, in order to 983 prevent the use of an address after it has become deprecated or 984 invalid. 986 Further details on this kind of extension are beyond the scope of 987 this document. 989 6. SECURITY CONSIDERATIONS 991 Stateless address autoconfiguration allows a host to connect to a 992 network, configure an address and start communicating with other 993 nodes without ever registering or authenticating itself with the 994 local site. Although this allows unauthorized users to connect to 995 and use a network, the threat is inherently present in the Internet 996 architecture. Any node with a physical attachment to a network can 997 generate an address (using a variety of ad hoc techniques) that 998 provides connectivity. 1000 The use of stateless address autoconfiguration and Duplicate Address 1001 Detection opens up the possibility of several denial of service 1002 attacks. For example, any node can respond to Neighbor Solicitations 1003 for a tentative address, causing the other node to reject the address 1004 as a duplicate. A separate document [12] discusses details about 1005 these attacks. These attacks can be addressed by requiring that 1006 Neighbor Discovery packets be authenticated [1]. However, it should 1007 be noted that [12] points out the use of IP security is not always 1008 feasible depending on network environments. 1010 7. IANA CONSIDERATIONS 1012 This document has no actions for IANA. 1014 8. Acknowledgements 1016 The authors would like to thank the members of both the IPNG (which 1017 is now IPV6) and ADDRCONF working groups for their input. In 1018 particular, thanks to Jim Bound, Steve Deering, Richard Draves, and 1019 Erik Nordmark. Thanks also goes to John Gilmore for alerting the WG 1020 of the "0 Lifetime Prefix Advertisement" denial of service attack 1021 vulnerability; this document incorporates changes that address this 1022 vulnerability. 1024 A number of people have contributed to identifying issues on a 1025 previous version of this document and to proposing resolutions to the 1026 issues, on which this version is based. In addition to those listed 1027 above, the contributors include Jari Arkko, Brian E Carpenter, 1028 Gregory Daley, Ralph Droms, Christian Huitema, Soohong Daniel Park, 1029 Markku Savela, and Pekka Savola. 1031 Normative References 1033 [1] Kent, S. and R. Atkinson, "IP Authentication Header", RFC 2402, 1034 November 1998. 1036 [2] Crawford, M., "A Method for the Transmission of IPv6 Packets 1037 over Ethernet Networks", RFC 2464, December 1998. 1039 [3] Bradner, S., "Key words for use in RFCs to Indicate Requirement 1040 Levels", RFC 2119, March 1997. 1042 [4] Hinden, R. and S. Deering, "Internet Protocol Version 6 (IPv6) 1043 Addressing Architecture", RFC 3513, April 2003. 1045 [5] Narten, T., Nordmark, E. and W. Simpson, "Neighbor Discovery for 1046 IP Version 6 (IPv6)", RFC 2461, December 1998. 1048 Informative References 1050 [6] Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C. and M. 1051 Carney, "Dynamic Host Configuration Protocol for IPv6 1052 (DHCPv6)", RFC 3315, July 2003. 1054 [7] Droms, R., "Stateless Dynamic Host Configuration Protocol 1055 (DHCP) Service for IPv6", RFC 3736, April 2004. 1057 [8] Draves, R., "Default Address Selection for Internet Protocol 1058 version 6 (IPv6)", RFC 3484, February 2003. 1060 [9] Narten, T. and R. Draves, "Privacy Extensions for Stateless 1061 Address Autoconfiguration in IPv6", RFC 3041, January 2001. 1063 [10] Aura, T., "Cryptographically Generated Addresses (CGA)", 1064 draft-ietf-send-cga-06.txt (work in progress), April 2004. 1066 [11] Deering, S., Fenner, W. and B. Haberman, "Multicast Listener 1067 Discovery (MLD) for IPv6", RFC 2710, October 1999. 1069 [12] Nikander, P., Kempf, J. and E. Nordmark, "IPv6 Neighbor 1070 Discovery (ND) Trust Models and Threats", RFC 3756, May 2004. 1072 [13] Deering, S., "Host Extensions for IP Multicasting", RFC 1112, 1073 August 1989. 1075 [14] IEEE, "Wireless LAN Medium Access Control (MAC) and Physical 1076 Layer (PHY) Specifications", ANSI/IEEE STd 802.11, August 1999. 1078 Authors' Addresses 1080 Susan Thomson 1081 Cisco Systems 1083 EMail: sethomso@cisco.com 1084 Thomas Narten 1085 IBM Corporation 1086 P.O. Box 12195 1087 Research Triangle Park, NC 27709-2195 1088 USA 1090 Phone: +1 919-254-7798 1091 EMail: narten@us.ibm.com 1093 Tatuya Jinmei 1094 Corporate Research & Development Center, Toshiba Corporation 1095 1 Komukai Toshiba-cho, Saiwai-ku 1096 Kawasaki-shi, Kanagawa 212-8582 1097 Japan 1099 Phone: +81 44-549-2230 1100 EMail: jinmei@isl.rdc.toshiba.co.jp 1102 Appendix A. LOOPBACK SUPPRESSION & DUPLICATE ADDRESS DETECTION 1104 Determining whether a received multicast solicitation was looped back 1105 to the sender or actually came from another node is implementation- 1106 dependent. A problematic case occurs when two interfaces attached to 1107 the same link happen to have the same identifier and link-layer 1108 address, and they both send out packets with identical contents at 1109 roughly the same time (e.g., Neighbor Solicitations for a tentative 1110 address as part of Duplicate Address Detection messages). Although a 1111 receiver will receive both packets, it cannot determine which packet 1112 was looped back and which packet came from the other node by simply 1113 comparing packet contents (i.e., the contents are identical). In this 1114 particular case, it is not necessary to know precisely which packet 1115 was looped back and which was sent by another node; if one receives 1116 more solicitations than were sent, the tentative address is a 1117 duplicate. However, the situation may not always be this 1118 straightforward. 