idnits 2.17.1 draft-ietf-hip-multihoming-11.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 date (September 8, 2016) is 2786 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-14) exists of draft-ietf-hip-rfc5206-bis-12 Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group T. Henderson, Ed. 3 Internet-Draft University of Washington 4 Intended status: Standards Track C. Vogt 5 Expires: March 12, 2017 Independent 6 J. Arkko 7 Ericsson 8 September 8, 2016 10 Host Multihoming with the Host Identity Protocol 11 draft-ietf-hip-multihoming-11 13 Abstract 15 This document defines host multihoming extensions to the Host 16 Identity Protocol (HIP), by leveraging protocol components defined 17 for host mobility. 19 Status of This Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at http://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on March 12, 2017. 36 Copyright Notice 38 Copyright (c) 2016 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (http://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Introduction and Scope . . . . . . . . . . . . . . . . . . . 2 54 2. Terminology and Conventions . . . . . . . . . . . . . . . . . 3 55 3. Protocol Model . . . . . . . . . . . . . . . . . . . . . . . 3 56 4. Protocol Overview . . . . . . . . . . . . . . . . . . . . . . 4 57 4.1. Background . . . . . . . . . . . . . . . . . . . . . . . 4 58 4.2. Multiple Addresses . . . . . . . . . . . . . . . . . . . 5 59 4.3. Multiple Security Associations . . . . . . . . . . . . . 6 60 4.4. Host Multihoming for Fault Tolerance . . . . . . . . . . 7 61 4.5. Host Multihoming for Load Balancing . . . . . . . . . . . 8 62 4.6. Site Multihoming . . . . . . . . . . . . . . . . . . . . 9 63 4.7. Dual Host Multihoming . . . . . . . . . . . . . . . . . . 10 64 4.8. Combined Mobility and Multihoming . . . . . . . . . . . . 10 65 4.9. Initiating the Protocol in R1, I2, or R2 . . . . . . . . 11 66 4.10. Using LOCATOR_SETs across Addressing Realms . . . . . . . 12 67 4.11. Interaction with Security Associations . . . . . . . . . 13 68 5. Processing Rules . . . . . . . . . . . . . . . . . . . . . . 13 69 5.1. Sending LOCATOR_SETs . . . . . . . . . . . . . . . . . . 13 70 5.2. Handling Received LOCATOR_SETs . . . . . . . . . . . . . 15 71 5.3. Verifying Address Reachability . . . . . . . . . . . . . 17 72 5.4. Changing the Preferred Locator . . . . . . . . . . . . . 18 73 6. Security Considerations . . . . . . . . . . . . . . . . . . . 18 74 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20 75 8. Authors and Acknowledgments . . . . . . . . . . . . . . . . . 20 76 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 20 77 9.1. Normative references . . . . . . . . . . . . . . . . . . 20 78 9.2. Informative references . . . . . . . . . . . . . . . . . 21 79 Appendix A. Document Revision History . . . . . . . . . . . . . 22 80 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 23 82 1. Introduction and Scope 84 The Host Identity Protocol [RFC7401] (HIP) supports an architecture 85 that decouples the transport layer (TCP, UDP, etc.) from the 86 internetworking layer (IPv4 and IPv6) by using public/private key 87 pairs, instead of IP addresses, as host identities. When a host uses 88 HIP, the overlying protocol sublayers (e.g., transport layer sockets 89 and Encapsulating Security Payload (ESP) Security Associations (SAs)) 90 are instead bound to representations of these host identities, and 91 the IP addresses are only used for packet forwarding. However, each 92 host must also know at least one IP address at which its peers are 93 reachable. Initially, these IP addresses are the ones used during 94 the HIP base exchange. 96 One consequence of such a decoupling is that new solutions to 97 network-layer mobility and host multihoming are possible. Basic host 98 mobility is defined in [I-D.ietf-hip-rfc5206-bis] and covers the case 99 in which a host has a single address and changes its network point- 100 of-attachment while desiring to preserve the HIP-enabled security 101 association. Host multihoming is somewhat of a dual case to host 102 mobility, in that a host may simultaneously have more than one 103 network point-of-attachment. There are potentially many variations 104 of host multihoming possible. [I-D.ietf-hip-rfc5206-bis] specifies 105 the format of the HIP parameter (LOCATOR_SET parameter) used to 106 convey IP addressing information between peers, the procedures for 107 sending and processing this parameter to enable basic host mobility, 108 and procedures for an address verification mechanism. The scope of 109 this document encompasses messaging and elements of procedure for 110 some basic host multihoming scenarios of interest. 112 Another variation of multihoming that has been heavily studied is 113 site multihoming. Solutions for host multihoming in multihomed IPv6 114 networks have been specified by the IETF shim6 working group. The 115 shim6 protocol [RFC5533] bears many architectural similarities to HIP 116 but there are differences in the security model and in the protocol. 118 While HIP can potentially be used with transports other than the ESP 119 transport format [RFC7402], this document largely assumes the use of 120 ESP and leaves other transport formats for further study. 122 Finally, making underlying IP multihoming transparent to the 123 transport layer has implications on the proper response of transport 124 congestion control, path MTU selection, and Quality of Service (QoS). 125 Transport-layer mobility triggers, and the proper transport response 126 to a HIP multihoming address change, are outside the scope of this 127 document. 129 2. Terminology and Conventions 131 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 132 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 133 document are to be interpreted as described in RFC 2119 [RFC2119]. 135 The following terms used in this document are defined in 136 [I-D.ietf-hip-rfc5206-bis]: LOCATOR_SET, Locator, Address, Preferred 137 locator, and Credit Based Authorization. 139 3. Protocol Model 141 The protocol model for HIP support of host multihoming extends the 142 model for host mobility described in Section 3 of 144 [I-D.ietf-hip-rfc5206-bis]. This section only highlights the 145 differences. 