idnits 2.17.1 draft-ietf-hip-rfc4423-bis-01.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 == Line 1071 has weird spacing: '...ncement to th...' == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, 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 (August 24, 2010) is 4993 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) == Unused Reference: 'RFC4025' is defined on line 1051, but no explicit reference was found in the text ** Obsolete normative reference: RFC 5202 (Obsoleted by RFC 7402) ** Obsolete normative reference: RFC 5204 (Obsoleted by RFC 8004) ** Obsolete normative reference: RFC 5205 (Obsoleted by RFC 8005) -- Obsolete informational reference (is this intentional?): RFC 2535 (Obsoleted by RFC 4033, RFC 4034, RFC 4035) -- Obsolete informational reference (is this intentional?): RFC 2766 (Obsoleted by RFC 4966) -- Obsolete informational reference (is this intentional?): RFC 4306 (Obsoleted by RFC 5996) -- Obsolete informational reference (is this intentional?): RFC 4423 (Obsoleted by RFC 9063) Summary: 3 errors (**), 0 flaws (~~), 4 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group R. Moskowitz 3 Internet-Draft ICSA labs 4 Obsoletes: 4423 (if approved) August 24, 2010 5 Intended status: Standards Track 6 Expires: February 25, 2011 8 Host Identity Protocol Architecture 9 draft-ietf-hip-rfc4423-bis-01 11 Abstract 13 This memo describes a new namespace, the Host Identity namespace, and 14 a new protocol layer, the Host Identity Protocol, between the 15 internetworking and transport layers. Herein are presented the 16 basics of the current namespaces, their strengths and weaknesses, and 17 how a new namespace will add completeness to them. The roles of this 18 new namespace in the protocols are defined. 20 This document obsoletes RFC 4423 and addresses the concerns raised by 21 the IESG, particularly that of crypto agility. It incorporates 22 lessons learned from the implementations of RFC 5201 and goes further 23 to explain how HIP works as a secure signalling channel. 25 Status of this Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on February 25, 2011. 42 Copyright Notice 44 Copyright (c) 2010 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 This document may contain material from IETF Documents or IETF 58 Contributions published or made publicly available before November 59 10, 2008. The person(s) controlling the copyright in some of this 60 material may not have granted the IETF Trust the right to allow 61 modifications of such material outside the IETF Standards Process. 62 Without obtaining an adequate license from the person(s) controlling 63 the copyright in such materials, this document may not be modified 64 outside the IETF Standards Process, and derivative works of it may 65 not be created outside the IETF Standards Process, except to format 66 it for publication as an RFC or to translate it into languages other 67 than English. 69 Table of Contents 71 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 72 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . 5 73 2.1. Terms common to other documents . . . . . . . . . . . . . . 5 74 2.2. Terms specific to this and other HIP documents . . . . . . . 5 75 3. Background . . . . . . . . . . . . . . . . . . . . . . . . . 7 76 3.1. A desire for a namespace for computing platforms . . . . . . 7 77 4. Host Identity namespace . . . . . . . . . . . . . . . . . . 9 78 4.1. Host Identifiers . . . . . . . . . . . . . . . . . . . . . . 10 79 4.2. Storing Host Identifiers in DNS . . . . . . . . . . . . . . 10 80 4.3. Host Identity Tag (HIT) . . . . . . . . . . . . . . . . . . 11 81 4.4. Local Scope Identifier (LSI) . . . . . . . . . . . . . . . . 11 82 5. New stack architecture . . . . . . . . . . . . . . . . . . . 11 83 5.1. Transport associations and end-points . . . . . . . . . . . 12 84 6. End-host mobility and multi-homing . . . . . . . . . . . . . 13 85 6.1. Rendezvous mechanism . . . . . . . . . . . . . . . . . . . . 13 86 6.2. Protection against flooding attacks . . . . . . . . . . . . 14 87 7. HIP and IPsec . . . . . . . . . . . . . . . . . . . . . . . 14 88 8. HIP and NATs . . . . . . . . . . . . . . . . . . . . . . . . 15 89 8.1. HIP and TCP checksums . . . . . . . . . . . . . . . . . . . 16 90 9. Multicast . . . . . . . . . . . . . . . . . . . . . . . . . 16 91 10. HIP policies . . . . . . . . . . . . . . . . . . . . . . . . 16 92 11. Benefits of HIP . . . . . . . . . . . . . . . . . . . . . . 17 93 11.1. HIP's answers to NSRG questions . . . . . . . . . . . . . . 18 94 12. Changes from RFC 4423 . . . . . . . . . . . . . . . . . . . 20 95 13. Security considerations . . . . . . . . . . . . . . . . . . 20 96 13.1. HITs used in ACLs . . . . . . . . . . . . . . . . . . . . . 21 97 13.2. Non-security considerations . . . . . . . . . . . . . . . . 22 98 14. IANA considerations . . . . . . . . . . . . . . . . . . . . 23 99 15. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . 23 100 16. References . . . . . . . . . . . . . . . . . . . . . . . . . 23 101 16.1. Normative References . . . . . . . . . . . . . . . . . . . . 23 102 16.2. Informative references . . . . . . . . . . . . . . . . . . . 23 103 Author's Address . . . . . . . . . . . . . . . . . . . . . . 25 105 1. Introduction 107 The Internet has two important global namespaces: Internet Protocol 108 (IP) addresses and Domain Name Service (DNS) names. These two 109 namespaces have a set of features and abstractions that have powered 110 the Internet to what it is today. They also have a number of 111 weaknesses. Basically, since they are all we have, we try and do too 112 much with them. Semantic overloading and functionality extensions 113 have greatly complicated these namespaces. 115 The proposed Host Identity namespace fills an important gap between 116 the IP and DNS namespaces. The Host Identity namespace consists of 117 Host Identifiers (HI). A Host Identifier is cryptographic in its 118 nature; it is the public key of an asymmetric key-pair. Each host 119 will have at least one Host Identity, but it will typically have more 120 than one. Each Host Identity uniquely identifies a single host, 121 i.e., no two hosts have the same Host Identity. The Host Identity, 122 and the corresponding Host Identifier, can either be public (e.g. 123 published in the DNS), or unpublished. Client systems will tend to 124 have both public and unpublished Identities. 126 There is a subtle but important difference between Host Identities 127 and Host Identifiers. An Identity refers to the abstract entity that 128 is identified. An Identifier, on the other hand, refers to the 129 concrete bit pattern that is used in the identification process. 