idnits 2.17.1 draft-templin-intarea-vet-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.ii or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) 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 : ---------------------------------------------------------------------------- == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 1 instance of lines with multicast IPv4 addresses in the document. If these are generic example addresses, they should be changed to use the 233.252.0.x range defined in RFC 5771 Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (August 28, 2009) is 5347 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. 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: 'I-D.ietf-autoconf-manetarch' is defined on line 1697, but no explicit reference was found in the text == Unused Reference: 'RFC3056' is defined on line 1773, but no explicit reference was found in the text == Unused Reference: 'RFC3753' is defined on line 1779, but no explicit reference was found in the text == Outdated reference: A later version (-68) exists of draft-templin-intarea-seal-05 ** Downref: Normative reference to an Experimental draft: draft-templin-intarea-seal (ref. 'I-D.templin-intarea-seal') ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) ** Obsolete normative reference: RFC 3315 (Obsoleted by RFC 8415) ** Obsolete normative reference: RFC 3633 (Obsoleted by RFC 8415) ** Downref: Normative reference to an Informational RFC: RFC 5214 == Outdated reference: A later version (-15) exists of draft-cheshire-dnsext-multicastdns-07 == Outdated reference: A later version (-02) exists of draft-hain-ipv6-ulac-00 == Outdated reference: A later version (-05) exists of draft-ietf-csi-proxy-send-01 == Outdated reference: A later version (-13) exists of draft-ietf-dhc-subnet-alloc-09 == Outdated reference: A later version (-14) exists of draft-ietf-manet-smf-09 == Outdated reference: A later version (-04) exists of draft-ietf-v6ops-tunnel-security-concerns-01 == Outdated reference: A later version (-05) exists of draft-russert-rangers-00 == Outdated reference: A later version (-09) exists of draft-templin-ranger-07 -- Obsolete informational reference (is this intentional?): RFC 4941 (Obsoleted by RFC 8981) Summary: 6 errors (**), 0 flaws (~~), 15 warnings (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group F. Templin, Ed. 3 Internet-Draft Boeing Research & Technology 4 Intended status: Standards Track August 28, 2009 5 Expires: March 1, 2010 7 Virtual Enterprise Traversal (VET) 8 draft-templin-intarea-vet-03.txt 10 Status of this Memo 12 This Internet-Draft is submitted to IETF in full conformance with the 13 provisions of BCP 78 and BCP 79. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that 17 other groups may also distribute working documents as Internet- 18 Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet-Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt. 28 The list of Internet-Draft Shadow Directories can be accessed at 29 http://www.ietf.org/shadow.html. 31 This Internet-Draft will expire on March 1, 2010. 33 Copyright Notice 35 Copyright (c) 2009 IETF Trust and the persons identified as the 36 document authors. All rights reserved. 38 This document is subject to BCP 78 and the IETF Trust's Legal 39 Provisions Relating to IETF Documents in effect on the date of 40 publication of this document (http://trustee.ietf.org/license-info). 41 Please review these documents carefully, as they describe your rights 42 and restrictions with respect to this document. 44 Abstract 46 Enterprise networks connect routers over various link types, and may 47 also connect to provider networks and/or the global Internet. 48 Enterprise network nodes require a means to automatically provision 49 IP addresses/prefixes and support internetworking operation in a wide 50 variety of use cases including Small Office, Home Office (SOHO) 51 networks, Mobile Ad hoc Networks (MANETs), ISP networks, multi- 52 organizational corporate networks and the interdomain core of the 53 global Internet itself. This document specifies a Virtual Enterprise 54 Traversal (VET) abstraction for autoconfiguration and operation of 55 nodes in enterprise networks. VET can also be considered as version 56 2 of the Intra-Site Automatic Tunnel Addressing Protocol (i.e., 57 "ISATAPv2"). 59 Table of Contents 61 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 62 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6 63 3. Enterprise Characteristics . . . . . . . . . . . . . . . . . . 10 64 4. Autoconfiguration . . . . . . . . . . . . . . . . . . . . . . 12 65 4.1. Enterprise Router (ER) Autoconfiguration . . . . . . . . . 12 66 4.2. Enterprise Border Router (EBR) Autoconfiguration . . . . . 14 67 4.2.1. VET Interface Autoconfiguration . . . . . . . . . . . 14 68 4.2.2. Provider-Aggregated (PA) EID Prefix 69 Autoconfiguration . . . . . . . . . . . . . . . . . . 16 70 4.2.3. Provider-Independent (PI) EID Prefix 71 Autoconfiguration . . . . . . . . . . . . . . . . . . 17 72 4.3. Enterprise Border Gateway (EBG) Autoconfiguration . . . . 17 73 4.4. VET Host Autoconfiguration . . . . . . . . . . . . . . . . 18 74 5. Internetworking Operation . . . . . . . . . . . . . . . . . . 18 75 5.1. Routing Protocol Participation . . . . . . . . . . . . . . 18 76 5.2. RLOC-Based Communications . . . . . . . . . . . . . . . . 19 77 5.3. EID-Based Communications . . . . . . . . . . . . . . . . . 19 78 5.4. IPv6 Router and Prefix Discovery . . . . . . . . . . . . . 19 79 5.4.1. Router and Prefix Discovery . . . . . . . . . . . . . 19 80 5.4.2. Address Autoconfiguration on VET Interfaces . . . . . 20 81 5.4.3. PA Prefix Registration . . . . . . . . . . . . . . . . 22 82 5.4.4. PI Prefix Registration . . . . . . . . . . . . . . . . 22 83 5.4.5. Next-Hop Discovery . . . . . . . . . . . . . . . . . . 24 84 5.5. IPv4 Router and Prefix Discovery . . . . . . . . . . . . . 26 85 5.6. Forwarding Packets on VET Interfaces . . . . . . . . . . . 26 86 5.7. VET and SEAL Encapsulation . . . . . . . . . . . . . . . . 27 87 5.8. Generating Errors . . . . . . . . . . . . . . . . . . . . 28 88 5.9. Processing Errors . . . . . . . . . . . . . . . . . . . . 29 89 5.10. Mobility and Multihoming Considerations . . . . . . . . . 29 90 5.11. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 30 91 5.12. Service Discovery . . . . . . . . . . . . . . . . . . . . 32 92 5.13. Enterprise Partitioning . . . . . . . . . . . . . . . . . 32 93 5.14. EBG Prefix State Recovery . . . . . . . . . . . . . . . . 32 94 5.15. Support for Legacy ISATAP Services . . . . . . . . . . . . 33 95 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 33 96 7. Security Considerations . . . . . . . . . . . . . . . . . . . 33 97 8. Related Work . . . . . . . . . . . . . . . . . . . . . . . . . 34 98 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 35 99 10. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 35 100 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 35 101 11.1. Normative References . . . . . . . . . . . . . . . . . . . 35 102 11.2. Informative References . . . . . . . . . . . . . . . . . . 37 103 Appendix A. Duplicate Address Detection (DAD) Considerations . . 40 104 Appendix B. Link-Layer Multiplexing and Traffic Engineering . . . 41 105 Appendix C. Change Log . . . . . . . . . . . . . . . . . . . . . 43 106 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 44 108 1. Introduction 110 Enterprise networks [RFC4852] connect routers over various link types 111 (see [RFC4861], Section 2.2). The term "enterprise network" in this 112 context extends to a wide variety of use cases and deployment 113 scenarios. For example, an "enterprise" can be as small as a SOHO 114 network, as complex as a multi-organizational corporation, or as 115 large as the global Internet itself. ISP networks are another 116 example use case that fits well with the VET enterprise network 117 model. Mobile Ad hoc Networks (MANETs) [RFC2501] can also be 118 considered as a challenging example of an enterprise network, in that 119 their topologies may change dynamically over time and that they may 120 employ little/no active management by a centralized network 121 administrative authority. These specialized characteristics for 122 MANETs require careful consideration, but the same principles apply 123 equally to other enterprise network scenarios. 125 This document specifies a Virtual Enterprise Traversal (VET) 126 abstraction for autoconfiguration and internetworking operation, 127 where addresses of different scopes may be assigned on various types 128 of interfaces with diverse properties. Both IPv4 [RFC0791] and IPv6 129 [RFC2460] are discussed within this context. The use of standard 130 DHCP [RFC2131] [RFC3315] and neighbor discovery [RFC0826] [RFC1256] 131 [RFC4861] mechanisms is assumed unless otherwise specified. 133 Provider-Edge Interfaces 134 x x x 135 | | | 136 +--------------------+---+--------+----------+ E 137 | | | | | n 138 | I | | .... | | t 139 | n +---+---+--------+---+ | e 140 | t | +--------+ /| | r 141 | e I x----+ | Host | I /*+------+--< p I 142 | r n | |Function| n|**| | r n 143 | n t | +--------+ t|**| | i t 144 | a e x----+ V e|**+------+--< s e 145 | l r . | E r|**| . | e r 146 | f . | T f|**| . | f 147 | V a . | +--------+ a|**| . | I a 148 | i c . | | Router | c|**| . | n c 149 | r e x----+ |Function| e \*+------+--< t e 150 | t s | +--------+ \| | e s 151 | u +---+---+--------+---+ | r 152 | a | | .... | | i 153 | l | | | | o 154 +--------------------+---+--------+----------+ r 155 | | | 156 x x x 157 Enterprise-Edge Interfaces 159 Figure 1: Enterprise Router (ER) Architecture 161 Figure 1 above depicts the architectural model for an Enterprise 162 Router (ER). As shown in the figure, an ER may have a variety of 163 interface types including enterprise-edge, enterprise-interior, 164 provider-edge, internal-virtual, as well as VET interfaces used for 165 IP in IP encapsulation. The different types of interfaces are 166 defined, and the autoconfiguration mechanisms used for each type are 167 specified. This architecture applies equally for MANET routers, in 168 which enterprise-interior interfaces correspond to the wireless 169 multihop radio interfaces typically associated with MANETs. Out of 170 scope for this document is the autoconfiguration of provider 171 interfaces, which must be coordinated in a manner specific to the 172 service provider's network. 174 Enterprise networks must have a means for supporting both Provider- 175 Independent (PI) and Provider-Aggregated (PA) IP prefixes. This is 176 especially true for enterprise scenarios that involve mobility and 177 multihoming. Also in scope are ingress filtering for multihomed 178 sites, adaptation based on authenticated ICMP feedback from on-path 179 routers, effective tunnel path MTU mitigations, and routing scaling 180 suppression as required in many enterprise network scenarios. 182 Recognizing that one size does not fit all, the VET specification 183 provides adaptable mechanisms that address these issues, and more, in 184 a wide variety of enterprise network use cases. 186 VET represents a functional superset of 6over4 [RFC2529] and the 187 Intra-Site Automatic Tunnel Addressing Protocol (ISATAP) [RFC5214], 188 where VET can be considered as version 2 of the ISATAP protocol 189 (i.e., "ISATAPv2"). VET also works in conjunction with the 190 Subnetwork Encapsulation and Adaptation Layer (SEAL) 191 [I-D.templin-intarea-seal] and supports additional encapsulations 192 such as IPsec [RFC4301]. Together, these technologies serve as 193 functional building blocks for a new Internetworking architecture 194 known as Routing and Addressing in Next Generation EnteRprises 195 [I-D.templin-ranger] [I-D.russert-rangers]. 197 The VET principles can be either directly or indirectly traced to the 198 deliberations of the ROAD group in January 1992, and also to still 199 earlier works including NIMROD [RFC1753] and the Catenet model for 200 internetworking [CATENET] [IEN48] [RFC2775]. [RFC1955] captures the 201 high-level architectural aspects of the ROAD group deliberations in a 202 "New Scheme for Internet Routing and Addressing (ENCAPS) for IPNG". 204 VET is related to the present-day activities of the IETF INTAREA, 205 AUTOCONF, DHC, IPv6, MANET, and V6OPS working groups, as well as the 206 IRTF RRG working group. 208 2. Terminology 210 The mechanisms within this document build upon the fundamental 211 principles of IP in IP encapsulation. The terms "inner" and "outer" 212 are used to, respectively, refer to the innermost IP {address, 213 protocol, header, packet, etc.} *before* encapsulation, and the 214 outermost IP {address, protocol, header, packet, etc.} *after* 215 encapsulation. VET also uses the Subnetwork Encapsulation and 216 Adaptation Layer (SEAL) [I-D.templin-intarea-seal] as a "mid-layer" 217 encapsulation between the inner and outer IP headers, and also allows 218 for inclusion of other mid-layer encapsulations including IPSec 219 [RFC4301]. 