idnits 2.17.1 draft-templin-intarea-6706bis-60.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 (September 15, 2020) is 1320 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) -- Looks like a reference, but probably isn't: '1' on line 901 -- Looks like a reference, but probably isn't: '2' on line 901 -- Looks like a reference, but probably isn't: '3' on line 874 == Unused Reference: 'BGP' is defined on line 3176, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3189, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3206, but no explicit reference was found in the text == Outdated reference: A later version (-99) exists of draft-templin-6man-omni-interface-31 == Outdated reference: A later version (-31) exists of draft-bonica-6man-comp-rtg-hdr-22 == Outdated reference: A later version (-04) exists of draft-bonica-6man-crh-helper-opt-01 == Outdated reference: A later version (-13) exists of draft-ietf-intarea-tunnels-10 == Outdated reference: A later version (-26) exists of draft-ietf-rtgwg-atn-bgp-06 == Outdated reference: A later version (-11) exists of draft-templin-6man-dhcpv6-ndopt-10 == Outdated reference: A later version (-27) exists of draft-templin-v6ops-pdhost-26 -- Obsolete informational reference (is this intentional?): RFC 3330 (Obsoleted by RFC 5735) -- Obsolete informational reference (is this intentional?): RFC 6106 (Obsoleted by RFC 8106) -- Obsolete informational reference (is this intentional?): RFC 6347 (Obsoleted by RFC 9147) Summary: 0 errors (**), 0 flaws (~~), 12 warnings (==), 7 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: rfc5320, rfc5558, rfc5720, September 15, 2020 5 rfc6179, rfc6706 (if 6 approved) 7 Intended status: Standards Track 8 Expires: March 19, 2021 10 Asymmetric Extended Route Optimization (AERO) 11 draft-templin-intarea-6706bis-60 13 Abstract 15 This document specifies the operation of IP over Overlay Multilink 16 Network (OMNI) interfaces using the Asymmetric Extended Route 17 Optimization (AERO) internetworking and mobility management service. 18 AERO/OMNI use an IPv6 link-local address format that supports 19 operation of the IPv6 Neighbor Discovery (ND) protocol and links ND 20 to IP forwarding. Prefix delegation/registration services are 21 employed for network admission and to manage the routing system. 22 Multilink operation, mobility management, quality of service (QoS) 23 signaling and route optimization are naturally supported through 24 dynamic neighbor cache updates. Standard IP multicasting services 25 are also supported. AERO is a widely-applicable mobile 26 internetworking service especially well-suited to aviation services, 27 intelligent transportation systems, mobile Virtual Private Networks 28 (VPNs) and many other applications. 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at https://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on March 19, 2021. 47 Copyright Notice 49 Copyright (c) 2020 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 (https://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 . . . . . . . . . . . . . . . . . . . . . . . . 4 65 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 66 3. Asymmetric Extended Route Optimization (AERO) . . . . . . . . 10 67 3.1. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 10 68 3.2. The AERO Service over OMNI Links . . . . . . . . . . . . 11 69 3.2.1. AERO/OMNI Reference Model . . . . . . . . . . . . . . 11 70 3.2.2. Link-Local Addresses (LLAs) and Unique Local 71 Addresses (ULAs) . . . . . . . . . . . . . . . . . . 14 72 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 15 73 3.2.4. AERO Encapsulation . . . . . . . . . . . . . . . . . 16 74 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 18 75 3.2.6. Segment Routing To the OMNI Link . . . . . . . . . . 18 76 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 19 77 3.2.8. Segment Routing Header Compression . . . . . . . . . 21 78 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 21 79 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 23 80 3.4.1. AERO Server/Relay Behavior . . . . . . . . . . . . . 23 81 3.4.2. AERO Proxy Behavior . . . . . . . . . . . . . . . . . 24 82 3.4.3. AERO Client Behavior . . . . . . . . . . . . . . . . 24 83 3.4.4. AERO Bridge Behavior . . . . . . . . . . . . . . . . 24 84 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 24 85 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 26 86 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 28 87 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 28 88 3.9. OMNI Interface MTU and Fragmentation . . . . . . . . . . 28 89 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 28 90 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 29 91 3.10.2. Proxy Forwarding Algorithm . . . . . . . . . . . . . 30 92 3.10.3. Server/Relay Forwarding Algorithm . . . . . . . . . 31 93 3.10.4. Bridge Forwarding Algorithm . . . . . . . . . . . . 32 94 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 33 95 3.12. AERO Router Discovery, Prefix Delegation and 96 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 35 97 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 35 98 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 36 99 3.12.3. AERO Server Behavior . . . . . . . . . . . . . . . . 38 100 3.13. The AERO Proxy . . . . . . . . . . . . . . . . . . . . . 41 101 3.13.1. Combined Proxy/Servers . . . . . . . . . . . . . . . 43 102 3.13.2. Detecting and Responding to Server Failures . . . . 43 103 3.13.3. Point-to-Multipoint Server Coordination . . . . . . 44 104 3.14. AERO Route Optimization / Address Resolution . . . . . . 45 105 3.14.1. Route Optimization Initiation . . . . . . . . . . . 45 106 3.14.2. Relaying the NS . . . . . . . . . . . . . . . . . . 46 107 3.14.3. Processing the NS and Sending the NA . . . . . . . . 46 108 3.14.4. Relaying the NA . . . . . . . . . . . . . . . . . . 47 109 3.14.5. Processing the NA . . . . . . . . . . . . . . . . . 47 110 3.14.6. Route Optimization Maintenance . . . . . . . . . . . 47 111 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 48 112 3.16. Mobility Management and Quality of Service (QoS) . . . . 50 113 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 50 114 3.16.2. Announcing Link-Layer Address and/or QoS Preference 115 Changes . . . . . . . . . . . . . . . . . . . . . . 51 116 3.16.3. Bringing New Links Into Service . . . . . . . . . . 51 117 3.16.4. Removing Existing Links from Service . . . . . . . . 51 118 3.16.5. Moving to a New Server . . . . . . . . . . . . . . . 52 119 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 53 120 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 53 121 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 55 122 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 55 123 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 55 124 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 56 125 3.20. Transition Considerations . . . . . . . . . . . . . . . . 57 126 3.21. Detecting and Reacting to Server and Bridge Failures . . 57 127 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 58 128 3.22.1. Use of SEND and CGA . . . . . . . . . . . . . . . . 60 129 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 62 130 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 62 131 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 62 132 6. Security Considerations . . . . . . . . . . . . . . . . . . . 63 133 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 65 134 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 66 135 8.1. Normative References . . . . . . . . . . . . . . . . . . 66 136 8.2. Informative References . . . . . . . . . . . . . . . . . 68 137 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 73 138 A.1. Implementation Strategies for Route Optimization . . . . 73 139 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 74 140 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 74 141 A.4. AERO Critical Infrastructure Considerations . . . . . . . 74 142 A.5. AERO Server Failure Implications . . . . . . . . . . . . 75 143 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 76 144 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 78 145 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 79 147 1. Introduction 149 Asymmetric Extended Route Optimization (AERO) fulfills the 150 requirements of Distributed Mobility Management (DMM) [RFC7333] and 151 route optimization [RFC5522] for aeronautical networking and other 152 network mobility use cases such as intelligent transportation 153 systems. AERO is an internetworking and mobility management service 154 based on the Overlay Multilink Network Interface (OMNI) 155 [I-D.templin-6man-omni-interface] Non-Broadcast, Multiple Access 156 (NBMA) virtual link model. The OMNI link is a virtual overlay 157 configured over one or more underlying Internetworks, and nodes on 158 the link can exchange IP packets via tunneling. Multilink operation 159 allows for increased reliability, bandwidth optimization and traffic 160 path diversity. 162 The AERO service comprises Clients, Proxys, Servers and Relays that 163 are seen as OMNI link neighbors as well as Bridges that interconnect 164 OMNI link segments. Each node's OMNI interface uses an IPv6 link- 165 local address format that supports operation of the IPv6 Neighbor 166 Discovery (ND) protocol [RFC4861] and links ND to IP forwarding. A 167 node's OMNI interface can be configured over multiple underlying 168 interfaces, and may therefore appear as a single interface with 169 multiple link-layer addresses. Each link-layer address is subject to 170 change due to mobility and/or QoS fluctuations, and link-layer 171 address changes are signaled by ND messaging the same as for any IPv6 172 link. 174 AERO provides a cloud-based service where mobile nodes may use any 175 Server acting as a Mobility Anchor Point (MAP) and fixed nodes may 176 use any Relay on the link for efficient communications. Fixed nodes 177 forward packets destined to other AERO nodes to the nearest Relay, 178 which forwards them through the cloud. A mobile node's initial 179 packets are forwarded through the Server, while direct routing is 180 supported through asymmetric extended route optimization while data 181 packets are flowing. Both unicast and multicast communications are 182 supported, and mobile nodes may efficiently move between locations 183 while maintaining continuous communications with correspondents and 184 without changing their IP Address. 186 AERO Bridges are interconnected in a secured private BGP overlay 187 routing instance using encapsulation to provide a hybrid routing/ 188 bridging service that joins the underlying Internetworks of multiple 189 disjoint administrative domains into a single unified OMNI link. 190 Each OMNI link instance is characterized by the set of Mobility 191 Service Prefixes (MSPs) common to all mobile nodes. The link extends 192 to the point where a Relay/Server is on the optimal route from any 193 correspondent node on the link, and provides a conduit between the 194 underlying Internetwork and the OMNI link. To the underlying 195 Internetwork, the Relay/Server is the source of a route to the MSP, 196 and hence uplink traffic to the mobile node is naturally routed to 197 the nearest Relay/Server. 199 AERO assumes the use of PIM Sparse Mode in support of multicast 200 communication. In support of Source Specific Multicast (SSM) when a 201 Mobile Node is the source, AERO route optimization ensures that a 202 shortest-path multicast tree is established with provisions for 203 mobility and multilink operation. In all other multicast scenarios 204 there are no AERO dependencies. 206 AERO was designed for aeronautical networking for both manned and 207 unmanned aircraft, where the aircraft is treated as a mobile node 208 that can connect an Internet of Things (IoT). AERO is also 209 applicable to a wide variety of other use cases. For example, it can 210 be used to coordinate the Virtual Private Network (VPN) links of 211 mobile nodes (e.g., cellphones, tablets, laptop computers, etc.) that 212 connect into a home enterprise network via public access networks 213 using services such as OpenVPN [OVPN]. It can also be used to 214 facilitate vehicular and pedestrian communications services for 215 intelligent transportation systems. Other applicable use cases are 216 also in scope. 218 The following numbered sections present the AERO specification. The 219 appendices at the end of the document are non-normative. 221 2. Terminology 223 The terminology in the normative references applies; especially, the 224 terminology in the OMNI specification 225 [I-D.templin-6man-omni-interface] is used extensively throughout. 226 The following terms are defined within the scope of this document: 228 IPv6 Neighbor Discovery (ND) 229 an IPv6 control message service for coordinating neighbor 230 relationships between nodes connected to a common link. AERO uses 231 the ND service specified in [RFC4861]. 233 IPv6 Prefix Delegation 234 a networking service for delegating IPv6 prefixes to nodes on the 235 link. The nominal service is DHCPv6 [RFC8415], however alternate 236 services (e.g., based on ND messaging) are also in scope. Most 237 notably, a minimal form of prefix delegation known as "prefix 238 registration" can be used if the Client knows its prefix in 239 advance and can represent it in the IPv6 source address of an ND 240 message. 242 Access Network (ANET) 243 a node's first-hop data link service network (e.g., a radio access 244 network, cellular service provider network, corporate enterprise 245 network, etc.) that often provides link-layer security services 246 such as IEEE 802.1X and physical-layer security prevent 247 unauthorized access internally and with border network-layer 248 security services such as firewalls and proxies that prevent 249 unauthorized outside access. 251 ANET interface 252 a node's attachment to a link in an ANET. 254 Internetwork (INET) 255 a connected IP network topology with a coherent routing and 256 addressing plan and that provides a transit backbone service for 257 ANET end systems. INETs also provide an underlay service over 258 which the AERO virtual link is configured. Example INETs include 259 corporate enterprise networks, aviation networks, and the public 260 Internet itself. When there is no administrative boundary between 261 an ANET and the INET, the ANET and INET are one and the same. 263 INET Partition 264 frequently, INETs such as large corporate enterprise networks are 265 sub-divided internally into separate isolated partitions. Each 266 partition is fully connected internally but disconnected from 267 other partitions, and there is no requirement that separate 268 partitions maintain consistent Internet Protocol and/or addressing 269 plans. (Each INET partition is seen as a separate OMNI link 270 segment as discussed below.) 272 INET interface 273 a node's attachment to a link in an INET. 275 INET address 276 an IP address assigned to a node's interface connection to an 277 INET. 279 INET encapsulation 280 the encapsulation of a packet in an outer header or headers that 281 can be routed within the scope of the local INET partition. 283 OMNI link 284 the same as defined in [I-D.templin-6man-omni-interface], and 285 manifested by IPv6 encapsulation [RFC2473]. The OMNI link spans 286 underlying INET segments joined by virtual bridges in a spanning 287 tree the same as a bridged campus LAN. AERO nodes on the OMNI 288 link appear as single-hop neighbors even though they may be 289 separated by multiple underlying INET hops, and can use Segment 290 Routing [RFC8402] to cause packets to visit selected waypoints on 291 the link. 293 OMNI Interface 294 a node's attachment to an OMNI link. Since the addresses assigned 295 to an OMNI interface are managed for uniqueness, OMNI interfaces 296 do not require Duplicate Address Detection (DAD) and therefore set 297 the administrative variable 'DupAddrDetectTransmits' to zero 298 [RFC4862]. 300 OMNI Link-Local Address (LLA) 301 a link local IPv6 address per [RFC4291] constructed as specified 302 in Section 3.2.2. 304 OMNI Unique-Local Address (ULA) 305 a unique local IPv6 address per [RFC4193] constructed as specified 306 in Section 3.2.2. OMNI ULAs are statelessly derived from OMNI 307 LLAs, and vice-versa. 309 underlying interface 310 an ANET or INET interface over which an OMNI interface is 311 configured. 313 Mobility Service Prefix (MSP) 314 an IP prefix assigned to the OMNI link and from which more- 315 specific Mobile Network Prefixes (MNPs) are derived. 317 Mobile Network Prefix (MNP) 318 an IP prefix allocated from an MSP and delegated to an AERO Client 319 or Relay. 321 AERO node 322 a node that is connected to an OMNI link and participates in the 323 AERO internetworking and mobility service. 325 AERO Client ("Client") 326 an AERO node that connects over one or more underlying interfaces 327 and requests MNP delegation/registration service from AERO 328 Servers. The Client assigns a Client LLA to the OMNI interface 329 for use in ND exchanges with other AERO nodes and forwards packets 330 to correspondents according to OMNI interface neighbor cache 331 state. 333 AERO Server ("Server") 334 an INET node that configures an OMNI interface to provide default 335 forwarding and mobility/multilink services for AERO Clients. The 336 Server assigns an administratively-provisioned LLA to its OMNI 337 interface to support the operation of the ND services, and 338 advertises all of its associated MNPs via BGP peerings with 339 Bridges. 341 AERO Relay ("Relay") 342 an AERO Server that also provides forwarding services between 343 nodes reached via the OMNI link and correspondents on other links. 344 AERO Relays are provisioned with MNPs (i.e., the same as for an 345 AERO Client) and run a dynamic routing protocol to discover any 346 non-MNP IP routes. In both cases, the Relay advertises the MSP(s) 347 to its downstream networks, and distributes all of its associated 348 MNPs and non-MNP IP routes via BGP peerings with Bridges (i.e., 349 the same as for an AERO Server). 351 AERO Bridge ("Bridge") 352 a node that provides hybrid routing/bridging services (as well as 353 a security trust anchor) for nodes on an OMNI link. As a router, 354 the Bridge forwards packets using standard IP forwarding. As a 355 bridge, the Bridge forwards packets over the OMNI link without 356 decrementing the IPv6 Hop Limit. AERO Bridges peer with Servers 357 and other Bridges to discover the full set of MNPs for the link as 358 well as any non-MNPs that are reachable via Relays. 360 AERO Proxy ("Proxy") 361 a node that provides proxying services between Clients in an ANET 362 and Servers in external INETs. The AERO Proxy is a conduit 363 between the ANET and external INETs in the same manner as for 364 common web proxies, and behaves in a similar fashion as for ND 365 proxies [RFC4389]. A node may be configured to act as either a 366 Proxy and/or a Server, depending on Client Server selection 367 criteria. 369 ingress tunnel endpoint (ITE) 370 an OMNI interface endpoint that injects encapsulated packets into 371 an OMNI link. 373 egress tunnel endpoint (ETE) 374 an OMNI interface endpoint that receives encapsulated packets from 375 an OMNI link. 377 link-layer address 378 an IP address used as an encapsulation header source or 379 destination address from the perspective of the OMNI interface. 380 When an upper layer protocol (e.g., UDP) is used as part of the 381 encapsulation, the port number is also considered as part of the 382 link-layer address. 384 network layer address 385 the source or destination address of an encapsulated IP packet 386 presented to the OMNI interface. 388 end user network (EUN) 389 an internal virtual or external edge IP network that an AERO 390 Client or Relay connects to the rest of the network via the OMNI 391 interface. The Client/Relay sees each EUN as a "downstream" 392 network, and sees the OMNI interface as the point of attachment to 393 the "upstream" network. 395 Mobile Node (MN) 396 an AERO Client and all of its downstream-attached networks that 397 move together as a single unit, i.e., an end system that connects 398 an Internet of Things. 400 Mobile Router (MR) 401 a MN's on-board router that forwards packets between any 402 downstream-attached networks and the OMNI link. 404 Route Optimization Source (ROS) 405 the AERO node nearest the source that initiates route 406 optimization. The ROS may be a Server or Proxy acting on behalf 407 of the source Client. 409 Route Optimization responder (ROR) 410 the AERO node nearest the target destination that responds to 411 route optimization requests. The ROR may be a Server acting on 412 behalf of a target MNP Client, or a Relay for a non-MNP 413 destination. 415 MAP List 416 a geographically and/or topologically referenced list of addresses 417 of all Servers within the same OMNI link. There is a single MAP 418 list for the entire OMNI link. 420 Distributed Mobility Management (DMM) 421 a BGP-based overlay routing service coordinated by Servers and 422 Bridges that tracks all Server-to-Client associations. 424 Mobility Service (MS) 425 the collective set of all Servers, Proxys, Bridges and Relays that 426 provide the AERO Service to Clients. 428 Mobility Service Endpoint MSE) 429 an individual Server, Proxy, Bridge or Relay in the Mobility 430 Service. 432 Throughout the document, the simple terms "Client", "Server", 433 "Bridge", "Proxy" and "Relay" refer to "AERO Client", "AERO Server", 434 "AERO Bridge", "AERO Proxy" and "AERO Relay", respectively. 435 Capitalization is used to distinguish these terms from other common 436 Internetworking uses in which they appear without capitalization. 438 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 439 the names of node variables, messages and protocol constants) is used 440 throughout this document. The terms "All-Routers multicast", "All- 441 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 442 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 443 generically refer to either Internet Protocol version, i.e., IPv4 444 [RFC0791] or IPv6 [RFC8200]. 446 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 447 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 448 "OPTIONAL" in this document are to be interpreted as described in BCP 449 14 [RFC2119][RFC8174] when, and only when, they appear in all 450 capitals, as shown here. 452 3. Asymmetric Extended Route Optimization (AERO) 454 The following sections specify the operation of IP over OMNI links 455 using the AERO service: 457 3.1. AERO Node Types 459 AERO Clients are Mobile Nodes (MNs) that connect via underlying 460 interfaces with addresses that may change when the Client moves to a 461 new network connection point. AERO Clients register their Mobile 462 Network Prefixes (MNPs) with the AERO service, and distribute the 463 MNPs to nodes on EUNs. AERO Bridges, Servers, Proxys and Relays are 464 critical infrastructure elements in fixed (i.e., non-mobile) INET 465 deployments and hence have permanent and unchanging INET addresses. 466 Together, they constitute the AERO service which provides an OMNI 467 link virtual overlay for connecting AERO Clients. 