idnits 2.17.1 draft-rja-ilnp-intro-06.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, you can and should remove the disclaimer; otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (18 August 2010) is 4999 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) ** Obsolete normative reference: RFC 3484 (Obsoleted by RFC 6724) == Outdated reference: A later version (-11) exists of draft-rja-ilnp-nonce-05 == Outdated reference: A later version (-11) exists of draft-rja-ilnp-dns-06 == Outdated reference: A later version (-11) exists of draft-rja-ilnp-icmp-04 -- Obsolete informational reference (is this intentional?): RFC 1631 (Obsoleted by RFC 3022) -- Obsolete informational reference (is this intentional?): RFC 3775 (Obsoleted by RFC 6275) -- Obsolete informational reference (is this intentional?): RFC 4941 (Obsoleted by RFC 8981) Summary: 2 errors (**), 0 flaws (~~), 4 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Draft R. Atkinson 3 draft-rja-ilnp-intro-06.txt Consultant 4 Expires: 18 FEB 2011 18 August 2010 5 Category: Experimental 7 ILNP Concept of Operations 8 draft-rja-ilnp-intro-06.txt 10 Status of this Memo 12 Distribution of this memo is unlimited. 14 Copyright (c) 2010 IETF Trust and the persons identified as the 15 document authors. All rights reserved. 17 This document is subject to BCP 78 and the IETF Trust's Legal 18 Provisions Relating to IETF Documents 19 (http://trustee.ietf.org/license-info) in effect on the date of 20 publication of this document. Please review these documents 21 carefully, as they describe your rights and restrictions with 22 respect to this document. Code Components extracted from this 23 document must include Simplified BSD License text as described in 24 Section 4.e of the Trust Legal Provisions and are provided 25 without warranty as described in the Simplified BSD License. 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 This document may contain material from IETF Documents or IETF 31 Contributions published or made publicly available before 32 November 10, 2008. The person(s) controlling the copyright in 33 some of this material may not have granted the IETF Trust the 34 right to allow modifications of such material outside the IETF 35 Standards Process. Without obtaining an adequate license from 36 the person(s) controlling the copyright in such materials, this 37 document may not be modified outside the IETF Standards Process, 38 and derivative works of it may not be created outside the IETF 39 Standards Process, except to format it for publication as an 40 RFC or to translate it into languages other than English. 42 Internet-Drafts are working documents of the Internet 43 Engineering Task Force (IETF), its areas, and its working 44 groups. Note that other groups may also distribute working 45 documents as Internet-Drafts. 47 Internet-Drafts are draft documents valid for a maximum of six 48 months and may be updated, replaced, or obsoleted by other 49 documents at any time. It is inappropriate to use 50 Internet-Drafts as reference material or to cite them other 51 than as "work in progress." 53 The list of current Internet-Drafts can be accessed at 54 http://www.ietf.org/1id-abstracts.html 56 The list of Internet-Draft Shadow Directories can be accessed at 57 http://www.ietf.org/shadow.html 59 This document is not on the IETF standards-track and does not 60 specify any level of standard. This document merely provides 61 information for the Internet community. 63 This document has had extensive review within the IRTF Routing 64 Research Group, and is part of the ILNP document set. ILNP is 65 one of the recommendations made by the RG Chairs. Separately, 66 various refereed research papers on ILNP have also been published 67 during this decade. So the ideas contained herein have had much 68 broader review than the IRTF Routing RG. The views in this 69 document were considered controversial by the Routing RG, 70 but the RG reached a consensus that the document still should be 71 published. The Routing RG has had remarkably little consensus 72 on anything, so virtually all Routing RG outputs are considered 73 controversial. 75 Abstract 77 This document describes the Concept of Operations for the 78 Identifier Locator Network Protocol (ILNP), which is an 79 experimental extension to IP. This is a product of the 80 IRTF Routing RG. 82 Table of Contents 84 1. Introduction ...............................................2 85 2. Locators & Identifiers......................................4 86 3. Transport Protocols.........................................8 87 4. Mobility....................................................9 88 5. Multi-Homing...............................................12 89 6. Localised Addressing.......................................13 90 7. IP Security Enhancements...................................14 91 8. DNS Enhancements...........................................15 92 9. Referrals & Application Programming Interfaces.............17 93 10. Backwards Compatibility....................................18 94 11. Incremental Deployment.....................................19 95 12. Implementation Considerations..............................20 96 13. Security Considerations ...................................21 97 14. IANA Considerations .......................................26 98 15. References ................................................26 100 1. INTRODUCTION 102 At present, the research and development community are exploring 103 various approaches to evolving the Internet Architecture. 104 Several different classes of evolution are being considered. One 105 class is often called "Map and Encapsulate", where traffic would 106 be mapped and then tunnelled through the inter-domain core of the 107 Internet. Another class being considered is sometimes known as 108 "Identifier/Locator Split". This document relates to a proposal 109 that is in the latter class of evolutionary approaches. 111 There has been substantial research relating to naming in the 112 Internet through the years. [IEN 1] [IEN 19] [IEN 23] [IEN 31] 113 [RFC 814] [RFC 1498] More recently, mindful of that important 114 prior work, and starting well before the Routing RG was 115 re-chartered to focus on inter-domain routing scalability, the 116 author has been examining enhancements to certain naming aspects 117 of the Internet Architecture. [MobiArch07] [MobiWAC07] 118 [MobiArch08] [MILCOM08] [MILCOM09] [TeleSys] 120 The architectural concept behind ILNP derives originally from a 121 June 1994 note by Bob Smart to the IETF SIPP WG mailing list. 122 [SIPP94] In January 1995, Dave Clark sent a note to the IETF IPng 123 WG mailing list suggesting that the IPv6 address be split into 124 separate Identifier and Locator fields. [IPng95] 126 Afterwards, Mike O'Dell pursued this concept in Internet-Drafts 127 describing "8+8" or "GSE".[8+8] [GSE] More recently, the IRTF 128 Namespace Research Group (NSRG) studied this matter. Unusually 129 for an IRTF RG, the NSRG operated on the principle that unanimity 130 was required for the NSRG to make a recommendation. The author 131 was a member of the IRTF NSRG. At least one other proposal, the 132 Host Identity Protocol (HIP), also derives in part from the IRTF 133 NSRG studies (and related antecedent work). This current 134 proposal differs from O'Dell's work in various ways. 136 The crux of this proposal is to have different names for the 137 identity of a node and the location of its subnet, with crisp 138 semantics for each. This enhances the Internet Architecture 139 by adding crisp and clear semantics for the Identifier and 140 for the Locator, removing the semantically-muddled concept 141 of the IP address, and updating end system protocols slightly, 142 without requiring router changes. 144 With these naming enhancements, we have improved the Internet 145 Architecture by adding explicit support not only for 146 multi-homing, but also for mobility, localised addressing 147 (e.g. NAT/NAPT), and IP Security. 149 ILNP is an architecture, and can have more than one engineering 150 instantiation. The term ILNPv4 refers precisely to an instance 151 of ILNP that is based upon and backwards compatible with IPv4. 152 The term ILNPv6 refers precisely to an instance of ILNP that is 153 based upon and backwards compatible with IPv6. The following 154 two subsections provide brief overview of ILNPv6 and ILNPv4, 155 respectively. A full specification for either ILNPv4 or ILNPv6 156 is beyond the scope of this document. 158 Readers are referred to other related ILNP documents for details 159 not described here. [ILNP-DNS] describes additional DNS resource 160 records that support the Identifier/Locator split mode of 161 operation. [ILNP-ICMP] describes a new ICMPv6 Locator Update 162 message used by an ILNP node to inform its correspondent nodes 163 or any changes to its set of valid Locators. [ILNP-Nonce] 164 describes a new IPv6 Nonce Destination Option used by ILNP 165 nodes (1) to indicate to ILNP correspondent nodes (by inclusion 166 within the initial packets of an ILNP session) that the node 167 is operating in the Identifier/Locator split mode and 168 (2) to authenticate ICMP messages, for example the ICMPv6 169 Locator Update message, that are exchanged with ILNP 170 correspondent nodes. 172 1.1 Terminology 174 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL 175 NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and 176 "OPTIONAL" in this document are to be interpreted as described 177 in RFC 2119. [RFC 2119] 179 2. LOCATORS & IDENTIFIERS 181 ILNP deprecates the semantically muddled concept of an "IP 182 Address" and replaces it with 2 new concepts, the "Locator" 183 and the "Identifier". 