221 The terminology in the normative references apply; the following 222 terms are defined within the scope of this document: 224 subnetwork 225 the same as defined in [RFC3819]. 227 enterprise 228 the same as defined in [RFC4852]. An enterprise is also 229 understood to refer to a cooperative networked collective with a 230 commonality of business, social, political, etc. interests. 231 Minimally, the only commonality of interest in some enterprise 232 network scenarios may be the cooperative provisioning of 233 connectivity itself. 235 site 236 a logical and/or physical grouping of interfaces that connect a 237 topological area less than or equal to an enterprise in scope. A 238 site within an enterprise can, in some sense, be considered as an 239 enterprise unto itself. 241 Mobile Ad hoc Network (MANET) 242 a connected topology of mobile or fixed routers that maintain a 243 routing structure among themselves over dynamic links, where a 244 wide variety of MANETs share common properties with enterprise 245 networks. The characteristics of MANETs are defined in [RFC2501], 246 Section 3. 248 enterprise/site/MANET 249 throughout the remainder of this document, the term "enterprise" 250 is used to collectively refer to any of {enterprise, site, MANET}, 251 i.e., the VET mechanisms and operational principles can be applied 252 to enterprises, sites, and MANETs of any size or shape. 254 Enterprise Router (ER) 255 As depicted in Figure 1, an Enterprise Router (ER) is a fixed or 256 mobile router that comprises a router function, a host function, 257 one or more enterprise-interior interfaces, and zero or more 258 internal virtual, enterprise-edge, provider-edge, and VET 259 interfaces. At a minimum, an ER forwards outer IP packets over 260 one or more sets of enterprise-interior interfaces, where each set 261 connects to a distinct enterprise. 263 Enterprise Border Router (EBR) 264 an ER that connects edge networks to the enterprise and/or 265 connects multiple enterprises together. An EBR is a tunnel 266 endpoint router, and it configures a separate VET interface over 267 each set of enterprise-interior interfaces that connect the EBR to 268 each distinct enterprise. In particular, an EBR may configure 269 multiple VET interfaces - one for each distinct enterprise. All 270 EBRs are also ERs. 272 Enterprise Border Gateway (EBG) 273 an EBR that connects VET interfaces configured over child 274 enterprises to a provider network - either directly via a 275 provider-edge interface or indirectly via another VET interface 276 configured over a parent enterprise. EBRs may act as EBGs on some 277 VET interfaces and as ordinary EBRs on other VET interfaces. All 278 EBGs are also EBRs. 280 enterprise-interior interface 281 an ER's attachment to a link within an enterprise. Packets sent 282 over enterprise-interior interfaces may be forwarded over multiple 283 additional enterprise-interior interfaces within the enterprise 284 before they are forwarded via an enterprise-edge interface, 285 provider-edge interface, or a VET interface configured over a 286 different enterprise. Enterprise-interior interfaces connect 287 laterally within the IP network hierarchy. 289 enterprise-edge interface 290 an EBR's attachment to a link (e.g., an Ethernet, a wireless 291 personal area network, etc.) on an arbitrarily complex edge 292 network that the EBR connects to an enterprise and/or provider 293 network. Enterprise-edge interfaces connect to lower levels 294 within the IP network hierarchy. 296 provider-edge interface 297 an EBR's attachment to the Internet or to a provider network 298 outside of the enterprise via which the Internet can be reached. 299 Provider-edge interfaces connect to higher levels within the IP 300 network hierarchy. 302 internal-virtual interface 303 an interface that is internal to an EBR and does not in itself 304 directly attach to a tangible physical link, e.g., an Ethernet 305 cable. Examples include a loopback interface, a virtual private 306 network interface, or some form of tunnel interface. 308 Virtual Enterprise Traversal (VET) 309 an abstraction that uses IP in IP encapsulation to create an 310 overlay that spans an enterprise in a single (inner) IP hop. VET 311 can be considered as version 2 of the ISATAP protocol (i.e., 312 "ISATAPv2"). 314 VET interface 315 an EBR's tunnel virtual interface used for Virtual Enterprise 316 Traversal. The EBR configures a VET interface over a set of 317 underlying interfaces belonging to the same enterprise. When 318 there are multiple distinct enterprises (each with their own 319 distinct set of underlying interfaces), the EBR configures a 320 separate VET interface over each set of underlying interfaces, 321 i.e., the EBR configures multiple VET interfaces. VET interfaces 322 natively use the Subnetwork Encapsulation and Adaptation Layer 323 (SEAL). 325 The VET interface encapsulates each inner IP packet in any mid- 326 layer headers followed by the SEAL header followed by an outer IP 327 header, then forwards the packet on an underlying interface such 328 that the Time to Live (TTL) - Hop Limit in the inner header is not 329 decremented as the packet traverses the enterprise. The VET 330 interface therefore presents an automatic tunneling abstraction 331 that represents the enterprise as a single IP hop. 333 VET interfaces in non-multicast environments are Non-Broadcast, 334 Multiple Access (NBMA); VET interfaces in multicast environments 335 are multicast capable. 337 VET address 338 an IPv6 address format associated with a VET interface that use 339 IPv6 and IPv4 as the inner and outer IP protocols, respectively. 340 VET addresses are formed exactly as specified for ISATAP addresses 341 in Sections 6.1 and 6.2 of [RFC5214]. 343 VET host 344 any node (host or router) that configures a VET interface for host 345 operation only. Note that a single node may configure some of its 346 VET interfaces as host interfaces and others as router interfaces. 348 VET node 349 any node that configures and uses a VET interface. 351 Provider-Independent (PI) prefix 352 an IPv6 or IPv4 prefix (e.g., 2001:DB8::/48, 192.0.2/24, etc.) 353 that is either self-generated by an ER or delegated to an 354 enterprise by a registry. 356 Provider Aggregated (PA) prefix 357 an IPv6 or IPv4 prefix that is delegated to an enterprise by a 358 provider network. 360 Routing Locator (RLOC) 361 a non-link-local IPv4 or IPv6 address taken from a PI/PA prefix 362 that can appear in enterprise-interior and/or interdomain routing 363 tables. Global-scope RLOC prefixes are delegated to specific 364 enterprises and routable within both the enterprise-interior and 365 interdomain routing regions. Enterprise-local-scope RLOC prefixes 366 (e.g., IPv6 Unique Local Addresses [RFC4193], IPv4 privacy 367 addresses [RFC1918], etc.) are self-generated by individual 368 enterprises and routable only within the enterprise-interior 369 routing region. 371 ERs use RLOCs for operating the enterprise-interior routing 372 protocol and for next-hop determination in forwarding packets 373 addressed to other RLOCs. End systems use RLOCs as addresses for 374 communications between endpoints within the same enterprise. VET 375 interfaces treat RLOCs as *outer* IP addresses during IP in IP 376 encapsulation. 378 Endpoint Interface iDentifier (EID) 379 an IPv4 or IPv6 address taken from a PI/PA prefix that is routable 380 within an enterprise-edge or VET overlay network scope, and may 381 also appear in enterprise-interior and/or interdomain mapping 382 tables. EID prefixes are typically separate and distinct from any 383 RLOC prefix space. 385 Edge network routers use EIDs for operating the enterprise-edge or 386 VET overlay network routing protocol and for next-hop 387 determination in forwarding packets addressed to other EIDs. End 388 systems use EIDs as addresses for communications between endpoints 389 either within the same enterprise or within different enterprises. 390 VET interfaces treat EIDs as *inner* IP addresses during IP in IP 391 encapsulation. 393 The following additional acronyms are used throughout the document: 395 CGA - Cryptographically Generated Address 396 DHCP(v4, v6) - Dynamic Host Configuration Protocol 397 FIB - Forwarding Information Base 398 ISATAP - Intra-Site Automatic Tunnel Addressing Protocol 399 NBMA - Non-Broadcast, Multiple Access 400 ND - Neighbor Discovery 401 PIO - Prefix Information Option 402 PRL - Potential Router List 403 PRLNAME - Identifying name for the PRL (default is "isatapv2") 404 RIO - Route Information Option 405 RS/RA - IPv6 ND Router Solicitation/Advertisement 406 SEAL - Subnetwork Encapsulation and Adaptation Layer 407 SLAAC - IPv6 StateLess Address AutoConfiguation 409 3. Enterprise Characteristics 411 Enterprises consist of links that are connected by Enterprise Routers 412 (ERs) as depicted in Figure 1. ERs typically participate in a 413 routing protocol over enterprise-interior interfaces to discover 414 routes that may include multiple Layer 2 or Layer 3 forwarding hops. 416 Enterprise Border Routers (EBRs) are ERs that connect edge networks 417 to the enterprise and/or join multiple enterprises together. 418 Enterprise Border Gateways (EBGs) are EBRs that either directly or 419 indirectly connect enterprises to provider networks. 421 Conceptually, an ER embodies both a host function and router 422 function. The host function supports Endpoint Interface iDentifier 423 (EID)-based and/or Routing LOCator (RLOC)-based communications 424 according to the weak end-system model [RFC1122]. The router 425 function engages in the enterprise-interior routing protocol, 426 connects any of the ER's edge networks to the enterprise, and may 427 also connect the enterprise to provider networks (see Figure 1). 429 An enterprise may be as simple as a small collection of ERs and their 430 attached edge networks; an enterprise may also contain other 431 enterprises and/or be a subnetwork of a larger enterprise. An 432 enterprise may further encompass a set of branch offices and/or 433 nomadic hosts connected to a home office over one or several service 434 providers, e.g., through Virtual Private Network (VPN) tunnels. 435 Finally, an enterprise may contain many internal partitions that are 436 logical groupings of nodes for the purpose of load balancing, 437 organizational separation, etc. In that case, each internal 438 partition resembles an individual segment of a bridged LAN. 440 Enterprises that comprise link types with sufficiently similar 441 properties (e.g., Layer 2 (L2) address formats, maximum transmission 442 units (MTUs), etc.) can configure a sub-IP layer routing service such 443 that IP sees the enterprise as an ordinary shared link the same as 444 for a (bridged) campus LAN. In that case, a single IP hop is 445 sufficient to traverse the enterprise without IP layer encapsulation. 446 Enterprises that comprise link types with diverse properties and/or 447 configure multiple IP subnets must also provide a routing service 448 that operates as an IP layer mechanism. In that case, multiple IP 449 hops may be necessary to traverse the enterprise such that care must 450 be taken to avoid multi-link subnet issues [RFC4903]. 452 In addition to other interface types, VET nodes configure VET 453 interfaces that view all other VET nodes in an enterprise as single- 454 hop neighbors attached to a virtual link. VET nodes configure a 455 separate VET interface for each distinct enterprise to which they 456 connect, and discover other EBRs on each VET interface that can be 457 used for forwarding packets to off-enterprise destinations. 459 For each distinct enterprise, an enterprise trust basis must be 460 established and consistently applied. For example, in enterprises in 461 which EBRs establish symmetric security associations, mechanisms such 462 as IPsec [RFC4301] can be used to assure authentication and 463 confidentiality. In other enterprise network scenarios, asymmetric 464 securing mechanisms such as SEcure Neighbor Discovery (SEND) 465 [RFC3971] may be necessary to authenticate exchanges based on trust 466 anchors. Still other enterprises may have sufficient infrastructure 467 trust basis (e.g., through proper deployment of filtering gateways at 468 enterprise borders) and may not require nodes to implement such 469 additional mechanisms. 471 Finally, in enterprises with a centralized management structure 472 (e.g., a corporate campus network), an enterprise mapping service and 473 a synchronized set of EBGs can provide sufficient infrastructure 474 support for virtual enterprise traversal. In that case, the EBGs can 475 provide a "default mapper" [I-D.jen-apt] service used for short-term 476 packet forwarding until EBR neighbor relationships can be 477 established. In enterprises with a distributed management structure 478 (e.g., MANETs), peer-to-peer coordination between the EBRs themselves 479 may be required. Recognizing that various use cases will entail a 480 continuum between a fully distributed and fully centralized approach, 481 the following sections present the mechanisms of Virtual Enterprise 482 Traversal as they apply to a wide variety of scenarios. 