idnits 2.17.1 draft-templin-aerolink-26.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 : ---------------------------------------------------------------------------- == The 'Obsoletes: ' line in the draft header should list only the _numbers_ of the RFCs which will be obsoleted by this document (if approved); it should not include the word 'RFC' in the list. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 9, 2014) is 3608 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC0768' is defined on line 1346, but no explicit reference was found in the text == Unused Reference: 'RFC0792' is defined on line 1352, but no explicit reference was found in the text == Unused Reference: 'RFC2460' is defined on line 1358, but no explicit reference was found in the text == Unused Reference: 'RFC4862' is defined on line 1382, but no explicit reference was found in the text == Unused Reference: 'RFC0879' is defined on line 1393, but no explicit reference was found in the text == Unused Reference: 'RFC6204' is defined on line 1452, but no explicit reference was found in the text == Unused Reference: 'RFC6691' is defined on line 1464, but no explicit reference was found in the text == Unused Reference: 'RFC6706' is defined on line 1467, but no explicit reference was found in the text == Unused Reference: 'RFC6980' is defined on line 1483, but no explicit reference was found in the text ** 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) ** Obsolete normative reference: RFC 6434 (Obsoleted by RFC 8504) -- Obsolete informational reference (is this intentional?): RFC 879 (Obsoleted by RFC 7805, RFC 9293) -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) -- Obsolete informational reference (is this intentional?): RFC 5996 (Obsoleted by RFC 7296) -- Obsolete informational reference (is this intentional?): RFC 6204 (Obsoleted by RFC 7084) -- Obsolete informational reference (is this intentional?): RFC 6691 (Obsoleted by RFC 9293) Summary: 4 errors (**), 0 flaws (~~), 11 warnings (==), 6 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 Obsoletes: rfc6706 (if approved) June 9, 2014 5 Intended status: Standards Track 6 Expires: December 11, 2014 8 Transmission of IPv6 Packets over AERO Links 9 draft-templin-aerolink-26.txt 11 Abstract 13 This document specifies the operation of IPv6 over tunnel virtual 14 Non-Broadcast, Multiple Access (NBMA) links using Asymmetric Extended 15 Route Optimization (AERO). Nodes attached to AERO links can exchange 16 packets via trusted intermediate routers that provide forwarding 17 services to reach off-link destinations and redirection services for 18 route optimization. AERO provides an IPv6 link-local address format 19 known as the AERO address that supports operation of the IPv6 20 Neighbor Discovery (ND) protocol and links IPv6 ND to IPv6 routing. 21 Admission control and provisioning are supported by the Dynamic Host 22 Configuration Protocol for IPv6 (DHCPv6), and node mobility is 23 naturally supported through dynamic neighbor cache updates. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on December 11, 2014. 42 Copyright Notice 44 Copyright (c) 2014 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 60 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 61 3. Asymmetric Extended Route Optimization (AERO) . . . . . . . . 5 62 3.1. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 5 63 3.2. AERO Addresses . . . . . . . . . . . . . . . . . . . . . 6 64 3.3. AERO Interface Characteristics . . . . . . . . . . . . . 6 65 3.3.1. Coordination of Multiple Underlying Interfaces . . . 8 66 3.4. AERO Interface Neighbor Cache Maintenace . . . . . . . . 8 67 3.5. AERO Interface Data Origin Authentication . . . . . . . . 10 68 3.6. AERO Interface MTU Considerations . . . . . . . . . . . . 10 69 3.7. AERO Interface Encapsulation, Re-encapsulation and 70 Decapsulation . . . . . . . . . . . . . . . . . . . . . . 12 71 3.8. AERO Router Discovery, Prefix Delegation and Address 72 Configuration . . . . . . . . . . . . . . . . . . . . . . 13 73 3.8.1. AERO Client Behavior . . . . . . . . . . . . . . . . 13 74 3.8.2. AERO Server Behavior . . . . . . . . . . . . . . . . 15 75 3.9. AERO Redirection . . . . . . . . . . . . . . . . . . . . 16 76 3.9.1. Reference Operational Scenario . . . . . . . . . . . 16 77 3.9.2. Classical Redirection Approaches . . . . . . . . . . 18 78 3.9.3. Concept of Operations . . . . . . . . . . . . . . . . 19 79 3.9.4. Message Format . . . . . . . . . . . . . . . . . . . 19 80 3.9.5. Sending Predirects . . . . . . . . . . . . . . . . . 20 81 3.9.6. Processing Predirects and Sending Redirects . . . . . 21 82 3.9.7. Re-encapsulating and Relaying Redirects . . . . . . . 23 83 3.9.8. Processing Redirects . . . . . . . . . . . . . . . . 23 84 3.9.9. Server-Oriented Redirection . . . . . . . . . . . . . 24 85 3.10. Neighbor Reachability Maintenance . . . . . . . . . . . . 24 86 3.11. Mobility Management . . . . . . . . . . . . . . . . . . . 25 87 3.12. Encapsulation Protocol Version Considerations . . . . . . 27 88 3.13. Multicast Considerations . . . . . . . . . . . . . . . . 27 89 3.14. Operation on AERO Links Without DHCPv6 Services . . . . . 27 90 3.15. Operation on Server-less AERO Links . . . . . . . . . . . 27 91 3.16. Other Considerations . . . . . . . . . . . . . . . . . . 28 92 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 28 93 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 28 94 6. Security Considerations . . . . . . . . . . . . . . . . . . . 28 95 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 29 96 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 29 97 8.1. Normative References . . . . . . . . . . . . . . . . . . 29 98 8.2. Informative References . . . . . . . . . . . . . . . . . 30 99 Appendix A. AERO Server and Relay Interworking . . . . . . . . . 32 100 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 34 102 1. Introduction 104 This document specifies the operation of IPv6 over tunnel virtual 105 Non-Broadcast, Multiple Access (NBMA) links using Asymmetric Extended 106 Route Optimization (AERO). The AERO link can be used for tunneling 107 to neighboring nodes on either IPv6 or IPv4 networks, i.e., AERO 108 views the IPv6 and IPv4 networks as equivalent links for tunneling. 109 Nodes attached to AERO links can exchange packets via trusted 110 intermediate routers that provide forwarding services to reach off- 111 link destinations and redirection services for route optimization 112 that addresses the requirements outlined in [RFC5522]. 114 AERO proivdes an IPv6 link-local address format known as the AERO 115 address that supports operation of the IPv6 Neighbor Discovery (ND) 116 [RFC4861] protocol and links IPv6 ND to IPv6 routing. Admission 117 control and provisioning are supported by the Dynamic Host 118 Configuration Protocol for IPv6 (DHCPv6) [RFC3315], and node mobility 119 is naturally supported through dynamic neighbor cache updates. The 120 remainder of this document presents the AERO specification. 122 2. Terminology 124 The terminology in the normative references applies; the following 125 terms are defined within the scope of this document: 127 AERO link 128 a Non-Broadcast, Multiple Access (NBMA) tunnel virtual overlay 129 configured over a node's attached IPv6 and/or IPv4 networks. All 130 nodes on the AERO link appear as single-hop neighbors from the 131 perspective of IPv6. 133 AERO interface 134 a node's attachment to an AERO link. 136 AERO address 137 an IPv6 link-local address constructed as specified in Section 3.2 138 and assigned to a Client's AERO interface. 140 AERO node 141 a node that is connected to an AERO link and that participates in 142 IPv6 Neighbor Discovery over the link. 144 AERO Client ("Client") 145 a node that assigns an AERO address on an AERO interface and 146 receives an IPv6 prefix delegation. 148 AERO Server ("Server") 149 a node that configures a router interface on an AERO link over 150 which it can provide default forwarding and redirection services 151 for AERO Clients. 153 AERO Relay ("Relay") 154 a node that relays IPv6 packets between Servers on the same AERO 155 link, and/or that forwards IPv6 packets between the AERO link and 156 the IPv6 Internet. An AERO Relay may or may not also be 157 configured as an AERO Server. 159 ingress tunnel endpoint (ITE) 160 an AERO interface endpoint that injects tunneled packets into an 161 AERO link. 163 egress tunnel endpoint (ETE) 164 an AERO interface endpoint that receives tunneled packets from an 165 AERO link. 167 underlying network 168 a connected IPv6 or IPv4 network routing region over which AERO 169 nodes tunnel IPv6 packets. 171 underlying interface 172 an AERO node's interface point of attachment to an underlying 173 network. 175 link-layer address 176 an IP address assigned to an AERO node's underlying interface. 177 When UDP encapsulation is used, the UDP port number is also 178 considered as part of the link-layer address. Link-layer 179 addresses are used as the encapsulation header source and 180 destination addresses. 182 network layer address 183 the source or destination address of the encapsulated IPv6 packet. 185 end user network (EUN) 186 an internal virtual or external edge IPv6 network that an AERO 187 Client connects to the AERO interface. 189 Throughout the document, the simple terms "Client", "Server" and 190 "Relay" refer to "AERO Client", "AERO Server" and "AERO Relay", 191 respectively. Capitalization is used to distinguish these terms from 192 DHCPv6 client/server/relay. This is an important distinction, since 193 an AERO Server may be a DHCPv6 relay, and an AERO Relay may be a 194 DHCPv6 server. 