147 In host multihoming, a host has multiple locators simultaneously 148 rather than sequentially, as in the case of mobility. By using the 149 LOCATOR_SET parameter defined in [I-D.ietf-hip-rfc5206-bis], a host 150 can inform its peers of additional (multiple) locators at which it 151 can be reached. When multiple locators are available and announced 152 to the peer, a host can designate a particular locator as a 153 "preferred" locator, meaning that the host prefers that its peer send 154 packets to the designated address before trying an alternative 155 address. Although this document defines a basic mechanism for 156 multihoming, it does not define all possible policies and procedures, 157 such as which locators to choose when more than one is available, the 158 operation of simultaneous mobility and multihoming, source address 159 selection policies (beyond those specified in [RFC6724]), and the 160 implications of multihoming on transport protocols. 162 4. Protocol Overview 164 In this section, we briefly introduce a number of usage scenarios for 165 HIP multihoming. These scenarios assume that HIP is being used with 166 the ESP transport [RFC7402], although other scenarios may be defined 167 in the future. To understand these usage scenarios, the reader 168 should be at least minimally familiar with the HIP protocol 169 specification [RFC7401], the use of the ESP transport format 170 [RFC7402], and the HIP mobility specification 171 [I-D.ietf-hip-rfc5206-bis]. However, for the (relatively) 172 uninitiated reader, it is most important to keep in mind that in HIP 173 the actual payload traffic is protected with ESP, and that the ESP 174 Security Parameter Index (SPI) acts as an index to the right host-to- 175 host context. 177 4.1. Background 179 The multihoming scenarios can be explained in contrast to the non- 180 multihoming case described in the base protocol specification. We 181 review the pertinent details here. In the base specification when 182 used with the ESP transport format, the HIP base exchange will set up 183 a single SA in each direction. The IP addresses associated with the 184 SAs are the same as those used to convey the HIP packets. For data 185 traffic, a security policy database (SPD) and security association 186 database (SAD) will likely exist, following the IPsec architecture. 187 One distinction between HIP and IPsec, however, is that the host IDs, 188 and not the IP addresses, are conceptually used as selectors in the 189 SPD. In the outbound direction, as a result of SPD processing, when 190 an outbound SA is selected, the correct IP destination address for 191 the peer must also be assigned. Therefore, outbound SAs are 192 conceptually associated with the peer IP address that must be used as 193 the destination IP address below the HIP layer. In the inbound 194 direction, the IP addresses may be used as selectors in the SAD to 195 look up the SA, but they are not strictly required; the ESP SPI may 196 be used alone. To summarize, in the non-multihoming case, there is 197 only one source IP address, one destination IP address, one inbound 198 SA, and one outbound SA. 200 The HIP readdressing protocol [I-D.ietf-hip-rfc5206-bis] is an 201 asymmetric protocol in which a mobile or multihomed host informs a 202 peer host about changes of IP addresses on affected SPIs. IP address 203 and ESP SPI information is carried in Locator data structures in a 204 HIP parameter called a LOCATOR_SET. The HIP mobility specification 205 [I-D.ietf-hip-rfc5206-bis] describes how the LOCATOR_SET is carried 206 in a HIP UPDATE packet. 208 To summarize the mobility elements of procedure, as background for 209 multihoming, the basic idea of host mobility is to communicate a 210 local IP address change to the peer when active HIP-maintained SAs 211 are in use. To do so, the IP address must be conveyed, any 212 association between the IP address and an inbound SA (via the SPI 213 index) may be conveyed, and protection against flooding attacks must 214 be ensured. The association of an IP address with an SPI is 215 performed by a Locator of type 1, which is a concatenation of an ESP 216 SPI with an IP address. 218 An address verification method is specified in 219 [I-D.ietf-hip-rfc5206-bis]. It is expected that addresses learned in 220 multihoming scenarios also are subject to the same verification 221 rules. The scenarios at times describe addresses as being in either 222 an ACTIVE, VERIFIED, or DEPRECATED state. From the perspective of a 223 host, newly-learned addresses of the peer must be verified before put 224 into active service, and addresses removed by the peer are put into a 225 deprecated state. Under limited conditions described in 226 [I-D.ietf-hip-rfc5206-bis], an UNVERIFIED address may be used. 228 With this background, we next describe additional protocol to 229 facilitate scenarios in which one or both hosts have multiple IP 230 addresses available. Increasingly, this is the common case with 231 network-connected hosts on the Internet. 233 4.2. Multiple Addresses 235 Hosts may have multiple IP addresses within different address 236 families (IPv4 and IPv6) and scopes available to support HIP 237 messaging and HIP-enabled SAs. The multiple addresses may be on a 238 single or multiple network interfaces. It is outside of the scope of 239 this document to specify how a host decides which of possibly 240 multiple addresses may be used to support a HIP association. Some IP 241 addresses may be held back from usage due to privacy, security, or 242 cost considerations. 244 When multiple IP addresses are shared with a peer, the procedures 245 described in the HIP mobility specification 246 [I-D.ietf-hip-rfc5206-bis] allow for a host to set a Preferred bit, 247 requesting that one of the multiple addresses be preferred for 248 control- or data-plane traffic. It is also permitted to leave the 249 Preferred bit unset for all addresses, allowing the peer to make 250 address selection decisions. 252 Hosts that use link-local addresses as source addresses in their HIP 253 handshakes may not be reachable by a mobile peer. Such hosts SHOULD 254 provide a globally routable address either in the initial handshake 255 or via the LOCATOR_SET parameter. 257 To support mobility, as described in the HIP mobility specification 258 [I-D.ietf-hip-rfc5206-bis], the LOCATOR_SET may be sent in a HIP 259 UPDATE packet. To support multihoming, the LOCATOR_SET may also be 260 sent in R1, I2, or R2 packets defined in the HIP protocol 261 specification [RFC7401]. The reason to consider to send LOCATOR_SET 262 parameters in the base exchange packets is to convey all usable 263 addresses for fault-tolerance or load balancing considerations. 