484 4. Autoconfiguration 486 ERs, EBRs, EBGs, and VET hosts configure themselves for operation as 487 specified in the following subsections. 489 4.1. Enterprise Router (ER) Autoconfiguration 491 ERs configure enterprise-interior interfaces and engage in any 492 routing protocols over those interfaces. 494 When an ER joins an enterprise, it first configures an IPv6 link- 495 local address on each enterprise-interior interface and configures an 496 IPv4 link-local address on each enterprise-interior interface that 497 requires an IPv4 link-local capability. IPv6 link-local address 498 generation mechanisms include Cryptographically Generated Addresses 499 (CGAs) [RFC3972], IPv6 Privacy Addresses [RFC4941], StateLess Address 500 AutoConfiguration (SLAAC) using EUI-64 interface identifiers 501 [RFC4291] [RFC4862], etc. The mechanisms specified in [RFC3927] 502 provide an IPv4 link-local address generation capability. 504 Next, the ER configures one or more RLOCs and engages in any routing 505 protocols on its enterprise-interior interfaces. The ER can 506 configure RLOCs via explicit management, DHCP autoconfiguration, 507 pseudo-random self-generation from a suitably large address pool, or 508 through an alternate autoconfiguration mechanism. The ER may 509 optionally configure and assign a separate RLOC for each underlying 510 interface, or it may configure only a single RLOC and assign it to a 511 VET interface configured over the underlying interfaces (see Section 512 4.2.1). In the latter case, the ER can use the VET interface for 513 link layer multiplexing and traffic engineering purposes as specified 514 in Appendix B. 516 Alternatively (or in addition), the ER can request RLOC prefix 517 delegations via an automated prefix delegation exchange over an 518 enterprise-interior interface and can assign the prefix(es) on 519 enterprise-edge interfaces. Note that in some cases, the same 520 enterprise-edge interfaces may assign both RLOC and EID addresses if 521 there is a means for source address selection. In other cases (e.g., 522 for separation of security domains), RLOCs and EIDs must be assigned 523 on separate sets of enterprise-edge interfaces. 525 Self-generation of RLOCs for IPv6 can be from a large public or 526 local-use IPv6 address range (e.g., IPv6 Unique Local Addresses 527 [RFC4193]). Self-generation of RLOCs for IPv4 can be from a large 528 public or private IPv4 private address range (e.g., [RFC1918]). When 529 self-generation is used alone, the ER must continuously monitor the 530 RLOCs for uniqueness, e.g., by monitoring the routing protocol. 532 DHCP generation of RLOCs may require support from relays within the 533 enterprise. For DHCPv6, relays that do not already know the RLOC of 534 a server within the enterprise forward requests to the 535 'All_DHCP_Servers' site-scoped IPv6 multicast group [RFC3315]. For 536 DHCPv4, relays that do not already know the RLOC of a server within 537 the enterprise forward requests to the site-scoped IPv4 multicast 538 group address 'All_DHCPv4_Servers', which should be set to 539 239.255.2.1 unless an alternate multicast group for the site is 540 known. DHCPv4 servers that delegate RLOCs should therefore join the 541 'All_DHCPv4_Servers' multicast group and service any DHCPv4 messages 542 received for that group. 544 A combined approach using both DHCP and self-generation is also 545 possible when the ER configures both a DHCP client and relay that are 546 connected, e.g., via a pair of back-to-back connected Ethernet 547 interfaces, a tun/tap interface, a loopback interface, inter-process 548 communication, etc. The ER first self-generates a temporary RLOC 549 used only for the purpose of procuring an actual RLOC taken from a 550 disjoint addressing range. The ER then engages in the routing 551 protocol and performs a DHCP client/relay exchange using the 552 temporary RLOC as the address of the relay. When the DHCP server 553 delegates an actual RLOC address/prefix, the ER abandons the 554 temporary RLOC and re-engages in the routing protocol using an RLOC 555 taken from the delegation. 557 In some enterprise use cases (e.g., MANETs), assignment of RLOCs on 558 enterprise-interior interfaces as singleton addresses (i.e., as 559 addresses with /32 prefix lengths for IPv4, or as addresses with /128 560 prefix lengths for IPv6) may be necessary to avoid multi-link subnet 561 issues. In other use cases, assignment of an RLOC on a VET interface 562 as specified in Appendix B can provide link layer multiplexing and 563 traffic engineering over multiple underlying interfaces using only a 564 single IP address. 566 4.2. Enterprise Border Router (EBR) Autoconfiguration 568 EBRs are ERs that configure VET interfaces over distinct sets of 569 underlying interfaces belonging to the same enterprise; an EBR can 570 connect to multiple enterprises, in which case it would configure 571 multiple VET interfaces. In addition to the ER autoconfiguration 572 procedures specified in Section 4.1, EBRs perform the following 573 autoconfiguration operations. 575 4.2.1. VET Interface Autoconfiguration 577 VET interface autoconfiguration entails: 1) interface initialization, 578 2) EBG discovery and enterprise identification, and 3) EID 579 configuration. These functions are specified in the following 580 sections. 582 4.2.1.1. Interface Initialization 584 EBRs configure a VET interface over a set of underlying interfaces 585 belonging to the same enterprise, where the VET interface presents a 586 virtual-link abstraction in which all EBRs in the enterprise appear 587 as single-hop neighbors through the use of IP in IP encapsulation. 588 After the EBR configures a VET interface, it initializes the 589 interface and assigns an IPv6 link-local address and an IPv4 link- 590 local address if necessary. The EBR also associates an RLOC obtained 591 as specified in Section 4.1 with the VET interface to serve as the 592 source address for outer IP packets. 594 When IPv6 and IPv4 are used as the inner/outer protocols 595 (respectively), the EBR autoconfigures an IPv6 link-local VET address 596 on the VET interface to support packet forwarding and operation of 597 the IPv6 neighbor discovery protocol. The link-local VET address is 598 formed exactly as specified in Sections 6.1 and 6.2 of [RFC5214]. 599 The link-local address need not be checked for uniqueness since the 600 IPv4 RLOC embedded in the address itself is managed for uniqueness 601 (see Section 4.1). 603 Link-local address configuration for other inner/outer IP protocol 604 combinations is through administrative configuration or through an 605 unspecified alternate method. Link-local address configuration for 606 other inner/outer IP protocol combinations may not be necessary if an 607 EID can be configured through other means (see Section 4.2.1.3). 609 After the EBR initializes a VET interface, it can communicate with 610 other VET nodes as single-hop neighbors on the VET interface from the 611 viewpoint of the inner IP protocol. The EBR can also configure the 612 VET interface for link-layer multiplexing and traffic engineering 613 purposes as specified in Appendix B. 615 4.2.1.2. Enterprise Border Gateway Discovery and Enterprise 616 Identification 618 The EBR next discovers a list of EBGs for each of its VET interfaces. 619 The list can be discovered through information conveyed in the 620 routing protocol, through the Potential Router List (PRL) discovery 621 mechanisms outlined in Section 8.3.2 of [RFC5214], through DHCP 622 options, etc. In multicast-capable enterprises, EBRs can also listen 623 for advertisements on the 'rasadv' [RASADV] multicast group address. 625 In particular, whether or not routing information is available, the 626 EBR can discover the list of EBGs by resolving an identifying name 627 for the PRL ('PRLNAME') formed as 'hostname.domainname', where 628 'hostname' is an enterprise-specific name string and 'domainname' is 629 an enterprise-specific DNS suffix. The EBR discovers 'PRLNAME' 630 through manual configuration, the DHCP Domain Name option [RFC2132], 631 'rasadv' protocol advertisements, link-layer information (e.g., an 632 IEEE 802.11 Service Set Identifier (SSID)), or through some other 633 means specific to the enterprise. 635 In the absence of other information, the EBR sets the 'hostname' 636 component of 'PRLNAME' to "isatapv2" and sets the 'domainname' 637 component to the enterprise-specific DNS suffix "example.com" (e.g., 638 as "isatapv2.example.com"). Note that this naming convention is 639 intentionally distinct from the convention specified in [RFC5214], 640 and is used by the EBR to distinguish between ISATAP and VET virtual 641 interfaces. 643 The global Internet interdomain routing core represents a specific 644 example of an enterprise network scenario, albeit on an enormous 645 scale. The 'PRLNAME' assigned to the global Internet interdomain 646 routing core for the purpose of VET is "isatapv2.net". 648 After discovering 'PRLNAME', the EBR can discover the list of EBGs by 649 resolving 'PRLNAME' to a list of RLOC addresses through a name 650 service lookup. For centrally managed enterprises, the EBR resolves 651 'PRLNAME' using an enterprise-local name service (e.g., the 652 enterprise-local DNS). For enterprises with a distributed management 653 structure, the EBR resolves 'PRLNAME' using Link-Local Multicast Name 654 Resolution (LLMNR) [RFC4795] over the VET interface. In that case, 655 all EBGs in the PRL respond to the LLMNR query, and the EBR accepts 656 the union of all responses. 658 Each distinct enterprise must have a unique identity that EBRs can 659 use to uniquely discern their enterprise affiliations. 'PRLNAME' as 660 well as the RLOCs of EBGs and the IP prefixes they aggregate serve as 661 an identifier for the enterprise. 663 4.2.1.3. EID Configuration 665 After EBG discovery, the EBR configures EIDs on its VET interfaces. 666 When IPv6 and IPv4 are used as the inner/outer protocols 667 (respectively), the EBR autoconfigures EIDs as specified in 668 Section 5.4. In particular, the EBR acts as a host on its VET 669 interfaces for router and prefix discovery purposes but acts as a 670 router on its VET interfaces for routing protocol operation and 671 packet forwarding purposes. 673 EID configuration for other inner/outer IP protocol combinations is 674 through administrative configuration or through an unspecified 675 alternate method; in some cases, such EID configuration can be 676 performed independently of EBG discovery. 678 4.2.2. Provider-Aggregated (PA) EID Prefix Autoconfiguration 680 EBRs can acquire Provider-Aggregated (PA) EID prefixes through 681 autoconfiguration exchanges with EBGs over VET interfaces, where each 682 EBG may be configured as either a DHCP relay or DHCP server. 684 For IPv4 EIDs, the EBR acquires prefixes via an automated IPv4 prefix 685 delegation exchange, explicit management, etc. 687 For IPv6 EIDs, the EBR acquires prefixes via DHCPv6 Prefix Delegation 688 exchanges. In particular, the EBR (acting as a requesting router) 689 can use DHCPv6 prefix delegation [RFC3633] over the VET interface to 690 obtain IPv6 EID prefixes from the server (acting as a delegating 691 router). 693 The EBR obtains prefixes using either a 2-message or 4-message DHCPv6 694 exchange [RFC3315]. For example, to perform the 2-message exchange, 695 the EBR's DHCPv6 client forwards a Solicit message with an IA_PD 696 option to its DHCPv6 relay, i.e., the EBR acts as a combined client/ 697 relay (see Section 4.1). The relay then forwards the message over 698 the VET interface to an EBG, which either services the request or 699 relays it further. The forwarded Solicit message will elicit a reply 700 from the server containing PA IPv6 prefix delegations. 702 The EBR can propose a specific prefix to the DHCPv6 server per 703 Section 7 of [RFC3633], e.g., if a prefix delegation hint is 704 available. The server will check the proposed prefix for consistency 705 and uniqueness, then return it in the reply to the EBR if it was able 706 to perform the delegation. 708 After the EBR receives PA prefix delegations, it can provision the 709 prefixes on enterprise-edge interfaces as well as on other VET 710 interfaces for which it is configured as an EBG. It can also 711 provision the prefixes on enterprise-interior interfaces as long as 712 other nodes on those interfaces unambiguously associate the prefixes 713 with the EBR. 715 4.2.3. Provider-Independent (PI) EID Prefix Autoconfiguration 717 Independent of any PA prefixes, EBRs can acquire and use Provider- 718 Independent (PI) EID prefixes that are self-configured (e.g., using 719 [RFC4193], etc.) and/or delegated by a registration authority (e.g., 720 through a regional Internet registry, through a centrally-assigned 721 unique local address delegation authority [I-D.hain-ipv6-ulac], 722 etc.). When an EBR acquires a PI prefix, it must also obtain 723 credentials that it can use to prove prefix ownership when it 724 registers the prefixes with EBGs within an enterprise (see 725 Section 5.4 and Section 5.5). 