idnits 2.17.1 draft-templin-iron-04.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 8, 2010) is 5042 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) == Outdated reference: A later version (-06) exists of draft-ietf-grow-va-02 == Outdated reference: A later version (-05) exists of draft-russert-rangers-03 == Outdated reference: A later version (-68) exists of draft-templin-intarea-seal-15 == Outdated reference: A later version (-40) exists of draft-templin-intarea-vet-14 Summary: 1 error (**), 0 flaws (~~), 5 warnings (==), 1 comment (--). 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: Informational June 8, 2010 5 Expires: December 10, 2010 7 The Internet Routing Overlay Network (IRON) 8 draft-templin-iron-04.txt 10 Abstract 12 The Internet routing system is experiencing a growth profile that has 13 led many to express concerns for unsustainable routing scaling. 14 Operational practices such as increased use of multihoming with IPv4 15 Provider-Independent (PI) addressing are resulting in more and more 16 fine-grained prefixes injected into the routing system from more and 17 more end user networks. Furthermore, depletion of the remaining 18 public IPv4 address space has raised concerns for both increased 19 deaggregation (leading to yet further routing scaling) and an 20 impending address space runout scenario. At the same time, the IPv6 21 routing system is finally beginning to see significant growth in IPv6 22 Provider-Aggregated (PA) prefixes but there does not seem to be a 23 solution on the near term horizon for IPv6 PI addressing. Since the 24 Internet must continue to support escalating growth due to increasing 25 demand, it is clear that current mechanisms and operational practices 26 must be updated. This document therefore proposes an Internet 27 Routing Overlay Network (IRON) for supporting sustainable growth 28 through PI addressing while requiring no changes to end systems and 29 no changes to the existing routing system. 31 Status of this Memo 33 This Internet-Draft is submitted in full conformance with the 34 provisions of BCP 78 and BCP 79. 36 Internet-Drafts are working documents of the Internet Engineering 37 Task Force (IETF). Note that other groups may also distribute 38 working documents as Internet-Drafts. The list of current Internet- 39 Drafts is at http://datatracker.ietf.org/drafts/current/. 41 Internet-Drafts are draft documents valid for a maximum of six months 42 and may be updated, replaced, or obsoleted by other documents at any 43 time. It is inappropriate to use Internet-Drafts as reference 44 material or to cite them other than as "work in progress." 46 This Internet-Draft will expire on December 10, 2010. 48 Copyright Notice 49 Copyright (c) 2010 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (http://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 65 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 66 3. IRON Routers . . . . . . . . . . . . . . . . . . . . . . . . . 4 67 4. The Internet Routing Overlay Network (IRON) . . . . . . . . . 5 68 5. IRON Initialization . . . . . . . . . . . . . . . . . . . . . 7 69 5.1. IR(VP) and IR(GW) Initialization . . . . . . . . . . . . . 7 70 5.2. IR(EP) Initialization . . . . . . . . . . . . . . . . . . 8 71 6. IRON Operation . . . . . . . . . . . . . . . . . . . . . . . . 8 72 6.1. IR(EP) Operation . . . . . . . . . . . . . . . . . . . . . 9 73 6.2. IR(VP) Operation . . . . . . . . . . . . . . . . . . . . . 10 74 6.3. IR(GW) Operation . . . . . . . . . . . . . . . . . . . . . 10 75 6.4. IRON Reference Operating Scenario . . . . . . . . . . . . 10 76 6.5. Mobility, Multihoming and Traffic Engineering . . . . . . 12 77 6.5.1. Mobility Management . . . . . . . . . . . . . . . . . 12 78 6.5.2. Multihoming . . . . . . . . . . . . . . . . . . . . . 12 79 6.5.3. Inbound Traffic Engineering . . . . . . . . . . . . . 13 80 6.5.4. Outbound Traffic Engineering . . . . . . . . . . . . . 13 81 7. Related Initiatives . . . . . . . . . . . . . . . . . . . . . 13 82 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 83 9. Security Considerations . . . . . . . . . . . . . . . . . . . 14 84 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 14 85 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 14 86 11.1. Normative References . . . . . . . . . . . . . . . . . . . 