idnits 2.17.1 draft-templin-intarea-6706bis-88.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 (February 1, 2021) is 1170 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 1005, but not defined == Missing Reference: 'ULA0' is mentioned on line 1005, but not defined == Missing Reference: 'ULA1' is mentioned on line 1005, but not defined == Missing Reference: 'ULA2' is mentioned on line 1005, but not defined == Unused Reference: 'RFC3971' is defined on line 3410, but no explicit reference was found in the text == Unused Reference: 'RFC3972' is defined on line 3415, but no explicit reference was found in the text == Unused Reference: 'BGP' is defined on line 3468, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-comp-rtg-hdr' is defined on line 3471, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-crh-helper-opt' is defined on line 3477, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3482, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3505, but no explicit reference was found in the text == Unused Reference: 'RFC4122' is defined on line 3583, but no explicit reference was found in the text == Unused Reference: 'RFC4982' is defined on line 3632, but no explicit reference was found in the text == Unused Reference: 'RFC6273' is defined on line 3688, but no explicit reference was found in the text == Unused Reference: 'RFC6355' is defined on line 3697, but no explicit reference was found in the text == Outdated reference: A later version (-99) exists of draft-templin-6man-omni-interface-69 == Outdated reference: A later version (-31) exists of draft-bonica-6man-comp-rtg-hdr-24 == 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 (-30) exists of draft-ietf-ipwave-vehicular-networking-19 == Outdated reference: A later version (-26) exists of draft-ietf-rtgwg-atn-bgp-10 -- 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, February 1, 2021 5 rfc6179, rfc6706 (if 6 approved) 7 Intended status: Standards Track 8 Expires: August 5, 2021 10 Asymmetric Extended Route Optimization (AERO) 11 draft-templin-intarea-6706bis-88 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 August 5, 2021. 47 Copyright Notice 49 Copyright (c) 2021 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) . . . . . . . . 11 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) . . . . . . 15 72 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 16 73 3.2.4. OMNI Link Encapsulation . . . . . . . . . . . . . . . 18 74 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 22 75 3.2.6. Segment Routing For OMNI Link Selection . . . . . . . 23 76 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 23 77 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 24 78 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 26 79 3.4.1. AERO Server/Relay Behavior . . . . . . . . . . . . . 26 80 3.4.2. AERO Proxy Behavior . . . . . . . . . . . . . . . . . 26 81 3.4.3. AERO Client Behavior . . . . . . . . . . . . . . . . 27 82 3.4.4. AERO Bridge Behavior . . . . . . . . . . . . . . . . 27 83 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 27 84 3.5.1. OMNI Neighbor Interface Attributes . . . . . . . . . 29 85 3.5.2. OMNI Neighbor Advertisement Message Flags . . . . . . 30 86 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 30 87 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 32 88 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 32 89 3.9. OMNI Adaptation Layer and OMNI Interface MTU . . . . . . 32 90 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 33 91 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 34 92 3.10.2. Proxy Forwarding Algorithm . . . . . . . . . . . . . 35 93 3.10.3. Server/Relay Forwarding Algorithm . . . . . . . . . 36 94 3.10.4. Bridge Forwarding Algorithm . . . . . . . . . . . . 38 96 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 39 97 3.12. AERO Router Discovery, Prefix Delegation and 98 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 41 99 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 41 100 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 42 101 3.12.3. AERO Server Behavior . . . . . . . . . . . . . . . . 44 102 3.13. The AERO Proxy . . . . . . . . . . . . . . . . . . . . . 47 103 3.13.1. Combined Proxy/Servers . . . . . . . . . . . . . . . 49 104 3.13.2. Detecting and Responding to Server Failures . . . . 49 105 3.13.3. Point-to-Multipoint Server Coordination . . . . . . 50 106 3.14. AERO Route Optimization / Address Resolution . . . . . . 51 107 3.14.1. Route Optimization Initiation . . . . . . . . . . . 51 108 3.14.2. Relaying the NS(AR) . . . . . . . . . . . . . . . . 52 109 3.14.3. Processing the NS(AR) and Sending the NA(AR) . . . . 52 110 3.14.4. Relaying the NA(AR) . . . . . . . . . . . . . . . . 53 111 3.14.5. Processing the NA(AR) . . . . . . . . . . . . . . . 54 112 3.14.6. Route Optimization Maintenance . . . . . . . . . . . 54 113 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 55 114 3.16. Mobility Management and Quality of Service (QoS) . . . . 56 115 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 57 116 3.16.2. Announcing Link-Layer Address and/or QoS Preference 117 Changes . . . . . . . . . . . . . . . . . . . . . . 58 118 3.16.3. Bringing New Links Into Service . . . . . . . . . . 59 119 3.16.4. Deactivating Existing Links . . . . . . . . . . . . 59 120 3.16.5. Moving Between Servers . . . . . . . . . . . . . . . 59 121 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 60 122 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 61 123 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 62 124 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 63 125 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 63 126 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 64 127 3.20. Transition Considerations . . . . . . . . . . . . . . . . 64 128 3.21. Detecting and Reacting to Server and Bridge Failures . . 65 129 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 66 130 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 68 131 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 69 132 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 69 133 6. Security Considerations . . . . . . . . . . . . . . . . . . . 69 134 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 71 135 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 73 136 8.1. Normative References . . . . . . . . . . . . . . . . . . 73 137 8.2. Informative References . . . . . . . . . . . . . . . . . 74 138 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 80 139 A.1. Implementation Strategies for Route Optimization . . . . 81 140 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 81 141 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 82 142 A.4. AERO Critical Infrastructure Considerations . . . . . . . 82 143 A.5. AERO Server Failure Implications . . . . . . . . . . . . 83 144 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 83 145 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 85 146 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 86 148 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 as single-hop neighbors via 160 encapsulation. The OMNI Adaptation Layer (OAL) supports multilink 161 operation for increased reliability, bandwidth optimization and 162 traffic path selection while accommodating Maximum Transmission Unit 163 (MTU) diversity. 165 The AERO service comprises Clients, Proxys, Servers and Relays that 166 are seen as OMNI link neighbors as well as Bridges that interconnect 167 OMNI link segments. Each node's OMNI interface uses an IPv6 link- 168 local address format that supports operation of the IPv6 Neighbor 169 Discovery (ND) protocol [RFC4861] and links ND to IP forwarding. A 170 node's OMNI interface can be configured over multiple underlying 171 interfaces, and may therefore appear as a single interface with 172 multiple link-layer addresses. Each link-layer address is subject to 173 change due to mobility and/or QoS fluctuations, and link-layer 174 address changes are signaled by ND messaging the same as for any IPv6 175 link. 177 AERO provides a cloud-based service where mobile node Clients may use 178 any Server acting as a Mobility Anchor Point (MAP) and fixed nodes 179 may use any Relay on the link for efficient communications. Fixed 180 nodes forward packets destined to other AERO nodes via the nearest 181 Relay, which forwards them through the cloud. A mobile node's 182 initial packets are forwarded through the Server, while direct 183 routing is supported through asymmetric extended route optimization 184 while data packets are flowing. Both unicast and multicast 185 communications are supported, and mobile nodes may efficiently move 186 between locations while maintaining continuous communications with 187 correspondents and without changing their IP Address. 189 AERO Bridges are interconnected in a secured private BGP overlay 190 routing instance using encapsulation to provide a hybrid routing/ 191 bridging service that joins the underlying Internetworks of multiple 192 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 can be used with OMNI links that span private-use Internetworks 203 and/or public Internetworks such as the global Internet. In the 204 latter case, some end systems may be located behind global Internet 205 Network Address Translators (NATs). A means for robust traversal of 206 NATs while avoiding "triangle routing" is therefore provided. 208 AERO assumes the use of PIM Sparse Mode in support of multicast 209 communication. In support of Source Specific Multicast (SSM) when a 210 Mobile Node is the source, AERO route optimization ensures that a 211 shortest-path multicast tree is established with provisions for 212 mobility and multilink operation. In all other multicast scenarios 213 there are no AERO dependencies. 215 AERO was designed for aeronautical networking for both manned and 216 unmanned aircraft, where the aircraft is treated as a mobile node 217 that can connect an Internet of Things (IoT). AERO is also 218 applicable to a wide variety of other use cases. For example, it can 219 be used to coordinate the links of mobile nodes (e.g., cellphones, 220 tablets, laptop computers, etc.) that connect into a home enterprise 221 network via public access networks using tunneling software such as 222 OpenVPN [OVPN] with VPN or non-VPN services enabled according to the 223 appropriate security model. AERO can also be used to facilitate 224 terrestrial vehicular and urban air mobility (as well as pedestrian 225 communication services) for future intelligent transportation systems 226 [I-D.ietf-ipwave-vehicular-networking][I-D.templin-ipwave-uam-its]. 227 Other applicable use cases are also in scope. 229 The following numbered sections present the AERO specification. The 230 appendices at the end of the document are non-normative. 232 2. Terminology 234 The terminology in the normative references applies; especially, the 235 terminology in the OMNI specification 236 [I-D.templin-6man-omni-interface] is used extensively throughout. 237 The following terms are defined within the scope of this document: 239 IPv6 Neighbor Discovery (ND) 240 an IPv6 control message service for coordinating neighbor 241 relationships between nodes connected to a common link. AERO uses 242 the ND service specified in [RFC4861]. 