727 After the EBR receives PI prefix delegations, it can provision the 728 prefixes on enterprise-edge interfaces as well as on other VET 729 interfaces for which it is configured as an EBG. It can also 730 provision the prefixes on enterprise-interior interfaces as long as 731 other nodes on those interfaces can unambiguously associate the 732 prefixes with the EBR. 734 The minimum-sized IPv6 PI prefix that an EBR may acquire is a /56. 736 The minimum-sized IPv4 PI prefix that an EBR may acquire is a /24. 738 4.3. Enterprise Border Gateway (EBG) Autoconfiguration 740 EBGs are EBRs that connect child enterprises to provider networks via 741 provider-edge interfaces and/or via VET interfaces configured over 742 parent enterprises. EBGs autoconfigure their provider-edge 743 interfaces in a manner that is specific to the provider connections, 744 and they autoconfigure their VET interfaces that were configured over 745 parent enterprises using the EBR autoconfiguration procedures 746 specified in Section 4.2. 748 For each of its VET interfaces configured over a child enterprise, 749 the EBG initializes the interface the same as for an ordinary EBR 750 (see Section 4.2.1). It must then arrange to add one or more of its 751 RLOCs associated with the child enterprise to the PRL, and it must 752 maintain these resource records in accordance with [RFC5214], Section 753 9. In particular, for each VET interface configured over a child 754 enterprise, the EBG adds the RLOCs to name-service resource records 755 for 'PRLNAME' ("isatapv2.example.com" by default). 757 EBGs respond to LLMNR queries for 'PRLNAME' on VET interfaces 758 configured over child enterprises with a distributed management 759 structure. 761 EBGs configure a DHCP relay/server on VET interfaces configured over 762 child enterprises that require DHCP services. 764 To avoid looping, EBGs must not configure a default route on a VET 765 interface configured over a child interface. 767 4.4. VET Host Autoconfiguration 769 Nodes that cannot be attached via an EBR's enterprise-edge interface 770 (e.g., nomadic laptops that connect to a home office via a Virtual 771 Private Network (VPN)) can instead be configured for operation as a 772 simple host connected to the VET interface. Such VET hosts perform 773 the same VET interface autoconfiguration procedures as specified for 774 EBRs in Section 4.2.1, but they configure their VET interfaces as 775 host interfaces (and not router interfaces). VET hosts can then send 776 packets to the EID addresses of other hosts on the VET interface, or 777 to off-enterprise EID destinations via a next-hop EBR. 779 Note that a node may be configured as a host on some VET interfaces 780 and as an EBR/EBG on other VET interfaces. 782 5. Internetworking Operation 784 Following the autoconfiguration procedures specified in Section 4, 785 ERs, EBRs, EBGs, and VET hosts engage in normal internetworking 786 operations as discussed in the following sections. 788 5.1. Routing Protocol Participation 790 Following autoconfiguration, ERs engage in any RLOC-based IP routing 791 protocols and forward IP packets with RLOC addresses. EBRs can 792 additionally engage in any EID-based IP routing protocols and forward 793 IP packets with EID addresses. Note that the EID-based IP routing 794 domains are separate and distinct from any RLOC-based IP routing 795 domains. 797 5.2. RLOC-Based Communications 799 When permitted by policy and supported by routing, end systems can 800 avoid VET interface encapsulation through communications that 801 directly invoke the outer IP protocol using RLOC addresses instead of 802 EID addresses. End systems can use source address selection rules to 803 determine whether to use EID or RLOC addresses based on, e.g., name 804 service information. 806 5.3. EID-Based Communications 808 In many enterprise scenarios, the use of EID-based communications 809 (i.e., instead of RLOC-based communications) may be necessary and/or 810 beneficial to support address scaling, NAT avoidance, security domain 811 separation, site multihoming, traffic engineering, etc. The 812 remainder of this section discusses internetworking operation for 813 EID-based communications using the VET interface abstraction. 815 5.4. IPv6 Router and Prefix Discovery 817 The following sections discuss router and prefix discovery 818 considerations for the case of IPv6 as the inner IP protocol and IPv4 819 as the outer protocol. Router discovery and prefix discovery for 820 other IP protocol combinations are out of scope. 822 5.4.1. Router and Prefix Discovery 824 VET nodes follow the router and prefix discovery procedures specified 825 in [RFC5214], Section 8.3. They discover EBGs within the enterprise 826 as specified in Section 4.2.1.2, then perform RS/RA exchanges with 827 the EBGs to establish and maintain routes and prefixes. Depending on 828 the enterprise network trust basis, VET nodes may be required to use 829 SEND to secure the RS/RA exchanges. 831 EBGs follow the router and prefix discovery procedures specified in 832 [RFC5214], Section 8.2. They send solicited RAs over VET interfaces 833 for which they are configured as gateways where the RAs include 834 Router Lifetimes, Prefix Information Options (PIOs) that contain PA 835 prefixes for SLAAC, and with other required options/parameters. The 836 RAs can also include PIOs with the 'L' bit set to 0 and with a prefix 837 such as '2001:DB8::/48' as a hint of an aggregated prefix from which 838 the EBG is willing to delegate longer PA prefixes. When PIOs that 839 contain PA prefixes for SLAAC are included, the 'M' flag in the RA 840 should also be set to 0. 842 When an EBG receives an RS on a VET interface, it authenticates the 843 message then proceeds according to whether/not the VET interface 844 maintains a neighbor cache. If the VET interface maintains a 845 neighbor cache, the EBG first creates or updates a neighbor cache 846 entry for the VET link-local source address in the RS according to 847 Section 6.2.6 of [RFC4861]. If the neighbor cache entry cannot be 848 created/updated (e.g., due to insufficient resources), the EBG 849 silently discards the RS message and does not send an RA. Otherwise, 850 the EBG creates/updates the neighbor cache entry, sets a "Time To 851 Live (TTL)" on the entry that is no shorter than the its advertised 852 Router Lifetime, and sends the RA response to the RS. If the 853 neighbor cache entry TTL subsequently expires before a new RS 854 arrives, the EBG deletes the neighbor cache entry. Note that if the 855 VET interface does not maintain a neighbor cache, the EBG simply 856 omits these neighbor cache manipulations and sends the RA response to 857 the RS. 859 When the VET node receives an RA on a VET interface, it authenticates 860 the message then configures a default route based on the Router 861 Lifetime. Thereafter, the VET node accepts packets that are 862 forwarded by EBGs for which it has current default routing 863 information (i.e., ingress filtering is based on the default router 864 trust relationship rather than a prefix-specific ingress filter 865 entry). If the RA also contains Prefix Information Options (PIOs) 866 with the 'A' and 'L' bits set to 1, the VET node autoconfigures IPv6 867 VET addresses from the advertised prefixes and assigns them to the 868 VET interface as specified in Section 5.4.2. 870 5.4.2. Address Autoconfiguration on VET Interfaces 872 VET nodes perform address autoconfiguration to generate both VET 873 addresses and other types of IPv6 addresses (e.g.,CGA addresses 874 [RFC3972], IPv6 Privacy addresses [RFC4941], etc.). 876 When a VET node generates a VET address, it first creates an ISATAP 877 interface identifer that embeds its IPv4 RLOC address as specified in 878 Section 6.1 of [RFC5214]. The node then configures IPv6 unicast VET 879 addresses from advertised on-link prefixes received in RA messages 880 according to [RFC4862] and assigns them to the VET interface, i.e., 881 it does not perform Duplicate Address Detection (DAD) on the 882 addresses since the embedded IPv4 RLOC address already provides 883 uniqueness. 885 When the node self-generates a non-VET IPv6 address from one of the 886 EBG's advertised on-link prefixes, it first verifies that the 887 interface identifier does not begin with the reserved tokens "00-00- 888 5E-FE" or "02-00-5E-FE"; otherwise, it repeats the self-generation 889 process until it obtains an interface identifier that does not 890 collide. The VET node next marks each self-generated non-VET address 891 as tentative and uses the address as the target address in an IPv6 892 Neighbor Solicitation (NS) message [RFC4861] used for DAD [RFC4862]. 894 If the self-generated address is a CGA, the node also includes SEND 895 credentials to prove address ownership. The VET node sets the IPv6 896 source address of the NS to 0 and sets the IPv6 destination address 897 to the VET link-local unicast address of the EBG that advertised the 898 prefix, but does not include a Source Link-Layer Address Option 899 (SLLAO) in the NS message. 901 When the EBG receives the NS, it checks for the VET link-local 902 address of the sender in its neighbor cache. If the EBG does not 903 have the address in the cache (i.e., the EBG has not received a 904 recent RS from this VET node), it silently discards the NS. 905 Otherwise, it checks for the target address in the NS mesage in its 906 neighbor cache. If the target address is not already in the cache, 907 the EBG first verifies the NS SEND credentials (if present) then 908 creates a new neighbor cache entry for the target address in the 909 STALE state and records the IPv4 RLOC source address of the 910 requesting node as the link layer address. It the target address is 911 already in the cache and its link-layer address matches the IPv4 RLOC 912 source address of the NS, the EBG updates the neighbor cache entry 913 TTL. If the target address is already in the neighbor cache and its 914 link-layer address does not match the IPv4 source address of the NS, 915 the VET node's tentative address is a duplicate and the NS does not 916 update the cache. 918 The EBG then prepares a Neighbor Advertisement (NA) message with the 919 IPv6 source address set to the EBG's VET link-local address, the IPv6 920 destination address set to the VET node's VET link-local address, the 921 target address set to the VET node's proposed self-generated address, 922 and with a Target Link-Layer Address Option (TLLAO) formatted using a 923 modified version of the form specified in Section 5 of [RFC2529], 924 i.e., as shown in Figure 2: 926 +-------+-------+-------+-------+-------+-------+-------+-------+ 927 | Type |Length | TTL | IPv4 Address | 928 +-------+-------+-------+-------+-------+-------+-------+-------+ 930 Figure 2: VET Link-Layer Address Option Format 932 The EBG sets "IPv4 address" in the TLLAO option to the IPv4 RLOC 933 address of the VET node if the address was not a duplicate, sets 934 "IPv4 address" to the EBG's own IPv4 RLOC address if the address was 935 a duplicate, or sets "IPv4 address" to 0 if the address could not 936 otherwise be assigned (e.g., due to incorrect SEND credentails, 937 insufficient resources, etc.). The EBG also sets "TTL" to the 938 maximum time in seconds that the VET node is permitted to use the 939 address, where the value '0' means that the address was either a 940 duplicate or cannot otherwise be used. The EBG then sends the NA 941 message to the VET node. 943 When the VET node receives the NA message, it does not update its 944 neighbor cache but rather checks the NA to verify that it is 945 authorized to use the non-VET address. In particular, if the TLLAO 946 contains a non-zero TTL and IPv4 address set to the VET node's IPv4 947 RLOC address, the VET node assigns the address to the VET interface 948 and can subsequently use it as the IPv6 source address for on- and 949 off-link communications. If the VET node wishes to subsequently 950 extend the lifetime of the non-VET address beyond TTL seconds, it 951 must send additional NS(DAD) messages as above to update the EBG's 952 neighbor cache. 