idnits 2.17.1 draft-ietf-zeroconf-ipv4-linklocal-15.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: ---------------------------------------------------------------------------- == The page length should not exceed 58 lines per page, but there was 30 longer pages, the longest (page 2) being 63 lines == It seems as if not all pages are separated by form feeds - found 0 form feeds but 31 pages Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 2 instances of too long lines in the document, the longest one being 4 characters in excess of 72. == There are 4 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == Line 587 has weird spacing: '...) on an inter...' == Line 588 has weird spacing: '...ost has confi...' == Line 589 has weird spacing: '...oes not match...' -- 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.) -- Couldn't find a document date in the document -- date freshness check skipped. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'RFC 959' is mentioned on line 1102, but not defined == Unused Reference: '1394' is defined on line 1215, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2434 (Obsoleted by RFC 5226) -- Obsolete informational reference (is this intentional?): RFC 2462 (Obsoleted by RFC 4862) == Outdated reference: A later version (-18) exists of draft-ietf-dhc-dna-ipv4-07 == Outdated reference: A later version (-47) exists of draft-ietf-dnsext-mdns-31 Summary: 3 errors (**), 0 flaws (~~), 11 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 ZEROCONF Working Group Stuart Cheshire 2 INTERNET-DRAFT Apple Computer 3 Category: Standards Track Bernard Aboba 4 Microsoft Corporation 5 7 June 2004 Erik Guttman 6 Sun Microsystems 8 Dynamic Configuration of IPv4 Link-Local Addresses 10 By submitting this Internet-Draft, I certify that any applicable 11 patent or other IPR claims of which I am aware have been disclosed, 12 and any of which I become aware will be disclosed, in accordance with 13 RFC 3667. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that other 17 groups may also distribute working documents as Internet-Drafts. 19 Internet-Drafts are draft documents valid for a maximum of six months 20 and may be updated, replaced, or obsoleted by other documents at any 21 time. It is inappropriate to use Internet-Drafts as reference 22 material or to cite them other than as "work in progress." 24 The list of current Internet-Drafts can be accessed at http:// 25 www.ietf.org/ietf/1id-abstracts.txt. 27 The list of Internet-Draft Shadow Directories can be accessed at 28 http://www.ietf.org/shadow.html. 30 This Internet-Draft will expire on November 22, 2004. 32 Copyright Notice 34 Copyright (C) The Internet Society (2004). All Rights Reserved. 36 Abstract 38 To participate in wide-area IP networking, a host needs to be 39 configured with IP addresses for its interfaces, either manually by 40 the user or automatically from a source on the network such as a DHCP 41 server. Unfortunately, such address configuration information may not 42 always be available. It is therefore beneficial for a host to be able 43 to depend on a useful subset of IP networking functions even when no 44 address configuration is available. This document describes how a 45 host may automatically configure an interface with an IPv4 address 46 within the 169.254/16 prefix that is valid for communication with 47 other devices connected to the same physical (or logical) link. 49 IPv4 Link-Local addresses are not suitable for communication with 50 devices not directly connected to the same physical (or logical) 51 link, and are only used where stable, routable addresses are not 52 available (such as on ad hoc or isolated networks). This document 53 does not recommend that IPv4 Link-Local addresses and routable 54 addresses be configured simultaneously on the same interface. 56 Table of Contents 58 1. Introduction.............................................. 3 59 1.1 Requirements ....................................... 3 60 1.2 Terminology ........................................ 3 61 1.3 Applicability....................................... 4 62 1.4 Application Layer Protocol Considerations........... 5 63 1.5 Autoconfiguration Issues ........................... 6 64 1.6 Alternate Use Prohibition .......................... 7 65 1.7 Multiple Interfaces ................................ 7 66 1.8 Communication with Routable Addresses............... 7 67 1.9 When to configure an IPv4 Link-Local Address ....... 7 68 2. Address Selection, Defense and Delivery................... 9 69 2.1 Link-Local Address Selection........................ 9 70 2.2 Claiming a Link-Local Address....................... 10 71 2.3 Shorter Timeouts ................................... 12 72 2.4 Announcing an Address............................... 12 73 2.5 Conflict Detection and Defense...................... 12 74 2.6 Address Usage and Forwarding Rules.................. 13 75 2.7 Link-Local Packets Are Not Forwarded................ 15 76 2.8 Link-Local Packets are Local........................ 15 77 2.9 Higher-Layer Protocol Considerations................ 15 78 2.10 Privacy Concerns.................................... 16 79 2.11 Interaction between DHCPv4 and IPv4 Link-Local 80 State Machines ..................................... 16 81 3. Considerations for Multiple Interfaces.................... 16 82 3.1 Scoped Addresses.................................... 17 83 3.2 Address Ambiguity................................... 18 84 3.3 Interaction with Hosts with Routable Addresses...... 18 85 3.4 Unintentional Autoimmune Response................... 19 86 4. Healing of Network Partitions ............................ 20 87 5. Security Considerations................................... 21 88 6. Application Programming Considerations.................... 22 89 6.1 Address Changes, Failure and Recovery............... 22 90 6.2 Limited Forwarding of Locators...................... 22 91 6.3 Address Ambiguity................................... 23 92 7. Router Considerations..................................... 23 93 8. IANA Considerations....................................... 23 94 9. Constants ................................................ 23 95 10. References ............................................... 24 96 10.1 Normative References ............................... 24 97 10.2 Informative References ............................. 24 98 Acknowledgments .............................................. 25 99 Authors' Addresses ........................................... 26 100 Appendix A - Prior Implementations............................ 27 101 Intellectual Property Statement .............................. 30 102 Full Copyright Statement ..................................... 31 103 1. Introduction 105 As the Internet Protocol continues to grow in popularity, it becomes 106 increasingly valuable to be able to use familiar IP tools such as FTP 107 not only for global communication, but for local communication as 108 well. For example, two people with laptop computers supporting IEEE 109 802.11 Wireless LANs [802.11] may meet and wish to exchange files. 110 It is desirable for these people to be able to use IP application 111 software without the inconvenience of having to manually configure 112 static IP addresses or set up a DHCP server [RFC2131]. 114 This document describes a method by which a host may automatically 115 configure an interface with an IPv4 address in the 169.254/16 prefix 116 that is valid for Link-Local communication on that interface. This 117 is especially valuable in environments where no other configuration 118 mechanism is available. The IPv4 prefix 169.254/16 is registered 119 with the IANA for this purpose. Allocation of IPv6 Link-Local 120 addresses is described in "IPv6 Stateless Address Autoconfiguration" 121 [RFC2462]. 123 Link-Local communication using IPv4 Link-Local addresses is only 124 suitable for communication with other devices connected to the same 125 physical (or logical) link. Link-Local communication using IPv4 126 Link-Local addresses is not suitable for communication with devices 127 not directly connected to the same physical (or logical) link. 129 Microsoft Windows 98 (and later) and Mac OS 8.5 (and later) already 130 support this capability. This document standardizes usage, 131 prescribing rules for how IPv4 Link-Local addresses are to be treated 132 by hosts and routers. In particular, it describes how routers are to 133 behave when receiving packets with IPv4 Link-Local addresses in the 134 source or destination address. With respect to hosts, it discusses 135 claiming and defending addresses, maintaining Link-Local and routable 136 IPv4 addresses on the same interface, and multi-homing issues. 138 1.1. Requirements 140 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 141 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 142 document are to be interpreted as described in [RFC2119]. 144 1.2. Terminology 146 This document describes Link-Local addressing, for IPv4 communication 147 between two hosts on a single link. A set of hosts is considered to 148 be "on the same link", if: 150 - when any host A from that set sends a packet to any other 151 host B in that set, using unicast, multicast, or broadcast, 152 the entire link-layer packet payload arrives unmodified, 153 and 155 - a broadcast sent over that link by any host from that set 156 of hosts can be received by every other host in that set 158 The link-layer *header* may be modified, such as in Token Ring Source 159 Routing [802.5], but not the link-layer *payload*. In particular, if 160 any device forwarding a packet modifies any part of the IP header or 161 IP payload then the packet is no longer considered to be on the same 162 link. This means that the packet may pass through devices such as 163 repeaters, bridges, hubs or switches and still be considered to be on 164 the same link for the purpose of this document, but not through a 165 device such as an IP router that decrements the TTL or otherwise 166 modifies the IP header. 