244 IPv6 Prefix Delegation 245 a networking service for delegating IPv6 prefixes to nodes on the 246 link. The nominal service is DHCPv6 [RFC8415], however alternate 247 services (e.g., based on ND messaging) are also in scope. Most 248 notably, a minimal form of prefix delegation known as "prefix 249 registration" can be used if the Client knows its prefix in 250 advance and can represent it in the IPv6 source address of an ND 251 message. 253 Access Network (ANET) 254 a node's first-hop data link service network (e.g., a radio access 255 network, cellular service provider network, corporate enterprise 256 network, etc.) that often provides link-layer security services 257 such as IEEE 802.1X and physical-layer security prevent 258 unauthorized access internally and with border network-layer 259 security services such as firewalls and proxies that prevent 260 unauthorized outside access. 262 ANET interface 263 a node's attachment to a link in an ANET. 265 Internetwork (INET) 266 a connected IP network topology with a coherent routing and 267 addressing plan and that provides a transit backbone service for 268 ANET end systems. INETs also provide an underlay service over 269 which the AERO virtual link is configured. Example INETs include 270 corporate enterprise networks, aviation networks, and the public 271 Internet itself. When there is no administrative boundary between 272 an ANET and the INET, the ANET and INET are one and the same. 274 INET Partition 275 frequently, INETs such as large corporate enterprise networks are 276 sub-divided internally into separate isolated partitions. Each 277 partition is fully connected internally but disconnected from 278 other partitions, and there is no requirement that separate 279 partitions maintain consistent Internet Protocol and/or addressing 280 plans. (Each INET partition is seen as a separate OMNI link 281 segment as discussed below.) 283 INET interface 284 a node's attachment to a link in an INET. 286 INET address 287 an IP address assigned to a node's interface connection to an 288 INET. 290 INET encapsulation 291 the encapsulation of a packet in an outer header or headers that 292 can be routed within the scope of the local INET partition. 294 OMNI link 295 the same as defined in [I-D.templin-6man-omni-interface], and 296 manifested by IPv6 encapsulation [RFC2473]. The OMNI link spans 297 underlying INET segments joined by virtual bridges in a spanning 298 tree the same as a bridged campus LAN. AERO nodes on the OMNI 299 link appear as single-hop neighbors even though they may be 300 separated by multiple underlying INET hops, and can use Segment 301 Routing [RFC8402] to cause packets to visit selected waypoints on 302 the link. 304 OMNI Interface 305 a node's attachment to an OMNI link. Since the addresses assigned 306 to an OMNI interface are managed for uniqueness, OMNI interfaces 307 do not require Duplicate Address Detection (DAD) and therefore set 308 the administrative variable 'DupAddrDetectTransmits' to zero 309 [RFC4862]. 311 OMNI Adaptation Layer (OAL) 312 an OMNI interface process whereby packets admitted into the 313 interface are wrapped in a mid-layer IPv6 header and fragmented/ 314 reassembled if necessary to support the OMNI link Maximum 315 Transmission Unit (MTU). The OAL is also responsible for 316 generating MTU-related control messages as necessary, and for 317 providing addressing context for spanning multiple segments of a 318 bridged OMNI link. 320 underlying interface 321 an ANET or INET interface over which an OMNI interface is 322 configured. 324 Mobility Service Prefix (MSP) 325 an aggregated IP Global Unicast Address (GUA) prefix (e.g., 326 2001:db8::/32, 192.0.2.0/24, etc.) assigned to the OMNI link and 327 from which more-specific Mobile Network Prefixes (MNPs) are 328 delegated. OMNI link administrators typically obtain MSPs from an 329 Internet address registry, however private-use prefixes can 330 alternatively be used subject to certain limitations (see: 331 [I-D.templin-6man-omni-interface]). OMNI links that connect to 332 the global Internet advertise their MSPs to their interdomain 333 routing peers. 335 Mobile Network Prefix (MNP) 336 a longer IP prefix delegated from an MSP (e.g., 337 2001:db8:1000:2000::/56, 192.0.2.8/30, etc.) and assigned to an 338 AERO Client or Relay. 340 Mobile Network Prefix Link Local Address (MNP-LLA) 341 an IPv6 Link Local Address that embeds the most significant 64 342 bits of an MNP in the lower 64 bits of fe80::/64, as specified in 343 [I-D.templin-6man-omni-interface]. 345 Mobile Network Prefix Unique Local Address (MNP-ULA) 346 an IPv6 Unique-Local Address derived from an MNP-LLA. 348 Administrative Link Local Address (ADM-LLA) 349 an IPv6 Link Local Address that embeds a 32-bit administratively- 350 assigned identification value in the lower 32 bits of fe80::/96, 351 as specified in [I-D.templin-6man-omni-interface]. 353 Administrative Unique Local Address (ADM-ULA) 354 an IPv6 Unique-Local Address derived from an ADM-LLA. 356 AERO node 357 a node that is connected to an OMNI link and participates in the 358 AERO internetworking and mobility service. 360 AERO Client ("Client") 361 an AERO node that connects over one or more underlying interfaces 362 and requests MNP delegation/registration service from AERO 363 Servers. The Client assigns an MNP-LLA to the OMNI interface for 364 use in ND exchanges with other AERO nodes and forwards packets to 365 correspondents according to OMNI interface neighbor cache state. 367 AERO Server ("Server") 368 an INET node that configures an OMNI interface to provide default 369 forwarding and mobility/multilink services for AERO Clients. The 370 Server assigns an ADM-LLA to its OMNI interface to support the 371 operation of the ND services, and advertises all of its associated 372 MNPs via BGP peerings with Bridges. 374 AERO Relay ("Relay") 375 an AERO Server that also provides forwarding services between 376 nodes reached via the OMNI link and correspondents on other links. 377 AERO Relays are provisioned with MNPs (i.e., the same as for an 378 AERO Client) and run a dynamic routing protocol to discover any 379 non-MNP IP GUA routes in service on its connected INET links. In 380 both cases, the Relay advertises the MSP(s) to its downstream 381 networks, and distributes all of its associated MNPs and non-MNP 382 IP GUA routes via BGP peerings with Bridges (i.e., the same as for 383 an AERO Server). 385 AERO Bridge ("Bridge") 386 a node that provides hybrid routing/bridging services (as well as 387 a security trust anchor) for nodes on an OMNI link. As a router, 388 the Bridge forwards packets using standard IP forwarding. As a 389 bridge, the Bridge forwards packets over the OMNI link without 390 decrementing the IPv6 Hop Limit. AERO Bridges peer with Servers 391 and other Bridges to discover the full set of MNPs for the link as 392 well as any non-MNP IP GUA routes that are reachable via Relays. 394 AERO Proxy ("Proxy") 395 a node that provides proxying services between Clients in an ANET 396 and Servers in external INETs. The AERO Proxy is a conduit 397 between the ANET and external INETs in the same manner as for 398 common web proxies, and behaves in a similar fashion as for ND 399 proxies [RFC4389]. A node may be configured to act as either a 400 Proxy and/or a Server, depending on Client Server selection 401 criteria. 403 ingress tunnel endpoint (ITE) 404 an OMNI interface endpoint that injects encapsulated packets into 405 an OMNI link. 407 egress tunnel endpoint (ETE) 408 an OMNI interface endpoint that receives encapsulated packets from 409 an OMNI link. 411 link-layer address 412 an IP address used as an encapsulation header source or 413 destination address from the perspective of the OMNI interface. 414 When an upper layer protocol (e.g., UDP) is used as part of the 415 encapsulation, the port number is also considered as part of the 416 link-layer address. 418 network layer address 419 the source or destination address of an encapsulated IP packet 420 presented to the OMNI interface. 422 end user network (EUN) 423 an internal virtual or external edge IP network that an AERO 424 Client or Relay connects to the rest of the network via the OMNI 425 interface. The Client/Relay sees each EUN as a "downstream" 426 network, and sees the OMNI interface as the point of attachment to 427 the "upstream" network. 429 Mobile Node (MN) 430 an AERO Client and all of its downstream-attached networks that 431 move together as a single unit, i.e., an end system that connects 432 an Internet of Things. 434 Mobile Router (MR) 435 a MN's on-board router that forwards packets between any 436 downstream-attached networks and the OMNI link. The MR is the MN 437 entity that hosts the AERO Client. 439 Route Optimization Source (ROS) 440 the AERO node nearest the source that initiates route 441 optimization. The ROS may be a Server or Proxy acting on behalf 442 of the source Client, or may be the Client itself if the Client is 443 connected to the INET either directly or through a NAT. 445 Route Optimization responder (ROR) 446 the AERO node nearest the target destination that responds to 447 route optimization requests. The ROR may be a Server acting on 448 behalf of a target MNP Client, or a Relay for a non-MNP 449 destination. 451 MAP List 452 a geographically and/or topologically referenced list of addresses 453 of all Servers within the same OMNI link. There is a single MAP 454 list for the entire OMNI link. 456 Distributed Mobility Management (DMM) 457 a BGP-based overlay routing service coordinated by Servers and 458 Bridges that tracks all Server-to-Client associations. 460 Mobility Service (MS) 461 the collective set of all Servers, Proxys, Bridges and Relays that 462 provide the AERO Service to Clients. 464 Mobility Service Endpoint MSE) 465 an individual Server, Proxy, Bridge or Relay in the Mobility 466 Service. 468 Throughout the document, the simple terms "Client", "Server", 469 "Bridge", "Proxy" and "Relay" refer to "AERO Client", "AERO Server", 470 "AERO Bridge", "AERO Proxy" and "AERO Relay", respectively. 471 Capitalization is used to distinguish these terms from other common 472 Internetworking uses in which they appear without capitalization. 474 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 475 the names of node variables, messages and protocol constants) is used 476 throughout this document. The terms "All-Routers multicast", "All- 477 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 478 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 479 generically refer to either Internet Protocol version, i.e., IPv4 480 [RFC0791] or IPv6 [RFC8200]. 482 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 483 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 484 "OPTIONAL" in this document are to be interpreted as described in BCP 485 14 [RFC2119][RFC8174] when, and only when, they appear in all 486 capitals, as shown here. 488 3. Asymmetric Extended Route Optimization (AERO) 490 The following sections specify the operation of IP over OMNI links 491 using the AERO service: 493 3.1. AERO Node Types 495 AERO Clients are Mobile Nodes (MNs) that connect via underlying 496 interfaces with addresses that may change when the Client moves to a 497 new network connection point. AERO Clients register their Mobile 498 Network Prefixes (MNPs) with the AERO service, and distribute the 499 MNPs to nodes on EUNs. AERO Bridges, Servers, Proxys and Relays are 500 critical infrastructure elements in fixed (i.e., non-mobile) INET 501 deployments and hence have permanent and unchanging INET addresses. 