196 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 197 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 198 document are to be interpreted as described in [RFC2119]. 200 3. Asymmetric Extended Route Optimization (AERO) 202 The following sections specify the operation of IPv6 over Asymmetric 203 Extended Route Optimization (AERO) links: 205 3.1. AERO Node Types 207 AERO Relays relay packets between nodes connected to the same AERO 208 link and also forward packets between the AERO link and the native 209 IPv6 network. The relaying process entails re-encapsulation of IPv6 210 packets that were received from a first AERO node and are to be 211 forwarded without modification to a second AERO node. 213 AERO Servers configure their AERO interfaces as router interfaces, 214 and provide default routing services to AERO Clients. AERO Servers 215 configure a DHCPv6 relay or server function and facilitate DHCPv6 216 Prefix Delegation (PD) exchanges. An AERO Server may also act as an 217 AERO Relay. 219 AERO Clients act as requesting routers to receive IPv6 prefixes 220 through a DHCPv6 PD exchange via AERO Servers over the AERO link. 221 (Each client MAY associate with multiple Servers, but associating 222 with many Servers may result in excessive control message overhead.) 223 Each AERO Client receives at least a /64 prefix delegation, and may 224 receive even shorter prefixes. 226 AERO Clients that act as routers configure their AERO interfaces as 227 router interfaces and sub-delegate portions of their received prefix 228 delegations to links on EUNs. End system applications on AERO 229 Clients that act as routers bind to EUN interfaces (i.e., and not the 230 AERO interface). 232 AERO Clients that act as ordinary hosts configure their AERO 233 interfaces as host interfaces and assign one or more IPv6 addresses 234 taken from their received prefix delegations to the AERO interface 235 but DO NOT assign the delegated prefix itself to the AERO interface. 236 Instead, the host assigns the delegated prefix to a "black hole" 237 route so that unused portions of the prefix are nullified. End 238 system applications on AERO Clients that act as hosts bind directly 239 to the AERO interface. 241 3.2. AERO Addresses 243 An AERO address is an IPv6 link-local address with an embedded IPv6 244 prefix and assigned to a Client's AERO interface. The AERO address 245 is formatted as follows: 247 fe80::[IPv6 prefix] 249 The AERO address begins with the prefix fe80::/64 and includes in its 250 interface identifier the base /64 prefix taken from the Client's 251 delegated IPv6 prefix. The base prefix is determined by masking the 252 delegated prefix with the prefix length. For example, if the AERO 253 Client receives the prefix delegation: 255 2001:db8:1000:2000::/56 257 it constructs its AERO address as: 259 fe80::2001:db8:1000:2000 261 The AERO address remains stable as the Client moves between 262 topological locations, i.e., even if its underlying addresses change. 264 3.3. AERO Interface Characteristics 266 AERO interfaces use IPv6-in-IPv6 encapsulation [RFC2473] to exchange 267 tunneled packets with AERO neighbors attached to an underlying IPv6 268 network, and use IPv6-in-IPv4 encapsulation [RFC4213] to exchange 269 tunneled packets with AERO neighbors attached to an underlying IPv4 270 network. AERO interfaces can also operate over secured tunnel types 271 such as IPsec [RFC4301] or TLS [RFC5246]. When Network Address 272 Translator (NAT) traversal and/or filtering middlebox traversal may 273 be necessary, a UDP header is further inserted immediately above the 274 IP encapsulation header. 276 Servers assign the address fe80:: to their AERO interfaces as a link- 277 local Subnet Router Anycast address. Servers and Relays also assign 278 a link-local address fe80::ID to support the operation of the IPv6 279 Neighbor Discovery protocol and the inter-Server/Relay routing system 280 (see: Appendix A). Each address fe80::ID MUST be unique, and MUST 281 NOT collide with any potential AERO addresses (e.g., fe80::1, 282 fe80::2, fe80::3, etc). 284 When a Client enables an AERO interface, it invokes DHCPv6 PD using 285 the temporary IPv6 link-local source address 286 fe80::ffff:ffff:ffff:ffff. After the Client receives a prefix 287 delegation, it assigns the corresponding AERO address to the AERO 288 interface and deprecates the temporary address, i.e., the Client 289 invokes DHCPv6 to bootstrap the provisioning of a unique link-local 290 address before invoking IPv6 ND. 292 AERO interfaces maintain a neighbor cache and use an adaptation of 293 standard unicast IPv6 ND messaging. AERO interfaces use unicast 294 Neighbor Solicitation (NS), Neighbor Advertisement (NA), Router 295 Solicitation (RS) and Router Advertisement (RA) messages the same as 296 for any IPv6 link. AERO interfaces use two redirection message types 297 -- the first being the standard Redirect message and the second known 298 as a Predirect message (see Section 3.9). AERO links further use 299 link-local-only addressing; hence, Clients ignore any Prefix 300 Information Options (PIOs) they may receive in RA messages. 302 AERO interface Redirect/Predirect messages include Target Link-Layer 303 Address Options (TLLAOs) formatted as shown in Figure 1: 305 0 1 2 3 306 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 307 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 308 | Type = 2 | Length = 3 | Reserved | 309 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 310 | Link ID | Preference | UDP Port Number (or 0) | 311 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 312 | | 313 +-- --+ 314 | | 315 +-- IP Address --+ 316 | | 317 +-- --+ 318 | | 319 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 321 Figure 1: AERO Target Link-Layer Address Option (TLLAO) Format 323 In this format, Link ID is an integer value between 0 and 255 324 corresponding to an underlying interface of the target node, and 325 Preference is an integer value between 0 and 255 indicating the 326 node's preference for this underlying interface, with 0 being highest 327 preference and 255 being lowest. UDP Port Number and IP Address are 328 set to the addresses used by the target node when it sends 329 encapsulated packets over the underlying interface. When no UDP 330 encapsulation is used, UDP Port Number is set to 0. When the 331 encapsulation IP address family is IPv4, IP Address is formed as an 332 IPv4-compatible IPv6 address [RFC4291]. 334 AERO interface Redirect/Predirect messages can both update and create 335 neighbor cache entries, including link-layer address information. 337 Redirect/Predirect messages SHOULD include a Timestamp option (see 338 Section 5.3 of [RFC3971]) that other AERO nodes can use to verify the 339 message time of origin. 341 AERO interface NS/NA/RS/RA messages update timers in existing 342 neighbor cache entires but do not update link-layer addresses nor 343 create new neighbor cache entries. NS/RS messages SHOULD include a 344 Nonce option (see Section 5.3 of [RFC3971]) that the recipient echoes 345 back in the corresponding NA/RA response. Unsolicited NA/RA messages 346 are not used on AERO interfaces, and SHOULD be ignored on receipt. 348 3.3.1. Coordination of Multiple Underlying Interfaces 350 AERO interfaces may be configured over multiple underlying 351 interfaces. For example, common handheld devices of the current era 352 have both wireless local area network (aka "WiFi") and cellular 353 wireless links. These links are typically used "one at a time" with 354 low-cost WiFi preferred and highly-available cellular wireless as a 355 standby. In a more complex example, aircraft frequently have many 356 wireless data link types (e.g. satellite-based, terrestrial, air-to- 357 air directional, etc.) with diverse performance and cost properties. 359 If a Client's multiple underlying interfaces are used "one at a time" 360 (i.e., all other interfaces are in standby mode while one interface 361 is active), then Predirect/Redirect messages include only a single 362 TLLAO with Link ID set to 0. 364 If the Client has multiple active underlying interfaces, then from 365 the perspective of IPv6 ND it would appear to have a single link- 366 local address with multiple link-layer addresses. In that case, 367 Predirect/Redirect messages MAY include multiple TLLAOs -- each with 368 a different Link ID that corresponds to an underlying interface of 369 the Client. Further details on coordination of multiple active 370 underlying interfaces are outside the scope of this specification. 372 3.4. AERO Interface Neighbor Cache Maintenace 374 Each AERO interface maintains a conceptual neighbor cache that 375 includes an entry for each neighbor it communicates with on the AERO 376 link, the same as for any IPv6 interface [RFC4861]. Neighbor cache 377 entries are created and maintained as follows: 379 When an AERO Server relays a DHCPv6 Reply message to an AERO Client, 380 it creates or updates a neighbor cache entry for the Client based on 381 the AERO address corresponding to the prefix in the IA_PD option as 382 the Client's network-layer address and with the Client's 383 encapsulation IP address and UDP port number as the link-layer 384 address. 386 When an AERO Client receives a DHCPv6 Reply message from an AERO 387 Server, it creates or updates a neighbor cache entry for the Server 388 based on the Reply message link-local source address as the network- 389 layer address, and the encapsulation IP source address and UDP source 390 port number as the link-layer address. 