954 This implies that EBGs that maintain a neighbor cache can provide an 955 address registration service for VET nodes that will autoconfigure 956 non-VET IPv6 addresses, and that the EBG sets the frequency with 957 which non-VET IPv6 addresses may be updated or deprecated (i.e. by 958 setting the TTL). The EBG must therefore maintain neighbor cache 959 entries indexed by the node's IPv4 RLOC address (i.e., as the link- 960 layer address) for each non-VET IPv6 address that the VET node 961 autoconfigures in addition to maintaining a neighbor cache entry for 962 the node's VET link-local address. EBG neighbor cache entries for 963 non-VET addresses are therefore purged under two possible 964 circumstances: 1) that the non-VET address expires due to no NS(DAD) 965 message being received within the TTL timeout period, or 2) that the 966 VET link-local address of the VET node expires due to no RS message 967 being received within the prefix lifetime timeout period. 969 5.4.3. PA Prefix Registration 971 After an EBR discovers default routes, it can use DHCP prefix 972 delegation to obtain PA prefixes via an EBG as specified in 973 Section 4.2.2. The DHCP server ensures that the delegations are 974 unique and that the EBG's router function will forward IP packets 975 over the VET interface to the correct EBR. In particular, the EBG 976 must register and track the PA prefixes that are delegated to each 977 EBR. 979 The PA prefix registrations remain active in the EBGs as long as the 980 EBR continues to issue DHCP renewals over the VET interface before 981 lease lifetimes expire. The lease lifetime also keeps the delegation 982 state active even if communications between the EBR and DHCP server 983 are disrupted for a period of time (e.g., due to an enterprise 984 network partition) before being reestablished (e.g., due to an 985 enterprise network merge). 987 5.4.4. PI Prefix Registration 989 After an EBR discovers default routes, it must register its PI 990 prefixes by sending RAs to a set of one or more EBGs with Route 991 Information Options (RIOs) [RFC4191] that contain the EBR's PI 992 prefixes. Each RA must include the RLOC of an EBG as the outer IP 993 destination address and a link-local address assigned to the VET 994 interface as the inner IP destination address. For enterprises that 995 use SEND, the RAs also include a CGA link-local inner source address, 996 SEND credentials, plus any certificates needed to prove ownership of 997 the PI prefixes. The EBR additionally tracks the set of EBGs to 998 which it sends RAs so that it can send subsequent RAs to the same 999 set. 1001 When the EBG receives the RA, it first authenticates the message; if 1002 the authentication fails, the EBG discards the RA. Otherwise, the 1003 EBG installs the PI prefixes with their respective lifetimes in its 1004 Forwarding Information Base (FIB) and configures them for both 1005 ingress filtering [RFC3704] and forwarding purposes. In particular, 1006 the EBG configures the FIB entries as ingress filter rules to accept 1007 packets received on the VET interface that have a source address 1008 taken from the PI prefixes. It also configures the FIB entries to 1009 permit forwarding of packets with a destination address taken from 1010 the PI prefixes to the EBR that registered the prefixes on the VET 1011 interface. 1013 The EBG then publishes the PI prefixes in a distributed mapping 1014 database (e.g., in a private instance of a routing protocol in which 1015 only EBGs participate, via an automated name-service update mechanism 1016 [RFC3007], etc.). For enterprises that are managed under a 1017 centralized administrative authority, the EBG also publishes the PI 1018 prefixes in the enterprise-local name-service (e.g., the enterprise- 1019 local DNS [RFC1035]). 1021 In particular, the EBG publishes each /56 prefix taken from the PI 1022 prefixes as a separate Fully Qualified Domain Name (FQDN) that 1023 consists of a sequence of 14 nibbles in reverse order (i.e., the same 1024 as in [RFC3596], Section 2.5) followed by the string 'ip6' followed 1025 by the string 'PRLNAME'. For example, when 'PRLNAME' is 1026 "isatapv2.example.com", the EBG publishes the prefix '2001:DB8::/56' 1027 as: 1028 '0.0.0.0.0.0.8.b.d.0.1.0.0.2.ip6.isatapv2.example.com'. 1030 The EBG includes the outer RLOC source address of the RA (e.g., in a 1031 DNS A resource record) in each prefix publication. For enterprises 1032 that use SEND, the EBG also includes the inner IPv6 CGA source 1033 address (e.g., in a DNS AAAA record) in each prefix publication. If 1034 the prefix was already installed in the distributed database, the EBG 1035 instead adds the outer RLOC source address (e.g., in an additional 1036 DNS A record) to the preexisting publication to support PI prefixes 1037 that are multihomed. For enterprises that use SEND, this latter 1038 provision requires all EBRs of a multihomed site that advertise the 1039 same PI prefixes in RAs to use the same CGA and the same SEND 1040 credentials. 1042 After the EBG authenticates the RA and publishes the PI prefixes, it 1043 next acts as a Neighbor Discovery proxy (NDProxy) [RFC4389] on the 1044 VET interfaces configured over any of its parent enterprises, and it 1045 relays a proxied RA to the EBGs on those interfaces. (For 1046 enterprises that use SEND, the EBG additionally acts as a SEcure 1047 Neighbor Discovery Proxy (SENDProxy) [I-D.ietf-csi-proxy-send].) 1048 EBGs in parent enterprises that receive the proxied RAs in turn act 1049 as NDProxys/SENDProxys to relay the RAs to EBGs on their parent 1050 enterprises, etc. The RA proxying and PI prefix publication recurses 1051 in this fashion and ends when an EBR attached to an interdomain 1052 routing core is reached. 1054 After the initial PI prefix registration, the EBR that owns the 1055 prefix(es) must periodically send additional RAs to its set of EBGs 1056 to refresh prefix lifetimes. Each such EBG tracks the set of EBGs in 1057 parent enterprises to which it relays the proxied RAs, and should 1058 relay subsequent RAs to the same set. 1060 This procedure has a direct analogy in the Teredo method of 1061 maintaining state in network middleboxes through the periodic 1062 transmission of "bubbles" [RFC4380]. 1064 5.4.5. Next-Hop Discovery 1066 VET nodes discover destination-specific next-hop EBRs within the 1067 enterprise by querying the name service for the /56 IPv6 PI prefix 1068 taken from a packet's destination address, by forwarding packets via 1069 a default route to an EBG, or by some other inner-IP to outer-IP 1070 address mapping mechanism. For example, for the IPv6 destination 1071 address '2001:DB8:1:2::1' and 'PRLNAME' "isatapv2.example.com" the 1072 VET node can lookup the domain name: 1073 '0.0.1.0.0.0.8.b.d.0.1.0.0.2.ip6.isatapv2.example.com'. 1075 If the name-service lookup succeeds, it will return RLOC addresses 1076 (e.g., in DNS A records) that correspond to next-hop EBRs to which 1077 the VET node can forward packets. (In enterprises that use SEND, it 1078 will also return an IPv6 CGA address, e.g., in a DNS AAAA record.) 1080 Name-service lookups in enterprises with a centralized management 1081 structure use an infrastructure-based service, e.g., an enterprise- 1082 local DNS. Name-service lookups in enterprises with a distributed 1083 management structure and/or that lack an infrastructure-based name- 1084 service instead use LLMNR over the VET interface. When LLMNR is 1085 used, the EBR that performs the lookup sends an LLMNR query (with the 1086 /56 prefix taken from the IP destination address encoded in dotted- 1087 nibble format as shown above) and accepts the union of all replies it 1088 receives from other EBRs on the VET interface. When an EBR receives 1089 an LLMNR query, it responds to the query IFF it aggregates an IP 1090 prefix that covers the prefix in the query. 1092 Alternatively, in enterprises with a stable and highly-available set 1093 of EBGs, the VET node can simply forward an initial packet via a 1094 default route to an EBG. The EBG will forward the packet to a next- 1095 hop EBR on the VET interface and return an ICMPv6 Redirect [RFC4861] 1096 (using SEND, if necessary). If the packet's source address is on- 1097 link on the VET interface, the EBG returns an ordinary "router-to- 1098 host" redirect with the source address of the packet as its 1099 destination. If the packet's source address is not on-link, the EBG 1100 instead returns a "router-to-router" redirect with the link-local VET 1101 address of the previous-hop EBR as its destination. 1103 When IPv4 is used as the outer IP protocol, the EBG includes in the 1104 redirect one or more IPv6 TLLAOs formatted as specified n Section 1105 5.4.2. The TLLAOs contain the IPv4 RLOCs of potential next-hop EBRs 1106 arranged in order from lowest to highest priority (i.e., the first 1107 TLLAO contains the lowest priority RLOC and the final TLLAO option 1108 contains the highest priority). For each such IPv6/IPv4 LLAO, the 1109 Type is set to 2 (for Target Link-Layer Address Option), Length is 1110 set to 1, and IPv4 Address is set to the IPv4 RLOC of the next-hop 1111 EBR. TTL is set to the time in seconds that the recipient may cache 1112 the RLOC, where the value 65535 represents infinity and the value 0 1113 suspends forwarding through this RLOC. 1115 When a VET host receives an ordinary "router-to-host" redirect, it 1116 processes the redirect exactly as specified in [RFC4861], Section 8. 1117 When an EBR receives a "router-to-router" redirect, it discovers the 1118 RLOC addresses of potential next-hop EBRs by examining the LLAOs 1119 included in the redirect. The EBR then installs a FIB entry that 1120 contains the /56 prefix of the destination address encoded in the 1121 redirect and the list of RLOCs of potential next-hop EBRs. The EBR 1122 then enables the FIB entry for forwarding to next-hop EBRs but DOES 1123 NOT enable it for ingress filtering acceptance of packets from next- 1124 hop EBRs (i.e., the forwarding determination is unidirectional). 1126 In enterprises in which spoofing is possible, after discovering 1127 potential next-hop EBRs (either through name-service lookup or ICMP 1128 redirect) the EBR must send authenticating credentials before 1129 forwarding packets via the next-hops. To do so, the EBR must send 1130 RAs over the VET interface (using SEND, if necessary) to the RLOCs of 1131 one or more of the potential next-hop EBRs. The RAs must include a 1132 Route Information Option (RIO) [RFC4191] that contains the /56 PI 1133 prefix of the original packet's source address. After sending the 1134 RAs, the EBR can either enable the new FIB entry for forwarding 1135 immediately or delay until it receives an explicit acknowledgement 1136 that a next-hop EBR received the RA (e.g., using the SEAL explicit 1137 acknowledgement mechanism -- see Section 5.7). 1139 When a next-hop EBR receives the RA, it authenticates the message 1140 then it performs a name-service lookup on the prefix in the RIO if 1141 further authenticating evidence is required. If the name service 1142 returns resource records that are consistent with the inner and outer 1143 IP addresses of the RA, the next-hop EBR then installs the prefix in 1144 the RIO in its FIB and enables the FIB entry for ingress filtering 1145 but DOES NOT enable it for forwarding purposes. After an EBR sends 1146 initial RAs following a redirect, it should send periodic RAs to 1147 refresh the next-hop EBR's ingress filter prefix lifetimes as long as 1148 traffic is flowing. 1150 EBRs retain the FIB entries created as a result of an ICMP redirect 1151 until all RLOC TTLs expire, or until no hints of forward progress 1152 through any of the associated RLOCs are received. In this way, RLOC 1153 liveness detection exactly parallels IPv6 Neighbor Unreachability 1154 Detection ([RFC4861], Section 3). 1156 5.5. IPv4 Router and Prefix Discovery 1158 When IPv4 is used as the inner IP protocol, router discovery and 1159 prefix registration exactly parallel the mechanisms specified for 1160 IPv6 in Section 5.4. To support this, modifications to the ICMPv4 1161 Router Advertisement [RFC1256] function to include SEND constructs 1162 and modifications to the ICMPv4 Redirect [RFC0792] function to 1163 support router-to-router redirects will be specified in a future 1164 document. Additionally, publications for IPv4 prefixes will be in 1165 dotted-nibble format in the 'ip4.isatapv2.example.com' domain. For 1166 example, the IPv4 prefix 192.0.2/24 would be represented as: 1167 '2.0.0.0.0.c.ip4.isatapv2.example.com' 1169 5.6. Forwarding Packets on VET Interfaces 1171 VET nodes forward packets by consulting the FIB to determine a route 1172 with a next-hop toward the destination, where the next-hop is the 1173 destination itself if the destination matches an interface's on-link 1174 prefix. When multiple routes are available, VET nodes can use 1175 default router preferences, routing protocol information, traffic 1176 engineering configurations, etc. to select the best route. When no 1177 routes other than "default" are available, VET nodes can discover the 1178 best next-hop through the mechanisms specified in Section 5.4 and 1179 Section 5.5. 