131 Although the Host Identifiers could be used in many authentication 132 systems, such as IKEv2 [RFC4306], the presented architecture 133 introduces a new protocol, called the Host Identity Protocol (HIP), 134 and a cryptographic exchange, called the HIP base exchange; see also 135 Section 7. The HIP protocols under development provide for limited 136 forms of trust between systems, enhance mobility, multi-homing and 137 dynamic IP renumbering, aid in protocol translation / transition, and 138 reduce certain types of denial-of-service (DoS) attacks. 140 When HIP is used, the actual payload traffic between two HIP hosts is 141 typically, but not necessarily, protected with IPsec. The Host 142 Identities are used to create the needed IPsec Security Associations 143 (SAs) and to authenticate the hosts. When IPsec is used, the actual 144 payload IP packets do not differ in any way from standard IPsec 145 protected IP packets. 147 Much has been learned about HIP since [RFC4423] was published. This 148 document expands Host Identities beyond use to enable IP connectivity 149 and security to general interhost secure signalling at any protocol 150 layer. The signal may establish a security association between the 151 hosts, or simply pass information within the channel. 153 2. Terminology 155 2.1. Terms common to other documents 157 +---------------+---------------------------------------------------+ 158 | Term | Explanation | 159 +---------------+---------------------------------------------------+ 160 | Public key | The public key of an asymmetric cryptographic key | 161 | | pair. Used as a publicly known identifier for | 162 | | cryptographic identity authentication. | 163 | | | 164 | Private key | The private or secret key of an asymmetric | 165 | | cryptographic key pair. Assumed to be known only | 166 | | to the party identified by the corresponding | 167 | | public key. Used by the identified party to | 168 | | authenticate its identity to other parties. | 169 | | | 170 | Public key | An asymmetric cryptographic key pair consisting | 171 | pair | of public and private keys. For example, | 172 | | Rivest-Shamir-Adelman (RSA) and Digital Signature | 173 | | Algorithm (DSA) key pairs are such key pairs. | 174 | | | 175 | End-point | A communicating entity. For historical reasons, | 176 | | the term 'computing platform' is used in this | 177 | | document as a (rough) synonym for end-point. | 178 +---------------+---------------------------------------------------+ 180 2.2. Terms specific to this and other HIP documents 182 It should be noted that many of the terms defined herein are 183 tautologous, self-referential or defined through circular reference 184 to other terms. This is due to the succinct nature of the 185 definitions. See the text elsewhere in this document for more 186 elaborate explanations. 188 +---------------+---------------------------------------------------+ 189 | Term | Explanation | 190 +---------------+---------------------------------------------------+ 191 | Computing | An entity capable of communicating and computing, | 192 | platform | for example, a computer. See the definition of | 193 | | 'End-point', above. | 194 | | | 195 | HIP base | A cryptographic protocol; see also Section 7. | 196 | exchange | | 197 | | | 198 | HIP packet | An IP packet that carries a 'Host Identity | 199 | | Protocol' message. | 200 | | | 201 | Host Identity | An abstract concept assigned to a 'computing | 202 | | platform'. See 'Host Identifier', below. | 203 | | | 204 | Host Identity | A name space formed by all possible Host | 205 | namespace | Identifiers. | 206 | | | 207 | Host Identity | A protocol used to carry and authenticate Host | 208 | Protocol | Identifiers and other information. | 209 | | | 210 | Host Identity | A 128-bit datum created by taking a cryptographic | 211 | Tag | hash over a Host Identifier. | 212 | | | 213 | Host | A public key used as a name for a Host Identity. | 214 | Identifier | | 215 | | | 216 | Local Scope | A 32-bit datum denoting a Host Identity. | 217 | Identifier | | 218 | | | 219 | Public Host | A published or publicly known Host Identfier used | 220 | Identifier | as a public name for a Host Identity, and the | 221 | and Identity | corresponding Identity. | 222 | | | 223 | Unpublished | A Host Identifier that is not placed in any | 224 | Host | public directory, and the corresponding Host | 225 | Identifier | Identity. Unpublished Host Identities are | 226 | and Identity | typically short lived in nature, being often | 227 | | replaced and possibly used just once. | 228 | | | 229 | Rendezvous | A mechanism used to locate mobile hosts based on | 230 | Mechanism | their HIT. | 231 +---------------+---------------------------------------------------+ 233 3. Background 235 The Internet is built from three principal components: computing 236 platforms (end-points), packet transport (i.e., internetworking) 237 infrastructure, and services (applications). The Internet exists to 238 service two principal components: people and robotic services 239 (silicon based people, if you will). All these components need to be 240 named in order to interact in a scalable manner. Here we concentrate 241 on naming computing platforms and packet transport elements. 243 There are two principal namespaces in use in the Internet for these 244 components: IP numbers, and Domain Names. Domain Names provide 245 hierarchically assigned names for some computing platforms and some 246 services. Each hierarchy is delegated from the level above; there is 247 no anonymity in Domain Names. Email, HTTP, and SIP addresses all 248 reference Domain Names. 250 IP numbers are a confounding of two namespaces, the names of a host's 251 networking interfaces and the names of the locations ('confounding' 252 is a term used in statistics to discuss metrics that are merged into 253 one with a gain in indexing, but a loss in informational value). The 254 names of locations should be understood as denoting routing direction 255 vectors, i.e., information that is used to deliver packets to their 256 destinations. 258 IP numbers name networking interfaces, and typically only when the 259 interface is connected to the network. Originally, IP numbers had 260 long-term significance. Today, the vast number of interfaces use 261 ephemeral and/or non-unique IP numbers. That is, every time an 262 interface is connected to the network, it is assigned an IP number. 264 In the current Internet, the transport layers are coupled to the IP 265 addresses. Neither can evolve separately from the other. IPng 266 deliberations were strongly shaped by the decision that a 267 corresponding TCPng would not be created. 269 There are three critical deficiencies with the current namespaces. 270 Firstly, dynamic readdressing cannot be directly managed. Secondly, 271 anonymity is not provided in a consistent, trustable manner. 