185 The Locator is used only to name the subnetwork a node is 186 connected to, while the Identifier is used only for node 187 identity. So the routing system uses Locators, while 188 upper-layer protocols (e.g. TCP/UDP pseudo-header checksum, 189 IPsec Security Association) use only the Identifier. 191 The same Identifier definition is used for both ILNPv4 and 192 ILNPv6. This is described in the next sub-section. Following 193 that is a description of ILNPv6, including a description of 194 the 64-bit Locator value used with ILNPv6. Then, there is a 195 description of ILNPv4, including a description of the 32-bit 196 Locator value used with ILNPv4. 198 2.1 Identifiers 200 With ILNP, the Identifier is an unsigned 64-bit number. 202 This provides a fixed-length non-topological name for a node. 203 Identifiers are bound to nodes, not to interfaces of a node. 204 All ILNP Identifiers MUST comply with the modified EUI-64 syntax 205 already specified for IPv6's "Interface Identifier" values. 206 [RFC 2460][RFC 4219][IEEE-EUI] 208 Identifiers have either global-scope or local-scope. 209 A reserved bit in the modified EUI-64 syntax clearly 210 indicates whether a given Identifier has global-scope or 211 local-scope.[RFC 4219][IEEE-EUI] A node is not required 212 to use a global-scope Identifier, although that is the 213 recommended practice. 215 Most commonly, Identifiers have global-scope and are derived 216 from one or more IEEE 802 or IEEE 1394 'MAC Addresses' (sic) 217 already associated with the node, following the procedure 218 already defined for IPv6.[RFC 4219] Global-scope identifiers 219 have a high probability of being globally unique. This approach 220 eliminates the need to manage Identifiers, among other benefits. 222 Local-scope Identifiers MUST be unique within the context of 223 their Locators. The existing mechanisms of the IPv4 Address 224 Resolution Protocol [RFC 826] and IPv6 Neighbour Discovery 225 Protocol [RFC 4861] automatically enforce this constraint. 227 For example, on an Ethernet-based IPv4 subnetwork the ARP Reply 228 message is sent via link-layer broadcast, thereby advertising 229 the current binding between an IPv4 address and a MAC address 230 to all nodes on that IPv4 subnetwork. (Note also that a 231 well-known, long standing, issue with ARP is that it cannot be 232 authenticated.) Local-scope Identifiers MUST NOT be used with 233 other Locators without first ensuring uniqueness in the context 234 of those other Locators (e.g. by using IPv6 Neighbour 235 Discovery's Duplicate Address Detection mechanism when using 236 ILNPv6 or by sending an ARP Request when using ILNPv4). 238 Other methods might be used to generate local-scope Identifiers. 239 For example, one might derive Identifiers using some form of 240 cryptographic generation or using the methods specified in the 241 IPv6 Privacy Extensions to State-Less Address Auto-Configuration 242 (SLAAC). [RFC 3972, RFC 4941] When cryptographic generation of 243 Identifiers using methods described in RFC-3972 is in use, only 244 the Identifier is included, never the Locator, thereby preserving 245 roaming capability. [RFC 3972] One could also imagine creating 246 a local-scope Identifier by taking a cryptographic hash of a 247 node's public key. Of course, in the very unlikely event of a 248 Identifier collision, for example when a node has chosen to use 249 a local-scope Identifier value, the node remains free to use 250 some other local-scope Identifier value(s). 252 2.2 ILNPv6 254 It is worth remembering here that an IPv6 address names a 255 specific network interface on a specific node, but an ILNP 256 Identifier names the node itself, not a specific interface 257 on the node. This difference in definition is essential 258 to providing seamless support for mobility and multi-homing, 259 which are discussed in more detail later in this note. 261 1 1 2 3 262 0 4 8 2 6 4 1 263 +---------------+-----------------+----------------+---------------+ 264 | Version| Traffic Class | Flow Label | 265 +---------------+-----------------+----------------+---------------+ 266 | Payload Length | Next Header | Hop Limit | 267 +---------------+-----------------+--------------------------------+ 268 | Source Address | 269 + + 270 | | 271 + + 272 | | 273 + + 274 | | 275 +---------------+-----------------+----------------+---------------+ 276 | Destination Address | 277 + + 278 | | 279 + + 280 | | 281 + + 282 | | 283 +---------------+-----------------+----------------+---------------+ 285 Figure 1: Existing ("Classic") IPv6 Header 287 The high-order 64-bits of the IPv6 address become the Locator. 288 The Locator indicates the subnetwork point of attachment for a 289 node. In essence, the Locator names a subnetwork. Locators are 290 also known as Routing Prefixes. Of course, backwards 291 compatibility requirements mean that ILNPv6 Locators use the same 292 number space as IPv6 routing prefixes. This ensures that no 293 changes are needed to deployed IPv6 routers when deploying 294 ILNPv6. 296 The low-order 64-bits of the IPv6 address become the Identifier. 297 Details of the Identifier were discussed just above. 299 1 1 2 3 300 0 4 8 2 6 4 1 301 +---------------+-----------------+----------------+---------------+ 302 | Version| Traffic Class | Flow Label | 303 +---------------+-----------------+----------------+---------------+ 304 | Payload Length | Next Header | Hop Limit | 305 +---------------+-----------------+----------------+---------------+ 306 | Source Locator | 307 + + 308 | | 309 +---------------+-----------------+----------------+---------------+ 310 | Source Identifier | 311 + + 312 | | 313 +---------------+-----------------+----------------+---------------+ 314 | Destination Locator | 315 + + 316 | | 317 +---------------+-----------------+----------------+---------------+ 318 | Destination Identifier | 319 + + 320 | | 321 +---------------+-----------------+----------------+---------------+ 323 Figure 2: ILNPv6 Header 325 2.3 ILNPv4 327 ILNPv4 is merely a different instantiation of the ILNP 328 Architecture, so it retains the crisp distinction between the 329 Locator and the Identifier. Also, as with ILNPv6, when ILNPv4 330 is used for a network-layer session, the upper-layer protocols 331 (e.g. TCP/UDP pseudo-header checksum, IPsec Security 332 Association) bind only to the Identifiers, never to the Locators. 334 As with ILNPv6, only the Locator values are used for routing 335 ILNPv4 packets. 337 Just as ILNPv6 is carefully engineered to be backwards- 338 compatible with IPv6, ILNPv4 is carefully engineered 339 to be backwards-compatible with IPv4. 341 The Source IP Address in the IPv4 header becomes the Source 342 ILNPv4 Locator value, while the Destination IP Address of the 343 IPv4 header becomes the Destination ILNPv4 Locator value. Of 344 course, backwards compatibility requirements mean that ILNPv4 345 Locators use the same number space as IPv4 routing prefixes. 347 ILNPv4 uses the same 64-bit Identifier, with the same modified 348 EUI-64 syntax, as ILNPv6. Because the IPv4 address is much 349 smaller than the IPv6 address, ILNPv4 cannot carry the 350 Identifier values in the fixed portion of the IPv4 header. 351 The obvious two ways to carry the ILNP Identifier with ILNPv4 352 are either as an IPv4 Option or as an IPv6-style Extension 353 Header placed after the IPv4 header and before the upper-layer 354 protocol (e.g. OSPF, TCP, UDP, SCTP). 356 At least some currently available IPv4 forwarding silicon is able 357 to parse past IPv4 options to examine the upper-layer protocol 358 header at wire-speed on reasonably fast (e.g. 1 Gbps or better) 359 network interfaces. By contrast, no existing silicon is able to 360 parse past a new Extension Header at all. So, for engineering 361 reasons, ILNPv4 uses a new IPv4 Option to carry the Identifier 362 values. The new IPv4 option also carries a nonce value, 363 performing the same function for ILNPv4 as the IPv6 Nonce 364 Destination Option [ILNP-Nonce] performs for ILNPv6. 366 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 367 |Version|IHL=12 |Type of Service| Total Length | 368 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 369 | Identification |Flags| Fragment Offset | 370 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 371 | Time to Live | Protocol | Header Checksum | 372 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 373 | Source Locator | 374 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 375 | Destination Locator | 376 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 377 | OT=ILNPv4_ID | OL=5 | Padding=0x0000 | 378 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 379 | | 380 + Source Identifier + 381 | | 382 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 383 | | 384 + Destination Identifier + 385 | | 386 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 387 |OT=ILNPv4_NONCE| OL=2 | top 16 bits of nonce | 388 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 389 | lower 32 bits of nonce | 390 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 392 Figure 3: ILNPv4 header with ILNP ID option 393 and ILNP Nonce option. 395 Notation for Figure 3: 396 IHL: Internet Header Length 397 OT: Option Type 398 OL: Option Length 400 The remainder of this note focuses on ILNP for IPv6, in the 401 interest of both clarity and brevity, however the same 402 architectural concepts and principles also apply to ILNP 403 for IPv4, albeit with slightly different engineering. 