14 87 11.2. Informative References . . . . . . . . . . . . . . . . . . 14 88 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 15 90 1. Introduction 92 The Internet routing system is experiencing a growth profile that has 93 led many to express concerns for unsustainable routing scaling. 94 Operational practices such as increased use of multihoming with IPv4 95 Provider-Independent (PI) addressing are resulting in more and more 96 fine-grained prefixes injected into the routing system from more and 97 more end user networks. Furthermore, depletion of the remaining 98 public IPv4 address space has raised concerns for both increased 99 deaggregation (leading to yet further routing scaling) and an 100 impending address space runout scenario. At the same time, the IPv6 101 routing system is finally beginning to see significant growth in IPv6 102 Provider-Aggregated (PA) prefixes but there does not seem to be a 103 solution on the near term horizon for IPv6 PI addressing. Since the 104 Internet must continue to support escalating growth due to increasing 105 demand, it is clear that current mechanisms and operational practices 106 must be updated. 108 Virtual Aggregation (VA) [I-D.ietf-grow-va] and Aggregation in 109 Increasing Scopes (AIS) [I-D.zhang-evolution] are global routing 110 proposals that introduce routing overlays with Virtual Prefixes (VPs) 111 for router Forwarding Information Base (FIB) and Routing Information 112 Base (RIB) scaling reduction. Routing and Addressing in Networks 113 with Global Enterprise Recursion (RANGER) [RFC5720] examines 114 recursive arrangements of enterprise networks that can apply to a 115 very broad set of use case scenarios [I-D.russert-rangers]. In 116 particular, RANGER supports encapsulation and secure redirection by 117 treating each layer in the recursive hierarchy as a virtual non- 118 broadcast, multiple access (NBMA) "link". RANGER is an architectural 119 framework that includes Virtual Enterprise Traversal (VET) 120 [I-D.templin-intarea-vet] and the Subnetwork Adaptation and 121 Encapsulation Layer (SEAL) [I-D.templin-intarea-seal] as its 122 functional building blocks. 124 This document proposes an Internet Routing Overlay Network (IRON) for 125 supporting sustainable growth while requiring no changes to the 126 existing routing system. IRON borrows concepts from VA, AIS and 127 RANGER, and further borrows concepts from the Internet Vastly 128 Improved Plumbing (Ivip) [I-D.whittle-ivip-arch] architecture 129 proposal. IRON specifically seeks to enable scalable Provider- 130 Independent (PI) addressing without changing the current BGP 131 [RFC4271] routing systems of the IPv4 and IPv6 Internets in any way. 133 IRON uses the IPv4 and IPv6 global Internet routing systems as 134 virtual NBMA links for tunneling inner network protocol packets that 135 use End User Network (EUN) PI Prefix (EP) source addresses within 136 outer IPv4 or IPv6 packets that use Routing LOCator (RLOC) addresses. 137 Moreover, inner packets can be either IPv4 or IPv6 without regard to 138 the address family used in the outer packet, and inner packets can 139 even be non-IP protocols such as OSI/CLNP. The following sections 140 discuss details of the IRON architecture. 142 2. Terminology 144 The following abbreviations correspond to terms used within this 145 document and elsewhere in common Internetworking nomenclature: 147 EP - End User Network PI Prefix 149 ETE - Egress Tunnel Endpoint 151 EUN - End User Network 153 ISP - Internet Service Provider 155 ITE - Ingress Tunnel Endpoint 157 MVP - Master Virtual Prefix (database) 159 NBMA - Non-Broadcast, Multiple Access 161 PA - Provider Aggregated 163 PI - Provider Independent 165 SCMP - the SEAL Control Message Protocol 167 SEAL_ID - an Identification value, randomly initialized and 168 monotonically incremented for each SEAL protocol packet 170 TE - Tunnel Endpoint (i.e., either ingress or egress) 172 VP - Virtual Prefix 174 3. IRON Routers 176 IRON introduces a new class of routers called IRON Routers (IRs) that 177 can be deployed on platforms ranging from high-end enterprise routers 178 to customer premises routers to simple commodity servers. Moreover, 179 IRs can be introduced incrementally and without affecting existing 180 infrastructure. The purpose of these new IRs is to