469 AERO Bridges provide hybrid routing/bridging services (as well as a 470 security trust anchor) for nodes on an OMNI link. Bridges use 471 standard IPv6 routing to forward packets both within the same INET 472 partitions and between disjoint INET partitions based on a mid-layer 473 IPv6 encapsulation per [RFC2473]. The inner IP layer experiences a 474 virtual bridging service since the inner IP TTL/Hop Limit is not 475 decremented during forwarding. Each Bridge also peers with Servers 476 and other Bridges in a dynamic routing protocol instance to provide a 477 Distributed Mobility Management (DMM) service for the list of active 478 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 479 one or more Mobility Service Prefixes (MSPs) and configure secured 480 tunnels with Servers, Relays, Proxys and other Bridges; they further 481 maintain IP forwarding table entries for each MNP and any other 482 reachable non-MNP prefixes. 484 AERO Servers provide default forwarding and mobility/multilink 485 services for AERO Client Mobile Nodes (MNs). Each Server also peers 486 with Bridges in a dynamic routing protocol instance to advertise its 487 list of associated MNPs (see Section 3.2.3). Servers facilitate 488 prefix delegation/registration exchanges with Clients, where each 489 delegated prefix becomes an MNP taken from an MSP. Servers forward 490 packets between OMNI interface neighbors and track each Client's 491 mobility profiles. Servers may further act as Servers for some sets 492 of Clients and as Proxies for others. 494 AERO Proxys provide a conduit for ANET Clients to associate with 495 Servers in external INETs. Client and Servers exchange control plane 496 messages via the Proxy acting as a bridge between the ANET/INET 497 boundary. The Proxy forwards data packets between Clients and the 498 OMNI link according to forwarding information in the neighbor cache. 499 The Proxy function is specified in Section 3.13. Proxys may further 500 act as Proxys for some sets of Clients and as Servers for others. 502 AERO Relays are Servers that provide forwarding services between the 503 OMNI interface and INET/EUN interfaces. Relays are provisioned with 504 MNPs the same as for an AERO Client, and also run a dynamic routing 505 protocol to discover any non-MNP IP routes. The Relay advertises the 506 MSP(s) to its connected networks, and distributes all of its 507 associated MNPs and non-MNP IP routes via BGP peerings with Bridges 509 3.2. The AERO Service over OMNI Links 511 3.2.1. AERO/OMNI Reference Model 513 Figure 1 presents the basic OMNI link reference model: 515 +----------------+ 516 | AERO Bridge B1 | 517 | Nbr: S1, S2, P1| 518 |(X1->S1; X2->S2)| 519 | MSP M1 | 520 +-+---------+--+-+ 521 +--------------+ | Secured | | +--------------+ 522 |AERO Server S1| | tunnels | | |AERO Server S2| 523 | Nbr: C1, B1 +-----+ | +-----+ Nbr: C2, B1 | 524 | default->B1 | | | default->B1 | 525 | X1->C1 | | | X2->C2 | 526 +-------+------+ | +------+-------+ 527 | OMNI link | | 528 X===+===+===================+==)===============+===+===X 529 | | | | 530 +-----+--------+ +--------+--+-----+ +--------+-----+ 531 |AERO Client C1| | AERO Proxy P1 | |AERO Client C2| 532 | Nbr: S1 | |(Proxy Nbr Cache)| | Nbr: S2 | 533 | default->S1 | +--------+--------+ | default->S2 | 534 | MNP X1 | | | MNP X2 | 535 +------+-------+ .--------+------. +-----+--------+ 536 | (- Proxyed Clients -) | 537 .-. `---------------' .-. 538 ,-( _)-. ,-( _)-. 539 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 540 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 541 `-(______)-' +-------+ +-------+ `-(______)-' 543 Figure 1: AERO/OMNI Reference Model 545 In this model: 547 o the OMNI link is an overlay network service configured over one or 548 more underlying INET partitions which may be managed by different 549 administrative authorities and have incompatible protocols and/or 550 addressing plans. 552 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 553 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 554 via BGP peerings over secured tunnels to Servers (S1, S2). 555 Bridges connect the disjoint segments of a partitioned OMNI link. 557 o AERO Servers/Relays S1 and S2 configure secured tunnels with 558 Bridge B1 and also provide mobility, multilink and default router 559 services for their associated Clients C1 and C2. 561 o AERO Clients C1 and C2 associate with Servers S1 and S2, 562 respectively. They receive Mobile Network Prefix (MNP) 563 delegations X1 and X2, and also act as default routers for their 564 associated physical or internal virtual EUNs. Simple hosts H1 and 565 H2 attach to the EUNs served by Clients C1 and C2, respectively. 567 o AERO Proxy P1 configures a secured tunnel with Bridge B1 and 568 provides proxy services for AERO Clients in secured enclaves that 569 cannot associate directly with other OMNI link neighbors. 571 An OMNI link configured over a single INET appears as a single 572 unified link with a consistent underlying network addressing plan. 573 In that case, all nodes on the link can exchange packets via simple 574 INET encapsulation, since the underlying INET is connected. In 575 common practice, however, an OMNI link may be partitioned into 576 multiple "segments", where each segment is a distinct INET 577 potentially managed under a different administrative authority (e.g., 578 as for worldwide aviation service providers such as ARINC, SITA, 579 Inmarsat, etc.). Individual INETs may also themselves be partitioned 580 internally, in which case each internal partition is seen as a 581 separate segment. 583 The addressing plan of each segment is consistent internally but will 584 often bear no relation to the addressing plans of other segments. 585 Each segment is also likely to be separated from others by network 586 security devices (e.g., firewalls, proxies, packet filtering 587 gateways, etc.), and in many cases disjoint segments may not even 588 have any common physical link connections. Therefore, nodes can only 589 be assured of exchanging packets directly with correspondents in the 590 same segment, and not with those in other segments. The only means 591 for joining the segments therefore is through inter-domain peerings 592 between AERO Bridges. 594 The same as for traditional campus LANs, multiple OMNI link segments 595 can be joined into a single unified link via a virtual bridging 596 service using a mid-layer IPv6 encpasulation per [RFC2473] known as 597 the "SPAN header" that supports inter-segment forwarding (i.e., 598 bridging) without decrementing the network-layer TTL/Hop Limit. This 599 bridging of OMNI link segments is shown in Figure 2: 601 . . . . . . . . . . . . . . . . . . . . . . . 602 . . 603 . .-(::::::::) . 604 . .-(::::::::::::)-. +-+ . 605 . (:::: Segment A :::)--|B|---+ . 606 . `-(::::::::::::)-' +-+ | . 607 . `-(::::::)-' | . 608 . | . 609 . .-(::::::::) | . 610 . .-(::::::::::::)-. +-+ | . 611 . (:::: Segment B :::)--|B|---+ . 612 . `-(::::::::::::)-' +-+ | . 613 . `-(::::::)-' | . 614 . | . 615 . .-(::::::::) | . 616 . .-(::::::::::::)-. +-+ | . 617 . (:::: Segment C :::)--|B|---+ . 618 . `-(::::::::::::)-' +-+ | . 619 . `-(::::::)-' | . 620 . | . 621 . ..(etc).. x . 622 . . 623 . . 624 . <- OMNI link Bridged by encapsulation -> . 625 . . . . . . . . . . . . . .. . . . . . . . . 627 Figure 2: Bridging OMNI Link Segments 629 Bridges, Servers, Relays and Proxys connect via secured INET tunnels 630 over their respecitve segments in a spanning tree topology rooted at 631 the Bridges. The secured spanning tree supports strong 632 authentication for IPv6 ND control messages and may also be used to 633 convey the initial data packets in a flow. Route optimization can 634 then be employed to cause data packets to take more direct paths 635 between OMNI link neighbors without having to strictly follow the 636 spanning tree. 638 3.2.2. Link-Local Addresses (LLAs) and Unique Local Addresses (ULAs) 640 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 641 fe80::/10 [RFC4193] to assign LLAs used for network-layer addresses 642 in IPv6 ND and data messages. They also use the Unique Local Address 643 (ULA) prefix fc80::/10 [RFC4193] to form ULAs used for SPAN header 644 source and desitnation addresses. See 645 [I-D.templin-6man-omni-interface] for a full specification of the 646 LLAs and ULAs used by AERO nodes on OMNI links. 648 For routing system organization (see Section 3.2.3), ULAs are 649 organized in partition prefixes, e.g., fc80::1000/116. For each such 650 partition prefix, the Bridge(s) that connect that segment assign the 651 all-zero's address of the prefix as a Subnet Router Anycast address. 652 For example, the Subnet Router Anycast address for fc80::1000/116 is 653 simply fc80::1000. 655 3.2.3. AERO Routing System 657 The AERO routing system comprises a private instance of the Border 658 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 659 and Servers and does not interact with either the public Internet BGP 660 routing system or any underlying INET routing systems. 662 In a reference deployment, each Server is configured as an Autonomous 663 System Border Router (ASBR) for a stub Autonomous System (AS) using 664 an AS Number (ASN) that is unique within the BGP instance, and each 665 Server further uses eBGP to peer with one or more Bridges but does 666 not peer with other Servers. Each INET of a multi-segment OMNI link 667 must include one or more Bridges, which peer with the Servers and 668 Proxys within that INET. All Bridges within the same INET are 669 members of the same hub AS using a common ASN, and use iBGP to 670 maintain a consistent view of all active MNPs currently in service. 671 The Bridges of different INETs peer with one another using eBGP. 673 Bridges advertise the OMNI link's MSPs and any non-MNP routes to each 674 of their Servers. This means that any aggregated non-MNPs (including 675 "default") are advertised to all Servers. Each Bridge configures a 676 black-hole route for each of its MSPs. By black-holing the MSPs, the 677 Bridge will maintain forwarding table entries only for the MNPs that 678 are currently active, and packets destined to all other MNPs will 679 correctly incur Destination Unreachable messages due to the black- 680 hole route. In this way, Servers have only partial topology 681 knowledge (i.e., they know only about the MNPs of their directly 682 associated Clients) and they forward all other packets to Bridges 683 which have full topology knowledge. 685 Each OMNI link segment assigns a unique sub-prefix of fc80::/96 known 686 as the ULA partition prefix. For example, a first segment could 687 assign fc80::1000/116, a second could assign fc80::2000/116, a third 688 could assign fc80::3000/116, etc. The administrative authorities for 689 each segment must therefore coordinate to assure mutually-exclusive 690 partiton prefix assignments, but internal provisioning of each prefix 691 is an independent local consideration for each administrative 692 authority. 694 ULA partition prefixes are statitcally represented in Bridge 695 forwarding tables. Bridges join multiple segments into a unified 696 OMNI link over multiple diverse administrative domains. They support 697 a bridging function by first establishing forwarding table entries 698 for their partiion prefixes either via standard BGP routing or static 699 routes. For example, if three Bridges ('A', 'B' and 'C') from 700 different segments serviced fc80::1000/116, fc80::2000/116 and 701 fc80::3000/116 respectively, then the forwarding tables in each 702 Bridge are as follows: 704 A: fc80::1000/116->local, fc80::2000/116->B, fc80::3000/116->C 706 B: fc80::1000/116->A, fc80::2000/116->local, fc80::3000/116->C 708 C: fc80::1000/116->A, fc80::2000/116->B, fc80::3000/116->local 710 These forwarding table entries are permanent and never change, since 711 they correspond to fixed infrastructure elements in their respective 712 segments. 714 ULA Client prefixes are instead dynamically advertised in the AERO 715 routing system by Servers and Relays that provide service for their 716 corresponding MNPs. For example, if three Servers ('D', 'E' and 'F') 717 service the MNPs 2001:db8:1000:2000::/56, 2001:db8:3000:4000::/56 and 718 2001:db8:5000:6000::/56 then the routing system would include: 720 D: fc80:2001:db8:1000:2000::/72 722 E: fc80:2001:db8:3000:4000::/72 724 F: fc80:2001:db8:5000:6000::/72 726 A full discussion of the BGP-based routing system used by AERO is 727 found in [I-D.ietf-rtgwg-atn-bgp]. 729 3.2.4. AERO Encapsulation 731 With the Client and partition prefixes in place in each Bridge's 732 forwarding table, control and data packets sent between AERO nodes in 733 different segments can therefore be carried over the spanning treee 734 via mid-layer encapsulation using the SPAN header. For example, when 735 an AERO service node forwards a packet with IPv6 address 736 2001:db8:1:2::1 to a target AERO node with IPv6 address 737 2001:db8:1000:2000::1, it first encapsulates the packet in a SPAN 738 header with source address set to its own SPAN address (e.g., 739 fc80::1000:2000) and destination address set to 740 fc80:2001:db8:1000:2000::. Next, it encapsulates the resulting SPAN 741 packet in an INET header with source address set to its own INET 742 address (e.g., 192.0.2.100) and destination set to the INET address 743 of a Bridge (e.g., 192.0.2.1). 745 SPAN encapsulation is based on Generic Packet Tunneling in IPv6 746 [RFC2473]; the encapsulation format in the above example is shown in 747 Figure 3: 749 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 750 | INET Header | 751 | src = 192.0.2.100 | 752 | dst = 192.0.2.1 | 753 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 754 | SPAN Header | 755 | src = fc80::1000:2000 | 756 | dst=fc80:2001:db8:1000:2000:: | 757 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 758 | Inner IP Header | 759 | src = 2001:db8:1:2::1 | 760 | dst = 2001:db8:1000:2000::1 | 761 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 762 | | 763 ~ ~ 764 ~ Inner Packet Body ~ 765 ~ ~ 766 | | 767 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 769 Figure 3: SPAN Encapsulation 771 In this format, the inner IP header and packet body are the original 772 IP packet, the SPAN header is an IPv6 header prepared according to 773 [RFC2473], and the INET header is prepared as discussed in 774 Section 3.6. 776 This gives rise to a routing system that contains both Client prefix 777 routes that may change dynamically due to regional node mobility and 778 partion prefix routes that rarely if ever change. The Bridges can 779 therefore provide link-layer bridging by sending packets over the 780 spanning tree instead of network-layer routing according to MNP 781 routes. As a result, opportunities for packet loss due to node 782 mobility between different segments are mitigated. 784 In normal operations, IPv6 ND messages are conveyed over secured 785 paths between OMNI link neighbors so that specific Proxys, Servers or 786 Relays can be addressed without being subject to mobility events. 787 Conversely, only the first few packets destined to Clients need to 788 traverse secured paths until route optimization can determine a more 789 direct path. 791 3.2.5. Segment Routing Topologies (SRTs) 793 The 16-bit sub-prefixes of fc80::/10 identify up to 64 distinct 794 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 795 OMNI link overlay instance using a mutually-exclusive set of ULAs, 796 and emulates a Virtual LAN (VLAN) service for the OMNI link. In some 797 cases (e.g., when redundant topologies are needed for fault tolerance 798 and reliability) it may be beneficial to deploy multiple SRTs that 799 act as independent overlay instances. A communication failure in one 800 instance therefore will not affect communications in other instances. 802 Each SRT is identified by a distinct value in bits 10-15 of fc80::10, 803 i.e., as fc80::/16, fc81::/16, fc82::/16, etc. This document asserts 804 that up to four SRTs provide a level of safety sufficient for 805 critical communications such as civil aviation. Each SRT is 806 designated with a color that identifies a different OMNI link 807 instance as follows: 809 o Red - corresponds to fc80::/16 811 o Green - corresponds to fc81::/16 813 o Blue-1 - corresponds to fc82::/16 815 o Blue-2 - corresponds to fc83::/16 817 o fc84::/16 through fcbf::/16 are available for additional SRTs. 819 Each OMNI interface assigns an anycast ULA corresponding to its SRT 820 prefix. For example, the anycast ULA for the Green SRT is simply 821 fc81::. The anycast ULA is used for OMNI interface determination in 822 Safety-Based Multilink (SBM) as discussed in 823 [I-D.templin-6man-omni-interface]. Each OMNI interface further 824 applies Performance-Based Multilink (PBM) internally. 826 3.2.6. Segment Routing To the OMNI Link 828 An original IPv6 source can direct a packet to an OMNI link Client by 829 including a Segment Routing Header (SRH) with the anycast ULA for the 830 selected SRT as either the IPv6 destination or as an intermediate hop 831 within the SRH. This allows the original source to determine the 832 specific topology a packet will traverse when there may be multiple 833 alternatives to choose from. Since the SRH contains no useful 834 information for the destination, the Client may elect to delete the 835 SRH before forwarding in order to reduce overhead. This form of 836 Segment Routing supports Safety-Based Multilink (SBM), and can be 837 exercised through general-purpose SRH types such as [RFC8754]. 839 3.2.7. Segment Routing Within the OMNI Link 841 AERO nodes that insert a SPAN header can use Segment Routing within 842 the OMNI link when necessary to influence the path of packets 843 destined to targets in remote segments without requiring all packets 844 to traverse strict spanning tree paths. 846 When a Client, Proxy or Server has a packet to send to a target 847 discovered through route optimization located in the same OMNI link 848 segment, it encapsulates the packet in a SPAN header with the ULA of 849 the target as the destination address if fragmentation is necessary; 850 otherwise, it may omit the SPAN header. The node then uses the 851 target's Link Layer Address (L2ADDR) information for INET 852 encapsulation without including an SRH. 854 When a Client, Proxy or Server has a packet to send to a route 855 optimization target located in a remote OMNI link segment, it 856 encapsulates the packet in a SPAN header with its own ULA as the 857 source address. The node then SHOULD include an SRH [RFC8754] while 858 forwarding the packet to a Bridge. 860 When the SRH is omitted, the node sets the SPAN destination address 861 to the ULA of the target Client/Proxy/Server and packet forwarding is 862 via spanning tree paths. When the SRH is included, the node first 863 sets the destination address to the ULA Subnet Router Anycast address 864 of the remote segment and sets the lower 32 bits of the ULA of the 865 target's Proxy/Server as the Last Hop Segment (LHS). The node also 866 includes an AERO Route Optimization specification in the SRH TLV 867 section as shown in Figure 4: 869 0 1 2 3 870 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 1 871 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 872 | Type=TBD | Length | MNPlen|V| FMT | MNP[1] | 873 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 874 | MNP[2] | MNP[3] | ... | MNP[i] | 875 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 876 ~ Link Layer Address (L2ADDR) ~ 877 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 879 Figure 4: AERO Route Optimization SRH TLV 881 In this format: 883 o Type is TBD to be assigned according to the Segment Routing Header 884 TLV registry [RFC8754]. 886 o Length is the length of the body of the TLV in bytes, excluding 887 the Type and Length fields. 889 o MNPlen encodes a value between 0 and 15 that, when added to 1, 890 indicates the number of octets of the IPv4/IPv6 MNP prefix that 891 follows. (For example, when MNPlen encodes the value 2 the number 892 of octets 'i' is set to the value 3.) 894 o V indicates the IP protocol version of the MNP that follows. V is 895 set to 0 for IPv4 or 1 for IPv6. 897 o FMT is a three bit code that determines the context and format of 898 the L2ADDR exactly as specified in the OMNI Interface Attributes 899 sub-option [I-D.templin-6man-omni-interface]. 901 o MNP[1], MNP[2], etc. up to MNP[i] encode the leading 'i' octets of 902 the MNP, beginning with the most significant octet followed by the 903 next most significant octet, etc. The number of MNP octets to be 904 included is determined by the number of trailing zero octets in 905 the prefix. For example, for the IPv6 MNP 2001:db8:1:2::/64, 'i' 906 is set to 8 and only the leftmost 8 octets of the MNP are 907 included. In the same way, for the IPv4 MNP 192.0.2/24, 'i' is 908 set to 3 and only the leftmost 3 octets of the MNP are included. 910 o Link Layer Address (L2ADDR) is a UDP Port Number and IP address 911 encoded according to FMT exactly as specified in 912 [I-D.templin-6man-omni-interface]. 914 The node then forwards the packet via a local Bridge, which will 915 eventually direct it to a Bridge on the same segment as the target. 917 When a Bridge receives a packet with Segments Left=1 and with LHS on 918 a local segment, it checks to see if there is an AERO Route 919 Optimization TLV. If so, the Bridge creates a ULA destination 920 according to FMT. If FMT indicates that L2ADDR corresponds to a 921 target Proxy/Server, the Bridge concatenates the SRT fc*::/96 prefix 922 with the 32 bit LHS value to form the ULA destination. Otherwise, 923 the Bridge concatenates the SRT fc*::/16 prefix with the leading 924 MNPlen octets of the MNP and sets the remaining rightmost bits to 0 925 to form a Subnet Router Anycast ULA destination. The Bridge then 926 writes the ULA into the SPAN header destination address and 927 encapsulates the packet in an INET header with the target's L2ADDR as 928 the destination then forwards the packet. Since the SRH contains no 929 useful information for the destination, the Bridge may elect to 930 delete the SRH before forwarding in order to reduce overhead. 932 In this way, the Bridge participates in route optimization to reduce 933 traffic load and suboptimal routing through strict spanning tree 934 paths. Note that if the Bridge does not recognize the AERO Route 935 Optimization TLV, it instead places the SRT fc*::/96 prefix 936 concatenated with the 32 bit LHS in the IPv6 destination address and 937 forwards according to the spanning tree. (Note that this is the same 938 behavior that would occur if the AERO Route Optimization TLV were not 939 present). 941 3.2.8. Segment Routing Header Compression 943 In the Segment Routing use cases discussed above, the segment routing 944 headers must be kept to a minimum size since source and target 945 Clients may be located behind low-end wireless links (e.g., 1Mbps or 946 less). The Compressed Routing Header (CRH) 947 [I-D.bonica-6man-comp-rtg-hdr] provides a compact form that reduces 948 the header size by omitting invariant information. The CRH Helper 949 option [I-D.bonica-6man-crh-helper-opt] can be used to encode the 950 AERO Route Optimization TLV, and the final hop Bridge that performs 951 route optimization may remove the CRH and its helper before 952 encapsulating and forwarding to the target. 954 The CRH and its companion helper option are therefore seen as 955 critical architectural elements that should be quickly progressed 956 through the standards process. Implementations SHOULD use the CRH 957 and its companion helper option instead of other Routing Header types 958 whenever possible to conserve bandwidth. 960 3.3. OMNI Interface Characteristics 962 OMNI interfaces are virtual interfaces configured over one or more 963 underlying interfaces classified as follows: 965 o INET interfaces connect to an INET either natively or through one 966 or several IPv4 Network Address Translators (NATs). Native INET 967 interfaces have global IP addresses that are reachable from any 968 INET correspondent. All Server, Relay and Bridge interfaces are 969 native interfaces, as are INET-facing interfaces of Proxys. NATed 970 INET interfaces connect to a private network behind one or more 971 NATs that provide INET access. Clients that are behind a NAT are 972 required to send periodic keepalive messages to keep NAT state 973 alive when there are no data packets flowing. 