405 3. TRANSPORT PROTOCOLS 407 At present, commonly deployed transport protocols include a 408 pseudo-header checksum that includes certain network-layer 409 fields, the IP addresses used for the session, in its 410 calculation. This inclusion of network-layer information 411 within the transport-layer session state creates issues for 412 multi-homing, mobility, IP Security, and localised addressing 413 (e.g. using Network Address Translation). [RFC 1631][RFC 3022] 415 This unfortunate aspect of the TCP pseudo-header checksum 416 has been understood to be an architectural problem at least 417 since 1977, well before the transition from NCP to 418 IPv4.[IEN 1][IEN 19][IEN 23][IEN 31][RFC 1498] 420 With this proposal, transport protocols include only the 421 Identifier in their pseudo-header calculations, but do not 422 include the Locator in their pseudo-header calculations. 424 To minimise the changes required within transport protocol 425 implementations, when this proposal is in use for a 426 communications session, the Locator fields are zeroed for 427 the purpose of transport-layer pseudo-header calculations. 429 Later in this document, methods for incremental deployment 430 of this change and backwards compatibility with non-upgraded 431 nodes are described. 433 4. MOBILITY 435 First, please recall that mobility and multi-homing actually 436 present the same set of issues. In each case, the set of 437 Locators associated with a node or site changes. The reason 438 for the change might be different, but the effects on the 439 network and on correspondents is identical. 441 There are no standardised mechanisms to update most transport 442 protocols with new IP addresses in use for the session. 443 Exceptionally, the Stream Control Transport Protocol (SCTP) 444 recently added this capability.[RFC 5061] In July 2008, Mark 445 Handley at UCL proposed adding such a capability to TCP during 446 a presentation at the IRTF Routing RG in Dublin, Ireland. 447 His Multi-Path TCP concept is being considered in the IETF 448 as of this writing. 450 This creates various issues for mobility. For example, there 451 is no method at present to update the IP addresses associated 452 with a transport layer session when one of the nodes in that 453 session moves (i.e. changes one of its points of network 454 attachment). 456 So, the several approaches to IP mobility seek to hide the 457 change in location (and corresponding change in IP addresses) 458 via tunnelling, home agents, foreign agents, and so forth. 459 [RFC 3775] All of this can add substantial complexity to IP 460 mobility approaches, both in the initial deployment and also 461 in ongoing operation. 463 By contrast, this ILNP proposal hides each node's location 464 information from the transport layer protocols at all times, 465 by removing location information from the transport session 466 state (e.g. pseudo-header checksum calculations). 468 In this proposal, mobility and multi-homing are supported 469 using a common set of mechanisms. In both cases, different 470 Locator values are used to identify different IP subnetworks. 471 Also, ILNP nodes are assumed to have a Fully Qualified 472 Domain Name (FQDN) stored in the Domain Name System (DNS), 473 as is already done within the deployed Internet. 475 To handle the move of a node, we add a new ICMP control message. 476 The ICMP Locator Update message is used by a node to inform its 477 existing correspondents that the set of valid Locators for the 478 node has changed. This mechanism can be used to add newly valid 479 Locators, to remove no longer valid Locators, or to do both at 480 the same time. Further, the node uses Secure Dynamic DNS Update 481 [RFC 3007] to correct the set of Locator (i.e. L32, L64) records 482 in the DNS for that node.[ILNP-DNS] This enables any new 483 correspondents to correctly initiate a new session with the 484 node at its new location. 486 This use of DNS for initial rendezvous with mobile node was 487 independently proposed by others [PHG02] and then separately 488 re-invented by the current author later on. 490 (The Locator Update control message could be an entirely new 491 protocol running over UDP, for example, but there is no obvious 492 advantage to creating a new protocol rather than using a new 493 ICMP message.) 495 With ILNP, network mobility (as well as node mobility) 496 is considered a special case of multihoming. That is, 497 when a network moves, it uses a new Locator value for all 498 of its communications sessions. So, the same mechanism, 499 using a new or additional Locator value, also supports 500 network mobility. Similarly, when a multi-homed site 501 or multi-homed node changes its set of upstream links, 502 the Locators associated with that site or node change. 504 So in ILNP, when a connectivity change affects the set of 505 valid Locators, the affected node(s) actively: 507 (1) use the ICMP Locator Update message to inform their 508 existing correspondents with the updated information 509 about their currently valid Locator(s). [ILNP-ICMP] 511 AND also 513 (2) update their DNS entries, most commonly by using 514 the Secure Dynamic DNS Update mechanism. [RFC 3007] 516 In the unlikely event of simultaneous motion which changes 517 both nodes' Locators within a very small time period, a 518 node can use the DNS to discover the new Locator value(s) 519 for the other node. 521 As a DNS performance optimisation, the "LP" DNS resource record 522 MAY be used to avoid requiring each node on a subnetwork to 523 update its DNS L64 record entries when that subnetwork's location 524 (e.g. upstream connectivity) changes. In this case, the nodes on 525 the subnetwork each would have an "LP" record pointing to a 526 common domain-name used to name that subnetwork. In turn, that 527 subnetwork's domain name would have one or more L64 record(s) in 528 the DNS. Since the contents of an "LP" record are stable, 529 relatively long DNS TTL values can be associated with these 530 records facilitating DNS caching. By contrast, the DNS TTL 531 of an L32 or L64 record for a mobile or multi-homed node 532 should be small. Experimental work at the University of 533 St Andrews indicates that the DNS continues to work well 534 even with very low DNS TTL values. [Bhatti10] 536 Correspondents of a node on that subnetwork would perform a "L64" 537 record query for that target node (or an "ID" query for that 538 target node) and receive the "LP" records as additional data in 539 the DNS reply. Then the correspondent would perform an L64 540 record lookup on the domain-name pointed to by that LP record, 541 in order to learn the Locator value to use to reach that 542 target node. 544 For bi-directional flows, such as a TCP session, each node knows 545 whether the current path in use is working by the reception of 546 data packets, acknowledgements, or both. As with TCP/IP, 547 TCP/ILNP does not need special path probes. UDP/ILNP sessions 548 with acknowledgements work similarly, and also don't need special 549 path probes. 551 In the deployed Internet, the sending node for a UDP/IP session 552 without acknowledgements does not know for certain that all 553 packets are received by the intended receiving node. Such 554 UDP/ILNP sessions fare no worse than UDP/IP sessions. The 555 receiver(s) of such a UDP session SHOULD send a gratuitous 556 IP packet containing an ILNP Nonce option to the sender, 557 in order to enable the receiver to subsequently send ICMP 558 Locator Updates if appropriate. [ILNP-Nonce] In this case, 559 UDP/ILNP sessions fare better than UDP/IP sessions, 560 still without using network path probes. 562 One might wonder what happens if a mobile node is moving more 563 quickly than DNS can be updated. This situation is unlikely, 564 particularly given the widespread use of link-layer mobility 565 mechanisms (e.g. GSM, IEEE 802 bridging) in combination with 566 network-layer mobility. However, the situation is functionally 567 equivalent to the situation where a traditional IP node is moving 568 faster than the Mobile IPv4 or Mobile IPv6 agents/servers can be 569 updated with the mobile node's new location. So the issue is not 570 new in any way. In all cases, Mobile IPv4 and Mobile IPv6 and 571 ILNP, a node moving that quickly might be temporarily unreachable 572 until it remains at a given network-layer location (e.g. IP 573 subnetwork) long enough for the location update mechanisms (for 574 Mobile IPv4, for Mobile IPv6, or ILNP) to catch up. ILNP is 575 prospectively better than either form of Mobile IP with respect 576 to key management, given that ILNP is using Secure Dynamic DNS 577 Update and given that neither form of Mobile IP has a 578 standardised and scalable key management mechanism at present. 580 5. MULTI-HOMING 582 Conceptually, there are two kinds of multi-homing. Site 583 multi-homing is when all nodes at a site are multi-homed at the 584 same time. This is what most people mean when they talk about 585 multi-homing. However, there is also a separate concept of node 586 multi-homing, where only a single node is multi-homed. Kindly 587 recall, that multiple transport-layer sessions might currently 588 share a single current network-layer (e.g. IP or ILNP) session. 590 5.2 Node Multi-Homing 592 At present, node multi-homing is not common in the deployed 593 Internet. When TCP or UDP are in use for an IP session, node 594 multi-homing cannot provide session resilience, because the 595 transport pseudo-header checksum binds the session to a single 596 address of the multi-homed node, and hence to a single interface. 