272 Finally, authentication for systems and datagrams is not provided. 273 All of these deficiencies arise because computing platforms are not 274 well named with the current namespaces. 276 3.1. A desire for a namespace for computing platforms 278 An independent namespace for computing platforms could be used in 279 end-to-end operations independent of the evolution of the 280 internetworking layer and across the many internetworking layers. 282 This could support rapid readdressing of the internetworking layer 283 because of mobility, rehoming, or renumbering. 285 If the namespace for computing platforms is based on public-key 286 cryptography, it can also provide authentication services. If this 287 namespace is locally created without requiring registration, it can 288 provide anonymity. 290 Such a namespace (for computing platforms) and the names in it should 291 have the following characteristics: 293 o The namespace should be applied to the IP 'kernel'. The IP kernel 294 is the 'component' between applications and the packet transport 295 infrastructure. 297 o The namespace should fully decouple the internetworking layer from 298 the higher layers. The names should replace all occurrences of IP 299 addresses within applications (like in the Transport Control 300 Block, TCB). This may require changes to the current APIs. In 301 the long run, it is probable that some new APIs are needed. 303 o The introduction of the namespace should not mandate any 304 administrative infrastructure. Deployment must come from the 305 bottom up, in a pairwise deployment. 307 o The names should have a fixed length representation, for easy 308 inclusion in datagram headers and existing programming interfaces 309 (e.g the TCB). 311 o Using the namespace should be affordable when used in protocols. 312 This is primarily a packet size issue. There is also a 313 computational concern in affordability. 315 o Name collisions should be avoided as much as possible. The 316 mathematics of the birthday paradox can be used to estimate the 317 chance of a collision in a given population and hash space. In 318 general, for a random hash space of size n bits, we would expect 319 to obtain a collision after approximately 1.2*sqrt(2**n) hashes 320 were obtained. For 64 bits, this number is roughly 4 billion. A 321 hash size of 64 bits may be too small to avoid collisions in a 322 large population; for example, there is a 1% chance of collision 323 in a population of 640M. For 100 bits (or more), we would not 324 expect a collision until approximately 2**50 (1 quadrillion) 325 hashes were generated. 327 o The names should have a localized abstraction so that it can be 328 used in existing protocols and APIs. 330 o It must be possible to create names locally. This can provide 331 anonymity at the cost of making resolvability very difficult. 333 * Sometimes the names may contain a delegation component. This 334 is the cost of resolvability. 336 o The namespace should provide authentication services. 338 o The names should be long lived, but replaceable at any time. This 339 impacts access control lists; short lifetimes will tend to result 340 in tedious list maintenance or require a namespace infrastructure 341 for central control of access lists. 343 In this document, a new namespace approaching these ideas is called 344 the Host Identity namespace. Using Host Identities requires its own 345 protocol layer, the Host Identity Protocol, between the 346 internetworking and transport layers. The names are based on public- 347 key cryptography to supply authentication services. Properly 348 designed, it can deliver all of the above stated requirements. 350 4. Host Identity namespace 352 A name in the Host Identity namespace, a Host Identifier (HI), 353 represents a statistically globally unique name for naming any system 354 with an IP stack. This identity is normally associated with, but not 355 limited to, an IP stack. A system can have multiple identities, some 356 'well known', some unpublished or 'anonymous'. A system may self- 357 assert its own identity, or may use a third-party authenticator like 358 DNSSEC [RFC2535], PGP, or X.509 to 'notarize' the identity assertion. 359 It is expected that the Host Identifiers will initially be 360 authenticated with DNSSEC and that all implementations will support 361 DNSSEC as a minimal baseline. 363 In theory, any name that can claim to be 'statistically globally 364 unique' may serve as a Host Identifier. However, in the authors' 365 opinion, a public key of a 'public key pair' makes the best Host 366 Identifier. As will be specified in the Host Identity Protocol 367 specification, a public-key-based HI can authenticate the HIP packets 368 and protect them for man-in-the-middle attacks. Since authenticated 369 datagrams are mandatory to provide much of HIP's denial-of-service 370 protection, the Diffie-Hellman exchange in HIP has to be 371 authenticated. Thus, only public-key HI and authenticated HIP 372 messages are supported in practice. In this document, the non- 373 cryptographic forms of HI and HIP are presented to complete the 374 theory of HI, but they should not be implemented as they could 375 produce worse denial-of-service attacks than the Internet has without 376 Host Identity. 378 4.1. Host Identifiers 380 Host Identity adds two main features to Internet protocols. The 381 first is a decoupling of the internetworking and transport layers; 382 see Section 5. This decoupling will allow for independent evolution 383 of the two layers. Additionally, it can provide end-to-end services 384 over multiple internetworking realms. The second feature is host 385 authentication. Because the Host Identifier is a public key, this 386 key can be used for authentication in security protocols like IPsec. 388 The only completely defined structure of the Host Identity is that of 389 a public/private key pair. In this case, the Host Identity is 390 referred to by its public component, the public key. Thus, the name 391 representing a Host Identity in the Host Identity namespace, i.e., 392 the Host Identifier, is the public key. In a way, the possession of 393 the private key defines the Identity itself. If the private key is 394 possessed by more than one node, the Identity can be considered to be 395 a distributed one. 397 Architecturally, any other Internet naming convention might form a 398 usable base for Host Identifiers. However, non-cryptographic names 399 should only be used in situations of high trust - low risk. That is 400 any place where host authentication is not needed (no risk of host 401 spoofing) and no use of IPsec. However, at least for interconnected 402 networks spanning several operational domains, the set of 403 environments where the risk of host spoofing allowed by non- 404 cryptographic Host Identifiers is acceptable is the null set. Hence, 405 the current HIP documents do not specify how to use any other types 406 of Host Identifiers but public keys. 408 The actual Host Identities are never directly used in any Internet 409 protocols. The corresponding Host Identifiers (public keys) may be 410 stored in various DNS or LDAP directories as identified elsewhere in 411 this document, and they are passed in the HIP base exchange. A Host 412 Identity Tag (HIT) is used in other protocols to represent the Host 413 Identity. Another representation of the Host Identities, the Local 414 Scope Identifier (LSI), can also be used in protocols and APIs. 416 4.2. Storing Host Identifiers in DNS 418 The public Host Identifiers should be stored in DNS; the unpublished 419 Host Identifiers should not be stored anywhere (besides the 420 communicating hosts themselves). The (public) HI is stored in a new 421 RR type. This RR type is defined in HIP DNS Extension [RFC5205]. 