265 4.3. Multiple Security Associations 267 When multiple addresses are available between peer hosts, a question 268 that arises is whether to use one or multiple SAs. The intent of 269 this specification is to support different use cases but to leave the 270 policy decision to the hosts. 272 When one host has n addresses and the other host has m addresses, it 273 is possible to set up as many as (n * m) SAs in each direction. In 274 such a case, every combination of source and destination IP address 275 would have a unique SA, and the possibility of reordering of 276 datagrams on each SA will be lessened (ESP SAs may have an anti- 277 replay window [RFC4303] sensitive to reordering). However, the 278 downside to creating a mesh of SAs is the signaling overhead required 279 (for exchanging UPDATE messages conveying ESP_INFO parameters) and 280 the state maintenance required in the SPD/SAD. 282 For load balancing, when multiple paths are to be used in parallel, 283 it may make sense to create different SAs for different paths. In 284 this use case, while a full mesh of 2 * (n * m) SAs may not be 285 required, it may be beneficial to create one SA pair per load- 286 balanced path to avoid anti-replay window issues. 288 For fault tolerance, it is more likely that a single SA can be used 289 and multiple IP addresses associated with that SA, and the 290 alternative addresses used only upon failure detection of the 291 addresses in use. Techniques for path failure detection are outside 292 the scope of this specification. An implementation may use ICMP 293 interactions, reachability checks, or other means to detect the 294 failure of a locator. 296 In summary, whether and how a host decides to leverage additional 297 addresses in a load-balancing or fault-tolerant manner is outside the 298 scope of the specification. However, in general, this document 299 recommends that for fault tolerance, it is likely sufficient to use a 300 single SA pair for all addresses, and for load balancing, to support 301 a different SA pair for all active paths being balanced across. 303 4.4. Host Multihoming for Fault Tolerance 305 A (mobile or stationary) host may have more than one interface or 306 global address. The host may choose to notify the peer host of the 307 additional interface or address by using the LOCATOR_SET parameter. 308 The LOCATOR_SET parameter may be included in an I2, R1, or R2 packet, 309 or may be conveyed, after the base exchange completes in an UPDATE 310 packet. 312 When more than one locator is provided to the peer host, the host MAY 313 indicate which locator is preferred (the locator on which the host 314 prefers to receive traffic). By default, the address that a host 315 uses in the base exchange is its preferred locator (for the address 316 family and address scope in use during the base exchange) until 317 indicated otherwise. It may be the case that the host does not 318 express any preferred locators. 320 In the multihoming case, the sender may also have multiple valid 321 locators from which to source traffic. In practice, a HIP 322 association in a multihoming configuration may have both a preferred 323 peer locator and a preferred local locator. The host should try to 324 use the peer's preferred locator unless policy or other circumstances 325 prevent such usage. A preferred local locator may be overridden if 326 source address selection rules on the destination address (peer's 327 preferred locator) suggest the use of a different source address. 329 Although the protocol may allow for configurations in which there is 330 an asymmetric number of SAs between the hosts (e.g., one host has two 331 interfaces and two inbound SAs, while the peer has one interface and 332 one inbound SA), it is RECOMMENDED that inbound and outbound SAs be 333 created pairwise between hosts. When an ESP_INFO arrives to rekey a 334 particular outbound SA, the corresponding inbound SA should be also 335 rekeyed at that time. 337 Consider the case of two hosts, one single-homed and one multihomed. 338 The multihomed host may decide to inform the single-homed host about 339 its other address(es). It may choose to do so as follows. 341 If the multihomed host wishes to convey the additional address(es) 342 for fault tolerance, it should include all of its addresses in 343 Locator records, indicating the Traffic Type, Locator Type, and 344 Preferred Locator for each address. If it wishes to bind any 345 particular address to an existing SPI, it may do so by using a 346 Locator of Type 1 as specified in the HIP mobility specification 347 [I-D.ietf-hip-rfc5206-bis]. It does not need to rekey the existing 348 SA or request additional SAs at this time. 350 Figure 1 illustrates this scenario. 352 Multi-homed Host Peer Host 354 UPDATE(LOCATOR_SET, SEQ) 355 -----------------------------------> 356 UPDATE(ACK) 357 <----------------------------------- 359 Figure 1: Basic Multihoming Scenario 361 In this scenario, the peer host associates the multiple addresses 362 with the SA pair between it and the multihomed host. It may also 363 undergo address verification procedures to transition the addresses 364 to ACTIVE state. For inbound data traffic, it may choose to use the 365 addresses along with the SPI as selectors. For outbound data 366 traffic, it must choose among the available addresses of the 367 multihomed host, considering the state of address verification 368 [I-D.ietf-hip-rfc5206-bis] of each address, and also considering 369 available information about whether an address is in a working state. 371 4.5. Host Multihoming for Load Balancing 373 A multihomed host may decide to set up new SA pairs corresponding to 374 new addresses, for the purpose of load balancing. The decision to 375 load balance and the mechanism for splitting load across multiple SAs 376 is out of scope of this document. The scenario can be supported by 377 sending the LOCATOR_SET parameter with one or more ESP_INFO 378 parameters to initiate new ESP SAs. To do this, the multihomed host 379 sends a LOCATOR_SET with an ESP_INFO, indicating the request for a 380 new SA by setting the OLD SPI value to zero, and the NEW SPI value to 381 the newly created incoming SPI. A Locator Type of "1" is used to 382 associate the new address with the new SPI. The LOCATOR_SET 383 parameter also contains a second Type "1" locator, that of the 384 original address and SPI. To simplify parameter processing and avoid 385 explicit protocol extensions to remove locators, each LOCATOR_SET 386 parameter MUST list all locators in use on a connection (a complete 387 listing of inbound locators and SPIs for the host). The multihomed 388 host waits for a corresponding ESP_INFO (new outbound SA) from the 389 peer and an ACK of its own UPDATE. As in the mobility case, the peer 390 host must perform an address verification before actively using the 391 new address. 