597 SCTP has a protocol-specific mechanism to support node 598 multi-homing; SCTP can support session resilience both at present 599 and also without change in the proposed approach. [RFC 5061] 601 In the new scheme, when a node is multi-homed, then the node 602 typically has more than one valid Locator value. When one 603 upstream connection fails, the node sends an ICMP Locator Update 604 message to each existing correspondent node to remove the 605 no-longer-valid Locator from the set of valid Locators. 606 [ILNP-ICMP] Also, the node can use Secure Dynamic DNS Update 607 to alter the set of currently valid L64 records associated with 608 that node. [RFC 3007] This second step ensures that any new 609 correspondents can reach the node. 611 5.2 Site Multi-Homing 613 At present, site multi-homing is common in the deployed Internet. 614 This is primarily achieved by advertising the site's routing 615 prefix(es) to more than one upstream Internet service provider 616 at a given time. In turn, this requires de-aggregation of 617 routing prefixes within the inter-domain routing system. 618 In turn, this increases the entropy of the inter-domain 619 routing system (e.g. RIB/FIB size increases beyond the 620 minimal RIB/FIB size that would be required to reach all sites). 622 In the new scheme, site multi-homing is similar to node 623 multi-homing, but with nodes within the site having one Locator 624 for each upstream connection to the Internet. To avoid a DNS 625 Update burst when a site or (sub)network moves location, a DNS 626 record optimisation is possible. This would change the number of 627 DNS Updates required from Order(number of nodes at the 628 site/subnetwork that moved) to Order(1). [ILNP-DNS] 630 Additionally, since the transport-protocol session state no 631 longer includes the Locators, a site could choose to perform 632 Locator rewriting at its site border routers, possibly in 633 combination with applying site traffic engineering policy on 634 which upstream link to use for which packets. Since the site 635 border router(s) are in the middle of any exterior packet flow, 636 they also can send proxy Locator Update messages on behalf of 637 nodes inside that site, and can even include the appropriate 638 Nonce value in such proxy Locator Updates, if desired by that 639 site's administration. 641 5.3 Multi-Path Support 643 Because ILNP decouples the transport-layer information from 644 the Locator values being used for a given session (e.g. TCP 645 pseudo-header checksum includes Identifier values, but not 646 Locator values, when ILNP is in use), ILNP can enable 647 multi-path transport-layer sessions without requiring any 648 changes to existing transport-layer protocols (e.g. TCP, UDP). 649 Note that this approach also does not interfere with SCTP's 650 existing support for multi-path transport nor with the 651 proposed TCP multi-path extensions. 653 With ILNP, any transport-layer session can use multiple paths 654 concurrently, simply by using multiple (valid) Locator values 655 in that session's ILNP packets. Obviously for any given ILNP 656 packet a single Source Locator and a single Destination Locator 657 is in use. 659 As an example, if one considers TCP with an originator using 660 Locators (A, B, C) and a responder using Locators (W, X, Y), then 661 the originator can choose which Source Locators to use and also 662 which Destination Locators on a packet-by-packet basis. So 663 different TCP segments (or TCP ACKs, or other TCP information) 664 within a single TCP session can use different Locator pairs. 666 Again, purely as an example, the originator could send 667 packets using these Locator values in this simple sequence: 668 (A, W) 669 (B, X) 670 (C, Y) 671 or any other sequence that it wishes to. Similarly, the 672 responder can use any valid combination of Locators that it 673 wishes to use. 675 In any case, the TCP implementations at either end are unaware 676 that multiple Locators are being used (i.e. because the 677 transport-layer pseudo-header checksum only includes Identifiers, 678 never Locators). In turn, this is why not special multi-path 679 TCP (or UDP or SCTP or other) transport-layer modifications 680 are required. (Caveat: Of course, the ILNP stack upgrade is 681 needed in the first place.) 683 The same concepts and general approach also apply to UDP and/or SCTP. 685 6. LOCALISED ADDRESSING 687 As the Locator value no longer forms part of the node session 688 state (e.g. TCP pseudo-header), it is easier to support 689 localised addressing, which is sometimes also called "Private 690 Addressing", based on the use of local values of the Locator. 691 This would be either in place of, or to supplement, existing 692 NAT-based schemes. [RFC 1631] [RFC 3022] 694 For example, a site that desires to use private addresses 695 internally might deploy IPv6 Unique Local Addressing 696 (ULA) for localised addressing, along with some form of ILNP/ 697 IPv6 Network Address Translation at a site border gateway. 698 [ID-ULA] [RFC 4193] This example is described in detail 699 in [MILCOM09], both as a mechanism for site multi-homing 700 and also as a mechanism to support site-controlled traffic 701 engineering. 703 In the simplest case, an ILNP capable NAT only would need to 704 change the value of the Source Locator in an outbound packet, 705 and the value of the Destination Locator for an inbound packet. 706 Identifier values would not need to change, nor would 707 transport-layer checksums, so a true end-to-end session 708 could be maintained. 710 If a site using localised addressing chooses to deploy a 711 split-horizon DNS server, then the DNS server would advertise 712 the global-scope Locator(s) of the site border routers outside 713 the site to DNS clients outside the site, and would advertise 714 the local-scope Locator(s) specific to that internal node to 715 DNS clients inside the site. Such deployments of split-horizon 716 DNS servers are not unusual in the IPv4 Internet today. If an 717 internal node (e.g. portable computer) moves outside the site, 718 it would follow the normal ILNP methods to update its 719 authoritative DNS server with its current Locator set. In this 720 deployment model, the authoritative DNS server for that mobile 721 device will be either the split-horizon DNS server itself or the 722 master DNS server providing data to the split-horizon DNS server. 724 If a site using localised addressing chooses not to deploy a 725 split-horizon DNS server, then all internal nodes would 726 advertise the global-scope Locator(s) of the site border routers. 727 To deliver packets from one internal node to another internal 728 node, the site would either choose to use layer-2 bridging 729 (e.g. IEEE Spanning Tree, IEEE Rapid Spanning Tree, or a 730 link-state layer-2 algorithm such as the IETF TRILL group or 731 IEEE 802.1 are developing), or the interior routers would 732 forward packets up to the nearest site border router, 733 which in turn would then rewrite the Locators to appropriate 734 local-scope values, and forward the packet towards the interior 735 destination node. 737 Alternately, for sites using localised addressing but not 738 deploying a split-horizon DNS server, the DNS server could 739 return all global-scope and local-scope Locators to all queriers, 740 and to assume that correspondent hosts would use address 741 selection to choose the best Locator to use to reach a given 742 correspondent. [RFC 3484] Hosts within the same site as the 743 correspondent node would only have a ULA configured, and hence 744 would select the ULA destination Locator for the correspondent. 745 Hosts outside the site would not have the same ULA configured. 746 Note that RFC 3484 probably needs to be updated to indicate 747 that the longest-prefix matching rule is inadequate when 748 comparing ULA-based Locators with global-scope Locators: 749 to choose a ULA for a correspondent, a node must have a 750 Locator that matches all 48 ULA bits of the target Locator 751 value. 753 We note that a deployment using private/local addressing can 754 also provide site multi-homing by deploying site border 755 routers in this manner. 757 Please note that with this proposal, localised addressing 758 (e.g. using Network Address Translation on the Locator bits) 759 would work in harmony with multihoming, mobility, and IP 760 Security.[MobiWAC07][MILCOM08][MILCOM09] 762 7. IP SECURITY ENHANCEMENTS 764 A current issue is that the IP Security protocols, AH and ESP, 765 have Security Associations that include the IP addresses of 766 the secure session endpoints. This was understood to be a 767 problem when AH and ESP were originally defined, however the 768 limited set of namespaces in the Internet Architecture did not 769 provide any better choices at that time. 771 Operationally, this binding causes problems for the use of the 772 IPsec protocols through Network Address Translation devices, 773 with mobile nodes (because the mobile node's IP address changes 774 at each network-layer handoff), and with multi-homed nodes 775 (because the session is bound to a particular interface of the 776 multi-homed node, rather than being bound to the node 777 itself).[RFC 3027][RFC 3715] 779 To resolve the issue of IPsec interoperability through a 780 NAT deployment, UDP encapsulation of IPsec is commonly 781 used today.[RFC 3948] 783 With this proposal, the IP Security protocols, AH and ESP, 784 are enhanced to bind Security Associations only to 785 Identifier values and never to Locator values (and also 786 not to an entire 128-bit IPv6 address). 788 Similarly, key management protocols used with IPsec would be 789 enhanced to deprecate use of IP addresses as identifiers and 790 to substitute the use of the new Identifier for that 791 purpose. 