392 When an AERO Client receives a valid Predirect message it creates or 393 updates a neighbor cache entry for the Predirect target network-layer 394 and link-layer addresses, and also creates an IPv6 forwarding table 395 entry for the predirected (source) prefix. The node then sets an 396 "ACCEPT" timer for the neighbor and uses this timer to determine 397 whether messages received from the predirected neighbor can be 398 accepted. 400 When an AERO Client receives a valid Redirect message it creates or 401 updates a neighbor cache entry for the Redirect target network-layer 402 and link-layer addresses, and also creates an IPv6 forwarding table 403 entry for the redirected (destination) prefix. The node then sets a 404 "FORWARD" timer for the neighbor and uses this timer to determine 405 whether packets can be sent directly to the redirected neighbor. The 406 node also maintains a constant value MAX_RETRY to limit the number of 407 keepalives sent when a neighbor may have gone unreachable. 409 When an AERO Client receives a valid NS message it (re)sets the 410 ACCEPT timer for the neighbor to ACCEPT_TIME. 412 When an AERO Client receives a valid NA message, it (re)sets the 413 FORWARD timer for the neighbor to FORWARD_TIME. 415 It is RECOMMENDED that FORWARD_TIME be set to the default constant 416 value 30 seconds to match the default REACHABLE_TIME value specified 417 for IPv6 neighbor discovery [RFC4861]. 419 It is RECOMMENDED that ACCEPT_TIME be set to the default constant 420 value 40 seconds to allow a 10 second window so that the AERO 421 redirection procedure can converge before the ACCEPT timer decrements 422 below FORWARD_TIME. 424 It is RECOMMENDED that MAX_RETRY be set to 3 the same as described 425 for IPv6 neighbor discovery address resolution in Section 7.3.3 of 426 [RFC4861]. 428 Different values for FORWARD_TIME, ACCEPT_TIME, and MAX_RETRY MAY be 429 administratively set, if necessary, to better match the AERO link's 430 performance characteristics; however, if different values are chosen, 431 all nodes on the link MUST consistently configure the same values. 432 In particular, ACCEPT_TIME SHOULD be set to a value that is 433 sufficiently longer than FORWARD_TIME to allow the AERO redirection 434 procedure to converge. 436 3.5. AERO Interface Data Origin Authentication 438 AERO nodes use a simple data origin authentication for encapsulated 439 packets they receive from other nodes. In particular, AERO nodes 440 accept encapsulated packets with a link-layer source address 441 belonging to one of their current AERO Servers and accept 442 encapsulated packets with a link-layer source address that is correct 443 for the network-layer source address. 445 The AERO node considers the link-layer source address correct for the 446 network-layer source address if there is an IPv6 forwarding table 447 entry that matches the network-layer source address as well as a 448 neighbor cache entry corresponding to the next hop that includes the 449 link-layer address and the ACCEPT timer is non-zero. 451 Note that this simple data origin authentication only applies to 452 environments in which link-layer addresses cannot be spoofed. 453 Additional security mitigations may be necessary in other 454 environments. 456 3.6. AERO Interface MTU Considerations 458 The AERO link Maximum Transmission Unit (MTU) is 64KB minus the 459 encapsulation overhead for IPv4 [RFC0791] and 4GB minus the 460 encapsulation overhead for IPv6 [RFC2675]. This is the most that 461 IPv4 and IPv6 (respectively) can convey within the constraints of 462 protocol constants, but actual sizes available for tunneling will 463 frequently be much smaller. 465 The base tunneling specifications for IPv4 and IPv6 typically set a 466 static MTU on the tunnel interface to 1500 bytes minus the 467 encapsulation overhead or smaller still if the tunnel is likely to 468 incur additional encapsulations on the path. This can result in path 469 MTU related black holes when packets that are too large to be 470 accommodated over the AERO link are dropped, but the resulting ICMP 471 Packet Too Big (PTB) messages are lost on the return path. As a 472 result, AERO nodes use the following MTU mitigations to accommodate 473 larger packets. 475 AERO nodes set their AERO interface MTU to the larger of the 476 underlying interface MTU minus the encapsulation overhead, and 1500 477 bytes. (If there are multiple underlying interfaces, the node sets 478 the AERO interface MTU according to the largest underlying interface 479 MTU, or 64KB /4G minus the encapsulation overhead if the largest MTU 480 cannot be determined.) AERO nodes optionally cache other per- 481 neighbor MTU values in the underlying IP path MTU discovery cache 482 initialized to the underlying interface MTU. 484 AERO nodes admit packets that are no larger than 1280 bytes minus the 485 encapsulation overhead (*) as well as packets that are larger than 486 1500 bytes into the tunnel without fragmentation, i.e., as long as 487 they are no larger than the AERO interface MTU before encapsulation 488 and also no larger than the cached per-neighbor MTU following 489 encapsulation. For IPv4, the node sets the "Don't Fragment" (DF) bit 490 to 0 for packets no larger than 1280 bytes minus the encapsulation 491 overhead (*) and sets the DF bit to 1 for packets larger than 1500 492 bytes. If a large packet is lost in the path, the node may 493 optionally cache the MTU reported in the resulting PTB message or may 494 ignore the message, e.g., if there is a possibility that the message 495 is spurious. 497 For packets destined to an AERO node that are larger than 1280 bytes 498 minus the encapsulation overhead (*) but no larger than 1500 bytes, 499 the node uses IP fragmentation to fragment the encapsulated packet 500 into two pieces (where the first fragment contains 1024 bytes of the 501 original IPv6 packet) then admits the fragments into the tunnel. If 502 the encapsulation protocol is IPv4, the node admits each fragment 503 into the tunnel with DF set to 0 and subject to rate limiting to 504 avoid reassembly errors [RFC4963][RFC6864]. For both IPv4 and IPv6, 505 the node also sends a 1500 byte probe message (**) to the neighbor, 506 subject to rate limiting. 508 To construct a probe, the node prepares an NS message with a Nonce 509 option plus trailing padding octets added to a length of 1500 bytes 510 without including the length of the padding in the IPv6 Payload 511 Length field. The node then encapsulates the NS in the encapsulation 512 headers (while including the length of the padding in the 513 encapsulation header length fields), sets DF to 1 (for IPv4) and 514 sends the padded NS message to the neighbor. If the neighbor returns 515 an NA message with a correct Nonce value, the node may then send 516 whole packets within this size range and (for IPv4) relax the rate 517 limiting requirement. (Note that the trailing padding SHOULD NOT be 518 included within the Nonce option itself but rather as padding beyond 519 the last option in the NS message; otherwise, the (large) Nonce 520 option would be echoed back in the solicited NA message and may be 521 lost at a link with a small MTU along the reverse path.) 523 AERO nodes MUST be capable of reassembling packets up to 1500 bytes 524 plus the encapsulation overhead length. It is therefore RECOMMENDED 525 that AERO nodes be capable of reassembling at least 2KB. 527 (*) Note that if it is known without probing that the minimum Path 528 MTU to an AERO node is MINMTU bytes (where 1280 < MINMTU < 1500) then 529 MINMTU can be used instead of 1280 in the fragmentation threshold 530 considerations listed above. 532 (**) It is RECOMMENDED that no probes smaller than 1500 bytes be used 533 for MTU probing purposes, since smaller probes may be fragmented if 534 there is a nested tunnel somewhere on the path to the neighbor. 535 Probe sizes larger than 1500 bytes MAY be used, but may be 536 unnecessary since original sources are expected to implement 537 [RFC4821] when sending large packets. 539 3.7. AERO Interface Encapsulation, Re-encapsulation and Decapsulation 541 AERO interfaces encapsulate IPv6 packets according to whether they 542 are entering the AERO interface for the first time or if they are 543 being forwarded out the same AERO interface that they arrived on. 544 This latter form of encapsulation is known as "re-encapsulation". 546 AERO interfaces encapsulate packets per the specifications in 547 [RFC2473][RFC4213][RFC4301][RFC5246] except that the interface copies 548 the "Hop Limit", "Traffic Class" and "Congestion Experienced" values 549 in the packet's IPv6 header into the corresponding fields in the 550 encapsulation header. For packets undergoing re-encapsulation, the 551 AERO interface instead copies the "TTL/Hop Limit", "Type of Service/ 552 Traffic Class" and "Congestion Experienced" values in the original 553 encapsulation header into the corresponding fields in the new 554 encapsulation header (i.e., the values are transferred between 555 encapsulation headers and *not* copied from the encapsulated packet's 556 network-layer header). 558 When AERO UDP encapsulation is used, the AERO interface encapsulates 559 the packet per the specifications in [RFC2473][RFC4213] except that 560 it inserts a UDP header between the encapsulation header and IPv6 561 packet header. The AERO interface sets the UDP source port to a 562 constant value that it will use in each successive packet it sends, 563 sets the UDP checksum field to zero (see: [RFC6935][RFC6936]) and 564 sets the UDP length field to the length of the IPv6 packet plus 8 565 bytes for the UDP header itself. For packets sent via a Server, the 566 AERO interface sets the UDP destination port to 8060 (i.e., the IANA- 567 registered port number for AERO) when AERO-only encapsulation is 568 used. For packets sent to a neighboring Client, the AERO interface 569 sets the UDP destination port to the port value stored in the 570 neighbor cache entry for this neighbor. 