423 Alternatively, or in addition to storing Host Identifiers in the DNS, 424 they may be stored in various kinds of Public Key Infrastructure 425 (PKI). Such a practice may allow them to be used for purposes other 426 than pure host identification. 428 4.3. Host Identity Tag (HIT) 430 A Host Identity Tag is a 128-bit representation for a Host Identity. 431 It is created by taking a cryptographic hash over the corresponding 432 Host Identifier. There are two advantages of using a hash over using 433 the Host Identifier in protocols. Firstly, its fixed length makes 434 for easier protocol coding and also better manages the packet size 435 cost of this technology. Secondly, it presents the identity in a 436 consistent format to the protocol independent of the cryptographic 437 algorithms used. 439 In the HIP packets, the HITs identify the sender and recipient of a 440 packet. Consequently, a HIT should be unique in the whole IP 441 universe as long as it is being used. In the extremely rare case of 442 a single HIT mapping to more than one Host Identity, the Host 443 Identifiers (public keys) will make the final difference. If there 444 is more than one public key for a given node, the HIT acts as a hint 445 for the correct public key to use. 447 4.4. Local Scope Identifier (LSI) 449 An LSI is a 32-bit localized representation for a Host Identity. The 450 purpose of an LSI is to facilitate using Host Identities in existing 451 protocols and APIs. LSI's advantage over HIT is its size; its 452 disadvantage is its local scope. 454 Examples of how LSIs can be used include: as the address in an FTP 455 command and as the address in a socket call. Thus, LSIs act as a 456 bridge for Host Identities into IPv4-based protocols and APIs. LSIs 457 also make it possible for some IPv4 applications to run over an IPv6 458 network. 460 5. New stack architecture 462 One way to characterize Host Identity is to compare the proposed new 463 architecture with the current one. As discussed above, the IP 464 addresses can be seen to be a confounding of routing direction 465 vectors and interface names. Using the terminology from the IRTF 466 Name Space Research Group Report [I-D.irtf-nsrg-report] and, e.g., 467 the unpublished Internet-Draft Endpoints and Endpoint Names 468 [chiappa-endpoints], the IP addresses currently embody the dual role 469 of locators and end-point identifiers. That is, each IP address 470 names a topological location in the Internet, thereby acting as a 471 routing direction vector, or locator. At the same time, the IP 472 address names the physical network interface currently located at the 473 point-of-attachment, thereby acting as a end-point name. 475 In the HIP architecture, the end-point names and locators are 476 separated from each other. IP addresses continue to act as locators. 477 The Host Identifiers take the role of end-point identifiers. It is 478 important to understand that the end-point names based on Host 479 Identities are slightly different from interface names; a Host 480 Identity can be simultaneously reachable through several interfaces. 482 The difference between the bindings of the logical entities are 483 illustrated in Figure 1. 485 Service ------ Socket Service ------ Socket 486 | | 487 | | 488 | | 489 | | 490 End-point | End-point --- Host Identity 491 \ | | 492 \ | | 493 \ | | 494 \ | | 495 Location --- IP address Location --- IP address 497 Figure 1 499 5.1. Transport associations and end-points 501 Architecturally, HIP provides for a different binding of transport- 502 layer protocols. That is, the transport-layer associations, i.e., 503 TCP connections and UDP associations, are no longer bound to IP 504 addresses but to Host Identities. 506 It is possible that a single physical computer hosts several logical 507 end-points. With HIP, each of these end-points would have a distinct 508 Host Identity. Furthermore, since the transport associations are 509 bound to Host Identities, HIP provides for process migration and 510 clustered servers. That is, if a Host Identity is moved from one 511 physical computer to another, it is also possible to simultaneously 512 move all the transport associations without breaking them. 513 Similarly, if it is possible to distribute the processing of a single 514 Host Identity over several physical computers, HIP provides for 515 cluster based services without any changes at the client end-point. 517 6. End-host mobility and multi-homing 519 HIP decouples the transport from the internetworking layer, and binds 520 the transport associations to the Host Identities (through actually 521 either the HIT or LSI). Consequently, HIP can provide for a degree 522 of internetworking mobility and multi-homing at a low infrastructure 523 cost. HIP mobility includes IP address changes (via any method) to 524 either party. Thus, a system is considered mobile if its IP address 525 can change dynamically for any reason like PPP, DHCP, IPv6 prefix 526 reassignments, or a NAT device remapping its translation. Likewise, 527 a system is considered multi-homed if it has more than one globally 528 routable IP address at the same time. HIP links IP addresses 529 together, when multiple IP addresses correspond to the same Host 530 Identity, and if one address becomes unusable, or a more preferred 531 address becomes available, existing transport associations can easily 532 be moved to another address. 534 When a node moves while communication is already on-going, address 535 changes are rather straightforward. The peer of the mobile node can 536 just accept a HIP or an integrity protected IPsec packet from any 537 address and ignore the source address. However, as discussed in 538 Section 6.2 below, a mobile node must send a HIP readdress packet to 539 inform the peer of the new address(es), and the peer must verify that 540 the mobile node is reachable through these addresses. This is 541 especially helpful for those situations where the peer node is 542 sending data periodically to the mobile node (that is re-starting a 543 connection after the initial connection). 