1120 The IPv4 multicast specification [13] recommends that the service 1121 interface provide a way for an upper-layer protocol to inhibit local 1122 delivery of packets sent to a multicast group that the sending host 1123 is a member of. Some applications know that there will be no other 1124 group members on the same host, and suppressing loopback prevents 1125 them from having to receive (and discard) the packets they themselves 1126 send out. A straightforward way to implement this facility is to 1127 disable loopback at the hardware level (if supported by the 1128 hardware), with packets looped back (if requested) by software. On 1129 interfaces in which the hardware itself suppresses loopbacks, a node 1130 running Duplicate Address Detection simply counts the number of 1131 Neighbor Solicitations received for a tentative address and compares 1132 them with the number expected. If there is a mismatch, the tentative 1133 address is a duplicate. 1135 In those cases where the hardware cannot suppress loopbacks, however, 1136 one possible software heuristic to filter out unwanted loopbacks is 1137 to discard any received packet whose link-layer source address is the 1138 same as the receiving interface's. There is even a link-layer 1139 specification that requires to discard any such packets [14]. 1140 Unfortunately, use of that criteria also results in the discarding of 1141 all packets sent by another node using the same link-layer address. 1142 Duplicate Address Detection will fail on interfaces that filter 1143 received packets in this manner: 1145 o If a node performing Duplicate Address Detection discards received 1146 packets having the same source link-layer address as the receiving 1147 interface, it will also discard packets from other nodes also 1148 using the same link-layer address, including Neighbor 1149 Advertisement and Neighbor Solicitation messages required to make 1150 Duplicate Address Detection work correctly. This particular 1151 problem can be avoided by temporarily disabling the software 1152 suppression of loopbacks while a node performs Duplicate Address 1153 Detection, if it is possible to disable the suppression. 1155 o If a node that is already using a particular IP address discards 1156 received packets having the same link-layer source address as the 1157 interface, it will also discard Duplicate Address 1158 Detection-related Neighbor Solicitation messages sent by another 1159 node also using the same link-layer address. Consequently, 1160 Duplicate Address Detection will fail, and the other node will 1161 configure a non-unique address. Since it is generally impossible 1162 to know when another node is performing Duplicate Address 1163 Detection, this scenario can be avoided only if software 1164 suppression of loopback is permanently disabled. 1166 Thus, to perform Duplicate Address Detection correctly in the case 1167 where two interfaces are using the same link-layer address, an 1168 implementation must have a good understanding of the interface's 1169 multicast loopback semantics, and the interface cannot discard 1170 received packets simply because the source link-layer address is the 1171 same as the interfaces. It should also be noted that a link-layer 1172 specification can conflict with the condition necessary to make 1173 Duplicate Address Detection work. 1175 Appendix B. CHANGES SINCE RFC 1971 1177 o Changed document to use term "interface identifier" rather than 1178 "interface token" for consistency with other IPv6 documents. 1180 o Clarified definition of deprecated address to make clear it is OK 1181 to continue sending to or from deprecated addresses. 1183 o Added rules to Section 5.5.3 Router Advertisement processing to 1184 address potential denial-of-service attack when prefixes are 1185 advertised with very short Lifetimes. 1187 o Clarified wording in Section 5.5.4 to make clear that all upper 1188 layer protocols must process (i.e., send and receive) packets sent 1189 to deprecated addresses. 1191 Appendix C. CHANGE HISTORY 1193 Changes since RFC 2462 are: 1195 o Fixed a typo in Section 2. 1197 o Updated references and categorized them into normative and 1198 informative ones. 1200 o Removed redundant code in denial of service protection in Section 1201 5.5.3. 1203 o Clarified that a unicasted NS or NA should be discarded while 1204 performing Duplicate Address Detection. 1206 o Replaced the word "StoredLifetime" with "RemainingLifetime" with a 1207 precise definition to avoid confusion. 1209 o Removed references to site-local and revise wording around the 1210 keyword. 1212 o Added a note about source address selection with regards to 1213 deprecated vs insufficient-scope addresses, etc. Added a reference 1214 to RFC 3484 for further details. 1216 o Clarified what "new communication" means in Section 5.5.4. 1218 o Added a new subsection (5.7) to mention the possibility to use 1219 stable storage to retain configured addresses for stability. 1221 o Revised the Security Considerations section with a reference to 1222 RFC 3756 and a note that the use of IP security is not always 1223 feasible. 