572 The AERO interface next sets the IP protocol number in the 573 encapsulation header to the appropriate value for the first protocol 574 layer within the encapsulation (e.g., IPv6, UDP, IPsec, etc.). When 575 IPv6 is used as the encapsulation protocol, the interface then sets 576 the flow label value in the encapsulation header the same as 577 described in [RFC6438]. When IPv4 is used as the encapsulation 578 protocol, the AERO interface sets the DF bit as discussed in 579 Section 3.6. 581 AERO interfaces decapsulate packets destined either to the node 582 itself or to a destination reached via an interface other than the 583 receiving AERO interface. When AERO UDP encapsulation is used (i.e., 584 when a UDP header with destination port 8060 is present) the 585 interface examines the first octet of the encapsulated packet. If 586 the most significant four bits of the first octet encode the value 587 '0110' (i.e., the version number value for IPv6), the packet is 588 accepted and the encapsulating UDP header is discarded; otherwise, 589 the packet is discarded. 591 Further decapsulation then proceeds according to the appropriate 592 tunnel type [RFC2473][RFC4213][RFC4301][RFC5246]. 594 3.8. AERO Router Discovery, Prefix Delegation and Address Configuration 596 3.8.1. AERO Client Behavior 598 AERO Clients observe the IPv6 node requirements defined in [RFC6434]. 599 AERO Clients first discover the link-layer addresses of AERO Servers 600 via static configuration, or through an automated means such as DNS 601 name resolution. In the absence of other information, the Client 602 resolves the Fully-Qualified Domain Name (FQDN) 603 "linkupnetworks.domainname", where "domainname" is the DNS domain 604 appropriate for the Client's attached underlying network. After 605 discovering the link-layer addresses, the Client associates with one 606 or more of the corresponding Servers. 608 To associate with a Server, the Client acts as a requesting router to 609 request an IPv6 prefix through DHCPv6 PD [RFC3315][RFC3633][RFC6355] 610 using fe80::ffff:ffff:ffff:ffff as the IPv6 source address (see 611 Section 3.3), 'All_DHCP_Relay_Agents_and_Servers' as the IPv6 612 destination address and the link-layer address of the Server as the 613 link-layer destination address. The Client includes a DHCPv6 Unique 614 Identifier (DUID) in the Client Identifier option of its DHCPv6 615 messages (as well as a DHCPv6 authentication option if necessary) to 616 identify itself to the DHCPv6 server. If the Client is pre- 617 provisioned with an IPv6 prefix associated with the AERO service, it 618 MAY also include the prefix in an IA_PD option in its DHCPv6 Request 619 to indicate its preferred prefix to the DHCPv6 server. The Client 620 then sends the encapsulated DHCPv6 request via an underlying 621 interface. 623 When the Client receives its prefix delegation via a Reply from the 624 DHCPv6 server, it creates a neighbor cache entry with the Server's 625 link-local address (i.e., fe80::ID) as the network-layer address and 626 the Server's encapsulation address as the link-layer addresses. 627 Next, the Client assigns the AERO address derived from the delegated 628 prefix to the AERO interface and sub-delegates the prefix to nodes 629 and links within its attached EUNs (the AERO address thereafter 630 remains stable as the Client moves). The Client also sets both the 631 ACCEPT and FORWARD timers for each Server to infinity, since the 632 Client will remain with this Server unless it explicitly terminates 633 the association. The Client further renews its prefix delegation by 634 performing DHCPv6 Renew/Reply exchanges with its AERO address as the 635 IPv6 source address, 'All_DHCP_Relay_Agents_and_Servers' as the IPv6 636 destination address, the link-layer address of a Server as the link- 637 layer destination address and the same DUID and authentication 638 information. If the Client wishes to associate with multiple 639 Servers, it can perform DHCPv6 Renew/Reply exchanges via each of the 640 Servers. 642 The Client then sends an RS message to each of its associated Servers 643 to receive an RA message with a default router lifetime and any other 644 link-specific parameters. When the Client receives an RA message, it 645 configures or updates a default route according to the default router 646 lifetime but ignores any Prefix Information Options (PIOs) included 647 in the RA message since the AERO link is link-local-only. The Client 648 further ignores any RS messages it might receive, since only Servers 649 may process RS messages. 651 The Client then sends periodic RS messages to each Server (subject to 652 rate limiting) to obtain new RA messages for Neighbor Unreachability 653 Detection (NUD), to refresh any network state, and to update the 654 default router lifetime and any other link-specific parameters. The 655 Client can also forward IPv6 packets destined to networks beyond its 656 local EUNs via a Server as an IPv6 default router. The Server may in 657 turn return a redirection message informing the Client of a neighbor 658 on the AERO link that is topologically closer to the final 659 destination (see Section 3.9). 661 Note that, since the Client's AERO address is configured from the 662 unique DHCPv6 prefix delegation it receives, there is no need for 663 Duplicate Address Detection (DAD) on AERO links. Other nodes 664 maliciously attempting to hijack an authorized Client's AERO address 665 will be denied access to the network by the DHCPv6 server due to an 666 unacceptable link-layer address and/or security parameters (see: 667 Security Considerations). 669 3.8.2. AERO Server Behavior 671 AERO Servers observe the IPv6 router requirements defined in 672 [RFC6434] and further configure a DHCPv6 relay function on their AERO 673 links. AERO Servers arrange to add their (link-layer) IP Addresses 674 to the DNS resource records for the FQDN "linkupnetworks.domainname" 675 before entering service. 677 When an AERO Server relays a prospective Client's DHCPv6 PD messages 678 to the DHCPv6 server, it wraps each message in a "Relay-forward" 679 message per [RFC3315] and includes a DHCPv6 Interface Identifier 680 option that encodes a value that identifies the AERO link to the 681 DHCPv6 server. Without creating internal state, the Server then 682 includes the Client's link-layer address in a DHCPv6 Client Link 683 Layer Address Option (CLLAO) [RFC6939] with the link-layer address 684 format shown in Figure 1 (i.e., Link ID followed by Preference 685 followed by UDP Port Number followed by IP Address). The Server sets 686 the CLLAO 'option-length' field to 22 (2 plus the length of the link- 687 layer address) and sets the 'link-layer type' field to TBD (see: IANA 688 Considerations). The Server finally includes a DHCPv6 Echo Request 689 Option (ERO) [RFC4994] that encodes the option code for the CLLAO in 690 a 'requested-option-code-n' field, then relays the message to the 691 DHCPv6 server. The CLLAO information will therefore subsequently be 692 echoed back in the DHCPv6 server's "Relay-reply" message. 694 When the DHCPv6 server issues the IPv6 prefix delegation in a "Relay- 695 reply" message via the AERO Server (acting as a DHCPv6 relay), the 696 Server obtains the Client's link-layer address from the echoed CLLAO 697 option and obtains the Client's delegated prefix from the included 698 IA_PD option. The Server then creates a neighbor cache entry for the 699 Client's AERO address with the Client's link-layer address as the 700 link-layer address for the neighbor cache entry. The neighbor cache 701 entry is created with both ACCEPT and FORWARD timers set to infinity, 702 since the Client will remain with this Server unless it explicitly 703 terminates the association (however, the Server SHOULD set a finite 704 expiration timer for the neighbor cache entry itself in case the 705 Client goes unreachable for an extended period of time). 707 The Server also configures an IPv6 forwarding table entry that lists 708 the Client's AERO address as the next hop toward the delegated IPv6 709 prefix with a lifetime derived from the DHCPv6 lease lifetime. The 710 Server finally injects the Client's prefix as an IPv6 route into the 711 inter-Server/Relay routing system (see: Appendix A) then relays the 712 DHCPv6 message to the Client while using fe80::ID as the IPv6 source 713 address, the link-local address found in the "peer address" field of 714 the Relay-reply message as the IPv6 destination address, and the 715 Client's link-layer address as the destination link-layer address. 717 Servers respond to NS/RS messages from Clients on their AERO 718 interfaces by returning an NA/RA message. The Server SHOULD NOT 719 include PIOs in the RA messages it sends to Clients, since the Client 720 will ignore any such options. 722 Servers ignore any RA messages they may receive from a Client. 723 Servers MAY examine RA messages received from other Servers for 724 consistency verification purposes. 726 When the Server forwards a packet via the same AERO interface on 727 which it arrived, it initiates an AERO route optimization procedure 728 as specified in Section 3.9. 730 3.9. AERO Redirection 732 3.9.1. Reference Operational Scenario 734 Figure 2 depicts the AERO redirection reference operational scenario. 735 The figure shows an AERO Server('A'), two AERO Clients ('B', 'C') and 736 three ordinary IPv6 hosts ('D', 'E', 'F'): 738 .-(::::::::) 739 .