idnits 2.17.1 draft-templin-intarea-6706bis-79.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 (December 16, 2020) is 1226 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) == Missing Reference: 'ULA' is mentioned on line 958, but not defined == Missing Reference: 'ULA0' is mentioned on line 964, but not defined == Missing Reference: 'ULA1' is mentioned on line 978, but not defined == Missing Reference: 'ULA2' is mentioned on line 968, but not defined == Missing Reference: 'ULA3' is mentioned on line 970, but not defined == Unused Reference: 'RFC3971' is defined on line 3265, but no explicit reference was found in the text == Unused Reference: 'RFC3972' is defined on line 3270, but no explicit reference was found in the text == Unused Reference: 'BGP' is defined on line 3318, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-comp-rtg-hdr' is defined on line 3321, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-crh-helper-opt' is defined on line 3327, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3332, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3349, but no explicit reference was found in the text == Unused Reference: 'RFC4982' is defined on line 3471, but no explicit reference was found in the text == Unused Reference: 'RFC6273' is defined on line 3527, but no explicit reference was found in the text == Outdated reference: A later version (-99) exists of draft-templin-6man-omni-interface-59 == Outdated reference: A later version (-31) exists of draft-bonica-6man-comp-rtg-hdr-23 == Outdated reference: A later version (-04) exists of draft-bonica-6man-crh-helper-opt-02 == 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 (~~), 23 warnings (==), 4 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, December 16, 2020 5 rfc6179, rfc6706 (if 6 approved) 7 Intended status: Standards Track 8 Expires: June 19, 2021 10 Asymmetric Extended Route Optimization (AERO) 11 draft-templin-intarea-6706bis-79 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 June 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 . . . . . . . . . . . . . . . . . . . . . 11 68 3.2. The AERO Service over OMNI Links . . . . . . . . . . . . 12 69 3.2.1. AERO/OMNI Reference Model . . . . . . . . . . . . . . 12 70 3.2.2. Link-Local Addresses (LLAs), Unique Local Addresses 71 (ULAs) and Global Unicast Addresses (GUA) . . . . . . 14 72 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 15 73 3.2.4. OMNI Link Encapsulation . . . . . . . . . . . . . . . 16 74 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 20 75 3.2.6. Segment Routing For OMNI Link Selection . . . . . . . 21 76 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 21 77 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 22 78 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 24 79 3.4.1. AERO Server/Relay Behavior . . . . . . . . . . . . . 24 80 3.4.2. AERO Proxy Behavior . . . . . . . . . . . . . . . . . 25 81 3.4.3. AERO Client Behavior . . . . . . . . . . . . . . . . 25 82 3.4.4. AERO Bridge Behavior . . . . . . . . . . . . . . . . 25 83 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 25 84 3.5.1. OMNI Neighbor Interface Attributes . . . . . . . . . 27 85 3.5.2. OMNI Neighbor Advertisement Message Flags . . . . . . 28 86 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 28 87 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 30 88 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 30 89 3.9. OMNI Adaptation Layer and OMNI Interface MTU . . . . . . 30 90 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 31 91 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 32 92 3.10.2. Proxy Forwarding Algorithm . . . . . . . . . . . . . 32 93 3.10.3. Server/Relay Forwarding Algorithm . . . . . . . . . 33 94 3.10.4. Bridge Forwarding Algorithm . . . . . . . . . . . . 34 96 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 35 97 3.12. AERO Router Discovery, Prefix Delegation and 98 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 38 99 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 38 100 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 38 101 3.12.3. AERO Server Behavior . . . . . . . . . . . . . . . . 40 102 3.13. The AERO Proxy . . . . . . . . . . . . . . . . . . . . . 43 103 3.13.1. Combined Proxy/Servers . . . . . . . . . . . . . . . 45 104 3.13.2. Detecting and Responding to Server Failures . . . . 46 105 3.13.3. Point-to-Multipoint Server Coordination . . . . . . 46 106 3.14. AERO Route Optimization / Address Resolution . . . . . . 47 107 3.14.1. Route Optimization Initiation . . . . . . . . . . . 48 108 3.14.2. Relaying the NS . . . . . . . . . . . . . . . . . . 48 109 3.14.3. Processing the NS and Sending the NA . . . . . . . . 48 110 3.14.4. Relaying the NA . . . . . . . . . . . . . . . . . . 49 111 3.14.5. Processing the NA . . . . . . . . . . . . . . . . . 50 112 3.14.6. Route Optimization Maintenance . . . . . . . . . . . 50 113 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 51 114 3.16. Mobility Management and Quality of Service (QoS) . . . . 52 115 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 53 116 3.16.2. Announcing Link-Layer Address and/or QoS Preference 117 Changes . . . . . . . . . . . . . . . . . . . . . . 54 118 3.16.3. Bringing New Links Into Service . . . . . . . . . . 55 119 3.16.4. Deactivating Existing Links . . . . . . . . . . . . 55 120 3.16.5. Moving Between Servers . . . . . . . . . . . . . . . 55 121 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 56 122 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 57 123 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 58 124 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 59 125 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 59 126 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 60 127 3.20. Transition Considerations . . . . . . . . . . . . . . . . 60 128 3.21. Detecting and Reacting to Server and Bridge Failures . . 61 129 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 62 130 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 64 131 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 65 132 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 65 133 6. Security Considerations . . . . . . . . . . . . . . . . . . . 65 134 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 67 135 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 68 136 8.1. Normative References . . . . . . . . . . . . . . . . . . 68 137 8.2. Informative References . . . . . . . . . . . . . . . . . 70 138 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 76 139 A.1. Implementation Strategies for Route Optimization . . . . 76 140 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 76 141 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 77 142 A.4. AERO Critical Infrastructure Considerations . . . . . . . 77 143 A.5. AERO Server Failure Implications . . . . . . . . . . . . 78 144 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 79 145 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 81 146 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 82 148 1. Introduction 150 Asymmetric Extended Route Optimization (AERO) fulfills the 151 requirements of Distributed Mobility Management (DMM) [RFC7333] and 152 route optimization [RFC5522] for aeronautical networking and other 153 network mobility use cases such as intelligent transportation 154 systems. AERO is an internetworking and mobility management service 155 that employs the Overlay Multilink Network Interface (OMNI) 156 [I-D.templin-6man-omni-interface] Non-Broadcast, Multiple Access 157 (NBMA) virtual link model. The OMNI link is a virtual overlay 158 configured over one or more underlying Internetworks, and nodes on 159 the link can exchange IP packets via tunneling. The OMNI Adaptation 160 Layer (OAL) supports multilink operation for increased reliability, 161 bandwidth optimization and traffic path selection while accommodating 162 Maximum Transmission Unit (MTU) diversity. 164 The AERO service comprises Clients, Proxys, Servers and Relays that 165 are seen as OMNI link neighbors as well as Bridges that interconnect 166 OMNI link segments. Each node's OMNI interface uses an IPv6 link- 167 local address format that supports operation of the IPv6 Neighbor 168 Discovery (ND) protocol [RFC4861] and links ND to IP forwarding. A 169 node's OMNI interface can be configured over multiple underlying 170 interfaces, and may therefore appear as a single interface with 171 multiple link-layer addresses. Each link-layer address is subject to 172 change due to mobility and/or QoS fluctuations, and link-layer 173 address changes are signaled by ND messaging the same as for any IPv6 174 link. 176 AERO provides a cloud-based service where mobile node Clients may use 177 any Server acting as a Mobility Anchor Point (MAP) and fixed nodes 178 may use any Relay on the link for efficient communications. Fixed 179 nodes forward packets destined to other AERO nodes to the nearest 180 Relay, which forwards them through the cloud. A mobile node's 181 initial packets are forwarded through the Server, while direct 182 routing is supported through asymmetric extended route optimization 183 while data packets are flowing. Both unicast and multicast 184 communications are supported, and mobile nodes may efficiently move 185 between locations while maintaining continuous communications with 186 correspondents and without changing their IP Address. 188 AERO Bridges are interconnected in a secured private BGP overlay 189 routing instance using encapsulation to provide a hybrid routing/ 190 bridging service that joins the underlying Internetworks of multiple 191 disjoint administrative domains into a single unified OMNI link. 193 Each OMNI link instance is characterized by the set of Mobility 194 Service Prefixes (MSPs) common to all mobile nodes. The link extends 195 to the point where a Relay/Server is on the optimal route from any 196 correspondent node on the link, and provides a conduit between the 197 underlying Internetwork and the OMNI link. To the underlying 198 Internetwork, the Relay/Server is the source of a route to the MSP, 199 and hence uplink traffic to the mobile node is naturally routed to 200 the nearest Relay/Server. 202 AERO assumes the use of PIM Sparse Mode in support of multicast 203 communication. In support of Source Specific Multicast (SSM) when a 204 Mobile Node is the source, AERO route optimization ensures that a 205 shortest-path multicast tree is established with provisions for 206 mobility and multilink operation. In all other multicast scenarios 207 there are no AERO dependencies. 209 AERO was designed for aeronautical networking for both manned and 210 unmanned aircraft, where the aircraft is treated as a mobile node 211 that can connect an Internet of Things (IoT). AERO is also 212 applicable to a wide variety of other use cases. For example, it can 213 be used to coordinate the Virtual Private Network (VPN) links of 214 mobile nodes (e.g., cellphones, tablets, laptop computers, etc.) that 215 connect into a home enterprise network via public access networks 216 using services such as OpenVPN [OVPN]. It can also be used to 217 facilitate vehicular and pedestrian communications services for 218 intelligent transportation systems. Other applicable use cases are 219 also in scope. 221 The following numbered sections present the AERO specification. The 222 appendices at the end of the document are non-normative. 224 2. Terminology 226 The terminology in the normative references applies; especially, the 227 terminology in the OMNI specification 228 [I-D.templin-6man-omni-interface] is used extensively throughout. 229 The following terms are defined within the scope of this document: 231 IPv6 Neighbor Discovery (ND) 232 an IPv6 control message service for coordinating neighbor 233 relationships between nodes connected to a common link. AERO uses 234 the ND service specified in [RFC4861]. 236 IPv6 Prefix Delegation 237 a networking service for delegating IPv6 prefixes to nodes on the 238 link. The nominal service is DHCPv6 [RFC8415], however alternate 239 services (e.g., based on ND messaging) are also in scope. Most 240 notably, a minimal form of prefix delegation known as "prefix 241 registration" can be used if the Client knows its prefix in 242 advance and can represent it in the IPv6 source address of an ND 243 message. 245 Access Network (ANET) 246 a node's first-hop data link service network (e.g., a radio access 247 network, cellular service provider network, corporate enterprise 248 network, etc.) that often provides link-layer security services 249 such as IEEE 802.1X and physical-layer security prevent 250 unauthorized access internally and with border network-layer 251 security services such as firewalls and proxies that prevent 252 unauthorized outside access. 254 ANET interface 255 a node's attachment to a link in an ANET. 257 Internetwork (INET) 258 a connected IP network topology with a coherent routing and 259 addressing plan and that provides a transit backbone service for 260 ANET end systems. INETs also provide an underlay service over 261 which the AERO virtual link is configured. Example INETs include 262 corporate enterprise networks, aviation networks, and the public 263 Internet itself. When there is no administrative boundary between 264 an ANET and the INET, the ANET and INET are one and the same. 266 INET Partition 267 frequently, INETs such as large corporate enterprise networks are 268 sub-divided internally into separate isolated partitions. Each 269 partition is fully connected internally but disconnected from 270 other partitions, and there is no requirement that separate 271 partitions maintain consistent Internet Protocol and/or addressing 272 plans. (Each INET partition is seen as a separate OMNI link 273 segment as discussed below.) 275 INET interface 276 a node's attachment to a link in an INET. 278 INET address 279 an IP address assigned to a node's interface connection to an 280 INET. 282 INET encapsulation 283 the encapsulation of a packet in an outer header or headers that 284 can be routed within the scope of the local INET partition. 286 OMNI link 287 the same as defined in [I-D.templin-6man-omni-interface], and 288 manifested by IPv6 encapsulation [RFC2473]. The OMNI link spans 289 underlying INET segments joined by virtual bridges in a spanning 290 tree the same as a bridged campus LAN. AERO nodes on the OMNI 291 link appear as single-hop neighbors even though they may be 292 separated by multiple underlying INET hops, and can use Segment 293 Routing [RFC8402] to cause packets to visit selected waypoints on 294 the link. 296 OMNI Interface 297 a node's attachment to an OMNI link. Since the addresses assigned 298 to an OMNI interface are managed for uniqueness, OMNI interfaces 299 do not require Duplicate Address Detection (DAD) and therefore set 300 the administrative variable 'DupAddrDetectTransmits' to zero 301 [RFC4862]. 303 OMNI Adaptation Layer (OAL) 304 an OMNI interface process whereby packets admitted into the 305 interface are wrapped in a mid-layer IPv6 header and fragmented/ 306 reassembled if necessary to support the OMNI link Maximum 307 Transmission Unit (MTU). The OAL is also responsible for 308 generating MTU-related control messages as necessary, and for 309 providing addressing context for spanning multiple segments of a 310 bridged OMNI link. 312 OMNI Link-Local Address (LLA) 313 a link local IPv6 address per [RFC4291] constructed as specified 314 in Section 3.2.2. 316 OMNI Unique-Local Address (ULA) 317 an IPv6 address from the IPv6 ULA prefix fd00::/8 [RFC4193], and 318 constructed as specified in [I-D.templin-6man-omni-interface]. 319 OMNI ULAs are statelessly derived from OMNI LLAs, and vice-versa. 321 underlying interface 322 an ANET or INET interface over which an OMNI interface is 323 configured. 325 Mobility Service Prefix (MSP) 326 an aggregated IP Global Unicast Address (GUA) prefix (e.g., 327 2001:db8::/32, 192.0.2.0/24, etc.) assigned to the OMNI link and 328 from which more-specific Mobile Network Prefixes (MNPs) are 329 delegated. OMNI link administrators typically obtain MSPs from an 330 Internet address registry, however private-use prefixes can 331 alternatively be used subject to certain limitations (see: 332 [I-D.templin-6man-omni-interface]). OMNI links that connect to 333 the global Internet advertise their MSPs to their interdomain 334 routing peers. 336 Mobile Network Prefix (MNP) 337 a longer IP prefix delegated from an MSP (e.g., 338 2001:db8:1000:2000::/56, 192.0.2.8/30, etc.) and assigned to an 339 AERO Client or Relay. 341 AERO node 342 a node that is connected to an OMNI link and participates in the 343 AERO internetworking and mobility service. 345 AERO Client ("Client") 346 an AERO node that connects over one or more underlying interfaces 347 and requests MNP delegation/registration service from AERO 348 Servers. The Client assigns a Client LLA to the OMNI interface 349 for use in ND exchanges with other AERO nodes and forwards packets 350 to correspondents according to OMNI interface neighbor cache 351 state. 353 AERO Server ("Server") 354 an INET node that configures an OMNI interface to provide default 355 forwarding and mobility/multilink services for AERO Clients. The 356 Server assigns an administratively-provisioned LLA to its OMNI 357 interface to support the operation of the ND services, and 358 advertises all of its associated MNPs via BGP peerings with 359 Bridges. 361 AERO Relay ("Relay") 362 an AERO Server that also provides forwarding services between 363 nodes reached via the OMNI link and correspondents on other links. 364 AERO Relays are provisioned with MNPs (i.e., the same as for an 365 AERO Client) and run a dynamic routing protocol to discover any 366 non-MNP IP GUA routes in service on its connected INET links. In 367 both cases, the Relay advertises the MSP(s) to its downstream 368 networks, and distributes all of its associated MNPs and non-MNP 369 IP GUA routes via BGP peerings with Bridges (i.e., the same as for 370 an AERO Server). 372 AERO Bridge ("Bridge") 373 a node that provides hybrid routing/bridging services (as well as 374 a security trust anchor) for nodes on an OMNI link. As a router, 375 the Bridge forwards packets using standard IP forwarding. As a 376 bridge, the Bridge forwards packets over the OMNI link without 377 decrementing the IPv6 Hop Limit. AERO Bridges peer with Servers 378 and other Bridges to discover the full set of MNPs for the link as 379 well as any non-MNP IP GUA routes that are reachable via Relays. 381 AERO Proxy ("Proxy") 382 a node that provides proxying services between Clients in an ANET 383 and Servers in external INETs. The AERO Proxy is a conduit 384 between the ANET and external INETs in the same manner as for 385 common web proxies, and behaves in a similar fashion as for ND 386 proxies [RFC4389]. A node may be configured to act as either a 387 Proxy and/or a Server, depending on Client Server selection 388 criteria. 390 ingress tunnel endpoint (ITE) 391 an OMNI interface endpoint that injects encapsulated packets into 392 an OMNI link. 394 egress tunnel endpoint (ETE) 395 an OMNI interface endpoint that receives encapsulated packets from 396 an OMNI link. 398 link-layer address 399 an IP address used as an encapsulation header source or 400 destination address from the perspective of the OMNI interface. 401 When an upper layer protocol (e.g., UDP) is used as part of the 402 encapsulation, the port number is also considered as part of the 403 link-layer address. 405 network layer address 406 the source or destination address of an encapsulated IP packet 407 presented to the OMNI interface. 409 end user network (EUN) 410 an internal virtual or external edge IP network that an AERO 411 Client or Relay connects to the rest of the network via the OMNI 412 interface. The Client/Relay sees each EUN as a "downstream" 413 network, and sees the OMNI interface as the point of attachment to 414 the "upstream" network. 416 Mobile Node (MN) 417 an AERO Client and all of its downstream-attached networks that 418 move together as a single unit, i.e., an end system that connects 419 an Internet of Things. 421 Mobile Router (MR) 422 a MN's on-board router that forwards packets between any 423 downstream-attached networks and the OMNI link. The MR is the MN 424 entity that hosts the AERO Client. 426 Route Optimization Source (ROS) 427 the AERO node nearest the source that initiates route 428 optimization. The ROS may be a Server or Proxy acting on behalf 429 of the source Client. 431 Route Optimization responder (ROR) 432 the AERO node nearest the target destination that responds to 433 route optimization requests. The ROR may be a Server acting on 434 behalf of a target MNP Client, or a Relay for a non-MNP 435 destination. 437 MAP List 438 a geographically and/or topologically referenced list of addresses 439 of all Servers within the same OMNI link. There is a single MAP 440 list for the entire OMNI link. 442 Distributed Mobility Management (DMM) 443 a BGP-based overlay routing service coordinated by Servers and 444 Bridges that tracks all Server-to-Client associations. 446 Mobility Service (MS) 447 the collective set of all Servers, Proxys, Bridges and Relays that 448 provide the AERO Service to Clients. 450 Mobility Service Endpoint MSE) 451 an individual Server, Proxy, Bridge or Relay in the Mobility 452 Service. 454 Throughout the document, the simple terms "Client", "Server", 455 "Bridge", "Proxy" and "Relay" refer to "AERO Client", "AERO Server", 456 "AERO Bridge", "AERO Proxy" and "AERO Relay", respectively. 457 Capitalization is used to distinguish these terms from other common 458 Internetworking uses in which they appear without capitalization. 460 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 461 the names of node variables, messages and protocol constants) is used 462 throughout this document. The terms "All-Routers multicast", "All- 463 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 464 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 465 generically refer to either Internet Protocol version, i.e., IPv4 466 [RFC0791] or IPv6 [RFC8200]. 468 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 469 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 470 "OPTIONAL" in this document are to be interpreted as described in BCP 471 14 [RFC2119][RFC8174] when, and only when, they appear in all 472 capitals, as shown here. 474 3. Asymmetric Extended Route Optimization (AERO) 476 The following sections specify the operation of IP over OMNI links 477 using the AERO service: 479 3.1. AERO Node Types 481 AERO Clients are Mobile Nodes (MNs) that connect via underlying 482 interfaces with addresses that may change when the Client moves to a 483 new network connection point. AERO Clients register their Mobile 484 Network Prefixes (MNPs) with the AERO service, and distribute the 485 MNPs to nodes on EUNs. AERO Bridges, Servers, Proxys and Relays are 486 critical infrastructure elements in fixed (i.e., non-mobile) INET 487 deployments and hence have permanent and unchanging INET addresses. 488 Together, they constitute the AERO service which provides an OMNI 489 link virtual overlay for connecting AERO Clients. 491 AERO Bridges provide hybrid routing/bridging services (as well as a 492 security trust anchor) for nodes on an OMNI link. Bridges use 493 standard IPv6 routing to forward packets both within the same INET 494 partitions and between disjoint INET partitions based on a mid-layer 495 IPv6 encapsulation per [RFC2473]. The inner IP layer experiences a 496 virtual bridging service since the inner IP TTL/Hop Limit is not 497 decremented during forwarding. Each Bridge also peers with Servers 498 and other Bridges in a dynamic routing protocol instance to provide a 499 Distributed Mobility Management (DMM) service for the list of active 500 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 501 one or more Mobility Service Prefixes (MSPs) and configure secured 502 tunnels with Servers, Relays, Proxys and other Bridges; they further 503 maintain IP forwarding table entries for each MNP and any other 504 reachable non-MNP prefixes. 506 AERO Servers provide default forwarding and mobility/multilink 507 services for AERO Client Mobile Nodes (MNs). Each Server also peers 508 with Bridges in a dynamic routing protocol instance to advertise its 509 list of associated MNPs (see Section 3.2.3). Servers facilitate 510 prefix delegation/registration exchanges with Clients, where each 511 delegated prefix becomes an MNP taken from an MSP. Servers forward 512 packets between OMNI interface neighbors and track each Client's 513 mobility profiles. Servers may further act as Servers for some sets 514 of Clients and as Proxies for others. 