provide waypoints 181 (or "cairns") for navigating the IRON so that packets with 182 destination addresses taken from End User Network PI prefixes (EPs) 183 can be delivered to the correct End User Networks (EUNs) through the 184 use of encapsulation with minimum path stretch for initial packets 185 and optimized routes for non-initial packets. The different 186 categories of IRs includes: 188 o IR - an IRON Router of any kind 190 o IR(VP) - a tunnel endpoint router that is owned by a VP company 191 and that aggregates VPs from which it sub-delegates more-specific 192 EPs to EUNs. 194 o IR(EP) - a tunnel endpoint router (or host with embedded gateway 195 function) that obtains an EP from a VP company, and that connects 196 an EUN to the IRON. An IR(EP) will typically be a customer 197 premises equipment (CPE) device that connects the EUN to its 198 ISP(s), but may also be a router or even a singleton host within 199 the EUN. 201 o IR(GW) - a router that acts as a gateway between the IRON and the 202 non-IRON Internet. Each VP company configures one or more IR(GWs) 203 which advertise the company's VPs into the IPv4 and/or IPv6 global 204 Internet DFZs. An IR(GW) may be configured on the same physical 205 platform as IR(VPs), or as a separate standalone platform. An 206 IR(GW) will typically be a BGP router that is capable of sourcing 207 encapsulated packets. 209 IRON observes the Internet Protocol standards [RFC0791][RFC2460]. 210 Other network layer protocols that can be encapsulated within IP 211 packets (e.g., OSI/CLNP [RFC1070], etc.) are also within scope. 213 4. The Internet Routing Overlay Network (IRON) 215 The Internet Routing Overlay Network (IRON) consists of IRON Routers 216 (IRs) that use Virtual Enterprise Traversal (VET) and the Subnetwork 217 Encapsulation and Adaptation Layer (SEAL) for the purpose of 218 forwarding encapsulated inner network layer packets over the IPv4 and 219 IPv6 Internets. Each such IR views the IPv4 and IPv6 global 220 Internets as monolithic virtual NBMA "links", and connects to the 221 links via a VET interface used for automatic tunneling. Each IR 222 therefore sees all other IRs as virtual single-hop neighbors on the 223 link from the standpoint of the inner network layer protocol, while 224 they may be separated by many physical outer IP hops. IRs are 225 deployed incrementally and without disturbing the existing Internet 226 routing system. 228 The IRON is manifested through a business model in which VP companies 229 own and manage a set of IR(VPs) that are dispersed throughout the 230 Internet and that serve a set of highly-aggregated VPs. Each VP 231 company sets up a service in which it leases EPs taken from the VPs 232 to customer EUNs. These EUNs may be located within the same network 233 as the VP company's IR(VP) routers, or they may be located elsewhere 234 within the Internet. The VP company acts as a virtual enterprise 235 network which EUNs loosely consider as their "home" network even 236 though they physically arrange for basic connectivity via one or more 237 ISP networks that may have no affiliation with the VP company. VP 238 companies can therefore open for business and begin serving their 239 customers immediately without the need to coordinate their activities 240 with ISPs or with other VP companies. 242 Each VP company also establishes a set of IR(GW) routers that connect 243 to the IPv4 and/or IPv6 Internet DFZs. The IR(GW) advertises all of 244 the VP company's IPv4 VPs into the IPv4 DFZ and advertises all of its 245 IPv6 VPs into the IPv6 DFZ. Each IR(GW) forwards any packets coming 246 from the DFZ to an IR(VP) that can encapsulate the packet and forward 247 it to the appropriate IR(EP). In this way, end systems that use PA 248 addresses can communicate with other end systems that use PI 249 addresses taken from an IRON VP. 251 EUNs establish at least one IR(EP) that connects the EUN to the IRON. 252 The IR(EP) uses encapsulation to forward packets with EP source 253 addresses to an IR(VP) belonging to its VP company as a default 254 router. The VP company's IR(VP) then forwards the packets toward 255 their final destination, and returns a SEAL Control Message Protocol 256 (SCMP) redirect message to inform the IR(EP) of a better next hop if 257 necessary. In this way, IR(EPs) experience reasonable path stretch 258 for initial packets and can discover route-optimized paths for 259 subsequent packets. 