502 Together, they constitute the AERO service which provides an OMNI 503 link virtual overlay for connecting AERO Clients. 505 AERO Bridges provide hybrid routing/bridging services (as well as a 506 security trust anchor) for nodes on an OMNI link. Bridges use 507 standard IPv6 routing to forward packets both within the same INET 508 partitions and between disjoint INET partitions based on a mid-layer 509 IPv6 encapsulation per [RFC2473]. The inner IP layer experiences a 510 virtual bridging service since the inner IP TTL/Hop Limit is not 511 decremented during forwarding. Each Bridge also peers with Servers 512 and other Bridges in a dynamic routing protocol instance to provide a 513 Distributed Mobility Management (DMM) service for the list of active 514 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 515 one or more Mobility Service Prefixes (MSPs) and configure secured 516 tunnels with Servers, Relays, Proxys and other Bridges; they further 517 maintain IP forwarding table entries for each MNP and any other 518 reachable non-MNP prefixes. 520 AERO Servers provide default forwarding and mobility/multilink 521 services for AERO Client Mobile Nodes (MNs). Each Server also peers 522 with Bridges in a dynamic routing protocol instance to advertise its 523 list of associated MNPs (see Section 3.2.3). Servers facilitate 524 prefix delegation/registration exchanges with Clients, where each 525 delegated prefix becomes an MNP taken from an MSP. Servers forward 526 packets between OMNI interface neighbors and track each Client's 527 mobility profiles. Servers may further act as Servers for some sets 528 of Clients and as Proxies for others. 530 AERO Proxys provide a conduit for ANET Clients to associate with 531 Servers in external INETs. Client and Servers exchange control plane 532 messages via the Proxy acting as a bridge between the ANET/INET 533 boundary. The Proxy forwards data packets between Clients and the 534 OMNI link according to forwarding information in the neighbor cache. 535 The Proxy function is specified in Section 3.13. Proxys may further 536 act as Proxys for some sets of Clients and as Servers for others. 538 AERO Relays are Servers that provide forwarding services between the 539 OMNI interface and INET/EUN interfaces. Relays are provisioned with 540 MNPs the same as for an AERO Client, and also run a dynamic routing 541 protocol to discover any non-MNP IP routes. The Relay advertises the 542 MSP(s) to its connected networks, and distributes all of its 543 associated MNPs and non-MNP IP GUA routes via BGP peerings with 544 Bridges 546 3.2. The AERO Service over OMNI Links 548 3.2.1. AERO/OMNI Reference Model 550 Figure 1 presents the basic OMNI link reference model: 552 +----------------+ 553 | AERO Bridge B1 | 554 | Nbr: S1, S2, P1| 555 |(X1->S1; X2->S2)| 556 | MSP M1 | 557 +-+---------+--+-+ 558 +--------------+ | Secured | | +--------------+ 559 |AERO Server S1| | tunnels | | |AERO Server S2| 560 | Nbr: C1, B1 +-----+ | +-----+ Nbr: C2, B1 | 561 | default->B1 | | | default->B1 | 562 | X1->C1 | | | X2->C2 | 563 +-------+------+ | +------+-------+ 564 | OMNI link | | 565 X===+===+===================+==)===============+===+===X 566 | | | | 567 +-----+--------+ +--------+--+-----+ +--------+-----+ 568 |AERO Client C1| | AERO Proxy P1 | |AERO Client C2| 569 | Nbr: S1 | |(Proxy Nbr Cache)| | Nbr: S2 | 570 | default->S1 | +--------+--------+ | default->S2 | 571 | MNP X1 | | | MNP X2 | 572 +------+-------+ .--------+------. +-----+--------+ 573 | (- Proxyed Clients -) | 574 .-. `---------------' .-. 575 ,-( _)-. ,-( _)-. 576 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 577 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 578 `-(______)-' +-------+ +-------+ `-(______)-' 580 Figure 1: AERO/OMNI Reference Model 582 In this model: 584 o the OMNI link is an overlay network service configured over one or 585 more underlying INET partitions which may be managed by different 586 administrative authorities and have incompatible protocols and/or 587 addressing plans. 589 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 590 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 591 via BGP peerings over secured tunnels to Servers (S1, S2). 592 Bridges connect the disjoint segments of a partitioned OMNI link. 594 o AERO Servers/Relays S1 and S2 configure secured tunnels with 595 Bridge B1 and also provide mobility, multilink and default router 596 services for their associated Clients C1 and C2. 598 o AERO Clients C1 and C2 associate with Servers S1 and S2, 599 respectively. They receive Mobile Network Prefix (MNP) 600 delegations X1 and X2, and also act as default routers for their 601 associated physical or internal virtual EUNs. Simple hosts H1 and 602 H2 attach to the EUNs served by Clients C1 and C2, respectively. 604 o AERO Proxy P1 configures a secured tunnel with Bridge B1 and 605 provides proxy services for AERO Clients in secured enclaves that 606 cannot associate directly with other OMNI link neighbors. 608 An OMNI link configured over a single INET appears as a single 609 unified link with a consistent underlying network addressing plan. 610 In that case, all nodes on the link can exchange packets via simple 611 INET encapsulation, since the underlying INET is connected. In 612 common practice, however, an OMNI link may be partitioned into 613 multiple "segments", where each segment is a distinct INET 614 potentially managed under a different administrative authority (e.g., 615 as for worldwide aviation service providers such as ARINC, SITA, 616 Inmarsat, etc.). Individual INETs may also themselves be partitioned 617 internally, in which case each internal partition is seen as a 618 separate segment. 620 The addressing plan of each segment is consistent internally but will 621 often bear no relation to the addressing plans of other segments. 622 Each segment is also likely to be separated from others by network 623 security devices (e.g., firewalls, proxies, packet filtering 624 gateways, etc.), and in many cases disjoint segments may not even 625 have any common physical link connections. Therefore, nodes can only 626 be assured of exchanging packets directly with correspondents in the 627 same segment, and not with those in other segments. The only means 628 for joining the segments therefore is through inter-domain peerings 629 between AERO Bridges. 631 The same as for traditional campus LANs, multiple OMNI link segments 632 can be joined into a single unified link via a virtual bridging 633 service using the OMNI Adaptation Layer (OAL) which inserts a mid- 634 layer IPv6 encapsulation per [RFC2473] that supports inter-segment 635 forwarding (i.e., bridging) without decrementing the network-layer 636 TTL/Hop Limit. This bridging of OMNI link segments is shown in 637 Figure 2: 639 . . . . . . . . . . . . . . . . . . . . . . . 640 . . 641 . .-(::::::::) . 642 . .-(::::::::::::)-. +-+ . 643 . (:::: Segment A :::)--|B|---+ . 644 . `-(::::::::::::)-' +-+ | . 645 . `-(::::::)-' | . 646 . | . 647 . .-(::::::::) | . 648 . .-(::::::::::::)-. +-+ | . 649 . (:::: Segment B :::)--|B|---+ . 650 . `-(::::::::::::)-' +-+ | . 651 . `-(::::::)-' | . 652 . | . 653 . .-(::::::::) | . 654 . .-(::::::::::::)-. +-+ | . 655 . (:::: Segment C :::)--|B|---+ . 656 . `-(::::::::::::)-' +-+ | . 657 . `-(::::::)-' | . 658 . | . 659 . ..(etc).. x . 660 . . 661 . . 662 . <- OMNI link Bridged by encapsulation -> . 663 . . . . . . . . . . . . . .. . . . . . . . . 665 Figure 2: Bridging OMNI Link Segments 667 Bridges, Servers, Relays and Proxys connect via secured INET tunnels 668 over their respective segments in a spanning tree topology rooted at 669 the Bridges. The secured spanning tree supports strong 670 authentication for IPv6 ND control messages and may also be used to 671 convey the initial data packets in a flow. Route optimization can 672 then be employed to cause data packets to take more direct paths 673 between OMNI link neighbors without having to strictly follow the 674 spanning tree. 676 3.2.2. Link-Local Addresses (LLAs), Unique Local Addresses (ULAs) and 677 Global Unicast Addresses (GUA) 679 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 680 fe80::/64 [RFC4291] to assign LLAs used for network-layer addresses 681 in link-scoped IPv6 ND and data messages. AERO Clients use LLAs 682 constructed from MNPs (i.e., "MNP-LLAs") while other AERO nodes use 683 LLAs constructed from administrative identification values ("ADM- 684 LLAs") as specified in [I-D.templin-6man-omni-interface]. 686 AERO nodes also use the Unique Local Address (ULA) prefix fd00::/8 687 followed by a pseudo-random 40-bit global ID to form the prefix 688 [ULA]::/48, then include a 16-bit OMNI link identifier '*' to form 689 the prefix [ULA*]::/64 [RFC4291]. The AERO node then uses the prefix 690 [ULA*]::/64 to form "MNP-ULAs" or "ADM-ULA"s as specified in 691 [I-D.templin-6man-omni-interface] to support OAL addressing. 693 AERO Clients also use Temporary LLAs/ULAs constructed per 694 [I-D.templin-6man-omni-interface], where the addresses are typically 695 used only in initial control message exchanges until a stable MNP- 696 LLA/ULA is assigned. 698 Finally, the MSPs and MNPs used by AERO nodes are typically based on 699 Global Unicast Addresses (GUAs), but in some cases may be based on 700 private-use addresses. See for a full specification of LLAs, ULAs 701 and GUAs used by AERO nodes on OMNI links. 703 3.2.3. AERO Routing System 705 The AERO routing system comprises a private instance of the Border 706 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 707 and Servers and does not interact with either the public Internet BGP 708 routing system or any underlying INET routing systems. 710 In a reference deployment, each Server is configured as an Autonomous 711 System Border Router (ASBR) for a stub Autonomous System (AS) using a 712 32-bit AS Number (ASN) [RFC4271] that is unique within the BGP 713 instance, and each Server further uses eBGP to peer with one or more 714 Bridges but does not peer with other Servers. Each INET of a multi- 715 segment OMNI link must include one or more Bridges, which peer with 716 the Servers and Proxys within that INET. All Bridges within the same 717 INET are members of the same hub AS, and use iBGP to maintain a 718 consistent view of all active MNPs currently in service. The Bridges 719 of different INETs peer with one another using eBGP. 721 Bridges advertise the OMNI link's MSPs and any non-MNP routes to each 722 of their Servers. This means that any aggregated non-MNPs (including 723 "default") are advertised to all Servers. Each Bridge configures a 724 black-hole route for each of its MSPs. By black-holing the MSPs, the 725 Bridge will maintain forwarding table entries only for the MNPs that 726 are currently active, and packets destined to all other MNPs will 727 correctly incur Destination Unreachable messages due to the black- 728 hole route. In this way, Servers have only partial topology 729 knowledge (i.e., they know only about the MNPs of their directly 730 associated Clients) and they forward all other packets to Bridges 731 which have full topology knowledge. 