975 o ANET interfaces connect to an ANET that is separated from the open 976 INET by a Proxy. Proxys can actively issue control messages over 977 the INET on behalf of the Client to reduce ANET congestion. 979 o VPNed interfaces use security encapsulation over the INET to a 980 Virtual Private Network (VPN) server that also acts as a Server or 981 Proxy. Other than the link-layer encapsulation format, VPNed 982 interfaces behave the same as Direct interfaces. 984 o Direct interfaces connect a Client directly to a Server or Proxy 985 without crossing any ANET/INET paths. An example is a line-of- 986 sight link between a remote pilot and an unmanned aircraft. The 987 same Client considerations apply as for VPNed interfaces. 989 OMNI interfaces use SPAN encapsulation as necessary as discussed in 990 Section 3.2.4. OMNI interfaces use link-layer encapsulation (see: 991 Section 3.6) to exchange packets with OMNI link neighbors over INET 992 or VPNed interfaces. OMNI interfaces do not use link-layer 993 encapsulation over ANET and Direct underlying interfaces. 995 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 996 state the same as for any interface. OMNI interfaces use ND messages 997 including Router Solicitation (RS), Router Advertisement (RA), 998 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 999 neighbor cache management. 1001 OMNI interfaces send ND messages with an OMNI option formatted as 1002 specified in [I-D.templin-6man-omni-interface]. The OMNI option 1003 includes prefix registration information and Interface Attributes 1004 containing link information parameters for the OMNI interface's 1005 underlying interfaces. Each OMNI option may include multiple 1006 Interface Attributes sub-options, each idenfiied by an ifIndex value. 1008 A Client's OMNI interface may be configured over multiple underlying 1009 interface connections. For example, common mobile handheld devices 1010 have both wireless local area network ("WLAN") and cellular wireless 1011 links. These links are often used "one at a time" with low-cost WLAN 1012 preferred and highly-available cellular wireless as a standby, but a 1013 simultaneous-use capability could provide benefits. In a more 1014 complex example, aircraft frequently have many wireless data link 1015 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1016 directional, etc.) with diverse performance and cost properties. 1018 If a Client's multiple underlying interfaces are used "one at a time" 1019 (i.e., all other interfaces are in standby mode while one interface 1020 is active), then ND message OMNI options include only a single 1021 Interface Attribues sub-option set to constant values. In that case, 1022 the Client would appear to have a single interface but with a 1023 dynamically changing link-layer address. 1025 If the Client has multiple active underlying interfaces, then from 1026 the perspective of ND it would appear to have multiple link-layer 1027 addresses. In that case, ND message OMNI options MAY include 1028 multiple Interface Attributes sub-options - each with values that 1029 correspond to a specific interface. Every ND message need not 1030 include Interface Attributes for all underlying interfaces; for any 1031 attributes not included, the neighbor considers the status as 1032 unchanged. 1034 Bridge, Server and Proxy OMNI interfaces may be configured over one 1035 or more secured tunnel interfaces. The OMNI interface configures 1036 both an LLA and its corresponding ULA, while the underlying secured 1037 tunnel interfaces are either unnumbered or configure the same ULA. 1038 The OMNI interface encapsulates each IP packet in a SPAN header and 1039 presents the packet to the underlying secured tunnel interface. 1040 Routing protocols such as BGP that run over the OMNI interface do not 1041 employ SPAN encapsulation, but rather present the routing protocol 1042 messages directly to the underlying secured tunnels while using the 1043 ULA as the source address. This distinction must be honored 1044 consistently according to each node's configuration so that the IP 1045 forwarding table will associate discovered IP routes with the correct 1046 interface. 1048 3.4. OMNI Interface Initialization 1050 AERO Servers, Proxys and Clients configure OMNI interfaces as their 1051 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1052 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1053 to ensure that packets with destination addresses covered by an MNP 1054 not explicitly assigned to a non-OMNI interface are directed to the 1055 OMNI interface. 1057 OMNI interface initialization procedures for Servers, Proxys, Clients 1058 and Bridges are discussed in the following sections. 1060 3.4.1. AERO Server/Relay Behavior 1062 When a Server enables an OMNI interface, it assigns an LLA/ULA 1063 appropriate for the given OMNI link segment. The Server also 1064 configures secured tunnels with one or more neighboring Bridges and 1065 engages in a BGP routing protocol session with each Bridge. 1067 The OMNI interface provides a single interface abstraction to the IP 1068 layer, but internally comprises multiple secured tunnels as well as 1069 an NBMA nexus for sending encapsulated data packets to OMNI interface 1070 neighbors. The Server further configures a service to facilitate ND 1071 exchanges with AERO Clients and manages per-Client neighbor cache 1072 entries and IP forwarding table entries based on control message 1073 exchanges. 1075 Relays are simply Servers that run a dynamic routing protocol to 1076 redistribute routes between the OMNI interface and INET/EUN 1077 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1078 networks on the INET/EUN interfaces (i.e., the same as a Client would 1079 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1080 interfaces. The Relay further provides an attachment point of the 1081 OMNI link to a non-MNP-based global topology. 1083 3.4.2. AERO Proxy Behavior 1085 When a Proxy enables an OMNI interface, it assigns an LLA/ULA and 1086 configures permanent neighbor cache entries the same as for Servers. 1087 The Proxy also configures secured tunnels with one or more 1088 neighboring Bridges and maintains per-Client neighbor cache entries 1089 based on control message exchanges. Importantly Proxys are often 1090 configured to act as Servers, and vice-versa. 1092 3.4.3. AERO Client Behavior 1094 When a Client enables an OMNI interface, it sends RS messages with ND 1095 parameters over its underlying interfaces to a Server, which returns 1096 an RA message with corresponding parameters. (The RS/RA messages may 1097 pass through a Proxy in the case of a Client's ANET interface, or 1098 through one or more NATs in the case of a Client's INET interface.) 1100 3.4.4. AERO Bridge Behavior 1102 AERO Bridges configure an OMNI interface and assign the ULA Subnet 1103 Router Anycast address for each OMNI link segment they connect to. 1104 Bridges configure secured tunnels with Servers, Proxys and other 1105 Bridges; they also configure LLAs/ULAs and permanent neighbor cache 1106 entries the same as Servers. Bridges engage in a BGP routing 1107 protocol session with a subset of the Servers and other Bridges on 1108 the spanning tree (see: Section 3.2.3). 1110 3.5. OMNI Interface Neighbor Cache Maintenance 1112 Each OMNI interface maintains a conceptual neighbor cache that 1113 includes an entry for each neighbor it communicates with on the OMNI 1114 link per [RFC4861]. OMNI interface neighbor cache entries are said 1115 to be one of "permanent", "symmetric", "asymmetric" or "proxy". 1117 Permanent neighbor cache entries are created through explicit 1118 administrative action; they have no timeout values and remain in 1119 place until explicitly deleted. AERO Bridges maintain permanent 1120 neighbor cache entries for their associated Proxys/Servers (and vice- 1121 versa). Each entry maintains the mapping between the neighbor's 1122 network-layer LLA and corresponding INET address. 1124 Symmetric neighbor cache entries are created and maintained through 1125 RS/RA exchanges as specified in Section 3.12, and remain in place for 1126 durations bounded by prefix lifetimes. AERO Servers maintain 1127 symmetric neighbor cache entries for each of their associated 1128 Clients, and AERO Clients maintain symmetric neighbor cache entries 1129 for each of their associated Servers. 1131 Asymmetric neighbor cache entries are created or updated based on 1132 route optimization messaging as specified in Section 3.14, and are 1133 garbage-collected when keepalive timers expire. AERO ROSs maintain 1134 asymmetric neighbor cache entries for active targets with lifetimes 1135 based on ND messaging constants. Asymmetric neighbor cache entries 1136 are unidirectional since only the ROS (and not the ROR) creates an 1137 entry. 1139 Proxy neighbor cache entries are created and maintained by AERO 1140 Proxys when they process Client/Server ND exchanges, and remain in 1141 place for durations bounded by ND and prefix lifetimes. AERO Proxys 1142 maintain proxy neighbor cache entries for each of their associated 1143 Clients. Proxy neighbor cache entries track the Client state and the 1144 address of the Client's associated Server(s). 1146 To the list of neighbor cache entry states in Section 7.3.2 of 1147 [RFC4861], Proxy and Server OMNI interfaces add an additional state 1148 DEPARTED that applies to symmetric and proxy neighbor cache entries 1149 for Clients that have recently departed. The interface sets a 1150 "DepartTime" variable for the neighbor cache entry to "DEPART_TIME" 1151 seconds. DepartTime is decremented unless a new ND message causes 1152 the state to return to REACHABLE. While a neighbor cache entry is in 1153 the DEPARTED state, packets destined to the target Client are 1154 forwarded to the Client's new location instead of being dropped. 1155 When DepartTime decrements to 0, the neighbor cache entry is deleted. 1156 It is RECOMMENDED that DEPART_TIME be set to the default constant 1157 value REACHABLE_TIME plus 10 seconds (40 seconds by default) to allow 1158 a window for packets in flight to be delivered while stale route 1159 optimization state may be present. 1161 When an ROR receives an authentic NS message used for route 1162 optimization, it searches for a symmetric neighbor cache entry for 1163 the target Client. The ROR then returns a solicited NA message 1164 without creating a neighbor cache entry for the ROS, but creates or 1165 updates a target Client "Report List" entry for the ROS and sets a 1166 "ReportTime" variable for the entry to REPORT_TIME seconds. The ROR 1167 resets ReportTime when it receives a new authentic NS message, and 1168 otherwise decrements ReportTime while no authentic NS messages have 1169 been received. It is RECOMMENDED that REPORT_TIME be set to the 1170 default constant value REACHABLE_TIME plus 10 seconds (40 seconds by 1171 default) to allow a window for route optimization to converge before 1172 ReportTime decrements below REACHABLE_TIME. 1174 When the ROS receives a solicited NA message response to its NS 1175 message used for route optimization, it creates or updates an 1176 asymmetric neighbor cache entry for the target network-layer and 1177 link-layer addresses. The ROS then (re)sets ReachableTime for the 1178 neighbor cache entry to REACHABLE_TIME seconds and uses this value to 1179 determine whether packets can be forwarded directly to the target, 1180 i.e., instead of via a default route. The ROS otherwise decrements 1181 ReachableTime while no further solicited NA messages arrive. It is 1182 RECOMMENDED that REACHABLE_TIME be set to the default constant value 1183 30 seconds as specified in [RFC4861]. 1185 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1186 of NS keepalives sent when a correspondent may have gone unreachable, 1187 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1188 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1189 to limit the number of unsolicited NAs that can be sent based on a 1190 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1191 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1192 same as specified in [RFC4861]. 1194 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1195 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1196 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1197 different values are chosen, all nodes on the link MUST consistently 1198 configure the same values. Most importantly, DEPART_TIME and 1199 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1200 REACHABLE_TIME to avoid packet loss due to stale route optimization 1201 state. 1203 3.6. OMNI Interface Encapsulation and Re-encapsulation 1205 OMNI interfaces insert a mid-layer IPv6 header known as the SPAN 1206 header when necessary as discussed in the following sections. After 1207 either inserting or omitting the SPAN header, the OMNI interface also 1208 inserts or omits an outer encapsulation header as discussed below. 1210 OMNI interfaces avoid outer encapsulation over Direct underlying 1211 interfaces and ANET underlying interfaces for which the first-hop 1212 access router is connected to the same underlying link. Otherwise, 1213 OMNI interfaces encapsulate packets according to whether they are 1214 entering the OMNI interface from the network layer or if they are 1215 being re-admitted into the same OMNI link they arrived on. This 1216 latter form of encapsulation is known as "re-encapsulation". 1218 For packets entering the OMNI interface from the network layer, the 1219 OMNI interface copies the "TTL/Hop Limit", "Type of Service/Traffic 1220 Class" [RFC2983], "Flow Label"[RFC6438] (for IPv6) and "Congestion 1221 Experienced" [RFC3168] values in the inner packet's IP header into 1222 the corresponding fields in the SPAN and outer encapsulation 1223 header(s). 1225 For packets undergoing re-encapsulation, the OMNI interface instead 1226 copies these values from the original encapsulation header into the 1227 new encapsulation header, i.e., the values are transferred between 1228 encapsulation headers and *not* copied from the encapsulated packet's 1229 network-layer header. (Note especially that by copying the TTL/Hop 1230 Limit between encapsulation headers the value will eventually 1231 decrement to 0 if there is a (temporary) routing loop.) 1233 OMNI interfaces configured over ANET underlying interfaces which 1234 employ a different IP protocol version and/or may be located multiple 1235 IP hops from the nearest Proxy/Server use IP-in-IP encapsulation so 1236 that the inner packet can traverse the ANET. IPv6 underlying ANET 1237 interfaces use [RFC2473] encapsulation, while IPv4 interfaces use the 1238 appropriate encapsulation per one of [RFC2529][RFC5214][RFC2003]. 1240 OMNI interfaces configured over INET underlying interfaces 1241 encapsulate packets in INET headers according to the next hop 1242 determined in the forwarding algorithm in Section 3.10. If the next 1243 hop is reached via a secured tunnel, the OMNI interface uses an 1244 encapsulation format specific to the secured tunnel type (see: 1245 Section 6). If the next hop is reached via an unsecured INET 1246 interface, the OMNI interface instead uses UDP/IP encapsulation per 1247 [RFC4380] and as extended in [RFC6081]. 1249 When UDP/IP encapsulation is used, the OMNI interface next sets the 1250 UDP source port to a constant value that it will use in each 1251 successive packet it sends, and sets the UDP length field to the 1252 length of the encapsulated packet plus 8 bytes for the UDP header 1253 itself plus the length of any included extension headers or trailers. 1254 The encapsulated packet may be either IPv6 or IPv4, as distinguished 1255 by the version number found in the first four bits. 1257 For UDP/IP-encapsulated packets sent to a Server, Relay or Bridge, 1258 the OMNI interface sets the UDP destination port to 8060, i.e., the 1259 IANA-registered port number for AERO. For packets sent to a Client, 1260 the OMNI interface sets the UDP destination port to the port value 1261 stored in the neighbor cache entry for this Client. The OMNI 1262 interface finally includes/omits the UDP checksum according to 1263 [RFC6935][RFC6936]. 1265 3.7. OMNI Interface Decapsulation 1267 OMNI interfaces decapsulate packets destined either to the AERO node 1268 itself or to a destination reached via an interface other than the 1269 OMNI interface the packet was received on. When the encapsulated 1270 packet arrives in multiple SPAN fragments, the OMNI interface 1271 reassembles as discussed in Section 3.9. Further decapsulation steps 1272 are performed according to the appropriate encapsulation format 1273 specification. 1275 3.8. OMNI Interface Data Origin Authentication 1277 AERO nodes employ simple data origin authentication procedures. In 1278 particular: 1280 o AERO Bridges, Servers and Proxys accept encapsulated data packets 1281 and control messages received from the (secured) spanning tree. 1283 o AERO Proxys and Clients accept packets that originate from within 1284 the same secured ANET. 1286 o AERO Clients and Relays accept packets from downstream network 1287 correspondents based on ingress filtering. 1289 o AERO Clients, Relays and Servers verify the outer UDP/IP 1290 encapsulation addresses according to [RFC4380]. 1292 AERO nodes silently drop any packets that do not satisfy the above 1293 data origin authentication procedures. Further security 1294 considerations are discussed in Section 6. 1296 3.9. OMNI Interface MTU and Fragmentation 1298 The OMNI interface observes the link nature of tunnels, including the 1299 Maximum Transmission Unit (MTU) and the role of fragmentation and 1300 reassembly[I-D.ietf-intarea-tunnels]. 1302 OMNI interface MTU and fragmentation/reassembly procedures are 1303 specified in [I-D.templin-6man-omni-interface]. 1305 3.10. OMNI Interface Forwarding Algorithm 1307 IP packets enter a node's OMNI interface either from the network 1308 layer (i.e., from a local application or the IP forwarding system) or 1309 from the link layer (i.e., from an OMNI interface neighbor). All 1310 packets entering a node's OMNI interface first undergo data origin 1311 authentication as discussed in Section 3.8. Packets that satisfy 1312 data origin authentication are processed further, while all others 1313 are dropped silently. OMNI interfaces wrap accepted packets in a 1314 SPAN header and SRH if necessary as discussed above. 1316 Packets that enter the OMNI interface from the network layer are 1317 forwarded to an OMNI interface neighbor. Packets that enter the OMNI 1318 interface from the link layer are either re-admitted into the OMNI 1319 link or forwarded to the network layer where they are subject to 1320 either local delivery or IP forwarding. In all cases, the OMNI 1321 interface itself MUST NOT decrement the network layer TTL/Hop-count 1322 since its forwarding actions occur below the network layer. 1324 OMNI interfaces may have multiple underlying interfaces and/or 1325 neighbor cache entries for neighbors with multiple underlying 1326 interfaces (see Section 3.3). The OMNI interface uses traffic 1327 classifiers (e.g., DSCP value, port number, etc.) to select an 1328 outgoing underlying interface for each packet based on the node's own 1329 QoS preferences, and also to select a destination link-layer address 1330 based on the neighbor's underlying interface with the highest 1331 preference. AERO implementations SHOULD allow for QoS preference 1332 values to be modified at runtime through network management. 1334 If multiple outgoing interfaces and/or neighbor interfaces have a 1335 preference of "high", the AERO node replicates the packet and sends 1336 one copy via each of the (outgoing / neighbor) interface pairs; 1337 otherwise, the node sends a single copy of the packet via an 1338 interface with the highest preference. AERO nodes keep track of 1339 which underlying interfaces are currently "reachable" or 1340 "unreachable", and only use "reachable" interfaces for forwarding 1341 purposes. 1343 The following sections discuss the OMNI interface forwarding 1344 algorithms for Clients, Proxys, Servers and Bridges. In the 1345 following discussion, a packet's destination address is said to 1346 "match" if it is the same as a cached address, or if it is covered by 1347 a cached prefix (which may be encoded in an LLA). 1349 3.10.1. Client Forwarding Algorithm 1351 When an IP packet enters a Client's OMNI interface from the network 1352 layer the Client searches for an asymmetric neighbor cache entry that 1353 matches the destination. If there is a match, the Client uses one or 1354 more "reachable" neighbor interfaces in the entry for packet 1355 forwarding. If there is no asymmetric neighbor cache entry, the 1356 Client instead forwards the packet toward a Server (the packet is 1357 intercepted by a Proxy if there is a Proxy on the path). The Client 1358 encapsulates the packet in a SPAN header and SRH if necessary and 1359 fragments according to MTU requirements (see: Section 3.9). 1361 When an IP packet enters a Client's OMNI interface from the link- 1362 layer, if the destination matches one of the Client's MNPs or link- 1363 local addresses the Client reassembles and decapsulates as necessary 1364 and delivers the inner packet to the network layer. Otherwise, the 1365 Client drops the packet and MAY return a network-layer ICMP 1366 Destination Unreachable message subject to rate limiting (see: 1367 Section 3.11). 1369 3.10.2. Proxy Forwarding Algorithm 1371 For control messages originating from or destined to a Client, the 1372 Proxy intercepts the message and updates its proxy neighbor cache 1373 entry for the Client. The Proxy then forwards a (proxyed) copy of 1374 the control message. (For example, the Proxy forwards a proxied 1375 version of a Client's NS/RS message to the target neighbor, and 1376 forwards a proxied version of the NA/RA reply to the Client.) 1378 When the Proxy receives a data packet from a Client within the ANET, 1379 the Proxy reassembles and re-fragments if necessary then searches for 1380 an asymmetric neighbor cache entry that matches the destination and 1381 forwards as follows: 1383 o if the destination matches an asymmetric neighbor cache entry, the 1384 Proxy uses one or more "reachable" neighbor interfaces in the 1385 entry for packet forwarding using SPAN encapsulation and including 1386 a SRH if necessary according to the cached link-layer address 1387 information. If the neighbor interface is in the same SPAN 1388 segment, the Proxy forwards the packet directly to the neighbor; 1389 otherwise, it forwards the packet to a Bridge. 1391 o else, the Proxy uses SPAN encapsulation and forwards the packet to 1392 a Bridge while using the ULA corresponding to the packet's 1393 destination as the SPAN destination address. 1395 When the Proxy receives an encapsulated data packet from an INET 1396 neighbor or from a secured tunnel from a Bridge, it accepts the 1397 packet only if data origin authentication succeeds and if there is a 1398 proxy neighbor cache entry that matches the inner destination. Next, 1399 the Proxy reassembles the packet (if necessary) and continues 1400 processing. If the reassembly is complete and the neighbor cache 1401 state is REACHABLE, the Proxy then returns a PTB if necessary (see: 1402 Section 3.9) then either drops or forwards the packet to the Client 1403 while performing SPAN encapsulation and re-fragmentation to the ANET 1404 MTU size if necessary. If the neighbor cache entry state is 1405 DEPARTED, the Proxy instead changes the SPAN destination address to 1406 the address of the new Server and forwards it to a Bridge while 1407 performing re-fragmentation to 1280 bytes if necessary. 