-(::: IPv6 :::)-. +-------------+ 740 (:::: Internet ::::)--| Host F | 741 `-(::::::::::::)-' +-------------+ 742 `-(::::::)-' 2001:db8:2::1 743 | 744 +--------------+ 745 | AERO Server A| 746 | (D->B; E->C) | 747 +--------------+ 748 fe80::ID 749 L2(A) 750 | 751 X-----+-----------+-----------+--------X 752 | AERO Link | 753 L2(B) L2(C) 754 fe80::2001:db8:0:0 fe80::2001:db8:1:0 .-. 755 +--------------+ +--------------+ ,-( _)-. 756 | AERO Client B| | AERO Client C| .-(_ IPv6 )-. 757 | (default->A) | | (default->A) |--(__ EUN ) 758 +--------------+ +--------------+ `-(______)-' 759 2001:DB8:0::/48 2001:DB8:1::/48 | 760 | 2001:db8:1::1 761 .-. +-------------+ 762 ,-( _)-. 2001:db8:0::1 | Host E | 763 .-(_ IPv6 )-. +-------------+ +-------------+ 764 (__ EUN )--| Host D | 765 `-(______)-' +-------------+ 767 Figure 2: AERO Reference Operational Scenario 769 In Figure 2, AERO Server ('A') connects to the AERO link and connects 770 to the IPv6 Internet, either directly or via an AERO Relay (not 771 shown). Server ('A') assigns the address fe80::ID to its AERO 772 interface with link-layer address L2(A). Server ('A') next arranges 773 to add L2(A) to a published list of valid Servers for the AERO link. 775 AERO Client ('B') receives the IPv6 prefix 2001:db8:0::/48 in a 776 DHCPv6 PD exchange via AERO Server ('A') then assigns the address 777 fe80::2001:db8:0:0 to its AERO interface with link-layer address 778 L2(B). Client ('B') configures a default route and neighbor cache 779 entry via the AERO interface with next-hop address fe80::ID and link- 780 layer address L2(A), then sub-delegates the prefix 2001:db8:0::/48 to 781 its attached EUNs. IPv6 host ('D') connects to the EUN, and 782 configures the address 2001:db8:0::1. 784 AERO Client ('C') receives the IPv6 prefix 2001:db8:1::/48 in a 785 DHCPv6 PD exchange via AERO Server ('A') then assigns the address 786 fe80::2001:db8:1:0 to its AERO interface with link-layer address 787 L2(C). Client ('C') configures a default route and neighbor cache 788 entry via the AERO interface with next-hop address fe80::ID and link- 789 layer address L2(A), then sub-delegates the prefix 2001:db8:1::/48 to 790 its attached EUNs. IPv6 host ('E') connects to the EUN, and 791 configures the address 2001:db8:1::1. 793 Finally, IPv6 host ('F') connects to an IPv6 network outside of the 794 AERO link domain. Host ('F') configures its IPv6 interface in a 795 manner specific to its attached IPv6 link, and assigns the address 796 2001:db8:2::1 to its IPv6 link interface. 798 3.9.2. Classical Redirection Approaches 800 With reference to Figure 2, when the IPv6 source host ('D') sends a 801 packet to an IPv6 destination host ('E'), the packet is first 802 forwarded via the EUN to AERO Client ('B'). Client ('B') then 803 forwards the packet over its AERO interface to AERO Server ('A'), 804 which then re-encapsulates and forwards the packet to AERO Client 805 ('C'), where the packet is finally forwarded to the IPv6 destination 806 host ('E'). When Server ('A') re-encapsulates and forwards the 807 packet back out on its advertising AERO interface, it must arrange to 808 redirect Client ('B') toward Client ('C') as a better next-hop node 809 on the AERO link that is closer to the final destination. However, 810 this redirection process applied to AERO interfaces must be more 811 carefully orchestrated than on ordinary links since the parties may 812 be separated by potentially many underlying network routing hops. 814 Consider a first alternative in which Server ('A') informs Client 815 ('B') only and does not inform Client ('C') (i.e., "classical 816 redirection"). In that case, Client ('C') has no way of knowing that 817 Client ('B') is authorized to forward packets from the claimed source 818 address, and it may simply elect to drop the packets. Also, Client 819 ('B') has no way of knowing whether Client ('C') is performing some 820 form of source address filtering that would reject packets arriving 821 from a node other than a trusted default router, nor whether Client 822 ('C') is even reachable via a direct path that does not involve 823 Server ('A'). 825 Consider a second alternative in which Server ('A') informs both 826 Client ('B') and Client ('C') separately, via independent redirection 827 control messages (i.e., "augmented redirection"). In that case, if 828 Client ('B') receives the redirection control message but Client 829 ('C') does not, subsequent packets sent by Client ('B') could be 830 dropped due to filtering since Client ('C') would not have a route to 831 verify the claimed source address. Also, if Client ('C') receives 832 the redirection control message but Client ('B') does not, subsequent 833 packets sent in the reverse direction by Client ('C') would be lost. 835 Since both of these alternatives have shortcomings, a new redirection 836 technique (i.e., "AERO redirection") is needed. 838 3.9.3. Concept of Operations 840 Again, with reference to Figure 2, when source host ('D') sends a 841 packet to destination host ('E'), the packet is first forwarded over 842 the source host's attached EUN to Client ('B'), which then forwards 843 the packet via its AERO interface to Server ('A'). 845 Server ('A') then re-encapsulates and forwards the packet out the 846 same AERO interface toward Client ('C') and also sends an AERO 847 "Predirect" message forward to Client ('C') as specified in 848 Section 3.9.5. The Predirect message includes Client ('B')'s 849 network- and link-layer addresses as well as information that Client 850 ('C') can use to determine the IPv6 prefix used by Client ('B') . 851 After Client ('C') receives the Predirect message, it process the 852 message and returns an AERO Redirect message destined for Client 853 ('B') via Server ('A') as specified in Section 3.9.6. During the 854 process, Client ('C') also creates or updates a neighbor cache entry 855 for Client ('B') and creates an IPv6 forwarding table entry for 856 Client ('B')'s IPv6 prefix. 858 When Server ('A') receives the Redirect message, it re-encapsulates 859 the message and forwards it on to Client ('B') as specified in 860 Section 3.9.7. The message includes Client ('C')'s network- and 861 link-layer addresses as well as information that Client ('B') can use 862 to determine the IPv6 prefix used by Client ('C'). After Client 863 ('B') receives the Redirect message, it processes the message as 864 specified in Section 3.9.8. During the process, Client ('B') also 865 creates or updates a neighbor cache entry for Client ('C') and 866 creates an IPv6 forwarding table entry for Client ('C')'s IPv6 867 prefix. 869 Following the above Predirect/Redirect message exchange, forwarding 870 of packets from Client ('B') to Client ('C') without involving Server 871 ('A) as an intermediary is enabled. The mechanisms that support this 872 exchange are specified in the following sections. 874 3.9.4. Message Format 876 AERO Redirect/Predirect messages use the same format as for ICMPv6 877 Redirect messages depicted in Section 4.5 of [RFC4861], but also 878 include a new "Prefix Length" field taken from the low-order 8 bits 879 of the Redirect message Reserved field (valid values for the Prefix 880 Length field are 0 through 64). The Redirect/Predirect messages are 881 formatted as shown in Figure 3: 883 0 1 2 3 884 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 885 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 886 | Type (=137) | Code (=0/1) | Checksum | 887 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 888 | Reserved | Prefix Length | 889 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 890 | | 891 + + 892 | | 893 + Target Address + 894 | | 895 + + 896 | | 897 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 898 | | 899 + + 900 | | 901 + Destination Address + 902 | | 903 + + 904 | | 905 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 906 | Options ... 907 +-+-+-+-+-+-+-+-+-+-+-+- 909 Figure 3: AERO Redirect/Predirect Message Format 911 3.9.5. Sending Predirects 913 When a Server forwards a packet out the same AERO interface that it 914 arrived on, the Server sends a Predirect message forward toward the 915 AERO Client nearest the destination instead of sending a Redirect 916 message back to the Client nearest the source. 918 In the reference operational scenario, when Server ('A') forwards a 919 packet sent by Client ('B') toward Client ('C'), it also sends a 920 Predirect message forward toward Client ('C'), subject to rate 921 limiting (see Section 8.2 of [RFC4861]). Server ('A') prepares the 922 Predirect message as follows: 924 o the link-layer source address is set to 'L2(A)' (i.e., the 925 underlying address of Server ('A')). 927 o the link-layer destination address is set to 'L2(C)' (i.e., the 928 underlying address of Client ('C')). 930 o the network-layer source address is set to fe80::2001:db8:0:0 931 (i.e., the AERO address of Client ('B')). 933 o the network-layer destination address is set to fe80::2001:db8:1:0 934 (i.e., the AERO address of Client ('C')). 936 o the Type is set to 137. 938 o the Code is set to 1 to indicate "Predirect". 940 o the Prefix Length is set to the length of the prefix to be applied 941 to the Target Address. 943 o the Target Address is set to fe80::2001:db8:0:0 (i.e., the AERO 944 address of Client ('B')). 946 o the Destination Address is set to the IPv6 source address of the 947 packet that triggered the Predirection event. 949 o the message includes a TLLAO with Link ID and Preference set to 950 appropriate values for Client ('B')'s underlying interface, and 951 with UDP Port Number and IP Address set to 'L2(B)'. 953 o the message includes a Timestamp option. 955 o the message includes a Redirected Header Option (RHO) that 956 contains the originating packet truncated to ensure that at least 957 the network-layer header is included but the size of the message 958 does not exceed 1280 bytes. 