733 Each OMNI link segment assigns a unique ADM-ULA sub-prefix of 734 [ULA*]::/96. For example, a first segment could assign 735 [ULA*]::1000/116, a second could assign [ULA*]::2000/116, a third 736 could assign [ULA*]::3000/116, etc. Within each segment, each Proxy/ 737 Server configures an ADM-ULA within the segment's prefix, e.g., the 738 Proxy/Servers within [ULA*]::2000/116 could assign the ADM-ULAs 739 [ULA*]::2011/116, [ULA*]::2026/116, [ULA*]::2003/116, etc. 741 The administrative authorities for each segment must therefore 742 coordinate to assure mutually-exclusive ADM-ULA prefix assignments, 743 but internal provisioning of ADM-ULAs an independent local 744 consideration for each administrative authority. For each ADM-ULA 745 prefix, the Bridge(s) that connect that segment assign the all-zero's 746 address of the prefix as a Subnet Router Anycast address. For 747 example, the Subnet Router Anycast address for [ULA*]::1023/116 is 748 simply [ULA*]::1000. 750 ADM-ULA prefixes are statically represented in Bridge forwarding 751 tables. Bridges join multiple segments into a unified OMNI link over 752 multiple diverse administrative domains. They support a bridging 753 function by first establishing forwarding table entries for their 754 ADM-ULA prefixes either via standard BGP routing or static routes. 755 For example, if three Bridges ('A', 'B' and 'C') from different 756 segments serviced [ULA*]::1000/116, [ULA*]::2000/116 and 757 [ULA*]::3000/116 respectively, then the forwarding tables in each 758 Bridge are as follows: 760 A: [ULA*]::1000/116->local, [ULA*]::2000/116->B, [ULA*]::3000/116->C 762 B: [ULA*]::1000/116->A, [ULA*]::2000/116->local, [ULA*]::3000/116->C 764 C: [ULA*]::1000/116->A, [ULA*]::2000/116->B, [ULA*]::3000/116->local 766 These forwarding table entries are permanent and never change, since 767 they correspond to fixed infrastructure elements in their respective 768 segments. 770 MNP ULAs are instead dynamically advertised in the AERO routing 771 system by Servers and Relays that provide service for their 772 corresponding MNPs. For example, if three Servers ('D', 'E' and 'F') 773 service the MNPs 2001:db8:1000:2000::/56, 2001:db8:3000:4000::/56 and 774 2001:db8:5000:6000::/56 then the routing system would include: 776 D: [ULA*]:2001:db8:1000:2000/120 778 E: [ULA*]:2001:db8:3000:4000/120 780 F: [ULA*]:2001:db8:5000:6000/120 781 A full discussion of the BGP-based routing system used by AERO is 782 found in [I-D.ietf-rtgwg-atn-bgp]. 784 3.2.4. OMNI Link Encapsulation 786 With the Client and partition prefixes in place in Bridge forwarding 787 tables, the OMNI interface sends control and data packets toward AERO 788 destination nodes located in different OMNI link segments over the 789 spanning tree via mid-layer encapsulation using the OMNI Adaptation 790 Layer (OAL) header based on Generic Packet Tunneling in IPv6 791 [RFC2473]. When necessary, the OMNI interface also includes an OMNI 792 Routing Header (ORH) as an extension to the OAL header if final 793 segment forwarding information is available, e.g., in the neighbor 794 cache. The ORH is formatted as shown in Figure 3: 796 0 1 2 3 797 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 1 798 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 799 | Next Header | Hdr Ext Len | Routing Type | SRT | FMT | 800 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 801 | Last Hop Segment-id (LHS) | 802 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 803 ~ Link Layer Address (L2ADDR) ~ 804 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 805 ~ Destination Suffix (if necessary) ~ 806 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 807 ~ Null Padding (if necessary) ~ 808 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 810 Figure 3: OMNI Routing Header (ORH) Format 812 In this format: 814 o Next Header identifies the type of header immediately following 815 the ORH. 817 o Hdr Ext Len is the length of the Routing header in 8-octet units 818 (not including the first 8 octets), with trailing padding added if 819 necessary to produce an integral number of 8-octet units. 821 o Routing Type is set to TBD (see IANA Considerations). 823 o Segments Left is omitted, and replaced by a 5-bit SRT and 3-bit 824 FMT field. 826 o SRT - a 5-bit Segment Routing Topology prefix length value that 827 (when added to 96) determines the prefix length to apply to the 828 ADM-ULA formed from concatenating [ULA*]::/96 with the 32 bit LHS 829 value that follows. For example, the value 16 corresponds to the 830 prefix length 112. 832 o FMT - a 3-bit "Framework/Mode/Type" code corresponding to the 833 included Link Layer Address as follows: 835 * When the most significant bit (i.e., "Framework") is set to 1, 836 L2ADDR is the INET encapsulation address for the target Client 837 itself; otherwise is the address of the Server/Proxy named in 838 the LHS. 840 * When the next most significant bit (i.e., "Mode") is set to 1, 841 the target Client is (likely) located behind an INET Network 842 Address Translator (NAT); otherwise, it is reachable only via 843 the Server/Proxy or directly addressable via INET addressing. 845 * When the least significant bit (i.e., "Type") is set to 0, 846 L2ADDR includes a UDP Port Number followed by an IPv4 address; 847 otherwise, it includes a UDP Port Number followed by an IPv6 848 address. 850 o LHS - the 32 bit ID of a node in the Last Hop Segment that 851 services the target. When SRT and LHS are both set to 0, the LHS 852 is considered unspecified. When SRT is set to 0 and LHS is non- 853 zero, the prefix length is set to 128. SRT and LHS provide 854 guidance to the OMNI interface forwarding algorithm. 855 Specifically, if SRT/LHS is located in the local OMNI link segment 856 then the OMNI interface can omit OAL/ORH encapsulation and 857 (following any necessary NAT traversal messaging) send directly to 858 the target using INET encapsulation according to FMT/L2ADDR. 859 Otherwise, it must perform INET/OAL/ORH encapsulation and forward 860 according to the OMNI link spanning tree. 862 o Link Layer Address (L2ADDR) - Formatted according to FMT, and 863 identifies the link-layer address (i.e., the encapsulation 864 address) of the target. The UDP Port Number appears in the first 865 two octets and the IP address appears in the next 4 octets for 866 IPv4 or 16 octets for IPv6. The Port Number and IP address are 867 recorded in ones-compliment "obfuscated" form per [RFC4380]. The 868 OMNI interface forwarding algorithm uses FMT/L2ADDR to determine 869 the INET encapsulation address for local forwarding when SRT/LHS 870 is located in the same OMNI link segment. Note that if the target 871 is behind a NAT, L2ADDR will contain the mapped INET address 872 stored in the NAT; otherwise, L2ADDR will contain the native INET 873 information of the target itself. 875 o Destination Suffix is a 64-bit field included only for OAL- 876 encapsulated packets that require OAL fragmentation. Present only 877 when Hdr Ext Len indicates that at least 8 bytes follow L2ADDR. 878 When present, encodes the 64-bit MNP-ULA suffix for the target 879 Client. For example, if the MNP-ULA is [ULA*]:2001:db8:1:2 then 880 the Destination suffix encodes the value 2001:db8:1:2. 882 o Null Padding contains zero-valued octets as necessary to pad the 883 ORH to an integral number of 8-octet units. 885 AERO nodes use OAL encapsulation when the target destination is 886 located in a different OMNI link segment, or when OAL fragmentation 887 is necessary (otherwise, OAL encapsulation is omitted and simple INET 888 encapsulation is used). When an AERO node uses OAL encapsulation for 889 a packet with source address 2001:db8:1:2::1 and destination address 890 2001:db8:1234:5678::1, it sets the OAL header source address to its 891 own ULA (e.g., [ULA*]::1000:2000) and sets the destination address to 892 the MNP-ULA corresponding to the IP destination address (e.g., 893 [ULA*]::2001:db8:1234:5678). 895 If the neighbor cache information indicates that the target is in a 896 different segment, the node next inserts an ORH immediately following 897 the OAL header while including the correct SRT, FMT, LHS, L2ADDR and 898 Destination Suffix if fragmentation is needed (in this case, the 899 Destination Suffix is 2001:db8:1234:5678). Next, the node overwrites 900 the OAL header destination address with the LHS Subnet Router Anycast 901 address (for example, for LHS 3000:4567 with SRT 16, the Subnet 902 Router Anycast address is [ULA*]::3000:0000). (Note: if ADM-ULA of 903 the last-hop Proxy/Server is known but the SRT, FMT, LHS and L2ADDR 904 are not (yet) known, the node instead overwrites the OAL header 905 destination address with the ADM-ULA.) 907 The node then fragments the OAL/ORH packet if necessary, with each 908 resulting fragment including the OAL/ORH headers while only the first 909 fragment includes the original IPv6 header. The node finally 910 encapsulates each resulting OAL/ORH packet/fragment in an INET header 911 with source address set to its own INET address (e.g., 192.0.2.100) 912 and destination set to the INET address of a Bridge (e.g., 913 192.0.2.1). 915 The encapsulation format in the above example is shown in Figure 4: 917 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 918 | INET Header | 919 | src = 192.0.2.100 | 920 | dst = 192.0.2.1 | 921 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 922 | OAL Header (if necessary) | 923 | src = [ULA*]::1000:2000 | 924 | dst= [ULA*]::3000:0000 | 925 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 926 | ORH Header (if necessary) | 927 | Destination Suffix: | 928 | 2001:db8:1234:5678 | 929 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 930 | Inner IP Header | 931 | src = 2001:db8:1:2::1 | 932 | dst = 2001:db8:1234:5678::1 | 933 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 934 | | 935 ~ ~ 936 ~ Inner Packet Body ~ 937 ~ ~ 938 | | 939 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 941 Figure 4: OAL/ORH Encapsulation 943 In this format, the inner IP header and packet body are the original 944 IP packet, the OAL header is an IPv6 header prepared according to 945 [RFC2473], the ORH is a Routing Header extension of the OAL header, 946 and the INET header is prepared as discussed in Section 3.6. 948 This gives rise to a routing system that contains both Client prefix 949 routes that may change dynamically due to regional node mobility and 950 partition prefix routes that rarely if ever change. The Bridges can 951 therefore provide link-layer bridging by sending packets over the 952 spanning tree instead of network-layer routing according to MNP 953 routes. As a result, opportunities for packet loss due to node 954 mobility between different segments are mitigated. 956 In normal operations, IPv6 ND messages are conveyed over secured 957 paths between OMNI link neighbors so that specific Proxys, Servers or 958 Relays can be addressed without being subject to mobility events. 