168 This document uses the term "routable address" to refer to all valid 169 unicast IPv4 addresses outside the 169.254/16 prefix that may be 170 forwarded via routers. This includes all global IP addresses and 171 private addresses such as Net 10/8 [RFC1918], but not loopback 172 addresses such as 127.0.0.1. 174 Wherever this document uses the term "host" when describing use of 175 IPv4 Link-Local addresses, the text applies equally to routers when 176 they are the source of or intended destination of packets containing 177 IPv4 Link-Local source or destination addresses. 179 Wherever this document uses the term "sender IP address" or "target 180 IP address" in the context of an ARP packet, it is referring to the 181 fields of the ARP packet identified in the ARP specification [RFC826] 182 as "ar$spa" (Sender Protocol Address) and "ar$tpa" (Target Protocol 183 Address) respectively. For the usage of ARP described in this 184 document, each of these fields always contains an IP address. 186 In this document, the term "ARP Probe" is used to refer to an ARP 187 Request packet, broadcast on the local link, with an all-zero 'sender 188 IP address'. The 'sender hardware address' MUST contain the hardware 189 address of the interface sending the packet. The 'target hardware 190 address' field is ignored and SHOULD be set to all zeroes. The 191 'target IP address' field MUST be set to the address being probed. 193 In this document, the term "ARP Announcement" is used to refer to an 194 ARP Request packet, broadcast on the local link, identical to the ARP 195 probe described above, except that both the sender and target IP 196 address fields contain the IP address being announced. 198 Constants are introduced in all capital letters. Their values are 199 given in Section 9. 201 1.3. Applicability 203 This specification applies to all IEEE 802 Local Area Networks (LANs) 204 [802], including Ethernet [802.3], Token-Ring [802.5] and IEEE 802.11 205 wireless LANs [802.11], as well as to other link-layer technologies 206 that operate at data rates of at least 1 Mbps, have a round-trip 207 latency of at most one second, and support ARP [RFC826]. Wherever 208 this document uses the term "IEEE 802", the text applies equally to 209 any of these network technologies. 211 Link-layer technologies that support ARP but operate at rates below 1 212 Mbps or latencies above one second may need to specify different 213 values for the following parameters. 215 (a) the number of, and interval between, ARP probes, 216 see PROBE_NUM, PROBE_MIN, PROBE_MAX defined in Section 2.2.1 218 (b) the number of, and interval between, ARP announcements, 219 see ANNOUNCE_N and ANNOUNCE_INTERVAL defined in Section 2.4 221 (c) the maximum rate at which address claiming may be attempted, 222 see RATE_LIMIT_INTERVAL and MAX_COLLISIONS defined in Section 2.2.1 224 (d) the time interval between conflicting ARPs below which a host 225 MUST reconfigure instead of attempting to defend its address, 226 see WATCH_WAIT defined in Section 2.5 228 Link-layer technologies that do not support ARP may be able to use 229 other techniques for determining whether a particular IP address is 230 currently in use. However, the application of claim-and-defend 231 mechanisms to such networks is outside the scope of this document. 233 This specification is intended for use with small ad-hoc networks - a 234 single link containing only a few hosts. Although 65024 IPv4 Link- 235 Local addresses are available in principle, attempting to use all 236 those addresses on a single link would result a high probability of 237 an address conflict, requiring a host to take an inordinate amount of 238 time to find an available address. 240 Network operators with more than 1300 hosts on a single link may want 241 to consider dividing that single link into two or more subnets. A 242 host connecting to a link that already has 1300 hosts, selecting an 243 IPv4 Link-Local address at random, has a 98% chance of selecting an 244 unused IPv4 Link-Local address on the first try. A host has a 99.96% 245 chance of selecting an unused IPv4 Link-Local address within two 246 tries. The probability that it will have to try more than ten times 247 is about 1 in 10^17. 249 1.4. Application Layer Protocol Considerations 251 IPv4 Link-Local addresses and their dynamic configuration have 252 profound implications upon applications which use them. This is 253 discussed in Section 6. Many applications fundamentally assume that 254 addresses of communicating peers are routable, relatively unchanging 255 and unique. These assumptions no longer hold with IPv4 Link-Local 256 addresses, or a mixture of Link-Local and routable IPv4 addresses. 258 Therefore while many applications will work properly with IPv4 Link- 259 Local addresses, or a mixture of Link-Local and routable IPv4 260 addresses, others may do so only after modification, or will exhibit 261 reduced or partial functionality. 263 In some cases it may be infeasible for the application to be modified 264 to operate under such conditions. 266 IPv4 Link-Local addresses should therefore only be used where stable, 267 routable addresses are not available (such as on ad hoc or isolated 268 networks) or in controlled situations where these limitations and 269 their impact on applications are understood and accepted. This 270 document does not recommend that IPv4 Link-Local addresses and 271 routable addresses be configured simultaneously on the same 272 interface. 274 Use of IPv4 Link-Local addresses in off-link communication is likely 275 to cause application failures. This can occur within any application 276 that includes embedded addresses, if an IPv4 Link-Local address is 277 embedded when communicating with a host that is not on the link. 278 Examples of applications that include embedded addresses are found in 279 [RFC3027]. This includes IPsec, Kerberos 4/5, FTP, RSVP, SMTP, SIP, 280 X-Windows/Xterm/Telnet, Real Audio, H.323, and SNMP. 282 To preclude use of IPv4 Link-Local addresses in off-link 283 communication, the following cautionary measures are advised: 285 a. IPv4 Link-Local addresses MUST NOT be configured in the DNS. 287 b. Names that are globally resolvable to routable addresses should 288 be used within applications whenever they are available. Names 289 that are resolvable only on the local link (such as through use 290 of protocols such as Link Local Multicast Name Resolution 291 [LLMNR]) MUST NOT be used in off-link communication. IPv4 292 addresses and names which can only be resolved on the local 293 link SHOULD NOT be forwarded beyond the local link. IPv4 294 Link-Local addresses SHOULD only be sent when a Link-Local 295 address is used as the source address. This strong advice 296 should hinder limited scope addresses and names from leaving 297 the context in which they apply. 299 c. If names resolvable to globally routable addresses are not 300 available, but the globally routable addresses are, they should 301 be used instead of IPv4 Link-Local addresses. 303 1.5. Autoconfiguration Issues 305 Implementations of IPv4 Link-Local address autoconfiguration MUST 306 expect address collisions, and MUST be prepared to handle them 307 gracefully by automatically selecting a new address whenever a 308 collision is detected, as described in Section 2. This requirement 309 to detect and handle address collisions applies during the entire 310 period that a host is using a 169.254/16 IPv4 Link-Local address, not 311 just during initial interface configuration. For example, address 312 collisions can occur well after a host has completed booting if two 313 previously separate networks are joined, as described in Section 4. 315 1.6. Alternate Use Prohibition 317 Note that addresses in the 169.254/16 prefix SHOULD NOT be configured 318 manually or by a DHCP server. Manual or DHCP configuration may cause 319 a host to use an address in the 169.254/16 prefix without following 320 the special rules regarding duplicate detection and automatic 321 configuration that pertain to addresses in this prefix. While 322 [RFC2131] indicates that a DHCP client SHOULD probe a newly received 323 address with ARP, this is not mandatory. Similarly, while [RFC2131] 324 recommends that a DHCP server SHOULD probe an address using an ICMP 325 Echo Request before allocating it, this is also not mandatory, and 326 even if the server does this, IPv4 Link-Local addresses are not 327 routable, so a DHCP server not directly connected to a link cannot 328 detect whether a host on that link is already using the desired IPv4 329 Link-Local address. 331 Administrators wishing to configure their own local addresses (using 332 manual configuration, a DHCP server, or any other mechanism not 333 described in this document) should use one of the existing private 334 address prefixes [RFC1918], not the 169.254/16 prefix. 336 1.7. Multiple Interfaces 338 Additional considerations apply to hosts that support more than one 339 active interface where one or more of these interfaces support IPv4 340 Link-Local address configuration. These considerations are 341 discussed in Section 3. 343 1.8. Communication with Routable Addresses 345 There will be cases when devices with a configured Link-Local address 346 will need to communicate with a device with a routable address 347 configured on the same physical link, and vice versa. The rules in 348 Section 2.6 allow this communication. 350 This allows, for example, a laptop computer with only a routable 351 address to communicate with web servers world-wide using its 352 globally-routable address while at the same time printing those web 353 pages on a local printer that has only an IPv4 Link-Local address. 