393 Figure 2 illustrates this scenario. 395 Multi-homed Host Peer Host 397 UPDATE(ESP_INFO, LOCATOR_SET, SEQ, [DIFFIE_HELLMAN]) 398 -----------------------------------> 399 UPDATE(ESP_INFO, SEQ, ACK, [DIFFIE_HELLMAN,] ECHO_REQUEST) 400 <----------------------------------- 401 UPDATE(ACK, ECHO_RESPONSE) 402 -----------------------------------> 404 Figure 2: Host Multihoming for Load Balancing 406 In multihoming scenarios, it is important that hosts receiving 407 UPDATEs associate them correctly with the destination address used in 408 the packet carrying the UPDATE. When processing inbound LOCATOR_SETs 409 that establish new security associations on an interface with 410 multiple addresses, a host uses the destination address of the UPDATE 411 containing the LOCATOR_SET as the local address to which the 412 LOCATOR_SET plus ESP_INFO is targeted. This is because hosts may 413 send UPDATEs with the same (locator) IP address to different peer 414 addresses -- this has the effect of creating multiple inbound SAs 415 implicitly affiliated with different peer source addresses. 417 4.6. Site Multihoming 419 A host may have an interface that has multiple globally routable IP 420 addresses. Such a situation may be a result of the site having 421 multiple upper Internet Service Providers, or just because the site 422 provides all hosts with both IPv4 and IPv6 addresses. The host 423 should stay reachable at all or any subset of the currently available 424 global routable addresses, independent of how they are provided. 426 This case is handled the same as if there were different IP 427 addresses, described above in Section 4.4 and Section 4.5. Note that 428 a single interface may have addresses corresponding to site 429 multihoming while the host itself may also have multiple network 430 interfaces. 432 Note that a host may be multihomed and mobile simultaneously, and 433 that a multihomed host may want to protect the location of some of 434 its interfaces while revealing the real IP address of some others. 436 This document does not presently additional site multihoming 437 extensions to HIP; such extensions are for further study. 439 4.7. Dual Host Multihoming 441 Consider the case in which both hosts are multihomed and would like 442 to notify the peer of an additional address after the base exchange 443 completes. It may be the case that both hosts choose to simply 444 announce the second address in a LOCATOR_SET parameter using an 445 UPDATE message exchange. It may also be the case that one or both 446 hosts decide to ask for new SA pairs to be created using the newly 447 announced address. In the case that both hosts request this, the 448 result will be a full mesh of SAs as depicted in Figure 3. In such a 449 scenario, consider that host1, which used address addr1a in the base 450 exchange to set up SPI1a and SPI2a, wants to add address addr1b. It 451 would send an UPDATE with LOCATOR_SET (containing the address addr1b) 452 to host2, using destination address addr2a, and a new ESP_INFO, and a 453 new set of SPIs would be added between hosts 1 and 2 (call them SPI1b 454 and SPI2b; not shown in the figure). Next, consider host2 deciding 455 to add addr2b to the relationship. Host2 must select one of host1's 456 addresses towards which to initiate an UPDATE. It may choose to 457 initiate an UPDATE to addr1a, addr1b, or both. If it chooses to send 458 to both, then a full mesh (four SA pairs) of SAs would exist between 459 the two hosts. This is the most general case; the protocol is 460 flexible enough to accommodate this choice. 462 -<- SPI1a -- -- SPI2a ->- 463 host1 < > addr1a <---> addr2a < > host2 464 ->- SPI2a -- -- SPI1a -<- 466 addr1b <---> addr2a (second SA pair) 467 addr1a <---> addr2b (third SA pair) 468 addr1b <---> addr2b (fourth SA pair) 470 Figure 3: Dual Multihoming Case in which Each Host Uses LOCATOR_SET 471 to Add a Second Address 473 4.8. Combined Mobility and Multihoming 475 It looks likely that in the future, many mobile hosts will be 476 simultaneously mobile and multihomed, i.e., have multiple mobile 477 interfaces. Furthermore, if the interfaces use different access 478 technologies, it is fairly likely that one of the interfaces may 479 appear stable (retain its current IP address) while some other(s) may 480 experience mobility (undergo IP address change). 482 The use of LOCATOR_SET plus ESP_INFO should be flexible enough to 483 handle most such scenarios, although more complicated scenarios have 484 not been studied so far. 486 4.9. Initiating the Protocol in R1, I2, or R2 488 A Responder host MAY include a LOCATOR_SET parameter in the R1 packet 489 that it sends to the Initiator. This parameter MUST be protected by 490 the R1 signature. If the R1 packet contains LOCATOR_SET parameters 491 with a new Preferred locator, the Initiator SHOULD directly set the 492 new Preferred locator to status ACTIVE without performing address 493 verification first, and MUST send the I2 packet to the new Preferred 494 locator. The I1 destination address and the new Preferred locator 495 may be identical. All new non-preferred locators must still undergo 496 address verification once the base exchange completes. It is also 497 possible for the host to send the LOCATOR_SET without any Preferred 498 bits set, in which case the exchange will continue as normal and the 499 newly-learned addresses will be in an UNVERIFIED state at the 500 initiator. 