959 Conversely, only the first few packets destined to Clients need to 960 traverse secured paths until route optimization can determine a more 961 direct path. 963 Note: When the source OAL has multiple inner IP packets to send to 964 the same destination, it can perform "packing" to generate a "super- 965 packet" as discussed in Section 5.2 of 966 [I-D.templin-6man-omni-interface]. In that case, the OAL/ORH super- 967 packet may include up to N inner packets as long as the total length 968 of the super-packet does not exceed the OMNI interface MTU. 970 Note: Use of an IPv6 "minimal encapsulation" format (i.e., an IPv6 971 variant of [RFC2004]) based on extensions to the ORH was considered 972 and abandoned. In the approach, the ORH would be inserted as an 973 extension header to the original IPv6 packet header. The IPv6 974 destination address would then be written into the ORH, and the ULA 975 corresponding to the destination would be overwritten in the IPv6 976 destination address. This would seemingly convey enough forwarding 977 information so that OAL encapsulation could be avoided. However, 978 this "minimal encapsulation" IPv6 packet would then have a non-ULA 979 source address and ULA destination address, an incorrect value in 980 upper layer protocol checksums, and a Hop Limit that is decremented 981 within the spanning tree when it should not be. The insertion and 982 removal of the ORH would also entail rewriting the Payload Length and 983 Next Header fields - again, invalidating upper layer checksums. 984 These irregularities would result in implementation challenges and 985 the potential for operational issues, e.g., since actionable ICMPv6 986 error reports could not be delivered to the original source. In 987 order to address the issues, still more information such as the 988 original IPv6 source address could be written into the ORH. However, 989 with the additional information the benefit of the "minimal 990 encapsulation" savings quickly diminishes, and becomes overshadowed 991 by the implementation and operational irregularities. 993 3.2.5. Segment Routing Topologies (SRTs) 995 The 64-bit sub-prefixes of [ULA]::/48 identify up to 2^16 distinct 996 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 997 OMNI link overlay instance using a distinct set of ULAs, and emulates 998 a Virtual LAN (VLAN) service for the OMNI link. In some cases (e.g., 999 when redundant topologies are needed for fault tolerance and 1000 reliability) it may be beneficial to deploy multiple SRTs that act as 1001 independent overlay instances. A communication failure in one 1002 instance therefore will not affect communications in other instances. 1004 Each SRT is identified by a distinct value in bits 48-63 of 1005 [ULA]::/48, i.e., as [ULA0]::/64, [ULA1]::/64, [ULA2]::/64, etc. 1006 Each OMNI interface is identified by a unique interface name (e.g., 1007 omni0, omni1, omni2, etc.) and assigns an anycast ADM-ULA 1008 corresponding to its SRT prefix length. The anycast ADM-ULA is used 1009 for OMNI interface determination in Safety-Based Multilink (SBM) as 1010 discussed in [I-D.templin-6man-omni-interface]. Each OMNI interface 1011 further applies Performance-Based Multilink (PBM) internally. 1013 3.2.6. Segment Routing For OMNI Link Selection 1015 An original IPv6 source can direct an IPv6 packet to an AERO node by 1016 including a standard IPv6 Segment Routing Header (SRH) [RFC8754] with 1017 the anycast ADM-ULA for the selected SRT as either the IPv6 1018 destination or as an intermediate hop within the SRH. This allows 1019 the original source to determine the specific OMNI link topology a 1020 packet will traverse when there may be multiple alternatives. 1022 When the AERO node processes the SRH and forwards the packet to the 1023 correct OMNI interface, the OMNI interface writes the next IPv6 1024 address from the SRH into the IPv6 destination address and decrements 1025 Segments Left. If decrementing would cause Segments Left to become 1026 0, the OMNI interface deletes the SRH before forwarding. This form 1027 of Segment Routing supports Safety-Based Multilink (SBM). 1029 3.2.7. Segment Routing Within the OMNI Link 1031 AERO node OMNI interfaces can insert OAL/ORH headers for Segment 1032 Routing within the OMNI link to influence the paths of packets 1033 destined to targets in remote segments without requiring all packets 1034 to traverse strict spanning tree paths. 1036 When an AERO node's OMNI interface has a packet to send to a target 1037 discovered through route optimization located in the same OMNI link 1038 segment, it encapsulates the packet in OAL/ORH headers if necessary 1039 as discussed above. The node then uses the target's Link Layer 1040 Address (L2ADDR) information for INET encapsulation. 1042 When an AERO node's OMNI interface has a packet to send to a route 1043 optimization target located in a remote OMNI link segment, it 1044 encapsulates the packet in OAL/ORH headers as discussed above while 1045 forwarding the packet to a Bridge with destination set to the Subnet 1046 Router Anycast address for the final OMNI link segment. 1048 When a Bridge receives a packet destined to its Subnet Router Anycast 1049 address with an OAL/ORH with SRT/LHS values corresponding to the 1050 local segment, it examines the L2ADDR according to FMT and removes 1051 the ORH from the packet; if the packet is not a fragment, the Bridge 1052 also removes the OAL header. If the packet was a fragment, the 1053 Bridge instead writes the MNP-ULA corresponding to the ORH 1054 Destination Suffix into the OAL destination address. The Bridge then 1055 encapsulates the packet/fragment in an INET header according to 1056 L2ADDR and forwards the packet within the INET either to the LHS 1057 Server/Proxy or directly to the target Client itself. In this way, 1058 the Bridge participates in route optimization to reduce traffic load 1059 and suboptimal routing through strict spanning tree paths. 1061 3.3. OMNI Interface Characteristics 1063 OMNI interfaces are virtual interfaces configured over one or more 1064 underlying interfaces classified as follows: 1066 o INET interfaces connect to an INET either natively or through one 1067 or several IPv4 NATs. Native INET interfaces have global IP 1068 addresses that are reachable from any INET correspondent. All 1069 Server, Relay and Bridge interfaces are native interfaces, as are 1070 INET-facing interfaces of Proxys. NATed INET interfaces connect 1071 to a private network behind one or more NATs that provide INET 1072 access. Clients that are behind a NAT are required to send 1073 periodic keepalive messages to keep NAT state alive when there are 1074 no data packets flowing. 1076 o ANET interfaces connect to an ANET that is separated from the open 1077 INET by a Proxy. Proxys can actively issue control messages over 1078 the INET on behalf of the Client to reduce ANET congestion. 1080 o VPNed interfaces use security encapsulation over the INET to a 1081 Virtual Private Network (VPN) server that also acts as a Server or 1082 Proxy. Other than the link-layer encapsulation format, VPNed 1083 interfaces behave the same as Direct interfaces. 1085 o Direct interfaces connect a Client directly to a Server or Proxy 1086 without crossing any ANET/INET paths. An example is a line-of- 1087 sight link between a remote pilot and an unmanned aircraft. The 1088 same Client considerations apply as for VPNed interfaces. 1090 OMNI interfaces use OAL/ORH encapsulation as necessary as discussed 1091 in Section 3.2.4. OMNI interfaces use link-layer encapsulation (see: 1092 Section 3.6) to exchange packets with OMNI link neighbors over INET 1093 or VPNed interfaces as well as over ANET interfaces for which the 1094 Client and Proxy may be multiple IP hops away. OMNI interfaces do 1095 not use link-layer encapsulation over Direct underlying interfaces or 1096 ANET interfaces when the Client and Proxy are known to be on the same 1097 underlying link. 1099 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 1100 state the same as for any interface. OMNI interfaces use ND messages 1101 including Router Solicitation (RS), Router Advertisement (RA), 1102 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 1103 neighbor cache management. 1105 OMNI interfaces send ND messages with an OMNI option formatted as 1106 specified in [I-D.templin-6man-omni-interface]. The OMNI option 1107 includes prefix registration information and Interface Attributes 1108 containing link information parameters for the OMNI interface's 1109 underlying interfaces. Each OMNI option may include multiple 1110 Interface Attributes sub-options, each identified by an ifIndex 1111 value. 1113 A Client's OMNI interface may be configured over multiple underlying 1114 interface connections. For example, common mobile handheld devices 1115 have both wireless local area network ("WLAN") and cellular wireless 1116 links. These links are often used "one at a time" with low-cost WLAN 1117 preferred and highly-available cellular wireless as a standby, but a 1118 simultaneous-use capability could provide benefits. In a more 1119 complex example, aircraft frequently have many wireless data link 1120 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1121 directional, etc.) with diverse performance and cost properties. 1123 If a Client's multiple underlying interfaces are used "one at a time" 1124 (i.e., all other interfaces are in standby mode while one interface 1125 is active), then ND message OMNI options include only a single 1126 Interface Attributes sub-option set to constant values. In that 1127 case, the Client would appear to have a single interface but with a 1128 dynamically changing link-layer address. 1130 If the Client has multiple active underlying interfaces, then from 1131 the perspective of ND it would appear to have multiple link-layer 1132 addresses. In that case, ND message OMNI options MAY include 1133 multiple Interface Attributes sub-options - each with values that 1134 correspond to a specific interface. Every ND message need not 1135 include Interface Attributes for all underlying interfaces; for any 1136 attributes not included, the neighbor considers the status as 1137 unchanged. 1139 Bridge, Server and Proxy OMNI interfaces may be configured over one 1140 or more secured tunnel interfaces. The OMNI interface configures 1141 both an ADM-LLA and its corresponding ADM-ULA, while the underlying 1142 secured tunnel interfaces are either unnumbered or configure the same 1143 ULA. The OMNI interface encapsulates each IP packet in OAL/ORH 1144 headers and presents the packet to the underlying secured tunnel 1145 interface. Routing protocols such as BGP that run over the OMNI 1146 interface do not employ OAL/ORH encapsulation, but rather present the 1147 routing protocol messages directly to the underlying secured tunnels 1148 while using the ULA as the source address. This distinction must be 1149 honored consistently according to each node's configuration so that 1150 the IP forwarding table will associate discovered IP routes with the 1151 correct interface. 1153 3.4. OMNI Interface Initialization 1155 AERO Servers, Proxys and Clients configure OMNI interfaces as their 1156 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1157 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1158 to ensure that packets with destination addresses covered by an MNP 1159 not explicitly assigned to a non-OMNI interface are directed to the 1160 OMNI interface. 1162 OMNI interface initialization procedures for Servers, Proxys, Clients 1163 and Bridges are discussed in the following sections. 1165 3.4.1. AERO Server/Relay Behavior 1167 When a Server enables an OMNI interface, it assigns an ADM-{LLA,ULA} 1168 appropriate for the given OMNI link segment. The Server also 1169 configures secured tunnels with one or more neighboring Bridges and 1170 engages in a BGP routing protocol session with each Bridge. 