960 Server ('A') then sends the message forward to Client ('C'). 962 3.9.6. Processing Predirects and Sending Redirects 964 When Client ('C') receives a Predirect message, it accepts the 965 message only if the message has a link-layer source address of the 966 Server, i.e. 'L2(A)'. Client ('C') further accepts the message only 967 if it is willing to serve as a redirection target. Next, Client 968 ('C') validates the message according to the ICMPv6 Redirect message 969 validation rules in Section 8.1 of [RFC4861], except that it accepts 970 the message even though the network-layer source address is not that 971 of it's current first-hop router. 973 In the reference operational scenario, when Client ('C') receives a 974 valid Predirect message, it either creates or updates a neighbor 975 cache entry that stores the Target Address of the message as the 976 network-layer address of Client ('B') and stores the link-layer 977 address found in the TLLAO as the link-layer address(es) of Client 978 ('B'). Client ('C') then sets the neighbor cache entry ACCEPT timer 979 with timeout value ACCEPT_TIME. Next, Client ('C') applies the 980 Prefix Length to the Destination Address and records the resulting 981 IPv6 prefix in its IPv6 forwarding table. 983 After processing the message, Client ('C') prepares a Redirect 984 message response as follows: 986 o the link-layer source address is set to 'L2(C)' (i.e., the link- 987 layer address of Client ('C')). 989 o the link-layer destination address is set to 'L2(A)' (i.e., the 990 link-layer address of Server ('A')). 992 o the network-layer source address is set to fe80::2001:db8:1:0 993 (i.e., the AERO address of Client ('C')). 995 o the network-layer destination address is set to fe80::2001:db8:0:0 996 (i.e., the AERO address of Client ('B')). 998 o the Type is set to 137. 1000 o the Code is set to 0 to indicate "Redirect". 1002 o the Prefix Length is set to the length of the prefix to be applied 1003 to the Target Address. 1005 o the Target Address is set to fe80::2001:db8:1:0 (i.e., the AERO 1006 address of Client ('C')). 1008 o the Destination Address is set to the IPv6 destination address of 1009 the packet that triggered the Redirection event. 1011 o the message includes a TLLAO with Link ID and Preference set to 1012 appropriate values for Client ('C')'s underlying interface, and 1013 with UDP Port Number and IP Address set to '0'. 1015 o the message includes a Timestamp option. 1017 o the message includes as much of the RHO copied from the 1018 corresponding AERO Predirect message as possible such that at 1019 least the network-layer header is included but the size of the 1020 message does not exceed 1280 bytes. 1022 After Client ('C') prepares the Redirect message, it sends the 1023 message to Server ('A'). 1025 3.9.7. Re-encapsulating and Relaying Redirects 1027 When Server ('A') receives a Redirect message from Client ('C'), it 1028 validates the message according to the ICMPv6 Redirect message 1029 validation rules in Section 8.1 of [RFC4861] and also verifies that 1030 Client ('C') is authorized to use the Prefix Length in the Redirect 1031 message when applied to the AERO address in the network-layer source 1032 of the Redirect message by searching for the AERO address' embedded 1033 prefix in the IPv6 routing table. If validation fails, Server ('A') 1034 discards the message; otherwise, it copies the correct UDP Port 1035 number and IP Address into the TLLAO supplied by Client ('C'). 1037 Server ('A') then examines the network-layer destination address of 1038 the message to determine the IPv6 next hop toward the prefix of 1039 Client ('B') by searching for the AERO address' embedded prefix in 1040 the IPv6 routing table. If the next hop is reached via the AERO 1041 interface, Server ('A') re-encapsulates the Redirect and relays it on 1042 to Client ('B') by changing the link-layer source address of the 1043 message to 'L2(A)' and changing the link-layer destination address to 1044 'L2(B)'. Server ('A') finally forwards the re-encapsulated message 1045 to Client ('B') without decrementing the network-layer IPv6 header 1046 Hop Limit field. 1048 While not shown in Figure 2, AERO Relays relay Redirect and Predirect 1049 messages in exactly this same fashion described above. See Figure 4 1050 in Appendix A for an extension of the reference operational scenario 1051 that includes Relays. 1053 3.9.8. Processing Redirects 1055 When Client ('B') receives the Redirect message, it accepts the 1056 message only if it has a link-layer source address of the Server, 1057 i.e. 'L2(A)'. Next, Client ('B') validates the message according to 1058 the ICMPv6 Redirect message validation rules in Section 8.1 of 1059 [RFC4861], except that it accepts the message even though the 1060 network-layer source address is not that of it's current first-hop 1061 router. Following validation, Client ('B') then processes the 1062 message as follows. 1064 In the reference operational scenario, when Client ('B') receives the 1065 Redirect message, it either creates or updates a neighbor cache entry 1066 that stores the Target Address of the message as the network-layer 1067 address of Client ('C') and stores the link-layer address found in 1068 the TLLAO as the link-layer address of Client ('C'). Client ('B') 1069 then sets the neighbor cache entry FORWARD timer with timeout value 1070 FORWARD_TIME. Next, Client ('B') applies the Prefix Length to the 1071 Destination Address and records the resulting IPv6 prefix in its IPv6 1072 forwarding table. 1074 Now, Client ('B') has an IPv6 forwarding table entry for 1075 Client('C')'s prefix and a neighbor cache entry with a valid FORWARD 1076 time, while Client ('C') has an IPv6 forwarding table entry for 1077 Client ('B')'s prefix with a valid ACCEPT time. Thereafter, Client 1078 ('B') may forward ordinary network-layer data packets directly to 1079 Client ("C") without involving Server ('A') and Client ('C') can 1080 verify that the packets came from an acceptable source. (In order 1081 for Client ('C') to forward packets to Client ('B') a corresponding 1082 Predirect/Redirect message exchange is required in the reverse 1083 direction.) 1085 3.9.9. Server-Oriented Redirection 1087 In some environments, the Server nearest the Client may need to serve 1088 as the redirection target, e.g., if direct Client-to-Client 1089 communications are not possible. In that case, the Redirect message 1090 Target Address encodes the link-local address of the Server instead 1091 of the link-local address of the Client. 1093 3.10. Neighbor Reachability Maintenance 1095 AERO nodes send unicast NS messages to elicit NA messages from 1096 neighbors the same as described for Neighbor Unreachability Detection 1097 (NUD) in [RFC4861]. When an AERO node sends an NS/NA message, it 1098 MUST use its AERO address as the IPv6 source address and the link- 1099 local address of the neighbor as the IPv6 destination address. When 1100 an AERO node receives an NS/NA message, it accepts the message if it 1101 has a neighbor cache entry for the neighbor; otherwise, it ignores 1102 the message. 1104 When a source Client is redirected to a target Client it SHOULD test 1105 the direct path to the target by sending an initial NS message to 1106 elicit a solicited NA response. While testing the path, the source 1107 Client can either continue sending packets via the Server or maintain 1108 a small queue of packets until target reachability is confirmed. The 1109 source Client SHOULD thereafter continue to test the direct path to 1110 the target Client (see Section 7.3 of [RFC4861]) periodically in 1111 order to keep neighbor cache entries alive. In particular, the 1112 source Client sends NS messages to the target Client subject to rate 1113 limiting in order to receive solicited NA messages. If at any time 1114 the direct path appears to be failing, the source Client can resume 1115 sending packets via the Server which may or may not result in a new 1116 redirection event. 1118 When a target Client receives an NS message from a source Client, it 1119 resets the ACCEPT timer to ACCEPT_TIME if a neighbor cache entry 1120 exists; otherwise, it discards the NS message. 1122 When a source Client receives a solicited NA message from a target 1123 Client, it resets the FORWARD timer to FORWARD_TIME if a neighbor 1124 cache entry exists; otherwise, it discards the NA message. 1126 When the FORWARD timer on a neighbor cache entry expires, the source 1127 Client resumes sending any subsequent packets via the Server and may 1128 (eventually) receive a new Redirect message. When the ACCEPT timer 1129 on a neighbor cache entry expires, the target Client discards any 1130 subsequent packets received directly from the source Client. When 1131 both the FORWARD and ACCEPT timers on a neighbor cache entry expire, 1132 the Client deletes both the neighbor cache entry and the 1133 corresponding IPv6 forwarding table entry. 1135 If the source Client is unable to elicit an NA response from the 1136 target Client after MAX_RETRY attempts, it SHOULD consider the direct 1137 path unusable for forwarding purposes. Otherwise, the source Client 1138 considers the path usable and SHOULD thereafter process any link- 1139 layer errors as a hint that the direct path to the target Client has 1140 either failed or has become intermittent. 1142 3.11. Mobility Management 1144 When a Client needs to change its link-layer address (e.g., due to a 1145 mobility event), it performs an immediate DHCPv6 Renew/Reply via each 1146 of its Servers using the new link-layer address as the source. The 1147 DHCPv6 Renew/Reply exchange will update each Server's neighbor cache. 