545 6.1. Rendezvous mechanism 547 Making a contact to a mobile node is slightly more involved. In 548 order to start the HIP exchange, the initiator node has to know how 549 to reach the mobile node. Although infrequently moving HIP nodes 550 could use Dynamic DNS [RFC2136] to update their reachability 551 information in the DNS, an alternative to using DNS in this fashion 552 is to use a piece of new static infrastructure to facilitate 553 rendezvous between HIP nodes. 555 The mobile node keeps the rendezvous infrastructure continuously 556 updated with its current IP address(es). The mobile nodes must trust 557 the rendezvous mechanism to properly maintain their HIT and IP 558 address mappings. 560 The rendezvous mechanism is also needed if both of the nodes happen 561 to change their address at the same time, either because they are 562 mobile and happen to move at the same time, because one of them is 563 off-line for a while, or because of some other reason. In such a 564 case, the HIP readdress packets will cross each other in the network 565 and never reach the peer node. 567 The HIP rendezvous mechanism is defined in HIP Rendezvous [RFC5204]. 569 6.2. Protection against flooding attacks 571 Although the idea of informing about address changes by simply 572 sending packets with a new source address appears appealing, it is 573 not secure enough. That is, even if HIP does not rely on the source 574 address for anything (once the base exchange has been completed), it 575 appears to be necessary to check a mobile node's reachability at the 576 new address before actually sending any larger amounts of traffic to 577 the new address. 579 Blindly accepting new addresses would potentially lead to flooding 580 Denial-of-Service attacks against third parties [RFC4225]. In a 581 distributed flooding attack an attacker opens high volume HIP 582 connections with a large number of hosts (using unpublished HIs), and 583 then claims to all of these hosts that it has moved to a target 584 node's IP address. If the peer hosts were to simply accept the move, 585 the result would be a packet flood to the target node's address. To 586 close this attack, HIP includes an address check mechanism where the 587 reachability of a node is separately checked at each address before 588 using the address for larger amounts of traffic. 590 Whenever HIP is used between two hosts that fully trust each other, 591 the hosts may optionally decide to skip the address tests. However, 592 such performance optimization must be restricted to peers that are 593 known to be trustworthy and capable of protecting themselves from 594 malicious software. 596 7. HIP and IPsec 598 The preferred way of implementing HIP is to use IPsec to carry the 599 actual data traffic. As of today, the only completely defined method 600 is to use IPsec Encapsulated Security Payload (ESP) to carry the data 601 packets [RFC5202]. In the future, other ways of transporting payload 602 data may be developed, including ones that do not use cryptographic 603 protection. 605 In practice, the HIP base exchange uses the cryptographic Host 606 Identifiers to set up a pair of ESP Security Associations (SAs) to 607 enable ESP in an end-to-end manner. This is implemented in a way 608 that can span addressing realms. 610 While it would be possible, at least in theory, to use some existing 611 cryptographic protocol, such as IKEv2 together with Host Identifiers, 612 to establish the needed SAs, HIP defines a new protocol. There are a 613 number of historical reasons for this, and there are also a few 614 architectural reasons. First, IKE (and IKEv2) were not designed with 615 middle boxes in mind. As adding a new naming layer allows one to 616 potentially add a new forwarding layer (see Section 8, below), it is 617 very important that the HIP protocols are friendly towards any middle 618 boxes. 620 Second, from a conceptual point of view, the IPsec Security Parameter 621 Index (SPI) in ESP provides a simple compression of the HITs. This 622 does require per-HIT-pair SAs (and SPIs), and a decrease of policy 623 granularity over other Key Management Protocols, such as IKE and 624 IKEv2. In particular, the current thinking is limited to a situation 625 where, conceptually, there is only one pair of SAs between any given 626 pair of HITs. In other words, from an architectural point of view, 627 HIP only supports host-to-host (or endpoint-to-endpoint) Security 628 Associations. If two hosts need more pairs of parallel SAs, they 629 should use separate HITs for that. However, future HIP extensions 630 may provide for more granularity and creation of several ESP SAs 631 between a pair of HITs. 633 Since HIP is designed for host usage, not for gateways or so called 634 Bump-in-the-Wire (BITW) implementations, only ESP transport mode is 635 supported. An ESP SA pair is indexed by the SPIs and the two HITs 636 (both HITs since a system can have more than one HIT). The SAs need 637 not to be bound to IP addresses; all internal control of the SA is by 638 the HITs. Thus, a host can easily change its address using Mobile 639 IP, DHCP, PPP, or IPv6 readdressing and still maintain the SAs. 640 Since the transports are bound to the SA (via an LSI or a HIT), any 641 active transport is also maintained. Thus, real-world conditions 642 like loss of a PPP connection and its re-establishment or a mobile 643 handover will not require a HIP negotiation or disruption of 644 transport services [Bel1998]. 646 Since HIP does not negotiate any SA lifetimes, all lifetimes are 647 local policy. The only lifetimes a HIP implementation must support 648 are sequence number rollover (for replay protection), and SA timeout. 649 An SA times out if no packets are received using that SA. 650 Implementations may support lifetimes for the various ESP transforms. 652 8. HIP and NATs 654 Passing packets between different IP addressing realms requires 655 changing IP addresses in the packet header. This may happen, for 656 example, when a packet is passed between the public Internet and a 657 private address space, or between IPv4 and IPv6 networks. The 658 address translation is usually implemented as Network Address 659 Translation (NAT) [RFC3022] or NAT Protocol translation (NAT-PT) 660 [RFC2766]. 662 In a network environment where identification is based on the IP 663 addresses, identifying the communicating nodes is difficult when NAT 664 is used. With HIP, the transport-layer end-points are bound to the 665 Host Identities. Thus, a connection between two hosts can traverse 666 many addressing realm boundaries. The IP addresses are used only for 667 routing purposes; they may be changed freely during packet traversal. 