1181 When the VET node selects a route with a next-hop configured on a VET 1182 interface that uses IPv6-in-SEAL-in-IPv4 encapsulation, it next 1183 performs next-hop address resolution. For VET addresses, the VET 1184 node performs address resolution through static extraction of the 1185 embedded IPv4 RLOC in the VET interface identifier. For non-VET 1186 addresses, the VET node first checks for the next-hop address in its 1187 neigbhbor cache. If the address is in the cache, address resolution 1188 is through static extraction of the IPv4 RLOC address recorded in the 1189 link-layer address. Otherwise, if the node is not the EBG it sends a 1190 Neighbor Solicitation (NS) message to the EBG with its VET interface 1191 link-local address as the IPv6 source, the VET link-local address of 1192 the EBG as the destination, and the next-hop address as the target. 1194 When the EBG receives the NS message, it does not update or create a 1195 neighbor cache entry for the source of the solicitation, but rather 1196 returns an immediate Neighbor Advertisement (NA) message. If the EBG 1197 has an entry for the target address in its neighbor cache, it returns 1198 an NA message with the next-hop address as the target, the link-local 1199 address of the EBG as the IPv6 source, the link-local address of the 1200 VET node as the IPv6 destination, and with a Target Link Layer 1201 Address Option (TLLAO) formatted as specified in Section 5.4.2 that 1202 encodes a TTL and the IPv4 RLOC address of the VET node that owns the 1203 next-hop address. If the EBG does not have a neighbor cache entry 1204 for the target address, it returns an NA message as above except that 1205 the TLLAO encodes the value 0 in both the TTL and IPv4 address 1206 fields. 1208 When the VET node receives the NA message, it can immediately 1209 determine whether address resolution has succeeded or failed by 1210 examining the results recorded in the TLLAO. If the address 1211 resolution succeeds, the VET node records the address in its neighbor 1212 cache and retains the neighbor cache entry for up to the duration 1213 recorded in the TTL. If the address resolution fails, the VET node 1214 discards the packet and/or selects an alternate route (if one is 1215 available). 1217 Note that the above address resolution is peformed only for IPv6-in- 1218 SEAL-in-IPv4 encapsulation. For other encapsulations, address 1219 resolutuion is through administrative configuration or through an 1220 unspecified alternate method. 1222 5.7. VET and SEAL Encapsulation 1224 After address resolution, the VET interface encapsulates the inner IP 1225 packet in any mid-layer headers (e.g., IPsec [RFC4301]) followed a 1226 SEAL header [I-D.templin-intarea-seal] followed by an outer IP 1227 header; it next submits the encapsulated packet to the outer IP 1228 forwarding engine for transmission on an underlying interface. 1230 VET interfaces use SEAL encapsulation to accommodate path MTU 1231 diversity, to defeat source address spoofing, and to enable sub-IP 1232 layer hints of forward progress that can be piggybacked on ordinary 1233 data messages. SEAL encapsulation maintains a unidirectional and 1234 monotonically incrementing per-packet identification value known as 1235 the 'SEAL_ID'. When a VET node that uses SEAL encapsulation sends a 1236 SEND-protected Router Advertisement (RA) or Router Solicitation (RS) 1237 message to another VET node, both nodes can cache the new SEAL_ID as 1238 per-tunnel state used for maintaining a window of unacknowledged 1239 SEAL_IDs. 1241 In terms of security, when a VET node receives an ICMP message or a 1242 SEAL error message, it can confirm that the packet-in-error within 1243 the message corresponds to one of its recently sent packets by 1244 examining the SEAL_ID along with source and destination addresses, 1245 etc. Additionally, a next-hop EBR can track the SEAL_ID in packets 1246 received from EBRs for which there is an ingress filter entry and 1247 discard packets that have SEAL_ID values outside of the current 1248 window. 1250 In terms of next-hop reachability, an EBR can set the SEAL 1251 "Acknowledgement Requested" bit in messages to receive confirmation 1252 that a next-hop EBR is reachable. (Note that this is a mid-layer 1253 reachability confirmation, and not an L2 reachability indication.) 1254 Setting the "Acknowledgement Requested" bit is also used as the 1255 method for maintaining the window of outstanding SEAL_IDs. 1257 5.8. Generating Errors 1259 When an EBR receives an IPv6 packet over a VET interface and there is 1260 no matching ingress filter entry, it drops the packet and returns an 1261 ICMPv6 [RFC4443] "Destination Unreachable; Source address failed 1262 ingress/egress policy" message to the previous-hop EBR subject to 1263 rate limiting. 1265 When an EBR receives an IPv6 packet over a VET interface, and there 1266 is no longest-prefix-match FIB entry for the destination, it returns 1267 an ICMPv6 "Destination Unreachable; No route to destination" message 1268 to the previous hop EBR subject to rate limiting. 1270 When an EBR receives an IPv6 packet over a VET interface and the 1271 longest-prefix-match FIB entry for the destination is via a next-hop 1272 configured over the same VET interface the packet arrived on, the EBR 1273 forwards the packet. If the FIB prefix is longer than ::/0, the EBR 1274 then sends a router-to-router ICMPv6 Redirect message (using SEND, if 1275 necessary) to the previous-hop EBR as specified in Section 5.4.5. 1277 Generation of other ICMP messages [RFC0792] [RFC4443] is the same as 1278 for any IP interface. 1280 5.9. Processing Errors 1282 When an EBR receives an ICMPv6 "Destination Unreachable; Source 1283 address failed ingress/egress policy" message from a next-hop EBR, 1284 and there is a longest-prefix-match FIB entry for the original 1285 packet's destination that is more specific than ::/0, the EBR 1286 discards the message and marks the FIB entry for the destination as 1287 "forwarding suspended" for the RLOC taken from the source address of 1288 the ICMPv6 message. The EBR should then allow subsequent packets to 1289 flow through different RLOCs associated with the FIB entry until it 1290 forwards a new RA to the suspended RLOC. If the EBR receives 1291 excessive ICMPv6 ingress/egress policy errors through multiple RLOCs 1292 associated with the same FIB entry, it should delete the FIB entry 1293 and allow subsequent packets to flow through an EBG if supported in 1294 the specific enterprise scenario. 1296 When a VET node receives an ICMPv6 "Destination Unreachable; No route 1297 to destination" message from a next-hop EBR, it forwards the ICMPv6 1298 message to the source of the original packet as normal. If the EBR 1299 has a longest-prefix-match FIB entry for the original packet's 1300 destination that is more specific than ::/0, the EBR also deletes the 1301 FIB entry. 1303 When an EBR receives an authentic ICMPv6 Redirect, it processes the 1304 packet as specified in Section 5.4.5. 1306 When an EBG receives new mapping information for a specific 1307 destination prefix, it can propagate the update to other EBRs/EBGs by 1308 sending an ICMPv6 redirect message to the 'All Routers' link-local 1309 multicast address with an LLAO with the TTL for the unreachable LLAO 1310 set to zero, and with a NULL packet in error. 1312 Additionally, a VET node may receive ICMP "Destination Unreachable; 1313 net / host unreachable" messages from an ER indicating that the path 1314 to a VET neighbor may be failing. The VET node should first check, 1315 e.g., the SEAL_ID, IPsec sequence number, source address of the 1316 original packet if available, etc. to obtain reasonable assurance 1317 that the ICMP message is authentic, then should mark the longest- 1318 prefix-match FIB entry for the destination as "forwarding suspended" 1319 for the RLOC destination address of the ICMP packet-in-error. If the 1320 VET node receives excessive ICMP unreachable errors through multiple 1321 RLOCs associated with the same FIB entry, it should delete the FIB 1322 entry and allow subsequent packets to flow through a different route. 1324 5.10. Mobility and Multihoming Considerations 1326 EBRs that travel between distinct enterprise networks must either 1327 abandon their PA prefixes that are relative to the "old" enterprise 1328 and obtain new ones relative to the "new" enterprise or somehow 1329 coordinate with a "home" enterprise to retain ownership of the 1330 prefixes. In the first instance, the EBR would be required to 1331 coordinate a network renumbering event using the new PA prefixes 1332 [RFC4192]. In the second instance, an ancillary mobility management 1333 mechanism must be used. 1335 EBRs can retain their PI prefixes as they travel between distinct 1336 enterprise networks as long as they register the prefixes with new 1337 EBGs and (preferably) withdraw the prefixes from old EBGs prior to 1338 departure. Prefix registration with new EBGs is coordinated exactly 1339 as specified in Section 5.4.4; prefix withdrawal from old EBGs is 1340 simply through re-announcing the PI prefixes with zero lifetimes. 1342 Since EBRs can move about independently of one another, stale FIB 1343 entry state may be left in VET nodes when a neighboring EBR departs. 1344 Additionally, EBRs can lose state for various reasons, e.g., power 1345 failure, machine reboot, etc. For this reason, EBRs are advised to 1346 set relatively short PI prefix lifetimes in RIO options, and to send 1347 additional RAs to refresh lifetimes before they expire. (EBRs should 1348 place conservative limits on the RAs they send to reduce congestion, 1349 however.) 1351 EBRs may register their PI prefixes with multiple EBGs for 1352 multihoming purposes. EBRs should only forward packets via EBGs with 1353 which it has registered its PI prefixes, since other EBGs may drop 1354 the packets and return ICMPv6 "Destination Unreachable; Source 1355 address failed ingress/egress policy" messages. 1357 EBRs can also act as delegating routers to sub-delegate portions of 1358 their PI prefixes to requesting routers on their enterprise-edge 1359 interfaces and on VET interfaces for which they are configured as 1360 EBGs. In this sense, the sub-delegations of an EBR's PI prefixes 1361 become the PA prefixes for downstream-dependent nodes. 1363 The EBGs of a multihomed enterprise should participate in a private 1364 inner IP routing protocol instance between themselves (possibly over 1365 an alternate topology) to accommodate enterprise partitions/merges as 1366 well as intra-enterprise mobility events. These peer EBGs should 1367 accept packets from one another without respect to the destination 1368 (i.e., ingress filtering is based on the peering relationship rather 1369 than a prefix-specific ingress filter entry). 1371 5.11. Multicast 1373 In multicast-capable deployments, ERs provide an enterprise-wide 1374 multicasting service (e.g., Simplified Multicast Forwarding (SMF) 1375 [I-D.ietf-manet-smf], Protocol Independent Multicast (PIM) routing, 1376 Distance Vector Multicast Routing Protocol (DVMRP) routing, etc.) 1377 over their enterprise-interior interfaces such that outer IP 1378 multicast messages of site-scope or greater scope will be propagated 1379 across the enterprise. For such deployments, VET nodes can also 1380 provide an inner IP multicast/broadcast capability over their VET 1381 interfaces through mapping of the inner IP multicast address space to 1382 the outer IP multicast address space. In that case, operation of 1383 link-scoped (or greater scoped) inner IP multicasting services (e.g., 1384 a link-scoped neighbor discovery protocol) over the VET interface is 1385 available, but link-scoped services should be used sparingly to 1386 minimize enterprise-wide flooding. 1388 VET nodes encapsulate inner IP multicast messages sent over the VET 1389 interface in any mid-layer headers (e.g., IPsec, etc.) followed by a 1390 SEAL header followed by an outer IP header with a site-scoped outer 1391 IP multicast address as the destination. For the case of IPv6 and 1392 IPv4 as the inner/outer protocols (respectively), [RFC2529] provides 1393 mappings from the IPv6 multicast address space to a site-scoped IPv4 1394 multicast address space (for other encapsulations, mappings are 1395 established through administrative configuration or through an 1396 unspecified alternate static mapping). 1398 Multicast mapping for inner IP multicast groups over outer IP 1399 multicast groups can be accommodated, e.g., through VET interface 1400 snooping of inner multicast group membership and routing protocol 1401 control messages. To support inner-to-outer IP multicast mapping, 1402 the VET interface acts as a virtual outer IP multicast host connected 1403 to its underlying interfaces. When the VET interface detects that an 1404 inner IP multicast group joins or leaves, it forwards corresponding 1405 outer IP multicast group membership reports on an underlying 1406 interface over which the VET interface is configured. If the VET 1407 node is configured as an outer IP multicast router on the underlying 1408 interfaces, the VET interface forwards locally looped-back group 1409 membership reports to the outer IP multicast routing process. If the 1410 VET node is configured as a simple outer IP multicast host, the VET 1411 interface instead forwards actual group membership reports (e.g., 1412 IGMP messages) directly over an underlying interface. 