1409 3.10.3. Server/Relay Forwarding Algorithm 1411 For control messages destined to a target Client's LLA that are 1412 received from a secured tunnel, the Server intercepts the message and 1413 sends a Proxyed response on behalf of the Client. (For example, the 1414 Server sends a Proxyed NA message reply in response to an NS message 1415 directed to one of its associated Clients.) If the Client's neighbor 1416 cache entry is in the DEPARTED state, however, the Server instead 1417 forwards the packet to the Client's new Server as discussed in 1418 Section 3.16. 1420 When the Server receives an encapsulated data packet from an INET 1421 neighbor or from a secured tunnel, it accepts the packet only if data 1422 origin authentication succeeds. The Server then continues processing 1423 as follows: 1425 o if the network layer destination matches a symmetric neighbor 1426 cache entry in the REACHABLE state the Server prepares the packet 1427 for forwarding to the destination Client. The Server first 1428 reassembles (if necessary) and forwards the packet (while re- 1429 fragmenting if necessary) as specified in Section 3.9. 1431 o else, if the destination matches a symmetric neighbor cache entry 1432 in the DEPARETED state the Server re-encapsulates the packet and 1433 forwards it using the ULA of the Client's new Server as the SPAN 1434 destination. 1436 o else, if the destination matches an asymmetric neighbor cache 1437 entry, the Server uses one or more "reachable" neighbor interfaces 1438 in the entry for packet forwarding via the local INET if the 1439 neighbor is in the same OMNI link segment or using SPAN 1440 encapsulation and Segment Routing if necessary with the final 1441 destination set to the neighbor's ULA otherwise. 1443 o else, if the destination matches a non-MNP route in the IP 1444 forwarding table or an LLA assigned to the Server's OMNI 1445 interface, the Server reassembles if necessary, decapsulates the 1446 packet and releases it to the network layer for local delivery or 1447 IP forwarding. 1449 o else, the Server drops the packet. 1451 When the Server's OMNI interface receives a data packet from the 1452 network layer or from a VPNed or Direct Client, it performs SPAN 1453 encapsulation and fragmentation if necessary, then processes the 1454 packet according to the network-layer destination address as follows: 1456 o if the destination matches a symmetric or asymmetric neighbor 1457 cache entry the Server processes the packet as above. 1459 o else, the Server encapsulates the packet in a SPAN header and 1460 forwards it to a Bridge using its own ULA as the source and the 1461 ULA corresponding to the destination as the destination. 1463 3.10.4. Bridge Forwarding Algorithm 1465 Bridges forward SPAN-encapsulated packets over secured tunnels the 1466 same as any IP router. When the Bridge receives a SPAN-encapsulated 1467 packet via a secured tunnel, it removes the outer INET header and 1468 searches for a forwarding table entry that matches the SPAN 1469 destination address. The Bridge then processes the packet as 1470 follows: 1472 o if the destination matches its ULA Subnet Router Anycast address, 1473 the Bridge checks for a SRH. If there is a SRH with Segments 1474 Left=1, with the ULA of a Proxy/Server on the local segment as the 1475 LHS, and with an AERO Route Optimization TLV, the Bridge examines 1476 the FMT to determine if the target is behind a NAT. If no NAT is 1477 indicated, the Bridge copies the MNP Subnet Router Anycast address 1478 if an MNP is included (otherwise copies the Proxy/Server ULA) into 1479 the destination address then forwards the packet directly to the 1480 L2ADDR using link-layer (UDP/IP) encapsulation. If a NAT is 1481 indicated, the Bridge MAY perform NAT traversal procedures by 1482 sending bubbles per [RFC4380]. The Bridge then either applies 1483 AERO route optimization if NAT traversal procedures have been 1484 successfully applied, or forwards the packet directly to the 1485 Server. 1487 o if the destination matches one of the Bridge's own addresses, the 1488 Bridge submits the packet for local delivery. 1490 o else, if the destination matches a forwarding table entry the 1491 Bridge forwards the packet via a secured tunnel to the next hop. 1492 If the destination matches an MSP without matching an MNP, 1493 however, the Bridge instead drops the packet and returns an ICMP 1494 Destination Unreachable message subject to rate limiting (see: 1495 Section 3.11). 1497 o else, the Bridge drops the packet and returns an ICMP Destination 1498 Unreachable as above. 1500 As for any IP router, the Bridge decrements the TTL/Hop Limit when it 1501 forwards the packet. Therefore, only the Hop Limit in the SPAN 1502 header is decremented, and not the TTL/Hop Limit in the inner packet 1503 header. Bridges do not insert SPAN headers themselves; instead, they 1504 act as IPv6 routers and forward packets based on the destination 1505 address found in the SPAN headers of packets they receive. 1507 3.11. OMNI Interface Error Handling 1509 When an AERO node admits a packet into the OMNI interface, it may 1510 receive link-layer or network-layer error indications. 1512 A link-layer error indication is an ICMP error message generated by a 1513 router in the INET on the path to the neighbor or by the neighbor 1514 itself. The message includes an IP header with the address of the 1515 node that generated the error as the source address and with the 1516 link-layer address of the AERO node as the destination address. 1518 The IP header is followed by an ICMP header that includes an error 1519 Type, Code and Checksum. Valid type values include "Destination 1520 Unreachable", "Time Exceeded" and "Parameter Problem" 1521 [RFC0792][RFC4443]. (OMNI interfaces ignore all link-layer IPv4 1522 "Fragmentation Needed" and IPv6 "Packet Too Big" messages since they 1523 only emit packets that are guaranteed to be no larger than the IP 1524 minimum link MTU as discussed in Section 3.9.) 1526 The ICMP header is followed by the leading portion of the packet that 1527 generated the error, also known as the "packet-in-error". For 1528 ICMPv6, [RFC4443] specifies that the packet-in-error includes: "As 1529 much of invoking packet as possible without the ICMPv6 packet 1530 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1531 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1532 "Internet Header + 64 bits of Original Data Datagram", however 1533 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1534 ICMP datagram SHOULD contain as much of the original datagram as 1535 possible without the length of the ICMP datagram exceeding 576 1536 bytes". 1538 The link-layer error message format is shown in Figure 5 (where, "L2" 1539 and "L3" refer to link-layer and network-layer, respectively): 1541 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1542 ~ ~ 1543 | L2 IP Header of | 1544 | error message | 1545 ~ ~ 1546 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1547 | L2 ICMP Header | 1548 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1549 ~ ~ P 1550 | IP and other encapsulation | a 1551 | headers of original L3 packet | c 1552 ~ ~ k 1553 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1554 ~ ~ t 1555 | IP header of | 1556 | original L3 packet | i 1557 ~ ~ n 1558 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1559 ~ ~ e 1560 | Upper layer headers and | r 1561 | leading portion of body | r 1562 | of the original L3 packet | o 1563 ~ ~ r 1564 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1566 Figure 5: OMNI Interface Link-Layer Error Message Format 1568 The AERO node rules for processing these link-layer error messages 1569 are as follows: 1571 o When an AERO node receives a link-layer Parameter Problem message, 1572 it processes the message the same as described as for ordinary 1573 ICMP errors in the normative references [RFC0792][RFC4443]. 1575 o When an AERO node receives persistent link-layer Time Exceeded 1576 messages, the IP ID field may be wrapping before earlier fragments 1577 awaiting reassembly have been processed. In that case, the node 1578 should begin including integrity checks and/or institute rate 1579 limits for subsequent packets. 1581 o When an AERO node receives persistent link-layer Destination 1582 Unreachable messages in response to encapsulated packets that it 1583 sends to one of its asymmetric neighbor correspondents, the node 1584 should process the message as an indication that a path may be 1585 failing, and optionally initiate NUD over that path. If it 1586 receives Destination Unreachable messages over multiple paths, the 1587 node should allow future packets destined to the correspondent to 1588 flow through a default route and re-initiate route optimization. 1590 o When an AERO Client receives persistent link-layer Destination 1591 Unreachable messages in response to encapsulated packets that it 1592 sends to one of its symmetric neighbor Servers, the Client should 1593 mark the path as unusable and use another path. If it receives 1594 Destination Unreachable messages on many or all paths, the Client 1595 should associate with a new Server and release its association 1596 with the old Server as specified in Section 3.16.5. 1598 o When an AERO Server receives persistent link-layer Destination 1599 Unreachable messages in response to encapsulated packets that it 1600 sends to one of its symmetric neighbor Clients, the Server should 1601 mark the underlying path as unusable and use another underlying 1602 path. 1604 o When an AERO Server or Proxy receives link-layer Destination 1605 Unreachable messages in response to an encapsulated packet that it 1606 sends to one of its permanent neighbors, it treats the messages as 1607 an indication that the path to the neighbor may be failing. 1608 However, the dynamic routing protocol should soon reconverge and 1609 correct the temporary outage. 1611 When an AERO Bridge receives a packet for which the network-layer 1612 destination address is covered by an MSP, if there is no more- 1613 specific routing information for the destination the Bridge drops the 1614 packet and returns a network-layer Destination Unreachable message 1615 subject to rate limiting. The Bridge writes the network-layer source 1616 address of the original packet as the destination address and uses 1617 one of its non link-local addresses as the source address of the 1618 message. 1620 When an AERO node receives an encapsulated packet for which the 1621 reassembly buffer it too small, it drops the packet and returns a 1622 network-layer Packet Too Big (PTB) message. The node first writes 1623 the MRU value into the PTB message MTU field, writes the network- 1624 layer source address of the original packet as the destination 1625 address and writes one of its non link-local addresses as the source 1626 address. 1628 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1630 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1631 coordinated as discussed in the following Sections. 1633 3.12.1. AERO Service Model 1635 Each AERO Server on the OMNI link is configured to facilitate Client 1636 prefix delegation/registration requests. Each Server is provisioned 1637 with a database of MNP-to-Client ID mappings for all Clients enrolled 1638 in the AERO service, as well as any information necessary to 1639 authenticate each Client. The Client database is maintained by a 1640 central administrative authority for the OMNI link and securely 1641 distributed to all Servers, e.g., via the Lightweight Directory 1642 Access Protocol (LDAP) [RFC4511], via static configuration, etc. 1643 Clients receive the same service regardless of the Servers they 1644 select. 1646 AERO Clients and Servers use ND messages to maintain neighbor cache 1647 entries. AERO Servers configure their OMNI interfaces as advertising 1648 NBMA interfaces, and therefore send unicast RA messages with a short 1649 Router Lifetime value (e.g., ReachableTime seconds) in response to a 1650 Client's RS message. Thereafter, Clients send additional RS messages 1651 to keep Server state alive. 1653 AERO Clients and Servers include prefix delegation and/or 1654 registration parameters in RS/RA messages (see 1655 [I-D.templin-6man-omni-interface]). The ND messages are exchanged 1656 between Client and Server according to the prefix management schedule 1657 required by the service. If the Client knows its MNP in advance, it 1658 can employ prefix registration by including its LLA as the source 1659 address of an RS message and with an OMNI option with valid prefix 1660 registration information for the MNP. If the Server (and Proxy) 1661 accept the Client's MNP assertion, they inject the prefix into the 1662 routing system and establish the necessary neighbor cache state. 1664 The following sections specify the Client and Server behavior. 1666 3.12.2. AERO Client Behavior 1668 AERO Clients discover the addresses of Servers in a similar manner as 1669 described in [RFC5214]. Discovery methods include static 1670 configuration (e.g., from a flat-file map of Server addresses and 1671 locations), or through an automated means such as Domain Name System 1672 (DNS) name resolution [RFC1035]. Alternatively, the Client can 1673 discover Server addresses through a layer 2 data link login exchange, 1674 or through a unicast RA response to a multicast/anycast RS as 1675 described below. In the absence of other information, the Client can 1676 resolve the DNS Fully-Qualified Domain Name (FQDN) 1677 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1678 text string and "[domainname]" is a DNS suffix for the OMNI link 1679 (e.g., "example.com"). 1681 To associate with a Server, the Client acts as a requesting router to 1682 request MNPs. The Client prepares an RS message with prefix 1683 management parameters and includes a Nonce and Timestamp option if 1684 the Client needs to correlate RA replies. If the Client already 1685 knows the Server's LLA, it includes the LLA as the network-layer 1686 destination address; otherwise, it includes (link-local) All-Routers 1687 multicast as the network-layer destination. If the Client already 1688 knows its own LLA, it uses the LLA as the network-layer source 1689 address; otherwise, it uses the unspecified IPv6 address (::/128) as 1690 the network-layer source address. 1692 The Client next includes an OMNI option in the RS message to register 1693 its link-layer information with the Server. The Client sets the OMNI 1694 option prefix registration information according to the MNP, and 1695 includes Interface Attributes corresponding to the underlying 1696 interface over which the Client will send the RS message. The Client 1697 MAY include additional Interface Attributes specific to other 1698 underlying interfaces. 1700 The Client then sends the RS message (either directly via Direct 1701 interfaces, via a VPN for VPNed interfaces, via a Proxy for ANET 1702 interfaces or via INET encapsulation for INET interfaces) and waits 1703 for an RA message reply (see Section 3.12.3). The Client retries up 1704 to MAX_RTR_SOLICITATIONS times until an RA is received. If the 1705 Client receives no RAs, or if it receives an RA with Router Lifetime 1706 set to 0, the Client SHOULD abandon this Server and try another 1707 Server. Otherwise, the Client processes the prefix information found 1708 in the RA message. 1710 Next, the Client creates a symmetric neighbor cache entry with the 1711 Server's LLA as the network-layer address and the Server's 1712 encapsulation and/or link-layer addresses as the link-layer address. 1713 The Client records the RA Router Lifetime field value in the neighbor 1714 cache entry as the time for which the Server has committed to 1715 maintaining the MNP in the routing system via this underlying 1716 interface, and caches the other RA configuration information 1717 including Cur Hop Limit, M and O flags, Reachable Time and Retrans 1718 Timer. The Client then autoconfigures LLAs for each of the delegated 1719 MNPs and assigns them to the OMNI interface. The Client also caches 1720 any MSPs included in Route Information Options (RIOs) [RFC4191] as 1721 MSPs to associate with the OMNI link, and assigns the MTU value in 1722 the MTU option to the underlying interface. 1724 The Client then registers additional underlying interfaces with the 1725 Server by sending RS messages via each additional interface. The RS 1726 messages include the same parameters as for the initial RS/RA 1727 exchange, but with destination address set to the Server's LLA. 1729 Following autoconfiguration, the Client sub-delegates the MNPs to its 1730 attached EUNs and/or the Client's own internal virtual interfaces as 1731 described in [I-D.templin-v6ops-pdhost] to support the Client's 1732 downstream attached "Internet of Things (IoT)". The Client 1733 subsequently sends additional RS messages over each underlying 1734 interface before the Router Lifetime received for that interface 1735 expires. 1737 After the Client registers its underlying interfaces, it may wish to 1738 change one or more registrations, e.g., if an interface changes 1739 address or becomes unavailable, if QoS preferences change, etc. To 1740 do so, the Client prepares an RS message to send over any available 1741 underlying interface. The RS includes an OMNI option with prefix 1742 registration information specific to its MNP, with Interface 1743 Attributes specific to the selected underlying interface, and with 1744 any additional Interface Attributes specific to other underlying 1745 interfaces. When the Client receives the Server's RA response, it 1746 has assurance that the Server has been updated with the new 1747 information. 1749 If the Client wishes to discontinue use of a Server it issues an RS 1750 message over any underlying interface with an OMNI option with a 1751 prefix release indication. When the Server processes the message, it 1752 releases the MNP, sets the symmetric neighbor cache entry state for 1753 the Client to DEPARTED and returns an RA reply with Router Lifetime 1754 set to 0. After a short delay (e.g., 2 seconds), the Server 1755 withdraws the MNP from the routing system. 1757 3.12.3. AERO Server Behavior 1759 AERO Servers act as IP routers and support a prefix delegation/ 1760 registration service for Clients. Servers arrange to add their LLAs 1761 to a static map of Server addresses for the link and/or the DNS 1762 resource records for the FQDN "linkupnetworks.[domainname]" before 1763 entering service. Server addresses should be geographically and/or 1764 topologically referenced, and made available for discovery by Clients 1765 on the OMNI link. 1767 When a Server receives a prospective Client's RS message on its OMNI 1768 interface, it SHOULD return an immediate RA reply with Router 1769 Lifetime set to 0 if it is currently too busy or otherwise unable to 1770 service the Client. Otherwise, the Server authenticates the RS 1771 message and processes the prefix delegation/registration parameters. 1772 The Server first determines the correct MNPs to provide to the Client 1773 by searching the Client database. When the Server returns the MNPs, 1774 it also creates a forwarding table entry for each MNP so that the 1775 MNPs are propagated into the routing system (see: Section 3.2.3). 1776 For IPv6, the Server creates an IPv6 forwarding table entry for each 1777 MNP. For IPv4, the Server creates an IPv6 forwarding table entry 1778 with the SPAN Compatibility Prefix (SCP) corresponding to the IPv4 1779 address. 1781 The Server next creates a symmetric neighbor cache entry for the 1782 Client using the base LLA as the network-layer address and with 1783 lifetime set to no more than the smallest prefix lifetime. Next, the 1784 Server updates the neighbor cache entry by recording the information 1785 in each Interface Attributes sub-option in the RS OMNI option. The 1786 Server also records the actual SPAN/INET addresses in the neighbor 1787 cache entry. 1789 Next, the Server prepares an RA message using its LLA as the network- 1790 layer source address and the network-layer source address of the RS 1791 message as the network-layer destination address. The Server sets 1792 the Router Lifetime to the time for which it will maintain both this 1793 underlying interface individually and the symmetric neighbor cache 1794 entry as a whole. The Server also sets Cur Hop Limit, M and O flags, 1795 Reachable Time and Retrans Timer to values appropriate for the OMNI 1796 link. The Server includes the MNPs, any other prefix management 1797 parameters and an OMNI option with no Interface Attributes. The 1798 Server then includes one or more RIOs that encode the MSPs for the 1799 OMNI link, plus an MTU option (see Section 3.9). The Server finally 1800 forwards the message to the Client using SPAN/INET, INET, or NULL 1801 encapsulation as necessary. 1803 After the initial RS/RA exchange, the Server maintains a 1804 ReachableTime timer for each of the Client's underlying interfaces 1805 individually (and for the Client's symmetric neighbor cache entry 1806 collectively) set to expire after ReachableTime seconds. If the 1807 Client (or Proxy) issues additional RS messages, the Server sends an 1808 RA response and resets ReachableTime. If the Server receives an ND 1809 message with a prefix release indication it sets the Client's 1810 symmetric neighbor cache entry to the DEPARTED state and withdraws 1811 the MNP from the routing system after a short delay (e.g., 2 1812 seconds). If ReachableTime expires before a new RS is received on an 1813 individual underlying interface, the Server marks the interface as 1814 DOWN. If ReachableTime expires before any new RS is received on any 1815 individual underlying interface, the Server sets the symmetric 1816 neighbor cache entry state to STALE and sets a 10 second timer. If 1817 the Server has not received a new RS or ND message with a prefix 1818 release indication before the 10 second timer expires, it deletes the 1819 neighbor cache entry and withdraws the MNP from the routing system. 1821 The Server processes any ND messages pertaining to the Client and 1822 returns an NA/RA reply in response to solicitations. The Server may 1823 also issue unsolicited RA messages, e.g., with reconfigure parameters 1824 to cause the Client to renegotiate its prefix delegation/ 1825 registrations, with Router Lifetime set to 0 if it can no longer 1826 service this Client, etc. Finally, If the symmetric neighbor cache 1827 entry is in the DEPARTED state, the Server deletes the entry after 1828 DepartTime expires. 1830 Note: Clients SHOULD notify former Servers of their departures, but 1831 Servers are responsible for expiring neighbor cache entries and 1832 withdrawing routes even if no departure notification is received 1833 (e.g., if the Client leaves the network unexpectedly). Servers 1834 SHOULD therefore set Router Lifetime to ReachableTime seconds in 1835 solicited RA messages to minimize persistent stale cache information 1836 in the absence of Client departure notifications. A short Router 1837 Lifetime also ensures that proactive Client/Server RS/RA messaging 1838 will keep any NAT state alive (see above). 