793 This small change enables IPsec to work in harmony with 794 multihoming, mobility, and localised addressing. [MILCOM08] 795 [MILCOM09] Further, it would obviate the need for specialised 796 IPsec NAT Traversal mechanisms, thus simplifying IPsec 797 implementations while enhancing deployability and 798 interoperability. [RFC 3948] 800 This change does not reduce the security provided by the 801 IP Security protocols. 803 8. DNS ENHANCEMENTS 805 As part of this proposal, additional DNS Resource Records have 806 been proposed in a separate document. [ILNP-DNS] These new 807 records store the Identifier and Locator values for nodes that 808 have been upgraded to support the Identifier-Locator Split Mode. 810 With this proposal, mobile or multi-homed nodes and sites are 811 expected to use the existing "Secure Dynamic DNS Update" protocol 812 to keep their Identifier and Locator records correct in its 813 authoritative DNS server(s). [RFC 3007] 815 While some might be surprised, Secure Dynamic DNS Update is 816 available now in a very wide range of existing deployed systems. 817 For example, Microsoft Windows XP (and later versions), the 818 freely distributable BIND DNS software package (used in Apple 819 MacOS X and in most UNIX systems), and the commercial Nominum DNS 820 server all implement support for Secure Dynamic DNS Update and 821 are known to interoperate. [Liu-DNS] There are credible reports 822 that when a site deploys Microsoft's Active Directory, the site 823 (silently) automatically deploys Secure Dynamic DNS 824 Update. [Liu-DNS] So it appears that many sites have already 825 deployed Secure Dynamic DNS Update even though they might not be 826 aware they have already deployed that protocol. [Liu-DNS] 828 Reverse DNS lookups, to find a node's Fully Qualified Domain Name 829 from the combination of a Locator and related Identifier value, 830 can be performed as at present. 832 Previous research by others indicates that DNS caching is largely 833 ineffective, with the exception of NS records and the addresses 834 of DNS servers referred to by NS records.[SBK2002] This means DNS 835 caching performance will not be adversely affected by assigning 836 very short time-to-live (TTL) values to the Locator records of 837 typical nodes.[Bhatti10] It also means that it is preferable 838 to deploy the DNS server function on nodes that have longer 839 DNS TTL values, rather than on nodes that have shorter DNS 840 TTL values. 842 As discussed previously, LP records normally are stable, 843 even if the L32 or L64 records they point to aren't stable, 844 so LP records normally can be given very long DNS TTL values. 846 Identifier values might be very long-lived (e.g. days) when they 847 have been generated from an IEEE MAC Address on the system. 848 Identifier values might have a shorter lifetime (e.g. hours) if 849 they have been cryptographically-generated [RFC 3972], or have 850 been created by the IPv6 Privacy Extensions [RFC 4941], or 851 otherwise have the EUI-64 scope bit at the "local-scope" value. 852 Note that when ILNP is used, the cryptographic generation 853 method described in RFC-3972 is used only for the Identifier, 854 omitting the Locator, thereby preserving roaming capability. 855 Note that a given ILNP session normally will use a single 856 Identifier value for the life of that session. 858 Existing DNS specifications require that DNS clients and DNS 859 resolvers obey the TTL values provided by the DNS servers. In 860 the context of this proposal, short DNS TTL values are assigned 861 to particular DNS records to ensure that the ubiquitous DNS 862 caching resolvers do not cache volatile values (e.g. Locator 863 records of a mobile node) and consequently return stale 864 information to new requestors. 866 As a practical matter, it is not sensible to flush all Locator 867 values associated with an existing session's correspondent node. 868 Instead, Locator values cached for a correspondent node (in the 869 ILNP Correspondent Cache, described in Section 12.1) SHOULD be 870 marked as "aged" when their TTL has expired until either the next 871 Locator Update message is received or there is other indication 872 that a given Locator is not working any longer. 874 During a long transition period, a node that is I/L-enabled 875 SHOULD have not only ID and L64 (or ID and LP) records present in 876 its authoritative DNS server, but also SHOULD have AAAA records 877 in the DNS for the benefit of non-upgraded nodes. This 878 capability might be implemented strictly inside a DNS server, 879 whereby the DNS server synthesised a set of AAAA records to 880 advertise from the ID and Locator (i.e., L32, L64, or LP) values 881 that the node has kept updated in that DNS server. 883 Existing DNS specifications require that a DNS resolver or DNS 884 client ignore unrecognised DNS record types. So gratuitously 885 appending ID and Locator (i.e., L32, L64, or LP) records as 886 "additional data" in DNS responses to AAAA queries ought not 887 create any operational issues. 889 9. REFERRALS & APPLICATION PROGRAMMING INTERFACES 891 This section is concerned with support for using 892 existing ("legacy") applications over ILNP, including 893 both referrals and APIs. 895 9.1 BSD Sockets APIs 897 The existing BSD Sockets API can continue to be used with 898 ILNP underneath the API. That API can be implemented in a 899 manner that hides the underlying protocol changes from the 900 applications. For example, the combination of a Locator 901 and an Identifier can be used with the API in the place 902 of an IPv6 address. 904 So it is believed that existing IP address referrals can 905 continue to work properly in most cases. For a rapidly 906 moving target node, referrals might break in at least some 907 cases. The potential for referral breakage is necessarily 908 dependent upon the specific application and implementation 909 being considered. 911 It is suggested, however, that a new, optional, more abstract, 912 C language API be created so that new applications may avoid 913 delving into low-level details of the underlying network 914 protocols. Such an API would be useful today, even with 915 the existing IPv4 and IPv6 Internet, whether or not ILNP 916 were ever widely deployed. 918 9.2 Java APIs 920 Most existing Java APIs already use abstracted network 921 programming interfaces, for example in the java.Net.URL class. 922 Because these APIs already hide the low-level network-protocol 923 details from the applications, the applications using these APIs 924 (and the APIs themselves) don't need any modification to work 925 equally well with IPv4, IPv6, ILNP, and probably also HIP. 927 9.3 Referrals in the Future 929 The approach proposed in [ID-Referral] appears to be very 930 suitable for use with ILNP, in addition to being suitable 931 for use with the deployed Internet. Protocols using that 932 approach would not need modification to have their referrals 933 work well with IPv4, IPv6, ILNP, and probably also other 934 network protocols (e.g. HIP). 936 A more sensible approach to referrals would be to use 937 Fully-Qualified Domain Names (FQDNs), as is commonly done 938 today with web URLs. This is approach is highly portable 939 across different network protocols, even with both the IPv4 940 Internet or the IPv6 Internet. 942 10. BACKWARDS COMPATIBILITY 944 First, if one compares Figure 1 and Figure 2, one can see 945 that IPv6 with the Identifier/Locator Split enhancement is 946 fully backwards compatible with existing IPv6. This means 947 that no router software or silicon changes are necessary to 948 support the proposed enhancements. A router would be 949 unaware whether the packet being forwarded were classic IPv6 950 or the proposed enhanced version of IPv6. So no changes to 951 IPv6 routers is required to deploy this proposal. 953 Further, IPv6 Neighbour Discovery should work fine as is. 955 If a node that has been enhanced to support the Identifier/ 956 Locator Split mode initiates an IP session with another node, 957 normally it will first perform a DNS lookup on the responding 958 node's DNS name. If the initiator node does not find any ID 959 or L64 DNS resource records for the responder node, then the 960 initiator uses the Classical IPv6 mode of operation for the 961 new session with the responder, rather than trying to use 962 the I/L Split mode for that session. Of course, multiple 963 transport-layer sessions can concurrently share a single 964 network-layer (e.g. IP or ILNP) session. 966 If the responder node for a new IP session has not been enhanced 967 to support the I/L Split mode and receives initial packet(s) 968 containing the Nonce Destination Option, the responder will drop 969 the packet and send an ICMP Parameter Problem error message back 970 to the initiator. A responder node that has been upgraded to 971 support the I/L Split mode that receives initial packet(s) 972 containing the Nonce Destination Option knows those packets are 973 ILNP packets by the presence of that Nonce Destination Option. 975 If the initiator node does not receive a response from the 976 responder in a timely manner (e.g. within TCP timeout for a TCP 977 session) and also does not receive an ICMP Unreachable error 978 message for that packet, OR if the initiator receives an ICMP 979 Parameter Problem error message for that packet, then the 980 initiator knows that the responder is not able to support the I/L 981 Split Operating mode. In this case, the initiator node SHOULD 982 try again to create the new IP session but this time OMITTING the 983 Nonce Destination Option, and this time operating in Classic IPv6 984 mode, rather than I/L Split mode. 986 Finally, since an ILNP node is also a fully-capable IPv6 node, 987 then the upgraded node can use any standardised IPv6 mechanisms 988 for communicating with a legacy IPv6 node (i.e. an IPv6 node 989 without ILNP capability enhancements). So ILNP will in no case 990 be worse than existing IPv6, and in many cases ILNP will out 991 perform existing IPv6. 