355 1.9. When to configure an IPv4 Link-Local address 357 Having addresses of multiple different scopes assigned to an 358 interface, with no adequate way to determine in what circumstances 359 each address should be used, leads to complexity for applications and 360 confusion for users. A host with an address on a link can 361 communicate with all other devices on that link, whether those 362 devices use Link- Local addresses, or routable addresses. For these 363 reasons, a host SHOULD NOT have both an operable routable address and 364 an IPv4 Link-Local address configured on the same interface. The term 365 "operable address" is used to mean an address which works effectively 366 for communication in the current network context (see below). When 367 an operable routable address is available on an interface, the host 368 SHOULD NOT also assign an IPv4 Link-Local address on that interface. 369 However, during the transition (in either direction) between using 370 routable and IPv4 Link-Local addresses both MAY be in use at once 371 subject to these rules: 373 1. The assignment of an IPv4 Link-Local address on an interface is 374 based solely on the state of the interface, and is independent 375 of any other protocols such as DHCP. A host MUST NOT alter its 376 behavior and use of other protocols such as DHCP because the 377 host has assigned an IPv4 Link-Local address to an interface. 379 2. If a host finds that an interface that was previously 380 configured with an IPv4 Link-Local address now has an operable 381 routable address available, the host MUST use the routable 382 address when initiating new communications, and MUST cease 383 advertising the availability of the IPv4 Link-Local address 384 through whatever mechanisms that address had been made known 385 to others. The host SHOULD continue to use the IPv4 Link-Local 386 address for communications already underway, and MAY continue 387 to accept new communications addressed to the IPv4 Link-Local 388 address. Ways in which an operable routable address might 389 become available on an interface include: 391 * Manual configuration 392 * Address assignment through DHCP 393 * Roaming of the host to a network on which a 394 previously assigned address becomes operable 396 3. If a host finds that an interface no longer has an operable 397 routable address available, the host MAY identify a usable 398 IPv4 Link-Local address (as described in section 2) and assign 399 that address to the interface. Ways in which an operable 400 routable address might cease to be available on an interface 401 include: 403 * Removal of the address from the interface through 404 manual configuration 405 * Expiration of the lease on the address assigned 406 through DHCP 407 * Roaming of the host to a new network on which the 408 address is no longer operable. 410 The determination by the system of whether an address is "operable" 411 is not clear cut and many changes in the system context (e.g. router 412 changes) may affect the operability of an address. In particular 413 roaming of a host from one network to another is likely - but not 414 certain - to change the operability of a configured address but 415 detecting such a move is not always trivial. 417 "Detection of Network Attachment (DNA) in IPv4" [DNAv4] provides 418 further discussion of address assignment and operability 419 determination. 421 2. Address Selection, Defense and Delivery 423 The following section explains the IPv4 Link-Local address selection 424 algorithm, how IPv4 Link-Local addresses are defended, and how IPv4 425 packets with IPv4 Link-Local addresses are delivered. 427 Windows and Mac OS hosts that already implement Link-Local IPv4 428 address auto-configuration are compatible with the rules presented in 429 this section. However, should any interoperability problem be 430 discovered, this document, not any prior implementation, defines the 431 standard. 433 2.1. Link-Local Address Selection 435 When a host wishes to configure an IPv4 Link-Local address, it 436 selects an address using a pseudo-random number generator with a 437 uniform distribution in the range from 169.254.1.0 to 438 169.254.254.255. The IPv4 prefix 169.254/16 is registered with the 439 IANA for this purpose. The first 256 and last 256 addresses in the 440 169.254/16 prefix are reserved for future use and MUST NOT be 441 selected by a host using this dynamic configuration mechanism. 443 The pseudo-random number generation algorithm MUST be chosen so that 444 different hosts do not generate the same sequence of numbers. If the 445 host has access to persistent information that is different for each 446 host, such as its IEEE 802 MAC address, then the pseudo-random number 447 generator SHOULD be seeded using a value derived from this 448 information. This means that even without using any other persistent 449 storage, a host will usually select the same IPv4 Link-Local address 450 each time it is booted, which can be convenient for debugging and 451 other operational reasons. Seeding the pseudo-random number 452 generator using the real-time clock or any other information which is 453 (or may be) identical in every host is NOT suitable for this purpose, 454 because a group of hosts that are all powered on at the same time 455 might then all generate the same sequence, resulting in a never- 456 ending series of conflicts as the hosts move in lock-step though 457 exactly the same pseudo-random sequence, conflicting on every address 458 they probe. 460 Hosts that are equipped with persistent storage MAY, for each 461 interface, record the IPv4 address they have selected. On booting, 462 hosts with a previously recorded address SHOULD use that address as 463 their first candidate when probing. This increases the stability of 464 addresses. For example, if a group of hosts are powered off at 465 night, then when they are powered on the next morning they will all 466 resume using the same addresses, instead of picking different 467 addresses and potentially having to resolve conflicts that arise. 469 2.2. Claiming a Link-Local Address 471 After it has selected an IPv4 Link-Local address, a host MUST test to 472 see if the IPv4 Link-Local address is already in use before beginning 473 to use it. When a network interface transitions from an inactive to 474 an active state, the host does not have knowledge of what IPv4 Link- 475 Local addresses may currently be in use on that link, since the point 476 of attachment may have changed or the network interface may have been 477 inactive when a conflicting address was claimed. 479 Were the host to immediately begin using an IPv4 Link-Local address 480 which is already in use by another host, this would be disruptive to 481 that other host. Since it is possible that the host has changed its 482 point of attachment, a routable address may be obtainable on the new 483 network, and therefore it cannot be assumed that an IPv4 Link-Local 484 address is to be preferred. 486 Before using the IPv4 Link-Local address (e.g. using it as the source 487 address in an IPv4 packet, or as the Sender IPv4 address in an ARP 488 packet) a host MUST perform the probing test described below to 489 achieve better confidence that using the IPv4 Link-Local address will 490 not cause disruption. 492 Examples of events that involve an interface becoming active include: 494 Reboot/startup 495 Wake from sleep (if network interface was inactive during sleep) 496 Bringing up previously inactive network interface 497 IEEE 802 hardware link-state change (appropriate for the 498 media type and security mechanisms which apply) indicates 499 that an interface has become active. 500 Association with a wireless base station or adhoc network. 502 A host MUST NOT perform this check periodically as a matter of 503 course. This would be a waste of network bandwidth, and is 504 unnecessary due to the ability of hosts to passively discover 505 conflicts, as described in Section 2.5. 507 2.2.1. Probe details 509 On a link-layer such as IEEE 802 that supports ARP, conflict 510 detection is done using ARP probes. On link-layer technologies that 511 do not support ARP other techniques may be available for determining 512 whether a particular IPv4 address is currently in use. However, the 513 application of claim-and-defend mechanisms to such networks is 514 outside the scope of this document. 516 A host probes to see if an address is already in use by broadcasting 517 an ARP Request for the desired address. The client MUST fill in the 518 'sender hardware address' field of the ARP Request with the hardware 519 address of the interface through which it is sending the packet. The 520 'sender IP address' field MUST be set to all zeroes, to avoid 521 polluting ARP caches in other hosts on the same link in the case 522 where the address turns out to be already in use by another host. 523 The 'target hardware address' field is ignored and SHOULD be set to 524 all zeroes. The 'target IP address' field MUST be set to the address 525 being probed. An ARP Request constructed this way with an all-zero 526 'sender IP address' is referred to as an "ARP probe". 528 When ready to begin probing, the host should then wait for a random 529 time interval selected uniformly in the range zero to START_WAIT 530 seconds, and should then send PROBE_NUM probe packets, each of these 531 probe packets spaced randomly, PROBE_MIN to PROBE_MAX seconds apart. 