261 The IRON additionally requires a global mapping database to allow IRs 262 to map EPs/VPs to RLOCs assigned to the interfaces of other IRs. 263 Each VP in the IRON is therefore represented in a globally 264 distributed Master VP (MVP) mapping database. The MVP database is 265 maintained by a globally-managed assigned numbers authority in the 266 same manner as the Internet Assigned Numbers Authority (IANA) 267 currently maintains the master list of all top-level IPv4 and IPv6 268 delegations. The database can be replicated across multiple servers 269 for load balancing much in the same way that FTP mirror sites are 270 used to manage software distributions. Each VP in the MVP database 271 is encoded as the tuple: "{address family, prefix, prefix-length, 272 FQDN}", where: 274 o "address family" is one of IPv4, IPv6, OSI/CLNP, etc. 276 o "prefix" is the VP, e.g., 2001:DB8::/32 (IPv6), 192.2/16 (IPv4), 277 etc. 279 o "prefix-length" is the length (in bits) of the associated VP 281 o FQDN is a DNS Fully-Qualified Domain Name 283 For each VP entry in the MVP database, the VP company maintains a 284 FQDN in the DNS to map the VP to a list of IR(VP)s that serve it. 285 The FQDN is resolved by both other IR(VP)s and by IR(EP)s that hold 286 EP delegations from the VP into a list of resource records. Each 287 resource record corresponds to an individual IR(VP), and encodes the 288 tuple : "{address family, RLOC address, WGS 84 coordinates}" where 289 "address family" is the address family of the RLOC, "RLOC" is the 290 routing locator assigned to an IR(VP), and "WGS 84 coordinates" 291 identify the physical location of the IR(VP). Together, the MVP 292 database and the FQDNs in the global DNS provide sufficient mapping 293 capabilities to support navigation of the IRON. 295 5. IRON Initialization 297 IRON initialization entails the startup actions of VP company and EUN 298 equipment. The following sections discuss these startups procedures: 300 5.1. IR(VP) and IR(GW) Initialization 302 Upon startup, each IR(VP) and IR(GW) owned by the VP company 303 discovers the full set of VPs for the IRON by reading the MVP 304 database (see Section 4). These VPs may be IPv4 or IPv6, but they 305 may also be prefixes of other network layer protocols (e.g., OSI/CLNP 306 NSAP [RFC4548], etc.). Each IR(VP/GW) reads the MVP database from a 307 nearby server upon startup time, and periodically checks for deltas 308 on the server since the database was last read. Upon reading the MVP 309 database, the IR(VP/GW) resolves the FQDN corresponding to each VP 310 into an RLOC and a physical location. Each RLOC address is an IPv4 311 or IPv6 RLOC address assigned to the IR(VP) within the DFZ. 313 For each VP, the IR(VP/GW) sorts the list of RLOCs in order of 314 "geographic closeness", and inserts each "VP->RLOC" mapping into its 315 Forwarding Information Base (FIB) with a priority corresponding to 316 geographic closeness. Specifically, the FIB entries must be 317 configured such that packets with destination addresses covered by 318 the VP are forwarded to the corresponding RLOC using encapsulation of 319 the inner network layer packet in an outer IP header. Note that the 320 VP and RLOC may be of different address families; hence, possible 321 encapsulations include IPv6-in-IPv4, IPv4-in-IPv6, IPv6-in-IPv6, 322 IPv4-in-IPv4, OSI/CLNP-in-IPv6, OSI/CLNP-in-IPv4, etc. After each 323 IR(VP/GW) reads in the list of VPs and sorts the information 324 accordingly, it is said to be "synchronized with the IRON". Each 325 IR(VP) next installs all EPs derived from its VPs into its FIB based 326 on the mapping information received from each of its EUN customers. 