502 Initiator Responder 504 R1 with LOCATOR_SET 505 <----------------------------------- 506 record additional addresses 507 change responder address 508 I2 sent to newly indicated preferred address 509 -----------------------------------> 510 (process normally) 511 R2 512 <----------------------------------- 513 (process normally, later verification of non-preferred locators) 515 Figure 4: LOCATOR_SET Inclusion in R1 517 An Initiator MAY include one or more LOCATOR_SET parameters in the I2 518 packet, independent of whether or not there was a LOCATOR_SET 519 parameter in the R1. These parameters MUST be protected by the I2 520 signature. Even if the I2 packet contains LOCATOR_SET parameters, 521 the Responder MUST still send the R2 packet to the source address of 522 the I2. The new Preferred locator, if set, SHOULD be identical to 523 the I2 source address. If the I2 packet contains LOCATOR_SET 524 parameters, all new locators must undergo address verification as 525 usual, and the ESP traffic that subsequently follows should use the 526 Preferred locator. 528 Initiator Responder 530 I2 with LOCATOR_SET 531 -----------------------------------> 532 (process normally) 533 record additional addresses 534 R2 sent to source address of I2 535 <----------------------------------- 536 (process normally) 538 Figure 5: LOCATOR_SET Inclusion in I2 540 The I1 and I2 may be arriving from different source addresses if the 541 LOCATOR_SET parameter is present in R1. In this case, 542 implementations simultaneously using multiple pre-created R1s, 543 indexed by Initiator IP addresses, may inadvertently fail the puzzle 544 solution of I2 packets due to a perceived puzzle mismatch. See, for 545 instance, the example in Appendix A of [RFC7401]. As a solution, the 546 Responder's puzzle indexing mechanism must be flexible enough to 547 accommodate the situation when R1 includes a LOCATOR_SET parameter. 549 Finally, the R2 may be used to carry the LOCATOR_SET parameter. In 550 this case, the LOCATOR_SET is covered by the HIP_MAC_2 and 551 HIP_SIGNATURE. Including LOCATOR_SET in R2 as opposed to R1 may have 552 some advantages when a host prefers not to divulge additional 553 locators until after the I2 is successfully processed. 555 When the LOCATOR_SET parameter is sent in an UPDATE packet, then the 556 receiver will respond with an UPDATE acknowledgment. When the 557 LOCATOR_SET parameter is sent in an R1, I2, or R2 packet, the base 558 exchange retransmission mechanism will confirm its successful 559 delivery. 561 4.10. Using LOCATOR_SETs across Addressing Realms 563 It is possible for HIP associations to use these mechanisms to 564 migrate their HIP associations and security associations from 565 addresses in the IPv4 addressing realm to IPv6 or vice versa. It may 566 be possible for a state to arise in which both hosts are only using 567 locators in different addressing realms, but in such a case, some 568 type of mechanism for interworking between the different realms must 569 be employed; such techniques are outside the scope of the present 570 text. 572 4.11. Interaction with Security Associations 574 A host may establish any number of security associations (or SPIs) 575 with a peer. The main purpose of having multiple SPIs with a peer is 576 to group the addresses into collections that are likely to experience 577 fate sharing, or to perform load balancing. 579 A basic property of HIP SAs is that the inbound IP address is not 580 used to lookup the incoming SA. However, the use of different source 581 and destination addresses typically leads to different paths, with 582 different latencies in the network, and if packets were to arrive via 583 an arbitrary destination IP address (or path) for a given SPI, the 584 reordering due to different latencies may cause some packets to fall 585 outside of the ESP anti-replay window. For this reason, HIP provides 586 a mechanism to affiliate destination addresses with inbound SPIs, 587 when there is a concern that anti-replay windows might be violated. 588 In this sense, we can say that a given inbound SPI has an "affinity" 589 for certain inbound IP addresses, and this affinity is communicated 590 to the peer host. Each physical interface SHOULD have a separate SA, 591 unless the ESP anti-replay window is extended or disabled. 593 Moreover, even when the destination addresses used for a particular 594 SPI are held constant, the use of different source interfaces may 595 also cause packets to fall outside of the ESP anti-replay window, 596 since the path traversed is often affected by the source address or 597 interface used. A host has no way to influence the source interface 598 on which a peer sends its packets on a given SPI. A host SHOULD 599 consistently use the same source interface and address when sending 600 to a particular destination IP address and SPI. For this reason, a 601 host may find it useful to change its SPI or at least reset its ESP 602 anti-replay window when the peer host readdresses. 604 5. Processing Rules 606 Basic processing rules for the LOCATOR_SET parameter are specified in 607 [I-D.ietf-hip-rfc5206-bis]. This document focuses on multihoming- 608 specific rules. 610 5.1. Sending LOCATOR_SETs 612 The decision of when to send a LOCATOR_SET, and which addresses to 613 include, is a local policy issue. [I-D.ietf-hip-rfc5206-bis] 614 recommends that a host should send a LOCATOR_SET whenever it 615 recognizes a change of its IP addresses in use on an active HIP 616 association, and assumes that the change is going to last at least 617 for a few seconds. It is possible to delay the exposure of 618 additional locators to the peer, and to send data from previously 619 unannounced locators, as might arise in certain mobility or 620 multihoming situations. 622 When a host decides to inform its peers about changes in its IP 623 addresses, it has to decide how to group the various addresses with 624 SPIs. The grouping should consider also whether middlebox 625 interaction requires sending the same LOCATOR_SET in separate UPDATEs 626 on different paths. Since each SPI is associated with a different 627 Security Association, the grouping policy may also be based on ESP 628 anti-replay protection considerations. In the typical case, simply 629 basing the grouping on actual kernel level physical and logical 630 interfaces may be the best policy. Grouping policy is outside of the 631 scope of this document. 