533 If during this period, from the beginning of the probing process 534 until DEFEND_INTERVAL seconds after the last probe packet is sent, 535 the host receives any ARP packet (Request *or* Reply) on the 536 interface where the probe is being performed where the packet's 537 'sender IP address' is the address being probed for, then the host 538 MUST treat this address as being in use by some other host, and MUST 539 select a new pseudo-random address and repeat the process. In 540 addition, if during this period the host receives any ARP probe where 541 the packet's 'target IP address' is the address being probed for, and 542 the packet's 'sender hardware address' is not the hardware address of 543 the interfaces the host is attempting to configure, then the host 544 MUST similarly treat this as an address collision and select a new 545 address as above. This can occur if two (or more) hosts attempt to 546 configure the same IPv4 Link-Local address at the same time. 548 A host should maintain a counter of the number of address collisions 549 it has experienced in the process of trying to acquire an address, 550 and if the number of collisions exceeds MAX_COLLISIONS then the host 551 MUST limit the rate at which it probes for new addresses to no more 552 than one new address per RATE_LIMIT_INTERVAL. This is to prevent 553 catastrophic ARP storms in pathological failure cases, such as a 554 rogue host that answers all ARP probes, causing legitimate hosts to 555 go into an infinite loop attempting to select a usable address. 557 If, by DEFEND_INTERVAL seconds after the transmission of the last ARP 558 probe no conflicting ARP Reply or ARP probe has been received, then 559 the host has successfully claimed the desired IPv4 Link-Local 560 address. 562 2.3. Shorter timeouts 564 Network technologies may emerge for which shorter delays are 565 appropriate than those required by this document. A subsequent IETF 566 publication may be produced providing guidelines for different timer 567 settings for PROBE_MIN and PROBE_MAX on those technologies. 569 2.4. Announcing an Address 571 The host MUST then announce its claimed address by broadcasting 572 ANNOUNCE_N ARP announcements, spaced ANNOUNCE_INTERVAL seconds apart. 573 This time interval is not modified by the shorter timeouts described 574 above in Section 2.3. An ARP announcement is identical to the ARP 575 probe described above, except that now the sender and target IP 576 addresses are both set to the host's newly selected IPv4 address. 577 The purpose of these ARP announcements is to make sure that other 578 hosts on the link do not have stale ARP cache entries left over from 579 some other host that may previously have been using the same address. 581 2.5. Conflict Detection and Defense 583 Address collision detection is not limited to the address selection 584 phase, when a host is sending ARP probes. Address collision 585 detection is an ongoing process that is in effect for as long as a 586 host is using an IPv4 Link-Local address. At any time, if a host 587 receives an ARP packet (request *or* reply) on an interface where 588 the 'sender IP address' is the IP address the host has configured 589 for that interface, but the 'sender hardware address' does not match 590 the hardware address of that interface, then this is a conflicting 591 ARP packet, indicating an address collision. 593 A host MUST respond to a conflicting ARP packet as described in 594 either (a) or (b) below: 596 (a) Upon receiving a conflicting ARP packet, a host MAY elect to 597 immediately configure a new IPv4 Link-Local address as described 598 above, or 600 (b) If a host currently has active TCP connections or other reasons 601 to prefer to keep the same IPv4 address, and it has not seen any 602 other conflicting ARP packets recently (for IEEE 802, within the last 603 WATCH_WAIT seconds) then it MAY elect to attempt to defend its 604 address, by recording the time that the conflicting ARP packet was 605 received, and then broadcasting one single ARP announcement, giving 606 its own IP and hardware addresses as the sender addresses of the ARP. 607 Having done this, the host can then continue to use the address 608 normally without any further special action. However, if this is not 609 the first conflicting ARP packet the host has seen, and the time 610 recorded for the previous conflicting ARP packet is recent (within 611 WATCH_WAIT seconds for IEEE 802) then the host MUST immediately cease 612 using this address and configure a new IPv4 Link-Local address as 613 described above. This is necessary to ensure that two hosts do not 614 get stuck in an endless loop with both hosts trying to defend the 615 same address. 617 A host MUST respond to conflicting ARP packets as described in either 618 (a) or (b) above. A host MUST NOT ignore conflicting ARP packets. 620 Forced address reconfiguration may be disruptive, causing TCP 621 connections to be broken. However, it is expected that such 622 disruptions will be rare, and if inadvertent address duplication 623 happens, then disruption of communication is inevitable, no matter 624 how the addresses were assigned. It is not possible for two 625 different hosts using the same IP address on the same network to 626 operate reliably. 628 Before abandoning an address due to a conflict, hosts SHOULD actively 629 attempt to reset any existing connections using that address. This 630 mitigates some security threats posed by address reconfiguration, as 631 discussed in Section 5. 633 Immediately configuring a new address as soon as the conflict is 634 detected is the best way to restore useful communication as quickly 635 as possible. The mechanism described above of broadcasting a single 636 ARP announcement to defend the address mitigates the problem 637 somewhat, by helping to improve the chance that one of the two 638 conflicting hosts may be able to retain its address. 640 All ARP packets (*replies* as well as requests) that contain a Link- 641 Local 'sender IP address' MUST be sent using link-layer broadcast 642 instead of link-layer unicast. This aids timely detection of 643 duplicate addresses. An example illustrating how this helps is given 644 in Section 4. 646 2.6. Address Usage and Forwarding Rules 648 A host implementing this specification has additional rules to 649 conform to, whether or not it has an interface configured with an 650 IPv4 Link-Local address. 652 2.6.1. Source Address Usage 654 Since each interface on a host may have an IPv4 Link-Local address in 655 addition to zero or more other addresses configured by other means 656 (e.g. manually or via a DHCP server), a host may have to make a 657 choice about what source address to use when it sends a packet or 658 initiates a TCP connection. 660 The host SHOULD use a routable address in preference to an IPv4 Link- 661 Local address except for communication to peers for which the host 662 has an existing TCP connection at the time in which the host obtained 663 a routable address configuration. For more details, see Section 1.7. 665 A multi-homed host needs to select an outgoing interface whether or 666 not the destination is an IPv4 Link-Local address. Details of that 667 process are beyond the scope of this specification. After selecting 668 an interface, the multi-homed host should send packets involving IPv4 669 Link-Local addresses as specified in this document, as if the 670 selected interface were the host's only interface. See Section 3 for 671 further discussion of multi-homed hosts. 673 2.6.2. Forwarding Rules 675 Whichever interface is used, if the destination address is in the 676 169.254/16 prefix (excluding the address 169.254.255.255, which is 677 the broadcast address for the Link-Local prefix), then the sender 678 MUST ARP for the destination address and then send its packet 679 directly to the destination on the same physical link. This MUST be 680 done whether the interface is configured with a Link-Local or a 681 routable IPv4 address. 683 In many network stacks, achieving this functionality may be as simple 684 as adding a routing table entry indicating that 169.254/16 is 685 directly reachable on the local link. This approach will not work 686 for routers or multi-homed hosts. Refer to section 3 for more 687 discussion of multi-homed hosts. 689 The host MUST NOT send a packet with an IPv4 Link-Local destination 690 address to any router for forwarding. 692 If the destination address is a unicast address outside the 693 169.254/16 prefix, then the host SHOULD use an appropriate routable 694 IPv4 source address, if it can. If for any reason the host chooses 695 to send the packet with an IPv4 Link-Local source address (e.g. no 696 routable address is available on the selected interface), then it 697 MUST ARP for the destination address and then send its packet, with 698 an IPv4 Link-Local source address and a routable destination IPv4 699 address, directly to its destination on the same physical link. The 700 host MUST NOT send the packet to any router for forwarding. 702 In the case of a device with a single interface and only an Link- 703 Local IPv4 address, this requirement can be paraphrased as "ARP for 704 everything". 706 In many network stacks, achieving this "ARP for everything" behavior 707 may be as simple as having no primary IP router configured, having 708 the primary IP router address configured to 0.0.0.0, or having the 709 primary IP router address set to be the same as the host's own Link- 710 Local IPv4 address. For suggested behavior in multi-homed hosts, see 711 Section 3. 713 2.7. Link-Local Packets Are Not Forwarded 715 A sensible default for applications which are sending from an IPv4 716 Link-Local address is to explicitly set the IPv4 TTL to 1. This is 717 not appropriate in all cases as some applications may require that 718 the IPv4 TTL be set to other values. 720 An IPv4 packet whose source and/or destination address is in the 721 169.254/16 prefix MUST NOT be sent to any router for forwarding, and 722 any network device receiving such a packet MUST NOT forward it, 723 regardless of the TTL in the IPv4 header. Similarly, a router or 724 other host MUST NOT indiscriminately answer all ARP Requests for 725 addresses in the 169.254/16 prefix. A router may of course answer 726 ARP Requests for one or more IPv4 Link-Local address(es) that it has 727 legitimately claimed for its own use according to the claim-and- 728 defend protocol described in this document. 