516 AERO Proxys provide a conduit for ANET Clients to associate with 517 Servers in external INETs. Client and Servers exchange control plane 518 messages via the Proxy acting as a bridge between the ANET/INET 519 boundary. The Proxy forwards data packets between Clients and the 520 OMNI link according to forwarding information in the neighbor cache. 521 The Proxy function is specified in Section 3.13. Proxys may further 522 act as Proxys for some sets of Clients and as Servers for others. 524 AERO Relays are Servers that provide forwarding services between the 525 OMNI interface and INET/EUN interfaces. Relays are provisioned with 526 MNPs the same as for an AERO Client, and also run a dynamic routing 527 protocol to discover any non-MNP IP routes. The Relay advertises the 528 MSP(s) to its connected networks, and distributes all of its 529 associated MNPs and non-MNP IP GUA routes via BGP peerings with 530 Bridges 532 3.2. The AERO Service over OMNI Links 534 3.2.1. AERO/OMNI Reference Model 536 Figure 1 presents the basic OMNI link reference model: 538 +----------------+ 539 | AERO Bridge B1 | 540 | Nbr: S1, S2, P1| 541 |(X1->S1; X2->S2)| 542 | MSP M1 | 543 +-+---------+--+-+ 544 +--------------+ | Secured | | +--------------+ 545 |AERO Server S1| | tunnels | | |AERO Server S2| 546 | Nbr: C1, B1 +-----+ | +-----+ Nbr: C2, B1 | 547 | default->B1 | | | default->B1 | 548 | X1->C1 | | | X2->C2 | 549 +-------+------+ | +------+-------+ 550 | OMNI link | | 551 X===+===+===================+==)===============+===+===X 552 | | | | 553 +-----+--------+ +--------+--+-----+ +--------+-----+ 554 |AERO Client C1| | AERO Proxy P1 | |AERO Client C2| 555 | Nbr: S1 | |(Proxy Nbr Cache)| | Nbr: S2 | 556 | default->S1 | +--------+--------+ | default->S2 | 557 | MNP X1 | | | MNP X2 | 558 +------+-------+ .--------+------. +-----+--------+ 559 | (- Proxyed Clients -) | 560 .-. `---------------' .-. 561 ,-( _)-. ,-( _)-. 562 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 563 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 564 `-(______)-' +-------+ +-------+ `-(______)-' 566 Figure 1: AERO/OMNI Reference Model 568 In this model: 570 o the OMNI link is an overlay network service configured over one or 571 more underlying INET partitions which may be managed by different 572 administrative authorities and have incompatible protocols and/or 573 addressing plans. 575 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 576 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 577 via BGP peerings over secured tunnels to Servers (S1, S2). 578 Bridges connect the disjoint segments of a partitioned OMNI link. 580 o AERO Servers/Relays S1 and S2 configure secured tunnels with 581 Bridge B1 and also provide mobility, multilink and default router 582 services for their associated Clients C1 and C2. 584 o AERO Clients C1 and C2 associate with Servers S1 and S2, 585 respectively. They receive Mobile Network Prefix (MNP) 586 delegations X1 and X2, and also act as default routers for their 587 associated physical or internal virtual EUNs. Simple hosts H1 and 588 H2 attach to the EUNs served by Clients C1 and C2, respectively. 590 o AERO Proxy P1 configures a secured tunnel with Bridge B1 and 591 provides proxy services for AERO Clients in secured enclaves that 592 cannot associate directly with other OMNI link neighbors. 594 An OMNI link configured over a single INET appears as a single 595 unified link with a consistent underlying network addressing plan. 596 In that case, all nodes on the link can exchange packets via simple 597 INET encapsulation, since the underlying INET is connected. In 598 common practice, however, an OMNI link may be partitioned into 599 multiple "segments", where each segment is a distinct INET 600 potentially managed under a different administrative authority (e.g., 601 as for worldwide aviation service providers such as ARINC, SITA, 602 Inmarsat, etc.). Individual INETs may also themselves be partitioned 603 internally, in which case each internal partition is seen as a 604 separate segment. 606 The addressing plan of each segment is consistent internally but will 607 often bear no relation to the addressing plans of other segments. 608 Each segment is also likely to be separated from others by network 609 security devices (e.g., firewalls, proxies, packet filtering 610 gateways, etc.), and in many cases disjoint segments may not even 611 have any common physical link connections. Therefore, nodes can only 612 be assured of exchanging packets directly with correspondents in the 613 same segment, and not with those in other segments. The only means 614 for joining the segments therefore is through inter-domain peerings 615 between AERO Bridges. 617 The same as for traditional campus LANs, multiple OMNI link segments 618 can be joined into a single unified link via a virtual bridging 619 service using the OMNI Adaptation Layer (OAL) which inserts a mid- 620 layer IPv6 encapsulation per [RFC2473] that supports inter-segment 621 forwarding (i.e., bridging) without decrementing the network-layer 622 TTL/Hop Limit. This bridging of OMNI link segments is shown in 623 Figure 2: 625 . . . . . . . . . . . . . . . . . . . . . . . 626 . . 627 . .-(::::::::) . 628 . .-(::::::::::::)-. +-+ . 629 . (:::: Segment A :::)--|B|---+ . 630 . `-(::::::::::::)-' +-+ | . 631 . `-(::::::)-' | . 632 . | . 633 . .-(::::::::) | . 634 . .-(::::::::::::)-. +-+ | . 635 . (:::: Segment B :::)--|B|---+ . 636 . `-(::::::::::::)-' +-+ | . 637 . `-(::::::)-' | . 638 . | . 639 . .-(::::::::) | . 640 . .-(::::::::::::)-. +-+ | . 641 . (:::: Segment C :::)--|B|---+ . 642 . `-(::::::::::::)-' +-+ | . 643 . `-(::::::)-' | . 644 . | . 645 . ..(etc).. x . 646 . . 647 . . 648 . <- OMNI link Bridged by encapsulation -> . 649 . . . . . . . . . . . . . .. . . . . . . . . 651 Figure 2: Bridging OMNI Link Segments 653 Bridges, Servers, Relays and Proxys connect via secured INET tunnels 654 over their respective segments in a spanning tree topology rooted at 655 the Bridges. The secured spanning tree supports strong 656 authentication for IPv6 ND control messages and may also be used to 657 convey the initial data packets in a flow. Route optimization can 658 then be employed to cause data packets to take more direct paths 659 between OMNI link neighbors without having to strictly follow the 660 spanning tree. 662 3.2.2. Link-Local Addresses (LLAs), Unique Local Addresses (ULAs) and 663 Global Unicast Addresses (GUA) 665 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 666 fe80::/64 [RFC4291] to assign LLAs used for network-layer addresses 667 in link-scoped IPv6 ND and data messages. They also use the Unique 668 Local Address (ULA) prefix fd00::/8 followed by a pseudo-random 669 40-bit global ID to form the prefix [ULA]::/48, then include a 16-bit 670 OMNI link identifier '*' to form the prefix [ULA*]::/64 to assign OAL 671 addresses. Finally, the MSPs and MNPs used by AERO nodes are 672 typically based on Global Unicast Addresses (GUAs), but in some cases 673 may be based on private-use addresses. See 674 [I-D.templin-6man-omni-interface] for a full specification of LLAs, 675 ULAs and GUAs used by AERO nodes on OMNI links. 677 3.2.3. AERO Routing System 679 The AERO routing system comprises a private instance of the Border 680 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 681 and Servers and does not interact with either the public Internet BGP 682 routing system or any underlying INET routing systems. 684 In a reference deployment, each Server is configured as an Autonomous 685 System Border Router (ASBR) for a stub Autonomous System (AS) using 686 an AS Number (ASN) that is unique within the BGP instance, and each 687 Server further uses eBGP to peer with one or more Bridges but does 688 not peer with other Servers. Each INET of a multi-segment OMNI link 689 must include one or more Bridges, which peer with the Servers and 690 Proxys within that INET. All Bridges within the same INET are 691 members of the same hub AS using a common ASN, and use iBGP to 692 maintain a consistent view of all active MNPs currently in service. 693 The Bridges of different INETs peer with one another using eBGP. 695 Bridges advertise the OMNI link's MSPs and any non-MNP routes to each 696 of their Servers. This means that any aggregated non-MNPs (including 697 "default") are advertised to all Servers. Each Bridge configures a 698 black-hole route for each of its MSPs. By black-holing the MSPs, the 699 Bridge will maintain forwarding table entries only for the MNPs that 700 are currently active, and packets destined to all other MNPs will 701 correctly incur Destination Unreachable messages due to the black- 702 hole route. In this way, Servers have only partial topology 703 knowledge (i.e., they know only about the MNPs of their directly 704 associated Clients) and they forward all other packets to Bridges 705 which have full topology knowledge. 707 Each OMNI link segment assigns a unique sub-prefix of [ULA*]::/96 708 known as the ULA partition prefix. For example, a first segment 709 could assign [ULA*]::1000/116, a second could assign 710 [ULA*]::2000/116, a third could assign [ULA*]::3000/116, etc. The 711 administrative authorities for each segment must therefore coordinate 712 to assure mutually-exclusive partition prefix assignments, but 713 internal provisioning of each prefix is an independent local 714 consideration for each administrative authority. For each partition 715 prefix, the Bridge(s) that connect that segment assign the all-zero's 716 address of the prefix as a Subnet Router Anycast address. For 717 example, the Subnet Router Anycast address for [ULA*]::1000/116 is 718 simply [ULA*]::1000. 720 ULA partition prefixes are statically represented in Bridge 721 forwarding tables. Bridges join multiple segments into a unified 722 OMNI link over multiple diverse administrative domains. They support 723 a bridging function by first establishing forwarding table entries 724 for their partition prefixes either via standard BGP routing or 725 static routes. For example, if three Bridges ('A', 'B' and 'C') from 726 different segments serviced [ULA*]::1000/116, [ULA*]::2000/116 and 727 [ULA*]::3000/116 respectively, then the forwarding tables in each 728 Bridge are as follows: 730 A: [ULA*]::1000/116->local, [ULA*]::2000/116->B, [ULA*]::3000/116->C 732 B: [ULA*]::1000/116->A, [ULA*]::2000/116->local, [ULA*]::3000/116->C 734 C: [ULA*]::1000/116->A, [ULA*]::2000/116->B, [ULA*]::3000/116->local 736 These forwarding table entries are permanent and never change, since 737 they correspond to fixed infrastructure elements in their respective 738 segments. 740 ULA Client prefixes are instead dynamically advertised in the AERO 741 routing system by Servers and Relays that provide service for their 742 corresponding MNPs. For example, if three Servers ('D', 'E' and 'F') 743 service the MNPs 2001:db8:1000:2000::/56, 2001:db8:3000:4000::/56 and 744 2001:db8:5000:6000::/56 then the routing system would include: 746 D: [ULA*]:2001:db8:1000:2000/120 748 E: [ULA*]:2001:db8:3000:4000/120 750 F: [ULA*]:2001:db8:5000:6000/120 752 A full discussion of the BGP-based routing system used by AERO is 753 found in [I-D.ietf-rtgwg-atn-bgp]. 755 3.2.4. OMNI Link Encapsulation 757 With the Client and partition prefixes in place in Bridge forwarding 758 tables, the OMNI interface sends control and data packets toward AERO 759 destination nodes located in different OMNI link segments over the 760 spanning tree via mid-layer encapsulation using the OMNI Adaptation 761 Layer (OAL) header based on Generic Packet Tunneling in IPv6 762 [RFC2473]. When necessary, the OMNI interface also includes an OMNI 763 Routing Header (ORH) as an extension to the OAL header if final 764 segment forwarding information is available, e.g., in the neighbor 765 cache. The ORH is formatted as shown in Figure 3: 767 0 1 2 3 768 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 1 769 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 770 | Next Header | Hdr Ext Len | Routing Type | SRT | FMT | 771 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 772 | Last Hop Segment-id (LHS) | 773 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 774 ~ Link Layer Address (L2ADDR) ~ 775 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 776 ~ Destination Suffix (if necessary) ~ 777 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 778 ~ Null Padding (if necessary) ~ 779 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 781 Figure 3: OMNI Routing Header (ORH) Format 783 In this format: 785 o Next Header identifies the type of header immediately following 786 the ORH. 788 o Hdr Ext Len is the length of the Routing header in 8-octet units 789 (not including the first 8 octets), with trailing padding added if 790 necessary to produce an integral number of 8-octet units. 792 o Routing Type is set to TBD (see IANA Considerations). 794 o Segments Left is omitted, and replaced by a 5-bit SRT and 3-bit 795 FMT field. 797 o SRT - a 5-bit Segment Routing Topology prefix length value that 798 (when added to 96) determines the prefix length to apply to the 799 ULA formed from concatenating [ULA*]::/96 with the 32 bit LHS MSID 800 value that follows. For example, the value 16 corresponds to the 801 prefix length 112. 803 o FMT - a 3-bit "Framework/Mode/Type" code corresponding to the 804 included Link Layer Address as follows: 806 * When the most significant bit (i.e., "Framework") is set to 0, 807 L2ADDR is the INET encapsulation address of the Server/Proxy 808 named in the LHS; otherwise, it is the address for the target 809 Client, and the Destination Suffix is included. 811 * When the next most significant bit (i.e., "Mode") is set to 0, 812 the Source/Target L2ADDR is on the open INET; otherwise, it is 813 (likely) located behind a Network Address Translator (NAT). 815 * When the least significant bit (i.e., "Type") is set to 0, 816 L2ADDR includes a UDP Port Number followed by an IPv4 address; 817 otherwise, it includes a UDP Port Number followed by an IPv6 818 address. 820 o LHS - the 32 bit MSID of a service node in the Last Hop Segment on 821 the path to the target. When SRT and LHS are both set to 0, the 822 LHS is considered unspecified. When SRT is set to 0 and LHS is 823 non-zero, the prefix length is set to 128. SRT and LHS provide 824 guidance to the OMNI interface forwarding algorithm. 825 Specifically, if SRT/LHS is located in the local OMNI link segment 826 then the OMNI interface can omit OAL/ORH encapsualtion and send 827 directly to the target using INET encapsulation according to FMT/ 828 L2ADDR; else, it must perform INET/OAL/ORH encapsulation and 829 forward according to the OMNI link spanning tree. 831 o Link Layer Address (L2ADDR) - Formatted according to FMT, and 832 identifies the link-layer address (i.e., the encapsulation 833 address) of the source/target. The UDP Port Number appears in the 834 first two octets and the IP address appears in the next 4 octets 835 for IPv4 or 16 octets for IPv6. The Port Number and IP address 836 are recorded in ones-compliment "obfuscated" form per [RFC4380]. 837 The OMNI interface forwarding algorithm uses FMT/L2ADDR to 838 determine the INET encapsulation address for local forwarding when 839 SRT/LHS is located in the same OMNI link segment. 841 o Destination Suffix is a 64-bit field included only for OAL- 842 encapsulated packets that are destined directly to the ULA of the 843 Client (i.e., according to the FMT code). When present, 844 Destination Suffix encodes the 64-bit ULA suffix for the Client 845 that will receive packet. For example, if the Client ULA is 846 [ULA*]:2001:db8:1:2 then the Destination suffix encodes the value 847 2001:db8:1:2. 849 o Null Padding contains zero-valued octets as necessary to pad the 850 ORH to an integral number of 8-octet units. 852 When an AERO node uses OAL encapsulation for a packet with addresses 853 such as 2001:db8:1:2::1 and 2001:db8:1234:5678::1, it sets the OAL 854 header source address to its own ULA address (e.g., 855 [ULA*]::1000:2000). The node also sets the destination address to 856 the ULA of the Client (e.g., [ULA*]::2001:db8:1234:5678) when the 857 Client is addressed directly; otherwise, it sets the destination 858 address to the ULA of the Client's Proxy/Server (e.g., 860 [ULA*]::4321:9876). If the neighbor cache includes Last Hop Segment 861 information for the target destination, the node next inserts an ORH 862 immediately following the OAL header while including the correct SRT, 863 FMT, LHS, L2ADDR and (if necessary) Destination Suffix information. 864 Next, the node overwrites the OAL header destination address with the 865 LHS Subnet Router Anycast address (for example, for LHS 1000:2000 866 with SRT 16, the Subnet Router Anycast address is [ULA*]::1000:0000). 868 The node then fragments the OAL/ORH packet if necessary, with each 869 resulting fragment including the OAL/ORH headers while only the first 870 fragment includes the original IPv6 header. The node finally 871 encapsulates each resulting OAL/ORH packet/fragment in an INET header 872 with source address set to its own INET address (e.g., 192.0.2.100) 873 and destination set to the INET address of a Bridge (e.g., 874 192.0.2.1). 876 The encapsulation format in the above example is shown in Figure 4: 878 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 879 | INET Header | 880 | src = 192.0.2.100 | 881 | dst = 192.0.2.1 | 882 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 883 | OAL Header | 884 | src = [ULA*]::1000:2000 | 885 | dst= ULA for inner IP dst | 886 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 887 | ORH Header | 888 | (if necessary) | 889 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 890 | Inner IP Header | 891 | src = 2001:db8:1:2::1 | 892 | dst = 2001:db8:1234:5678::1 | 893 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 894 | | 895 ~ ~ 896 ~ Inner Packet Body ~ 897 ~ ~ 898 | | 899 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 901 Figure 4: OAL/ORH Encapsulation 903 In this format, the inner IP header and packet body are the original 904 IP packet, the OAL header is an IPv6 header prepared according to 905 [RFC2473], the ORH is a Routing Header extension of the OAL header, 906 and the INET header is prepared as discussed in Section 3.6. 908 This gives rise to a routing system that contains both Client prefix 909 routes that may change dynamically due to regional node mobility and 910 partition prefix routes that rarely if ever change. The Bridges can 911 therefore provide link-layer bridging by sending packets over the 912 spanning tree instead of network-layer routing according to MNP 913 routes. As a result, opportunities for packet loss due to node 914 mobility between different segments are mitigated. 916 In normal operations, IPv6 ND messages are conveyed over secured 917 paths between OMNI link neighbors so that specific Proxys, Servers or 918 Relays can be addressed without being subject to mobility events. 919 Conversely, only the first few packets destined to Clients need to 920 traverse secured paths until route optimization can determine a more 921 direct path. 923 Note: An IPv6 "minimal encapsulation" format (i.e., an IPv6 variant 924 of [RFC2004]) based on extensions to the ORH was considered, analyzed 925 and rejected. In the approach, the ORH would be inserted as an 926 extension header to the original IPv6 packet header. The IPv6 927 destination address would then be written into the ORH, and the ULA 928 corresponding to the destination would be overwritten in the IPv6 929 destination address. This would seemingly convey enough forwarding 930 information so that OAL encapsulation could be avoided. However, 931 this "minimal encapsulation" IPv6 packet would then have a non-ULA 932 source address and ULA destination address, an incorrect value in 933 upper layer protocol checksums, and a Hop Limit that is decremented 934 within the spanning tree when it should not be. The insertion and 935 removal of the ORH would also entail rewriting the Payload Length and 936 Next Header fields - again, invalidating upper layer checksums. 937 These irregularities would result in implementation challenges and 938 the potential for operational issues, e.g., since actionable ICMPv6 939 error reports could not be delivered to the original source. In 940 order to address the issues, still more information such as the 941 original IPv6 source address could be written into the ORH. However, 942 with the additional information the benefit of the "minimal 943 encapsulation" savings quickly diminishes, and becomes overshadowed 944 by the implementation and operational irregularities. 946 3.2.5. Segment Routing Topologies (SRTs) 948 The 64-bit sub-prefixes of [ULA]::/48 identify up to 2^16 distinct 949 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 950 OMNI link overlay instance using a distinct set of ULAs, and emulates 951 a Virtual LAN (VLAN) service for the OMNI link. In some cases (e.g., 952 when redundant topologies are needed for fault tolerance and 953 reliability) it may be beneficial to deploy multiple SRTs that act as 954 independent overlay instances. A communication failure in one 955 instance therefore will not affect communications in other instances. 957 Each SRT is identified by a distinct value in bits 48-63 of 958 [ULA]::/48, i.e., as [ULA0]::/64, [ULA1]::/64, [ULA2]::/64, etc. 959 This document asserts that up to four SRTs provide a level of safety 960 sufficient for critical communications such as civil aviation. Each 961 SRT is designated with a color that identifies a different OMNI link 962 instance as follows: 964 o Red - corresponds to [ULA0]::/64 966 o Green - corresponds to [ULA1]::/64 968 o Blue-1 - corresponds to [ULA2]::/64 970 o Blue-2 - corresponds to [ULA3]::/64 972 o the remaining [ULA*]::/64 sub prefixes are available for 973 additional SRTs. 975 Each OMNI interface is identified by a unique interface name (e.g., 976 omni0, omni1, omni2, etc.) and assigns an anycast ULA corresponding 977 to its SRT prefix. For example, the anycast ULA for the Green SRT is 978 simply [ULA1]::. The anycast ULA is used for OMNI interface 979 determination in Safety-Based Multilink (SBM) as discussed in 980 [I-D.templin-6man-omni-interface]. Each OMNI interface further 981 applies Performance-Based Multilink (PBM) internally. 983 3.2.6. Segment Routing For OMNI Link Selection 985 An original IPv6 source can direct an IPv6 packet to an AERO node by 986 including a standard IPv6 Segment Routing Header (SRH) [RFC8754] with 987 the anycast ULA for the selected SRT as either the IPv6 destination 988 or as an intermediate hop within the SRH. This allows the original 989 source to determine the specific OMNI link topology a packet will 990 traverse when there may be multiple alternatives. 992 When the AERO node processes the SRH and forwards the packet to the 993 correct OMNI interface, the OMNI interface writes the next IPv6 994 address from the SRH into the IPv6 destination address and decrements 995 Segments Left. If decrementing would cause Segments Left to become 996 0, the OMNI interface deletes the SRH before forwarding. This form 997 of Segment Routing supports Safety-Based Multilink (SBM). 999 3.2.7. Segment Routing Within the OMNI Link 1001 AERO node OMNI interfaces can insert OAL/ORH headers for Segment 1002 Routing within the OMNI link to influence the paths of packets 1003 destined to targets in remote segments without requiring all packets 1004 to traverse strict spanning tree paths. 1006 When an AERO node's OMNI interface has a packet to send to a target 1007 discovered through route optimization located in the same OMNI link 1008 segment, it encapsulates the packet in OAL/ORH headers if necessary 1009 as discussed above. The node then uses the target's Link Layer 1010 Address (L2ADDR) information for INET encapsulation. 1012 When an AERO node's OMNI interface has a packet to send to a route 1013 optimization target located in a remote OMNI link segment, it 1014 encapsulates the packet in OAL/ORH headers as discussed above while 1015 forwarding the packet to a Bridge with destination set to the Subnet 1016 Router Anycast address for the final OMNI link segment. 