993 11. INCREMENTAL DEPLOYMENT 995 If a node has been enhanced to support the Identifier/ Locator 996 Split operating mode, that node's fully-qualified domain name 997 will normally have one or more ID records and one or more Locator 998 (i.e. L32, L64, and LP) records associated with the node within 999 the DNS. 1001 When a host ("initiator") initiates a new IP session with a 1002 correspondent ("responder"), it normally will perform a DNS 1003 lookup to determine the address(es) of the responder. A host 1004 that has been enhanced to support the Identifier/ Locator Split 1005 operating mode normally will look for Identifier ("ID") and 1006 Locator (i.e. L32, L64, and LP) records in any received DNS 1007 replies. DNS servers that support ID and Locator (i.e. L32, L64, 1008 and LP) records SHOULD include them (when they exist) as 1009 additional data in all DNS replies to queries for DNS AAAA 1010 records.[ILNP-DNS] 1012 If the initiator supports the I/L Split mode and from DNS 1013 information learns that the responder also supports the 1014 I/L Split mode, then the initiator will generate an 1015 unpredictable nonce value, store that value in a local 1016 Correspondent Cache, which is described in more detail below, 1017 and will include the Nonce Destination Option in its 1018 initial packet(s) to the responder.[ILNP-Nonce] 1020 If the responder supports the I/L Split mode and receives 1021 initial packet(s) containing the Nonce Destination Option, 1022 the responder will thereby know that the initiator supports 1023 the I/L Split mode and the responder will also operate in 1024 I/L Split mode for this new IP session. 1026 If the responder supports the I/L Split mode and receives 1027 initial packet(s) NOT containing the Nonce Destination Option, 1028 the responder will thereby know that the initiator does NOT 1029 support the I/L Split mode and the responder will operate 1030 in classic IPv6 mode for this new IP session. 1032 The previous section described how interoperability between 1033 enhanced nodes and non-enhanced nodes is retained even if a 1034 non-enhanced node erroneously has ID and/or L64 DNS resource 1035 records in place (e.g. due to some accident). 1037 The mobility capabilities of ILNP might be the most applicable 1038 to the deployment world. Despite substantial good efforts by many, 1039 neither Mobile IPv4 nor Mobile IPv6 are widely used at present. 1040 There are credible reports of specialised deployments 1041 of Mobile IPv4 and/or Mobile IPv6 within some wireless networks 1042 built using some mobile telephony standards (e.g. CDMA2000). 1043 However, much of the recent work in operating systems has focused 1044 on support for mobile devices (e.g. mobile telephone handsets, 1045 hand-held music players, hand-held organisers). Those devices 1046 probably represent the fastest growth segment of the Internet at 1047 present. Moreover, many vendors of such devices have included 1048 significant networking protocol improvements in incremental 1049 operating system updates, rather than always waiting for a new 1050 major release to add networking facilities. 1052 Data center operators might be interested in using ILNP to 1053 facilitate virtual machine mobility between VLANs within a data 1054 centre site or to a separate disaster recovery site. 1056 However, other users or vendors might be more interested by the 1057 new security models enabled by having Identifiers different from 1058 Locators, or they might be more interested in the ability to 1059 provide node-specific multi-homing, rather than always 1060 multi-homing an entire site. 1062 In the end, the marketplace has myriad users with various 1063 functional needs. The set of improvements offered by ILNP is 1064 broad, and should appeal to a wide range of vendors and users. 1066 12. IMPLEMENTATION CONSIDERATIONS 1068 This section discusses implementation considerations that 1069 are not otherwise discussed in the ILNP Internet-Drafts. 1071 12.1 ILNP Correspondent Cache 1073 An ILNP-capable node will need to modify its network protocol 1074 implementation to add an ILNP Correspondent Cache. In theory, 1075 this cache is within the ILNP network-layer. However, many 1076 network protocol implementations do not have strict protocol 1077 separation or layering. In the interest of efficient 1078 implementation, and to avoid unduly restricting implementers, 1079 an ILNP implementation is not required to limit the 1080 accessibility of ILNP Correspondent Cache to the network-layer. 1082 The ILNP Correspondent Cache contains at least the following 1083 inter-related data elements for the node itself: 1085 Set of Local Locator(s) 1086 & Preference value for each Locator 1087 Set of Local Identifier(s) 1089 and also the following per-correspondent data elements: 1090 Set of Correspondent's Locator(s) 1091 & Preference value for each Locator 1092 Set of Correspondent's Identifier(s) 1093 Nonce used from the local node to that correspondent 1094 Nonce used from that correspondent to the local node 1095 Valid Time 1097 For packets containing an ILNP Nonce Destination Option, lookups 1098 in the ILNP Correspondent Cache normally use an (Correspondent 1099 Identifier, Nonce) tuple as a lookup key. This facilitates 1100 situations where, perhaps due to deployment of Local-scope 1101 Identifiers, more than one correspondent node is using the same 1102 Identifier value. 1104 For other ILNP packets, the lookup key should be (Correspondent 1105 Locator, Correspondent Identifier). The Correspondent Locator 1106 is needed in the lookup key to distinguish between different 1107 correspondents that coincidentally are using the same 1108 Correspondent Identifier value. 1110 The Valid Time field indicates the remaining lifetime for which 1111 this ILNP session information is valid. For time, a node might 1112 use UTC (e.g. via Network Time Protocol) or perhaps some 1113 node-specific time (e.g. seconds since node boot). A table entry 1114 entry is current if the node's current time is less than or equal 1115 to the time in the Valid Time field, while a table entry is aged 1116 if the node's current time is greater than the time in the Valid 1117 Time field. 1119 While Locators are omitted from the transport-layer checksum, 1120 an implementation may use Locator values to distinguish between 1121 correspondents coincidentally using the same ID value when 1122 demultiplexing to determine which application(s) should receive 1123 the user data delivered by the transport-layer protocol. 1125 12.2 ICMP Locator Updates 1127 While ILNP's ICMP Locator Update message is defined in a 1128 separate document [ILNP-ICMP], it is worth mentioning that 1129 received authenticated Locator Update messages cause the 1130 ILNP Correspondent Cache described just above to be updated. 1131 Implementers should keep in mind that a node or site might 1132 have a large number of concurrent Locators, and should 1133 ensure that a system fault does not arise if the system 1134 receives an authentic ICMP Locator Update containing a 1135 large number of Locator values. 1137 13. SECURITY CONSIDERATIONS 1139 This proposal outlines a proposed evolution for the Internet 1140 Architecture to provide improved capabilities. This section 1141 discusses security considerations for this proposal. Note that 1142 ILNP provides security equivalent to IP for similar threats when 1143 similar mitigations (e.g. IPsec or not) are in use. In some 1144 cases, but not all, ILNP exceeds that objective and has less 1145 security risk than IP. 1147 13.1 Authentication of Locator Updates 1149 A separate document [ILNP-Nonce] proposes a new IPv6 1150 Destination Option that can be used to carry a session nonce 1151 end-to-end between communicating nodes. That nonce provides 1152 protection against off-path attacks on an Identifier/Locator 1153 session. The Nonce Destination Option is used ONLY for IP 1154 sessions in the Identifier/Locator Split mode. The nonce 1155 values are exchanged in the initial packets of an ILNP 1156 session. 1158 Ordinary IPv6 is vulnerable to on-path attacks unless 1159 the IP Authentication Header or IP Encapsulating Security 1160 Payload is in use. So the Nonce Destination Option 1161 only seeks to provide protection against off-path attacks 1162 on an IP session -- equivalent to ordinary IPv6 when 1163 not using IP Security. 1165 When the Identifier/Locator split mode is in use for an 1166 existing IP session, the Nonce Destination Option MUST be 1167 included in any ICMP control messages (e.g. ICMP Unreachable, 1168 ICMP Locator Update) sent with regard to that IP session. 1170 It is common to have non-symmetric paths between two nodes 1171 on the Internet. To reduce the number of on-path nodes that 1172 know the Nonce value for a given session when the I/L split 1173 mode is in use, a nonce value is unidirectional, not 1174 bidirectional. For example, for a session between two nodes 1175 A and B, one nonce value is used from A to B and a different 1176 nonce value is used from B to A. 1178 When in the I/L Split operating mode for an existing IP 1179 session, ICMP control messages received without a Nonce 1180 Destination Option MUST be discarded as forgeries. This 1181 security event SHOULD be logged. 