1172 The OMNI interface provides a single interface abstraction to the IP 1173 layer, but internally comprises multiple secured tunnels as well as 1174 an NBMA nexus for sending encapsulated data packets to OMNI interface 1175 neighbors. The Server further configures a service to facilitate ND 1176 exchanges with AERO Clients and manages per-Client neighbor cache 1177 entries and IP forwarding table entries based on control message 1178 exchanges. 1180 Relays are simply Servers that run a dynamic routing protocol to 1181 redistribute routes between the OMNI interface and INET/EUN 1182 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1183 networks on the INET/EUN interfaces (i.e., the same as a Client would 1184 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1185 interfaces. The Relay further provides an attachment point of the 1186 OMNI link to a non-MNP-based global topology. 1188 3.4.2. AERO Proxy Behavior 1190 When a Proxy enables an OMNI interface, it assigns an ADM-{LLA, ULA} 1191 and configures permanent neighbor cache entries the same as for 1192 Servers. The Proxy also configures secured tunnels with one or more 1193 neighboring Bridges and maintains per-Client neighbor cache entries 1194 based on control message exchanges. Importantly Proxys are often 1195 configured to act as Servers, and vice-versa. 1197 3.4.3. AERO Client Behavior 1199 When a Client enables an OMNI interface, it assigns either an 1200 MNP-{LLA, ULA} or a Temporary {LLA, ULA} and sends RS messages with 1201 ND parameters over its underlying interfaces to a Server, which 1202 returns an RA message with corresponding parameters. The RS/RA 1203 messages may pass through a Proxy in the case of a Client's ANET 1204 interface, or through one or more NATs in the case of a Client's INET 1205 interface. (Note: if the Client used a Temporary {LLA, ULA} in its 1206 initial RS message, it will discover an MNP-{LLA, ULA} in the 1207 corresponding RA that it receives from the Server and begin using 1208 these new addresses while deprecating the Temporary addresses.) 1210 3.4.4. AERO Bridge Behavior 1212 AERO Bridges configure an OMNI interface and assign the ADM-ULA 1213 Subnet Router Anycast address for each OMNI link segment they connect 1214 to. Bridges configure secured tunnels with Servers, Proxys and other 1215 Bridges, and engage in a BGP routing protocol session with neighbors 1216 on the spanning tree (see: Section 3.2.3). 1218 3.5. OMNI Interface Neighbor Cache Maintenance 1220 Each OMNI interface maintains a conceptual neighbor cache that 1221 includes an entry for each neighbor it communicates with on the OMNI 1222 link per [RFC4861]. OMNI interface neighbor cache entries are said 1223 to be one of "permanent", "symmetric", "asymmetric" or "proxy". 1225 Permanent neighbor cache entries are created through explicit 1226 administrative action; they have no timeout values and remain in 1227 place until explicitly deleted. AERO Bridges maintain permanent 1228 neighbor cache entries for their associated Proxys/Servers (and vice- 1229 versa). Each entry maintains the mapping between the neighbor's 1230 network-layer LLA and corresponding INET address. 1232 Symmetric neighbor cache entries are created and maintained through 1233 RS/RA exchanges as specified in Section 3.12, and remain in place for 1234 durations bounded by prefix lifetimes. AERO Servers maintain 1235 symmetric neighbor cache entries for each of their associated 1236 Clients, and AERO Clients maintain symmetric neighbor cache entries 1237 for each of their associated Servers. 1239 Asymmetric neighbor cache entries are created or updated based on 1240 route optimization messaging as specified in Section 3.14, and are 1241 garbage-collected when keepalive timers expire. AERO ROSs maintain 1242 asymmetric neighbor cache entries for active targets with lifetimes 1243 based on ND messaging constants. Asymmetric neighbor cache entries 1244 are unidirectional since only the ROS (and not the ROR) creates an 1245 entry. 1247 Proxy neighbor cache entries are created and maintained by AERO 1248 Proxys when they process Client/Server ND exchanges, and remain in 1249 place for durations bounded by ND and prefix lifetimes. AERO Proxys 1250 maintain proxy neighbor cache entries for each of their associated 1251 Clients. Proxy neighbor cache entries track the Client state and the 1252 address of the Client's associated Server(s). 1254 To the list of neighbor cache entry states in Section 7.3.2 of 1255 [RFC4861], Proxy and Server OMNI interfaces add an additional state 1256 DEPARTED that applies to symmetric and proxy neighbor cache entries 1257 for Clients that have recently departed. The interface sets a 1258 "DepartTime" variable for the neighbor cache entry to "DEPART_TIME" 1259 seconds. DepartTime is decremented unless a new ND message causes 1260 the state to return to REACHABLE. While a neighbor cache entry is in 1261 the DEPARTED state, packets destined to the target Client are 1262 forwarded to the Client's new location instead of being dropped. 1263 When DepartTime decrements to 0, the neighbor cache entry is deleted. 1264 It is RECOMMENDED that DEPART_TIME be set to the default constant 1265 value REACHABLE_TIME plus 10 seconds (40 seconds by default) to allow 1266 a window for packets in flight to be delivered while stale route 1267 optimization state may be present. 1269 When an ROR receives an authentic NS message used for route 1270 optimization, it searches for a symmetric neighbor cache entry for 1271 the target Client. The ROR then returns a solicited NA message 1272 without creating a neighbor cache entry for the ROS, but creates or 1273 updates a target Client "Report List" entry for the ROS and sets a 1274 "ReportTime" variable for the entry to REPORT_TIME seconds. The ROR 1275 resets ReportTime when it receives a new authentic NS message, and 1276 otherwise decrements ReportTime while no authentic NS messages have 1277 been received. It is RECOMMENDED that REPORT_TIME be set to the 1278 default constant value REACHABLE_TIME plus 10 seconds (40 seconds by 1279 default) to allow a window for route optimization to converge before 1280 ReportTime decrements below REACHABLE_TIME. 1282 When the ROS receives a solicited NA message response to its NS 1283 message used for route optimization, it creates or updates an 1284 asymmetric neighbor cache entry for the target network-layer and 1285 link-layer addresses. The ROS then (re)sets ReachableTime for the 1286 neighbor cache entry to REACHABLE_TIME seconds and uses this value to 1287 determine whether packets can be forwarded directly to the target, 1288 i.e., instead of via a default route. The ROS otherwise decrements 1289 ReachableTime while no further solicited NA messages arrive. It is 1290 RECOMMENDED that REACHABLE_TIME be set to the default constant value 1291 30 seconds as specified in [RFC4861]. 1293 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1294 of NS keepalives sent when a correspondent may have gone unreachable, 1295 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1296 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1297 to limit the number of unsolicited NAs that can be sent based on a 1298 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1299 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1300 same as specified in [RFC4861]. 1302 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1303 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1304 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1305 different values are chosen, all nodes on the link MUST consistently 1306 configure the same values. Most importantly, DEPART_TIME and 1307 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1308 REACHABLE_TIME to avoid packet loss due to stale route optimization 1309 state. 1311 3.5.1. OMNI Neighbor Interface Attributes 1313 OMNI interface IPv6 ND messages include OMNI options 1314 [I-D.templin-6man-omni-interface] with Interface Attributes that 1315 provide Link-Layer Address and QoS Preference information for the 1316 neighbor's underlying interfaces. This information is stored in the 1317 neighbor cache and provides the basis for the forwarding algorithm 1318 specified in Section 3.10. The information is cumulative and 1319 reflects the union of the OMNI information from the most recent ND 1320 messages received from the neighbor; it is therefore not required 1321 that each ND message contain all neighbor information. 1323 The OMNI option Interface Attributes for each underlying interface 1324 includes a two-part "Link-Layer Address" consisting of a simple IP 1325 encapsulation address determined by the FMT and L2ADDR fields and an 1326 ADM-ULA determined by the SRT and LHS fields. If the neighbor is 1327 located in the local OMNI link segment (and, if any necessary NAT 1328 state has been established) forwarding via simple IP encapsulation 1329 can be used; otherwise, OAL encapsulation must be used. Underlying 1330 interfaces are further selected based on their associated preference 1331 values "high", "medium", "low" or "disabled". 1333 Note: the OMNI option is distinct from any Source/Target Link-Layer 1334 Address Options (S/TLLAOs) that may appear in an ND message according 1335 to the appropriate IPv6 over specific link layer specification (e.g., 1336 [RFC2464]). If both an OMNI option and S/TLLAO appear, the former 1337 pertains to encapsulation addresses while the latter pertains to the 1338 native L2 address format of the underlying media. 1340 3.5.2. OMNI Neighbor Advertisement Message Flags 1342 As discussed in Section 4.4 of [RFC4861] NA messages include three 1343 flag bits R, S and O. OMNI interface NA messages treat the flags as 1344 follows: 1346 o R: The R ("Router") flag is set to 1 in the NA messages sent by 1347 all AERO/OMNI node types. Simple hosts that would set R to 0 do 1348 not occur on the OMNI link itself, but may occur on the downstream 1349 links of Clients and Relays. 1351 o S: The S ("Solicited") flag is set exactly as specified in 1352 Section 4.4. of [RFC4861], i.e., it is set to 1 for Solicited NAs 1353 and set to 0 for Unsolicited NAs (both unicast and multicast). 1355 o O: The O ("Override") flag is set to 0 for solicited proxy NAs and 1356 set to 1 for all other solicited and unsolicited NAs. For further 1357 study is whether solicited NAs for anycast targets apply for OMNI 1358 links. Since MNP-LLAs must be uniquely assigned to Clients to 1359 support correct ND protocol operation, however, no role is 1360 currently seen for assigning the same MNP-LLA to multiple Clients. 1362 3.6. OMNI Interface Encapsulation and Re-encapsulation 1364 The OMNI Adaptation Layer (OAL) inserts mid-layer IPv6 headers known 1365 as the OAL/ORH headers when necessary as discussed in the following 1366 sections. After either inserting or omitting the OAL/ORH headers, 1367 the OMNI interface also inserts or omits an outer ANET/INET 1368 encapsulation header as discussed below. 1370 OMNI interfaces avoid outer encapsulation over Direct underlying 1371 interfaces and ANET underlying interfaces for which the Client and 1372 Proxy are connected to the same underlying link. Otherwise, OMNI 1373 interfaces encapsulate packets according to whether they are entering 1374 the OMNI interface from the network layer or if they are being re- 1375 admitted into the same OMNI link they arrived on. This latter form 1376 of encapsulation is known as "re-encapsulation". 