328 5.2. IR(EP) Initialization 330 Upon startup, each IR(EP) must register its EP-to-RLOC binding with 331 the company that owns the corresponding VP, where the RLOC is an IPv4 332 or IPv6 address assigned to the IR(EP) by an ISP network. For 333 example, if an IR(EP) owns the EP 192.2.1/24 (IPv4) and the RLOC 334 assigned to the IR(EP) by the ISP is 2001:DB8::1 (IPv6), the IR(EP) 335 informs the VP company that the route 192.2.1/24 with 2001:DB8::1 as 336 the L2 address of the next-hop must be added to the FIB in each of 337 its IR(VPs) that aggregates the EP. The IR(EP) typically informs the 338 VP company by using an authenticated short transaction protocol 339 (e.g., http(s) with username/password) to register its EP-to-RLOC 340 mapping information. (The exact specification for the short 341 transaction is up to the VP company and need only be communicated to 342 the IR(EP); the IR(EP) also uses the same EP-to-RLOC registration 343 procedure to inform its VP company of a change in RLOC, e.g., due to 344 a mobility event, a change in its primary ISP, etc.). After the 345 IR(EP) registers its mapping information, the VP company then 346 propagates it to each of its IR(VPs) that aggregates the EP, e.g., 347 via a routing protocol that all of the VP company's IR(VP)s engage 348 in. 350 After the IR(EP) informs the VP company of its EP->RLOC mapping, it 351 resolves a FQDN for the VP company in order to discover the RLOC 352 addresses and geographic locations of the IR(VPs) owned by the 353 company. (This resolution closely resembles the ISATAP Potential 354 Router List (PRL) resolution procedure [RFC5214].) The IR(EP) then 355 selects the closest subset of these RLOC addresses (typically 2-4 356 routers chosen, e.g., based on geographic distance), and adds them to 357 a default router list of FIB entries that each points to a VET 358 interface with the RLOC as the L2 address of the next-hop. The 359 IR(EP) will then use these routes in the default router list as the 360 means for forwarding encapsulated packets with EID source addresses 361 toward the final destination via encapsulation. 363 6. IRON Operation 365 Following IRON initialization, IRs engage in the steady-state process 366 of receiving and forwarding packets. Except in instances when it 367 forwards an unencapsulated packet to the public Internet, the IR 368 encapsulates each forwarded packet using the mechanisms of VET 369 [I-D.templin-intarea-vet] and SEAL [I-D.templin-intarea-seal]. IRs 370 also use the SEAL Control Message Protocol (SCMP) to test liveness of 371 other IRs and to receive redirect messages informing them of a more 372 optimal route. Each IR operates as specified in the following 373 sections: 375 6.1. IR(EP) Operation 377 After an IR(EP) is initialized, it sends periodic beacons to at least 378 2-4 of its VP company's IR(VP)s which serve as default routers. Each 379 beacon is a SEAL Control Message Protocol (SCMP) Router Solicitation 380 (RS) message, and will elicit an SCMP Router Advertisement (RA) 381 message from the IR(VP). If the IR(EP) ceases to receive RA messages 382 from an IR(VP), it marks the IR(VP) as unreachable and selects a 383 different IR(VP). If the IR(EP) ceases to receive RA messages from 384 multiple IR(VPs), it marks the ISP connection as failed/failing and 385 uses an RLOC assigned by a different ISP to re-register its EP-to- 386 RLOC mapping. 388 When an end system in an EUN has a packet to send, the packet is 389 forwarded through the EUN until it reaches the IR(EP). The source 390 IR(EP) then forwards the packet either to an IR(VP) or to a 391 destination IR(EP). The source IR(EP) first checks its FIB for the 392 longest matching prefix. If the longest matching prefix is more- 393 specific than "default", the source IR(EP) forwards the packet to the 394 next-hop the same as for ordinary IP forwarding. If the longest 395 match is "default", however, the source IR(EP) forwards the packet to 396 one of the IR(VP)s serving as its default routers. 398 The source IR(EP) uses VET and SEAL to encapsulate each forwarded 399 packet in an outer IP header with the IP address of the next-hop IR 400 as the destination address. The source IR(EP) further uses SCMP to 401 test liveness and/or to accept redirect messages from the next-hop 402 IR. When the source IR(EP) receives an SCMP redirect, it checks the 403 SEAL_ID field of the encapsulated message to verify that the redirect 404 corresponds to a packet that it had previously sent to the neighbor 405 and accepts the redirect if there is a match. Thereafter, subsequent 406 packets forwarded by the source IR(EP) will follow a route-optimized 407 path. 409 An IR(EP) that accepts redirects may be redirected by an IR(VP) in 410 its home VP company network to one or more IR(VP)s in a "foreign" 411 network. In that case, the IR(EP) has no way of