1018 When a Bridge receives a packet destined to its Subnet Router Anycast 1019 address with an OAL/ORH with SRT/LHS values corresponding to the 1020 local segment, it examines the L2ADDR according to FMT and removes 1021 the ORH from the packet. If the ORH includes a saved Destination 1022 Suffix, the Bridge then writes the corresponding ULA into the OAL 1023 destination address; otherwise, it writes the ULA corresponding to 1024 the SRT/LHS fields into the destination. The Bridge then 1025 encapsulates the packet in an INET header according to L2ADDR and 1026 forwards the packet within the INET either to the LHS Server/Proxy or 1027 directly to the destination itself. In this way, the Bridge 1028 participates in route optimization to reduce traffic load and 1029 suboptimal routing through strict spanning tree paths. 1031 3.3. OMNI Interface Characteristics 1033 OMNI interfaces are virtual interfaces configured over one or more 1034 underlying interfaces classified as follows: 1036 o INET interfaces connect to an INET either natively or through one 1037 or several IPv4 Network Address Translators (NATs). Native INET 1038 interfaces have global IP addresses that are reachable from any 1039 INET correspondent. All Server, Relay and Bridge interfaces are 1040 native interfaces, as are INET-facing interfaces of Proxys. NATed 1041 INET interfaces connect to a private network behind one or more 1042 NATs that provide INET access. Clients that are behind a NAT are 1043 required to send periodic keepalive messages to keep NAT state 1044 alive when there are no data packets flowing. 1046 o ANET interfaces connect to an ANET that is separated from the open 1047 INET by a Proxy. Proxys can actively issue control messages over 1048 the INET on behalf of the Client to reduce ANET congestion. 1050 o VPNed interfaces use security encapsulation over the INET to a 1051 Virtual Private Network (VPN) server that also acts as a Server or 1052 Proxy. Other than the link-layer encapsulation format, VPNed 1053 interfaces behave the same as Direct interfaces. 1055 o Direct interfaces connect a Client directly to a Server or Proxy 1056 without crossing any ANET/INET paths. An example is a line-of- 1057 sight link between a remote pilot and an unmanned aircraft. The 1058 same Client considerations apply as for VPNed interfaces. 1060 OMNI interfaces use OAL/ORH encapsulation as necessary as discussed 1061 in Section 3.2.4. OMNI interfaces use link-layer encapsulation (see: 1062 Section 3.6) to exchange packets with OMNI link neighbors over INET 1063 or VPNed interfaces as well as over ANET interfaces for which the 1064 Client and Proxy may be multiple IP hops away. OMNI interfaces do 1065 not use link-layer encapsulation over Direct underlying interfaces or 1066 ANET interfaces when the Client and Proxy are known to be on the same 1067 underlying link. 1069 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 1070 state the same as for any interface. OMNI interfaces use ND messages 1071 including Router Solicitation (RS), Router Advertisement (RA), 1072 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 1073 neighbor cache management. 1075 OMNI interfaces send ND messages with an OMNI option formatted as 1076 specified in [I-D.templin-6man-omni-interface]. The OMNI option 1077 includes prefix registration information and Interface Attributes 1078 containing link information parameters for the OMNI interface's 1079 underlying interfaces. Each OMNI option may include multiple 1080 Interface Attributes sub-options, each identified by an ifIndex 1081 value. 1083 A Client's OMNI interface may be configured over multiple underlying 1084 interface connections. For example, common mobile handheld devices 1085 have both wireless local area network ("WLAN") and cellular wireless 1086 links. These links are often used "one at a time" with low-cost WLAN 1087 preferred and highly-available cellular wireless as a standby, but a 1088 simultaneous-use capability could provide benefits. In a more 1089 complex example, aircraft frequently have many wireless data link 1090 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1091 directional, etc.) with diverse performance and cost properties. 1093 If a Client's multiple underlying interfaces are used "one at a time" 1094 (i.e., all other interfaces are in standby mode while one interface 1095 is active), then ND message OMNI options include only a single 1096 Interface Attributes sub-option set to constant values. In that 1097 case, the Client would appear to have a single interface but with a 1098 dynamically changing link-layer address. 1100 If the Client has multiple active underlying interfaces, then from 1101 the perspective of ND it would appear to have multiple link-layer 1102 addresses. In that case, ND message OMNI options MAY include 1103 multiple Interface Attributes sub-options - each with values that 1104 correspond to a specific interface. Every ND message need not 1105 include Interface Attributes for all underlying interfaces; for any 1106 attributes not included, the neighbor considers the status as 1107 unchanged. 1109 Bridge, Server and Proxy OMNI interfaces may be configured over one 1110 or more secured tunnel interfaces. The OMNI interface configures 1111 both an LLA and its corresponding ULA, while the underlying secured 1112 tunnel interfaces are either unnumbered or configure the same ULA. 1113 The OMNI interface encapsulates each IP packet in OAL/ORH headers and 1114 presents the packet to the underlying secured tunnel interface. 1115 Routing protocols such as BGP that run over the OMNI interface do not 1116 employ OAL/ORH encapsulation, but rather present the routing protocol 1117 messages directly to the underlying secured tunnels while using the 1118 ULA as the source address. This distinction must be honored 1119 consistently according to each node's configuration so that the IP 1120 forwarding table will associate discovered IP routes with the correct 1121 interface. 1123 3.4. OMNI Interface Initialization 1125 AERO Servers, Proxys and Clients configure OMNI interfaces as their 1126 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1127 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1128 to ensure that packets with destination addresses covered by an MNP 1129 not explicitly assigned to a non-OMNI interface are directed to the 1130 OMNI interface. 1132 OMNI interface initialization procedures for Servers, Proxys, Clients 1133 and Bridges are discussed in the following sections. 1135 3.4.1. AERO Server/Relay Behavior 1137 When a Server enables an OMNI interface, it assigns an LLA/ULA 1138 appropriate for the given OMNI link segment. The Server also 1139 configures secured tunnels with one or more neighboring Bridges and 1140 engages in a BGP routing protocol session with each Bridge. 1142 The OMNI interface provides a single interface abstraction to the IP 1143 layer, but internally comprises multiple secured tunnels as well as 1144 an NBMA nexus for sending encapsulated data packets to OMNI interface 1145 neighbors. The Server further configures a service to facilitate ND 1146 exchanges with AERO Clients and manages per-Client neighbor cache 1147 entries and IP forwarding table entries based on control message 1148 exchanges. 1150 Relays are simply Servers that run a dynamic routing protocol to 1151 redistribute routes between the OMNI interface and INET/EUN 1152 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1153 networks on the INET/EUN interfaces (i.e., the same as a Client would 1154 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1155 interfaces. The Relay further provides an attachment point of the 1156 OMNI link to a non-MNP-based global topology. 1158 3.4.2. AERO Proxy Behavior 1160 When a Proxy enables an OMNI interface, it assigns an LLA/ULA and 1161 configures permanent neighbor cache entries the same as for Servers. 1162 The Proxy also configures secured tunnels with one or more 1163 neighboring Bridges and maintains per-Client neighbor cache entries 1164 based on control message exchanges. Importantly Proxys are often 1165 configured to act as Servers, and vice-versa. 1167 3.4.3. AERO Client Behavior 1169 When a Client enables an OMNI interface, it sends RS messages with ND 1170 parameters over its underlying interfaces to a Server, which returns 1171 an RA message with corresponding parameters. (The RS/RA messages may 1172 pass through a Proxy in the case of a Client's ANET interface, or 1173 through one or more NATs in the case of a Client's INET interface.) 1175 3.4.4. AERO Bridge Behavior 1177 AERO Bridges configure an OMNI interface and assign the ULA Subnet 1178 Router Anycast address for each OMNI link segment they connect to. 1179 Bridges configure secured tunnels with Servers, Proxys and other 1180 Bridges; they also configure LLAs/ULAs and permanent neighbor cache 1181 entries the same as Servers. Bridges engage in a BGP routing 1182 protocol session with a subset of the Servers and other Bridges on 1183 the spanning tree (see: Section 3.2.3). 1185 3.5. OMNI Interface Neighbor Cache Maintenance 1187 Each OMNI interface maintains a conceptual neighbor cache that 1188 includes an entry for each neighbor it communicates with on the OMNI 1189 link per [RFC4861]. OMNI interface neighbor cache entries are said 1190 to be one of "permanent", "symmetric", "asymmetric" or "proxy". 1192 Permanent neighbor cache entries are created through explicit 1193 administrative action; they have no timeout values and remain in 1194 place until explicitly deleted. AERO Bridges maintain permanent 1195 neighbor cache entries for their associated Proxys/Servers (and vice- 1196 versa). Each entry maintains the mapping between the neighbor's 1197 network-layer LLA and corresponding INET address. 1199 Symmetric neighbor cache entries are created and maintained through 1200 RS/RA exchanges as specified in Section 3.12, and remain in place for 1201 durations bounded by prefix lifetimes. AERO Servers maintain 1202 symmetric neighbor cache entries for each of their associated 1203 Clients, and AERO Clients maintain symmetric neighbor cache entries 1204 for each of their associated Servers. 1206 Asymmetric neighbor cache entries are created or updated based on 1207 route optimization messaging as specified in Section 3.14, and are 1208 garbage-collected when keepalive timers expire. AERO ROSs maintain 1209 asymmetric neighbor cache entries for active targets with lifetimes 1210 based on ND messaging constants. Asymmetric neighbor cache entries 1211 are unidirectional since only the ROS (and not the ROR) creates an 1212 entry. 1214 Proxy neighbor cache entries are created and maintained by AERO 1215 Proxys when they process Client/Server ND exchanges, and remain in 1216 place for durations bounded by ND and prefix lifetimes. AERO Proxys 1217 maintain proxy neighbor cache entries for each of their associated 1218 Clients. Proxy neighbor cache entries track the Client state and the 1219 address of the Client's associated Server(s). 1221 To the list of neighbor cache entry states in Section 7.3.2 of 1222 [RFC4861], Proxy and Server OMNI interfaces add an additional state 1223 DEPARTED that applies to symmetric and proxy neighbor cache entries 1224 for Clients that have recently departed. The interface sets a 1225 "DepartTime" variable for the neighbor cache entry to "DEPART_TIME" 1226 seconds. DepartTime is decremented unless a new ND message causes 1227 the state to return to REACHABLE. While a neighbor cache entry is in 1228 the DEPARTED state, packets destined to the target Client are 1229 forwarded to the Client's new location instead of being dropped. 1230 When DepartTime decrements to 0, the neighbor cache entry is deleted. 1231 It is RECOMMENDED that DEPART_TIME be set to the default constant 1232 value REACHABLE_TIME plus 10 seconds (40 seconds by default) to allow 1233 a window for packets in flight to be delivered while stale route 1234 optimization state may be present. 1236 When an ROR receives an authentic NS message used for route 1237 optimization, it searches for a symmetric neighbor cache entry for 1238 the target Client. The ROR then returns a solicited NA message 1239 without creating a neighbor cache entry for the ROS, but creates or 1240 updates a target Client "Report List" entry for the ROS and sets a 1241 "ReportTime" variable for the entry to REPORT_TIME seconds. The ROR 1242 resets ReportTime when it receives a new authentic NS message, and 1243 otherwise decrements ReportTime while no authentic NS messages have 1244 been received. It is RECOMMENDED that REPORT_TIME be set to the 1245 default constant value REACHABLE_TIME plus 10 seconds (40 seconds by 1246 default) to allow a window for route optimization to converge before 1247 ReportTime decrements below REACHABLE_TIME. 1249 When the ROS receives a solicited NA message response to its NS 1250 message used for route optimization, it creates or updates an 1251 asymmetric neighbor cache entry for the target network-layer and 1252 link-layer addresses. The ROS then (re)sets ReachableTime for the 1253 neighbor cache entry to REACHABLE_TIME seconds and uses this value to 1254 determine whether packets can be forwarded directly to the target, 1255 i.e., instead of via a default route. The ROS otherwise decrements 1256 ReachableTime while no further solicited NA messages arrive. It is 1257 RECOMMENDED that REACHABLE_TIME be set to the default constant value 1258 30 seconds as specified in [RFC4861]. 1260 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1261 of NS keepalives sent when a correspondent may have gone unreachable, 1262 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1263 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1264 to limit the number of unsolicited NAs that can be sent based on a 1265 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1266 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1267 same as specified in [RFC4861]. 1269 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1270 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1271 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1272 different values are chosen, all nodes on the link MUST consistently 1273 configure the same values. Most importantly, DEPART_TIME and 1274 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1275 REACHABLE_TIME to avoid packet loss due to stale route optimization 1276 state. 1278 3.5.1. OMNI Neighbor Interface Attributes 1280 OMNI interface IPv6 ND messages include OMNI options 1281 [I-D.templin-6man-omni-interface] with Interface Attributes that 1282 provide Link-Layer Address and QoS Preference information for the 1283 neighbor's underlying interfaces. This information is stored in the 1284 neighbor cache and provides the basis for the forwarding algorithm 1285 specified in Section 3.10. The information is cumulative and 1286 reflects the union of the OMNI information from the most recent ND 1287 messages received from the neighbor; it is therefore not required 1288 that each ND message contain all neighbor information. 1290 The OMNI option Interface Attributes for each underlying interface 1291 includes a two-part "Link-Layer Address" consisting of a simple IP 1292 encapsulation address determined by the FMT and L2ADDR fields and an 1293 OAL ULA determined by the SRT and LHS fields. If the neighbor is 1294 located in the local OMNI link segment (and, if any necessary NAT 1295 state has been established) forwarding via simple IP encapsulation 1296 can be used; otherwise, OAL encapsulation must be used. Underlying 1297 interfaces are further selected based on their associated preference 1298 values "high", "medium", "low" or "disabled". 1300 Note: the OMNI option is distinct from any Source/Target Link-Layer 1301 Address Options (S/TLLAOs) that may appear in an ND message according 1302 to the appropriate IPv6 over specific link layer specification (e.g., 1303 [RFC2464]). If both an OMNI option and S/TLLAO appear, the former 1304 pertains to encapsulation addresses while the latter pertains to the 1305 native L2 address format of the underlying media. 1307 3.5.2. OMNI Neighbor Advertisement Message Flags 1309 As discussed in Section 4.4 of [RFC4861] NA messages include three 1310 flag bits R, S and O. OMNI interface NA messages treat the flags as 1311 follows: 1313 o R: The R ("Router") flag is set to 1 in the NA messages sent by 1314 all AERO/OMNI node types. Simple hosts that would set R to 0 do 1315 not occur on the OMNI link itself, but may occur on the downstream 1316 links of Clients and Relays. 1318 o S: The S ("Solicited") flag is set exactly as specified in 1319 Section 4.4. of [RFC4861], i.e., it is set to 1 for Solicited NAs 1320 and set to 0 for Unsolicited NAs (both unicast and multicast). 1322 o O: The O ("Override") flag is set to 0 for solicited proxy NAs and 1323 set to 1 for all other solicited and unsolicited NAs. For further 1324 study is whether solicited NAs for anycast targets apply for OMNI 1325 links. Since OMNI LLAs must be uniquely assigned to Clients to 1326 support correct ND protocol operation, however, no role is 1327 currently seen for assigning the same OMNI LLA to multiple 1328 Clients. 1330 3.6. OMNI Interface Encapsulation and Re-encapsulation 1332 The OMNI Adaptation Layer (OAL) inserts mid-layer IPv6 headers known 1333 as the OAL/ORH headers when necessary as discussed in the following 1334 sections. After either inserting or omitting the OAL/ORH headers, 1335 the OMNI interface also inserts or omits an outer encapsulation 1336 header as discussed below. 1338 OMNI interfaces avoid outer encapsulation over Direct underlying 1339 interfaces and ANET underlying interfaces for which the Client and 1340 Proxy are connected to the same underlying link. Otherwise, OMNI 1341 interfaces encapsulate packets according to whether they are entering 1342 the OMNI interface from the network layer or if they are being re- 1343 admitted into the same OMNI link they arrived on. This latter form 1344 of encapsulation is known as "re-encapsulation". 1346 For packets entering the OMNI interface from the network layer, the 1347 OMNI interface copies the "TTL/Hop Limit", "Type of Service/Traffic 1348 Class" [RFC2983], "Flow Label"[RFC6438] (for IPv6) and "Congestion 1349 Experienced" [RFC3168] values in the inner packet's IP header into 1350 the corresponding fields in the OAL and outer encapsulation 1351 header(s). 1353 For packets undergoing re-encapsulation, the OMNI interface instead 1354 copies these values from the original encapsulation header into the 1355 new encapsulation header, i.e., the values are transferred between 1356 encapsulation headers and *not* copied from the encapsulated packet's 1357 network-layer header. (Note especially that by copying the TTL/Hop 1358 Limit between encapsulation headers the value will eventually 1359 decrement to 0 if there is a (temporary) routing loop.) 1361 OMNI interfaces configured over ANET underlying interfaces which 1362 employ a different IP protocol version (and/or when the Client and 1363 Proxy may be separated by multiple ANET IP hops) use IP-in-IP 1364 encapsulation so that the inner packet can traverse the ANET without 1365 decrementing the TTL/Hop-Limit. IPv6 underlying ANET interfaces use 1366 [RFC2473] encapsulation, while IPv4 interfaces use the appropriate 1367 encapsulation per one of [RFC5214][RFC2003]. 1369 OMNI interfaces configured over INET underlying interfaces 1370 encapsulate packets in INET headers according to the next hop 1371 determined in the forwarding algorithm in Section 3.10. If the next 1372 hop is reached via a secured tunnel, the OMNI interface uses an 1373 encapsulation format specific to the secured tunnel type (see: 1374 Section 6). If the next hop is reached via an unsecured INET 1375 interface, the OMNI interface instead uses UDP/IP encapsulation per 1376 [RFC4380] and as extended in [RFC6081]. 1378 When UDP/IP encapsulation is used, the OMNI interface next sets the 1379 UDP source port to a constant value that it will use in each 1380 successive packet it sends, and sets the UDP length field to the 1381 length of the encapsulated packet plus 8 bytes for the UDP header 1382 itself plus the length of any included extension headers or trailers. 1383 The encapsulated packet may be either IPv6 or IPv4, as distinguished 1384 by the version number found in the first four bits. 1386 For UDP/IP-encapsulated packets sent to a Server, Relay or Bridge, 1387 the OMNI interface sets the UDP destination port to 8060, i.e., the 1388 IANA-registered port number for AERO. For packets sent to a Client, 1389 the OMNI interface sets the UDP destination port to the port value 1390 stored in the neighbor cache entry for this Client. The OMNI 1391 interface finally includes/omits the UDP checksum according to 1392 [RFC6935][RFC6936]. 1394 3.7. OMNI Interface Decapsulation 1396 OMNI interfaces decapsulate packets destined either to the AERO node 1397 itself or to a destination reached via an interface other than the 1398 OMNI interface the packet was received on. When the encapsulated 1399 packet arrives in multiple OAL fragments, the OMNI interface 1400 reassembles as discussed in Section 3.9. Further decapsulation steps 1401 are performed according to the appropriate encapsulation format 1402 specification. 1404 3.8. OMNI Interface Data Origin Authentication 1406 AERO nodes employ simple data origin authentication procedures. In 1407 particular: 1409 o AERO Bridges, Servers and Proxys accept encapsulated data packets 1410 and control messages received from the (secured) spanning tree. 1412 o AERO Proxys and Clients accept packets that originate from within 1413 the same secured ANET. 1415 o AERO Clients and Relays accept packets from downstream network 1416 correspondents based on ingress filtering. 1418 o AERO Clients, Relays and Servers verify the outer UDP/IP 1419 encapsulation addresses according to [RFC4380]. 1421 AERO nodes silently drop any packets that do not satisfy the above 1422 data origin authentication procedures. Further security 1423 considerations are discussed in Section 6. 1425 3.9. OMNI Adaptation Layer and OMNI Interface MTU 1427 The OMNI interface observes the link nature of tunnels, including the 1428 Maximum Transmission Unit (MTU), Maximum Reassembly Unit (MRU) and 1429 the role of fragmentation and reassembly [I-D.ietf-intarea-tunnels]. 1430 The OMNI interface employs an OMNI Adaptation Layer (OAL) for 1431 accommodating multiple underlying links with diverse MTUs. The 1432 functions of the OAL and the OMNI interface MTU/MRU are specified in 1433 Section 5 of [I-D.templin-6man-omni-interface], with MTU/MRU both set 1434 to the constant value 9180 bytes. 1436 When the network layer presents an IP packet to the OMNI interface, 1437 the OAL encapsulates the packet in OAL/ORH headers. The OAL then 1438 fragments the encapsulated packet if necessary such that the OAL/ORH 1439 headers appear in each fragment while the original IP packet header 1440 appears only in the first fragment. The OAL then transmits each OAL/ 1441 ORH packet/fragment over an underlying linterface connected to either 1442 a physical link such as Ethernet, WiFi and the like or a virtual link 1443 such as an Internet or higher-layer tunnel (see the definition of 1444 link in [RFC8200]). 1446 Note: Although the ORH may be removed by a Bridge on the path (see: 1447 Section 3.10.4), this does not interfere with the destination's 1448 ability to reassemble in the event that the packet was fragmented. 1449 This is due to the fact that the ORH is not included in the 1450 fragmentabe part; therefore, its removal does not invalidate the 1451 offset values in any fragment headers. 1453 3.10. OMNI Interface Forwarding Algorithm 1455 IP packets enter a node's OMNI interface either from the network 1456 layer (i.e., from a local application or the IP forwarding system) or 1457 from the link layer (i.e., from an OMNI interface neighbor). All 1458 packets entering a node's OMNI interface first undergo data origin 1459 authentication as discussed in Section 3.8. Packets that satisfy 1460 data origin authentication are processed further, while all others 1461 are dropped silently. The OMNI interface OAL wraps accepted packets 1462 in OAL/ORH headers if necessary as discussed above. 1464 Packets that enter the OMNI interface from the network layer are 1465 forwarded to an OMNI interface neighbor. Packets that enter the OMNI 1466 interface from the link layer are either re-admitted into the OMNI 1467 link or forwarded to the network layer where they are subject to 1468 either local delivery or IP forwarding. In all cases, the OMNI 1469 interface itself MUST NOT decrement the network layer TTL/Hop-count 1470 since its forwarding actions occur below the network layer. 