1183 When in the I/L Split operating mode for an existing IP 1184 session, ICMP control messages received without a correct 1185 nonce value inside the Nonce Destination Option MUST be 1186 discarded as forgeries. This security event SHOULD be logged. 1188 When in the I/L Split operating mode for an existing IP 1189 session, and a node changes its Locator set, it should 1190 include the Nonce Destination Option in the first few 1191 data packets sent using a new Locator value, so that 1192 the recipient can validate the received data packets 1193 as valid (despite having an unexpected Source Locator 1194 value). 1196 For ID/Locator Split mode sessions operating in higher risk 1197 environments, the use of the cryptographic authentication 1198 provided by IP Authentication Header is recommended 1199 *in addition* to concurrent use of the Nonce Destination 1200 Option. 1202 It is important to note that at present an IPv6 session is 1203 entirely vulnerable to on-path attacks unless IPsec is in use 1204 for that particular IPv6 session, so the security properties 1205 of the new proposal are never worse than for existing IPv6. 1207 13.2 Forged Identifier Attacks 1209 In the deployed Internet, active attacks using packets with a 1210 forged Source IP Address have been publicly known at least since 1211 early 1995.[CA-1995-01] While these exist in the deployed 1212 Internet, they have not been widespread. This is equivalent to 1213 the issue of a forged Identifier value and demonstrates that this 1214 is not a new threat created by the Identifier/Locator-split mode 1215 of operation. 1217 One mitigation for these attacks has been to deploy Source IP 1218 Address Filtering.[RFC 2827] [RFC 3704] Jun Bi at U. Tsinghua 1219 cites Arbor Networks as reporting that this mechanism has less 1220 than 50% deployment and cites an MIT analysis indicating that at 1221 least 25% of the deployed Internet permits forged source IP 1222 addresses. 1224 Other parts of this document discuss the probability of an 1225 accidental duplicate Identifier being used on the Internet. 1226 However, this sub-section instead focuses on methods for 1227 mitigating attacks based on packets containing deliberately 1228 forged Source Identifier values. 1230 First, the recommendations of [RFC 2827] & [RFC 3704] remain. 1231 So any packets that have a forged Locator value can be easily 1232 filtered using existing widely available mechanisms. 1234 Second, the receiving node does not blindly accept any packet 1235 with the proper Source Identifier and proper Destination 1236 Identifier as an authentic packet. Instead, each node operating 1237 the I/L-split mode maintains an ILNP Correspondent Cache for each 1238 of its correspondents, as described above. This cache contains 1239 two unidirectional nonce values (one used in control messages 1240 sent by this node, a different one used to authenticate messages 1241 from the other node). The correspondent cache also contains 1242 the currently valid set of Locators and set of Identifiers for 1243 each correspondent node. If a received packet contains valid 1244 Identifier values and a valid Destination Locator, but contains 1245 a Source Locator value that is not present in the correspondent 1246 cache, the packet is dropped without further processing as an 1247 invalid packet, unless the packet also contains a Nonce 1248 Destination Option with the correct value used for packets from 1249 the node with that Source Identifier to this node. This prevents 1250 an off-path attacker from stealing an existing session. 1252 Third, any node can distinguish different nodes using the same 1253 Identifier value by other properties of their sessions. For 1254 example, IPv6 Neighbour Discovery prevents more than one node 1255 from using the same source (Locator + Identifier) pair at the 1256 same time on the same link. So cases of different nodes using 1257 the same Identifier value will involve nodes that have different 1258 sets of valid Locator values. A node can thus demux based on the 1259 combination of Source Locator and Source Identifier if necessary. 1260 If IP Security is in use, the combination of the Source 1261 Identifier and the SPI value would be sufficient to demux two 1262 different sessions. 1264 Fourth, deployments in high threat environments also SHOULD use 1265 the IP Authentication Header to authenticate control traffic and 1266 data traffic. Because in the I/L-split mode, IP Security binds 1267 only to the Identifier values, and never to the Locator values, 1268 this enables a mobile or multi-homed node to use IPsec even when 1269 its Locator value(s) have just changed. 1271 Last, note well that ordinary IPv4, ordinary IPv6, Mobile IPv4, 1272 and also Mobile IPv6 already are vulnerable to forged Identifier 1273 and/or forged IP address attacks. An attacker on the same link 1274 as the intended victim simply forges the victims MAC address and 1275 the victim's IP address. With IPv6, when SEND and CGAs are in 1276 use, the victim node can defend its use of its IPv6 address using 1277 SEND. With ILNP, when SEND and CGIs are in use, the victim node 1278 also can defend its use of its IPv6 address using SEND. There 1279 are no standard mechanisms to authenticate ARP messages, so IPv4 1280 is especially vulnerable to this sort of attack. These attacks 1281 also work against Mobile IPv4 and Mobile IPv6. In fact, when 1282 either form of Mobile IP is in use, there are additional risks, 1283 because the attacks work not only when the attacker has access to 1284 the victim's current IP subnetwork but also when the attacker has 1285 access to the victim's home IP subnetwork. So the risks of 1286 using ILNP are not greater than exist today with IP or Mobile IP. 1288 13.3 IP Security Enhancements 1290 The IP Security standards are enhanced here by binding IPsec 1291 Security Associations to the Identifiers of the session 1292 endpoints, rather than binding IPsec Security Associations 1293 to the IP Addresses as at present. This change enhances the 1294 deployability and interoperability of the IP Security standards, 1295 but does not decrease the security provided by those protocols. 1297 Also, the IP Authentication Header omits the Source Locator and 1298 Destination Locator fields from its authentication calculations 1299 when ILNP is in use. This enables IP AH to work well even 1300 through a NAT or other situation where a Locator value might 1301 change during transit. 1303 13.4 DNS Security 1305 The DNS enhancements proposed here are entirely compatible with, 1306 and can be protected using, the existing IETF standards for DNS 1307 Security.[RFC 4033] The Secure DNS Dynamic Update mechanism used 1308 here is also used unchanged.[RFC 3007] So there is no change to 1309 the security properties of the Domain Name System or of DNS 1310 servers due to ILNP. 1312 13.5 Firewall Considerations 1314 In the proposed new scheme, stateful firewalls are able to 1315 authenticate ICMP control messages arriving on the external 1316 interface. This enables more thoughtful handling of ICMP 1317 messages by firewalls than is commonly the case at present. As 1318 the firewall is along the path between the communicating nodes, 1319 the firewall can snoop on the Session Nonce being carried in the 1320 initial packets of an I/L Split mode session. The firewall can 1321 verify the correct nonce is present on incoming control packets, 1322 dropping any control packets that lack the correct nonce value. 1324 By always including the nonce in ILNP control messages, even when 1325 IP Security is also in use, the firewall can filter out off-path 1326 attacks against those ILNP messages. In any event, a forged 1327 packet from an on-path attacker will still be detected when the 1328 IPsec input processing occurs in the receiving node; this will 1329 cause that forged packet to be dropped rather than acted upon. 1331 13.6 Neighbour Discovery Authentication 1333 Nothing in this proposal prevents sites from using the Secure 1334 Neighbour Discovery (SEND) proposal for authenticating IPv6 1335 Neighbour Discovery. [RFC 3971] 1337 13.7 Site Topology Obfuscation 1339 A site that wishes to obscure its internal topology information 1340 MAY do so by deploying site border routers that rewrite the 1341 Locator values for the site as packets enter or leave the site. 1343 For example, a site might choose to use a ULA prefix internally 1344 for this reason.[RFC 4193] [ID-ULA] In this case, the site border 1345 routers would rewrite the Source Locator of ILNP packets leaving 1346 the site to a global-scope Locator associated with the site. 1347 Also, those site border routers would rewrite the Destination 1348 Locator of packets entering the site from the global-scope 1349 Locator to an appropriate interior ULA Locator for the 1350 destination node.[MILCOM08] 1352 13.8 Path Liveness 1354 Some perceive that an Identifier-Locator Split architecture 1355 creates a new issue that is sometimes called "Locator Liveness" 1356 or "Path Liveness". This refers to the question of whether an IP 1357 packet with a particular destination Locator value will be able 1358 to reach the intended destination or not, given that some 1359 otherwise valid paths might be unusable by the sending node 1360 (e.g. due to security policy or other administrative choice). 1361 In fact, this issue has existed in the IPv4 Internet for decades. 1363 For example, an IPv4 server might have multiple valid IP 1364 addresses, each advertised to the world via an DNS "A" record. 