669 For a HIP-based flow, a HIP-aware NAT or NAT-PT system tracks the 670 mapping of HITs, and the corresponding IPsec SPIs, to an IP address. 671 The NAT system has to learn mappings both from HITs and from SPIs to 672 IP addresses. Many HITs (and SPIs) can map to a single IP address on 673 a NAT, simplifying connections on address poor NAT interfaces. The 674 NAT can gain much of its knowledge from the HIP packets themselves; 675 however, some NAT configuration may be necessary. 677 NAT systems cannot touch the datagrams within the IPsec envelope, 678 thus application-specific address translation must be done in the end 679 systems. HIP provides for 'Distributed NAT', and uses the HIT or the 680 LSI as a placeholder for embedded IP addresses. 682 8.1. HIP and TCP checksums 684 There is no way for a host to know if any of the IP addresses in an 685 IP header are the addresses used to calculate the TCP checksum. That 686 is, it is not feasible to calculate the TCP checksum using the actual 687 IP addresses in the pseudo header; the addresses received in the 688 incoming packet are not necessarily the same as they were on the 689 sending host. Furthermore, it is not possible to recompute the 690 upper-layer checksums in the NAT/NAT-PT system, since the traffic is 691 IPsec protected. Consequently, the TCP and UDP checksums are 692 calculated using the HITs in the place of the IP addresses in the 693 pseudo header. Furthermore, only the IPv6 pseudo header format is 694 used. This provides for IPv4 / IPv6 protocol translation. 696 9. Multicast 698 There was little if any concrete thoughts about how HIP might affect 699 IP-layer or application-layer multicast. 701 10. HIP policies 703 There are a number of variables that will influence the HIP exchanges 704 that each host must support. All HIP implementations should support 705 at least 2 HIs, one to publish in DNS and an unpublished one for 706 anonymous usage. Although unpublished HIs will be rarely used as 707 responder HIs, they are likely be common for initiators. Support for 708 multiple HIs is recommended. 710 Many initiators would want to use a different HI for different 711 responders. The implementations should provide for a policy of 712 initiator HIT to responder HIT. This policy should also include 713 preferred transforms and local lifetimes. 715 Responders would need a similar policy, describing the hosts allowed 716 to participate in HIP exchanges, and the preferred transforms and 717 local lifetimes. 719 11. Benefits of HIP 721 In the beginning, the network layer protocol (i.e., IP) had the 722 following four "classic" invariants: 724 o Non-mutable: The address sent is the address received. 726 o Non-mobile: The address doesn't change during the course of an 727 "association". 729 o Reversible: A return header can always be formed by reversing the 730 source and destination addresses. 732 o Omniscient: Each host knows what address a partner host can use to 733 send packets to it. 735 Actually, the fourth can be inferred from 1 and 3, but it is worth 736 mentioning for reasons that will be obvious soon if not already. 738 In the current "post-classic" world, we are intentionally trying to 739 get rid of the second invariant (both for mobility and for multi- 740 homing), and we have been forced to give up the first and the fourth. 741 Realm Specific IP [RFC3102] is an attempt to reinstate the fourth 742 invariant without the first invariant. IPv6 is an attempt to 743 reinstate the first invariant. 745 Few systems on the Internet have DNS names that are meaningful. That 746 is, if they have a Fully Qualified Domain Name (FQDN), that name 747 typically belongs to a NAT device or a dial-up server, and does not 748 really identify the system itself but its current connectivity. 749 FQDNs (and their extensions as email names) are application-layer 750 names; more frequently naming services than a particular system. 751 This is why many systems on the Internet are not registered in the 752 DNS; they do not have services of interest to other Internet hosts. 754 DNS names are references to IP addresses. This only demonstrates the 755 interrelationship of the networking and application layers. DNS, as 756 the Internet's only deployed, distributed database is also the 757 repository of other namespaces, due in part to DNSSEC and application 758 specific key records. Although each namespace can be stretched (IP 759 with v6, DNS with KEY records), neither can adequately provide for 760 host authentication or act as a separation between internetworking 761 and transport layers. 763 The Host Identity (HI) namespace fills an important gap between the 764 IP and DNS namespaces. An interesting thing about the HI is that it 765 actually allows one to give up all but the 3rd network-layer 766 invariant. That is to say, as long as the source and destination 767 addresses in the network-layer protocol are reversible, then things 768 work ok because HIP takes care of host identification, and 769 reversibility allows one to get a packet back to one's partner host. 770 You do not care if the network-layer address changes in transit 771 (mutable) and you don't care what network-layer address the partner 772 is using (non-omniscient). 774 11.1. HIP's answers to NSRG questions 776 The IRTF Name Space Research Group has posed a number of evaluating 777 questions in their report [I-D.irtf-nsrg-report]. In this section, 778 we provide answers to these questions. 780 1. How would a stack name improve the overall functionality of the 781 Internet? 783 HIP decouples the internetworking layer from the transport 784 layer, allowing each to evolve separately. The decoupling 785 makes end-host mobility and multi-homing easier, also across 786 IPv4 and IPv6 networks. HIs make network renumbering easier, 787 and they also make process migration and clustered servers 788 easier to implement. Furthermore, being cryptographic in 789 nature, they provide the basis for solving the security 790 problems related to end-host mobility and multi-homing. 792 2. What does a stack name look like? 794 A HI is a cryptographic public key. However, instead of using 795 the keys directly, most protocols use a fixed size hash of the 796 public key. 798 3. What is its lifetime? 800 HIP provides both stable and temporary Host Identifiers. 801 Stable HIs are typically long lived, with a lifetime of years 802 or more. The lifetime of temporary HIs depends on how long 803 the upper-layer connections and applications need them, and 804 can range from a few seconds to years. 806 4. Where does it live in the stack? 