1472 OMNI interfaces may have multiple underlying interfaces and/or 1473 neighbor cache entries for neighbors with multiple underlying 1474 interfaces (see Section 3.3). The OMNI interface uses interface 1475 attributes and/or traffic classifiers (e.g., DSCP value, port number, 1476 etc.) to select an outgoing underlying interface for each packet 1477 based on the node's own QoS preferences, and also to select a 1478 destination link-layer address based on the neighbor's underlying 1479 interface with the highest preference. AERO implementations SHOULD 1480 allow for QoS preference values to be modified at runtime through 1481 network management. 1483 If multiple outgoing interfaces and/or neighbor interfaces have a 1484 preference of "high", the AERO node replicates the packet and sends 1485 one copy via each of the (outgoing / neighbor) interface pairs; 1486 otherwise, the node sends a single copy of the packet via an 1487 interface with the highest preference. AERO nodes keep track of 1488 which underlying interfaces are currently "reachable" or 1489 "unreachable", and only use "reachable" interfaces for forwarding 1490 purposes. 1492 The following sections discuss the OMNI interface forwarding 1493 algorithms for Clients, Proxys, Servers and Bridges. In the 1494 following discussion, a packet's destination address is said to 1495 "match" if it is the same as a cached address, or if it is covered by 1496 a cached prefix (which may be encoded in an LLA). 1498 3.10.1. Client Forwarding Algorithm 1500 When an IP packet enters a Client's OMNI interface from the network 1501 layer the Client searches for an asymmetric neighbor cache entry that 1502 matches the destination. If there is a match, the Client uses one or 1503 more "reachable" neighbor interfaces in the entry for packet 1504 forwarding. If there is no asymmetric neighbor cache entry, the 1505 Client instead forwards the packet toward a Server (the packet is 1506 intercepted by a Proxy if there is a Proxy on the path). The Client 1507 encapsulates the packet in OAL/ORH headers if necessary and fragments 1508 according to MTU requirements (see: Section 3.9). 1510 When an IP packet enters a Client's OMNI interface from the link- 1511 layer, if the destination matches one of the Client's MNPs or link- 1512 local addresses the Client reassembles and decapsulates as necessary 1513 and delivers the inner packet to the network layer. Otherwise, the 1514 Client drops the packet and MAY return a network-layer ICMP 1515 Destination Unreachable message subject to rate limiting (see: 1516 Section 3.11). 1518 3.10.2. Proxy Forwarding Algorithm 1520 For control messages originating from or destined to a Client, the 1521 Proxy intercepts the message and updates its proxy neighbor cache 1522 entry for the Client. The Proxy then forwards a (proxyed) copy of 1523 the control message. (For example, the Proxy forwards a proxied 1524 version of a Client's NS/RS message to the target neighbor, and 1525 forwards a proxied version of the NA/RA reply to the Client.) 1527 When the Proxy receives a data packet from a Client within the ANET, 1528 the Proxy reassembles and re-fragments if necessary then searches for 1529 an asymmetric neighbor cache entry that matches the destination and 1530 forwards as follows: 1532 o if the destination matches an asymmetric neighbor cache entry, the 1533 Proxy uses one or more "reachable" neighbor interfaces in the 1534 entry for packet forwarding using OAL/ORH encapsulation if 1535 necessary according to the cached link-layer address information. 1536 If the neighbor interface is in the same OMNI link segment, the 1537 Proxy forwards the packet directly to the neighbor; otherwise, it 1538 forwards the packet to a Bridge. 1540 o else, the Proxy uses OAL/ORH encapsulation and forwards the packet 1541 to a Bridge while using the ULA corresponding to the packet's 1542 destination as the destination address. 1544 When the Proxy receives an encapsulated data packet from an INET 1545 neighbor or from a secured tunnel from a Bridge, it accepts the 1546 packet only if data origin authentication succeeds and if there is a 1547 proxy neighbor cache entry that matches the inner destination. Next, 1548 the Proxy reassembles the packet (if necessary) and continues 1549 processing. If the reassembly is complete and the neighbor cache 1550 state is REACHABLE, the Proxy then returns a PTB if necessary (see: 1551 Section 3.9) then either drops or forwards the packet to the Client 1552 while performing OAL/ORH encapsulation and re-fragmentation if 1553 necessary. If the neighbor cache entry state is DEPARTED, the Proxy 1554 instead changes the destination address to the address of the new 1555 Server and forwards it to a Bridge while performing OAL/ORH re- 1556 fragmentation if necessary. 1558 3.10.3. Server/Relay Forwarding Algorithm 1560 For control messages destined to a target Client's LLA that are 1561 received from a secured tunnel, the Server intercepts the message and 1562 sends a Proxyed response on behalf of the Client. (For example, the 1563 Server sends a Proxyed NA message reply in response to an NS message 1564 directed to one of its associated Clients.) If the Client's neighbor 1565 cache entry is in the DEPARTED state, however, the Server instead 1566 forwards the packet to the Client's new Server as discussed in 1567 Section 3.16. 1569 When the Server receives an encapsulated data packet from an INET 1570 neighbor or from a secured tunnel, it accepts the packet only if data 1571 origin authentication succeeds. The Server then continues processing 1572 as follows: 1574 o if the network layer destination matches a symmetric neighbor 1575 cache entry in the REACHABLE state the Server prepares the packet 1576 for forwarding to the destination Client. The Server first 1577 reassembles (if necessary) and forwards the packet (while re- 1578 fragmenting if necessary) as specified in Section 3.9. 1580 o else, if the destination matches a symmetric neighbor cache entry 1581 in the DEPARETED state the Server re-encapsulates the packet and 1582 forwards it using the ULA of the Client's new Server as the 1583 destination. 1585 o else, if the destination matches an asymmetric neighbor cache 1586 entry, the Server uses one or more "reachable" neighbor interfaces 1587 in the entry for packet forwarding via the local INET if the 1588 neighbor is in the same OMNI link segment or using OAL/ORH 1589 encapsulation if necessary with the final destination set to the 1590 neighbor's ULA otherwise. 1592 o else, if the destination matches a non-MNP route in the IP 1593 forwarding table or an LLA assigned to the Server's OMNI 1594 interface, the Server reassembles if necessary, decapsulates the 1595 packet and releases it to the network layer for local delivery or 1596 IP forwarding. 1598 o else, the Server drops the packet. 1600 When the Server's OMNI interface receives a data packet from the 1601 network layer or from a VPNed or Direct Client, it performs OAL/ORH 1602 encapsulation and fragmentation if necessary, then processes the 1603 packet according to the network-layer destination address as follows: 1605 o if the destination matches a symmetric or asymmetric neighbor 1606 cache entry the Server processes the packet as above. 1608 o else, the Server encapsulates the packet in OLA/ORH headers and 1609 forwards it to a Bridge using its own ULA as the source and the 1610 ULA corresponding to the destination as the destination. 1612 3.10.4. Bridge Forwarding Algorithm 1614 Bridges forward OAL/ORH-encapsulated packets over secured tunnels the 1615 same as any IP router. When the Bridge receives an OAL/ORH- 1616 encapsulated packet via a secured tunnel, it removes the outer INET 1617 header and searches for a forwarding table entry that matches the 1618 destination address. The Bridge then processes the packet as 1619 follows: 1621 o if the destination matches its ULA Subnet Router Anycast address, 1622 the Bridge determines if the next header is an ORH. If so, the 1623 Bridge removes the ORH from the packet while decrementing the OAL 1624 header Payload Length field. If the ORH includes a Destination 1625 Suffix the Bridge also writes the ULA formed from the Destination 1626 Suffix into the OAL header destination address; otherwise, it 1627 writes the ULA formed from the SRT/LHS values. Next, the Bridge 1628 examines the FMT to determine if the target is behind a NAT. If 1629 no NAT is indicated, the Bridge forwards the packet directly to 1630 the L2ADDR using link-layer (UDP/IP) encapsulation. If a NAT is 1631 indicated, the Bridge MAY perform NAT traversal procedures by 1632 sending bubbles per [RFC4380]. The Bridge then either applies 1633 AERO route optimization after NAT traversal procedures have 1634 converged, or simply forwards the packet directly to the Server 1635 indicated by SRT/LHS. 1637 o if the destination matches one of the Bridge's own addresses, the 1638 Bridge submits the packet for local delivery. 1640 o else, if the destination matches a forwarding table entry the 1641 Bridge forwards the packet via a secured tunnel to the next hop. 1642 If the destination matches an MSP without matching an MNP, 1643 however, the Bridge instead drops the packet and returns an ICMP 1644 Destination Unreachable message subject to rate limiting (see: 1645 Section 3.11). 1647 o else, the Bridge drops the packet and returns an ICMP Destination 1648 Unreachable as above. 1650 As for any IP router, the Bridge decrements the TTL/Hop Limit when it 1651 forwards the packet. Therefore, when an OAL header is present only 1652 the Hop Limit in the OAL header is decremented and not the TTL/Hop 1653 Limit in the inner packet header. Bridges do not insert OAL/ORH 1654 headers themselves; instead, they act as IPv6 routers and forward 1655 packets based on the destination address found in the headers of 1656 packets they receive. 1658 3.11. OMNI Interface Error Handling 1660 When an AERO node admits a packet into the OMNI interface, it may 1661 receive link-layer or network-layer error indications. 1663 A link-layer error indication is an ICMP error message generated by a 1664 router in the INET on the path to the neighbor or by the neighbor 1665 itself. The message includes an IP header with the address of the 1666 node that generated the error as the source address and with the 1667 link-layer address of the AERO node as the destination address. 1669 The IP header is followed by an ICMP header that includes an error 1670 Type, Code and Checksum. Valid type values include "Destination 1671 Unreachable", "Time Exceeded" and "Parameter Problem" 1672 [RFC0792][RFC4443]. (OMNI interfaces ignore all link-layer IPv4 1673 "Fragmentation Needed" and IPv6 "Packet Too Big" messages since they 1674 only emit packets that are guaranteed to be no larger than the IP 1675 minimum link MTU as discussed in Section 3.9.) 1676 The ICMP header is followed by the leading portion of the packet that 1677 generated the error, also known as the "packet-in-error". For 1678 ICMPv6, [RFC4443] specifies that the packet-in-error includes: "As 1679 much of invoking packet as possible without the ICMPv6 packet 1680 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1681 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1682 "Internet Header + 64 bits of Original Data Datagram", however 1683 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1684 ICMP datagram SHOULD contain as much of the original datagram as 1685 possible without the length of the ICMP datagram exceeding 576 1686 bytes". 1688 The link-layer error message format is shown in Figure 5 (where, "L2" 1689 and "L3" refer to link-layer and network-layer, respectively): 1691 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1692 ~ ~ 1693 | L2 IP Header of | 1694 | error message | 1695 ~ ~ 1696 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1697 | L2 ICMP Header | 1698 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1699 ~ ~ P 1700 | IP and other encapsulation | a 1701 | headers of original L3 packet | c 1702 ~ ~ k 1703 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1704 ~ ~ t 1705 | IP header of | 1706 | original L3 packet | i 1707 ~ ~ n 1708 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1709 ~ ~ e 1710 | Upper layer headers and | r 1711 | leading portion of body | r 1712 | of the original L3 packet | o 1713 ~ ~ r 1714 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1716 Figure 5: OMNI Interface Link-Layer Error Message Format 1718 The AERO node rules for processing these link-layer error messages 1719 are as follows: 1721 o When an AERO node receives a link-layer Parameter Problem message, 1722 it processes the message the same as described as for ordinary 1723 ICMP errors in the normative references [RFC0792][RFC4443]. 1725 o When an AERO node receives persistent link-layer Time Exceeded 1726 messages, the IP ID field may be wrapping before earlier fragments 1727 awaiting reassembly have been processed. In that case, the node 1728 should begin including integrity checks and/or institute rate 1729 limits for subsequent packets. 1731 o When an AERO node receives persistent link-layer Destination 1732 Unreachable messages in response to encapsulated packets that it 1733 sends to one of its asymmetric neighbor correspondents, the node 1734 should process the message as an indication that a path may be 1735 failing, and optionally initiate NUD over that path. If it 1736 receives Destination Unreachable messages over multiple paths, the 1737 node should allow future packets destined to the correspondent to 1738 flow through a default route and re-initiate route optimization. 1740 o When an AERO Client receives persistent link-layer Destination 1741 Unreachable messages in response to encapsulated packets that it 1742 sends to one of its symmetric neighbor Servers, the Client should 1743 mark the path as unusable and use another path. If it receives 1744 Destination Unreachable messages on many or all paths, the Client 1745 should associate with a new Server and release its association 1746 with the old Server as specified in Section 3.16.5. 1748 o When an AERO Server receives persistent link-layer Destination 1749 Unreachable messages in response to encapsulated packets that it 1750 sends to one of its symmetric neighbor Clients, the Server should 1751 mark the underlying path as unusable and use another underlying 1752 path. 1754 o When an AERO Server or Proxy receives link-layer Destination 1755 Unreachable messages in response to an encapsulated packet that it 1756 sends to one of its permanent neighbors, it treats the messages as 1757 an indication that the path to the neighbor may be failing. 1758 However, the dynamic routing protocol should soon reconverge and 1759 correct the temporary outage. 1761 When an AERO Bridge receives a packet for which the network-layer 1762 destination address is covered by an MSP, if there is no more- 1763 specific routing information for the destination the Bridge drops the 1764 packet and returns a network-layer Destination Unreachable message 1765 subject to rate limiting. The Bridge writes the network-layer source 1766 address of the original packet as the destination address and uses 1767 one of its non link-local addresses as the source address of the 1768 message. 1770 When an AERO node receives an encapsulated packet for which the 1771 reassembly buffer it too small, it drops the packet and returns a 1772 network-layer Packet Too Big (PTB) message. The node first writes 1773 the MRU value into the PTB message MTU field, writes the network- 1774 layer source address of the original packet as the destination 1775 address and writes one of its non link-local addresses as the source 1776 address. 1778 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1780 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1781 coordinated as discussed in the following Sections. 1783 3.12.1. AERO Service Model 1785 Each AERO Server on the OMNI link is configured to facilitate Client 1786 prefix delegation/registration requests. Each Server is provisioned 1787 with a database of MNP-to-Client ID mappings for all Clients enrolled 1788 in the AERO service, as well as any information necessary to 1789 authenticate each Client. The Client database is maintained by a 1790 central administrative authority for the OMNI link and securely 1791 distributed to all Servers, e.g., via the Lightweight Directory 1792 Access Protocol (LDAP) [RFC4511], via static configuration, etc. 1793 Clients receive the same service regardless of the Servers they 1794 select. 1796 AERO Clients and Servers use ND messages to maintain neighbor cache 1797 entries. AERO Servers configure their OMNI interfaces as advertising 1798 NBMA interfaces, and therefore send unicast RA messages with a short 1799 Router Lifetime value (e.g., ReachableTime seconds) in response to a 1800 Client's RS message. Thereafter, Clients send additional RS messages 1801 to keep Server state alive. 1803 AERO Clients and Servers include prefix delegation and/or 1804 registration parameters in RS/RA messages (see 1805 [I-D.templin-6man-omni-interface]). The ND messages are exchanged 1806 between Client and Server according to the prefix management schedule 1807 required by the service. If the Client knows its MNP in advance, it 1808 can employ prefix registration by including its LLA as the source 1809 address of an RS message and with an OMNI option with valid prefix 1810 registration information for the MNP. If the Server (and Proxy) 1811 accept the Client's MNP assertion, they inject the prefix into the 1812 routing system and establish the necessary neighbor cache state. 1814 The following sections specify the Client and Server behavior. 1816 3.12.2. AERO Client Behavior 1818 AERO Clients discover the addresses of Servers in a similar manner as 1819 described in [RFC5214]. Discovery methods include static 1820 configuration (e.g., from a flat-file map of Server addresses and 1821 locations), or through an automated means such as Domain Name System 1822 (DNS) name resolution [RFC1035]. Alternatively, the Client can 1823 discover Server addresses through a layer 2 data link login exchange, 1824 or through a unicast RA response to a multicast/anycast RS as 1825 described below. In the absence of other information, the Client can 1826 resolve the DNS Fully-Qualified Domain Name (FQDN) 1827 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1828 text string and "[domainname]" is a DNS suffix for the OMNI link 1829 (e.g., "example.com"). 1831 To associate with a Server, the Client acts as a requesting router to 1832 request MNPs. The Client prepares an RS message with prefix 1833 management parameters and includes a Nonce and Timestamp option if 1834 the Client needs to correlate RA replies. If the Client already 1835 knows the Server's LLA, it includes the LLA as the network-layer 1836 destination address; otherwise, it includes (link-local) All-Routers 1837 multicast as the network-layer destination. If the Client already 1838 knows its own LLA, it uses the LLA as the network-layer source 1839 address; otherwise, it uses an OMNI Temporary LLA as the network- 1840 layer source address and includes a DHCP Unique Identifier (DUID) 1841 sub-option in the OMNI option (see: 1842 [I-D.templin-6man-omni-interface]). 1844 The Client next includes an OMNI option in the RS message to register 1845 its link-layer information with the Server. The Client sets the OMNI 1846 option prefix registration information according to the MNP, and 1847 includes Interface Attributes corresponding to the underlying 1848 interface over which the Client will send the RS message. The Client 1849 MAY include additional Interface Attributes specific to other 1850 underlying interfaces. 1852 The Client then sends the RS message (either directly via Direct 1853 interfaces, via a VPN for VPNed interfaces, via a Proxy for ANET 1854 interfaces or via INET encapsulation for INET interfaces) and waits 1855 for an RA message reply (see Section 3.12.3). The Client retries up 1856 to MAX_RTR_SOLICITATIONS times until an RA is received. If the 1857 Client receives no RAs, or if it receives an RA with Router Lifetime 1858 set to 0, the Client SHOULD abandon this Server and try another 1859 Server. Otherwise, the Client processes the prefix information found 1860 in the RA message. 1862 Next, the Client creates a symmetric neighbor cache entry with the 1863 Server's LLA as the network-layer address and the Server's 1864 encapsulation and/or link-layer addresses as the link-layer address. 1865 The Client records the RA Router Lifetime field value in the neighbor 1866 cache entry as the time for which the Server has committed to 1867 maintaining the MNP in the routing system via this underlying 1868 interface, and caches the other RA configuration information 1869 including Cur Hop Limit, M and O flags, Reachable Time and Retrans 1870 Timer. The Client then autoconfigures LLAs for each of the delegated 1871 MNPs and assigns them to the OMNI interface. The Client also caches 1872 any MSPs included in Route Information Options (RIOs) [RFC4191] as 1873 MSPs to associate with the OMNI link, and assigns the MTU value in 1874 the MTU option to the underlying interface. 1876 The Client then registers additional underlying interfaces with the 1877 Server by sending RS messages via each additional interface. The RS 1878 messages include the same parameters as for the initial RS/RA 1879 exchange, but with destination address set to the Server's LLA. 1881 Following autoconfiguration, the Client sub-delegates the MNPs to its 1882 attached EUNs and/or the Client's own internal virtual interfaces as 1883 described in [I-D.templin-v6ops-pdhost] to support the Client's 1884 downstream attached "Internet of Things (IoT)". The Client 1885 subsequently sends additional RS messages over each underlying 1886 interface before the Router Lifetime received for that interface 1887 expires. 1889 After the Client registers its underlying interfaces, it may wish to 1890 change one or more registrations, e.g., if an interface changes 1891 address or becomes unavailable, if QoS preferences change, etc. To 1892 do so, the Client prepares an RS message to send over any available 1893 underlying interface. The RS includes an OMNI option with prefix 1894 registration information specific to its MNP, with Interface 1895 Attributes specific to the selected underlying interface, and with 1896 any additional Interface Attributes specific to other underlying 1897 interfaces. When the Client receives the Server's RA response, it 1898 has assurance that the Server has been updated with the new 1899 information. 1901 If the Client wishes to discontinue use of a Server it issues an RS 1902 message over any underlying interface with an OMNI option with a 1903 prefix release indication. When the Server processes the message, it 1904 releases the MNP, sets the symmetric neighbor cache entry state for 1905 the Client to DEPARTED and returns an RA reply with Router Lifetime 1906 set to 0. After a short delay (e.g., 2 seconds), the Server 1907 withdraws the MNP from the routing system. 1909 3.12.3. AERO Server Behavior 1911 AERO Servers act as IP routers and support a prefix delegation/ 1912 registration service for Clients. Servers arrange to add their LLAs 1913 to a static map of Server addresses for the link and/or the DNS 1914 resource records for the FQDN "linkupnetworks.[domainname]" before 1915 entering service. Server addresses should be geographically and/or 1916 topologically referenced, and made available for discovery by Clients 1917 on the OMNI link. 1919 When a Server receives a prospective Client's RS message on its OMNI 1920 interface, it SHOULD return an immediate RA reply with Router 1921 Lifetime set to 0 if it is currently too busy or otherwise unable to 1922 service the Client. Otherwise, the Server authenticates the RS 1923 message and processes the prefix delegation/registration parameters. 1924 The Server first determines the correct MNPs to provide to the Client 1925 by searching the Client database. When the Server returns the MNPs, 1926 it also creates a forwarding table entry for the ULA corresponding to 1927 each MNP so that the MNPs are propagated into the routing system 1928 (see: Section 3.2.3). For IPv6, the Server creates an IPv6 1929 forwarding table entry for each MNP. For IPv4, the Server creates an 1930 IPv6 forwarding table entry with the IPv4-compatibility ULA prefix 1931 corresponding to the IPv4 address. 1933 The Server next creates a symmetric neighbor cache entry for the 1934 Client using the base LLA as the network-layer address and with 1935 lifetime set to no more than the smallest prefix lifetime. Next, the 1936 Server updates the neighbor cache entry by recording the information 1937 in each Interface Attributes sub-option in the RS OMNI option. The 1938 Server also records the actual OAL/INET addresses in the neighbor 1939 cache entry. 