1414 Since inner IP multicast groups are mapped to site-scoped outer IP 1415 multicast groups, the VET node must ensure that the site-scope outer 1416 IP multicast messages received on the underlying interfaces for one 1417 VET interface do not "leak out" to the underlying interfaces of 1418 another VET interface. This is accommodated through normal site- 1419 scoped outer IP multicast group filtering at enterprise boundaries. 1421 5.12. Service Discovery 1423 VET nodes can perform enterprise-wide service discovery using a 1424 suitable name-to-address resolution service. Examples of flooding- 1425 based services include the use of LLMNR [RFC4795] over the VET 1426 interface or multicast DNS (mDNS) [I-D.cheshire-dnsext-multicastdns] 1427 over an underlying interface. More scalable and efficient service 1428 discovery mechanisms are for further study. 1430 5.13. Enterprise Partitioning 1432 EBGs can physically partition an enterprise by configuring multiple 1433 VET interfaces over multiple distinct sets of underlying interfaces. 1434 In that case, each partition (i.e., each VET interface) must 1435 configure its own distinct 'PRLNAME' (e.g., 1436 'isatapv2.zone1.example.com', 'isatapv2.zone2.example.com', etc.). 1438 EBGs can logically partition an enterprise using a single VET 1439 interface by sending RAs with PIOs containing different IPv6 PA 1440 prefixes to group nodes into different logical partitions. EBGs can 1441 identify partitions, e.g., by examining RLOC prefixes, observing the 1442 interfaces over which RSs are received, etc. In that case, a single 1443 'PRLNAME' can cover all partitions. 1445 5.14. EBG Prefix State Recovery 1447 EBGs must retain explicit state that tracks the inner IP prefixes 1448 owned by EBRs within the enterprise, e.g., so that packets are 1449 delivered to the correct EBRs and not incorrectly "leaked out" of the 1450 enterprise via a default route. For PA prefixes, the state is 1451 maintained via an EBR's DHCP prefix delegation lease renewals, while 1452 for PI prefixes the state is maintained via an EBR's periodic prefix 1453 registration RAs. 1455 When an EBG loses some or all of its state (e.g., due to a power 1456 failure), it must recover the state so that packets can be forwarded 1457 over correct routes. If the EBG aggregates PA prefixes from which 1458 the IP prefixes of all EBRs in the enterprise are sub-delegated, then 1459 the EBG can recover state through DHCP prefix delegation lease 1460 renewals, through bulk lease queries, or through on-demand name- 1461 service lookups based on IP packet forwarding. If the EBG serves as 1462 an anchor for PI prefixes, however, care must be taken to avoid 1463 looping while state is recovered through prefix registration RAs from 1464 EBRs. In that case, when the EBG that is recovering state forwards 1465 an IP packet for which it has no explicit route other than ::/0, it 1466 must first perform an on-demand name-service lookup to refresh state. 1468 5.15. Support for Legacy ISATAP Services 1470 EBGs support legacy ISATAP services according to the specifications 1471 in [RFC5214]. In particular, EBGs can configure legacy ISATAP 1472 interfaces and VET interfaces over the same sets of underlying 1473 interface as long as the IPv6 prefixes associated with the ISATAP/VET 1474 interfaces are distinct. 1476 6. IANA Considerations 1478 There are no IANA considerations for this document. 1480 7. Security Considerations 1482 Security considerations for MANETs are found in [RFC2501]. 1484 The security considerations found in [RFC2529] [RFC5214] also apply 1485 to VET. In particular: 1487 o VET nodes must ensure that a VET interface does not span multiple 1488 sites as specified in Section 6.2 of [RFC5214]. 1490 o VET nodes must verify that the outer IP source address of a packet 1491 received on a VET interface is correct for the inner IP source 1492 address; for the case of IPv6-in-SEAL-in-IPv4 encapsulation, this 1493 is accomodated using the procedures specified in Section 7.3 of 1494 [RFC5214]. 1496 o EBRs/EBGs must implement both inner and outer IP ingress filtering 1497 in a manner that is consistent with [RFC2827] as well as 1498 ip-proto-41 filtering. When the node at the physical boundary of 1499 the enterprise is an ER (i.e., and not an EBR/EBG), the ER itself 1500 should implement filtering. 1502 Additionally, when an EBG receives an IPv6-in-SEAL-in-IPv4 1503 encapsulated packet with an IPv6 destination address whose prefix 1504 does not match one of the EBG's on-link prefixes, if the IPv6 1505 interface identifier encodes the value "0200:5EFE:V4ADDR" (where 1506 V4ADDR is one of the EBG's own IPv4 addresses) the EBG drops the 1507 packet. This mitigation is necessary to avoid a potential routing 1508 loop between two EBGs on different VET links. For EBGs that can 1509 maintain a coherent neighbor cache, an additional simple mitgation 1510 can be employed to avoid accidental routing loops within a site. In 1511 particular, when the EBG forwards a packet with a next-hop address 1512 that matches an on-link prefix on one of its VET interfaces, it 1513 should first verify that the VET interface has a neighbor cache entry 1514 corresponding to the next-hop as specified in Section 5.6. If the 1515 neighbor cache entries exist, the EBG forwards the packet; otherwise, 1516 it drops the packet. 1518 SEND [RFC3971], IPsec [RFC4301], and SEAL [I-D.templin-intarea-seal] 1519 provide additional securing mitigations to detect source address 1520 spoofing and bogus RA messages sent by rogue routers. 1522 Rogue routers can send bogus RA messages with spoofed RLOC source 1523 addresses that can consume network resources and cause EBGs to 1524 perform extra work. Nonetheless, EBGs should not "blacklist" such 1525 RLOCs, as that may result in a denial of service to the RLOCs' 1526 legitimate owners. 1528 8. Related Work 1530 Brian Carpenter and Cyndi Jung introduced the concept of intra-site 1531 automatic tunneling in [RFC2529]; this concept was later called: 1532 "Virtual Ethernet" and investigated by Quang Nguyen under the 1533 guidance of Dr. Lixia Zhang. Subsequent works by these authors and 1534 their colleagues have motivated a number of foundational concepts on 1535 which this work is based. 1537 Telcordia has proposed DHCP-related solutions for MANETs through the 1538 CECOM MOSAIC program. 1540 The Naval Research Lab (NRL) Information Technology Division uses 1541 DHCP in their MANET research testbeds. 1543 Security concerns pertaining to tunneling mechanisms are discussed in 1544 [I-D.ietf-v6ops-tunnel-security-concerns]. 1546 Default router and prefix information options for DHCPv6 are 1547 discussed in [I-D.droms-dhc-dhcpv6-default-router]. 1549 An automated IPv4 prefix delegation mechanism is proposed in 1550 [I-D.ietf-dhc-subnet-alloc]. 1552 RLOC prefix delegation for enterprise-edge interfaces is discussed in 1553 [I-D.clausen-manet-autoconf-recommendations]. 1555 MANET link types are discussed in [I-D.clausen-manet-linktype]. 1557 Various proposals within the IETF have suggested similar mechanisms. 1559 9. Acknowledgements 1561 The following individuals gave direct and/or indirect input that was 1562 essential to the work: Jari Arkko, Teco Boot, Emmanuel Bacelli, James 1563 Bound, Scott Brim, Brian Carpenter, Thomas Clausen, Claudiu Danilov, 1564 Chris Dearlove, Ralph Droms, Dino Farinacci, Vince Fuller, Thomas 1565 Goff, Joel Halpern, Bob Hinden, Sascha Hlusiak, Sapumal Jayatissa, 1566 Dan Jen, Darrel Lewis, Tony Li, Joe Macker, David Meyer, Gabi 1567 Nakibly, Thomas Narten, Pekka Nikander, Dave Oran, Alexandru 1568 Petrescu, John Spence, Jinmei Tatuya, Dave Thaler, Ole Troan, 1569 Michaela Vanderveen, Lixia Zhang, and others in the IETF AUTOCONF and 1570 MANET working groups. Many others have provided guidance over the 1571 course of many years. 1573 10. Contributors 1575 The following individuals have contributed to this document: 1577 Eric Fleischman (eric.fleischman@boeing.com) 1578 Thomas Henderson (thomas.r.henderson@boeing.com) 1579 Steven Russert (steven.w.russert@boeing.com) 1580 Seung Yi (seung.yi@boeing.com) 1582 Ian Chakeres (ian.chakeres@gmail.com) contributed to earlier versions 1583 of the document. 1585 Jim Bound's foundational work on enterprise networks provided 1586 significant guidance for this effort. We mourn his loss and honor 1587 his contributions. 1589 11. References 1591 11.1. Normative References 1593 [I-D.templin-intarea-seal] 1594 Templin, F., "The Subnetwork Encapsulation and Adaptation 1595 Layer (SEAL)", draft-templin-intarea-seal-05 (work in 1596 progress), July 2009. 1598 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 1599 September 1981. 1601 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 1602 RFC 792, September 1981. 1604 [RFC0826] Plummer, D., "Ethernet Address Resolution Protocol: Or 1605 converting network protocol addresses to 48.bit Ethernet 1606 address for transmission on Ethernet hardware", STD 37, 1607 RFC 826, November 1982. 1609 [RFC1035] Mockapetris, P., "Domain names - implementation and 1610 specification", STD 13, RFC 1035, November 1987. 1612 [RFC2131] Droms, R., "Dynamic Host Configuration Protocol", 1613 RFC 2131, March 1997. 1615 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1616 (IPv6) Specification", RFC 2460, December 1998. 1618 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 1619 Defeating Denial of Service Attacks which employ IP Source 1620 Address Spoofing", BCP 38, RFC 2827, May 2000. 1622 [RFC3007] Wellington, B., "Secure Domain Name System (DNS) Dynamic 1623 Update", RFC 3007, November 2000. 1625 [RFC3315] Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C., 1626 and M. Carney, "Dynamic Host Configuration Protocol for 1627 IPv6 (DHCPv6)", RFC 3315, July 2003. 1629 [RFC3596] Thomson, S., Huitema, C., Ksinant, V., and M. Souissi, 1630 "DNS Extensions to Support IP Version 6", RFC 3596, 1631 October 2003. 1633 [RFC3633] Troan, O. and R. Droms, "IPv6 Prefix Options for Dynamic 1634 Host Configuration Protocol (DHCP) version 6", RFC 3633, 1635 December 2003. 1637 [RFC3971] Arkko, J., Kempf, J., Zill, B., and P. Nikander, "SEcure 1638 Neighbor Discovery (SEND)", RFC 3971, March 2005. 1640 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 1641 RFC 3972, March 2005. 1643 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 1644 More-Specific Routes", RFC 4191, November 2005. 1646 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 1647 Architecture", RFC 4291, February 2006. 1649 [RFC4443] Conta, A., Deering, S., and M. Gupta, "Internet Control 1650 Message Protocol (ICMPv6) for the Internet Protocol 1651 Version 6 (IPv6) Specification", RFC 4443, March 2006. 1653 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 1654 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 1655 September 2007. 1657 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 1658 Address Autoconfiguration", RFC 4862, September 2007. 1660 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 1661 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 1662 March 2008. 1664 11.2. Informative References 1666 [CATENET] Pouzin, L., "A Proposal for Interconnecting Packet 1667 Switching Networks", May 1974. 1669 [I-D.cheshire-dnsext-multicastdns] 1670 Cheshire, S. and M. Krochmal, "Multicast DNS", 1671 draft-cheshire-dnsext-multicastdns-07 (work in progress), 1672 September 2008. 1674 [I-D.clausen-manet-autoconf-recommendations] 1675 Clausen, T. and U. Herberg, "MANET Router Configuration 1676 Recommendations", 1677 draft-clausen-manet-autoconf-recommendations-00 (work in 1678 progress), February 2009. 1680 [I-D.clausen-manet-linktype] 1681 Clausen, T., "The MANET Link Type", 1682 draft-clausen-manet-linktype-00 (work in progress), 1683 October 2008. 1685 [I-D.droms-dhc-dhcpv6-default-router] 1686 Droms, R. and T. Narten, "Default Router and Prefix 1687 Advertisement Options for DHCPv6", 1688 draft-droms-dhc-dhcpv6-default-router-00 (work in 1689 progress), March 2009. 1691 [I-D.hain-ipv6-ulac] 1692 Hain, T., Hinden, R., Huston, G., and T. Narten, 1693 "Centrally Assigned IPv6 Unicast Unique Local Address 1694 Prefixes", draft-hain-ipv6-ulac-00 (work in progress), 1695 July 2009. 1697 [I-D.ietf-autoconf-manetarch] 1698 Chakeres, I., Macker, J., and T. Clausen, "Mobile Ad hoc 1699 Network Architecture", draft-ietf-autoconf-manetarch-07 1700 (work in progress), November 2007. 1702 [I-D.ietf-csi-proxy-send] 1703 Krishnan, S., Laganier, J., and M. Bonola, "Secure Proxy 1704 ND Support for SEND", draft-ietf-csi-proxy-send-01 (work 1705 in progress), July 2009. 1707 [I-D.ietf-dhc-subnet-alloc] 1708 Johnson, R., Kumarasamy, J., Kinnear, K., and M. Stapp, 1709 "Subnet Allocation Option", draft-ietf-dhc-subnet-alloc-09 1710 (work in progress), March 2009. 