808 The HIs live between the transport and internetworking layers. 810 5. How is it used on the end points 812 The Host Identifiers may be used directly or indirectly (in 813 the form of HITs or LSIs) by applications when they access 814 network services. Additionally, the Host Identifiers, as 815 public keys, are used in the built in key agreement protocol, 816 called the HIP base exchange, to authenticate the hosts to 817 each other. 819 6. What administrative infrastructure is needed to support it? 821 In some environments, it is possible to use HIP 822 opportunistically, without any infrastructure. However, to 823 gain full benefit from HIP, the HIs must be stored in the DNS 824 or a PKI, and a new rendezvous mechanism is needed[RFC5205]. 826 7. If we add an additional layer would it make the address list in 827 SCTP unnecessary? 829 Yes 831 8. What additional security benefits would a new naming scheme 832 offer? 834 HIP reduces dependency on IP addresses, making the so called 835 address ownership [Nik2001] problems easier to solve. In 836 practice, HIP provides security for end-host mobility and 837 multi-homing. Furthermore, since HIP Host Identifiers are 838 public keys, standard public key certificate infrastructures 839 can be applied on the top of HIP. 841 9. What would the resolution mechanisms be, or what characteristics 842 of a resolution mechanisms would be required? 844 For most purposes, an approach where DNS names are resolved 845 simultaneously to HIs and IP addresses is sufficient. 847 However, if it becomes necessary to resolve HIs into IP 848 addresses or back to DNS names, a flat resolution 849 infrastructure is needed. Such an infrastructure could be 850 based on the ideas of Distributed Hash Tables, but would 851 require significant new development and deployment. 853 12. Changes from RFC 4423 855 This section will summarize the changes made from [RFC4423]. 857 13. Security considerations 859 HIP takes advantage of the new Host Identity paradigm to provide 860 secure authentication of hosts and to provide a fast key exchange for 861 IPsec. HIP also attempts to limit the exposure of the host to 862 various denial-of-service (DoS) and man-in-the-middle (MitM) attacks. 863 In so doing, HIP itself is subject to its own DoS and MitM attacks 864 that potentially could be more damaging to a host's ability to 865 conduct business as usual. 867 Resource exhausting denial-of-service attacks take advantage of the 868 cost of setting up a state for a protocol on the responder compared 869 to the 'cheapness' on the initiator. HIP allows a responder to 870 increase the cost of the start of state on the initiator and makes an 871 effort to reduce the cost to the responder. This is done by having 872 the responder start the authenticated Diffie-Hellman exchange instead 873 of the initiator, making the HIP base exchange 4 packets long. There 874 are more details on this process in the Host Identity Protocol under 875 development. 877 HIP optionally supports opportunistic negotiation. That is, if a 878 host receives a start of transport without a HIP negotiation, it can 879 attempt to force a HIP exchange before accepting the connection. 880 This has the potential for DoS attacks against both hosts. If the 881 method to force the start of HIP is expensive on either host, the 882 attacker need only spoof a TCP SYN. This would put both systems into 883 the expensive operations. HIP avoids this attack by having the 884 responder send a simple HIP packet that it can pre-build. Since this 885 packet is fixed and easily replayed, the initiator only reacts to it 886 if it has just started a connection to the responder. 888 Man-in-the-middle attacks are difficult to defend against, without 889 third-party authentication. A skillful MitM could easily handle all 890 parts of the HIP base exchange, but HIP indirectly provides the 891 following protection from a MitM attack. If the responder's HI is 892 retrieved from a signed DNS zone or secured by some other means, the 893 initiator can use this to authenticate the signed HIP packets. 894 Likewise, if the initiator's HI is in a secure DNS zone, the 895 responder can retrieve it and validate the signed HIP packets. 896 However, since an initiator may choose to use an unpublished HI, it 897 knowingly risks a MitM attack. The responder may choose not to 898 accept a HIP exchange with an initiator using an unknown HI. 900 In HIP, the Security Association for IPsec is indexed by the SPI; the 901 source address is always ignored, and the destination address may be 902 ignored as well. Therefore, HIP-enabled IPsec Encapsulated Security 903 Payload (ESP) is IP address independent. This might seem to make it 904 easier for an attacker, but ESP with replay protection is already as 905 well protected as possible, and the removal of the IP address as a 906 check should not increase the exposure of IPsec ESP to DoS attacks. 908 Since not all hosts will ever support HIP, ICMPv4 'Destination 909 Unreachable, Protocol Unreachable' and ICMPv6 'Parameter Problem, 910 Unrecognized Next Header' messages are to be expected and present a 911 DoS attack. Against an initiator, the attack would look like the 912 responder does not support HIP, but shortly after receiving the ICMP 913 message, the initiator would receive a valid HIP packet. Thus, to 914 protect against this attack, an initiator should not react to an ICMP 915 message until a reasonable time has passed, allowing it to get the 916 real responder's HIP packet. A similar attack against the responder 917 is more involved. 919 Another MitM attack is simulating a responder's administrative 920 rejection of a HIP initiation. This is a simple ICMP 'Destination 921 Unreachable, Administratively Prohibited' message. A HIP packet is 922 not used because it would either have to have unique content, and 923 thus difficult to generate, resulting in yet another DoS attack, or 924 just as spoofable as the ICMP message. Like in the previous case, 925 the defense against this attack is for the initiator to wait a 926 reasonable time period to get a valid HIP packet. If one does not 927 come, then the initiator has to assume that the ICMP message is 928 valid. Since this is the only point in the HIP base exchange where 929 this ICMP message is appropriate, it can be ignored at any other 930 point in the exchange. 932 13.1. HITs used in ACLs 934 It is expected that HITs will be used in ACLs. Future firewalls can 935 use HITs to control egress and ingress to networks, with an assurance 936 level difficult to achieve today. As discussed above in Section 7, 937 once a HIP session has been established, the SPI value in an IPsec 938 packet may be used as an index, indicating the HITs. In practice, 939 firewalls can inspect HIP packets to learn of the bindings between 940 HITs, SPI values, and IP addresses. They can even explicitly control 941 IPsec usage, dynamically opening IPsec ESP only for specific SPI 942 values and IP addresses. The signatures in HIP packets allow a 943 capable firewall to ensure that the HIP exchange is indeed happening 944 between two known hosts. This may increase firewall security. 