1941 Next, the Server prepares an RA message using its LLA as the network- 1942 layer source address and the network-layer source address of the RS 1943 message as the network-layer destination address. The Server sets 1944 the Router Lifetime to the time for which it will maintain both this 1945 underlying interface individually and the symmetric neighbor cache 1946 entry as a whole. The Server also sets Cur Hop Limit, M and O flags, 1947 Reachable Time and Retrans Timer to values appropriate for the OMNI 1948 link. The Server includes the MNPs, any other prefix management 1949 parameters and an OMNI option with no Interface Attributes. The 1950 Server then includes one or more RIOs that encode the MSPs for the 1951 OMNI link, plus an MTU option (see Section 3.9). The Server finally 1952 forwards the message to the Client using OAL/INET, INET, or NULL 1953 encapsulation as necessary. 1955 After the initial RS/RA exchange, the Server maintains a 1956 ReachableTime timer for each of the Client's underlying interfaces 1957 individually (and for the Client's symmetric neighbor cache entry 1958 collectively) set to expire after ReachableTime seconds. If the 1959 Client (or Proxy) issues additional RS messages, the Server sends an 1960 RA response and resets ReachableTime. If the Server receives an ND 1961 message with a prefix release indication it sets the Client's 1962 symmetric neighbor cache entry to the DEPARTED state and withdraws 1963 the MNP from the routing system after a short delay (e.g., 2 1964 seconds). If ReachableTime expires before a new RS is received on an 1965 individual underlying interface, the Server marks the interface as 1966 DOWN. If ReachableTime expires before any new RS is received on any 1967 individual underlying interface, the Server sets the symmetric 1968 neighbor cache entry state to STALE and sets a 10 second timer. If 1969 the Server has not received a new RS or ND message with a prefix 1970 release indication before the 10 second timer expires, it deletes the 1971 neighbor cache entry and withdraws the MNP from the routing system. 1973 The Server processes any ND messages pertaining to the Client and 1974 returns an NA/RA reply in response to solicitations. The Server may 1975 also issue unsolicited RA messages, e.g., with reconfigure parameters 1976 to cause the Client to renegotiate its prefix delegation/ 1977 registrations, with Router Lifetime set to 0 if it can no longer 1978 service this Client, etc. Finally, If the symmetric neighbor cache 1979 entry is in the DEPARTED state, the Server deletes the entry after 1980 DepartTime expires. 1982 Note: Clients SHOULD notify former Servers of their departures, but 1983 Servers are responsible for expiring neighbor cache entries and 1984 withdrawing routes even if no departure notification is received 1985 (e.g., if the Client leaves the network unexpectedly). Servers 1986 SHOULD therefore set Router Lifetime to ReachableTime seconds in 1987 solicited RA messages to minimize persistent stale cache information 1988 in the absence of Client departure notifications. A short Router 1989 Lifetime also ensures that proactive Client/Server RS/RA messaging 1990 will keep any NAT state alive (see above). 1992 Note: All Servers on an OMNI link MUST advertise consistent values in 1993 the RA Cur Hop Limit, M and O flags, Reachable Time and Retrans Timer 1994 fields the same as for any link, since unpredictable behavior could 1995 result if different Servers on the same link advertised different 1996 values. 1998 3.12.3.1. DHCPv6-Based Prefix Registration 2000 When a Client is not pre-provisioned with an OMNI LLA containing a 2001 MNP, it will need for the Server to select one or more MNPs on its 2002 behalf and set up the correct state in the AERO routing service. (A 2003 Client with a pre-provisioned MNP may also request the Server to 2004 select additional MNPs.) The DHCPv6 service [RFC8415] is used to 2005 support this requirement. 2007 When a Client needs to have the Server select MNPs, it sends an RS 2008 message with an OMNI option that includes a DHCPv6 message suboption 2009 with DHCPv6 Prefix Delegation (DHCPv6-PD) parameters. When the 2010 Server receives the RS message, it extracts the DHCPv6-PD message 2011 from the OMNI option. 2013 The Server then acts as a "Proxy DHCPv6 Client" in a message exchange 2014 with the locally-resident DHCPv6 server, which delegates MNPs and 2015 returns a DHCPv6-PD Reply message. (If the Server wishes to defer 2016 creation of MN state until the DHCPv6-PD Reply is received, it can 2017 instead act as a Lightweight DHCPv6 Relay Agent per [RFC6221] by 2018 encapsulating the DHCPv6-PD message in a Relay-forward/reply exchange 2019 with Relay Message and Interface ID options.) 2021 When the Server receives the DHCPv6-PD Reply, it adds a route to the 2022 routing system and creates an OMNI MN LLA based on the delegated MNP. 2023 The Server then sends an RA back to the Client with the (newly- 2024 created) OMNI MN LLA as the destination address and with the 2025 DHCPv6-PD Reply message coded in the OMNI option. When the Client 2026 receives the RA, it creates a default route, assigns the Subnet 2027 Router Anycast address and sets its OMNI LLA based on the delegated 2028 MNP. 2030 3.13. The AERO Proxy 2032 Clients may connect to protected-spectrum ANETs that employ physical 2033 and/or link-layer security services to facilitate communications to 2034 Servers in outside INETs. In that case, the ANET can employ an AERO 2035 Proxy. The Proxy is located at the ANET/INET border and listens for 2036 RS messages originating from or RA messages destined to ANET Clients. 2037 The Proxy acts on these control messages as follows: 2039 o when the Proxy receives an RS message from a new ANET Client, it 2040 first authenticates the message then examines the network-layer 2041 destination address. If the destination address is a Server's 2042 LLA, the Proxy proceeds to the next step. Otherwise, if the 2043 destination is (link-local) All-Routers multicast, the Proxy 2044 selects a "nearby" Server that is likely to be a good candidate to 2045 serve the Client and replaces the destination address with the 2046 Server's LLA. Next, the Proxy creates a proxy neighbor cache 2047 entry and caches the Client and Server link-layer addresses along 2048 with the OMNI option information and any other identifying 2049 information including Transaction IDs, Client Identifiers, Nonce 2050 values, etc. The Proxy finally encapsulates the (proxyed) RS 2051 message in an OAL header with source set to the Proxy's ULA and 2052 destination set to the Server's ULA. The Proxy also includes an 2053 OMNI header with an Interface Attributes option that includes its 2054 own INET address plus a unique Port Number for this Client, then 2055 forwards the message into the OMNI link spanning tree. 2057 o when the Server receives the RS, it authenticates the message then 2058 creates or updates a symmetric neighbor cache entry for the Client 2059 with the Proxy's ULA, INET address and Port Number as the link- 2060 layer address information. The Server then sends an RA message 2061 back to the Proxy via the spanning tree. 2063 o when the Proxy receives the RA, it authenticates the message and 2064 matches it with the proxy neighbor cache entry created by the RS. 2065 The Proxy then caches the prefix information as a mapping from the 2066 Client's MNPs to the Client's link-layer address, caches the 2067 Server's advertised Router Lifetime and sets the neighbor cache 2068 entry state to REACHABLE. The Proxy then optionally rewrites the 2069 Router Lifetime and forwards the (proxyed) message to the Client. 2070 The Proxy finally includes an MTU option (if necessary) with an 2071 MTU to use for the underlying ANET interface. 2073 After the initial RS/RA exchange, the Proxy forwards any Client data 2074 packets for which there is no matching asymmetric neighbor cache 2075 entry to a Bridge using OAL encapsulation with its own ULA as the 2076 source and the ULA corresponding to the Client as the destination. 2077 The Proxy instead forwards any Client data destined to an asymmetric 2078 neighbor cache target directly to the target according to the OAL/ 2079 link-layer information - the process of establishing asymmetric 2080 neighbor cache entries is specified in Section 3.14. 2082 While the Client is still attached to the ANET, the Proxy sends NS, 2083 RS and/or unsolicited NA messages to update the Server's symmetric 2084 neighbor cache entries on behalf of the Client and/or to convey QoS 2085 updates. This allows for higher-frequency Proxy-initiated RS/RA 2086 messaging over well-connected INET infrastructure supplemented by 2087 lower-frequency Client-initiated RS/RA messaging over constrained 2088 ANET data links. 2090 If the Server ceases to send solicited advertisements, the Proxy 2091 sends unsolicited RAs on the ANET interface with destination set to 2092 (link-local) All-Nodes multicast and with Router Lifetime set to zero 2093 to inform Clients that the Server has failed. Although the Proxy 2094 engages in ND exchanges on behalf of the Client, the Client can also 2095 send ND messages on its own behalf, e.g., if it is in a better 2096 position than the Proxy to convey QoS changes, etc. For this reason, 2097 the Proxy marks any Client-originated solicitation messages (e.g. by 2098 inserting a Nonce option) so that it can return the solicited 2099 advertisement to the Client instead of processing it locally. 2101 If the Client becomes unreachable, the Proxy sets the neighbor cache 2102 entry state to DEPARTED and retains the entry for DepartTime seconds. 2103 While the state is DEPARTED, the Proxy forwards any packets destined 2104 to the Client to a Bridge via OAL encapsulation with the Client's 2105 current Server as the destination. The Bridge in turn forwards the 2106 packets to the Client's current Server. When DepartTime expires, the 2107 Proxy deletes the neighbor cache entry and discards any further 2108 packets destined to this (now forgotten) Client. 2110 In some ANETs that employ a Proxy, the Client's MNP can be injected 2111 into the ANET routing system. In that case, the Client can send data 2112 messages without encapsulation so that the ANET routing system 2113 transports the unencapsulated packets to the Proxy. This can be very 2114 beneficial, e.g., if the Client connects to the ANET via low-end data 2115 links such as some aviation wireless links. 2117 If the first-hop ANET access router is on the same underlying link 2118 and recognizes the AERO/OMNI protocol, the Client can avoid 2119 encapsulation for both its control and data messages. When the 2120 Client connects to the link, it can send an unencapsulated RS message 2121 with source address set to its LLA and with destination address set 2122 to the LLA of the Client's selected Server or to (link-local) All- 2123 Routers multicast. The Client includes an OMNI option formatted as 2124 specified in [I-D.templin-6man-omni-interface]. 2126 The Client then sends the unencapsulated RS message, which will be 2127 intercepted by the AERO-Aware access router. The access router then 2128 encapsulates the RS message in an ANET header with its own address as 2129 the source address and the address of a Proxy as the destination 2130 address. The access router further remembers the address of the 2131 Proxy so that it can encapsulate future data packets from the Client 2132 via the same Proxy. If the access router needs to change to a new 2133 Proxy, it simply sends another RS message toward the Server via the 2134 new Proxy on behalf of the Client. 2136 In some cases, the access router and Proxy may be one and the same 2137 node. In that case, the node would be located on the same physical 2138 link as the Client, but its message exchanges with the Server would 2139 need to pass through a security gateway at the ANET/INET border. The 2140 method for deploying access routers and Proxys (i.e. as a single node 2141 or multiple nodes) is an ANET-local administrative consideration. 2143 3.13.1. Combined Proxy/Servers 2145 Clients may need to connect directly to Servers via INET, Direct and 2146 VPNed interfaces (i.e., non-ANET interfaces). If the Client's 2147 underlying interfaces all connect via the same INET partition, then 2148 it can connect to a single controlling Server via all interfaces. 2150 If some Client interfaces connect via different INET partitions, 2151 however, the Client still selects a set of controlling Servers and 2152 sends RS messages via their directly-connected Servers while using 2153 the LLA of the controlling Server as the destination. 2155 When a Server receives an RS with destination set to the LLA of a 2156 controlling Server, it acts as a Proxy to forward the message to the 2157 controlling Server while forwarding the corresponding RA reply to the 2158 Client. 2160 3.13.2. Detecting and Responding to Server Failures 2162 In environments where fast recovery from Server failure is required, 2163 Proxys SHOULD use proactive Neighbor Unreachability Detection (NUD) 2164 to track Server reachability in a similar fashion as for 2165 Bidirectional Forwarding Detection (BFD) [RFC5880]. Proxys can then 2166 quickly detect and react to failures so that cached information is 2167 re-established through alternate paths. The NUD control messaging is 2168 carried only over well-connected ground domain networks (i.e., and 2169 not low-end aeronautical radio links) and can therefore be tuned for 2170 rapid response. 2172 Proxys perform proactive NUD with Servers for which there are 2173 currently active ANET Clients by sending continuous NS messages in 2174 rapid succession, e.g., one message per second. The Proxy sends the 2175 NS message via the spanning tree with the Proxy's LLA as the source 2176 and the LLA of the Server as the destination. When the Proxy is also 2177 sending RS messages to the Server on behalf of ANET Clients, the 2178 resulting RA responses can be considered as equivalent hints of 2179 forward progress. This means that the Proxy need not also send a 2180 periodic NS if it has already sent an RS within the same period. If 2181 the Server fails (i.e., if the Proxy ceases to receive 2182 advertisements), the Proxy can quickly inform Clients by sending 2183 multicast RA messages on the ANET interface. 2185 The Proxy sends RA messages on the ANET interface with source address 2186 set to the Server's address, destination address set to (link-local) 2187 All-Nodes multicast, and Router Lifetime set to 0. The Proxy SHOULD 2188 send MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small 2189 delays [RFC4861]. Any Clients on the ANET that had been using the 2190 failed Server will receive the RA messages and associate with a new 2191 Server. 2193 3.13.3. Point-to-Multipoint Server Coordination 2195 In environments where Client messaging over ANETs is bandwidth- 2196 limited and/or expensive, Clients can enlist the services of the 2197 Proxy to coordinate with multiple Servers in a single RS/RA message 2198 exchange. The Client can send a single RS message to (link-local) 2199 All-Routers multicast that includes the ID's of multiple Servers in 2200 MS-Register sub-options of the OMNI option. 2202 When the Proxy receives the RS and processes the OMNI option, it 2203 sends a separate RS to each MS-Register Server ID. When the Proxy 2204 receives an RA, it can optionally return an immediate "singleton" RA 2205 to the Client or record the Server's ID for inclusion in a pending 2206 "aggregate" RA message. The Proxy can then return aggregate RA 2207 messages to the Client including multiple Server IDs in order to 2208 conserve bandwidth. Each RA includes a proper subset of the Server 2209 IDs from the original RS message, and the Proxy must ensure that the 2210 message contents of each RA are consistent with the information 2211 received from the (aggregated) Servers. 2213 Clients can thereafter employ efficient point-to-multipoint Server 2214 coordination under the assistance of the Proxy to reduce the number 2215 of messages sent over the ANET while enlisting the support of 2216 multiple Servers for fault tolerance. Clients can further include 2217 MS-Release sub-options in IPv6 ND messages to request the Proxy to 2218 release from former Servers via the procedures discussed in 2219 Section 3.16.5. 2221 The OMNI interface specification [I-D.templin-6man-omni-interface] 2222 provides further discussion of the Client/Proxy RS/RA messaging 2223 involved in point-to-multipoint coordination. 2225 3.14. AERO Route Optimization / Address Resolution 2227 While data packets are flowing between a source and target node, 2228 route optimization SHOULD be used. Route optimization is initiated 2229 by the first eligible Route Optimization Source (ROS) closest to the 2230 source as follows: 2232 o For Clients on VPNed and Direct interfaces, the Server is the ROS. 2234 o For Clients on ANET interfaces, the Proxy is the ROS. 2236 o For Clients on INET interfaces, the Client itself is the ROS. 2238 o For correspondent nodes on INET/EUN interfaces serviced by a 2239 Relay, the Relay is the ROS. 2241 The route optimization procedure is conducted between the ROS and the 2242 target Server/Relay acting as a Route Optimization Responder (ROR) in 2243 the same manner as for IPv6 ND Address Resolution and using the same 2244 NS/NA messaging. The target may either be a MNP Client serviced by a 2245 Server, or a non-MNP correspondent reachable via a Relay. 2247 The procedures are specified in the following sections. 2249 3.14.1. Route Optimization Initiation 2251 While data packets are flowing from the source node toward a target 2252 node, the ROS performs address resolution by sending an NS message 2253 for Address Resolution (NS(AR)) to receive a solicited NA message 2254 from the ROR. When the ROS sends an NS(AR), it includes: 2256 o the LLA of the ROS as the source address. 2258 o the data packet's destination as the Target Address. 2260 o the Solicited-Node multicast address [RFC4291] formed from the 2261 lower 24 bits of the data packet's destination as the destination 2262 address, e.g., for 2001:db8:1:2::10:2000 the NS destination 2263 address is ff02:0:0:0:0:1:ff10:2000. 2265 The NS(AR) message includes an OMNI option with no Interface 2266 Attributes, such that the target will not create a neighbor cache 2267 entry. The Prefix Length in the OMNI option is set to the Prefix 2268 Length associated with the ROS's LLA. 2270 The ROS then encapsulates the NS(AR) message in an OAL header with 2271 source set to its own ULA and destination set to the ULA 2272 corresponding to the target, then sends the message into the spanning 2273 tree without decrementing the network-layer TTL/Hop Limit field. 2274 (When the ROS is a Client, it instead securely sends the NS(AR) to 2275 one of its current Servers by including an Authentication option per 2276 [RFC4380]. The Server then forwards the message into the spanning 2277 tree on behalf of the Client, while setting the IPv6 source address 2278 and the OAL source address to the LLA and ULA of the Client, 2279 respectively.) 2281 3.14.2. Relaying the NS 2283 When the Bridge receives the NS(AR) message from the ROS, it discards 2284 the INET header and determines that the ROR is the next hop by 2285 consulting its standard IPv6 forwarding table for the OAL header 2286 destination address. The Bridge then forwards the message toward the 2287 ROR via the spanning tree the same as for any IPv6 router. The 2288 final-hop Bridge in the spanning tree will deliver the message via a 2289 secured tunnel to the ROR. 2291 3.14.3. Processing the NS and Sending the NA 2293 When the ROR receives the NS(AR) message, it examines the Target 2294 Address to determine whether it has a neighbor cache entry and/or 2295 route that matches the target. If there is no match, the ROR drops 2296 the message. Otherwise, the ROR continues processing as follows: 2298 o if the target belongs to an MNP Client neighbor in the DEPARTED 2299 state the ROR changes the NS(AR) message OAL destination address 2300 to the ULA of the Client's new Server, forwards the message into 2301 the spanning tree and returns from processing. 2303 o If the target belongs to an MNP Client neighbor in the REACHABLE 2304 state, the ROR instead adds the AERO source address to the target 2305 Client's Report List with time set to ReportTime. 2307 o If the target belongs to a non-MNP route, the ROR continues 2308 processing without adding an entry to the Report List. 2310 The ROR then prepares a solicited NA message to send back to the ROS 2311 but does not create a neighbor cache entry. The ROR sets the NA 2312 source address to the LLA corresponding to the target, sets the 2313 Target Address to the target of the solicitation, and sets the 2314 destination address to the source of the solicitation. The ROR then 2315 includes an OMNI option with Prefix Length set to the length 2316 associated with the LLA. 2318 If the target is an MNP Client, the ROR next includes Interface 2319 Attributes in the OMNI option for each of the target Client's 2320 underlying interfaces with current information for each interface and 2321 with the S/T-ifIndex field in the OMNI header set to 0 to indicate 2322 that the message originated from the ROR and not the Client. 2324 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2325 according to its own INET address for VPNed or Direct interfaces, to 2326 the INET address of the Proxy or to the Client's INET address for 2327 INET interfaces. The ROR then includes the lower 32 bits of its own 2328 ULA (or the ULA of the Proxy) as the LHS, encodes the ULA prefix 2329 length code in the SRT field and sets the FMT code accordingly as 2330 specified in Section 3.3. 2332 The ROR then sets the NA message R flag to 1 (as a router), S flag to 2333 1 (as a response to a solicitation), and O flag to 0 (as a proxy). 2334 The ROR finally encapsulates the NA message in an OAL header with 2335 source set to its own ULA and destination set to the source ULA of 2336 the NS(AR) message, then forwards the message into the spanning tree 2337 without decrementing the network-layer TTL/Hop Limit field. 2339 3.14.4. Relaying the NA 2341 When the Bridge receives the NA message from the ROR, it discards the 2342 INET header and determines that the ROS is the next hop by consulting 2343 its standard IPv6 forwarding table for the OAL header destination 2344 address. The Bridge then forwards the OAL-encapsulated NA message 2345 toward the ROS the same as for any IPv6 router. The final-hop Bridge 2346 in the spanning tree will deliver the message via a secured tunnel to 2347 the ROS. 2349 3.14.5. Processing the NA 2351 When the ROS receives the solicited NA message, it processes the 2352 message the same as for standard IPv6 Address Resolution [RFC4861]. 2353 In the process, it caches the source ULA then creates an asymmetric 2354 neighbor cache entry for the target and caches all information found 2355 in the OMNI option. The ROS finally sets the asymmetric neighbor 2356 cache entry lifetime to ReachableTime seconds. (When the ROS is a 2357 Client, the solicited NA message will first be delivered via the 2358 spanning tree to one of its current Servers, which then securely 2359 forwards the message to the Client by including an Authentication 2360 option per [RFC4380]. 2362 3.14.6. Route Optimization Maintenance 2364 Following route optimization, the ROS forwards future data packets 2365 destined to the target via the addresses found in the cached link- 2366 layer information. The route optimization is shared by all sources 2367 that send packets to the target via the ROS, i.e., and not just the 2368 source on behalf of which the route optimization was initiated. 2370 While new data packets destined to the target are flowing through the 2371 ROS, it sends additional NS(AR) messages to the ROR before 2372 ReachableTime expires to receive a fresh solicited NA message the 2373 same as described in the previous sections (route optimization 2374 refreshment strategies are an implementation matter, with a non- 2375 normative example given in Appendix A.1). The ROS uses the cached 2376 ULA of the ROR as the NS(AR) OAL destination address (i.e., instead 2377 of using the ULA corresponding to the target as was the case for the 2378 initial NS(AR)), and sends up to MAX_MULTICAST_SOLICIT NS(AR) 2379 messages separated by 1 second until an NA is received. If no NA is 2380 received, the ROS assumes that the current ROR has become unreachable 2381 and deletes the target neighbor cache entry. Subsequent data packets 2382 will trigger a new route optimization with an NS with OAL destination 2383 address set to the ULA corresponding to the target per Section 3.14.1 2384 to discover a new ROR while initial data packets travel over a 2385 suboptimal route. 