1840 Note: All Servers on an OMNI link MUST advertise consistent values in 1841 the RA Cur Hop Limit, M and O flags, Reachable Time and Retrans Timer 1842 fields the same as for any link, since unpredictable behavior could 1843 result if different Servers on the same link advertised different 1844 values. 1846 3.12.3.1. Lightweight DHCPv6 Relay Agent (LDRA) 1848 When the OMNI option includes a DHCPv6 Unique Identifier (DUID) sub- 1849 option, the use of DHCPv6 Prefix Delegation is indicated. AERO 1850 Clients and Servers are always on the same link (i.e., the OMNI link) 1851 from the perspective of DHCPv6, however in some implementations the 1852 DHCPv6 server and ND function may be located in separate modules. In 1853 that case, the Server's OMNI interface module can act as a 1854 Lightweight DHCPv6 Relay Agent (LDRA)[RFC6221] to relay prefix 1855 delegation messages to and from the DHCPv6 server module. 1857 When the LDRA receives an authentic RS message, it constructs a 1858 corresponding IPv6/UDP/DHCPv6 message. It sets the IPv6 source 1859 address to the source address of the RS message, sets the IPv6 1860 destination address to 'All_DHCP_Relay_Agents_and_Servers' and sets 1861 the UDP fields to values that will be understood by the DHCPv6 1862 server. 1864 The LDRA then wraps the message in a DHCPv6 'Relay-Forward' message 1865 header and includes an 'Interface-Id' option that includes enough 1866 information to allow the LDRA to forward the resulting Reply message 1867 back to the Client (e.g., the Client's link-layer addresses, a 1868 security association identifier, etc.). The LDRA also wraps the OMNI 1869 option into the Interface-Id option, then forwards the message to the 1870 DHCPv6 server. 1872 When the DHCPv6 server prepares a Reply message, it wraps the message 1873 in a 'Relay-Reply' message and echoes the Interface-Id option. The 1874 DHCPv6 server then delivers the Relay-Reply message to the LDRA, 1875 which discards the Relay-Reply wrapper and IPv6/UDP headers, then 1876 uses the DHCPv6 message to construct an RA response to the Client. 1877 The Server uses the information in the Interface-Id option to prepare 1878 the RA message and to cache the link-layer addresses taken from the 1879 OMNI option echoed in the Interface-Id option. 1881 3.13. The AERO Proxy 1883 Clients may connect to protected-spectrum ANETs that employ physical 1884 and/or link-layer security services to facilitate communications to 1885 Servers in outside INETs. In that case, the ANET can employ an AERO 1886 Proxy. The Proxy is located at the ANET/INET border and listens for 1887 RS messages originating from or RA messages destined to ANET Clients. 1888 The Proxy acts on these control messages as follows: 1890 o when the Proxy receives an RS message from a new ANET Client, it 1891 first authenticates the message then examines the network-layer 1892 destination address. If the destination address is a Server's 1893 LLA, the Proxy proceeds to the next step. Otherwise, if the 1894 destination is (link-local) All-Routers multicast, the Proxy 1895 selects a "nearby" Server that is likely to be a good candidate to 1896 serve the Client and replaces the destination address with the 1897 Server's LLA. Next, the Proxy creates a proxy neighbor cache 1898 entry and caches the Client and Server link-layer addresses along 1899 with the OMNI option information and any other identifying 1900 information including Transaction IDs, Client Identifiers, Nonce 1901 values, etc. The Proxy finally encapsulates the (proxyed) RS 1902 message in a SPAN header with source set to the Proxy's ULA and 1903 destination set to the Server's ULA then forwards the message into 1904 the SPAN. 1906 o when the Server receives the RS, it authenticates the message then 1907 creates or updates a symmetric neighbor cache entry for the Client 1908 with the Proxy's ULA as the link-layer address. The Server then 1909 sends an RA message back to the Proxy via the spanning tree. 1911 o when the Proxy receives the RA, it authenticates the message and 1912 matches it with the proxy neighbor cache entry created by the RS. 1913 The Proxy then caches the prefix information as a mapping from the 1914 Client's MNPs to the Client's link-layer address, caches the 1915 Server's advertised Router Lifetime and sets the neighbor cache 1916 entry state to REACHABLE. The Proxy then optionally rewrites the 1917 Router Lifetime and forwards the (proxyed) message to the Client. 1918 The Proxy finally includes an MTU option (if necessary) with an 1919 MTU to use for the underlying ANET interface. 1921 After the initial RS/RA exchange, the Proxy forwards any Client data 1922 packets for which there is no matching asymmetric neighbor cache 1923 entry to a Bridge using SPAN encapsulation with its own ULA as the 1924 source and the ULA corresponding to the Client as the destination. 1925 The Proxy instead forwards any Client data destined to an asymmetric 1926 neighbor cache target directly to the target according to the SPAN/ 1927 link-layer information - the process of establishing asymmetric 1928 neighbor cache entries is specified in Section 3.14. 1930 While the Client is still attached to the ANET, the Proxy sends NS, 1931 RS and/or unsolicited NA messages to update the Server's symmetric 1932 neighbor cache entries on behalf of the Client and/or to convey QoS 1933 updates. This allows for higher-frequency Proxy-initiated RS/RA 1934 messaging over well-connected INET infrastructure supplemented by 1935 lower-frequency Client-initiated RS/RA messaging over constrained 1936 ANET data links. 1938 If the Server ceases to send solicited advertisements, the Proxy 1939 sends unsolicited RAs on the ANET interface with destination set to 1940 (link-local) All-Nodes multicast and with Router Lifetime set to zero 1941 to inform Clients that the Server has failed. Although the Proxy 1942 engages in ND exchanges on behalf of the Client, the Client can also 1943 send ND messages on its own behalf, e.g., if it is in a better 1944 position than the Proxy to convey QoS changes, etc. For this reason, 1945 the Proxy marks any Client-originated solicitation messages (e.g. by 1946 inserting a Nonce option) so that it can return the solicited 1947 advertisement to the Client instead of processing it locally. 1949 If the Client becomes unreachable, the Proxy sets the neighbor cache 1950 entry state to DEPARTED and retains the entry for DepartTime seconds. 1951 While the state is DEPARTED, the Proxy forwards any packets destined 1952 to the Client to a Bridge via SPAN encapsulation with the Client's 1953 current Server as the destination. The Bridge in turn forwards the 1954 packets to the Client's current Server. When DepartTime expires, the 1955 Proxy deletes the neighbor cache entry and discards any further 1956 packets destined to this (now forgotten) Client. 1958 In some ANETs that employ a Proxy, the Client's MNP can be injected 1959 into the ANET routing system. In that case, the Client can send data 1960 messages without encapsulation so that the ANET routing system 1961 transports the unencapsulated packets to the Proxy. This can be very 1962 beneficial, e.g., if the Client connects to the ANET via low-end data 1963 links such as some aviation wireless links. 1965 If the first-hop ANET access router is on the same underlying link 1966 and recognizes the AERO/OMNI protocol, the Client can avoid 1967 encapsulation for both its control and data messages. When the 1968 Client connects to the link, it can send an unencapsulated RS message 1969 with source address set to its LLA and with destination address set 1970 to the LLA of the Client's selected Server or to (link-local) All- 1971 Routers multicast. The Client includes an OMNI option formatted as 1972 specified in [I-D.templin-6man-omni-interface]. 1974 The Client then sends the unencapsulated RS message, which will be 1975 intercepted by the AERO-Aware access router. The access router then 1976 encapsulates the RS message in an ANET header with its own address as 1977 the source address and the address of a Proxy as the destination 1978 address. The access router further remembers the address of the 1979 Proxy so that it can encapsulate future data packets from the Client 1980 via the same Proxy. If the access router needs to change to a new 1981 Proxy, it simply sends another RS message toward the Server via the 1982 new Proxy on behalf of the Client. 1984 In some cases, the access router and Proxy may be one and the same 1985 node. In that case, the node would be located on the same physical 1986 link as the Client, but its message exchanges with the Server would 1987 need to pass through a security gateway at the ANET/INET border. The 1988 method for deploying access routers and Proxys (i.e. as a single node 1989 or multiple nodes) is an ANET-local administrative consideration. 1991 3.13.1. Combined Proxy/Servers 1993 Clients may need to connect directly to Servers via INET, Direct and 1994 VPNed interfaces (i.e., non-ANET interfaces). If the Client's 1995 underlying interfaces all connect via the same INET partition, then 1996 it can connect to a single controlling Server via all interfaces. 1998 If some Client interfaces connect via different INET partitions, 1999 however, the Client still selects a set of controlling Servers and 2000 sends RS messages via their directly-connected Servers while using 2001 the LLA of the controlling Server as the destination. 2003 When a Server receives an RS with destination set to the LLA of a 2004 controlling Server, it acts as a Proxy to forward the message to the 2005 controlling Server while forwarding the corresponding RA reply to the 2006 Client. 2008 3.13.2. Detecting and Responding to Server Failures 2010 In environments where fast recovery from Server failure is required, 2011 Proxys SHOULD use proactive Neighbor Unreachability Detection (NUD) 2012 to track Server reachability in a similar fashion as for 2013 Bidirectional Forwarding Detection (BFD) [RFC5880]. Proxys can then 2014 quickly detect and react to failures so that cached information is 2015 re-established through alternate paths. The NUD control messaging is 2016 carried only over well-connected ground domain networks (i.e., and 2017 not low-end aeronautical radio links) and can therefore be tuned for 2018 rapid response. 2020 Proxys perform proactive NUD with Servers for which there are 2021 currently active ANET Clients by sending continuous NS messages in 2022 rapid succession, e.g., one message per second. The Proxy sends the 2023 NS message via the spanning tree with the Proxy's LLA as the source 2024 and the LLA of the Server as the destination. When the Proxy is also 2025 sending RS messages to the Server on behalf of ANET Clients, the 2026 resulting RA responses can be considered as equivalent hints of 2027 forward progress. This means that the Proxy need not also send a 2028 periodic NS if it has already sent an RS within the same period. If 2029 the Server fails (i.e., if the Proxy ceases to receive 2030 advertisements), the Proxy can quickly inform Clients by sending 2031 multicast RA messages on the ANET interface. 2033 The Proxy sends RA messages on the ANET interface with source address 2034 set to the Server's address, destination address set to (link-local) 2035 All-Nodes multicast, and Router Lifetime set to 0. The Proxy SHOULD 2036 send MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small 2037 delays [RFC4861]. Any Clients on the ANET that had been using the 2038 failed Server will receive the RA messages and associate with a new 2039 Server. 2041 3.13.3. Point-to-Multipoint Server Coordination 2043 In environments where Client messaging over ANETs is bandwidth- 2044 limited and/or expensive, Clients can enlist the services of the 2045 Proxy to coordinate with multiple Servers in a single RS/RA message 2046 exchange. The Client can send a single RS message to (link-local) 2047 All-Routers multicast that includes the ID's of multiple Servers in 2048 MS-Register sub-options of the OMNI option. 2050 When the Proxy receives the RS and processes the OMNI option, it 2051 sends a separate RS to each MS-Register Server ID. When the Proxy 2052 receives an RA, it can optionally return an immediate "singleton" RA 2053 to the Client or record the Server's ID for inclusion in a pending 2054 "aggregate" RA message. The Proxy can then return aggregate RA 2055 messages to the Client including multiple Server IDs in order to 2056 conserve bandwidth. Each RA includes a proper subset of the Server 2057 IDs from the original RS message, and the Proxy must ensure that the 2058 message contents of each RA are consistent with the information 2059 received from the (aggregated) Servers. 2061 Clients can thereafter employ efficient point-to-multipoint Server 2062 coordination under the assistance of the Proxy to reduce the number 2063 of messages sent over the ANET while enlisting the support of 2064 multiple Servers for fault tolerance. Clients can further include 2065 MS-Release suboptions in IPv6 ND messages to request the Proxy to 2066 release from former Servers via the procedures discussed in 2067 Section 3.16.5. 2069 The OMNI interface specification [I-D.templin-6man-omni-interface] 2070 provides further discussion of the Client/Proxy RS/RA messaging 2071 involved in point-to-multipoint coordination. 2073 3.14. AERO Route Optimization / Address Resolution 2075 While data packets are flowing between a source and target node, 2076 route optimization SHOULD be used. Route optimization is initiated 2077 by the first eligible Route Optimization Source (ROS) closest to the 2078 source as follows: 2080 o For Clients on VPNed and Direct interfaces, the Server is the ROS. 2082 o For Clients on ANET interfaces, the Proxy is the ROS. 2084 o For Clients on INET interfaces, the Client itself is the ROS. 2086 o For correspondent nodes on INET/EUN interfaces serviced by a 2087 Relay, the Relay is the ROS. 2089 The route optimization procedure is conducted between the ROS and the 2090 target Server/Relay acting as a Route Optimization Responder (ROR) in 2091 the same manner as for IPv6 ND Address Resolution and using the same 2092 NS/NA messaging. The target may either be a MNP Client serviced by a 2093 Server, or a non-MNP correspondent reachable via a Relay. 2095 The procedures are specified in the following sections. 2097 3.14.1. Route Optimization Initiation 2099 While data packets are flowing from the source node toward a target 2100 node, the ROS performs address resolution by sending an NS message 2101 for Address Resolution (NS(AR)) to receive a solicited NA message 2102 from the ROR. When the ROS sends an NS(AR), it includes: 2104 o the LLA of the ROS as the source address. 2106 o the data packet's destination as the Target Address. 2108 o the Solicited-Node multicast address [RFC4291] formed from the 2109 lower 24 bits of the data packet's destination as the destination 2110 address, e.g., for 2001:db8:1:2::10:2000 the NS destination 2111 address is ff02:0:0:0:0:1:ff10:2000. 2113 The NS(AR) message includes an OMNI option with no Interface 2114 Attributes, such that the target will not create a neighbor cache 2115 entry. The Prefix Length in the OMNI option is set to the Prefix 2116 Length associated with the ROS's LLA. 2118 The ROS then encapsulates the NS(AR) message in a SPAN header with 2119 source set to its own ULA and destination set to the ULA 2120 corresponding to the packet's final destination, then sends the 2121 message into the spanning tree without decrementing the network-layer 2122 TTL/Hop Limit field. 2124 3.14.2. Relaying the NS 2126 When the Bridge receives the NS(AR) message from the ROS, it discards 2127 the INET header and determines that the ROR is the next hop by 2128 consulting its standard IPv6 forwarding table for the SPAN header 2129 destination address. The Bridge then forwards the message toward the 2130 ROR via the spanning tree the same as for any IPv6 router. The 2131 final-hop Bridge in the spanning tree will deliver the message via a 2132 secured tunnel to the ROR. 2134 3.14.3. Processing the NS and Sending the NA 2136 When the ROR receives the NS(AR) message, it examines the Target 2137 Address to determine whether it has a neighbor cache entry and/or 2138 route that matches the target. If there is no match, the ROR drops 2139 the message. Otherwise, the ROR continues processing as follows: 2141 o if the target belongs to an MNP Client neighbor in the DEPARTED 2142 state the ROR changes the NS(AR) message SPAN destination address 2143 to the ULA of the Client's new Server, forwards the message into 2144 the spanning tree and returns from processing. 2146 o If the target belongs to an MNP Client neighbor in the REACHABLE 2147 state, the ROR instead adds the AERO source address to the target 2148 Client's Report List with time set to ReportTime. 2150 o If the target belongs to a non-MNP route, the ROR continues 2151 processing without adding an entry to the Report List. 2153 The ROR then prepares a solicited NA message to send back to the ROS 2154 but does not create a neighbor cache entry. The ROR sets the NA 2155 source address to the LLA corresponding to the target, sets the 2156 Target Address to the target of the solicitation, and sets the 2157 destination address to the source of the solicitation. The ROR then 2158 includes an OMNI option with Prefix Length set to the length 2159 associated with the LLA. 2161 If the target is an MNP Client, the ROR next includes Interface 2162 Attributes in the OMNI option for each of the target Client's 2163 underlying interfaces with current information for each interface and 2164 with the ifIndex field in the OMNI header set to 0 to indicate that 2165 the message originated from the ROR and not the Client. 2167 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2168 according to its own INET address for VPNed or Direct interfaces, to 2169 the INET address of the Proxy or to the Client's INET address for 2170 INET interfaces. The ROR then includes the lower 32 bits of its own 2171 ULA (or the ULA of the Proxy) as the LHS, encodes the ULA prefix 2172 length code in the SRT field and sets the FMT code accordingly as 2173 specified in Section 3.3. 2175 The ROR then sets the NA message R flag to 1 (as a router), S flag to 2176 1 (as a response to a solicitation), and O flag to 0 (as a proxy). 2177 The ROR finally encapsulates the NA message in a SPAN header with 2178 source set to its own ULA and destination set to the source ULA of 2179 the NS(AR) message, then forwards the message into the spanning tree 2180 without decrementing the network-layer TTL/Hop Limit field. 2182 3.14.4. Relaying the NA 2184 When the Bridge receives the NA message from the ROR, it discards the 2185 INET header and determines that the ROS is the next hop by consulting 2186 its standard IPv6 forwarding table for the SPAN header destination 2187 address. The Bridge then forwards the SPAN-encapsulated NA message 2188 toward the ROS the same as for any IPv6 router. The final-hop Bridge 2189 in the spanning tree will deliver the message via a secured tunnel to 2190 the ROS. 2192 3.14.5. Processing the NA 2194 When the ROS receives the solicited NA message, it processes the 2195 message the same as for standard IPv6 Address Resolution [RFC4861]. 2196 In the process, it caches the source ULA then creates an asymmetric 2197 neighbor cache entry for the target and caches all information found 2198 in the OMNI option. The ROS finally sets the asymmetric neighbor 2199 cache entry lifetime to ReachableTime seconds. 2201 3.14.6. Route Optimization Maintenance 2203 Following route optimization, the ROS forwards future data packets 2204 destined to the target via the addresses found in the cached link- 2205 layer information. The route optimization is shared by all sources 2206 that send packets to the target via the ROS, i.e., and not just the 2207 source on behalf of which the route optimization was initiated. 2209 While new data packets destined to the target are flowing through the 2210 ROS, it sends additional NS(AR) messages to the ROR before 2211 ReachableTime expires to receive a fresh solicited NA message the 2212 same as described in the previous sections (route optimization 2213 refreshment strategies are an implementation matter, with a non- 2214 normative example given in Appendix A.1). The ROS uses the cached 2215 ULA of the ROR as the NS(AR) SPAN destination address, and sends up 2216 to MAX_MULTICAST_SOLICIT NS(AR) messages separated by 1 second until 2217 an NA is received. If no NA is received, the ROS assumes that the 2218 current ROR has become unreachable and deletes the target neighbor 2219 cache entry. Subsequent data packets will trigger a new route 2220 optimization per Section 3.14.1 to discover a new ROR while initial 2221 data packets travel over a suboptimal route. 2223 If an NA is received, the ROS then updates the asymmetric neighbor 2224 cache entry to refresh ReachableTime, while (for MNP destinations) 2225 the ROR adds or updates the ROS address to the target's Report List 2226 and with time set to ReportTime. While no data packets are flowing, 2227 the ROS instead allows ReachableTime for the asymmetric neighbor 2228 cache entry to expire. When ReachableTime expires, the ROS deletes 2229 the asymmetric neighbor cache entry. Any future data packets flowing 2230 through the ROS will again trigger a new route optimization. 2232 The ROS may also receive unsolicited NA messages from the ROR at any 2233 time (see: Section 3.16). If there is an asymmetric neighbor cache 2234 entry for the target, the ROS updates the link-layer information but 2235 does not update ReachableTime since the receipt of an unsolicited NA 2236 does not confirm that any forward paths are working. If there is no 2237 asymmetric neighbor cache entry, the ROS simply discards the 2238 unsolicited NA. 2240 In this arrangement, the ROS holds an asymmetric neighbor cache entry 2241 for the target via the ROR, but the ROR does not hold an asymmetric 2242 neighbor cache entry for the ROS. The route optimization neighbor 2243 relationship is therefore asymmetric and unidirectional. If the 2244 target node also has packets to send back to the source node, then a 2245 separate route optimization procedure is performed in the reverse 2246 direction. But, there is no requirement that the forward and reverse 2247 paths be symmetric. 2249 3.15. Neighbor Unreachability Detection (NUD) 2251 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2252 [RFC4861] either reactively in response to persistent link-layer 2253 errors (see Section 3.11) or proactively to confirm reachability. 2254 The NUD algorithm is based on periodic control message exchanges. 2255 The algorithm may further be seeded by ND hints of forward progress, 2256 but care must be taken to avoid inferring reachability based on 2257 spoofed information. For example, authentic IPv6 ND message 2258 exchanges may be considered as acceptable hints of forward progress, 2259 while spurious data packets should not be. 2261 AERO Servers, Proxys and Relays can use (SPAN-encapsulated) standard 2262 NS/NA NUD exchanges sent over the spanning tree to securely test 2263 reachability without risk of DoS attacks from nodes pretending to be 2264 a neighbor; Proxys can further perform NUD to securely verify Server 2265 reachability on behalf of their proxyed Clients. However, a means 2266 for an ROS to test the unsecured forward directions of target route 2267 optimized paths is also necessary. 2269 When an ROR directs an ROS to a neighbor with one or more target 2270 link-layer addresses, the ROS can proactively test each such 2271 unsecured route optimized path by sending "loopback" NS(NUD) 2272 messages. While testing the paths, the ROS can optionally continue 2273 to send packets via the spanning tree, maintain a small queue of 2274 packets until target reachability is confirmed, or (optimistically) 2275 allow packets to flow via the route optimized paths. 