730 This restriction also applies to multicast packets. IPv4 packets with 731 a Link-Local source address MUST NOT be forwarded off the local link 732 even if they have a multicast destination address. 734 2.8. Link-Local Packets are Local 736 The non-forwarding rule means that hosts may assume that all 737 169.254/16 destination addresses are "on-link" and directly 738 reachable. The 169.254/16 address prefix MUST NOT be subnetted. 739 This specification utilizes ARP-based address collision detection, 740 which functions by broadcasting on the local subnet. Since such 741 broadcasts are not forwarded, were subnetting to be allowed then 742 address conflicts could remain undetected. 744 This does not mean that Link-Local devices are forbidden from any 745 communication outside the local link. IP hosts that implement both 746 Link-Local and conventional routable IPv4 addresses may still use 747 their routable addresses without restriction as they do today. 749 2.9. Higher-Layer Protocol Considerations 751 Similar considerations apply at layers above IP. 753 For example, designers of Web pages (including automatically 754 generated web pages) SHOULD NOT contain links with embedded IPv4 755 Link-Local addresses if those pages are viewable from hosts outside 756 the local link where the addresses are valid. 758 As IPv4 Link-Local addresses may change at any time and have limited 759 scope, IPv4 Link-Local addresses MUST NOT be stored in the DNS. 761 2.10. Privacy Concerns 763 Another reason to restrict leakage of IPv4 Link-Local addresses 764 outside the local link is privacy concerns. If IPv4 Link-Local 765 addresses are derived from a hash of the MAC address, some argue that 766 they could be indirectly associated with an individual, and thereby 767 used to track that individual's activities. Within the local link 768 the hardware addresses in the packets are all directly observable, so 769 as long as IPv4 Link-Local addresses don't leave the local link they 770 provide no more information to an intruder than could be gained by 771 direct observation of hardware addresses. 773 2.11. Interaction between DHCPv4 client and IPv4 Link-Local State 774 Machines 776 As documented in Appendix A, early implementations of IPv4 Link-Local 777 have modified the DHCP state machine. Field experience shows that 778 these modifications reduce the reliability of the DHCP service. 780 A device that implements both IPv4 Link-Local and a DHCPv4 client 781 should not alter the behavior of the DHCPv4 client to accommodate 782 IPv4 Link-Local configuration. In particular configuration of an 783 IPv4 Link-Local address, whether or not a DHCP server is currently 784 responding, is not sufficient reason to unconfigure a valid DHCP 785 lease, to stop the DHCP client from attempting to acquire a new IP 786 address, to change DHCP timeouts or to change the behavior of the 787 DHCP state machine in any other way. 789 Further discussion of this issue is provided in [DNAv4]. 791 3. Considerations for Multiple Interfaces 793 These considerations apply whenever a host has multiple IP addresses 794 whether or not it has multiple physical interfaces. Other examples 795 of multiple interfaces include different logical endpoints (tunnels, 796 virtual private networks etc.) and multiple logical networks on the 797 same physical medium. This is often referred to as "multi-homing". 799 Hosts which have more than one active interface and elect to 800 implement dynamic configuration of IPv4 Link-Local addresses on one 801 or more of those interfaces will face various problems. This section 802 lists these problems but does no more than indicate how one might 803 solve them. At the time of this writing, there is no silver bullet 804 which solves these problems in all cases, in a general way. 805 Implementors must think through these issues before implementing the 806 protocol specified in this document on a system which may have more 807 than one active interface as part of a TCP/IP stack capable of multi- 808 homing. 810 3.1. Scoped addresses 812 A host may be attached to more than one network at the same time. It 813 would be nice if there was a single address space used in every 814 network, but this is not the case. Addresses used in one network, be 815 it a network behind a NAT or a link on which IPv4 Link-Local 816 addresses are used, cannot be used in another network and have the 817 same effect. 819 It would also be nice if addresses were not exposed to applications, 820 but they are. Most software using TCP/IP which await messages 821 receives from any interface at a particular port number, for a 822 particular transport protocol. Applications are generally only aware 823 (and care) that they have received a message. The application knows 824 the address of the sender to which the application will reply. 826 The first scoped address problem is source address selection. A 827 multi-homed host has more than one address. Which address should be 828 used as the source address when sending to a particular destination? 829 This question is usually answered by referring to a routing table, 830 which expresses on which interface (with which address) to send, and 831 how to send (should one forward to a router, or send directly). The 832 choice is made complicated by scoped addresses because the address 833 range in which the destination lies may be ambiguous. The table may 834 not be able to yield a good answer. This problem is bound up with 835 next-hop selection, which is discussed in Section 3.2. 837 The second scoped address problem arises from scoped parameters 838 leaking outside their scope. This is discussed in Section 7. 840 It is possible to overcome these problems. One way is to expose scope 841 information to applications such that they are always aware of what 842 scope a peer is in. This way, the correct interface could be 843 selected, and a safe procedure could be followed with respect to 844 forwarding addresses and other scoped parameters. There are other 845 possible approaches. None of these methods have been standardized 846 for IPv4 nor are they specified in this document. A good API design 847 could mitigate the problems, either by exposing address scopes to 848 'scoped-address aware' applications or by cleverly encapsulating the 849 scoping information and logic so that applications do the right thing 850 without being aware of address scoping. 852 An implementer could undertake to solve these problems, but cannot 853 simply ignore them. With sufficient experience, it is hoped that 854 specifications will emerge explaining how to overcome scoped address 855 multi-homing problems. 857 3.2. Address Ambiguity 859 This is a core problem with respect to IPv4 Link-Local addresses 860 configured on more than one interface. What should a host do when it 861 needs to send to Link-Local destination L and L can be resolved using 862 ARP on more than one link? 864 Even if a Link-Local address can be resolved on only one link at a 865 given moment, there is no guarantee that it will remain unambiguous 866 in the future. Additional hosts on other interfaces may claim the 867 address L as well. 869 One possibility is to support this only in the case where the 870 application specifically expresses which interface to send from. 872 There is no standard or obvious solution to this problem. Existing 873 application software written for the IPv4 protocol suite is largely 874 incapable of dealing with address ambiguity. This does not preclude 875 an implementer from finding a solution, writing applications which 876 are able to use it, and providing a host which can support dynamic 877 configuration of IPv4 Link-Local addresses on more than one 878 interface. This solution will almost surely not be generally 879 applicable to existing software and transparent to higher layers, 880 however. 882 Given that the IP stack must have the outbound interface associated 883 with a packet that needs to be sent to a LL destination address, 884 interface selection must occur. The outbound interface cannot be 885 derived from the packet's header parameters such as src or dst 886 address (e.g. by using the forwarding table lookup). Therefore, 887 outbound interface association MUST be done explicitly through other 888 means. The specification does not stipulate those means. 890 3.3. Interaction with Hosts with Routable Addresses 892 Attention is paid in this specification to transition from the use of 893 IPv4 Link-Local addresses to routable addresses (see Section 1.5). 894 The intention is to allow a host with a single interface to first 895 support Link-Local configuration then gracefully transition to the 896 use of a routable address. Since the host transitioning to the use of 897 a routable address will not advertise scoped address information, the 898 scoped address issues described in Section 3.1 will apply. A host 899 which conforms to this specification will know that an IPv4 Link- 900 Local destination must be reached by forwarding to the destination, 901 not to a router, even if the host is sending from a routable address. 903 A host with an IPv4 Link-Local address may send to a destination 904 which does not have an IPv4 Link-Local address. If the host is not 905 multi-homed, the procedure is simple and unambiguous: Using ARP and 906 forwarding directly to on-link destinations is the default route. If 907 the host is multi-homed, however, the routing policy is more complex, 908 especially if one of the interfaces is configured with a routable 909 address and the default route is (sensibly) directed at a router 910 accessible through that interface. The following example illustrates 911 this problem and provides a common solution to it. 913 i1 +---------+ i2 i3 +-------+ 914 ROUTER-------= HOST1 =---------= HOST2 | 915 link1 +-------=-+ link2 +-------+ 917 In the figure above, HOST1 is connected to link1 and link2. Interface 918 i1 is configured with a routable address, while i2 is an IPv4 Link- 919 Local address. HOST1 has its default route set to ROUTER's address, 920 through i1. HOST1 will route to destinations in 169.254/16 to i2, 921 sending directly to the destination. 