1365 However, at a given moment in time, it is possible that a given 1366 sending node might not be able to use a given (otherwise valid) 1367 destination IPv4 address in an IP packet to reach that IPv4 1368 server. 1370 So we see that using an Identifier/Locator Split architecture 1371 does not create this issue, nor does it make this issue worse 1372 than it is with the deployed IPv4 Internet. 1374 In ILNP, the same conceptual approach described in [RFC 5534] can 1375 be reused. Alternatively, an ILNP node can reuse the existing 1376 IPv4 methods for determining whether a given path to the target 1377 destination is currently usable, which existing methods leverage 1378 transport-layer session state information that the communicating 1379 end systems are already keeping for transport-layer protocol 1380 reasons. 1382 Last, it is important for the reader to understand that the 1383 mechanism described in [ILNP-ICMP] is a performance optimisation, 1384 significantly shortening the layer-3 handoff time if/when a 1385 correspondent changes location. Architecturally, using ICMP 1386 is no different from using UDP, of course. 1388 14. IANA CONSIDERATIONS 1390 This document has no IANA considerations. 1392 15. REFERENCES 1394 This section provides both normative and informative 1395 references relating to this note. 1397 15.1. Normative References 1399 [RFC 826] D. Plummer, "Ethernet Address Resolution Protocol: 1400 Or Converting Network Protocol Addresses to 1401 48 bit Ethernet Address for Transmission on 1402 Ethernet Hardware", RFC 826, November 1982. 1404 [RFC 2119] Bradner, S., "Key words for use in RFCs to 1405 Indicate Requirement Levels", BCP 14, RFC 2119, 1406 March 1997. 1408 [RFC 2460] S. Deering & R. Hinden, "Internet Protocol 1409 Version 6 Specification", RFC-2460, 1410 December 1998. 1412 [RFC 3007] B. Wellington, "Secure Domain Name System 1413 Dynamic Update", RFC-3007, November 2000. 1415 [RFC 3484] R. Draves, "Derfault Address Selection for IPv6", 1416 RFC 3484, February 2003. 1418 [RFC 4033] R. Arends, et alia, "DNS Security Introduction 1419 and Requirements", RFC-4033, March 2005. 1421 [RFC 4219] R. Hinden & S. Deering, "IP Version 6 1422 Addressing Architecture", RFC-4219, 1423 February 2006. 1425 [RFC 4861] T. Narten, E. Nordmark, W. Simpson, & H. Soliman, 1426 "Neighbor Discovery for IP version 6 (IPv6)", 1427 RFC 4861, September 2007. 1429 15.2. Informative References 1431 [8+8] M. O'Dell, "8+8 - An Alternate Addressing 1432 Architecture for IPv6", Internet-Draft, 1433 October 1996. 1435 [Bhatti10] S. Bhatti, "Reducing DNS Caching (or 'How low 1436 can we go ?')", Presentation to 38th JANET 1437 Networkshop, 31st March 2010, UK Joint 1438 Academic Network (JANET), University of Manchester, 1439 Manchester, England, UK. 1441 [CA-1995-01] US CERT, "IP Spoofing Attacks and Hijacked 1442 Terminal Connections", CERT Advisory 1995-01, 1443 Issued 23 JAN 1995, Revised 23 SEP 1997. 1445 [GSE] M. O'Dell, "GSE - An Alternate Addressing 1446 Architecture for IPv6", Internet-Draft, 1447 February 1997. 1449 [ID-ULA] R. Hinden, G. Huston, & T. Narten, "Centrally 1450 Assigned Unique Local IPv6 Unicast Addresses", 1451 draft-ietf-ipv6-ula-central-02.txt, 15 June 2007. 1453 [ID-Referral] B. Carpenter and others, "A Generic Referral 1454 Object for Internet Entities", 1455 draft-carpenter-behave-referral-object-01, 1456 20 October 2009. 1458 [IEEE-EUI] IEEE Standards Association, "Guidelines for 1459 64-bit Global Identifier (EUI-64)", IEEE, 1460 2007. 1462 [IEN 1] C.J. Bennett, S.W. Edge, & A.J. Hinchley, 1463 "Issues in the Interconnection of Datagram 1464 Networks", Internet Experiment Note (IEN) 1, 1465 INDRA Note 637, PSPWN 76, University College 1466 London, London, England, UK, WC1E 6BT, 1467 29 July 1977. 1468 http://www.postel.org/ien/ien001.pdf 1470 [IEN 19] J. F. Shoch, "Inter-Network Naming, Addressing, 1471 and Routing", IEN-19, January 1978. 1473 [IEN 23] J. F. Shoch, "On Names, Addresses, and 1474 Routings", IEN-23, January 1978. 1476 [IEN 31] D. Cohen, "On Names, Addresses, and Routings 1477 (II)", IEN-31, April 1978. 1479 [ILNP-Nonce] R. Atkinson, "Nonce Destination Option", 1480 draft-rja-ilnp-nonce-05.txt, August 2010. 1482 [ILNP-DNS] R. Atkinson, "DNS Resource Records for ILNP", 1483 draft-rja-ilnp-dns-06.txt, August 2010. 1485 [ILNP-ICMP] R. Atkinson, "ICMP Locator Update message" 1486 draft-rja-ilnp-icmp-04.txt, August 2010. 1488 [IPng95] D. Clark, "A thought on addressing", 1489 electronic mail message to IETF IPng WG, 1490 Message-ID: 9501111901.AA28426@caraway.lcs.mit.edu, 1491 Laboratory for Computer Science, MIT, 1492 Cambridge, MA, USA, 11 January 1995. 1494 [Liu-DNS] C. Liu & P. Albitz, "DNS & Bind", 5th Edition, 1495 O'Reilly & Associates, Sebastopol, CA, USA, 1496 May 2006. ISBN 0-596-10057-4 1498 [MobiArch07] R. Atkinson, S. Bhatti, & S. Hailes, 1499 "Mobility as an Integrated Service Through 1500 the Use of Naming", Proceedings of 1501 ACM MobiArch 2007, August 2007, 1502 Kyoto, Japan. 1504 [MobiArch08] R. Atkinson, S. Bhatti, & S. Hailes, 1505 "Mobility Through Naming: Impact on DNS", 1506 Proceedings of ACM MobiArch 2008, August 2008, 1507 Seattle, WA, USA. 1509 [MobiWAC07] R. Atkinson, S. Bhatti, & S. Hailes, 1510 "A Proposal for Unifying Mobility with 1511 Multi-Homing, NAT, & Security", 1512 Proceedings of ACM MobiWAC 2007, Chania, 1513 Crete. ACM, October 2007. 1515 [MILCOM08] R. Atkinson, S. Bhatti, & S. Hailes, 1516 "Harmonised Resilience, Security, and Mobility 1517 Capability for IP", Proceedings of IEEE 1518 Military Communications (MILCOM) Conference, 1519 San Diego, CA, USA, November 2008. 1521 [MILCOM09] R. Atkinson, S. Bhatti, & S. Hailes, 1522 "Site-Controlled Secure Multi-Homing and 1523 Traffic Engineering For IP", Proceedings of 1524 IEEE Military Communications (MILCOM) Conference, 1525 Boston, MA, USA, October 2009. 1527 [PHG02] Pappas, A, S. Hailes, & R. Giaffreda, 1528 "Mobile Host Location Tracking through DNS", 1529 Proceedings of IEEE London Communications 1530 Symposium, IEEE, September 2002, London, 1531 England, UK. 1533 [SBK2002] Alex C. Snoeren, Hari Balakrishnan, & M. Frans 1534 Kaashoek, "Reconsidering Internet Mobility", 1535 Proceedings of 8th Workshop on Hot Topics in 1536 Operating Systems, 2002. 1538 [SIPP94] Bob Smart, "Re: IPng Directorate meeting in 1539 Chicago; possible SIPP changes", electronic 1540 mail to the IETF SIPP WG mailing list, 1541 Message-ID: 1542 199406020647.AA09887@shark.mel.dit.csiro.au, 1543 Commonwealth Scientific & Industrial Research 1544 Organisation (CSIRO), Melbourne, VIC, 3001, 1545 Australia, 2 June 1994. 1547 [RFC 814] D.D. Clark, "Names, Addresses, Ports, and 1548 Routes", RFC-814, July 1982. 1550 [RFC 1498] J.H. Saltzer, "On the Naming and Binding of 1551 Network Destinations", RFC-1498, August 1993. 1553 [RFC 1631] K. Egevang & P. Francis, "The IP Network 1554 Address Translator (NAT)", RFC-1631, May 1994. 1556 [RFC 2827] P. Ferguson & D. Senie, "Network Ingress Filtering: 1557 Defeating Denial of Service Attacks which employ 1558 IP Source Address Spoofing", RFC-2827, May 2000. 1560 [RFC 3022] P. Srisuresh & K. Egevang, "Traditional IP 1561 Network Address Translator", RFC-3022, 1562 January 2001. 1564 [RFC 3027] M. Holdrege and P Srisuresh, "Protocol 1565 Complications of the IP Network Address 1566 Translator", RFC-3027, January 2001. 1568 [RFC 3704] F. Baker & P. Savola, "Ingress Filtering for 1569 Multihomed Networks, RFC-3704, March 2004. 1571 [RFC 3715] B. Aboba and W. Dixon, "IPsec-Network Address 1572 Translation (NAT) Compatibility Requirements", 1573 RFC-3715, March 2004. 1575 [RFC 3775] D. Johnson, C. Perkins, and J. Arkko, "Mobility 1576 Support in IPv6", RFC-3775, June 2004. 1578 [RFC 3948] A. Huttunen, et alia, "UDP Encapsulation of 1579 IPsec ESP Packets", RFC-3948, January 2005. 1581 [RFC 3971] J. Arkko, J. Kempf, B. Zill, & P. Nikander, 1582 "SEcure Neighbor Discovery (SEND)", RFC-3971 1583 March 2005. 1585 [RFC 3972] T. Aura, "Cryptographically Generated Addresses 1586 (CGAs)", RFC-3972, March 2005. 1588 [RFC 4193] R. Hinden & B. Haberman, "Unique Local IPv6 1589 Unicast Addresses, RFC-4193, October 2005. 1591 [RFC 4941] T. Narten, R. Draves, & S. Krishnan, "Privacy 1592 Extensions for Stateless Address Autoconfiguration 1593 in IPv6", RFC-4941, September 2007. 1595 [RFC 5061] R. Stewart, Q. Xie, M. Tuexen, S. Maruyama, & 1596 M. Kozuka, "Stream Control Transmission Protocol 1597 (SCTP) Dynamic Address Reconfiguration", RFC-5061, 1598 September 2007. 1600 [RFC 5534] J. Arkko & I. van Beijnum, "Failure Detection and 1601 Locator Pair Exploration Protocol for IPv6 1602 Multihoming", RFC-5534, June 2009. 1604 [TeleSys] R. Atkinson, S. Bhatti, & S. Hailes, 1605 "ILNP: Mobility, Multi-Homing, Localised Addressing 1606 and Security Through Naming", Telecommunications 1607 Systems, Volume 42, Number 3-4, pp 273-291, 1608 Springer-Verlag, December 2009, ISSN 1018-4864. 1610 ACKNOWLEDGEMENTS 1612 Steve Blake, Mohamed Boucadair, Saleem Bhatti, Noel Chiappa, 1613 Steve Hailes, Joel Halpern, Mark Handley, Paul Jakma, Dae-Young 1614 Kim, Tony Li, Yakov Rehkter and Robin Whittle (in alphabetical 1615 order) provided review and feedback on earlier versions of this 1616 document. Steve Blake provided an especially thorough review 1617 of the entire ILNP document set, which was extremely helpful. 1618 Noel Chiappa graciously provided the author with copies of the 1619 original email messages cited here as [SIPP94] and [IPng95], 1620 which enabled the precise citation of those messages herein. 1622 AUTHOR'S ADDRESS 1624 R. Atkinson 1625 Consultant 1626 McLean, VA 1627 22102-9147 USA 1629 Email: rja.lists@gmail.com 1631 Expires: 18 FEB 2011