2387 If an NA is received, the ROS then updates the asymmetric neighbor 2388 cache entry to refresh ReachableTime, while (for MNP destinations) 2389 the ROR adds or updates the ROS address to the target's Report List 2390 and with time set to ReportTime. While no data packets are flowing, 2391 the ROS instead allows ReachableTime for the asymmetric neighbor 2392 cache entry to expire. When ReachableTime expires, the ROS deletes 2393 the asymmetric neighbor cache entry. Any future data packets flowing 2394 through the ROS will again trigger a new route optimization. 2396 The ROS may also receive unsolicited NA messages from the ROR at any 2397 time (see: Section 3.16). If there is an asymmetric neighbor cache 2398 entry for the target, the ROS updates the link-layer information but 2399 does not update ReachableTime since the receipt of an unsolicited NA 2400 does not confirm that any forward paths are working. If there is no 2401 asymmetric neighbor cache entry, the ROS simply discards the 2402 unsolicited NA. 2404 In this arrangement, the ROS holds an asymmetric neighbor cache entry 2405 for the target via the ROR, but the ROR does not hold an asymmetric 2406 neighbor cache entry for the ROS. The route optimization neighbor 2407 relationship is therefore asymmetric and unidirectional. If the 2408 target node also has packets to send back to the source node, then a 2409 separate route optimization procedure is performed in the reverse 2410 direction. But, there is no requirement that the forward and reverse 2411 paths be symmetric. 2413 3.15. Neighbor Unreachability Detection (NUD) 2415 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2416 [RFC4861] either reactively in response to persistent link-layer 2417 errors (see Section 3.11) or proactively to confirm reachability. 2418 The NUD algorithm is based on periodic control message exchanges. 2419 The algorithm may further be seeded by ND hints of forward progress, 2420 but care must be taken to avoid inferring reachability based on 2421 spoofed information. For example, authentic IPv6 ND message 2422 exchanges may be considered as acceptable hints of forward progress, 2423 while spurious data packets should not be. 2425 AERO Servers, Proxys and Relays can use (OAL-encapsulated) standard 2426 NS/NA NUD exchanges sent over the OMNI link spanning tree to securely 2427 test reachability without risk of DoS attacks from nodes pretending 2428 to be a neighbor (these NS/NA(NUD) messages use the unicast LLAs and 2429 ULAs of the two parties involved in the NUD test the same as for 2430 standard IPv6 ND). Proxys can further perform NUD to securely verify 2431 Server reachability on behalf of their proxyed Clients. However, a 2432 means for an ROS to test the unsecured forward directions of target 2433 route optimized paths is also necessary. 2435 When an ROR directs an ROS to a neighbor with one or more target 2436 link-layer addresses, the ROS can proactively test each such 2437 unsecured route optimized path by sending "loopback" NS(NUD) 2438 messages. While testing the paths, the ROS can optionally continue 2439 to send packets via the spanning tree, maintain a small queue of 2440 packets until target reachability is confirmed, or (optimistically) 2441 allow packets to flow via the route optimized paths. 2443 When the ROS sends a loopback NS(NUD) message, it uses its own LLA as 2444 both the IPv6 source and destination address, and the MNP Subnet- 2445 Router anycast address as the Target Address. The ROS includes a 2446 Nonce and Timestamp option, then encapsulates the message in OAL/INET 2447 headers with its own ULA as the source and the ULA of the route 2448 optimization target as the destination. The ROS then forwards the 2449 message to the target (either directly to the L2ADDR of the target if 2450 the target is in the same OMNI link segment, or via a Bridge if the 2451 target is in a different OMNI link segment). 2453 When the route optimization target receives the NS(NUD) message, it 2454 notices that the IPv6 destination address is the same as the source 2455 address. It then reverses the OAL header source and destination 2456 addresses and returns the message to the ROS (either directly or via 2457 the spanning tree). The route optimization target does not decrement 2458 the NS(NUD) message IPv6 Hop-Limit in the process, since the message 2459 has not exited the OMNI link. 2461 When the ROS receives the NS(NUD) message, it can determine from the 2462 Nonce, Timestamp and Target Address that the message originated from 2463 itself and that it transited the forward path. The ROS need not 2464 prepare an NA response, since the destination of the response would 2465 be itself and testing the route optimization path again would be 2466 redundant. 2468 The ROS marks route optimization target paths that pass these NUD 2469 tests as "reachable", and those that do not as "unreachable". These 2470 markings inform the OMNI interface forwarding algorithm specified in 2471 Section 3.10. 2473 Note that to avoid a DoS vector nodes MUST NOT return loopback 2474 NS(NUD) messages received from an unsecured link-layer source via the 2475 (secured) spanning tree. 2477 3.16. Mobility Management and Quality of Service (QoS) 2479 AERO is a Distributed Mobility Management (DMM) service. Each Server 2480 is responsible for only a subset of the Clients on the OMNI link, as 2481 opposed to a Centralized Mobility Management (CMM) service where 2482 there is a single network mobility collective entity for all Clients. 2483 Clients coordinate with their associated Servers via RS/RA exchanges 2484 to maintain the DMM profile, and the AERO routing system tracks all 2485 current Client/Server peering relationships. 2487 Servers provide default routing and mobility/multilink services for 2488 their dependent Clients. Clients are responsible for maintaining 2489 neighbor relationships with their Servers through periodic RS/RA 2490 exchanges, which also serves to confirm neighbor reachability. When 2491 a Client's underlying interface address and/or QoS information 2492 changes, the Client is responsible for updating the Server with this 2493 new information. Note that when there is a Proxy in the path, the 2494 Proxy can also perform some RS/RA exchanges on the Client's behalf. 2496 Mobility management messaging is based on the transmission and 2497 reception of unsolicited Neighbor Advertisement (uNA) messages. Each 2498 uNA message sets the IPv6 destination address to (link-local) All- 2499 Nodes multicast to convey a general update of Interface Attributes to 2500 (possibly) multiple recipients, or to a specific unicast LLA to 2501 announce a departure event to a specific recipient. Implementations 2502 must therefore examine the destination address to determine the 2503 nature of the mobility event (i.e., update vs departure). 2505 Mobility management considerations are specified in the following 2506 sections. 2508 3.16.1. Mobility Update Messaging 2510 Servers accommodate Client mobility, multilink and/or QoS change 2511 events by sending unsolicited NA (uNA) messages to each ROS in the 2512 target Client's Report List. When a Server sends a uNA message, it 2513 sets the IPv6 source address to the Client's LLA, sets the 2514 destination address to (link-local) All-Nodes multicast and sets the 2515 Target Address to the Client's Subnet-Router anycast address. The 2516 Server also includes an OMNI option with Prefix Length set to the 2517 length associated with the Client's LLA, with Interface Attributes 2518 for the target Client's underlying interfaces and with the OMNI 2519 header S/T-ifIndex set to 0. The Server then sets the NA R flag to 2520 1, the S flag to 0 and the O flag to 1, then encapsulates the message 2521 in an OAL header with source set to its own ULA and destination set 2522 to the ULA of the ROS and sends the message into the spanning tree. 2524 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2525 reception of uNA messages is unreliable but provides a useful 2526 optimization. In well-connected Internetworks with robust data links 2527 uNA messages will be delivered with high probability, but in any case 2528 the Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT uNAs 2529 to each ROS to increase the likelihood that at least one will be 2530 received. 2532 When the ROS receives a uNA message prepared as above, it ignores the 2533 message if there is no existing neighbor cache entry for the Client. 2534 Otherwise, it uses the included OMNI option information to update the 2535 neighbor cache entry, but does not reset ReachableTime since the 2536 receipt of an unsolicited NA message from the target Server does not 2537 provide confirmation that any forward paths to the target Client are 2538 working. 2540 If uNA messages are lost, the ROS may be left with stale address and/ 2541 or QoS information for the Client for up to ReachableTime seconds. 2542 During this time, the ROS can continue sending packets according to 2543 its stale neighbor cache information. When ReachableTime is close to 2544 expiring, the ROS will re-initiate route optimization and receive 2545 fresh link-layer address information. 2547 In addition to sending uNA messages to the current set of ROSs for 2548 the Client, the Server also sends uNAs to the ULA associated with the 2549 link-layer address for any underlying interface for which the link- 2550 layer address has changed. These uNA messages update an old Proxy/ 2551 Server that cannot easily detect (e.g., without active probing) when 2552 a formerly-active Client has departed. When the Server sends the 2553 uNA, it sets the IPv6 source address to the Client's LLA, sets the 2554 destination address to the old Proxy/Server's LLA, and sets the 2555 Target Address to the Client's Subnet-Router anycast address. The 2556 Server also includes an OMNI option with Prefix Length set to the 2557 length associated with the Client's LLA, with Interface Attributes 2558 for the changed underlying interface, and with the OMNI header S/ 2559 T-ifIndex set to 0. The Server then sets the NA R flag to 1, the S 2560 flag to 0 and the O flag to 1, then encapsulates the message in an 2561 OAL header with source set to its own ULA and destination set to the 2562 ULA of the old Proxy/Server and sends the message into the spanning 2563 tree. 2565 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2567 When a Client needs to change its underlying interface addresses and/ 2568 or QoS preferences (e.g., due to a mobility event), either the Client 2569 or its Proxys send RS messages to the Server via the spanning tree 2570 with an OMNI option that includes Interface attributes with the new 2571 link quality and address information. 2573 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2574 sending actual data packets in case one or more RAs are lost. If all 2575 RAs are lost, the Client SHOULD re-associate with a new Server. 2577 When the Server receives the Client's changes, it sends uNA messages 2578 to all nodes in the Report List the same as described in the previous 2579 section. 2581 3.16.3. Bringing New Links Into Service 2583 When a Client needs to bring new underlying interfaces into service 2584 (e.g., when it activates a new data link), it sends an RS message to 2585 the Server via the underlying interface with an OMNI option that 2586 includes Interface Attributes with appropriate link quality values 2587 and with link-layer address information for the new link. 2589 3.16.4. Deactivating Existing Links 2591 When a Client needs to deactivate an existing underlying interface, 2592 it sends an RS or uNA message to its Server with an OMNI option with 2593 appropriate Interface Attribute values - in particular, the link 2594 quality value 0 assures that neighbors will cease to use the link. 2596 If the Client needs to send RS/uNA messages over an underlying 2597 interface other than the one being deactivated, it MUST include 2598 Interface Attributes with appropriate link quality values for any 2599 underlying interfaces being deactivated. 2601 Note that when a Client deactivates an underlying interface, 2602 neighbors that have received the RS/uNA messages need not purge all 2603 references for the underlying interface from their neighbor cache 2604 entries. The Client may reactivate or reuse the underlying interface 2605 and/or its ifIndex at a later point in time, when it will send RS/uNA 2606 messages with fresh Interface Attributes to update any neighbors. 2608 3.16.5. Moving Between Servers 2610 The Client performs the procedures specified in Section 3.12.2 when 2611 it first associates with a new Server or renews its association with 2612 an existing Server. The Client also includes MS-Release identifiers 2613 in the RS message OMNI option per [I-D.templin-6man-omni-interface] 2614 if it wants the new Server to notify any old Servers from which the 2615 Client is departing. 2617 When the new Server receives the Client's RS message, it returns an 2618 RA as specified in Section 3.12.3 and sends up to 2619 MAX_NEIGHBOR_ADVERTIISEMENT uNA messages to any old Servers listed in 2620 OMNI option MS-Release identifiers. When the new Server sends a uNA 2621 message, it sets the IPv6 source address to the Client's LLA, sets 2622 the destination address to the old Server's LLA, and sets the Target 2623 Address to the Client's Subnet-Router anycast address. The new 2624 Server also includes an OMNI option with Prefix Length set to the 2625 length associated with the Client's LLA, with Interface Attributes 2626 for its own underlying interface, and with the OMNI header S/ 2627 T-ifIndex set to 0. The new Server then sets the NA R flag to 1, the 2628 S flag to 0 and the O flag to 1, then encapsulates the message in an 2629 OAL header with source set to its own ULA and destination set to the 2630 ULA of the old Server and sends the message into the spanning tree. 2632 When an old Server receives the uNA, it changes the Client's neighbor 2633 cache entry state to DEPARTED, sets the link-layer address of the 2634 Client to the new Server's ULA, and resets DepartTime. After a short 2635 delay (e.g., 2 seconds) the old Server withdraws the Client's MNP 2636 from the routing system. After DepartTime expires, the old Server 2637 deletes the Client's neighbor cache entry. 2639 The old Server also iteratively forwards a copy of the uNA message to 2640 each ROS in the Client's Report List by changing the OAL destination 2641 address to the ULA of the ROS while leaving all other fields of the 2642 message unmodified. When the ROS receives the uNA, it examines the 2643 Target address to determine the correct asymmetric neighbor cache 2644 entry and verifies that the IPv6 destination address matches the old 2645 Server. The ROS then caches the IPv6 source address as the new 2646 Server for the existing asymmetric neighbor cache entry and marks the 2647 entry as STALE. While in the STALE state, the ROS allows new data 2648 packets to flow according to any existing cached link-layer 2649 information and sends new NS(AR) messages using its own ULA as the 2650 OAL source and the ULA of the new Server as the OAL destination 2651 address to elicit NA messages that reset the asymmetric neighbor 2652 cache entry state to REACHABLE. If no new NA message is received for 2653 10 seconds while in the STALE state, the ROS deletes the neighbor 2654 cache entry. 2656 Clients SHOULD NOT move rapidly between Servers in order to avoid 2657 causing excessive oscillations in the AERO routing system. Examples 2658 of when a Client might wish to change to a different Server include a 2659 Server that has gone unreachable, topological movements of 2660 significant distance, movement to a new geographic region, movement 2661 to a new OMNI link segment, etc. 2663 When a Client moves to a new Server, some of the fragments of a 2664 multiple fragment packet may have already arrived at the old Server 2665 while others are en route to the new Server, however no special 2666 attention in the reassembly algorithm is necessary when re-routed 2667 fragments are simply treated as loss. 2669 3.17. Multicast 2671 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2672 [RFC3810] proxy service for its EUNs and/or hosted applications 2673 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2674 underlying interfaces for which group membership is required. The 2675 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2676 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2677 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2678 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2679 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 2680 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2681 INET/EUN networks. The behaviors identified in the following 2682 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2683 Multicast (ASM) operational modes. 2685 3.17.1. Source-Specific Multicast (SSM) 2687 When an ROS (i.e., an AERO Proxy/Server/Relay) "X" acting as PIM 2688 router receives a Join/Prune message from a node on its downstream 2689 interfaces containing one or more ((S)ource, (G)roup) pairs, it 2690 updates its Multicast Routing Information Base (MRIB) accordingly. 2691 For each S belonging to a prefix reachable via X's non-OMNI 2692 interfaces, X then forwards the (S, G) Join/Prune to any PIM routers 2693 on those interfaces per [RFC7761]. 2695 For each S belonging to a prefix reachable via X's OMNI interface, X 2696 originates a separate copy of the Join/Prune for each (S,G) in the 2697 message using its own LLA as the source address and ALL-PIM-ROUTERS 2698 as the destination address. X then encapsulates each message in an 2699 OAL header with source address set to the ULA of X and destination 2700 address set to S then forwards the message into the spanning tree, 2701 which delivers it to AERO Server/Relay "Y" that services S. At the 2702 same time, if the message was a Join, X sends a route-optimization NS 2703 message toward each S the same as discussed in Section 3.14. The 2704 resulting NAs will return the LLA for the prefix that matches S as 2705 the network-layer source address and with an OMNI option with the ULA 2706 corresponding to any underlying interfaces that are currently 2707 servicing S. 2709 When Y processes the Join/Prune message, if S located behind any 2710 INET, Direct, or VPNed interfaces Y acts as a PIM router and updates 2711 its MRIB to list X as the next hop in the reverse path. If S is 2712 located behind any Proxys "Z"*, Y also forwards the message to each 2713 Z* over the spanning tree while continuing to use the LLA of X as the 2714 source address. Each Z* then updates its MRIB accordingly and 2715 maintains the LLA of X as the next hop in the reverse path. Since 2716 the Bridges do not examine network layer control messages, this means 2717 that the (reverse) multicast tree path is simply from each Z* (and/or 2718 Y) to X with no other multicast-aware routers in the path. If any Z* 2719 (and/or Y) is located on the same OMNI link segment as X, the 2720 multicast data traffic sent to X directly using OAL/INET 2721 encapsulation instead of via a Bridge. 2723 Following the initial Join/Prune and NS/NA messaging, X maintains an 2724 asymmetric neighbor cache entry for each S the same as if X was 2725 sending unicast data traffic to S. In particular, X performs 2726 additional NS/NA exchanges to keep the neighbor cache entry alive for 2727 up to t_periodic seconds [RFC7761]. If no new Joins are received 2728 within t_periodic seconds, X allows the neighbor cache entry to 2729 expire. Finally, if X receives any additional Join/Prune messages 2730 for (S,G) it forwards the messages to each Y and Z* in the neighbor 2731 cache entry over the spanning tree. 2733 At some later time, Client C that holds an MNP for source S may 2734 depart from a first Proxy Z1 and/or connect via a new Proxy Z2. In 2735 that case, Y sends an unsolicited NA message to X the same as 2736 specified for unicast mobility in Section 3.16. When X receives the 2737 unsolicited NA message, it updates its asymmetric neighbor cache 2738 entry for the LLA for source S and sends new Join messages to any new 2739 Proxys Z2. There is no requirement to send any Prune messages to old 2740 Proxys Z1 since source S will no longer source any multicast data 2741 traffic via Z1. Instead, the multicast state for (S,G) in Proxy Z1 2742 will soon time out since no new Joins will arrive. 2744 After some later time, C may move to a new Server Y2 and depart from 2745 old Sever Y1. In that case, Y1 sends Join messages for any of C's 2746 active (S,G) groups to Y2 while including its own LLA as the source 2747 address. This causes Y2 to include Y1 in the multicast forwarding 2748 tree during the interim time that Y1's symmetric neighbor cache entry 2749 for C is in the DEPARTED state. At the same time, Y1 sends an 2750 unsolicited NA message to X with an OMNI option with S/T-ifIndex in 2751 the header set to 0 and a release indication to cause X to release 2752 its asymmetric neighbor cache entry. X then sends a new Join message 2753 to S via the spanning tree and re-initiates route optimization the 2754 same as if it were receiving a fresh Join message from a node on a 2755 downstream link. 2757 3.17.2. Any-Source Multicast (ASM) 2759 When an ROS X acting as a PIM router receives a Join/Prune from a 2760 node on its downstream interfaces containing one or more (*,G) pairs, 2761 it updates its Multicast Routing Information Base (MRIB) accordingly. 2762 X then forwards a copy of the message to the Rendezvous Point (RP) R 2763 for each G over the spanning tree. X uses its own LLA as the source 2764 address and ALL-PIM-ROUTERS as the destination address, then 2765 encapsulates each message in an OAL header with source address set to 2766 the ULA of X and destination address set to R, then sends the message 2767 into the spanning tree. At the same time, if the message was a Join 2768 X initiates NS/NA route optimization the same as for the SSM case 2769 discussed in Section 3.17.1. 2771 For each source S that sends multicast traffic to group G via R, the 2772 Proxy/Server Z* for the Client that aggregates S encapsulates the 2773 packets in PIM Register messages and forwards them to R via the 2774 spanning tree, which may then elect to send a PIM Join to Z*. This 2775 will result in an (S,G) tree rooted at Z* with R as the next hop so 2776 that R will begin to receive two copies of the packet; one native 2777 copy from the (S, G) tree and a second copy from the pre-existing (*, 2778 G) tree that still uses PIM Register encapsulation. R can then issue 2779 a PIM Register-stop message to suppress the Register-encapsulated 2780 stream. At some later time, if C moves to a new Proxy/Server Z*, it 2781 resumes sending packets via PIM Register encapsulation via the new 2782 Z*. 2784 At the same time, as multicast listeners discover individual S's for 2785 a given G, they can initiate an (S,G) Join for each S under the same 2786 procedures discussed in Section 3.17.1. Once the (S,G) tree is 2787 established, the listeners can send (S, G) Prune messages to R so 2788 that multicast packets for group G sourced by S will only be 2789 delivered via the (S, G) tree and not from the (*, G) tree rooted at 2790 R. All mobility considerations discussed for SSM apply. 2792 3.17.3. Bi-Directional PIM (BIDIR-PIM) 2794 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 2795 approach to ASM that treats the Rendezvous Point (RP) as a Designated 2796 Forwarder (DF). Further considerations for BIDIR-PIM are out of 2797 scope. 2799 3.18. Operation over Multiple OMNI Links 2801 An AERO Client can connect to multiple OMNI links the same as for any 2802 data link service. In that case, the Client maintains a distinct 2803 OMNI interface for each link, e.g., 'omni0' for the first link, 2804 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 2805 would include its own distinct set of Bridges, Servers and Proxys, 2806 thereby providing redundancy in case of failures. 2808 Each OMNI link could utilize the same or different ANET connections. 2809 The links can be distinguished at the link-layer via the SRT prefix 2810 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 2811 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 2812 MSPs on each link. This gives rise to the opportunity for supporting 2813 multiple redundant networked paths, with each VLAN distinguished by a 2814 different SRT "color" (see: Section 3.2.5). 