1712 [I-D.ietf-manet-smf] 1713 Macker, J. and S. Team, "Simplified Multicast Forwarding", 1714 draft-ietf-manet-smf-09 (work in progress), July 2009. 1716 [I-D.ietf-v6ops-tunnel-security-concerns] 1717 Hoagland, J., Krishnan, S., and D. Thaler, "Security 1718 Concerns With IP Tunneling", 1719 draft-ietf-v6ops-tunnel-security-concerns-01 (work in 1720 progress), October 2008. 1722 [I-D.jen-apt] 1723 Jen, D., Meisel, M., Massey, D., Wang, L., Zhang, B., and 1724 L. Zhang, "APT: A Practical Transit Mapping Service", 1725 draft-jen-apt-01 (work in progress), November 2007. 1727 [I-D.russert-rangers] 1728 Russert, S., Fleischman, E., and F. Templin, "RANGER 1729 Scenarios", draft-russert-rangers-00 (work in progress), 1730 May 2009. 1732 [I-D.templin-ranger] 1733 Templin, F., "Routing and Addressing in Next-Generation 1734 EnteRprises (RANGER)", draft-templin-ranger-07 (work in 1735 progress), February 2009. 1737 [IEN48] Cerf, V., "The Catenet Model for Internetworking", 1738 July 1978. 1740 [RASADV] Microsoft, "Remote Access Server Advertisement (RASADV) 1741 Protocol Specification", October 2008. 1743 [RFC1122] Braden, R., "Requirements for Internet Hosts - 1744 Communication Layers", STD 3, RFC 1122, October 1989. 1746 [RFC1256] Deering, S., "ICMP Router Discovery Messages", RFC 1256, 1747 September 1991. 1749 [RFC1753] Chiappa, J., "IPng Technical Requirements Of the Nimrod 1750 Routing and Addressing Architecture", RFC 1753, 1751 December 1994. 1753 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 1754 E. Lear, "Address Allocation for Private Internets", 1755 BCP 5, RFC 1918, February 1996. 1757 [RFC1955] Hinden, R., "New Scheme for Internet Routing and 1758 Addressing (ENCAPS) for IPNG", RFC 1955, June 1996. 1760 [RFC2132] Alexander, S. and R. Droms, "DHCP Options and BOOTP Vendor 1761 Extensions", RFC 2132, March 1997. 1763 [RFC2501] Corson, M. and J. Macker, "Mobile Ad hoc Networking 1764 (MANET): Routing Protocol Performance Issues and 1765 Evaluation Considerations", RFC 2501, January 1999. 1767 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 1768 Domains without Explicit Tunnels", RFC 2529, March 1999. 1770 [RFC2775] Carpenter, B., "Internet Transparency", RFC 2775, 1771 February 2000. 1773 [RFC3056] Carpenter, B. and K. Moore, "Connection of IPv6 Domains 1774 via IPv4 Clouds", RFC 3056, February 2001. 1776 [RFC3704] Baker, F. and P. Savola, "Ingress Filtering for Multihomed 1777 Networks", BCP 84, RFC 3704, March 2004. 1779 [RFC3753] Manner, J. and M. Kojo, "Mobility Related Terminology", 1780 RFC 3753, June 2004. 1782 [RFC3819] Karn, P., Bormann, C., Fairhurst, G., Grossman, D., 1783 Ludwig, R., Mahdavi, J., Montenegro, G., Touch, J., and L. 1784 Wood, "Advice for Internet Subnetwork Designers", BCP 89, 1785 RFC 3819, July 2004. 1787 [RFC3927] Cheshire, S., Aboba, B., and E. Guttman, "Dynamic 1788 Configuration of IPv4 Link-Local Addresses", RFC 3927, 1789 May 2005. 1791 [RFC4192] Baker, F., Lear, E., and R. Droms, "Procedures for 1792 Renumbering an IPv6 Network without a Flag Day", RFC 4192, 1793 September 2005. 1795 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 1796 Addresses", RFC 4193, October 2005. 1798 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1799 Internet Protocol", RFC 4301, December 2005. 1801 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 1802 Network Address Translations (NATs)", RFC 4380, 1803 February 2006. 1805 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 1806 Proxies (ND Proxy)", RFC 4389, April 2006. 1808 [RFC4795] Aboba, B., Thaler, D., and L. Esibov, "Link-local 1809 Multicast Name Resolution (LLMNR)", RFC 4795, 1810 January 2007. 1812 [RFC4852] Bound, J., Pouffary, Y., Klynsma, S., Chown, T., and D. 1813 Green, "IPv6 Enterprise Network Analysis - IP Layer 3 1814 Focus", RFC 4852, April 2007. 1816 [RFC4903] Thaler, D., "Multi-Link Subnet Issues", RFC 4903, 1817 June 2007. 1819 [RFC4941] Narten, T., Draves, R., and S. Krishnan, "Privacy 1820 Extensions for Stateless Address Autoconfiguration in 1821 IPv6", RFC 4941, September 2007. 1823 Appendix A. Duplicate Address Detection (DAD) Considerations 1825 A priori uniqueness determination (also known as "pre-service DAD") 1826 for an RLOC assigned on an enterprise-interior interface would 1827 require either flooding the entire enterprise or somehow discovering 1828 a link in the enterprise on which a node that configures a duplicate 1829 address is attached and performing a localized DAD exchange on that 1830 link. But, the control message overhead for such an enterprise-wide 1831 DAD would be substantial and prone to false-negatives due to packet 1832 loss and intermittent connectivity. An alternative to pre-service 1833 DAD is to autoconfigure pseudo-random RLOCs on enterprise-interior 1834 interfaces and employ a passive in-service DAD (e.g., one that 1835 monitors routing protocol messages for duplicate assignments). 1837 Pseudo-random IPv6 RLOCs can be generated with mechanisms such as 1838 CGAs, IPv6 privacy addresses, etc. with very small probability of 1839 collision. Pseudo-random IPv4 RLOCs can be generated through random 1840 assignment from a suitably large IPv4 prefix space. 1842 Consistent operational practices can assure uniqueness for EBG- 1843 aggregated addresses/prefixes, while statistical properties for 1844 pseudo-random address self-generation can assure uniqueness for the 1845 RLOCs assigned on an ER's enterprise-interior interfaces. Still, an 1846 RLOC delegation authority should be used when available, while a 1847 passive in-service DAD mechanism should be used to detect RLOC 1848 duplications when there is no RLOC delegation authority. 1850 Appendix B. Link-Layer Multiplexing and Traffic Engineering 1852 For each distinct enterprise that it connects to, an EBR configures a 1853 VET interface over possibly multiple underlying interfaces that all 1854 connect to the same enterprise. The VET interface therefore 1855 represents the EBR's logical point of attachment to the enterprise, 1856 and provides a logical interface for link-layer multiplexing over its 1857 underlying interfaces as described in Section 3.3.4.1 of [RFC1122]: 1859 "Finally, we note another possibility that is NOT multihoming: one 1860 logical interface may be bound to multiple physical interfaces, in 1861 order to increase the reliability or throughput between directly 1862 connected machines by providing alternative physical paths between 1863 them. For instance, two systems might be connected by multiple 1864 point-to-point links. We call this "link-layer multiplexing". 1865 With link-layer multiplexing, the protocols above the link layer 1866 are unaware that multiple physical interfaces are present; the 1867 link-layer device driver is responsible for multiplexing and 1868 routing packets across the physical interfaces." 1870 EBRs can support such a link-layer multiplexing capability across the 1871 enterprise in accordance with the Weak End System Model (see Section 1872 3.3.4.2 of [RFC1122]). In particular, when an EBR autoconfigures an 1873 RLOC address (see Section 4.1), it can associate it with the VET 1874 interface only instead of assigning it to an underlying interface. 1875 The EBR therefore only needs to obtain a single RLOC address even if 1876 there are multiple underlying interfaces, i.e., it does not need to 1877 obtain one for each underlying interface. The EBR can then leave the 1878 underlying interfaces unnumbered, or it can configure a randomly 1879 chosen IP link-local address (e.g., from the prefix 169.254/16 1880 [RFC3927] for IPv4) on underlying interfaces that require a 1881 configuration. The EBR need not check these link-local addresses for 1882 uniqueness within the enterprise, as they will not normally be used 1883 as the source address for packets. 1885 When the EBR engages in the enterprise-interior routing protocol, it 1886 uses the RLOC address assigned to the VET interface as the source 1887 address for all routing protocol control messages, however it must 1888 also supply an interface identifier (e.g., a small integer) that 1889 uniquely identifies the underlying interface that the control message 1890 is sent over. For example, if the underlying interfaces are known as 1891 "eth0", "eth1" and "eth7" the EBR can supply the token "7" when it 1892 sends a routing protocol control message over the "eth7" interface. 1893 This is necessary to ensure that other routers can determine the 1894 specific interface over which the EBR's routing protocol control 1895 message was sent, but the token need only be unique within the EBR 1896 itself and need not be unique throughout the enterprise. 1898 When the EBR discovers an RLOC route via the enterprise interior 1899 routing protocol, it configures a preferred route in the IP FIB that 1900 points to the VET interface instead of the underlying interface. At 1901 the same time, the EBR also configures an ancillary route that points 1902 to the underlying interface. If the EBR discovers that the same RLOC 1903 route is reachable via multiple underlying interfaces, it configures 1904 multiple ancillary routes (i.e., one for each interface). If the EBR 1905 discovers that the RLOC route is no longer reachable via any 1906 underlying interface, it removes the route in the IP FIB that points 1907 to the VET interface. 1909 With these arrangements, all locally-generated packets with RLOC 1910 destinations will flow through the VET interface (and thereby use the 1911 VET interface's RLOC address as the source address) instead of 1912 through the underlying interfaces. In the same fashion, all 1913 forwarded packets with RLOC destinations will flow through the VET 1914 interface instead of through the underlying interfaces. 1916 This arrangement has several operational advantages that enable a 1917 number of traffic engineering capabilities. First, the VET interface 1918 inserts the SEAL header so that ID-based duplicate packet detection 1919 is enabled within the enterprise. Secondly, SEAL can dynamically 1920 adjust its packet sizing parameters so that an optimum Maximum 1921 Transmission Unit (MTU) can be determined. This is true even if the 1922 VET interface reroutes traffic between underlying interfaces with 1923 different MTUs. 1925 Most importantly, the EBR can configure default and more-specific 1926 routes on the VET interface to direct traffic through a specific 1927 egress EBR (eEBR) that may be many outer IP hops away. Encapsulation 1928 will ensure that a specific eEBR is chosen, and the best eEBR can be 1929 chosen when multiple are available. Also, local applications see a 1930 stable IP source address even if there are multiple underlying 1931 interfaces. This link-layer multiplexing can therefore provide 1932 continuous operation across failovers between multiple links attached 1933 to the same enterprise without any need for readdressing. Finally, 1934 the VET interface can forward packets with RLOC-based destinations 1935 over an underlying interface without any encapsulation if 1936 encapsulation avoidance is desired. 1938 It must be specifically noted that the above arrangement constitutes 1939 a case in which the same RLOC may be used as both the inner and outer 1940 IP source address. This will not present a problem as long as both 1941 ends configure a VET interface in the same fashion. 1943 It must also be noted that EID-based communications can use the same 1944 VET interface arrangement, except that the EID-based next hop must be 1945 mapped to an RLOC-based next-hop within the VET interface. For IPvX- 1946 in-SEAL-in-IPvX encapsulation, as well as for IPv4-in-SEAL-in-Pv6 1947 encapsulation, this requires a VET interface specific address mapping 1948 database. For IPv6-in-SEAL-in-IPv4 encapsulation, the mapping is 1949 accomplished through simple static extraction of an IPv4 address 1950 embedded in a VET address. 1952 Appendix C. Change Log 1954 (Note to RFC editor - this section to be removed before publication 1955 as an RFC.) 1957 Changes from -02 to -03: 1959 o security consideration clarifications 1961 o new PRLNAME for VET is "isatav2.example.com" 1963 o VET now uses SEAL natively 1965 o EBGs can support both legacy ISATAP and VET over the same 1966 underlying interfaces. 1968 Changes from -01 to -02: 1970 o Defined CGA and privacy address configuration on VET interfaces 1972 o Interface identifiers added to routing protocol control messages 1973 for link-layer multiplexing 1975 Changes from -00 to -01: 1977 o Section 4.1 clarifications on link-local assignment and RLOC 1978 autoconfiguration. 1980 o Appendix B clarifications on Weak End System Model 1982 Changes from RFC5558 to -00: 1984 o New appendix on RLOC configuration on VET intefaces. 1986 Author's Address 1988 Fred L. Templin (editor) 1989 Boeing Research & Technology 1990 P.O. Box 3707 MC 7L-49 1991 Seattle, WA 98124 1992 USA 1994 Email: fltemplin@acm.org