knowing if these 412 foreign IR(VP)s are reachable and able to process encapsulated 413 packets. In that case, the IR(EP) should select multiple foreign 414 IR(VPs) (e.g., 2-4) and send "live" packets to one of them while 415 sending corresponding "blank" packets to the others. In turn, each 416 foreign IR(VP) accepts and forwards "live" packets, but drops "blank" 417 packets after sending a redirect. In this way, even if the original 418 packet is lost due to congestion or a short-term outage, the IR(EP) 419 will receive a redirect from at least one of the foreign IR(VP)s. 421 6.2. IR(VP) Operation 423 After an IR(VP) is initialized, it sends RA responses to the periodic 424 RS beacons sent by IR(EPs) as described in Section 6.1. When the 425 IR(VP) receives an encapsulated packet from another IR, it examines 426 the inner destination address then forwards the packet as follows: 428 o If the inner destination address matches an EP in its FIB, the 429 IR(VP) 'A' re-encapsulates the packet using VET/SEAL and forwards 430 it to the next-hop IR(EP) 'B'. If the source IR 'C' is accepting 431 redirects, 'A' also sends an SCMP redirect message back to 'C'. 432 'C' will then send subsequent packets directly to 'B'. 434 o If the inner destination address does not match an EP but matches 435 a VP in its FIB, the IR(VP) 'A' re-encapsulates the packet using 436 VET/SEAL and forwards it to the next-hop IR(VP) 'B' . If the 437 source IR 'C' is accepting redirects, 'A' also sends an SCMP 438 redirect message back to 'C'. 'C' will then send subsequent 439 packets directly to 'B'. 441 o if the inner destination address does not match an EP or a VP in 442 the FIB, the IR(VP) decapsulates the packet and forwards it to the 443 public Internet via a default or more-specific route. 445 An IR(VP) that accepts redirects may need to forward initial packets 446 via the IR(VP)s of a "foreign" network. In that case, the IR(VP) can 447 send a "live" packet in parallel with corresponding "blanks" the same 448 as for an IR(EP). 450 6.3. IR(GW) Operation 452 Each VP company must establish one or more IR(GW) routers which 453 advertise the full set of the company's VP's into the IPv4 and/or 454 IPv6 Internet BGP. The VPs will be seen as ordinary routing 455 information in the BGP, and any packets originating from the non-IRON 456 IPv4 or IPv6 Internet will be forwarded into the VP company's network 457 by an IR(GW). When an IR(GW) receives a packet from the non-IRON 458 Internet but destined to an EP destination, it consults its FIB to 459 determine the best next-hop toward the final destination. The IR(GW) 460 then either forwards the packet to an IR(VP) within the home network 461 or acts as an IR(VP) itself to forward the packet further. 463 6.4. IRON Reference Operating Scenario 465 With respect to the previous sections, a path between two EUNs can 466 potentially involve both the two IR(EPs) and the IR(VP)s of the two 467 VP companies that serve the EUNs. Route optimization based on 468 redirection will allow shortcuts that eliminate the IR(VP)s from the 469 path. The following figure depicts the IRON reference operating 470 scenario for communications between two EUNs: 472 +------------+ +------------+ 473 | | | | 474 /======>+ IR(VP(A)) +======>+ IR(VP(B)) +======\ 475 // | | | | \\ 476 // +------------+ +------------+ \\ 477 // V 478 +-----+-----+ +-----+-----+ 479 | IR(EP(A)) | ........................................>| IR(EP(B)) | 480 +-----+-----+ +-----+-----+ 481 | | 482 ........ ........ 483 ( EUN A ) ( EUN B ) 484 ........ ........ 485 | | 486 +---+----+ +---+----+ 487 | Host A | | Host B | 488 +--------+ +--------+ 490 Figure 1: IRON Reference Operating Scenario 492 In this reference scenario, VP companies A and B have established 493 IR(VP)s within the Internet that serve EPs to EUNs. EUN A has 494 procured an EP from VP company A, while EUN B has procured an EP from 495 VP company B. The hosts in both EUNs have assigned addresses taken 496 from their corresponding EPs on their EUN-interior interfaces, and 497 the IR(EPs) have assigned RLOC addresses taken from their ISPs on 498 their WAN interfaces. 