633 Locators corresponding to tunnel interfaces (e.g. IPsec tunnel 634 interfaces or Mobile IP home addresses) or other virtual interfaces 635 MAY be announced in a LOCATOR_SET, but implementations SHOULD avoid 636 announcing such locators as preferred locators if more direct paths 637 may be obtained by instead preferring locators from non-tunneling 638 interfaces if such locators provide a more direct path to the HIP 639 peer. 641 Hosts MUST NOT announce broadcast or multicast addresses in 642 LOCATOR_SETs. Link-local addresses MAY be announced to peers that 643 are known to be neighbors on the same link, such as when the IP 644 destination address of a peer is also link-local. The announcement 645 of link-local addresses in this case is a policy decision; link-local 646 addresses used as Preferred locators will create reachability 647 problems when the host moves to another link. In any case, link- 648 local addresses MUST NOT be announced to a peer unless that peer is 649 known to be on the same link. 651 Once the host has decided on the groups and assignment of addresses 652 to the SPIs, it creates a LOCATOR_SET parameter that serves as a 653 complete representation of the addresses and associated SPIs intended 654 for active use. We now describe a few cases introduced in Section 4. 655 We assume that the Traffic Type for each locator is set to "0" (other 656 values for Traffic Type may be specified in documents that separate 657 the HIP control plane from data plane traffic). Other mobility and 658 multihoming cases are possible but are left for further 659 experimentation. 661 1. Host multihoming (addition of an address). We only describe the 662 simple case of adding an additional address to a (previously) 663 single-homed, non-mobile host. The host MAY choose to simply 664 announce this address to the peer, for fault tolerance. To do 665 this, the multihomed host creates a LOCATOR_SET parameter 666 including the existing address and SPI as a Type "1" Locator, and 667 the new address as a Type "0" Locator. The host sends this in an 668 UPDATE message with SEQ parameter, which is acknowledged by the 669 peer. 671 2. The host MAY set up a new SA pair between this new address and an 672 address of the peer host. To do this, the multihomed host 673 creates a new inbound SA and creates a new SPI. For the outgoing 674 UPDATE message, it inserts an ESP_INFO parameter with an OLD SPI 675 field of "0", a NEW SPI field corresponding to the new SPI, and a 676 KEYMAT Index as selected by local policy. The host adds to the 677 UPDATE message a LOCATOR_SET with two Type "1" Locators: the 678 original address and SPI active on the association, and the new 679 address and new SPI being added (with the SPI matching the NEW 680 SPI contained in the ESP_INFO). The Preferred bit SHOULD be set 681 depending on the policy to tell the peer host which of the two 682 locators is preferred. The UPDATE also contains a SEQ parameter 683 and optionally a DIFFIE_HELLMAN parameter, and follows rekeying 684 procedures with respect to this new address. The UPDATE message 685 SHOULD be sent to the peer's Preferred address with a source 686 address corresponding to the new locator. 688 The sending of multiple LOCATOR_SETs is unsupported. Note that the 689 inclusion of LOCATOR_SET in an R1 packet requires the use of Type "0" 690 locators since no SAs are set up at that point. 692 5.2. Handling Received LOCATOR_SETs 694 A host SHOULD be prepared to receive a LOCATOR_SET parameter in the 695 following HIP packets: R1, I2, R2, and UPDATE. 697 This document describes sending both ESP_INFO and LOCATOR_SET 698 parameters in an UPDATE. The ESP_INFO parameter is included when 699 there is a need to rekey or key a new SPI, and can otherwise be 700 included for the possible benefit of HIP-aware middleboxes. The 701 LOCATOR_SET parameter contains a complete map of the locators that 702 the host wishes to make or keep active for the HIP association. 704 In general, the processing of a LOCATOR_SET depends upon the packet 705 type in which it is included. Here, we describe only the case in 706 which ESP_INFO is present and a single LOCATOR_SET and ESP_INFO are 707 sent in an UPDATE message; other cases are for further study. The 708 steps below cover each of the cases described in Section 5.1. 710 The processing of ESP_INFO and LOCATOR_SET parameters is intended to 711 be modular and support future generalization to the inclusion of 712 multiple ESP_INFO and/or multiple LOCATOR_SET parameters. A host 713 SHOULD first process the ESP_INFO before the LOCATOR_SET, since the 714 ESP_INFO may contain a new SPI value mapped to an existing SPI, while 715 a Type "1" locator will only contain a reference to the new SPI. 717 When a host receives a validated HIP UPDATE with a LOCATOR_SET and 718 ESP_INFO parameter, it processes the ESP_INFO as follows. The 719 ESP_INFO parameter indicates whether an SA is being rekeyed, created, 720 deprecated, or just identified for the benefit of middleboxes. The 721 host examines the OLD SPI and NEW SPI values in the ESP_INFO 722 parameter: 724 1. (no rekeying) If the OLD SPI is equal to the NEW SPI and both 725 correspond to an existing SPI, the ESP_INFO is gratuitous 726 (provided for middleboxes) and no rekeying is necessary. 728 2. (rekeying) If the OLD SPI indicates an existing SPI and the NEW 729 SPI is a different non-zero value, the existing SA is being 730 rekeyed and the host follows HIP ESP rekeying procedures by 731 creating a new outbound SA with an SPI corresponding to the NEW 732 SPI, with no addresses bound to this SPI. Note that locators in 733 the LOCATOR_SET parameter will reference this new SPI instead of 734 the old SPI. 736 3. (new SA) If the OLD SPI value is zero and the NEW SPI is a new 737 non-zero value, then a new SA is being requested by the peer. 738 This case is also treated like a rekeying event; the receiving 739 host must create a new SA and respond with an UPDATE ACK. 741 4. (deprecating the SA) If the OLD SPI indicates an existing SPI and 742 the NEW SPI is zero, the SA is being deprecated and all locators 743 uniquely bound to the SPI are put into the DEPRECATED state. 745 If none of the above cases apply, a protocol error has occurred and 746 the processing of the UPDATE is stopped. 748 Next, the locators in the LOCATOR_SET parameter are processed. For 749 each locator listed in the LOCATOR_SET parameter, check that the 750 address therein is a legal unicast or anycast address. That is, the 751 address MUST NOT be a broadcast or multicast address. Note that some 752 implementations MAY accept addresses that indicate the local host, 753 since it may be allowed that the host runs HIP with itself. 755 For each Type "1" address listed in the LOCATOR_SET parameter, the 756 host checks whether the address is already bound to the SPI 757 indicated. If the address is already bound, its lifetime is updated. 