946 There has been considerable bad experience with distributed ACLs that 947 contain public key related material, for example, with SSH. If the 948 owner of a key needs to revoke it for any reason, the task of finding 949 all locations where the key is held in an ACL may be impossible. If 950 the reason for the revocation is due to private key theft, this could 951 be a serious issue. 953 A host can keep track of all of its partners that might use its HIT 954 in an ACL by logging all remote HITs. It should only be necessary to 955 log responder hosts. With this information, the host can notify the 956 various hosts about the change to the HIT. There has been no attempt 957 to develop a secure method to issue the HIT revocation notice. 959 HIP-aware NATs, however, are transparent to the HIP aware systems by 960 design. Thus, the host may find it difficult to notify any NAT that 961 is using a HIT in an ACL. Since most systems will know of the NATs 962 for their network, there should be a process by which they can notify 963 these NATs of the change of the HIT. This is mandatory for systems 964 that function as responders behind a NAT. In a similar vein, if a 965 host is notified of a change in a HIT of an initiator, it should 966 notify its NAT of the change. In this manner, NATs will get updated 967 with the HIT change. 969 13.2. Non-security considerations 971 The definition of the Host Identifier states that the HI need not be 972 a public key. It implies that the HI could be any value; for example 973 a FQDN. This document does not describe how to support such a non- 974 cryptographic HI. A non-cryptographic HI would still offer the 975 services of the HIT or LSI for NAT traversal. It would be possible 976 to carry HITs in HIP packets that had neither privacy nor 977 authentication. Since such a mode would offer so little additional 978 functionality for so much addition to the IP kernel, it has not been 979 defined. Given how little public key cryptography HIP requires, HIP 980 should only be implemented using public key Host Identities. 982 If it is desirable to use HIP in a low security situation where 983 public key computations are considered expensive, HIP can be used 984 with very short Diffie-Hellman and Host Identity keys. Such use 985 makes the participating hosts vulnerable to MitM and connection 986 hijacking attacks. However, it does not cause flooding dangers, 987 since the address check mechanism relies on the routing system and 988 not on cryptographic strength. 990 14. IANA considerations 992 This document has no actions for IANA. 994 15. Acknowledgments 996 For the people historically involved in the early stages of HIP, see 997 the Acknowledgements section in the Host Identity Protocol 998 specification. 1000 During the later stages of this document, when the editing baton was 1001 transfered to Pekka Nikander, the comments from the early 1002 implementors and others, including Jari Arkko, Tom Henderson, Petri 1003 Jokela, Miika Komu, Mika Kousa, Andrew McGregor, Jan Melen, Tim 1004 Shepard, Jukka Ylitalo, and Jorma Wall, were invaluable. Finally, 1005 Lars Eggert, Spencer Dawkins and Dave Crocker provided valuable input 1006 during the final stages of publication, most of which was 1007 incorporated but some of which the authors decided to ignore in order 1008 to get this document published in the first place. 1010 The authors want to express their special thanks to Tom Henderson, 1011 who took the burden of editing the document in response to IESG 1012 comments at the time when both of the authors were busy doing other 1013 things. Without his perseverance this document might have never made 1014 it into an RFC form. 1016 16. References 1018 16.1. Normative References 1020 [RFC5202] Jokela, P., Moskowitz, R., and P. Nikander, "Using the 1021 Encapsulating Security Payload (ESP) Transport Format with 1022 the Host Identity Protocol (HIP)", RFC 5202, April 2008. 1024 [RFC5204] Laganier, J. and L. Eggert, "Host Identity Protocol (HIP) 1025 Rendezvous Extension", RFC 5204, April 2008. 1027 [RFC5205] Nikander, P. and J. Laganier, "Host Identity Protocol 1028 (HIP) Domain Name System (DNS) Extensions", RFC 5205, 1029 April 2008. 1031 16.2. Informative references 1033 [RFC2136] Vixie, P., Thomson, S., Rekhter, Y., and J. Bound, 1034 "Dynamic Updates in the Domain Name System (DNS UPDATE)", 1035 RFC 2136, April 1997. 1037 [RFC2535] Eastlake, D., "Domain Name System Security Extensions", 1038 RFC 2535, March 1999. 1040 [RFC2766] Tsirtsis, G. and P. Srisuresh, "Network Address 1041 Translation - Protocol Translation (NAT-PT)", RFC 2766, 1042 February 2000. 1044 [RFC3022] Srisuresh, P. and K. Egevang, "Traditional IP Network 1045 Address Translator (Traditional NAT)", RFC 3022, 1046 January 2001. 1048 [RFC3102] Borella, M., Lo, J., Grabelsky, D., and G. Montenegro, 1049 "Realm Specific IP: Framework", RFC 3102, October 2001. 1051 [RFC4025] Richardson, M., "A Method for Storing IPsec Keying 1052 Material in DNS", RFC 4025, March 2005. 1054 [RFC4225] Nikander, P., Arkko, J., Aura, T., Montenegro, G., and E. 1055 Nordmark, "Mobile IP Version 6 Route Optimization Security 1056 Design Background", RFC 4225, December 2005. 1058 [RFC4306] Kaufman, C., "Internet Key Exchange (IKEv2) Protocol", 1059 RFC 4306, December 2005. 1061 [RFC4423] Moskowitz, R. and P. Nikander, "Host Identity Protocol 1062 (HIP) Architecture", RFC 4423, May 2006. 1064 [I-D.irtf-nsrg-report] 1065 Lear, E. and R. Droms, "What's In A Name:Thoughts from the 1066 NSRG", draft-irtf-nsrg-report-10 (work in progress), 1067 September 2003. 1069 [chiappa-endpoints] 1070 Chiappa, J., "Endpoints and Endpoint Names: A Proposed 1071 Enhancement to the Internet Architecture", 1072 URL http://www.chiappa.net/~jnc/tech/endpoints.txt, 1999. 1074 [Nik2001] Nikander, P., "Denial-of-Service, Address Ownership, and 1075 Early Authentication in the IPv6 World", in Proceesings 1076 of Security Protocols, 9th International Workshop, 1077 Cambridge, UK, April 25-27 2001, LNCS 2467, pp. 12-26, 1078 Springer, 2002. 1080 [Bel1998] Bellovin, S., "EIDs, IPsec, and HostNAT", in Proceedings 1081 of 41th IETF, Los Angeles, CA, 1082 URL http://www1.cs.columbia.edu/~smb/talks/hostnat.pdf, 1083 March 1998. 1085 Author's Address 1087 Robert Moskowitz 1088 ICSA labs, An Independent Division of Verizon Business 1089 1000 Bent Creek Blvd, Suite 200 1090 Mechanicsburg, PA 1091 USA 1093 Email: robert.moskowitz@icsalabs.com