923 HOST2 has a configured (non-Link-Local) IPv4 address assigned to i3. 925 Using a name resolution or service discovery protocol HOST1 can 926 discover HOST2's address. Since HOST2's address is not in 169.254/16, 927 HOST1's routing policy will send datagrams to HOST2 via i1, to the 928 ROUTER. Unless there is a route from ROUTER to HOST2, the datagrams 929 sent from HOST1 to HOST2 will not reach it. 931 One solution to this problem is for a host to attempt to reach any 932 host locally (using ARP) for which it receives an unreachable ICMP 933 error message (ICMP message codes 0, 1, 6 or 7, see [RFC792]). The 934 host tries all its attached links in a round robin fashion. This has 935 been implemented successfully for some IPv6 hosts, to circumvent 936 exactly this problem. In terms of this example, HOST1 upon failing to 937 reach HOST2 via the ROUTER, will attempt to forward to HOST2 via i2 938 and succeed. 940 It may also be possible to overcome this problem using techniques 941 described in section 3.2, or other means not discussed here. This 942 specification does not provide a standard solution, nor does it 943 preclude implementers from supporting multi-homed configurations, 944 provided that they address the concerns in this section for the 945 applications which will be supported on the host. 947 3.4. Unintentional Autoimmune Response 949 Care must be taken if a multi-homed host can support more than one 950 interface on the same link, all of which support IPv4 Link-Local 951 autoconfiguration. If these interfaces attempt to allocate the same 952 address, they will defend the host against itself - causing the 953 claiming algorithm to fail. The simplest solution to this problem is 954 to run the algorithm independently on each interface configured with 955 IPv4 Link-Local addresses. 957 In particular, ARP packets which appear to claim an address which is 958 assigned to a specific interface, indicate conflict only if they are 959 received on that interface and their hardware address is of some 960 other interface. 962 If a host has two interfaces on the same network, then claiming and 963 defending on those interfaces must ensure that they end up with 964 different addresses just as if they were on different hosts. 966 4. Healing of Network Partitions 968 Hosts on disjoint network links may configure the same IPv4 Link- 969 Local address. If these separate network links are later joined or 970 bridged together, then there may be two hosts which are now on the 971 same link, trying to use the same address. When either host attempts 972 to communicate with any other host on the network, it will at some 973 point broadcast an ARP packet which will enable the hosts in question 974 to detect that there is an address conflict. 976 When these address conflicts are detected, the subsequent forced 977 reconfiguration may be disruptive, causing TCP connections to be 978 broken. However, it is expected that such disruptions will be rare. 979 It should be relatively uncommon for networks to be joined while 980 hosts on those networks are active. Also, 65024 addresses are 981 available for IPv4 Link-Local use, so even when two small networks 982 are joined, the chance of collision for any given host is fairly 983 small. 985 When joining two large networks (defined as networks with a 986 substantial number of hosts per segment) there is a greater chance of 987 collision. In such networks, it is likely that the joining of 988 previously separated segments will result in one or more hosts 989 needing to change their IPv4 Link-Local address, with subsequent loss 990 of TCP connections. In cases where separation and re-joining is 991 frequent, as in remotely bridged networks, this could prove 992 disruptive. However, unless the number of hosts on the joined 993 segments is very large, the traffic resulting from the join and 994 subsequent address conflict resolution will be small. 996 Sending ARP replies that have IPv4 Link-Local sender addresses via 997 broadcast instead of unicast ensures that these conflicts can be 998 detected as soon as they become potential problems, but no sooner. 999 For example, if two disjoint network links are joined, where hosts A 1000 and B have both configured the same Link-Local address, X, they can 1001 remain in this state until A, B or some other host attempts to 1002 initiate communication. If some other host C now sends an ARP request 1003 for address X, and hosts A and B were to both reply with conventional 1004 unicast ARP replies, then host C might be confused, but A and B still 1005 wouldn't know there is a problem because neither would have seen the 1006 other's packet. Sending these replies via broadcast allows A and B 1007 see each other's conflicting ARP packets and respond accordingly. 1009 Note that sending periodic gratuitous ARPs in an attempt to detect 1010 these conflicts sooner is not necessary, wastes network bandwidth, 1011 and may actually be detrimental. For example, if the network links 1012 were joined only briefly, and were separated again before any new 1013 communication involving A or B were initiated, then the temporary 1014 conflict would have been benign and no forced reconfiguration would 1015 have been required. Triggering an unnecessary forced reconfiguration 1016 in this case would not serve any useful purpose. Hosts SHOULD NOT 1017 send periodic gratuitous ARPs. 1019 5. Security Considerations 1021 The use of IPv4 Link-Local Addresses may open a network host to new 1022 attacks. In particular, a host that previously did not have an IP 1023 address, and no IP stack running, was not susceptible to IP-based 1024 attacks. By configuring a working address, the host may now be 1025 vulnerable to IP-based attacks. 1027 The ARP protocol [RFC826] is insecure. A malicious host may send 1028 fraudulent ARP packets on the network, interfering with the correct 1029 operation of other hosts. For example, it is easy for a host to 1030 answer all ARP requests with replies giving its own hardware address, 1031 thereby claiming ownership of every address on the network. 1033 NOTE: The existence of local links without physical security, such as 1034 wireless LANs, means that expecting all local links to be secure 1035 enough that normal precautions can be dispensed with is an extremely 1036 dangerous practice, which will expose users to considerable risks. 1038 A host implementing IPv4 Link-Local configuration has an additional 1039 vulnerability to selective reconfiguration and disruption. It is 1040 possible for an on-link attacker to issue ARP packets which would 1041 cause a host to break all its connections by switching to a new 1042 address. The attacker could force the host implementing IPv4 Link- 1043 Local configuration to select certain addresses, or prevent it from 1044 ever completing address selection. This is a distinct threat from 1045 that posed by spoofed ARPs, described in the preceding paragraph. 1047 Implementations and users should also note that a node that gives up 1048 an address and reconfigures, as required by section 2.5, allows the 1049 possibility that another node can easily successfully hijack existing 1050 TCP connections. 1052 Implementers are advised that the Internet Protocol architecture 1053 expects every networked device or host must implement security which 1054 is adequate to protect the resources to which the device or host has 1055 access, including the network itself, against known or credible 1056 threats. Even though use of IPv4 Link-Local addresses may reduce the 1057 number of threats to which a device is exposed, implementers of 1058 devices supporting the Internet Protocol must not assume that a 1059 customer's local network is free from security risks. 1061 While there may be particular kinds of devices, or particular 1062 environments, for which the security provided by the network is 1063 adequate to protect the resources that are accessible by the device, 1064 it would be misleading to make a general statement to the effect that 1065 the requirement to provide security is reduced for devices using IPv4 1066 Link-Local addresses as a sole means of access. 1068 In all cases, whether or not IPv4 Link-Local addresses are used, it 1069 is necessary for implementers of devices supporting the Internet 1070 Protocol to analyze the known and credible threats to which a 1071 specific host or device might be subjected, and to the extent that it 1072 is feasible, to provide security mechanisms which ameliorate or 1073 reduce the risks associated with such threats. 1075 6. Application Programming Considerations 1077 Use of IPv4 Link-Local autoconfigured addresses presents additional 1078 challenges to writers of applications and may result in existing 1079 application software failing. 1081 6.1. Address Changes, Failure and Recovery 1083 IPv4 Link-Local addresses used by an application may change over 1084 time. Some application software encountering an address change will 1085 fail. For example, existing client TCP connections will be aborted, 1086 servers whose addresses change will have to be rediscovered, blocked 1087 reads and writes will exit with an error condition, and so on. 1089 Vendors producing application software which will be used on IP 1090 implementations supporting IPv4 Link-Local address configuration 1091 SHOULD detect and cope with address change events. Vendors producing 1092 IPv4 implementations supporting IPv4 Link-Local address configuration 1093 SHOULD expose address change events to applications. 