2816 The Client's IP layer can select the outgoing OMNI interface 2817 appropriate for a given traffic profile while (in the reverse 2818 direction) correspondent nodes must have some way of steering their 2819 packets destined to a target via the correct OMNI link. 2821 In a first alternative, if each OMNI link services different MSPs, 2822 then the Client can receive a distinct MNP from each of the links. 2823 IP routing will therefore assure that the correct Red/Green/Blue/etc. 2824 network is used for both outbound and inbound traffic. This can be 2825 accomplished using existing technologies and approaches, and without 2826 requiring any special supporting code in correspondent nodes or 2827 Bridges. 2829 In a second alternative, if each OMNI link services the same MSP(s) 2830 then each link could assign a distinct "OMNI link Anycast" address 2831 that is configured by all Bridges on the link. Correspondent nodes 2832 can then perform Segment Routing to select the correct SRT, which 2833 will then direct the packet over multiple hops to the target. 2835 3.19. DNS Considerations 2837 AERO Client MNs and INET correspondent nodes consult the Domain Name 2838 System (DNS) the same as for any Internetworking node. When 2839 correspondent nodes and Client MNs use different IP protocol versions 2840 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 2841 A records for IPv4 address mappings to MNs which must then be 2842 populated in Relay NAT64 mapping caches. In that way, an IPv4 2843 correspondent node can send packets to the IPv4 address mapping of 2844 the target MN, and the Relay will translate the IPv4 header and 2845 destination address into an IPv6 header and IPv6 destination address 2846 of the MN. 2848 When an AERO Client registers with an AERO Server, the Server can 2849 return the address(es) of DNS servers in RDNSS options [RFC6106]. 2850 The DNS server provides the IP addresses of other MNs and 2851 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 2853 3.20. Transition Considerations 2855 OAL encapsulation ensures that dissimilar INET partitions can be 2856 joined into a single unified OMNI link, even though the partitions 2857 themselves may have differing protocol versions and/or incompatible 2858 addressing plans. However, a commonality can be achieved by 2859 incrementally distributing globally routable (i.e., native) IP 2860 prefixes to eventually reach all nodes (both mobile and fixed) in all 2861 OMNI link segments. This can be accomplished by incrementally 2862 deploying AERO Relays on each INET partition, with each Relay 2863 distributing its MNPs and/or discovering non-MNP IP GUA prefixes on 2864 its INET links. 2866 This gives rise to the opportunity to eventually distribute native IP 2867 addresses to all nodes, and to present a unified OMNI link view even 2868 if the INET partitions remain in their current protocol and 2869 addressing plans. In that way, the OMNI link can serve the dual 2870 purpose of providing a mobility/multilink service and a transition 2871 service. Or, if an INET partition is transitioned to a native IP 2872 protocol version and addressing scheme that is compatible with the 2873 OMNI link MNP-based addressing scheme, the partition and OMNI link 2874 can be joined by Relays. 2876 Relays that connect INETs/EUNs with dissimilar IP protocol versions 2877 may need to employ a network address and protocol translation 2878 function such as NAT64 [RFC6146]. 2880 3.21. Detecting and Reacting to Server and Bridge Failures 2882 In environments where rapid failure recovery is required, Servers and 2883 Bridges SHOULD use Bidirectional Forwarding Detection (BFD) 2884 [RFC5880]. Nodes that use BFD can quickly detect and react to 2885 failures so that cached information is re-established through 2886 alternate nodes. BFD control messaging is carried only over well- 2887 connected ground domain networks (i.e., and not low-end radio links) 2888 and can therefore be tuned for rapid response. 2890 Servers and Bridges maintain BFD sessions in parallel with their BGP 2891 peerings. If a Server or Bridge fails, BGP peers will quickly re- 2892 establish routes through alternate paths the same as for common BGP 2893 deployments. Similarly, Proxys maintain BFD sessions with their 2894 associated Bridges even though they do not establish BGP peerings 2895 with them. 2897 Proxys SHOULD use proactive NUD for Servers for which there are 2898 currently active ANET Clients in a manner that parallels BFD, i.e., 2899 by sending unicast NS messages in rapid succession to receive 2900 solicited NA messages. When the Proxy is also sending RS messages on 2901 behalf of ANET Clients, the RS/RA messaging can be considered as 2902 equivalent hints of forward progress. This means that the Proxy need 2903 not also send a periodic NS if it has already sent an RS within the 2904 same period. If a Server fails, the Proxy will cease to receive 2905 advertisements and can quickly inform Clients of the outage by 2906 sending multicast RA messages on the ANET interface. 2908 The Proxy sends multicast RA messages with source address set to the 2909 Server's address, destination address set to (link-local) All-Nodes 2910 multicast, and Router Lifetime set to 0. The Proxy SHOULD send 2911 MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small delays 2912 [RFC4861]. Any Clients on the ANET interface that have been using 2913 the (now defunct) Server will receive the RA messages and associate 2914 with a new Server. 2916 3.22. AERO Clients on the Open Internet 2918 AERO Clients that connect to the open Internet via INET interfaces 2919 can establish a VPN or direct link to securely connect to a Server in 2920 a "tethered" arrangement with all of the Client's traffic transiting 2921 the Server. Alternatively, the Client can associate with an INET 2922 Server using UDP/IP encapsulation and asymmetric securing services as 2923 discussed in the following sections. 2925 When a Client's OMNI interface enables an INET underlying interface, 2926 it first determines whether the interface is likely to be behind a 2927 NAT. For IPv4, the Client assumes it is on the open Internet if the 2928 INET address is not a special-use IPv4 address per [RFC3330]. 2929 Similarly for IPv6, the Client assumes it is on the open Internet if 2930 the INET address is not a link-local [RFC4291] or unique-local 2931 [RFC4193] IPv6 address. 2933 The Client then prepares a UDP/IP-encapsulated RS message with IPv6 2934 source address set to its LLA, with IPv6 destination set to (link- 2935 local) All-Routers multicast and with an OMNI option with underlying 2936 interface attributes. If the Client believes that it is on the open 2937 Internet, it SHOULD include Interface Attributes with the L2ADDR used 2938 for INET encapsulation (otherwise, it MAY omit L2ADDR). If the 2939 underlying address is IPv4, the Client includes the Port Number and 2940 IPv4 address written in obfuscated form [RFC4380] as discussed in 2941 Section 3.3. If the underlying interface address is IPv6, the Client 2942 instead includes the Port Number and IPv6 address in obfuscated form. 2943 The Client finally includes an Authentication option per [RFC4380] to 2944 provide message authentication, sets the UDP/IP source to its INET 2945 address and UDP port, sets the UDP/IP destination to the Server's 2946 INET address and the AERO service port number (8060), then sends the 2947 message to the Server. 2949 When the Server receives the RS, it authenticates the message and 2950 registers the Client's MNP and INET interface information according 2951 to the OMNI option parameters. If the RS message includes an L2ADDR 2952 in the OMNI option, the Server compares the encapsulation IP address 2953 and UDP port number with the (unobfuscated) values. If the values 2954 are the same, the Server caches the Client's information as "INET" 2955 addresses meaning that the Client is likely to accept direct messages 2956 without requiring NAT traversal exchanges. If the values are 2957 different (or, if the OMNI option did not include an L2ADDR) the 2958 Server instead caches the Client's information as "NAT" addresses 2959 meaning that NAT traversal exchanges may be necessary. 2961 The Server then returns an RA message with IPv6 source and 2962 destination set corresponding to the addresses in the RS, and with an 2963 Authentication option per [RFC4380]. For IPv4, the Server also 2964 includes an Origin option per [RFC4380] with the mapped and 2965 obfuscated Port Number and IPv4 address observed in the encapsulation 2966 headers. For IPv6, the Server instead includes an IPv6 Origin option 2967 per Figure 6 with the mapped and obfuscated observed Port Number and 2968 IPv6 address (note that the value 0x02 in the second octet 2969 differentiates from other [RFC4380] option types). 2971 +--------+--------+-----------------+ 2972 | 0x00 | 0x02 | Origin port # | 2973 +--------+--------+-----------------+ 2974 ~ Origin IPv6 address ~ 2975 +-----------------------------------+ 2977 Figure 6: IPv6 Origin Option 2979 When the Client receives the RA message, it compares the mapped Port 2980 Number and IP address from the Origin option with its own address. 2981 If the addresses are the same, the Client assumes the open Internet / 2982 Cone NAT principle; if the addresses are different, the Client 2983 instead assumes that further qualification procedures are necessary 2984 to detect the type of NAT and proceeds according to standard 2985 [RFC4380] procedures. 2987 After the Client has registered its INET interfaces in such RS/RA 2988 exchanges it sends periodic RS messages to receive fresh RA messages 2989 before the Router Lifetime received on each INET interface expires. 2990 The Client also maintains default routes via its Servers, i.e., the 2991 same as described in earlier sections. 2993 When the Client sends messages to target IP addresses, it also 2994 invokes route optimization per Section 3.14 using IPv6 ND address 2995 resolution messaging. The Client sends the NS(AR) message to the 2996 Server wrapped in a UDP/IP header with an Authentication option with 2997 the NS source address set to the Client's LLA and destination address 2998 set to the target solicited node multicast address. The Server 2999 authenticates the message and sends a corresponding NS(AR) message 3000 over the spanning tree the same as if it were the ROS, but with the 3001 OAL source address set to the Server's ULA and destination set to the 3002 ULA of the target. When the ROR receives the NS(AR), it adds the 3003 Server's ULA and Client's LLA to the target's Report List, and 3004 returns an NA with OMNI option information for the target. The 3005 Server then returns a UDP/IP encapsulated NA message with an 3006 Authentication option to the Client. 3008 Following route optimization for targets in the same OMNI link 3009 segment, if the target's L2ADDR is on the open INET, the Client 3010 forwards data packets directly to the target INET address. If the 3011 target is behind a NAT, the Client first establishes NAT state for 3012 the L2ADDR using the "bubble" mechanisms specified in 3013 [RFC6081][RFC4380]. The Client continues to send data packets via 3014 its Server until NAT state is populated, then begins forwarding 3015 packets via the direct path through the NAT to the target. For 3016 targets in different OMNI link segments, the Client uses OAL/ORH 3017 encapsulation and forwards data packets to the Bridge that returned 3018 the NA message. 3020 The ROR may return uNAs via the Server if the target moves, and the 3021 Server will send corresponding Authentication-protected uNAs to the 3022 Client. The Client can also send "loopback" NS(NUD) messages to test 3023 forward path reachability even though there is no security 3024 association between the Client and the target. 3026 The Client sends UDP/IP encapsulated IPv6 packets no larger than 1280 3027 bytes in one piece. In order to accommodate larger IPv6 packets (up 3028 to the OMNI interface MTU), the Client inserts an OAL header with 3029 source set to its own ULA and destination set to the ULA of the 3030 target and uses IPv6 fragmentation according to Section 3.9. The 3031 Client then encapsulates each fragment in a UDP/IP header and sends 3032 the fragments to the next hop. 3034 3.23. Time-Varying MNPs 3036 In some use cases, it is desirable, beneficial and efficient for the 3037 Client to receive a constant MNP that travels with the Client 3038 wherever it moves. For example, this would allow air traffic 3039 controllers to easily track aircraft, etc. In other cases, however 3040 (e.g., intelligent transportation systems), the MN may be willing to 3041 sacrifice a modicum of efficiency in order to have time-varying MNPs 3042 that can be changed every so often to defeat adversarial tracking. 3044 The DHCPv6 service offers a way for Clients that desire time-varying 3045 MNPs to obtain short-lived prefixes (e.g., on the order of a small 3046 number of minutes). In that case, the identity of the Client would 3047 not be bound to the MNP but rather the Client's identity would be 3048 bound to the DHCPv6 Device Unique Identifier (DUID) and used as the 3049 seed for Prefix Delegation. The Client would then be obligated to 3050 renumber its internal networks whenever its MNP (and therefore also 3051 its LLA) changes. This should not present a challenge for Clients 3052 with automated network renumbering services, however presents limits 3053 for the durations of ongoing sessions that would prefer to use a 3054 constant address. 3056 4. Implementation Status 3058 An early AERO implementation based on OpenVPN (https://openvpn.net/) 3059 was announced on the v6ops mailing list on January 10, 2018 and an 3060 initial public release of the AERO proof-of-concept source code was 3061 announced on the intarea mailing list on August 21, 2015. 3063 AERO Release-3.0.2 was tagged on October 15, 2020, and is undergoing 3064 internal testing. Additional releases expected Q42020, with first 3065 public release expected before year-end. 3067 5. IANA Considerations 3069 The IANA has assigned a 4-octet Private Enterprise Number "45282" for 3070 AERO in the "enterprise-numbers" registry. 3072 The IANA has assigned the UDP port number "8060" for an earlier 3073 experimental version of AERO [RFC6706]. This document obsoletes 3074 [RFC6706] and claims the UDP port number "8060" for all future use. 3076 The IANA is instructed to assign a new type value TBD in the IPv6 3077 Routing Types registry. 3079 No further IANA actions are required. 3081 6. Security Considerations 3083 AERO Bridges configure secured tunnels with AERO Servers, Relays and 3084 Proxys within their local OMNI link segments. Applicable secured 3085 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 3086 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 3087 segments in turn configure secured tunnels for their neighboring AERO 3088 Bridges in a spanning tree topology. Therefore, control messages 3089 exchanged between any pair of OMNI link neighbors on the spanning 3090 tree are already secured. 3092 AERO Servers, Relays and Proxys targeted by a route optimization may 3093 also receive data packets directly from arbitrary nodes in INET 3094 partitions instead of via the spanning tree. For INET partitions 3095 that apply effective ingress filtering to defeat source address 3096 spoofing, the simple data origin authentication procedures in 3097 Section 3.8 can be applied. 3099 For INET partitions that require strong security in the data plane, 3100 two options for securing communications include 1) disable route 3101 optimization so that all traffic is conveyed over secured tunnels, or 3102 2) enable on-demand secure tunnel creation between INET partition 3103 neighbors. Option 1) would result in longer routes than necessary 3104 and traffic concentration on critical infrastructure elements. 3105 Option 2) could be coordinated by establishing a secured tunnel on- 3106 demand instead of performing an NS/NA exchange in the route 3107 optimization procedures. Procedures for establishing on-demand 3108 secured tunnels are out of scope. 3110 AERO Clients that connect to secured ANETs need not apply security to 3111 their ND messages, since the messages will be intercepted by a 3112 perimeter Proxy that applies security on its INET-facing interface as 3113 part of the spanning tree (see above). AERO Clients connected to the 3114 open INET can use symmetric network and/or transport layer security 3115 services such as VPNs or can by some other means establish a direct 3116 link. When a VPN or direct link may be impractical, however, an 3117 asymmetric security service such as the Authentication option 3118 specified in [RFC4380] should be applied. The Authentication option 3119 requires a unique Client identifier, which can be obtained per the 3120 Universally Unique IDentifier (UUID) [RFC4122] service and also used 3121 as a DHCP Unique Identifier (DUID) per [RFC6355]. 3123 Application endpoints SHOULD use application-layer security services 3124 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 3125 protection as for critical secured Internet services. AERO Clients 3126 that require host-based VPN services SHOULD use symmetric network 3127 and/or transport layer security services such as IPsec, TLS/SSL, 3128 DTLS, etc. AERO Proxys and Servers can also provide a network-based 3129 VPN service on behalf of the Client, e.g., if the Client is located 3130 within a secured enclave and cannot establish a VPN on its own 3131 behalf. 3133 AERO Servers and Bridges present targets for traffic amplification 3134 Denial of Service (DoS) attacks. This concern is no different than 3135 for widely-deployed VPN security gateways in the Internet, where 3136 attackers could send spoofed packets to the gateways at high data 3137 rates. This can be mitigated by connecting Servers and Bridges over 3138 dedicated links with no connections to the Internet and/or when 3139 connections to the Internet are only permitted through well-managed 3140 firewalls. Traffic amplification DoS attacks can also target an AERO 3141 Client's low data rate links. This is a concern not only for Clients 3142 located on the open Internet but also for Clients in secured 3143 enclaves. AERO Servers and Proxys can institute rate limits that 3144 protect Clients from receiving packet floods that could DoS low data 3145 rate links. 3147 AERO Relays must implement ingress filtering to avoid a spoofing 3148 attack in which spurious messages with ULA addresses are injected 3149 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3150 that their connectivity is not used by unauthorized nodes on their 3151 EUNs to gain access to a protected network, i.e., AERO Clients that 3152 act as routers MUST NOT provide routing services for unauthorized 3153 nodes. (This concern is no different than for ordinary hosts that 3154 receive an IP address delegation but then "share" the address with 3155 other nodes via some form of Internet connection sharing such as 3156 tethering.) 3158 The MAP list MUST be well-managed and secured from unauthorized 3159 tampering, even though the list contains only public information. 3160 The MAP list can be conveyed to the Client in a similar fashion as in 3161 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3162 upload of a static file, DNS lookups, etc.). 3164 SRH authentication facilities are specified in [RFC8754]. 3166 Security considerations for accepting link-layer ICMP messages and 3167 reflected packets are discussed throughout the document. 3169 Security considerations for IPv6 fragmentation and reassembly are 3170 discussed in [I-D.templin-6man-omni-interface]. 3172 7. Acknowledgements 3174 Discussions in the IETF, aviation standards communities and private 3175 exchanges helped shape some of the concepts in this work. 3176 Individuals who contributed insights include Mikael Abrahamsson, Mark 3177 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3178 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3179 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3180 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3181 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3182 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3183 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3184 Wood and James Woodyatt. Members of the IESG also provided valuable 3185 input during their review process that greatly improved the document. 3186 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3187 for their shepherding guidance during the publication of the AERO 3188 first edition. 3190 This work has further been encouraged and supported by Boeing 3191 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3192 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3193 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3194 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3195 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3196 Gene MacLean III, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3197 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3198 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3199 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3200 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3201 Benson, Katie Tran and Eric Yeh are especially acknowledged for 3202 implementing the AERO functions as extensions to the public domain 3203 OpenVPN distribution. 3205 Earlier works on NBMA tunneling approaches are found in 3206 [RFC2529][RFC5214][RFC5569]. 3208 Many of the constructs presented in this second edition of AERO are 3209 based on the author's earlier works, including: 3211 o The Internet Routing Overlay Network (IRON) 3212 [RFC6179][I-D.templin-ironbis] 3214 o Virtual Enterprise Traversal (VET) 3215 [RFC5558][I-D.templin-intarea-vet] 3217 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3218 [RFC5320][I-D.templin-intarea-seal] 3220 o AERO, First Edition [RFC6706] 3222 Note that these works cite numerous earlier efforts that are not also 3223 cited here due to space limitations. The authors of those earlier 3224 works are acknowledged for their insights. 3226 This work is aligned with the NASA Safe Autonomous Systems Operation 3227 (SASO) program under NASA contract number NNA16BD84C. 3229 This work is aligned with the FAA as per the SE2025 contract number 3230 DTFAWA-15-D-00030. 3232 This work is aligned with the Boeing Commercial Airplanes (BCA) 3233 Internet of Things (IoT) and autonomy programs. 3235 This work is aligned with the Boeing Information Technology (BIT) 3236 MobileNet program. 3238 8. References 3240 8.1. Normative References 3242 [I-D.templin-6man-omni-interface] 3243 Templin, F. and T. Whyman, "Transmission of IP Packets 3244 over Overlay Multilink Network (OMNI) Interfaces", draft- 3245 templin-6man-omni-interface-59 (work in progress), 3246 December 2020. 3248 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3249 DOI 10.17487/RFC0791, September 1981, 3250 . 3252 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3253 RFC 792, DOI 10.17487/RFC0792, September 1981, 3254 . 3256 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3257 Requirement Levels", BCP 14, RFC 2119, 3258 DOI 10.17487/RFC2119, March 1997, 3259 . 3261 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3262 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3263 December 1998, . 3265 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3266 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3267 DOI 10.17487/RFC3971, March 2005, 3268 . 3270 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3271 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3272 . 3274 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3275 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3276 November 2005, . 3278 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3279 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3280 . 3282 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3283 Network Address Translations (NATs)", RFC 4380, 3284 DOI 10.17487/RFC4380, February 2006, 3285 . 3287 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3288 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3289 DOI 10.17487/RFC4861, September 2007, 3290 . 3292 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3293 Address Autoconfiguration", RFC 4862, 3294 DOI 10.17487/RFC4862, September 2007, 3295 . 3297 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3298 DOI 10.17487/RFC6081, January 2011, 3299 . 