2277 When the ROS sends a loopback NS(NUD) message, it uses its LLA as 2278 both the IPv6 source and destination address, and the MNP Subnet- 2279 Router anycast address as the Target Address. The ROS includes a 2280 Nonce and Timestamp option, then encapsulates the message in SPAN/ 2281 INET headers with its own ULA as the source and the ULA of the route 2282 optimization target as the destination. The ROS then forwards the 2283 message to the target (either directly to the L2ADDR of the target if 2284 the target is in the same OMNI link segment, or via a Bridge if the 2285 target is in a different OMNI link segment). 2287 When the route optimization target receives the NS(NUD) message, it 2288 notices that the IPv6 destination address is the same as the source 2289 address. It then reverses the SPAN source and destination addresses 2290 and returns the message to the ROS (either directly or via the 2291 spanning tree). The route optimization target does not decrement the 2292 NS(NUD) message IPv6 Hop-Limit in the process, since the message has 2293 not exited the OMNI link. 2295 When the ROS receives the NS(NUD) message, it can determine from the 2296 Nonce, Timestamp and Target Address that the message originated from 2297 itself and that it transited the forward path. The ROS need not 2298 prepare an NA response, since the destination of the response would 2299 be itself and testing the route optimization path again would be 2300 redundant. 2302 The ROS marks route optimization target paths that pass these NUD 2303 tests as "reachable", and those that do not as "unreachable". These 2304 markings inform the OMNI interface forwarding algorithm specified in 2305 Section 3.10. 2307 Note that to avoid a DoS vector nodes MUST NOT return loopback 2308 NS(NUD) messages received from an unsecured link-layer source via the 2309 spanning tree. 2311 3.16. Mobility Management and Quality of Service (QoS) 2313 AERO is a Distributed Mobility Management (DMM) service. Each Server 2314 is responsible for only a subset of the Clients on the OMNI link, as 2315 opposed to a Centralized Mobility Management (CMM) service where 2316 there is a single network mobility collective entity for all Clients. 2317 Clients coordinate with their associated Servers via RS/RA exchanges 2318 to maintain the DMM profile, and the AERO routing system tracks all 2319 current Client/Server peering relationships. 2321 Servers provide default routing and mobility/multilink services for 2322 their dependent Clients. Clients are responsible for maintaining 2323 neighbor relationships with their Servers through periodic RS/RA 2324 exchanges, which also serves to confirm neighbor reachability. When 2325 a Client's underlying interface address and/or QoS information 2326 changes, the Client is responsible for updating the Server with this 2327 new information. Note that when there is a Proxy in the path, the 2328 Proxy can also perform some RS/RA exchanges on the Client's behalf. 2330 Mobility management considerations are specified in the following 2331 sections. 2333 3.16.1. Mobility Update Messaging 2335 Servers accommodate Client mobility/multilink and/or QoS change 2336 events by sending unsolicited NA (uNA) messages to each ROS in the 2337 target Client's Report List. When a Server sends a uNA message, it 2338 sets the IPv6 source address to the Client's LLA, sets the 2339 destination address to (link-local) All-Nodes multicast and sets the 2340 Target Address to the Client's Subnet-Router anycast address. The 2341 Server also includes an OMNI option with Prefix Length set to the 2342 length associated with the Client's LLA, with Interface Attributes 2343 for the target Client's underlying interfaces and with ifIndex in the 2344 OMNI header set to 0. The Server sets the NA R flag to 1, the S flag 2345 to 0 and the O flag to 0, then encapsulates the message in a SPAN 2346 header with source set to its own ULA and destination set to the ULA 2347 of the ROS and sends the message into the spanning tree. 2349 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2350 reception of uNA messages is unreliable but provides a useful 2351 optimization. In well-connected Internetworks with robust data links 2352 uNA messages will be delivered with high probability, but in any case 2353 the Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT uNAs 2354 to each ROS to increase the likelihood that at least one will be 2355 received. 2357 When the ROS receives a uNA message, it ignores the message if there 2358 is no existing neighbor cache entry for the Client. Otherwise, it 2359 uses the included OMNI option information to update the neighbor 2360 cache entry, but does not reset ReachableTime since the receipt of an 2361 unsolicited NA message from the target Server does not provide 2362 confirmation that any forward paths to the target Client are working. 2364 If uNA messages are lost, the ROS may be left with stale address and/ 2365 or QoS information for the Client for up to ReachableTime seconds. 2366 During this time, the ROS can continue sending packets according to 2367 its stale neighbor cache information. When ReachableTime is close to 2368 expiring, the ROS will re-initiate route optimization and receive 2369 fresh link-layer address information. 2371 In addition to sending uNA messages to the current set of ROSs for 2372 the Client, the Server also sends uNAs to the former link-layer 2373 address for any underlying interface for which the link-layer address 2374 has changed. The uNA messages update Proxys that cannot easily 2375 detect (e.g., without active probing) when a formerly-active Client 2376 has departed. 2378 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2380 When a Client needs to change its underlying interface addresses and/ 2381 or QoS preferences (e.g., due to a mobility event), either the Client 2382 or its Proxys send RS messages to the Server via the spanning tree 2383 with an OMNI option that includes Interface attributes with the new 2384 link quality and address information. 2386 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2387 sending actual data packets in case one or more RAs are lost. If all 2388 RAs are lost, the Client SHOULD re-associate with a new Server. 2390 When the Server receives the Client's changes, it sends uNA messages 2391 to all nodes in the Report List the same as described in the previous 2392 section. 2394 3.16.3. Bringing New Links Into Service 2396 When a Client needs to bring new underlying interfaces into service 2397 (e.g., when it activates a new data link), it sends an RS message to 2398 the Server via the underlying interface with an OMNI option that 2399 includes Interface Attributes with appropriate link quality values 2400 and with link-layer address information for the new link. 2402 3.16.4. Removing Existing Links from Service 2404 When a Client needs to remove existing underlying interfaces from 2405 service (e.g., when it de-activates an existing data link), it sends 2406 an RS or uNA message to its Server with an OMNI option with 2407 appropriate link quality values. 2409 If the Client needs to send RS/uNA messages over an underlying 2410 interface other than the one being removed from service, it MUST 2411 include Interface Attributes with appropriate link quality values for 2412 any underlying interfaces being removed from service. 2414 3.16.5. Moving to a New Server 2416 When a Client associates with a new Server, it performs the Client 2417 procedures specified in Section 3.12.2. The Client also includes MS- 2418 Release identifiers in the RS message OMNI option per 2419 [I-D.templin-6man-omni-interface] if it wants the new Server to 2420 notify any old Servers from which the Client is departing. 2422 When the new Server receives the Client's RS message, it returns an 2423 RA as specified in Section 3.12.3 and sends up to 2424 MAX_NEIGHBOR_ADVERTIISEMENT uNA messages to any old Servers listed in 2425 OMNI option MS-Release identifiers. Each uNA message includes the 2426 Client's LLA as the source address, the old Server's LLA as the 2427 destination address, and an OMNI option with the old Server's ID in 2428 the MS-Release list. The new Server wraps the uNA in a SPAN header 2429 with its own ULA as the source and the old Server's ULA as the 2430 destination, then sends the message into the spanning tree. 2432 When an old Server receives the uNA, it changes the Client's neighbor 2433 cache entry state to DEPARTED, sets the link-layer address of the 2434 Client to the new Server's ULA, and resets DepartTime. After a short 2435 delay (e.g., 2 seconds) the old Server withdraws the Client's MNP 2436 from the routing system. After DepartTime expires, the old Server 2437 deletes the Client's neighbor cache entry. 2439 The old Server also sends unsolicited NA messages to all ROSs in the 2440 Client's Report List with an OMNI option with a single Interface 2441 Attributes sub-option with ifIndex set to 0, and with the Link Layer 2442 address information of the new Server. When the ROS receives the NA, 2443 it caches the address of the new Server in the existing asymmetric 2444 neighbor cache entry and marks the entry as STALE for a period of 10 2445 seconds after which the cache entry is deleted. While in the STALE 2446 state, subsequent data packets flow according to any existing cached 2447 link-layer information and trigger a new NS(AR)/NA exchange via the 2448 new Server. 2450 Clients SHOULD NOT move rapidly between Servers in order to avoid 2451 causing excessive oscillations in the AERO routing system. Examples 2452 of when a Client might wish to change to a different Server include a 2453 Server that has gone unreachable, topological movements of 2454 significant distance, movement to a new geographic region, movement 2455 to a new OMNI link segment, etc. 2457 When a Client moves to a new Server, some of the fragments of a 2458 multiple fragment packet may have already arrived at the old Server 2459 while others are en route to the new Server, however no special 2460 attention in the reassembly algorithm is necessary when re-routed 2461 fragments are simply treated as loss. 2463 3.17. Multicast 2465 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2466 [RFC3810] proxy service for its EUNs and/or hosted applications 2467 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2468 underlying interfaces for which group membership is required. The 2469 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2470 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2471 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2472 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2473 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 2474 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2475 INET/EUN networks. The behaviors identified in the following 2476 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2477 Multicast (ASM) operational modes. 2479 3.17.1. Source-Specific Multicast (SSM) 2481 When an ROS (i.e., an AERO Proxy/Server/Relay) "X" acting as PIM 2482 router receives a Join/Prune message from a node on its downstream 2483 interfaces containing one or more ((S)ource, (G)roup) pairs, it 2484 updates its Multicast Routing Information Base (MRIB) accordingly. 2485 For each S belonging to a prefix reachable via X's non-OMNI 2486 interfaces, X then forwards the (S, G) Join/Prune to any PIM routers 2487 on those interfaces per [RFC7761]. 2489 For each S belonging to a prefix reachable via X's OMNI interface, X 2490 originates a separate copy of the Join/Prune for each (S,G) in the 2491 message using its own LLA as the source address and ALL-PIM-ROUTERS 2492 as the destination address. X then encapsulates each message in a 2493 SPAN header with source address set to the ULA of X and destination 2494 address set to S then forwards the message into the spanning tree, 2495 which delivers it to AERO Server/Relay "Y" that services S. At the 2496 same time, if the message was a Join, X sends a route-optimization NS 2497 message toward each S the same as discussed in Section 3.14. The 2498 resulting NAs will return the LLA for the prefix that matches S as 2499 the network-layer source address and with an OMNI option with the ULA 2500 corresponding to any underlying interfaces that are currently 2501 servicing S. 2503 When Y processes the Join/Prune message, if S located behind any 2504 INET, Direct, or VPNed interfaces Y acts as a PIM router and updates 2505 its MRIB to list X as the next hop in the reverse path. If S is 2506 located behind any Proxys "Z"*, Y also forwards the message to each 2507 Z* over the spanning tree while continuing to use the LLA of X as the 2508 source address. Each Z* then updates its MRIB accordingly and 2509 maintains the LLA of X as the next hop in the reverse path. Since 2510 the Bridges do not examine network layer control messages, this means 2511 that the (reverse) multicast tree path is simply from each Z* (and/or 2512 Y) to X with no other multicast-aware routers in the path. If any Z* 2513 (and/or Y) is located on the same OMNI link segment as X, the 2514 multicast data traffic sent to X directly using SPAN/INET 2515 encapsulation instead of via a Bridge. 2517 Following the initial Join/Prune and NS/NA messaging, X maintains an 2518 asymmetric neighbor cache entry for each S the same as if X was 2519 sending unicast data traffic to S. In particular, X performs 2520 additional NS/NA exchanges to keep the neighbor cache entry alive for 2521 up to t_periodic seconds [RFC7761]. If no new Joins are received 2522 within t_periodic seconds, X allows the neighbor cache entry to 2523 expire. Finally, if X receives any additional Join/Prune messages 2524 for (S,G) it forwards the messages to each Y and Z* in the neighbor 2525 cache entry over the spanning tree. 2527 At some later time, Client C that holds an MNP for source S may 2528 depart from a first Proxy Z1 and/or connect via a new Proxy Z2. In 2529 that case, Y sends an unsolicited NA message to X the same as 2530 specified for unicast mobility in Section 3.16. When X receives the 2531 unsolicited NA message, it updates its asymmetric neighbor cache 2532 entry for the LLA for source S and sends new Join messages to any new 2533 Proxys Z2. There is no requirement to send any Prune messages to old 2534 Proxys Z1 since source S will no longer source any multicast data 2535 traffic via Z1. Instead, the multicast state for (S,G) in Proxy Z1 2536 will soon time out since no new Joins will arrive. 2538 After some later time, C may move to a new Server Y2 and depart from 2539 old Sever Y1. In that case, Y1 sends Join messages for any of C's 2540 active (S,G) groups to Y2 while including its own LLA as the source 2541 address. This causes Y2 to include Y1 in the multicast forwarding 2542 tree during the interim time that Y1's symmetric neighbor cache entry 2543 for C is in the DEPARTED state. At the same time, Y1 sends an 2544 unsolicited NA message to X with an OMNI option with ifIndex in the 2545 header set to 0 and a release indication to cause X to release its 2546 asymmetric neighbor cache entry. X then sends a new Join message to 2547 S via the spanning tree and re-initiates route optimization the same 2548 as if it were receiving a fresh Join message from a node on a 2549 downstream link. 2551 3.17.2. Any-Source Multicast (ASM) 2553 When an ROS X acting as a PIM router receives a Join/Prune from a 2554 node on its downstream interfaces containing one or more (*,G) pairs, 2555 it updates its Multicast Routing Information Base (MRIB) accordingly. 2556 X then forwards a copy of the message to the Rendezvous Point (RP) R 2557 for each G over the spanning tree. X uses its own LLA as the source 2558 address and ALL-PIM-ROUTERS as the destination address, then 2559 encapsulates each message in a SPAN header with source address set to 2560 the ULA of X and destination address set to R, then sends the message 2561 into the spanning tree. At the same time, if the message was a Join 2562 X initiates NS/NA route optimization the same as for the SSM case 2563 discussed in Section 3.17.1. 2565 For each source S that sends multicast traffic to group G via R, the 2566 Proxy/Server Z* for the Client that aggregates S encapsulates the 2567 packets in PIM Register messages and forwards them to R via the 2568 spanning tree, which may then elect to send a PIM Join to Z*. This 2569 will result in an (S,G) tree rooted at Z* with R as the next hop so 2570 that R will begin to receive two copies of the packet; one native 2571 copy from the (S, G) tree and a second copy from the pre-existing (*, 2572 G) tree that still uses PIM Register encapsulation. R can then issue 2573 a PIM Register-stop message to suppress the Register-encapsulated 2574 stream. At some later time, if C moves to a new Proxy/Server Z*, it 2575 resumes sending packets via PIM Register encapsulation via the new 2576 Z*. 2578 At the same time, as multicast listeners discover individual S's for 2579 a given G, they can initiate an (S,G) Join for each S under the same 2580 procedures discussed in Section 3.17.1. Once the (S,G) tree is 2581 established, the listeners can send (S, G) Prune messages to R so 2582 that multicast packets for group G sourced by S will only be 2583 delivered via the (S, G) tree and not from the (*, G) tree rooted at 2584 R. All mobility considerations discussed for SSM apply. 2586 3.17.3. Bi-Directional PIM (BIDIR-PIM) 2588 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 2589 approach to ASM that treats the Rendezvous Point (RP) as a Designated 2590 Forwarder (DF). Further considerations for BIDIR-PIM are out of 2591 scope. 2593 3.18. Operation over Multiple OMNI Links 2595 An AERO Client can connect to multiple OMNI links the same as for any 2596 data link service. In that case, the Client maintains a distinct 2597 OMNI interface for each link, e.g., 'omni0' for the first link, 2598 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 2599 would include its own distinct set of Bridges, Servers and Proxys, 2600 thereby providing redundancy in case of failures. 2602 Each OMNI link could utilize the same or different ANET connections. 2603 The links can be distinguished at the link-layer via the SRT prefix 2604 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 2605 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 2606 MSPs on each link. This gives rise to the opportunity for supporting 2607 multiple redundant networked paths, with each VLAN distinguished by a 2608 different SRT "color" (see: Section 3.2.5). 2610 The Client's IP layer can select the outgoing OMNI interface 2611 appropriate for a given traffic profile while (in the reverse 2612 direction) correspondent nodes must have some way of steering their 2613 packets destined to a target via the correct OMNI link. 2615 In a first alternative, if each OMNI link services different MSPs, 2616 then the Client can receive a distinct MNP from each of the links. 2617 IP routing will therefore assure that the correct Red/Green/Blue/etc. 2618 network is used for both outbound and inbound traffic. This can be 2619 accomplished using existing technologies and approaches, and without 2620 requiring any special supporting code in correspondent nodes or 2621 Bridges. 2623 In a second alternative, if each OMNI link services the same MSP(s) 2624 then each link could assign a distinct "OMNI link Anycast" address 2625 that is configured by all Bridges on the link. Correspondent nodes 2626 can then perform Segment Routing to select the correct SRT, which 2627 will then direct the packet over multiple hops to the target. 2629 3.19. DNS Considerations 2631 AERO Client MNs and INET correspondent nodes consult the Domain Name 2632 System (DNS) the same as for any Internetworking node. When 2633 correspondent nodes and Client MNs use different IP protocol versions 2634 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 2635 A records for IPv4 address mappings to MNs which must then be 2636 populated in Relay NAT64 mapping caches. In that way, an IPv4 2637 correspondent node can send packets to the IPv4 address mapping of 2638 the target MN, and the Relay will translate the IPv4 header and 2639 destination address into an IPv6 header and IPv6 destination address 2640 of the MN. 2642 When an AERO Client registers with an AERO Server, the Server can 2643 return the address(es) of DNS servers in RDNSS options [RFC6106]. 2644 The DNS server provides the IP addresses of other MNs and 2645 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 2647 3.20. Transition Considerations 2649 SPAN encapsulation ensures that dissimilar INET partitions can be 2650 joined into a single unified OMNI link, even though the partitions 2651 themselves may have differing protocol versions and/or incompatible 2652 addressing plans. However, a commonality can be achieved by 2653 incrementally distributing globally routable (i.e., native) IP 2654 prefixes to eventually reach all nodes (both mobile and fixed) in all 2655 OMNI link segments. This can be accomplished by incrementally 2656 deploying AERO Relays on each INET partition, with each Relay 2657 distributing its MNPs and/or discovering non-MNP prefixes on its INET 2658 links. 2660 This gives rise to the opportunity to eventually distribute native IP 2661 addresses to all nodes, and to present a unified OMNI link view even 2662 if the INET partitions remain in their current protocol and 2663 addressing plans. In that way, the OMNI link can serve the dual 2664 purpose of providing a mobility/multilink service and a transition 2665 service. Or, if an INET partition is transitioned to a native IP 2666 protocol version and addressing scheme that is compatible with the 2667 OMNI link MNP-based addressing scheme, the partition and OMNI link 2668 can be joined by Relays. 2670 Relays that connect INETs/EUNs with dissimilar IP protocol versions 2671 may need to employ a network address and protocol translation 2672 function such as NAT64[RFC6146]. 2674 3.21. Detecting and Reacting to Server and Bridge Failures 2676 In environments where rapid failure recovery is required, Servers and 2677 Bridges SHOULD use Bidirectional Forwarding Detection (BFD) 2678 [RFC5880]. Nodes that use BFD can quickly detect and react to 2679 failures so that cached information is re-established through 2680 alternate nodes. BFD control messaging is carried only over well- 2681 connected ground domain networks (i.e., and not low-end radio links) 2682 and can therefore be tuned for rapid response. 2684 Servers and Bridges maintain BFD sessions in parallel with their BGP 2685 peerings. If a Server or Bridge fails, BGP peers will quickly re- 2686 establish routes through alternate paths the same as for common BGP 2687 deployments. Similarly, Proxys maintain BFD sessions with their 2688 associated Bridges even though they do not establish BGP peerings 2689 with them. 2691 Proxys SHOULD use proactive NUD for Servers for which there are 2692 currently active ANET Clients in a manner that parallels BFD, i.e., 2693 by sending unicast NS messages in rapid succession to receive 2694 solicited NA messages. When the Proxy is also sending RS messages on 2695 behalf of ANET Clients, the RS/RA messaging can be considered as 2696 equivalent hints of forward progress. This means that the Proxy need 2697 not also send a periodic NS if it has already sent an RS within the 2698 same period. If a Server fails, the Proxy will cease to receive 2699 advertisements and can quickly inform Clients of the outage by 2700 sending multicast RA messages on the ANET interface. 