758 If the status of the address is DEPRECATED, the status is changed to 759 UNVERIFIED. If the address is not already bound, the address is 760 added, and its status is set to UNVERIFIED. If there exist remaining 761 addresses corresponding to the SPI that were NOT listed in the 762 LOCATOR_SET parameter, the host sets the status of such addresses to 763 DEPRECATED. 765 For each Type "0" address listed in the LOCATOR_SET parameter, if the 766 status of the address is DEPRECATED, or the address was not 767 previously known, the status is changed to UNVERIFIED. The host MAY 768 choose to associate this address with one or more SAs. The 769 association with different SAs is a local policy decision, unless the 770 peer has indicated that the address is Preferred, in which case the 771 address should be put into use on a SA that is prioritized in the 772 security policy database. 774 As a result, at the end of processing, the addresses listed in the 775 LOCATOR_SET parameter have either a state of UNVERIFIED or ACTIVE, 776 and any old addresses on the old SA not listed in the LOCATOR_SET 777 parameter have a state of DEPRECATED. 779 Once the host has processed the locators, if the LOCATOR_SET 780 parameter contains a new Preferred locator, the host SHOULD initiate 781 a change of the Preferred locator. This requires that the host first 782 verifies reachability of the associated address, and only then 783 changes the Preferred locator; see Section 5.4. 785 If a host receives a locator with an unsupported Locator Type, and 786 when such a locator is also declared to be the Preferred locator for 787 the peer, the host SHOULD send a NOTIFY error with a Notify Message 788 Type of LOCATOR_TYPE_UNSUPPORTED, with the Notification Data field 789 containing the locator(s) that the receiver failed to process. 790 Otherwise, a host MAY send a NOTIFY error if a (non-preferred) 791 locator with an unsupported Locator Type is received in a LOCATOR_SET 792 parameter. 794 5.3. Verifying Address Reachability 796 Address verification is defined in [I-D.ietf-hip-rfc5206-bis]. 798 When address verification is in progress for a new Preferred locator, 799 the host SHOULD select a different locator listed as ACTIVE, if one 800 such locator is available, to continue communications until address 801 verification completes. Alternatively, the host MAY use the new 802 Preferred locator while in UNVERIFIED status to the extent Credit- 803 Based Authorization permits. Credit-Based Authorization is explained 804 in [I-D.ietf-hip-rfc5206-bis]. Once address verification succeeds, 805 the status of the new Preferred locator changes to ACTIVE. 807 5.4. Changing the Preferred Locator 809 A host MAY want to change the Preferred outgoing locator for 810 different reasons, e.g., because traffic information or ICMP error 811 messages indicate that the currently used preferred address may have 812 become unreachable. Another reason may be due to receiving a 813 LOCATOR_SET parameter that has the "P" bit set. 815 To change the Preferred locator, the host initiates the following 816 procedure: 818 1. If the new Preferred locator has ACTIVE status, the Preferred 819 locator is changed and the procedure succeeds. 821 2. If the new Preferred locator has UNVERIFIED status, the host 822 starts to verify its reachability. The host SHOULD use a 823 different locator listed as ACTIVE until address verification 824 completes if one such locator is available. Alternatively, the 825 host MAY use the new Preferred locator, even though in UNVERIFIED 826 status, to the extent Credit-Based Authorization permits. Once 827 address verification succeeds, the status of the new Preferred 828 locator changes to ACTIVE and its use is no longer governed by 829 Credit-Based Authorization. 831 3. If the peer host has not indicated a preference for any address, 832 then the host picks one of the peer's ACTIVE addresses randomly 833 or according to policy. This case may arise if, for example, 834 ICMP error messages that deprecate the Preferred locator arrive, 835 but the peer has not yet indicated a new Preferred locator. 837 4. If the new Preferred locator has DEPRECATED status and there is 838 at least one non-deprecated address, the host selects one of the 839 non-deprecated addresses as a new Preferred locator and 840 continues. If the selected address is UNVERIFIED, the address 841 verification procedure described above will apply. 843 6. Security Considerations 845 This document extends the scope of host mobility solutions defined in 846 [I-D.ietf-hip-rfc5206-bis] to include also host multihoming, and as a 847 result, many of the same security considerations for mobility also 848 pertain to multihoming. In particular, [I-D.ietf-hip-rfc5206-bis] 849 describes how HIP host mobility is resistant to different types of 850 impersonation attacks and DoS attacks. 852 The security considerations for this document are similar to those of 853 [I-D.ietf-hip-rfc5206-bis] because the strong authentication 854 capabilities for mobility also carry over to end-host multihoming. 856 [RFC4218] provides a threat analysis for IPv6 multihoming, and the 857 remainder of this section describes how HIP host multihoming 858 addresses those threats. 860 The high-level threats discussed in [RFC4218] involve redirection 861 attacks for the purposes of packet recording, data manipulation, and 862 availability. There are a few types of attackers to consider: on- 863 path attackers, off-path attackers, and malicious hosts. 865 [RFC4218] also makes the comment that in identifier/locator split 866 solutions such as HIP, application security mechanisms should be tied 867 to the identifier, not the locator, and attacks on the identifier 868 mechanism and on the mechanism binding locators to the identifier are 869 of concern. This document does not consider the former issue 870 (application layer security bindings) to be within scope. The latter 871 issue (locator bindings to identifier) is directly addressed by the 872 cryptographic protections of the HIP protocol, in that locators 873 associated to an identifier are listed in HIP packets that are signed 874 using the identifier key. 876 Section 3.1 of [RFC4218] lists several classes of security 877 configurations in use in the Internet. HIP maps to the fourth 878 (strong identifier) and fifth ("leap-of-faith") categories, the 879 latter being associated with the optional opportunistic mode of HIP 880 operation. The remainder of Section 3 describe existing security 881 problems in the Internet, and comments that the goal of a multihoming 882 solution is not to solve them specifically but rather not to make any 883 of them worse. HIP multihoming should not increase the severity of 884 the identified risks. One concern for both HIP mobility and 885 multihoming is the susceptibility of the mechanisms to misuse for 886 flooding based redirections due to a malicious host. The mechanisms 887 described in [I-D.ietf-hip-rfc5206-bis] for address verification are 888 important in this regard. 