1148 Next, the Client sends a Predirect message to each of its active 1149 neighbors via a Server as follows: 1151 o the link-layer source address is set to the Client's new link- 1152 layer address. 1154 o the link-layer destination address is set to the link-layer 1155 address of the Server. 1157 o the network-layer source address is set to the Client's AERO 1158 address. 1160 o the network-layer destination address is set to the neighbor's 1161 AERO address. 1163 o the Type is set to 137. 1165 o the Code is set to 1 to indicate "Predirect". 1167 o the Prefix Length is set to the length of the prefix to be applied 1168 to the Target address. 1170 o the Target Address is set to the Client's AERO address. 1172 o the Destination Address contains the IPv6 source address of a NULL 1173 packet (i.e., a minimum-length IPv6 packet with Next Header set to 1174 'No Next Header') originating from an address within the Client's 1175 prefix. 1177 o the message includes a TLLAO with Link ID and Preference set to 1178 appropriate values for the underlying interface, and with UDP Port 1179 Number and IP Address set to 0. 1181 o the message includes a Timestamp option. 1183 o the message includes a Redirected Header Option (RHO) that 1184 contains the leading portion of the NULL packet without exceeding 1185 1280 bytes. 1187 When the Server receives the Predirect message, it copies the correct 1188 UDP port number and IP address into the TLLAO supplied by the Client, 1189 changes the link-layer source address to its own address, changes the 1190 link-layer destination address to the address of the neighbor, then 1191 forwards the Predirect message to the neighbor based on an IPv6 route 1192 matching the AERO address in the network-layer destination address. 1193 When the neighbor receives the Predirect message, it returns a 1194 Redirect message the same as specified in Section 3.9. 1196 When a Client associates with a new Server, it issues a new DHCPv6 1197 Renew message via the new Server as the DHCPv6 relay. The new Server 1198 then relays the message to the DHCPv6 server and processes the 1199 resulting exchange. After the Client receives the resulting DHCPv6 1200 Reply message, it sends an RS message to the new Server to receive a 1201 new RA message. 1203 When a Client disassociates with an existing Server, it sends a 1204 "terminating RS" message to the old Server. The terminating RS 1205 message is prepared exactly the same as for an ordinary RS message, 1206 except that the Code field contains the value '1'. When the old 1207 Server receives the terminating RS message, it withdraws the IPv6 1208 route from the routing system and deletes the neighbor cache entry 1209 and IPv6 forwarding table entry for the Client. The old Server then 1210 returns an RA message with default router lifetime set to 0 which the 1211 Client can use to verify that the termination signal has been 1212 processed. The client then deletes both the default route and the 1213 neighbor cache entry for the old Server. (Note that the Client and 1214 the old Server MAY impose a small delay before deleting the neighbor 1215 cache and IPv6 forwarding table entries so that any packets already 1216 in the system can still be delivered to the Client.) 1218 3.12. Encapsulation Protocol Version Considerations 1220 A source Client may connect only to an IPvX underlying network, while 1221 the target Client connects only to an IPvY underlying network. In 1222 that case, the target and source Clients have no means for reaching 1223 each other directly (since they connect to underlying networks of 1224 different IP protocol versions) and so must ignore any redirection 1225 messages and continue to send packets via the Server. 1227 3.13. Multicast Considerations 1229 When the underlying network does not support multicast, AERO nodes 1230 map IPv6 link-scoped multicast addresses (including 1231 'All_DHCP_Relay_Agents_and_Servers') to the underlying IP address of 1232 a Server. 1234 When the underlying network supports multicast, AERO nodes use the 1235 multicast address mapping specification found in [RFC2529] for IPv4 1236 underlying networks and use a direct multicast mapping for IPv6 1237 underlying networks. (In the latter case, "direct multicast mapping" 1238 means that if the IPv6 multicast destination address of the 1239 encapsulated packet is "M", then the IPv6 multicast destination 1240 address of the encapsulating header is also "M".) 1242 3.14. Operation on AERO Links Without DHCPv6 Services 1244 When the AERO link does not provide DHCPv6 services, operation can 1245 still be accommodated through administrative configuration of 1246 prefixes on AERO Clients. In that case, administrative 1247 configurations of IPv6 routes and AERO interface neighbor cache 1248 entries on both the Server and Client are also necessary. However, 1249 this may preclude the ability for Clients to dynamically change to 1250 new Servers, and can expose the AERO link to misconfigurations unless 1251 the administrative configurations are carefully coordinated. 1253 3.15. Operation on Server-less AERO Links 1255 In some AERO link scenarios, there may be no Servers on the link and/ 1256 or no need for Clients to use a Server as an intermediary trust 1257 anchor. In that case, each Client can then act as its own Server to 1258 establish neighbor cache entries and IPv6 forwarding table entries by 1259 performing direct Client-to-Client Predirect/Redirect exchanges, and 1260 some other form of trust basis must be applied so that each Client 1261 can verify that the prospective neighbor is authorized to use its 1262 claimed prefix. 1264 When there is no Server on the link, Clients must arrange to receive 1265 prefix delegations and publish the delegations via a secure alternate 1266 prefix delegation authority through some means outside the scope of 1267 this document. 1269 3.16. Other Considerations 1271 IPv6 hosts serviced by an AERO Client can reach IPv4-only services 1272 via a NAT64 gateway [RFC6146] within the IPv6 network. 1274 AERO nodes can use the Default Address Selection Policy with DHCPv6 1275 option [RFC7078] the same as on any IPv6 link. 1277 All other (non-multicast) functions that operate over ordinary IPv6 1278 links operate in the same fashion over AERO links. 1280 4. Implementation Status 1282 An application-layer implementation is in progress. 1284 5. IANA Considerations 1286 The IANA is instructed to assign a new 2-octet Hardware Type number 1287 for AERO in the "arp-parameters" registry per Section 2 of [RFC5494]. 1288 The number is assigned from the 2-octet Unassigned range with 1289 Hardware Type "AERO" and with this document as the reference. 1291 6. Security Considerations 1293 AERO link security considerations are the same as for standard IPv6 1294 Neighbor Discovery [RFC4861] except that AERO improves on some 1295 aspects. In particular, AERO uses a trust basis between Clients and 1296 Servers, where the Clients only engage in the AERO mechanism when it 1297 is facilitated by a trust anchor. AERO also uses DHCPv6 1298 authentication for Client authentication and admissions control. 1300 AERO links must be protected against link-layer address spoofing 1301 attacks in which an attacker on the link pretends to be a trusted 1302 neighbor. Links that provide link-layer securing mechanisms (e.g., 1303 WiFi networks) and links that provide physical security (e.g., 1304 enterprise network wired LANs) provide a first line of defense that 1305 is often sufficient. In other instances, additional securing 1306 mechanisms such as Secure Neighbor Discovery (SeND) [RFC3971], IPsec 1307 [RFC4301] or TLS [RFC5246] may be necessary. 1309 AERO Clients MUST ensure that their connectivity is not used by 1310 unauthorized nodes on EUNs to gain access to a protected network, 1311 i.e., AERO Clients that act as IPv6 routers MUST NOT provide routing 1312 services for unauthorized nodes. (This concern is no different than 1313 for ordinary hosts that receive an IP address delegation but then 1314 "share" the address with unauthorized nodes via an IPv6/IPv6 NAT 1315 function.) 1317 On some AERO links, establishment and maintenance of a direct path 1318 between neighbors requires secured coordination such as through the 1319 Internet Key Exchange (IKEv2) protocol [RFC5996] to establish a 1320 security association. 1322 7. Acknowledgements 1324 Discussions both on IETF lists and in private exchanges helped shape 1325 some of the concepts in this work. Individuals who contributed 1326 insights include Mikael Abrahamsson, Fred Baker, Stewart Bryant, 1327 Brian Carpenter, Wojciech Dec, Brian Haberman, Joel Halpern, Sascha 1328 Hlusiak, Lee Howard, Joe Touch and Bernie Volz. Members of the IESG 1329 also provided valuable input during their review process that greatly 1330 improved the document. Special thanks go to Stewart Bryant, Joel 1331 Halpern and Brian Haberman for their shepherding guidance. 1333 This work has further been encouraged and supported by Boeing 1334 colleagues including Keith Bartley, Dave Bernhardt, Cam Brodie, 1335 Balaguruna Chidambaram, Wen Fang, Anthony Gregory, Jeff Holland, Ed 1336 King, Gen MacLean, Kent Shuey, Mike Slane, Julie Wulff, Yueli Yang, 1337 and other members of the BR&T and BIT mobile networking teams. 1339 Earlier works on NBMA tunneling approaches are found in 1340 [RFC2529][RFC5214][RFC5569]. 1342 8. References 1344 8.1. Normative References 1346 [RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, 1347 August 1980. 1349 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, September 1350 1981. 1352 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 1353 RFC 792, September 1981. 