2702 The Proxy sends multicast RA messages with source address set to the 2703 Server's address, destination address set to (link-local) All-Nodes 2704 multicast, and Router Lifetime set to 0. The Proxy SHOULD send 2705 MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small delays 2706 [RFC4861]. Any Clients on the ANET interface that have been using 2707 the (now defunct) Server will receive the RA messages and associate 2708 with a new Server. 2710 3.22. AERO Clients on the Open Internet 2712 AERO Clients that connect to the open Internet via INET interfaces 2713 can establish a VPN or direct link to securely connect to a Server in 2714 a "tethered" arrangement with all of the Client's traffic transiting 2715 the Server. Alternatively, the Client can associate with an INET 2716 Server using UDP/IP encapsulation and asymmetric securing services as 2717 discussed in the following sections. 2719 When a Client's OMNI interface enables an INET underlying interface, 2720 it first determines whether the interface is likely to be behind a 2721 NAT. For IPv4, the Client assumes it is on the open Internet if the 2722 INET address is not a special-use IPv4 address per [RFC3330]. 2723 Similarly for IPv6, the Client assumes it is on the open Internet if 2724 the INET address is not a link-local [RFC4291] or unique-local 2725 [RFC4193] IPv6 address. 2727 The Client then prepares a UDP/IP-encapsulated RS message with IPv6 2728 source address set to its LLA, with IPv6 destination set to (link- 2729 local) All-Routers multicast and with an OMNI option with underlying 2730 interface parameters. If the Client believes that it is on the open 2731 Internet, it SHOULD include Interface Attributes with the L2ADDR used 2732 for INET encapsulation (otherwise, it MAY omit L2ADDR). If the 2733 underlying address is IPv4, the Client includes the Port Number and 2734 IPv4 address written in obfuscated form [RFC4380] as discussed in 2735 Section 3.3. If the underlying interface address is IPv6, the Client 2736 instead includes the Port Number and IPv6 address in obfuscated form. 2737 The Client finally includes an Authentication option per [RFC4380] to 2738 provide message authentication, sets the UDP/IP source to its INET 2739 address and UDP port, sets the UDP/IP destination to the Server's 2740 INET address and the AERO service port number (8060), then sends the 2741 message to the Server. 2743 When the Server receives the RS, it authenticates the message and 2744 registers the Client's MNP and INET interface information according 2745 to the OMNI option parameters. If the RS message includes an L2ADDR 2746 in the OMNI option, the Server compares the encapsulation IP address 2747 and UDP port number with the (unobfuscated) values. If the values 2748 are the same, the Server caches the Client's information as "INET" 2749 addresses meaning that the Client is likely to accept direct messages 2750 without requiring NAT traversal exchanges. If the values are 2751 different (or, if the OMNI option did not include an L2ADDR) the 2752 Server instead caches the Client's information as "NAT" addresses 2753 meaning that NAT traversal exchanges may be necessary. 2755 The Server then returns an RA message with IPv6 source and 2756 destination set corresponding to the addresses in the RS, and with an 2757 Authentication option per [RFC4380]. For IPv4, the Server also 2758 includes an Origin option per [RFC4380] with the mapped and 2759 obfuscated Port Number and IPv4 address observed in the encapsulation 2760 headers. For IPv6, the Server instead includes an IPv6 Origin option 2761 per Figure 6 with the mapped and obfuscated observed Port Number and 2762 IPv6 address (note that the value 0x02 in the second octet 2763 differentiates from other [RFC4380] option types). 2765 +--------+--------+-----------------+ 2766 | 0x00 | 0x02 | Origin port # | 2767 +--------+--------+-----------------+ 2768 ~ Origin IPv6 address ~ 2769 +-----------------------------------+ 2771 Figure 6: IPv6 Origin Option 2773 When the Client receives the RA message, it compares the mapped Port 2774 Number and IP address from the Origin option with its own address. 2775 If the addresses are the same, the Client assumes the open Internet / 2776 Cone NAT principle; if the addresses are different, the Client 2777 instead assumes that further qualification procedures are necessary 2778 to detect the type of NAT and proceeds according to standard 2779 [RFC4380] procedures. 2781 After the Client has registered its INET interfaces in such RS/RA 2782 exchanges it sends periodic RS messages to receive fresh RA messages 2783 before the Router Lifetime received on each INET interface expires. 2784 The Client also maintains default routes via its Servers, i.e., the 2785 same as described in earlier sections. 2787 When the Client sends messages to target IP addresses, it also 2788 invokes route optimization per Section 3.14 using IPv6 ND address 2789 resolution messaging. The Client sends the NS(AR) message to the 2790 Server wrapped in a UDP/IP header with an Authentication option with 2791 the NS source address set to the Client's LLA and destination address 2792 set to the target solicited node multicast address. The Server 2793 authenticates the message and sends a corresponding NS(AR) message 2794 over the spanning tree the same as if it were the ROS, but with the 2795 SPAN source address set to the Server's ULA and destination set to 2796 the ULA of the target. When the ROR receives the NS(AR), it adds the 2797 Server's ULA and Client's LLA to the target's Report List, and 2798 returns an NA with OMNI option information for the target. The 2799 Server then returns a UDP/IP encapsulated NA message with an 2800 Authentication option to the Client. 2802 Following route optimization for targets in the same OMNI link 2803 segment, if the target's L2ADDR is on the open INET, the Client 2804 forwards data packets directly to the target INET address. If the 2805 target is behind a NAT, the Client first establishes NAT state for 2806 the L2ADDR using the "bubble" mechanisms specified in 2807 [RFC6081][RFC4380]. The Client continues to send data packets via 2808 its Server until NAT state is populated, then begins forwarding 2809 packets via the direct path through the NAT to the target. For 2810 targets in different OMNI link segments, the Client inserts an SRH 2811 and forwards data packets to the Bridge that returned the NA message. 2813 The ROR may return uNAs via the Server if the target moves, and the 2814 Server will send corresponding Authentication-protected uNAs to the 2815 Client. The Client can also send "loopback" NS(NUD) messages to test 2816 forward path reachability even though there is no security 2817 association between the Client and the target. 2819 The Client sends UDP/IP encapsulated IPv6 packets no larger than 1280 2820 bytes in one piece. In order to accommodate larger IPv6 packets (up 2821 to the OMNI interface MTU), the Client inserts a SPAN header with 2822 source set to its own ULA and destination set to the ULA of the 2823 target and uses IPv6 fragmentation according to Section 3.9. The 2824 Client then encapsulates each fragment in a UDP/IP header and sends 2825 the fragments to the next hop. 2827 3.22.1. Use of SEND and CGA 2829 In some environments, use of the [RFC4380] Authentication option 2830 alone may be sufficient for assuring IPv6 ND message authentication 2831 between Clients and Servers. When additional protection is 2832 necessary, nodes should employ SEcure Neighbor Discovery (SEND) 2833 [RFC3971] with Cryptographically-Generated Addresses (CGA) [RFC3972]. 2835 When SEND/CGA are used, the Client prepares RS messages with its 2836 link-local CGA as the IPv6 source and (link-local) All-Routers 2837 multicast as the IPv6 Destination, includes any SEND options and 2838 wraps the message in a SPAN header. The Client sets the SPAN source 2839 address to its own ULA and sets the SPAN destination address to 2840 (site-local) All-Routers multicast. The Client then wraps the RS 2841 message in UDP/IP headers according to [RFC4380] and sends the 2842 message to the Server. 2844 When the Server receives the message, it first verifies the 2845 Authentication option (if present) then uses the SPAN source address 2846 to determine the MNP of the Client. The Server then processes the 2847 SEND options to authenticate the RS message and prepares an RA 2848 message response. The Server prepares the RA with its own link-local 2849 CGA as the IPv6 source and the CGA of the Client as the IPv6 2850 destination, includes any SEND options and wraps the message in a 2851 SPAN header. The Server sets the SPAN source address to its own ULA 2852 and sets the SPAN destination address to the Client's ULA. The 2853 Server then wraps the RA message in UDP/IP headers according to 2854 [RFC4380] and sends the message to the Client. Thereafter, the 2855 Client/Server send additional RS/RA messages to maintain their 2856 association and any NAT state. 2858 The Client and Server also may exchange NS/NA messages using their 2859 own CGA as the source and with SPAN encapsulation as above. When a 2860 Client sends an NS(AR), it sets the IPv6 source to its CGA and sets 2861 the IPv6 destination to the Solicited-Node Multicast address of the 2862 target. The Client then wraps the message in a SPAN header with its 2863 own ULA as the source and the ULA of the target as the destination 2864 and sends it to the Server. The Server authenticates the message, 2865 then changes the IPv6 source address to the Client's LLA, removes the 2866 SEND options, and sends a corresponding NS(AR) into the spanning 2867 tree. When the Server receives the corresponding SPAN-encapsulated 2868 NA, it changes the IPv6 destination address to the Client's CGA, 2869 inserts SEND options, then wraps the message in UDP/IP headers and 2870 sends it to the Client. 2872 When a Client sends a uNA, it sets the IPv6 source address to its own 2873 CGA and sets the IPv6 destination address to (link-local) All-Nodes 2874 multicast, includes SEND options, wraps the message in SPAN and UDP/ 2875 IP headers and sends the message to the Server. The Server 2876 authenticates the message, then changes the IPv6 address to the 2877 Client's LLA, removes the SEND options and forwards the message the 2878 same as discussed in Section 3.16.1. In the reverse direction, when 2879 the Server forwards a uNA to the Client, it changes the IPv6 address 2880 to its own CGA and inserts SEND options then forwards the message to 2881 the Client. 2883 When a Client sends an NS(NUD), it sets both the IPv6 source and 2884 destination address to its own LLA, wraps the message in a SPAN 2885 header and UDP/IP headers, then sends the message directly to the 2886 peer which will loop the message back. In this case alone, the 2887 Client does not use the Server as a trust broker for forwarding the 2888 ND message. 2890 3.23. Time-Varying MNPs 2892 In some use cases, it is desirable, beneficial and efficient for the 2893 Client to receive a constant MNP that travels with the Client 2894 wherever it moves. For example, this would allow air traffic 2895 controllers to easily track aircraft, etc. In other cases, however 2896 (e.g., intelligent transportation systems), the MN may be willing to 2897 sacrifice a modicum of efficiency in order to have time-varying MNPs 2898 that can be changed every so often to defeat adversarial tracking. 2900 The DHCPv6 service offers a way for Clients that desire time-varying 2901 MNPs to obtain short-lived prefixes (e.g., on the order of a small 2902 number of minutes). In that case, the identity of the Client would 2903 not be bound to the MNP but rather the Client's identity would be 2904 bound to the DHCPv6 Device Unique Identifier (DUID) and used as the 2905 seed for Prefix Delegation. The Client would then be obligated to 2906 renumber its internal networks whenever its MNP (and therefore also 2907 its LLA) changes. This should not present a challenge for Clients 2908 with automated network renumbering services, however presents limits 2909 for the durations of ongoing sessions that would prefer to use a 2910 constant address. 2912 4. Implementation Status 2914 An early AERO implementation based on OpenVPN (https://openvpn.net/) 2915 was announced on the v6ops mailing list on January 10, 2018 and an 2916 initial public release of the AERO proof-of-concept source code was 2917 announced on the intarea mailing list on August 21, 2015. 2919 AERO Release-3.0.1 was tagged on September 11, 2020, and is 2920 undergoing internal testing. Additional releases expected Q42020, 2921 with first public release expected before year-end. 2923 5. IANA Considerations 2925 The IANA has assigned a 4-octet Private Enterprise Number "45282" for 2926 AERO in the "enterprise-numbers" registry. 2928 The IANA has assigned the UDP port number "8060" for an earlier 2929 experimental version of AERO [RFC6706]. This document obsoletes 2930 [RFC6706] and claims the UDP port number "8060" for all future use. 2932 The IANA is instructed to assign a new type value TBD in the Segment 2933 Routing Header TLV registry [RFC8754]. 2935 No further IANA actions are required. 2937 6. Security Considerations 2939 AERO Bridges configure secured tunnels with AERO Servers, Realys and 2940 Proxys within their local OMNI link segments. Applicable secured 2941 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 2942 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 2943 segments in turn configure secured tunnels for their neighboring AERO 2944 Bridges in a spanning tree topology. Therefore, control messages 2945 exchanged between any pair of OMNI link neighbors on the spanning 2946 tree are already secured. 2948 AERO Servers, Relays and Proxys targeted by a route optimization may 2949 also receive data packets directly from arbitrary nodes in INET 2950 partitions instead of via the spanning tree. For INET partitions 2951 that apply effective ingress filtering to defeat source address 2952 spoofing, the simple data origin authentication procedures in 2953 Section 3.8 can be applied. 2955 For INET partitions that require strong security in the data plane, 2956 two options for securing communications include 1) disable route 2957 optimization so that all traffic is conveyed over secured tunnels, or 2958 2) enable on-demand secure tunnel creation between INET partition 2959 neighbors. Option 1) would result in longer routes than necessary 2960 and traffic concentration on critical infrastructure elements. 2961 Option 2) could be coordinated by establishing a secured tunnel on- 2962 demand instead of performing an NS/NA exchange in the route 2963 optimization procedures. Procedures for establishing on-demand 2964 secured tunnels are out of scope. 2966 AERO Clients that connect to secured ANETs need not apply security to 2967 their ND messages, since the messages will be intercepted by a 2968 perimeter Proxy that applies security on its INET-facing interface as 2969 part of the spanning tree (see above). AERO Clients connected to the 2970 open INET can use symmetric network and/or transport layer security 2971 services such as VPNs or can by some other means establish a direct 2972 link. When a VPN or direct link may be impractical, however, an 2973 asymmetric security service such as SEcure Neighbor Discovery (SEND) 2974 [RFC3971] with Cryptographically Generated Addresses (CGAs) [RFC3972] 2975 and/or the Authentication option [RFC4380] can be applied. 2977 Application endpoints SHOULD use application-layer security services 2978 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 2979 protection as for critical secured Internet services. AERO Clients 2980 that require host-based VPN services SHOULD use symmetric network 2981 and/or transport layer security services such as IPsec, TLS/SSL, 2982 DTLS, etc. AERO Proxys and Servers can also provide a network-based 2983 VPN service on behalf of the Client, e.g., if the Client is located 2984 within a secured enclave and cannot establish a VPN on its own 2985 behalf. 2987 AERO Servers and Bridges present targets for traffic amplification 2988 Denial of Service (DoS) attacks. This concern is no different than 2989 for widely-deployed VPN security gateways in the Internet, where 2990 attackers could send spoofed packets to the gateways at high data 2991 rates. This can be mitigated by connecting Servers and Bridges over 2992 dedicated links with no connections to the Internet and/or when 2993 connections to the Internet are only permitted through well-managed 2994 firewalls. Traffic amplification DoS attacks can also target an AERO 2995 Client's low data rate links. This is a concern not only for Clients 2996 located on the open Internet but also for Clients in secured 2997 enclaves. AERO Servers and Proxys can institute rate limits that 2998 protect Clients from receiving packet floods that could DoS low data 2999 rate links. 3001 AERO Relays must implement ingress filtering to avoid a spoofing 3002 attack in which spurious messages with ULA addresses are injected 3003 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3004 that their connectivity is not used by unauthorized nodes on their 3005 EUNs to gain access to a protected network, i.e., AERO Clients that 3006 act as routers MUST NOT provide routing services for unauthorized 3007 nodes. (This concern is no different than for ordinary hosts that 3008 receive an IP address delegation but then "share" the address with 3009 other nodes via some form of Internet connection sharing such as 3010 tethering.) 3012 The MAP list MUST be well-managed and secured from unauthorized 3013 tampering, even though the list contains only public information. 3014 The MAP list can be conveyed to the Client in a similar fashion as in 3015 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3016 upload of a static file, DNS lookups, etc.). 3018 Although public domain and commercial SEND implementations exist, 3019 concerns regarding the strength of the cryptographic hash algorithm 3020 have been documented [RFC6273] [RFC4982]. 3022 SRH authentication facilities are specified in [RFC8754]. 3024 Security considerations for accepting link-layer ICMP messages and 3025 reflected packets are discussed throughout the document. 3027 Security considerations for IPv6 fragmentation and reassembly are 3028 discussed in [I-D.templin-6man-omni-interface]. 3030 7. Acknowledgements 3032 Discussions in the IETF, aviation standards communities and private 3033 exchanges helped shape some of the concepts in this work. 3034 Individuals who contributed insights include Mikael Abrahamsson, Mark 3035 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3036 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3037 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3038 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3039 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3040 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3041 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3042 Wood and James Woodyatt. Members of the IESG also provided valuable 3043 input during their review process that greatly improved the document. 3044 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3045 for their shepherding guidance during the publication of the AERO 3046 first edition. 3048 This work has further been encouraged and supported by Boeing 3049 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3050 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3051 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3052 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3053 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3054 Gene MacLean III, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3055 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3056 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3057 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3058 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3059 Benson, Katie Tran and Eric Yeh are especially acknowledged for 3060 implementing the AERO functions as extensions to the public domain 3061 OpenVPN distribution. 3063 Earlier works on NBMA tunneling approaches are found in 3064 [RFC2529][RFC5214][RFC5569]. 3066 Many of the constructs presented in this second edition of AERO are 3067 based on the author's earlier works, including: 3069 o The Internet Routing Overlay Network (IRON) 3070 [RFC6179][I-D.templin-ironbis] 3072 o Virtual Enterprise Traversal (VET) 3073 [RFC5558][I-D.templin-intarea-vet] 3075 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3076 [RFC5320][I-D.templin-intarea-seal] 3078 o AERO, First Edition [RFC6706] 3080 Note that these works cite numerous earlier efforts that are not also 3081 cited here due to space limitations. The authors of those earlier 3082 works are acknowledged for their insights. 3084 This work is aligned with the NASA Safe Autonomous Systems Operation 3085 (SASO) program under NASA contract number NNA16BD84C. 3087 This work is aligned with the FAA as per the SE2025 contract number 3088 DTFAWA-15-D-00030. 3090 This work is aligned with the Boeing Commercial Airplanes (BCA) 3091 Internet of Things (IoT) and autonomy programs. 3093 This work is aligned with the Boeing Information Technology (BIT) 3094 MobileNet program. 3096 8. References 3098 8.1. Normative References 3100 [I-D.templin-6man-omni-interface] 3101 Templin, F. and T. Whyman, "Transmission of IPv6 Packets 3102 over Overlay Multilink Network (OMNI) Interfaces", draft- 3103 templin-6man-omni-interface-31 (work in progress), 3104 September 2020. 3106 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3107 DOI 10.17487/RFC0791, September 1981, 3108 . 3110 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3111 RFC 792, DOI 10.17487/RFC0792, September 1981, 3112 . 3114 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3115 Requirement Levels", BCP 14, RFC 2119, 3116 DOI 10.17487/RFC2119, March 1997, 3117 . 3119 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3120 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3121 December 1998, . 3123 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3124 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3125 DOI 10.17487/RFC3971, March 2005, 3126 . 3128 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3129 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3130 . 3132 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3133 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3134 November 2005, . 3136 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3137 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3138 . 3140 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3141 Network Address Translations (NATs)", RFC 4380, 3142 DOI 10.17487/RFC4380, February 2006, 3143 . 3145 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3146 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3147 DOI 10.17487/RFC4861, September 2007, 3148 . 3150 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3151 Address Autoconfiguration", RFC 4862, 3152 DOI 10.17487/RFC4862, September 2007, 3153 . 3155 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3156 DOI 10.17487/RFC6081, January 2011, 3157 . 3159 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3160 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3161 May 2017, . 3163 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3164 (IPv6) Specification", STD 86, RFC 8200, 3165 DOI 10.17487/RFC8200, July 2017, 3166 . 