890 Regarding the new types of threats introduced by multihoming 891 (Section 4 of [RFC4218]), HIP multihoming should not introduce new 892 concerns. Classic and premeditated redirection are prevented by the 893 strong authentication in HIP messages. Third-party denial-of-service 894 attacks are prevented by the address verification mechanism. Replay 895 attacks can be avoided via use of replay protection in Encapsulating 896 Security Payload (ESP) Security Associations (SAs). In addition, 897 accepting packets from unknown locators is protected by either the 898 strong authentication in the HIP control packets, or by the ESP-based 899 encryption in use for data packets. 901 Finally, the HIP mechanisms are designed to limit the ability to 902 introduce DoS on the mechanisms themselves (Section 7 of [RFC4218]). 903 Care is taken in the HIP base exchange to avoid creating state or 904 performing much work before hosts can authenticate one another. A 905 malicious host involved in HIP multihoming with another host might 906 attempt to misuse the mechanisms for multihoming by, for instance, 907 increasing the state required or inducing a resource limitation 908 attack by sending too many candidate locators to the peer host. 909 Therefore, implementations supporting the multihoming extensions 910 should consider to avoid accepting large numbers of peer locators, 911 and to rate limit any UPDATE messages being exchanged. 913 7. IANA Considerations 915 This document has no requests for IANA actions. 917 8. Authors and Acknowledgments 919 This document contains content that was originally included in 920 RFC5206. Pekka Nikander and Jari Arkko originated RFC5206, and 921 Christian Vogt and Thomas Henderson (editor) later joined as co- 922 authors. Also in RFC5206, Greg Perkins contributed the initial draft 923 of the security section, and Petri Jokela was a co-author of the 924 initial individual submission. 926 The authors thank Miika Komu, Mika Kousa, Jeff Ahrenholz, and Jan 927 Melen for many improvements to the document. Concepts from a paper 928 on host multihoming across address families, by Samu Varjonen, Miika 929 Komu, and Andrei Gurtov, contributed to this revised version. 931 9. References 933 9.1. Normative references 935 [I-D.ietf-hip-rfc5206-bis] 936 Henderson, T., Vogt, C., and J. Arkko, "Host Mobility with 937 the Host Identity Protocol", draft-ietf-hip-rfc5206-bis-12 938 (work in progress), May 2016. 940 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 941 Requirement Levels", BCP 14, RFC 2119, 942 DOI 10.17487/RFC2119, March 1997, 943 . 945 [RFC6724] Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown, 946 "Default Address Selection for Internet Protocol Version 6 947 (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012, 948 . 950 [RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T. 951 Henderson, "Host Identity Protocol Version 2 (HIPv2)", 952 RFC 7401, DOI 10.17487/RFC7401, April 2015, 953 . 955 [RFC7402] Jokela, P., Moskowitz, R., and J. Melen, "Using the 956 Encapsulating Security Payload (ESP) Transport Format with 957 the Host Identity Protocol (HIP)", RFC 7402, 958 DOI 10.17487/RFC7402, April 2015, 959 . 961 9.2. Informative references 963 [RFC4218] Nordmark, E. and T. Li, "Threats Relating to IPv6 964 Multihoming Solutions", RFC 4218, DOI 10.17487/RFC4218, 965 October 2005, . 967 [RFC4303] Kent, S., "IP Encapsulating Security Payload (ESP)", 968 RFC 4303, DOI 10.17487/RFC4303, December 2005, 969 . 971 [RFC5533] Nordmark, E. and M. Bagnulo, "Shim6: Level 3 Multihoming 972 Shim Protocol for IPv6", RFC 5533, DOI 10.17487/RFC5533, 973 June 2009, . 975 Appendix A. Document Revision History 977 To be removed upon publication 979 +----------+--------------------------------------------------------+ 980 | Revision | Comments | 981 +----------+--------------------------------------------------------+ 982 | draft-00 | Initial version with multihoming text imported from | 983 | | RFC5206. | 984 | | | 985 | draft-01 | Document refresh; no other changes. | 986 | | | 987 | draft-02 | Document refresh; no other changes. | 988 | | | 989 | draft-03 | Document refresh; no other changes. | 990 | | | 991 | draft-04 | Document refresh; no other changes. | 992 | | | 993 | draft-05 | Move remaining multihoming material from RFC5206-bis | 994 | | to this document | 995 | | | 996 | | Update lingering references to LOCATOR parameter to | 997 | | LOCATOR_SET | 998 | | | 999 | draft-06 | Document refresh with updated references. | 1000 | | | 1001 | draft-07 | Document refresh; no other changes. | 1002 | | | 1003 | draft-08 | issues 3 and 11: Address complaints of complexity due | 1004 | | to full mesh of SAs for multihoming. | 1005 | | | 1006 | | issue 5: Improve draft based on recommendations for | 1007 | | cross-family handovers in paper by Varjonen et. al. | 1008 | | | 1009 | | issue 7: Clarify and distinguish between load | 1010 | | balancing and fault tolerance use cases. | 1011 | | | 1012 | draft-09 | Update author affiliations, IPR boilerplate to | 1013 | | trust200902, and one stale reference. | 1014 | | | 1015 | draft-10 | Improve security considerations section by reviewing | 1016 | | RFC 4218. | 1017 | | | 1018 | | Clarified comment about applicability of shim6. | 1019 | | | 1020 | draft-11 | Editorial improvements based on last call comments. | 1021 +----------+--------------------------------------------------------+ 1023 Authors' Addresses 1025 Thomas R. Henderson (editor) 1026 University of Washington 1027 Campus Box 352500 1028 Seattle, WA 1029 USA 1031 EMail: tomhend@u.washington.edu 1033 Christian Vogt 1034 Independent 1035 3473 North First Street 1036 San Jose, CA 95134 1037 USA 1039 EMail: mail@christianvogt.net 1041 Jari Arkko 1042 Ericsson 1043 JORVAS FIN-02420 1044 FINLAND 1046 Phone: +358 40 5079256 1047 EMail: jari.arkko@piuha.net