1095 6.2. Limited Forwarding of Locators 1097 IPv4 Link-Local addresses MUST NOT be forwarded via an application 1098 protocol (for example in a URL), to a destination that is not on the 1099 same link. This is discussed further in Sections 2.9 and 3. 1101 Existing distributed application software that forwards address 1102 information may fail. For example, FTP [RFC 959] passive mode 1103 transmits the IPv4 address of the client. Assume a client starts up 1104 and obtains its *passive* IPv4 configuration at a time when the host 1105 has only a Link-Local address. Later, the host gets a global IP 1106 address configuration (for one of its interfaces). The client uses 1107 this global IPv4 address to contact an FTP server off of the local 1108 link for which it had (or still has) an IPv4 Link-Local address 1109 configured. If the FTP client transmits its stale out-date passive 1110 IPv4 configuration to the FTP server, the FTP server will be unable 1111 to reach the FTP client. The passive FTP operation will fail. 1113 6.3. Address Ambiguity 1115 Application software run on a multi-homed host that supports IPv4 1116 Link-Local address configuration on more than one interface may fail. 1118 This is because application software assumes that an IPv4 address is 1119 unambiguous, that it can refer to only one host. IPv4 Link-Local 1120 addresses are unique only on a single link. A host attached to 1121 multiple links can easily encounter a situation where the same 1122 address is present on more than one interface, or first on one 1123 interface, later on another; in any case associated with more than 1124 one host. Most existing software is not prepared for this ambiguity. 1125 In the future, application programming interfaces could be developed 1126 to prevent this problem. This issue is discussed in Section 3. 1128 7. Router Considerations 1130 A router MUST NOT forward a packet with an IPv4 Link-Local source or 1131 destination address, irrespective of the router's default route 1132 configuration or routes obtained from dynamic routing protocols. 1134 A router which receives a packet with an IPv4 Link-Local destination 1135 address on an interface which either has no IPv4 Link-Local address 1136 configured or is configured with a different address than the 1137 destination of the packet MUST NOT forward the packet. This prevents 1138 forwarding of packets back onto the network segment from which they 1139 originated, or to any other segment. 1141 8. IANA Considerations 1143 The IANA has allocated the prefix 169.254/16 for the use described in 1144 this document. The first and last 256 addresses in this range 1145 (169.254.0.x and 169.254.255.x) are allocated by Standards Action, as 1146 defined in [RFC2434] BCP 26. No other IANA services are required by 1147 this document. 1149 9. Constants 1151 The following timing constants are used in this protocol; they are 1152 not intended to be user configurable. 1154 START_WAIT 1 second (initial random delay) 1155 PROBE_MIN 1 second (minimum delay till repeated probe) 1156 PROBE_MAX 2 seconds (maximum delay till repeated probe) 1157 PROBE_NUM 3 (number of probe packets) 1158 ANNOUNCE_N 2 (number of announcement packets) 1159 ANNOUNCE_INTERVAL 2 seconds (time between announcement packets) 1160 MAX_COLLISIONS 10 (max collisions before rate limiting) 1161 RATE_LIMIT_INTERVAL 60 seconds (delay between successive attempts) 1162 DEFEND_INTERVAL 10 seconds (after probes, wait for defense) 1163 WATCH_WAIT 10 seconds (time to continue to defend your address, 1164 for IEEE 802 networks) 1166 10. References 1168 10.1. Normative References 1170 [RFC792] Postel, J., "Internet Control Message Protocol", RFC 792, 1171 September 1981. 1173 [RFC826] D. Plummer, "An Ethernet Address Resolution Protocol -or- 1174 Converting Network Addresses to 48-bit Ethernet Address for 1175 Transmission on Ethernet Hardware", STD 37, RFC 826, November 1176 1982. 1178 [RFC2119] S. Bradner, "Key words for use in RFCs to Indicate Requirement 1179 Levels", RFC 2119, March 1997. 1181 [RFC2434] Alvestrand, H. and T. Narten, "Guidelines for Writing an IANA 1182 Considerations Section in RFCs", BCP 26, RFC 2434, October 1183 1998. 1185 10.2. Informative References 1187 [802] IEEE Standards for Local and Metropolitan Area Networks: 1188 Overview and Architecture, ANSI/IEEE Std 802, 1990. 1190 [802.1d] ISO/IEC 15802-3 Information technology - Telecommunications 1191 and information exchange between systems - Local and 1192 metropolitan area networks - Common specifications - Part 3: 1193 Media access Control (MAC) Bridges, (also ANSI/IEEE Std 1194 802.1D-1998), 1998. 1196 [802.3] ISO/IEC 8802-3 Information technology - Telecommunications and 1197 information exchange between systems - Local and metropolitan 1198 area networks - Common specifications - Part 3: Carrier Sense 1199 Multiple Access with Collision Detection (CSMA/CD) Access 1200 Method and Physical Layer Specifications, (also ANSI/IEEE Std 1201 802.3- 1996), 1996. 1203 [802.5] ISO/IEC 8802-5 Information technology - Telecommunications and 1204 information exchange between systems - Local and metropolitan 1205 area networks - Common specifications - Part 5: Token ring 1206 access method and physical layer specifications, (also 1207 ANSI/IEEE Std 802.5-1998), 1998. 1209 [802.11] Information technology - Telecommunications and information 1210 exchange between systems - Local and metropolitan area 1211 networks - Specific Requirements Part 11: Wireless LAN Medium 1212 Access Control (MAC) and Physical Layer (PHY) Specifications, 1213 IEEE Std. 802.11-1999, 1999. 1215 [1394] Standard for a High Performance Serial Bus. Institute of 1216 Electrical and Electronic Engineers, IEEE Standard 1394-1995, 1217 1995. 1219 [RFC1918] Y. Rekhter et.al., "Address Allocation for Private Internets", 1220 RFC 1918, February 1996. 1222 [RFC2131] R. Droms, "Dynamic Host Configuration Protocol", RFC 2131, 1223 March 1997. 1225 [RFC2462] S. Thomson and T. Narten, "IPv6 Stateless Address 1226 Autoconfiguration", RFC 2462, December 1998. 1228 [RFC3027] Holdrege, M. and P. Srisuresh, "Protocol Complications with 1229 the IP Network Address Translator", RFC 3027, January 2001. 1231 [DNAv4] Aboba, B., "Detection of Network Attachment (DNA) in IPv4", 1232 draft-ietf-dhc-dna-ipv4-07.txt, Internet draft (work in 1233 progress), April 2004. 1235 [LLMNR] Esibov, L., Aboba, B. and D. Thaler, "Linklocal Multicast Name 1236 Resolution (LLMNR)", draft-ietf-dnsext-mdns-31.txt, Internet 1237 draft (work in progress), June 2004. 1239 Acknowledgments 1241 We would like to thank (in alphabetical order) Jim Busse, Pavani 1242 Diwanji, Donald Eastlake 3rd, Robert Elz, Peter Ford, Spencer 1243 Giacalone, Josh Graessley, Brad Hards, Myron Hattig, Hugh Holbrook, 1244 Christian Huitema, Richard Johnson, Kim Yong-Woon, Mika Liljeberg, 1245 Rod Lopez, Keith Moore, Satish Mundra, Thomas Narten, Erik Nordmark, 1246 Philip Nye, Howard Ridenour, Daniel Senie, Dieter Siegmund, Valery 1247 Smyslov and Ryan Troll for their contributions. 1249 Authors' Addresses 1251 Stuart Cheshire 1252 Apple Computer, Inc. 1253 1 Infinite Loop 1254 Cupertino 1255 California 95014, USA 1257 Phone: +1 408 974 3207 1258 EMail: rfc@stuartcheshire.org 1260 Bernard Aboba 1261 Microsoft Corporation 1262 One Microsoft Way 1263 Redmond, WA 98052 1265 Phone: +1 425 706 6605 1266 EMail: bernarda@microsoft.com 1268 Erik Guttman 1269 Sun Microsystems 1270 Eichhoelzelstr. 7 1271 74915 Waibstadt Germany 1273 Phone: +49 7263 911 701 1274 Email: erik.guttman@sun.com 1276 Appendix A - Prior Implementations 1278 A.1. Apple Mac OS 8.x and 9.x. 1280 Mac OS chooses the IP address on a pseudo-random basis. The selected 1281 address is saved in persistent storage for continued use after 1282 reboot, when possible. 1284 Mac OS sends nine DHCPDISCOVER packets, with an interval of two 1285 seconds between packets. If no response is received from any of these 1286 requests (18 seconds), it will autoconfigure. 1288 Upon finding that a selected address is in use, Mac OS will select a 1289 new random address and try again, at a rate limited to no more than 1290 one attempt every two seconds. 1292 Autoconfigured Mac OS systems check for the presence of a DHCP server 1293 every five minutes. If a DHCP server is found but Mac OS is not 1294 successful in obtaining a new lease, it keeps the existing 1295 autoconfigured IP address. If Mac OS is successful at obtaining a 1296 new lease, it drops all existing connections without warning. This 1297 may cause users to lose sessions in progress. Once a new lease is 1298 obtained, Mac OS will not allocate further connections using the 1299 autoconfigured IP address. 1301 Mac OS systems do not send packets addressed to a Link-Local address 1302 to the default gateway if one is present; these addresses are always 1303 resolved on the local segment. 1305 Mac OS systems by default send all outgoing unicast packets with a 1306 TTL of 255. All multicast and broadcast packets are also sent with a 1307 TTL of 255 if they have a source address in the 169.254/16 prefix. 1309 Mac OS implements media sense where the hardware (and driver 1310 software) supports this. As soon as network connectivity is 1311 detected, a DHCPDISCOVER will be sent on the interface. This means 1312 that systems will immediately transition out of autoconfigured mode 1313 as soon as connectivity is restored. 1315 A.2. Apple Mac OS X Version 10.2 1317 Mac OS X chooses the IP address on a pseudo-random basis. The 1318 selected address is saved in memory so that it can be re-used during 1319 subsequent autoconfiguration attempts during a single boot of the 1320 system. 1322 Autoconfiguration of a Link-Local address depends on the results of 1323 the DHCP process. DHCP sends two packets, with timeouts of one and 1324 two seconds. If no response is received (three seconds), it begins 1325 autoconfiguration. DHCP continues sending packets in parallel for a 1326 total time of 60 seconds. 