500 When Host A in EUN A has a packet to send to Host B in EUN B, normal 501 routing conveys the packet from Host A to IR(EP(A)). If IR(EP(A 502 ))does not have a more-specific route, it encapsulates the packet and 503 forwards it to an IR(VP) owned by VP company A. IR(VP(A )) 504 decapsulates the packet and checks its FIB for a route toward the 505 packet's destination address. If IR(VP(A)) does not have an EP route 506 to Host B in its FIB, it consults its full table of VP-to-RLOC 507 mappings to discover that the next-hop toward Host B is via 508 IR(VP(B)). IR(VP(A)) then re-encapsulates the packet and sends it to 509 IR(VP(B)) which has an EP route to Host B via IR(EP(B)). IR(VP(B)) 510 then re-encapsulates the packet and sends it to IR(EP(B)), which 511 decapsulates the packet and forwards it via EUN B to Host B. 513 In this process, when an IR(VP) re-encapsulates the packet and 514 forwards it to a next-hop IR, it also returns an SCMP redirect 515 message to the previous hop IR if the previous hop is willing to 516 accept redirects. The previous hop IR will then install a route in 517 its FIB that uses a more optimal next hop. For example, if IR(EP(A)) 518 is accepting redirects IR(VP(A)) will return a redirect message when 519 it forwards a packet to IR(VP(B)). IR(EP(A)) will then send 520 subsequent packets directly to IR(VP(B)), which will return a 521 redirect message when it forwards the packets to IR(EP(B)). Finally, 522 IR(EP(A)) will have an optimized route that lists IR(EP(B)) as the 523 next hop (shown as "....>" in the diagram). 525 Another redirection scenario arises when IR(VP(A)) is itself willing 526 to accept redirects. In that case, IR(EP(A)) may discover IR(EP(B)) 527 as a better next hop toward EUN A based solely on a redirect message 528 from IR(VP(A)) and without involving IR(VP(B)). Note however that 529 this may require IR(VP(A)) to carry thousands or even millions of EP 530 entries in its FIB for all EUNs that it has sent packets to recently, 531 which may negatively impact scalability. 533 6.5. Mobility, Multihoming and Traffic Engineering 535 While IR(VP)s can be considered as fixed infrastructure, IR(EP)s may 536 need to move between different network points of attachment, connect 537 to multiple ISPs, or explicitly manage their traffic flows. The 538 following sections discuss mobility, multihoming and traffic 539 engineering considerations for IR(EP)s: 541 6.5.1. Mobility Management 543 When an IR(EP) moves to a new topological location, it receives a new 544 RLOC address. The IR(EP) then registers the new EP-to-RLOC mapping 545 with its VP company the same as during its initialization phase as 546 described in Section 5.2. In this way, mobile networks are naturally 547 supported without the need for ancillary mechanisms. 549 Next, the IR(EP) sends Neighbor Advertisement (NA) messages to each 550 neighboring IR from which it has received packets recently. The NA 551 message includes the new RLOC as the outer source address and 552 includes the previous RLOC within an NA option field. The 553 neighboring IR will update its neighbor cache so that subsequent 554 packets will flow through the new RLOC. 556 6.5.2. Multihoming 558 An IR(EP) registers only the RLOC of its primary ISP with its VP 559 company even if it connects to multiple ISPs. If the IR(EP) later 560 needs to select a different ISP as its primary, it simply repeats the 561 EP-to-RLOC registration process the same as if it were reacting to a 562 mobility event as described above. 564 6.5.3. Inbound Traffic Engineering 566 When an IR(EP) receives packets via its primary ISP (i.e., the ISP 567 for which it is currently registered with the VP company), it may 568 wish to balance the load of incoming traffic between multiple ISP 569 connection points. In that case, the IR(EP) may send NA messages to 570 certain neighboring IRs the same as in the case of a mobility event 571 as described in Section 6.5.1. In that way, the IR(EP) can manage 572 its incoming traffic flows for best utilization of its multiple ISPs. 574 6.5.4. Outbound Traffic Engineering 576 IR(EP)s maintain a list of IR(VP)s that serve as default routers for 577 VET interface encapsulation of inner packets with source addresses 578 taken from an EP prefix. IR(EP)s also maintain a list of neighbors 579 on underlying interfaces that serve as default routers for packets 580 with non-EP source addresses. If the inner and outer protocols are 581 of different versions (e.g., OSI/CLNP as the inner version and IPv4 582 as the outer version) then the default routes of both versions are 583 mutually exclusive and no special arrangements are needed. 