1378 For packets entering the OMNI interface from the network layer, the 1379 OMNI interface copies the "TTL/Hop Limit", "Type of Service/Traffic 1380 Class" [RFC2983], "Flow Label"[RFC6438] (for IPv6) and "Congestion 1381 Experienced" [RFC3168] values in the inner packet's IP header into 1382 the corresponding fields in the OAL and outer encapsulation 1383 header(s). 1385 For packets undergoing re-encapsulation, the OMNI interface instead 1386 copies these values from the original encapsulation header into the 1387 new encapsulation header, i.e., the values are transferred between 1388 encapsulation headers and *not* copied from the encapsulated packet's 1389 network-layer header. (Note especially that by copying the TTL/Hop 1390 Limit between encapsulation headers the value will eventually 1391 decrement to 0 if there is a (temporary) routing loop.) 1393 OMNI interfaces configured over ANET underlying interfaces which 1394 employ a different IP protocol version (and/or when the Client and 1395 Proxy may be separated by multiple ANET IP hops) use IP-in-IP 1396 encapsulation so that the inner packet can traverse the ANET without 1397 decrementing the TTL/Hop-Limit. IPv6 underlying ANET interfaces use 1398 [RFC2473] encapsulation, while IPv4 interfaces use the appropriate 1399 encapsulation per one of [RFC5214][RFC2003]. 1401 OMNI interfaces configured over INET underlying interfaces 1402 encapsulate packets in INET headers according to the next hop 1403 determined in the forwarding algorithm in Section 3.10. If the next 1404 hop is reached via a secured tunnel, the OMNI interface uses an 1405 encapsulation format specific to the secured tunnel type (see: 1406 Section 6). If the next hop is reached via an unsecured INET 1407 interface, the OMNI interface instead uses UDP/IP encapsulation per 1408 [RFC4380] and as extended in [RFC6081]. 1410 When UDP/IP encapsulation is used, the OMNI interface next sets the 1411 UDP source port to a constant value that it will use in each 1412 successive packet it sends, and sets the UDP length field to the 1413 length of the encapsulated packet plus 8 bytes for the UDP header 1414 itself plus the length of any included extension headers or trailers. 1415 The encapsulated packet may be either IPv6 or IPv4, as distinguished 1416 by the version number found in the first four bits. 1418 For UDP/IP-encapsulated packets sent to a Server, Relay or Bridge, 1419 the OMNI interface sets the UDP destination port to 8060, i.e., the 1420 IANA-registered port number for AERO. For packets sent to a Client, 1421 the OMNI interface sets the UDP destination port to the port value 1422 stored in the neighbor cache entry for this Client. The OMNI 1423 interface finally includes/omits the UDP checksum according to 1424 [RFC6935][RFC6936]. 1426 When a Proxy, Relay or Server re-encapsulates a packet received from 1427 a Client that includes an OAL but no ORH, it inserts an ORH 1428 immediately following the OAL header and adjusts the OAL payload 1429 length and destination address field. The inserted ORH will be 1430 removed by the final-hop Bridge, but its insertion and removal will 1431 not interfere with reassembly at the final destination. 1433 3.7. OMNI Interface Decapsulation 1435 OMNI interfaces decapsulate packets destined either to the AERO node 1436 itself or to a destination reached via an interface other than the 1437 OMNI interface the packet was received on. When the encapsulated 1438 packet arrives in multiple OAL fragments, the OMNI interface 1439 reassembles as discussed in Section 3.9. Further decapsulation steps 1440 are performed according to the appropriate encapsulation format 1441 specification. 1443 3.8. OMNI Interface Data Origin Authentication 1445 AERO nodes employ simple data origin authentication procedures. In 1446 particular: 1448 o AERO Bridges, Servers and Proxys accept encapsulated data packets 1449 and control messages received from the (secured) spanning tree. 1451 o AERO Proxys and Clients accept packets that originate from within 1452 the same secured ANET. 1454 o AERO Clients and Relays accept packets from downstream network 1455 correspondents based on ingress filtering. 1457 o AERO Clients, Relays and Servers verify the outer UDP/IP 1458 encapsulation addresses according to [RFC4380]. 1460 AERO nodes silently drop any packets that do not satisfy the above 1461 data origin authentication procedures. Further security 1462 considerations are discussed in Section 6. 1464 3.9. OMNI Adaptation Layer and OMNI Interface MTU 1466 The OMNI interface observes the link nature of tunnels, including the 1467 Maximum Transmission Unit (MTU), Maximum Reassembly Unit (MRU) and 1468 the role of fragmentation and reassembly [I-D.ietf-intarea-tunnels]. 1469 The OMNI interface employs an OMNI Adaptation Layer (OAL) for 1470 accommodating multiple underlying links with diverse MTUs. The 1471 functions of the OAL and the OMNI interface MTU/MRU are specified in 1472 Section 5 of [I-D.templin-6man-omni-interface], with MTU/MRU both set 1473 to the constant value 9180 bytes. 1475 When the network layer presents an IP packet to the OMNI interface, 1476 the OAL encapsulates the packet in OAL/ORH headers. The OAL then 1477 fragments the encapsulated packet if necessary such that the OAL/ORH 1478 headers appear in each fragment while the original IP packet header 1479 appears only in the first fragment. The OAL then transmits each OAL/ 1480 ORH packet/fragment over an underlying interface connected to either 1481 a physical link such as Ethernet, WiFi and the like or a virtual link 1482 such as an Internet or higher-layer tunnel (see the definition of 1483 link in [RFC8200]). 1485 Note: Although the ORH may be removed by a Bridge on the path (see: 1486 Section 3.10.4), this does not interfere with the destination's 1487 ability to reassemble in the event that the packet was fragmented. 1488 This is due to the fact that the ORH is not included in the 1489 fragmentable part; therefore, its removal does not invalidate the 1490 offset values in any fragment headers. 1492 3.10. OMNI Interface Forwarding Algorithm 1494 IP packets enter a node's OMNI interface either from the network 1495 layer (i.e., from a local application or the IP forwarding system) or 1496 from the link layer (i.e., from an OMNI interface neighbor). All 1497 packets entering a node's OMNI interface first undergo data origin 1498 authentication as discussed in Section 3.8. Packets that satisfy 1499 data origin authentication are processed further, while all others 1500 are dropped silently. The OMNI interface OAL wraps accepted packets 1501 in OAL/ORH headers if necessary as discussed above. 1503 Packets that enter the OMNI interface from the network layer are 1504 forwarded to an OMNI interface neighbor. Packets that enter the OMNI 1505 interface from the link layer are either re-admitted into the OMNI 1506 link or forwarded to the network layer where they are subject to 1507 either local delivery or IP forwarding. In all cases, the OMNI 1508 interface itself MUST NOT decrement the network layer TTL/Hop-count 1509 since its forwarding actions occur below the network layer. 1511 OMNI interfaces may have multiple underlying interfaces and/or 1512 neighbor cache entries for neighbors with multiple underlying 1513 interfaces (see Section 3.3). The OMNI interface uses interface 1514 attributes and/or traffic classifiers (e.g., DSCP value, port number, 1515 flow specification, etc.) to select an outgoing underlying interface 1516 for each packet based on the node's own QoS preferences, and also to 1517 select a destination link-layer address based on the neighbor's 1518 underlying interface with the highest preference. AERO 1519 implementations SHOULD allow for QoS preference values to be modified 1520 at runtime through network management. 1522 If multiple outgoing interfaces and/or neighbor interfaces have a 1523 preference of "high", the AERO node replicates the packet and sends 1524 one copy via each of the (outgoing / neighbor) interface pairs; 1525 otherwise, the node sends a single copy of the packet via an 1526 interface with the highest preference. AERO nodes keep track of 1527 which underlying interfaces are currently "reachable" or 1528 "unreachable", and only use "reachable" interfaces for forwarding 1529 purposes. 1531 The following sections discuss the OMNI interface forwarding 1532 algorithms for Clients, Proxys, Servers and Bridges. In the 1533 following discussion, a packet's destination address is said to 1534 "match" if it is the same as a cached address, or if it is covered by 1535 a cached prefix (which may be encoded in an MNP-LLA). 1537 3.10.1. Client Forwarding Algorithm 1539 When an IP packet enters a Client's OMNI interface from the network 1540 layer the Client searches for an asymmetric neighbor cache entry that 1541 matches the destination. If there is a match, the Client uses one or 1542 more "reachable" neighbor interfaces in the entry for packet 1543 forwarding. If there is no asymmetric neighbor cache entry, the 1544 Client instead forwards the packet toward a Server (the packet is 1545 intercepted by a Proxy if there is a Proxy on the path). The Client 1546 encapsulates the packet in OAL/ORH headers if necessary and fragments 1547 according to MTU requirements (see: Section 3.9). If the Client has 1548 multiple IP packets to send to the same destination, it can 1549 concatenate them in a single super-packet as discussed in Section 5.2 1550 of [I-D.templin-6man-omni-interface]. 1552 If there is an asymmetric neighbor cache entry and the neighbor 1553 interface selected for forwarding is located on the same OMNI link 1554 segment and not behind a NAT, the Client forwards the packet/ 1555 fragments directly according to the L2ADDR information for the 1556 neighbor. If the neighbor interface is behind a NAT on the same OMNI 1557 link segment, the Client instead forwards the initial packet/ 1558 fragments to its Server and initiates NAT traversal procedures. If 1559 the Client's intended source underlying interface is also behind a 1560 NAT and located on the same OMNI link segment, it sends a "direct 1561 bubble" over the interface per [RFC6081][RFC4380] to the L2ADDR found 1562 in the neighbor cache in order to establish state in its own NAT by 1563 generating traffic toward the neighbor (note that no response to the 1564 bubble is expected). 1566 The Client next sends an NS(NUD) message toward the MNP-ULA of the 1567 neighbor via its Server as discussed in Section 3.15. If the Client 1568 receives an NA(NUD) from the neighbor over the underlying interface, 1569 it marks the neighbor interface as "trusted" and sends future 1570 packets/fragments directly to the L2ADDR information for the neighbor 1571 instead of indirectly via the Server. The Client must honor the 1572 neighbor cache maintenance procedure by sending additional direct 1573 bubbles and/or NS/NA(NUD) messages as discussed in [RFC6081][RFC4380] 1574 in order to keep NAT state alive as long as packets are still 1575 flowing. 