1355 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1356 Requirement Levels", BCP 14, RFC 2119, March 1997. 1358 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1359 (IPv6) Specification", RFC 2460, December 1998. 1361 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 1362 IPv6 Specification", RFC 2473, December 1998. 1364 [RFC3315] Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C., 1365 and M. Carney, "Dynamic Host Configuration Protocol for 1366 IPv6 (DHCPv6)", RFC 3315, July 2003. 1368 [RFC3633] Troan, O. and R. Droms, "IPv6 Prefix Options for Dynamic 1369 Host Configuration Protocol (DHCP) version 6", RFC 3633, 1370 December 2003. 1372 [RFC3971] Arkko, J., Kempf, J., Zill, B., and P. Nikander, "SEcure 1373 Neighbor Discovery (SEND)", RFC 3971, March 2005. 1375 [RFC4213] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms 1376 for IPv6 Hosts and Routers", RFC 4213, October 2005. 1378 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 1379 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 1380 September 2007. 1382 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 1383 Address Autoconfiguration", RFC 4862, September 2007. 1385 [RFC6434] Jankiewicz, E., Loughney, J., and T. Narten, "IPv6 Node 1386 Requirements", RFC 6434, December 2011. 1388 8.2. Informative References 1390 [IRON] Templin, F., "The Internet Routing Overlay Network 1391 (IRON)", Work in Progress, June 2012. 1393 [RFC0879] Postel, J., "TCP maximum segment size and related topics", 1394 RFC 879, November 1983. 1396 [RFC1930] Hawkinson, J. and T. Bates, "Guidelines for creation, 1397 selection, and registration of an Autonomous System (AS)", 1398 BCP 6, RFC 1930, March 1996. 1400 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 1401 Domains without Explicit Tunnels", RFC 2529, March 1999. 1403 [RFC2675] Borman, D., Deering, S., and R. Hinden, "IPv6 Jumbograms", 1404 RFC 2675, August 1999. 1406 [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway 1407 Protocol 4 (BGP-4)", RFC 4271, January 2006. 1409 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 1410 Architecture", RFC 4291, February 2006. 1412 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1413 Internet Protocol", RFC 4301, December 2005. 1415 [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU 1416 Discovery", RFC 4821, March 2007. 1418 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 1419 Errors at High Data Rates", RFC 4963, July 2007. 1421 [RFC4994] Zeng, S., Volz, B., Kinnear, K., and J. Brzozowski, 1422 "DHCPv6 Relay Agent Echo Request Option", RFC 4994, 1423 September 2007. 1425 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 1426 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 1427 March 2008. 1429 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1430 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 1432 [RFC5494] Arkko, J. and C. Pignataro, "IANA Allocation Guidelines 1433 for the Address Resolution Protocol (ARP)", RFC 5494, 1434 April 2009. 1436 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 1437 Route Optimization Requirements for Operational Use in 1438 Aeronautics and Space Exploration Mobile Networks", RFC 1439 5522, October 2009. 1441 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 1442 Infrastructures (6rd)", RFC 5569, January 2010. 1444 [RFC5996] Kaufman, C., Hoffman, P., Nir, Y., and P. Eronen, 1445 "Internet Key Exchange Protocol Version 2 (IKEv2)", RFC 1446 5996, September 2010. 1448 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 1449 NAT64: Network Address and Protocol Translation from IPv6 1450 Clients to IPv4 Servers", RFC 6146, April 2011. 1452 [RFC6204] Singh, H., Beebee, W., Donley, C., Stark, B., and O. 1453 Troan, "Basic Requirements for IPv6 Customer Edge 1454 Routers", RFC 6204, April 2011. 1456 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 1457 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, August 1458 2011. 1460 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 1461 for Equal Cost Multipath Routing and Link Aggregation in 1462 Tunnels", RFC 6438, November 2011. 1464 [RFC6691] Borman, D., "TCP Options and Maximum Segment Size (MSS)", 1465 RFC 6691, July 2012. 1467 [RFC6706] Templin, F., "Asymmetric Extended Route Optimization 1468 (AERO)", RFC 6706, August 2012. 1470 [RFC6864] Touch, J., "Updated Specification of the IPv4 ID Field", 1471 RFC 6864, February 2013. 1473 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 1474 UDP Checksums for Tunneled Packets", RFC 6935, April 2013. 1476 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 1477 for the Use of IPv6 UDP Datagrams with Zero Checksums", 1478 RFC 6936, April 2013. 1480 [RFC6939] Halwasia, G., Bhandari, S., and W. Dec, "Client Link-Layer 1481 Address Option in DHCPv6", RFC 6939, May 2013. 1483 [RFC6980] Gont, F., "Security Implications of IPv6 Fragmentation 1484 with IPv6 Neighbor Discovery", RFC 6980, August 2013. 1486 [RFC7078] Matsumoto, A., Fujisaki, T., and T. Chown, "Distributing 1487 Address Selection Policy Using DHCPv6", RFC 7078, January 1488 2014. 1490 Appendix A. AERO Server and Relay Interworking 1492 Figure 2 depicts a reference AERO operational scenario with a single 1493 Server on the AERO link. In order to support scaling to larger 1494 numbers of nodes, the AERO link can deploy multiple Servers and 1495 Relays, e.g., as shown in Figure 4. 1497 .-(::::::::) 1498 .-(::: IPv6 :::)-. 1499 (:: Internetwork ::) 1500 `-(::::::::::::)-' 1501 `-(::::::)-' 1502 | 1503 +--------------+ +------+-------+ +--------------+ 1504 |AERO Server C | | AERO Relay D | |AERO Server E | 1505 | (default->D) | | (A->C; G->E) | | (default->D) | 1506 | (A->B) | +-------+------+ | (G->F) | 1507 +-------+------+ | +------+-------+ 1508 | | | 1509 X---+---+-------------------+------------------+---+---X 1510 | AERO Link | 1511 +-----+--------+ +--------+-----+ 1512 |AERO Client B | |AERO Client F | 1513 | (default->C) | | (default->E) | 1514 +--------------+ +--------------+ 1515 .-. .-. 1516 ,-( _)-. ,-( _)-. 1517 .-(_ IPv6 )-. .-(_ IPv6 )-. 1518 (__ EUN ) (__ EUN ) 1519 `-(______)-' `-(______)-' 1520 | | 1521 +--------+ +--------+ 1522 | Host A | | Host G | 1523 +--------+ +--------+ 1525 Figure 4: AERO Server/Relay Interworking 1527 In this example, Client ('B') associates with Server ('C'), while 1528 Client ('F') associates with Server ('E'). Furthermore, Servers 1529 ('C') and ('E') do not associate with each other directly, but rather 1530 have an association with Relay ('D') (i.e., a router that has full 1531 topology information concerning its associated Servers and their 1532 Clients). Relay ('D') connects to the AERO link, and also connects 1533 to the native IPv6 Internetwork. 1535 When host ('A') sends a packet toward destination host ('G'), IPv6 1536 forwarding directs the packet through the EUN to Client ('B'), which 1537 forwards the packet to Server ('C') in absence of more-specific 1538 forwarding information. Server ('C') forwards the packet, and it 1539 also generates an AERO Predirect message that is then forwarded 1540 through Relay ('D') to Server ('E'). When Server ('E') receives the 1541 message, it forwards the message to Client ('F'). 1543 After processing the AERO Predirect message, Client ('F') sends an 1544 AERO Redirect message to Server ('E'). Server ('E'), in turn, 1545 forwards the message through Relay ('D') to Server ('C'). When 1546 Server ('C') receives the message, it forwards the message to Client 1547 ('B') informing it that host 'G's EUN can be reached via Client 1548 ('F'), thus completing the AERO redirection. 1550 The network-layer routing information shared between Servers and 1551 Relays must be carefully coordinated. In particular, Relays require 1552 full topology information, while individual Servers only require 1553 partial topology information, i.e., they only need to know the EUN 1554 prefixes associated with their current set of Clients. This can be 1555 accomplished in a number of ways, but a prominent example is through 1556 the use of an internal instance of the Border Gateway Protocol (BGP) 1557 [RFC4271] coordinated between Servers and Relays. This internal BGP 1558 instance does not interact with the public Internet BGP instance; 1559 therefore, the AERO link is presented to the Internet as a stable set 1560 of one or a few short IPv6 prefixes. 1562 In one possible BGP arrangement, each IRON Server is configured as an 1563 Autonomous System Border Router (ASBR) for a stub Autonomous System 1564 (AS) (possibly using a private AS Number (ASN) [RFC1930]), and each 1565 Server further peers with each Relay but does not peer with other 1566 Servers. Each Server maintains a working set of associated Clients, 1567 and dynamically announces new Client prefixes and withdraws departed 1568 Client prefixes in its BGP updates. Relays in turn run interior BGP 1569 (iBGP) between themselves the same as for ordinary ASBRs. The Relays 1570 therefore discover the full topology of the AERO link in terms of the 1571 working set of Clients associated with each Server. Since Clients 1572 are expected to remain associated with their current set of Servers 1573 for extended timeframes, the amount of BGP control messaging between 1574 Servers and Relays should be minimal. However, Servers SHOULD dampen 1575 any route oscillations caused by impatient Clients that associate and 1576 disassociate with the Server excessively. 1578 See [IRON] for further architectural discussion of routing 1579 coordination between Relays and Servers. 1581 Author's Address 1583 Fred L. Templin (editor) 1584 Boeing Research & Technology 1585 P.O. Box 3707 1586 Seattle, WA 98124 1587 USA 1589 Email: fltemplin@acm.org