585 If the inner and outer protocol versions are the same, however (e.g., 586 IPv6 as both the inner and outer protocol) then a special treatment 587 of the default route is necessary. In particular, the IR(EP) must 588 examine the source address of a packet to be forwarded to determine 589 the proper handling of "default". If the packet uses a source 590 address taken from an EP prefix, the IR(EP) forwards it to an IR(VP) 591 using encapsulation via a VET interface; otherwise, the IR(EP) 592 forwards the packet to a next hop on an underlying link. 594 Using this arrangement of default, when an IR(EP) forwards a packet 595 with an EP source address it can forward it to any of its associated 596 IR(VP)s using VET interface encapsulation over any of its underlying 597 interfaces. The choice of underlying interface can be managed, and 598 the source address assigned to the underlying interface will be used 599 as the outer source address of the encapsulated packet. Each 600 encapsulated packet can therefore be directed through the desired ISP 601 using a topologically-correct outer source address. 603 7. Related Initiatives 605 IRON builds upon the concepts RANGER architecture [RFC5720], and 606 therefore inherits the same set of related initiatives. 608 Virtual Aggregation (VA) [I-D.ietf-grow-va] and Aggregation in 609 Increasing Scopes (AIS) [I-D.zhang-evolution] provide the basis for 610 the Virtual Prefix concepts. 612 Internet vastly improved plumbing (Ivip) [I-D.whittle-ivip-arch] has 613 contributed valuable insights, including the use of real-time 614 mapping. 616 8. IANA Considerations 618 The IANA is instructed to create a Master Virtual Prefix (MVP) 619 registry for IRON. 621 9. Security Considerations 623 Security considerations for RANGER apply also to IRON. 625 10. Acknowledgements 627 This ideas behind this work have benefited greatly from discussions 628 with colleagues; some of which appear on the RRG and other IRTF/IETF 629 mailing lists. 631 11. References 633 11.1. Normative References 635 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 636 September 1981. 638 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 639 (IPv6) Specification", RFC 2460, December 1998. 641 11.2. Informative References 643 [I-D.ietf-grow-va] 644 Francis, P., Xu, X., Ballani, H., Jen, D., Raszuk, R., and 645 L. Zhang, "FIB Suppression with Virtual Aggregation", 646 draft-ietf-grow-va-02 (work in progress), March 2010. 648 [I-D.russert-rangers] 649 Russert, S., Fleischman, E., and F. Templin, "Operational 650 Scenarios for IRON and RANGER", draft-russert-rangers-03 651 (work in progress), June 2010. 653 [I-D.templin-intarea-seal] 654 Templin, F., "The Subnetwork Encapsulation and Adaptation 655 Layer (SEAL)", draft-templin-intarea-seal-15 (work in 656 progress), June 2010. 658 [I-D.templin-intarea-vet] 659 Templin, F., "Virtual Enterprise Traversal (VET)", 660 draft-templin-intarea-vet-14 (work in progress), 661 June 2010. 663 [I-D.whittle-ivip-arch] 664 Whittle, R., "Ivip (Internet Vastly Improved Plumbing) 665 Architecture", draft-whittle-ivip-arch-04 (work in 666 progress), March 2010. 668 [I-D.zhang-evolution] 669 Zhang, B. and L. Zhang, "Evolution Towards Global Routing 670 Scalability", draft-zhang-evolution-02 (work in progress), 671 October 2009. 673 [RFC1070] Hagens, R., Hall, N., and M. Rose, "Use of the Internet as 674 a subnetwork for experimentation with the OSI network 675 layer", RFC 1070, February 1989. 677 [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway 678 Protocol 4 (BGP-4)", RFC 4271, January 2006. 680 [RFC4548] Gray, E., Rutemiller, J., and G. Swallow, "Internet Code 681 Point (ICP) Assignments for NSAP Addresses", RFC 4548, 682 May 2006. 684 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 685 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 686 March 2008. 688 [RFC5720] Templin, F., "Routing and Addressing in Networks with 689 Global Enterprise Recursion (RANGER)", RFC 5720, 690 February 2010. 692 Author's Address 694 Fred L. Templin (editor) 695 Boeing Research & Technology 696 P.O. Box 3707 MC 7L-49 697 Seattle, WA 98124 698 USA 700 Email: fltemplin@acm.org