1225 o Added a note with a reference in Appendix A about the case where a 1226 link-layer filtering conflicts with a condition to make DAD work 1227 correctly. 1229 o Specified that a node performing Duplicate Address Detection delay 1230 joining the solicited-node multicast group, not just delay sending 1231 Neighbor Solicitations, explaining the detailed reason. 1233 o Clarified the reason why the interface should be disabled after an 1234 address duplicate is detected. Also clarified that the interface 1235 may continue to be used if the interface identifier is not based 1236 on the hardware address. 1238 o Clarified that the preferred lifetime for an existing configured 1239 address is always reset to the advertised value by Router 1240 Advertisement. 1242 o Updated the description of interface identifier considering the 1243 latest format. 1245 Changes since draft-ietf-ipv6-rfc2462bis-00.txt are: 1247 o Clarified how the length of interface identifiers should be 1248 determined, described the relationship with the prefix length 1249 advertised in Router Advertisements, and avoided using a 1250 particular length hard-coded in this document. 1252 o Added a note when an implementation uses stable storage for 1253 autoconfigured addresses. 1255 o Resolved conflict with the Multicast Listener Discovery 1256 specification about random delay for the first packet from the 1257 host. 1259 o Clarified the semantics of the M and O flags based on the latest 1260 standard and operational status. In particular, clarified that 1261 these flags show the availability of the stateful protocol instead 1262 of a trigger to invoke the stateful protocol. ManagedFlag and 1263 OtherConfigFlag, which were implementation-internal variables, 1264 were removed accordingly. 1266 o Recommended to perform Duplicate Address Detection for all unicast 1267 addresses more strongly, considering a variety of different 1268 interface identifiers, while keeping care of existing 1269 implementations. 1271 o Added a requirement for a random delay befor sending Neighbor 1272 Solicitations for Duplicate Address Detection if the address being 1273 checked is configured by a multicasted Router Advertisements. 1275 o Clarified that the prefix comparison in Section 5.5.3 is based on 1276 exact match. Also clarified the comparison described in this 1277 document concentrates on addresses configured by the stateless 1278 mechanism. 1280 o Revisited the author list. 1282 o Added IANA Considerations Section. 1284 Intellectual Property Statement 1286 The IETF takes no position regarding the validity or scope of any 1287 intellectual property or other rights that might be claimed to 1288 pertain to the implementation or use of the technology described in 1289 this document or the extent to which any license under such rights 1290 might or might not be available; neither does it represent that it 1291 has made any effort to identify any such rights. Information on the 1292 IETF's procedures with respect to rights in standards-track and 1293 standards-related documentation can be found in BCP-11. Copies of 1294 claims of rights made available for publication and any assurances of 1295 licenses to be made available, or the result of an attempt made to 1296 obtain a general license or permission for the use of such 1297 proprietary rights by implementors or users of this specification can 1298 be obtained from the IETF Secretariat. 1300 The IETF invites any interested party to bring to its attention any 1301 copyrights, patents or patent applications, or other proprietary 1302 rights which may cover technology that may be required to practice 1303 this standard. Please address the information to the IETF Executive 1304 Director. 1306 Full Copyright Statement 1308 Copyright (C) The Internet Society (2004). All Rights Reserved. 1310 This document and translations of it may be copied and furnished to 1311 others, and derivative works that comment on or otherwise explain it 1312 or assist in its implementation may be prepared, copied, published 1313 and distributed, in whole or in part, without restriction of any 1314 kind, provided that the above copyright notice and this paragraph are 1315 included on all such copies and derivative works. However, this 1316 document itself may not be modified in any way, such as by removing 1317 the copyright notice or references to the Internet Society or other 1318 Internet organizations, except as needed for the purpose of 1319 developing Internet standards in which case the procedures for 1320 copyrights defined in the Internet Standards process must be 1321 followed, or as required to translate it into languages other than 1322 English. 1324 The limited permissions granted above are perpetual and will not be 1325 revoked by the Internet Society or its successors or assignees. 1327 This document and the information contained herein is provided on an 1328 "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING 1329 TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING 1330 BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION 1331 HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF 1332 MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1334 Acknowledgement 1336 Funding for the RFC Editor function is currently provided by the 1337 Internet Society.