1328 At the start of autoconfiguration, it generates 10 unique random IP 1329 addresses, and probes each one in turn for 2 seconds. It stops 1330 probing after finding an address that is not in use, or the list of 1331 addresses is exhausted. 1333 If DHCP is not successful, it waits five minutes before starting over 1334 again. Once DHCP is successful, the autoconfigured Link-Local 1335 address is given up. The Link-Local subnet, however, remains 1336 configured. 1338 Autoconfiguration is only attempted on a single interface at any 1339 given moment in time. 1341 Mac OS X ensures that the connected interface with the highest 1342 priority is associated with the Link-Local subnet. Packets addressed 1343 to a Link-Local address are never sent to the default gateway, if one 1344 is present. Link-local addresses are always resolved on the local 1345 segment. 1347 Mac OS X implements media sense where the hardware and driver support 1348 it. When the network media indicates that it has been connected, the 1349 autoconfiguration process begins again, and attempts to re-use the 1350 previously assigned Link-Local address. When the network media 1351 indicates that it has been disconnected, the system waits four 1352 seconds before de-configuring the Link-Local address and subnet. If 1353 the connection is restored before that time, the autoconfiguration 1354 process begins again. If the connection is not restored before that 1355 time, the system chooses another interface to autoconfigure. 1357 Mac OS X by default sends all outgoing unicast packets with a TTL of 1358 255. All multicast and broadcast packets are also sent with a TTL of 1359 255 if they have a source address in the 169.254/16 prefix. 1361 A.3. Microsoft Windows 98/98SE 1363 Windows 98/98SE systems choose their IPv4 Link-Local address on a 1364 pseudo-random basis. This ensures that systems rebooting will obtain 1365 the same autoconfigured address, unless a conflict is detected. The 1366 address selection algorithm is based on computing a hash on the 1367 interface's MAC address, so that a large collection of hosts should 1368 obey the uniform probability distribution in choosing addresses 1369 within the 169.254/16 address space. 1371 When in INIT state, the Windows 98/98SE DHCP Client sends out a total 1372 of 4 DHCPDISCOVERs, with an inter-packet interval of 6 seconds. When 1373 no response is received after all 4 packets (24 seconds), it will 1374 autoconfigure an address. 1376 The autoconfigure retry count for Windows 98/98SE systems is 10. 1377 After trying 10 autoconfigured IPv4 addresses, and finding all are 1378 taken, the host will boot without an IPv4 address. 1380 Autoconfigured Windows 98/98SE systems check for the presence of a 1381 DHCP server every five minutes. If a DHCP server is found but 1382 Windows 98 is not successful in obtaining a new lease, it keeps the 1383 existing autoconfigured IPv4 Link-Local address. If Windows 98/98SE 1384 is successful at obtaining a new lease, it drops all existing 1385 connections without warning. This may cause users to lose sessions in 1386 progress. Once a new lease is obtained, Windows 98/98SE will not 1387 allocate further connections using the autoconfigured IPv4 Link-Local 1388 address. 1390 Windows 98/98SE systems with an IPv4 Link-Local address do not send 1391 packets addressed to an IPv4 Link-Local address to the default 1392 gateway if one is present; these addresses are always resolved on the 1393 local segment. 1395 Windows 98/98SE systems by default send all outgoing unicast packets 1396 with a TTL of 128. TTL configuration is performed by setting the 1397 Windows Registry Key 1398 HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services:\Tcpip\ 1399 Parameters\DefaultTTL of type REG_DWORD to the appropriate value. 1400 However, this default TTL will apply to all packets. While this 1401 facility could be used to set the default TTL to 255, it cannot be 1402 used to set the default TTL of IPv4 Link-Local packets to one (1), 1403 while allowing other packets to be sent with a TTL larger than one. 1405 Windows 98/98SE systems do not implement media sense. This means that 1406 network connectivity issues (such as a loose cable) may prevent a 1407 system from contacting the DHCP server, thereby causing it to auto- 1408 configure. When the connectivity problem is fixed (such as when the 1409 cable is re-connected) the situation will not immediately correct 1410 itself. Since the system will not sense the re-connection, it will 1411 remain in autoconfigured mode until an attempt is made to reach the 1412 DHCP server. 1414 The DHCP server included with Windows 98SE Internet Connection 1415 Sharing (ICS) (a NAT implementation) allocates out of the 192.168/16 1416 private address space by default. 1418 However, it is possible to change the allocation prefix via a 1419 registry key, and no checks are made to prevent allocation out of the 1420 IPv4 Link-Local prefix. When configured to do so, Windows 98SE ICS 1421 will NAT packets from the IPv4 Link-Local prefix off the local link. 1422 Windows 98SE ICS does not automatically route for the IPv4 Link-Local 1423 prefix, so that hosts obtaining addresses via DHCP cannot communicate 1424 with autoconfigured-only devices. 1426 Other home gateways exist that allocate addresses out of the IPv4 1427 Link-Local prefix by default. Windows 98/98SE systems can use a 1428 169.254/16 IPv4 Link-Local address as the source address when 1429 communicating with non-Link-Local hosts. However, Windows 98/98SE 1430 does not support router solicitation/advertisement. This means that 1431 Windows 98/98SE systems will typically not be able to discover a 1432 default gateway when in autoconfigured mode. 1434 A.4. Windows XP, 2000 and ME 1436 The autoconfiguration behavior of Windows XP, Windows 2000 and 1437 Windows ME systems is identical to Windows 98/98SE except in the 1438 following respects: 1440 Media Sense 1441 Router Discovery 1442 Silent RIP 1444 Windows XP, 2000 and ME implement media sense. As soon as network 1445 connectivity is detected, a DHCPREQUEST or DHCPDISCOVER will be sent 1446 on the interface. This means that systems will immediately 1447 transition out of autoconfigured mode as soon as connectivity is 1448 restored. 1450 Windows XP, 2000 and ME also support router discovery, although it is 1451 turned off by default. Windows XP and 2000 also support a RIP 1452 listener. This means that it is possible to discover a default 1453 gateway while in autoconfigured mode. 1455 ICS on Windows XP/2000/ME behaves identically to Windows 98SE with 1456 respect to address allocation and NATing of Link-Local prefixes. 1458 Intellectual Property Statement 1460 The IETF has been notified of intellectual property rights claimed in 1461 regard to some or all of the specification contained in this 1462 document. For more information consult the online list of claimed 1463 rights. 1465 The IETF takes no position regarding the validity or scope of any 1466 intellectual property or other rights that might be claimed to 1467 pertain to the implementation or use of the technology described in 1468 this document or the extent to which any license under such rights 1469 might or might not be available; neither does it represent that it 1470 has made any effort to identify any such rights. Information on the 1471 IETF's procedures with respect to rights in standards-track and 1472 standards-related documentation can be found in BCP-11. Copies of 1473 claims of rights made available for publication and any assurances of 1474 licenses to be made available, or the result of an attempt made to 1475 obtain a general license or permission for the use of such 1476 proprietary rights by implementers or users of this specification can 1477 be obtained from the IETF Secretariat. 1479 The IETF invites any interested party to bring to its attention any 1480 copyrights, patents or patent applications, or other proprietary 1481 rights that may cover technology that may be required to practice 1482 this standard. Please address the information to the IETF Executive 1483 Director. 1485 Full Copyright Statement 1487 Copyright (C) The Internet Society (2004). All Rights Reserved. 1488 This document and translations of it may be copied and furnished to 1489 others, and derivative works that comment on or otherwise explain it 1490 or assist in its implementation may be prepared, copied, published 1491 and distributed, in whole or in part, without restriction of any 1492 kind, provided that the above copyright notice and this paragraph are 1493 included on all such copies and derivative works. However, this 1494 document itself may not be modified in any way, such as by removing 1495 the copyright notice or references to the Internet Society or other 1496 Internet organizations, except as needed for the purpose of 1497 developing Internet standards in which case the procedures for 1498 copyrights defined in the Internet Standards process must be 1499 followed, or as required to translate it into languages other than 1500 English. The limited permissions granted above are perpetual and 1501 will not be revoked by the Internet Society or its successors or 1502 assigns. This document and the information contained herein is 1503 provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE 1504 INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR 1505 IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 1506 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1507 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1509 Open Issues 1511 Open issues with this specification are tracked on the following web 1512 site: 1514 http://www.drizzle.com/~aboba/ZEROCONF/issues.html