3168 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3169 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3170 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3171 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3172 . 3174 8.2. Informative References 3176 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3177 2016. 3179 [I-D.bonica-6man-comp-rtg-hdr] 3180 Bonica, R., Kamite, Y., Niwa, T., Alston, A., and L. 3181 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3182 bonica-6man-comp-rtg-hdr-22 (work in progress), May 2020. 3184 [I-D.bonica-6man-crh-helper-opt] 3185 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3186 Routing Header (CRH) Helper Option", draft-bonica-6man- 3187 crh-helper-opt-01 (work in progress), May 2020. 3189 [I-D.ietf-intarea-frag-fragile] 3190 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 3191 and F. Gont, "IP Fragmentation Considered Fragile", draft- 3192 ietf-intarea-frag-fragile-17 (work in progress), September 3193 2019. 3195 [I-D.ietf-intarea-tunnels] 3196 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3197 Architecture", draft-ietf-intarea-tunnels-10 (work in 3198 progress), September 2019. 3200 [I-D.ietf-rtgwg-atn-bgp] 3201 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3202 Moreno, "A Simple BGP-based Mobile Routing System for the 3203 Aeronautical Telecommunications Network", draft-ietf- 3204 rtgwg-atn-bgp-06 (work in progress), June 2020. 3206 [I-D.templin-6man-dhcpv6-ndopt] 3207 Templin, F., "A Unified Stateful/Stateless Configuration 3208 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-10 3209 (work in progress), June 2020. 3211 [I-D.templin-intarea-seal] 3212 Templin, F., "The Subnetwork Encapsulation and Adaptation 3213 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3214 progress), January 2014. 3216 [I-D.templin-intarea-vet] 3217 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3218 templin-intarea-vet-40 (work in progress), May 2013. 3220 [I-D.templin-ironbis] 3221 Templin, F., "The Interior Routing Overlay Network 3222 (IRON)", draft-templin-ironbis-16 (work in progress), 3223 March 2014. 3225 [I-D.templin-v6ops-pdhost] 3226 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3227 Models", draft-templin-v6ops-pdhost-26 (work in progress), 3228 June 2020. 3230 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3232 [RFC1035] Mockapetris, P., "Domain names - implementation and 3233 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3234 November 1987, . 3236 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3237 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3238 . 3240 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3241 DOI 10.17487/RFC2003, October 1996, 3242 . 3244 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3245 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3246 . 3248 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3249 Domains without Explicit Tunnels", RFC 2529, 3250 DOI 10.17487/RFC2529, March 1999, 3251 . 3253 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3254 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3255 . 3257 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3258 of Explicit Congestion Notification (ECN) to IP", 3259 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3260 . 3262 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3263 DOI 10.17487/RFC3330, September 2002, 3264 . 3266 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3267 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3268 DOI 10.17487/RFC3810, June 2004, 3269 . 3271 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3272 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3273 January 2006, . 3275 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3276 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3277 DOI 10.17487/RFC4271, January 2006, 3278 . 3280 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3281 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3282 2006, . 3284 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3285 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3286 December 2005, . 3288 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3289 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3290 2006, . 3292 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3293 Control Message Protocol (ICMPv6) for the Internet 3294 Protocol Version 6 (IPv6) Specification", STD 89, 3295 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3296 . 3298 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3299 Protocol (LDAP): The Protocol", RFC 4511, 3300 DOI 10.17487/RFC4511, June 2006, 3301 . 3303 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3304 "Considerations for Internet Group Management Protocol 3305 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3306 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3307 . 3309 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3310 "Internet Group Management Protocol (IGMP) / Multicast 3311 Listener Discovery (MLD)-Based Multicast Forwarding 3312 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3313 August 2006, . 3315 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3316 Algorithms in Cryptographically Generated Addresses 3317 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3318 . 3320 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3321 "Bidirectional Protocol Independent Multicast (BIDIR- 3322 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3323 . 3325 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3326 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3327 DOI 10.17487/RFC5214, March 2008, 3328 . 3330 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3331 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3332 February 2010, . 3334 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3335 Route Optimization Requirements for Operational Use in 3336 Aeronautics and Space Exploration Mobile Networks", 3337 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3338 . 3340 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3341 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3342 . 3344 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3345 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3346 January 2010, . 3348 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3349 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3350 . 3352 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3353 "IPv6 Router Advertisement Options for DNS Configuration", 3354 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3355 . 3357 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3358 NAT64: Network Address and Protocol Translation from IPv6 3359 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3360 April 2011, . 3362 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3363 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3364 . 3366 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3367 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3368 DOI 10.17487/RFC6221, May 2011, 3369 . 3371 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3372 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3373 DOI 10.17487/RFC6273, June 2011, 3374 . 3376 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3377 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3378 January 2012, . 3380 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3381 for Equal Cost Multipath Routing and Link Aggregation in 3382 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3383 . 3385 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3386 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3387 . 3389 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3390 UDP Checksums for Tunneled Packets", RFC 6935, 3391 DOI 10.17487/RFC6935, April 2013, 3392 . 3394 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3395 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3396 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3397 . 3399 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3400 Korhonen, "Requirements for Distributed Mobility 3401 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3402 . 3404 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3405 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3406 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3407 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3408 2016, . 3410 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3411 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3412 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3413 July 2018, . 3415 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3416 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3417 . 3419 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 3420 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3421 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 3422 . 3424 [WG] Wireguard, "Wireguard, https://www.wireguard.com", August 3425 2020. 3427 Appendix A. Non-Normative Considerations 3429 AERO can be applied to a multitude of Internetworking scenarios, with 3430 each having its own adaptations. The following considerations are 3431 provided as non-normative guidance: 3433 A.1. Implementation Strategies for Route Optimization 3435 Route optimization as discussed in Section 3.14 results in the route 3436 optimization source (ROS) creating an asymmetric neighbor cache entry 3437 for the target neighbor. The neighbor cache entry is maintained for 3438 at most ReachableTime seconds and then deleted unless updated. In 3439 order to refresh the neighbor cache entry lifetime before the 3440 ReachableTime timer expires, the specification requires 3441 implementations to issue a new NS/NA exchange to reset ReachableTime 3442 while data packets are still flowing. However, the decision of when 3443 to initiate a new NS/NA exchange and to perpetuate the process is 3444 left as an implementation detail. 3446 One possible strategy may be to monitor the neighbor cache entry 3447 watching for data packets for (ReachableTime - 5) seconds. If any 3448 data packets have been sent to the neighbor within this timeframe, 3449 then send an NS to receive a new NA. If no data packets have been 3450 sent, wait for 5 additional seconds and send an immediate NS if any 3451 data packets are sent within this "expiration pending" 5 second 3452 window. If no additional data packets are sent within the 5 second 3453 window, delete the neighbor cache entry. 3455 The monitoring of the neighbor data packet traffic therefore becomes 3456 an asymmetric ongoing process during the neighbor cache entry 3457 lifetime. If the neighbor cache entry expires, future data packets 3458 will trigger a new NS/NA exchange while the packets themselves are 3459 delivered over a longer path until route optimization state is re- 3460 established. 3462 A.2. Implicit Mobility Management 3464 OMNI interface neighbors MAY provide a configuration option that 3465 allows them to perform implicit mobility management in which no ND 3466 messaging is used. In that case, the Client only transmits packets 3467 over a single interface at a time, and the neighbor always observes 3468 packets arriving from the Client from the same link-layer source 3469 address. 3471 If the Client's underlying interface address changes (either due to a 3472 readdressing of the original interface or switching to a new 3473 interface) the neighbor immediately updates the neighbor cache entry 3474 for the Client and begins accepting and sending packets according to 3475 the Client's new address. This implicit mobility method applies to 3476 use cases such as cellphones with both WiFi and Cellular interfaces 3477 where only one of the interfaces is active at a given time, and the 3478 Client automatically switches over to the backup interface if the 3479 primary interface fails. 3481 A.3. Direct Underlying Interfaces 3483 When a Client's OMNI interface is configured over a Direct interface, 3484 the neighbor at the other end of the Direct link can receive packets 3485 without any encapsulation. In that case, the Client sends packets 3486 over the Direct link according to QoS preferences. If the Direct 3487 interface has the highest QoS preference, then the Client's IP 3488 packets are transmitted directly to the peer without going through an 3489 ANET/INET. If other interfaces have higher QoS preferences, then the 3490 Client's IP packets are transmitted via a different interface, which 3491 may result in the inclusion of Proxys, Servers and Bridges in the 3492 communications path. Direct interfaces must be tested periodically 3493 for reachability, e.g., via NUD. 3495 A.4. AERO Critical Infrastructure Considerations 3497 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 3498 IP routers or virtual machines in the cloud. Bridges must be 3499 provisioned, supported and managed by the INET administrative 3500 authority, and connected to the Bridges of other INETs via inter- 3501 domain peerings. Cost for purchasing, configuring and managing 3502 Bridges is nominal even for very large OMNI links. 3504 AERO Servers can be standard dedicated server platforms, but most 3505 often will be deployed as virtual machines in the cloud. The only 3506 requirements for Servers are that they can run the AERO user-level 3507 code and have at least one network interface connection to the INET. 3508 As with Bridges, Servers must be provisioned, supported and managed 3509 by the INET administrative authority. Cost for purchasing, 3510 configuring and managing Servers is nominal especially for virtual 3511 Servers hosted in the cloud. 3513 AERO Proxys are most often standard dedicated server platforms with 3514 one network interface connected to the ANET and a second interface 3515 connected to an INET. As with Servers, the only requirements are 3516 that they can run the AERO user-level code and have at least one 3517 interface connection to the INET. Proxys must be provisioned, 3518 supported and managed by the ANET administrative authority. Cost for 3519 purchasing, configuring and managing Proxys is nominal, and borne by 3520 the ANET administrative authority. 3522 AERO Relays can be any dedicated server or COTS router platform 3523 connected to INETs and/or EUNs. The Relay connects to the OMNI link 3524 and engages in eBGP peering with one or more Bridges as a stub AS. 3525 The Relay then injects its MNPs and/or non-MNP prefixes into the BGP 3526 routing system, and provisions the prefixes to its downstream- 3527 attached networks. The Relay can perform ROS/ROR services the same 3528 as for any Server, and can route between the MNP and non-MNP address 3529 spaces. 3531 A.5. AERO Server Failure Implications 3533 AERO Servers may appear as a single point of failure in the 3534 architecture, but such is not the case since all Servers on the link 3535 provide identical services and loss of a Server does not imply 3536 immediate and/or comprehensive communication failures. Although 3537 Clients typically associate with a single Server at a time, Server 3538 failure is quickly detected and conveyed by Bidirectional Forward 3539 Detection (BFD) and/or proactive NUD allowing Clients to migrate to 3540 new Servers. 3542 If a Server fails, ongoing packet forwarding to Clients will continue 3543 by virtue of the asymmetric neighbor cache entries that have already 3544 been established in route optimization sources (ROSs). If a Client 3545 also experiences mobility events at roughly the same time the Server 3546 fails, unsolicited NA messages may be lost but proxy neighbor cache 3547 entries in the DEPARTED state will ensure that packet forwarding to 3548 the Client's new locations will continue for up to DepartTime 3549 seconds. 3551 If a Client is left without a Server for an extended timeframe (e.g., 3552 greater than ReachableTime seconds) then existing asymmetric neighbor 3553 cache entries will eventually expire and both ongoing and new 3554 communications will fail. The original source will continue to 3555 retransmit until the Client has established a new Server 3556 relationship, after which time continuous communications will resume. 3558 Therefore, providing many Servers on the link with high availability 3559 profiles provides resilience against loss of individual Servers and 3560 assurance that Clients can establish new Server relationships quickly 3561 in event of a Server failure. 3563 A.6. AERO Client / Server Architecture 3565 The AERO architectural model is client / server in the control plane, 3566 with route optimization in the data plane. The same as for common 3567 Internet services, the AERO Client discovers the addresses of AERO 3568 Servers and selects one Server to connect to. The AERO service is 3569 analogous to common Internet services such as google.com, yahoo.com, 3570 cnn.com, etc. However, there is only one AERO service for the link 3571 and all Servers provide identical services. 3573 Common Internet services provide differing strategies for advertising 3574 server addresses to clients. The strategy is conveyed through the 3575 DNS resource records returned in response to name resolution queries. 3576 As of January 2020 Internet-based 'nslookup' services were used to 3577 determine the following: 3579 o When a client resolves the domainname "google.com", the DNS always 3580 returns one A record (i.e., an IPv4 address) and one AAAA record 3581 (i.e., an IPv6 address). The client receives the same addresses 3582 each time it resolves the domainname via the same DNS resolver, 3583 but may receive different addresses when it resolves the 3584 domainname via different DNS resolvers. But, in each case, 3585 exactly one A and one AAAA record are returned. 3587 o When a client resolves the domainname "ietf.org", the DNS always 3588 returns one A record and one AAAA record with the same addresses 3589 regardless of which DNS resolver is used. 3591 o When a client resolves the domainname "yahoo.com", the DNS always 3592 returns a list of 4 A records and 4 AAAA records. Each time the 3593 client resolves the domainname via the same DNS resolver, the same 3594 list of addresses are returned but in randomized order (i.e., 3595 consistent with a DNS round-robin strategy). But, interestingly, 3596 the same addresses are returned (albeit in randomized order) when 3597 the domainname is resolved via different DNS resolvers. 3599 o When a client resolves the domainname "amazon.com", the DNS always 3600 returns a list of 3 A records and no AAAA records. As with 3601 "yahoo.com", the same three A records are returned from any 3602 worldwide Internet connection point in randomized order. 3604 The above example strategies show differing approaches to Internet 3605 resilience and service distribution offered by major Internet 3606 services. The Google approach exposes only a single IPv4 and a 3607 single IPv6 address to clients. Clients can then select whichever IP 3608 protocol version offers the best response, but will always use the 3609 same IP address according to the current Internet connection point. 3610 This means that the IP address offered by the network must lead to a 3611 highly-available server and/or service distribution point. In other 3612 words, resilience is predicated on high availability within the 3613 network and with no client-initiated failovers expected (i.e., it is 3614 all-or-nothing from the client's perspective). However, Google does 3615 provide for worldwide distributed service distribution by virtue of 3616 the fact that each Internet connection point responds with a 3617 different IPv6 and IPv4 address. The IETF approach is like google 3618 (all-or-nothing from the client's perspective), but provides only a 3619 single IPv4 or IPv6 address on a worldwide basis. This means that 3620 the addresses must be made highly-available at the network level with 3621 no client failover possibility, and if there is any worldwide service 3622 distribution it would need to be conducted by a network element that 3623 is reached via the IP address acting as a service distribution point. 3625 In contrast to the Google and IETF philosophies, Yahoo and Amazon 3626 both provide clients with a (short) list of IP addresses with Yahoo 3627 providing both IP protocol versions and Amazon as IPv4-only. The 3628 order of the list is randomized with each name service query 3629 response, with the effect of round-robin load balancing for service 3630 distribution. With a short list of addresses, there is still 3631 expectation that the network will implement high availability for 3632 each address but in case any single address fails the client can 3633 switch over to using a different address. The balance then becomes 3634 one of function in the network vs function in the end system. 3636 The same implications observed for common highly-available services 3637 in the Internet apply also to the AERO client/server architecture. 3638 When an AERO Client connects to one or more ANETs, it discovers one 3639 or more AERO Server addresses through the mechanisms discussed in 3640 earlier sections. Each Server address presumably leads to a fault- 3641 tolerant clustering arrangement such as supported by Linux-HA, 3642 Extended Virtual Synchrony or Paxos. Such an arrangement has 3643 precedence in common Internet service deployments in lightweight 3644 virtual machines without requiring expensive hardware deployment. 3645 Similarly, common Internet service deployments set service IP 3646 addresses on service distribution points that may relay requests to 3647 many different servers. 3649 For AERO, the expectation is that a combination of the Google/IETF 3650 and Yahoo/Amazon philosophies would be employed. The AERO Client 3651 connects to different ANET access points and can receive 1-2 Server 3652 LLAs at each point. It then selects one AERO Server address, and 3653 engages in RS/RA exchanges with the same Server from all ANET 3654 connections. The Client remains with this Server unless or until the 3655 Server fails, in which case it can switch over to an alternate 3656 Server. The Client can likewise switch over to a different Server at 3657 any time if there is some reason for it to do so. So, the AERO 3658 expectation is for a balance of function in the network and end 3659 system, with fault tolerance and resilience at both levels. 3661 Appendix B. Change Log 3663 << RFC Editor - remove prior to publication >> 3665 Changes from draft-templin-intarea-6706bis-59 to draft-templin- 3666 intrea-6706bis-60: 3668 o Removed all references to S/TLLAO - all Interface Attributes are 3669 now maintained completely in the OMNI option. 3671 Changes from draft-templin-intarea-6706bis-58 to draft-templin- 3672 intrea-6706bis-59: 3674 o The term "Relay"used in older draft versions is now "Bridge". 3675 "Relay" now refers to what was formally called: "Gateway". 3677 o Fine-grained cleanup of Forwarding Algorithm; IPv6 ND message 3678 addressing; OMNI Prefix Lengths, etc. 3680 Changes from draft-templin-intarea-6706bis-54 to draft-templin- 3681 intrea-6706bis-55: 3683 o Updates on Segment Routing and S/TLLAO contents. 3685 o Various editorials and addressing cleanups. 3687 Changes from draft-templin-intarea-6706bis-52 to draft-templin- 3688 intrea-6706bis-53: 3690 o Normative reference to the OMNI spec, and remove portions that are 3691 already specified in OMNI. 3693 o Renamed "AERO interface/link" to "OMIN interface/link" throughout 3694 the document. 3696 o Truncated obsolete back section matter. 3698 Author's Address 3700 Fred L. Templin (editor) 3701 Boeing Research & Technology 3702 P.O. Box 3707 3703 Seattle, WA 98124 3704 USA 3706 Email: fltemplin@acm.org