3301 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3302 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3303 May 2017, . 3305 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3306 (IPv6) Specification", STD 86, RFC 8200, 3307 DOI 10.17487/RFC8200, July 2017, 3308 . 3310 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3311 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3312 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3313 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3314 . 3316 8.2. Informative References 3318 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3319 2016. 3321 [I-D.bonica-6man-comp-rtg-hdr] 3322 Bonica, R., Kamite, Y., Alston, A., Henriques, D., and L. 3323 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3324 bonica-6man-comp-rtg-hdr-23 (work in progress), October 3325 2020. 3327 [I-D.bonica-6man-crh-helper-opt] 3328 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3329 Routing Header (CRH) Helper Option", draft-bonica-6man- 3330 crh-helper-opt-02 (work in progress), October 2020. 3332 [I-D.ietf-intarea-frag-fragile] 3333 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 3334 and F. Gont, "IP Fragmentation Considered Fragile", draft- 3335 ietf-intarea-frag-fragile-17 (work in progress), September 3336 2019. 3338 [I-D.ietf-intarea-tunnels] 3339 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3340 Architecture", draft-ietf-intarea-tunnels-10 (work in 3341 progress), September 2019. 3343 [I-D.ietf-rtgwg-atn-bgp] 3344 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3345 Moreno, "A Simple BGP-based Mobile Routing System for the 3346 Aeronautical Telecommunications Network", draft-ietf- 3347 rtgwg-atn-bgp-06 (work in progress), June 2020. 3349 [I-D.templin-6man-dhcpv6-ndopt] 3350 Templin, F., "A Unified Stateful/Stateless Configuration 3351 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-10 3352 (work in progress), June 2020. 3354 [I-D.templin-intarea-seal] 3355 Templin, F., "The Subnetwork Encapsulation and Adaptation 3356 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3357 progress), January 2014. 3359 [I-D.templin-intarea-vet] 3360 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3361 templin-intarea-vet-40 (work in progress), May 2013. 3363 [I-D.templin-ironbis] 3364 Templin, F., "The Interior Routing Overlay Network 3365 (IRON)", draft-templin-ironbis-16 (work in progress), 3366 March 2014. 3368 [I-D.templin-v6ops-pdhost] 3369 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3370 Models", draft-templin-v6ops-pdhost-26 (work in progress), 3371 June 2020. 3373 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3375 [RFC1035] Mockapetris, P., "Domain names - implementation and 3376 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3377 November 1987, . 3379 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3380 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3381 . 3383 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3384 DOI 10.17487/RFC2003, October 1996, 3385 . 3387 [RFC2004] Perkins, C., "Minimal Encapsulation within IP", RFC 2004, 3388 DOI 10.17487/RFC2004, October 1996, 3389 . 3391 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3392 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3393 . 3395 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 3396 Networks", RFC 2464, DOI 10.17487/RFC2464, December 1998, 3397 . 3399 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3400 Domains without Explicit Tunnels", RFC 2529, 3401 DOI 10.17487/RFC2529, March 1999, 3402 . 3404 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3405 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3406 . 3408 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3409 of Explicit Congestion Notification (ECN) to IP", 3410 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3411 . 3413 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3414 DOI 10.17487/RFC3330, September 2002, 3415 . 3417 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3418 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3419 DOI 10.17487/RFC3810, June 2004, 3420 . 3422 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 3423 Unique IDentifier (UUID) URN Namespace", RFC 4122, 3424 DOI 10.17487/RFC4122, July 2005, 3425 . 3427 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3428 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3429 January 2006, . 3431 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3432 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3433 DOI 10.17487/RFC4271, January 2006, 3434 . 3436 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3437 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3438 2006, . 3440 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3441 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3442 December 2005, . 3444 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3445 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3446 2006, . 3448 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3449 Control Message Protocol (ICMPv6) for the Internet 3450 Protocol Version 6 (IPv6) Specification", STD 89, 3451 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3452 . 3454 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3455 Protocol (LDAP): The Protocol", RFC 4511, 3456 DOI 10.17487/RFC4511, June 2006, 3457 . 3459 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3460 "Considerations for Internet Group Management Protocol 3461 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3462 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3463 . 3465 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3466 "Internet Group Management Protocol (IGMP) / Multicast 3467 Listener Discovery (MLD)-Based Multicast Forwarding 3468 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3469 August 2006, . 3471 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3472 Algorithms in Cryptographically Generated Addresses 3473 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3474 . 3476 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3477 "Bidirectional Protocol Independent Multicast (BIDIR- 3478 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3479 . 3481 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3482 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3483 DOI 10.17487/RFC5214, March 2008, 3484 . 3486 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3487 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3488 February 2010, . 3490 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3491 Route Optimization Requirements for Operational Use in 3492 Aeronautics and Space Exploration Mobile Networks", 3493 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3494 . 3496 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3497 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3498 . 3500 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3501 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3502 January 2010, . 3504 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3505 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3506 . 3508 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3509 "IPv6 Router Advertisement Options for DNS Configuration", 3510 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3511 . 3513 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3514 NAT64: Network Address and Protocol Translation from IPv6 3515 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3516 April 2011, . 3518 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3519 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3520 . 3522 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3523 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3524 DOI 10.17487/RFC6221, May 2011, 3525 . 3527 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3528 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3529 DOI 10.17487/RFC6273, June 2011, 3530 . 3532 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3533 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3534 January 2012, . 3536 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 3537 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, 3538 DOI 10.17487/RFC6355, August 2011, 3539 . 3541 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3542 for Equal Cost Multipath Routing and Link Aggregation in 3543 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3544 . 3546 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3547 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3548 . 3550 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3551 UDP Checksums for Tunneled Packets", RFC 6935, 3552 DOI 10.17487/RFC6935, April 2013, 3553 . 3555 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3556 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3557 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3558 . 3560 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3561 Korhonen, "Requirements for Distributed Mobility 3562 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3563 . 3565 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3566 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3567 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3568 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3569 2016, . 3571 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3572 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3573 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3574 July 2018, . 3576 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3577 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3578 . 3580 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 3581 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3582 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 3583 . 3585 [WG] Wireguard, "Wireguard, https://www.wireguard.com", August 3586 2020. 3588 Appendix A. Non-Normative Considerations 3590 AERO can be applied to a multitude of Internetworking scenarios, with 3591 each having its own adaptations. The following considerations are 3592 provided as non-normative guidance: 3594 A.1. Implementation Strategies for Route Optimization 3596 Route optimization as discussed in Section 3.14 results in the route 3597 optimization source (ROS) creating an asymmetric neighbor cache entry 3598 for the target neighbor. The neighbor cache entry is maintained for 3599 at most ReachableTime seconds and then deleted unless updated. In 3600 order to refresh the neighbor cache entry lifetime before the 3601 ReachableTime timer expires, the specification requires 3602 implementations to issue a new NS/NA exchange to reset ReachableTime 3603 while data packets are still flowing. However, the decision of when 3604 to initiate a new NS/NA exchange and to perpetuate the process is 3605 left as an implementation detail. 3607 One possible strategy may be to monitor the neighbor cache entry 3608 watching for data packets for (ReachableTime - 5) seconds. If any 3609 data packets have been sent to the neighbor within this timeframe, 3610 then send an NS to receive a new NA. If no data packets have been 3611 sent, wait for 5 additional seconds and send an immediate NS if any 3612 data packets are sent within this "expiration pending" 5 second 3613 window. If no additional data packets are sent within the 5 second 3614 window, delete the neighbor cache entry. 3616 The monitoring of the neighbor data packet traffic therefore becomes 3617 an asymmetric ongoing process during the neighbor cache entry 3618 lifetime. If the neighbor cache entry expires, future data packets 3619 will trigger a new NS/NA exchange while the packets themselves are 3620 delivered over a longer path until route optimization state is re- 3621 established. 3623 A.2. Implicit Mobility Management 3625 OMNI interface neighbors MAY provide a configuration option that 3626 allows them to perform implicit mobility management in which no ND 3627 messaging is used. In that case, the Client only transmits packets 3628 over a single interface at a time, and the neighbor always observes 3629 packets arriving from the Client from the same link-layer source 3630 address. 3632 If the Client's underlying interface address changes (either due to a 3633 readdressing of the original interface or switching to a new 3634 interface) the neighbor immediately updates the neighbor cache entry 3635 for the Client and begins accepting and sending packets according to 3636 the Client's new address. This implicit mobility method applies to 3637 use cases such as cellphones with both WiFi and Cellular interfaces 3638 where only one of the interfaces is active at a given time, and the 3639 Client automatically switches over to the backup interface if the 3640 primary interface fails. 3642 A.3. Direct Underlying Interfaces 3644 When a Client's OMNI interface is configured over a Direct interface, 3645 the neighbor at the other end of the Direct link can receive packets 3646 without any encapsulation. In that case, the Client sends packets 3647 over the Direct link according to QoS preferences. If the Direct 3648 interface has the highest QoS preference, then the Client's IP 3649 packets are transmitted directly to the peer without going through an 3650 ANET/INET. If other interfaces have higher QoS preferences, then the 3651 Client's IP packets are transmitted via a different interface, which 3652 may result in the inclusion of Proxys, Servers and Bridges in the 3653 communications path. Direct interfaces must be tested periodically 3654 for reachability, e.g., via NUD. 3656 A.4. AERO Critical Infrastructure Considerations 3658 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 3659 IP routers or virtual machines in the cloud. Bridges must be 3660 provisioned, supported and managed by the INET administrative 3661 authority, and connected to the Bridges of other INETs via inter- 3662 domain peerings. Cost for purchasing, configuring and managing 3663 Bridges is nominal even for very large OMNI links. 3665 AERO Servers can be standard dedicated server platforms, but most 3666 often will be deployed as virtual machines in the cloud. The only 3667 requirements for Servers are that they can run the AERO user-level 3668 code and have at least one network interface connection to the INET. 3669 As with Bridges, Servers must be provisioned, supported and managed 3670 by the INET administrative authority. Cost for purchasing, 3671 configuring and managing Servers is nominal especially for virtual 3672 Servers hosted in the cloud. 3674 AERO Proxys are most often standard dedicated server platforms with 3675 one network interface connected to the ANET and a second interface 3676 connected to an INET. As with Servers, the only requirements are 3677 that they can run the AERO user-level code and have at least one 3678 interface connection to the INET. Proxys must be provisioned, 3679 supported and managed by the ANET administrative authority. Cost for 3680 purchasing, configuring and managing Proxys is nominal, and borne by 3681 the ANET administrative authority. 3683 AERO Relays can be any dedicated server or COTS router platform 3684 connected to INETs and/or EUNs. The Relay connects to the OMNI link 3685 and engages in eBGP peering with one or more Bridges as a stub AS. 3686 The Relay then injects its MNPs and/or non-MNP prefixes into the BGP 3687 routing system, and provisions the prefixes to its downstream- 3688 attached networks. The Relay can perform ROS/ROR services the same 3689 as for any Server, and can route between the MNP and non-MNP address 3690 spaces. 3692 A.5. AERO Server Failure Implications 3694 AERO Servers may appear as a single point of failure in the 3695 architecture, but such is not the case since all Servers on the link 3696 provide identical services and loss of a Server does not imply 3697 immediate and/or comprehensive communication failures. Although 3698 Clients typically associate with a single Server at a time, Server 3699 failure is quickly detected and conveyed by Bidirectional Forward 3700 Detection (BFD) and/or proactive NUD allowing Clients to migrate to 3701 new Servers. 3703 If a Server fails, ongoing packet forwarding to Clients will continue 3704 by virtue of the asymmetric neighbor cache entries that have already 3705 been established in route optimization sources (ROSs). If a Client 3706 also experiences mobility events at roughly the same time the Server 3707 fails, unsolicited NA messages may be lost but proxy neighbor cache 3708 entries in the DEPARTED state will ensure that packet forwarding to 3709 the Client's new locations will continue for up to DepartTime 3710 seconds. 3712 If a Client is left without a Server for an extended timeframe (e.g., 3713 greater than ReachableTime seconds) then existing asymmetric neighbor 3714 cache entries will eventually expire and both ongoing and new 3715 communications will fail. The original source will continue to 3716 retransmit until the Client has established a new Server 3717 relationship, after which time continuous communications will resume. 3719 Therefore, providing many Servers on the link with high availability 3720 profiles provides resilience against loss of individual Servers and 3721 assurance that Clients can establish new Server relationships quickly 3722 in event of a Server failure. 3724 A.6. AERO Client / Server Architecture 3726 The AERO architectural model is client / server in the control plane, 3727 with route optimization in the data plane. The same as for common 3728 Internet services, the AERO Client discovers the addresses of AERO 3729 Servers and selects one Server to connect to. The AERO service is 3730 analogous to common Internet services such as google.com, yahoo.com, 3731 cnn.com, etc. However, there is only one AERO service for the link 3732 and all Servers provide identical services. 3734 Common Internet services provide differing strategies for advertising 3735 server addresses to clients. The strategy is conveyed through the 3736 DNS resource records returned in response to name resolution queries. 3737 As of January 2020 Internet-based 'nslookup' services were used to 3738 determine the following: 3740 o When a client resolves the domainname "google.com", the DNS always 3741 returns one A record (i.e., an IPv4 address) and one AAAA record 3742 (i.e., an IPv6 address). The client receives the same addresses 3743 each time it resolves the domainname via the same DNS resolver, 3744 but may receive different addresses when it resolves the 3745 domainname via different DNS resolvers. But, in each case, 3746 exactly one A and one AAAA record are returned. 3748 o When a client resolves the domainname "ietf.org", the DNS always 3749 returns one A record and one AAAA record with the same addresses 3750 regardless of which DNS resolver is used. 3752 o When a client resolves the domainname "yahoo.com", the DNS always 3753 returns a list of 4 A records and 4 AAAA records. Each time the 3754 client resolves the domainname via the same DNS resolver, the same 3755 list of addresses are returned but in randomized order (i.e., 3756 consistent with a DNS round-robin strategy). But, interestingly, 3757 the same addresses are returned (albeit in randomized order) when 3758 the domainname is resolved via different DNS resolvers. 3760 o When a client resolves the domainname "amazon.com", the DNS always 3761 returns a list of 3 A records and no AAAA records. As with 3762 "yahoo.com", the same three A records are returned from any 3763 worldwide Internet connection point in randomized order. 3765 The above example strategies show differing approaches to Internet 3766 resilience and service distribution offered by major Internet 3767 services. The Google approach exposes only a single IPv4 and a 3768 single IPv6 address to clients. Clients can then select whichever IP 3769 protocol version offers the best response, but will always use the 3770 same IP address according to the current Internet connection point. 3771 This means that the IP address offered by the network must lead to a 3772 highly-available server and/or service distribution point. In other 3773 words, resilience is predicated on high availability within the 3774 network and with no client-initiated failovers expected (i.e., it is 3775 all-or-nothing from the client's perspective). However, Google does 3776 provide for worldwide distributed service distribution by virtue of 3777 the fact that each Internet connection point responds with a 3778 different IPv6 and IPv4 address. The IETF approach is like google 3779 (all-or-nothing from the client's perspective), but provides only a 3780 single IPv4 or IPv6 address on a worldwide basis. This means that 3781 the addresses must be made highly-available at the network level with 3782 no client failover possibility, and if there is any worldwide service 3783 distribution it would need to be conducted by a network element that 3784 is reached via the IP address acting as a service distribution point. 3786 In contrast to the Google and IETF philosophies, Yahoo and Amazon 3787 both provide clients with a (short) list of IP addresses with Yahoo 3788 providing both IP protocol versions and Amazon as IPv4-only. The 3789 order of the list is randomized with each name service query 3790 response, with the effect of round-robin load balancing for service 3791 distribution. With a short list of addresses, there is still 3792 expectation that the network will implement high availability for 3793 each address but in case any single address fails the client can 3794 switch over to using a different address. The balance then becomes 3795 one of function in the network vs function in the end system. 3797 The same implications observed for common highly-available services 3798 in the Internet apply also to the AERO client/server architecture. 3799 When an AERO Client connects to one or more ANETs, it discovers one 3800 or more AERO Server addresses through the mechanisms discussed in 3801 earlier sections. Each Server address presumably leads to a fault- 3802 tolerant clustering arrangement such as supported by Linux-HA, 3803 Extended Virtual Synchrony or Paxos. Such an arrangement has 3804 precedence in common Internet service deployments in lightweight 3805 virtual machines without requiring expensive hardware deployment. 3806 Similarly, common Internet service deployments set service IP 3807 addresses on service distribution points that may relay requests to 3808 many different servers. 3810 For AERO, the expectation is that a combination of the Google/IETF 3811 and Yahoo/Amazon philosophies would be employed. The AERO Client 3812 connects to different ANET access points and can receive 1-2 Server 3813 LLAs at each point. It then selects one AERO Server address, and 3814 engages in RS/RA exchanges with the same Server from all ANET 3815 connections. The Client remains with this Server unless or until the 3816 Server fails, in which case it can switch over to an alternate 3817 Server. The Client can likewise switch over to a different Server at 3818 any time if there is some reason for it to do so. So, the AERO 3819 expectation is for a balance of function in the network and end 3820 system, with fault tolerance and resilience at both levels. 3822 Appendix B. Change Log 3824 << RFC Editor - remove prior to publication >> 3826 Changes from draft-templin-intarea-6706bis-61 to draft-templin- 3827 intrea-6706bis-62: 3829 o New sub-section on OMNI Neighbor Interface Attributes 3831 Changes from draft-templin-intarea-6706bis-59 to draft-templin- 3832 intrea-6706bis-60: 3834 o Removed all references to S/TLLAO - all Interface Attributes are 3835 now maintained completely in the OMNI option. 3837 Changes from draft-templin-intarea-6706bis-58 to draft-templin- 3838 intrea-6706bis-59: 3840 o The term "Relay" used in older draft versions is now "Bridge". 3841 "Relay" now refers to what was formally called: "Gateway". 3843 o Fine-grained cleanup of Forwarding Algorithm; IPv6 ND message 3844 addressing; OMNI Prefix Lengths, etc. 3846 Changes from draft-templin-intarea-6706bis-54 to draft-templin- 3847 intrea-6706bis-55: 3849 o Updates on Segment Routing and S/TLLAO contents. 3851 o Various editorials and addressing cleanups. 3853 Changes from draft-templin-intarea-6706bis-52 to draft-templin- 3854 intrea-6706bis-53: 3856 o Normative reference to the OMNI spec, and remove portions that are 3857 already specified in OMNI. 3859 o Renamed "AERO interface/link" to "OMIN interface/link" throughout 3860 the document. 3862 o Truncated obsolete back section matter. 3864 Author's Address 3866 Fred L. Templin (editor) 3867 Boeing Research & Technology 3868 P.O. Box 3707 3869 Seattle, WA 98124 3870 USA 3872 Email: fltemplin@acm.org