1577 When an IP packet enters a Client's OMNI interface from the link- 1578 layer, if the destination matches one of the Client's MNPs or link- 1579 local addresses the Client reassembles and decapsulates as necessary 1580 and delivers the inner packet to the network layer. Otherwise, the 1581 Client drops the packet and MAY return a network-layer ICMP 1582 Destination Unreachable message subject to rate limiting (see: 1583 Section 3.11). 1585 3.10.2. Proxy Forwarding Algorithm 1587 For control messages originating from or destined to a Client, the 1588 Proxy intercepts the message and updates its proxy neighbor cache 1589 entry for the Client. The Proxy then forwards a (proxyed) copy of 1590 the control message. (For example, the Proxy forwards a proxied 1591 version of a Client's NS/RS message to the target neighbor, and 1592 forwards a proxied version of the neighbor's NA/RA reply to the 1593 Client.) 1595 When the Proxy receives a data packet from a Client within the ANET, 1596 the Proxy reassembles and re-fragments if necessary then searches for 1597 an asymmetric neighbor cache entry that matches the destination and 1598 forwards as follows: 1600 o if the destination matches an asymmetric neighbor cache entry, the 1601 Proxy uses one or more "reachable" neighbor interfaces in the 1602 entry for packet forwarding using OAL/ORH encapsulation if 1603 necessary according to the cached link-layer address information. 1604 If the neighbor interface is in a different OMNI link segment, the 1605 Proxy forwards the packet to a Bridge; otherwise, it forwards the 1606 packet directly to the neighbor. If the neighbor is behind a NAT, 1607 the Proxy instead forwards initial packets via a Bridge while 1608 sending an NS(NUD) to the neighbor. When the Proxy receives the 1609 NA(NUD), it can begin forwarding packets directly to the neighbor 1610 the same as discussed in Section 3.10.1 while sending additional 1611 NUD messages as necessary to maintain NAT state. Note that no 1612 direct bubbles are necessary since the Proxy is by definition not 1613 located behind a NAT. 1615 o else, the Proxy uses OAL/ORH encapsulation and forwards the packet 1616 to a Bridge while using the MNP-ULA corresponding to the packet's 1617 destination as the destination address. 1619 When the Proxy receives an encapsulated data packet from an INET 1620 neighbor or from a secured tunnel from a Bridge, it accepts the 1621 packet only if data origin authentication succeeds and if there is a 1622 proxy neighbor cache entry that matches the inner destination. Next, 1623 the Proxy reassembles the packet (if necessary) and continues 1624 processing. If the reassembly is complete and the neighbor cache 1625 state is REACHABLE, the Proxy then returns a PTB if necessary (see: 1626 Section 3.9) then either drops or forwards the packet to the Client 1627 while performing OAL/ORH re-encapsulation and re-fragmentation if 1628 necessary. If the neighbor cache entry state is DEPARTED, the Proxy 1629 instead changes the destination address to the address of the new 1630 Server and forwards it to a Bridge while performing OAL/ORH re- 1631 encapsulation if necessary. 1633 Note: When the Proxy receives fragmented OAL packets with destination 1634 address set to the MNP-ULA of the Client, it can either reassemble 1635 first and then re-encapsulate/re-fragment before forwarding to the 1636 Client or forward the raw fragments on to the Client which then must 1637 reassemble. In the former case, the Proxy can re-fragment to a size 1638 that better matches the link MTU for the Client, which may be 1639 important for low-end links with large MTUs. In the latter case, the 1640 Client may receive fragments that are smaller than its link MTU but 1641 can still be reassembled; this case may provide an important 1642 performance benefit to Proxys by permitting them to avoid excessive 1643 reassembly and re-fragmentation overhead.) 1645 Note: If the Proxy has multiple IP packets to send to the same 1646 destination, it can concatenate them in a single super-packet as 1647 discussed in Section 5.2 of [I-D.templin-6man-omni-interface]. 1649 3.10.3. Server/Relay Forwarding Algorithm 1651 For control messages destined to a target Client's MNP-LLA that are 1652 received from a secured tunnel, the Server intercepts the message and 1653 returns a Proxyed response on behalf of the Client. (For example, 1654 the Server sends a Proxyed NA message reply in response to an NS 1655 message directed to one of its associated Clients.) If the Client's 1656 neighbor cache entry is in the DEPARTED state, however, the Server 1657 instead forwards the packet to the Client's new Server using OAL/ORH 1658 re-encapsulation the same as discussed for Proxys in Section 3.10.2. 1660 When the Server receives an encapsulated data packet from an INET 1661 neighbor or from a secured tunnel, it accepts the packet only if data 1662 origin authentication succeeds. If the packet was an OAL fragment 1663 with destination set to the Server's ADM-ULA, the Server submits the 1664 packet for reassembly (if the destination was the MNP-ULA of one of 1665 its Clients, the Server instead either submits the packet for 1666 reassembly or forwards the fragment directly to the target Client). 1667 If reassembly is complete (or if no reassembly was required), the 1668 Server then (re)fragments if necessary and forwards as follows: 1670 o if the network layer destination matches a symmetric neighbor 1671 cache entry in the REACHABLE state the Server prepares the packet 1672 for forwarding to the destination Client. The Server then 1673 forwards the packet as specified in Section 3.9. 1675 o else, if the destination matches a symmetric neighbor cache entry 1676 in the DEPARETED state the Server re-encapsulates the packet and 1677 forwards it using the ADM-ULA of the Client's new Server as the 1678 destination. 1680 o else, if the destination matches an asymmetric neighbor cache 1681 entry, the Server proceeds the same as specified for a Proxy in 1682 Section 3.10.2. Again, the Server need not send direct bubbles 1683 since it is not located behind a NAT. 1685 o else, if the destination matches a non-MNP route in the IP 1686 forwarding table or an ADM-LLA assigned to the Server's OMNI 1687 interface, the Server decapsulates the packet and releases it to 1688 the network layer for local delivery or IP forwarding. 1690 o else, the Server drops the packet. 1692 When the Server's OMNI interface receives a data packet from the 1693 network layer or from a VPNed or Direct Client, it performs OAL/ORH 1694 encapsulation and fragmentation if necessary, then processes the 1695 packet according to the network-layer destination address as follows: 1697 o if the destination matches a symmetric or asymmetric neighbor 1698 cache entry the Server processes the packet as above. 1700 o else, the Server encapsulates the packet in OLA/ORH headers and 1701 forwards it to a Bridge using its own ADM-ULA as the source and 1702 the MNP-ULA corresponding to the destination as the destination. 1704 Note: The same as for Proxys, when the Server receives fragmented OAL 1705 packets with destination address set to the MNP-ULA of the target, it 1706 can either reassemble first and then re-encapsulate/re-fragment 1707 before sending or forward the raw fragments on to the target which 1708 then must reassemble. In the former case, the Server can re-fragment 1709 to a size that better matches the link MTU for the Client, which may 1710 be important for low-end links with large MTUs. In the latter case, 1711 the target may receive fragments that are smaller than its link MTU 1712 but can still be reassembled; this case may provide an important 1713 performance benefit to Servers by permitting them to avoid excessive 1714 reassembly and re-fragmentation overhead.) 1716 Note: The same as for Proxys, if the Server has multiple IP packets 1717 to send to the same destination, it can concatenate them in a single 1718 super-packet as discussed in Section 5.2 of 1719 [I-D.templin-6man-omni-interface]. 1721 3.10.4. Bridge Forwarding Algorithm 1723 Bridges forward OAL/ORH-encapsulated packets over secured tunnels the 1724 same as any IP router. When the Bridge receives an OAL/ORH- 1725 encapsulated packet via a secured tunnel, it removes the outer INET 1726 header and searches for a forwarding table entry that matches the OAL 1727 destination address. The Bridge then processes the packet as 1728 follows: 1730 o if the destination matches its ADM-ULA Subnet Router Anycast 1731 address the Bridge processes the packet locally before forwarding. 1732 The Bridge drops the packet if the next header is not an ORH; 1733 otherwise, for NA(NUD) messages the Bridge replaces the OMNI 1734 option Interface Attributes sub-option with information for its 1735 own interface while retaining the ifIndex value supplied by the 1736 NA(NUD) message source. For all packet types, the Bridge next 1737 examines the ORH FMT code. If the code indicates the destination 1738 is a Client on the open INET (or, a Client behind a NAT for which 1739 NAT traversal procedures have already converged) the Bridge 1740 removes the ORH then writes the MNP-ULA formed from the ORH 1741 Destination Suffix into the OAL destination if the packet is a 1742 fragment (otherwise, the Bridge also removes the OAL header). The 1743 Bridge then forwards the packet/fragment via INET encapsulation to 1744 the ORH L2ADDR. For all other destination cases, the Bridge 1745 instead writes the ADM-ULA formed from the ORH SRT/LHS into the 1746 OAL destination address and forwards the OAL/ORH encapsulated 1747 packet to the ADM-ULA Server while invoking NAT traversal 1748 procedures the same as for Proxys and Servers if necessary, noting 1749 that no direct bubbles are necessary since only the Client and not 1750 the Bridge is behind a NAT. 1752 o else, if the destination matches one of the Bridge's own 1753 addresses, the Bridge submits the packet for local delivery. 1755 o else, if the destination matches a forwarding table entry the 1756 Bridge forwards the packet via a secured tunnel to the next hop. 1757 If the destination matches an MSP without matching an MNP, 1758 however, the Bridge instead drops the packet and returns an ICMP 1759 Destination Unreachable message subject to rate limiting (see: 1760 Section 3.11). 1762 o else, the Bridge drops the packet and returns an ICMP Destination 1763 Unreachable as above. 1765 As for any IP router, the Bridge decrements the TTL/Hop Limit when it 1766 forwards the packet. Therefore, when an OAL header is present only 1767 the Hop Limit in the OAL header is decremented and not the TTL/Hop 1768 Limit in the inner packet header. Bridges do not insert OAL/ORH 1769 headers themselves; instead, they act as IPv6 routers and forward 1770 packets based on the destination address found in the headers of 1771 packets they receive. 1773 3.11. OMNI Interface Error Handling 1775 When an AERO node admits a packet into the OMNI interface, it may 1776 receive link-layer or network-layer error indications. 1778 A link-layer error indication is an ICMP error message generated by a 1779 router in the INET on the path to the neighbor or by the neighbor 1780 itself. The message includes an IP header with the address of the 1781 node that generated the error as the source address and with the 1782 link-layer address of the AERO node as the destination address. 1784 The IP header is followed by an ICMP header that includes an error 1785 Type, Code and Checksum. Valid type values include "Destination 1786 Unreachable", "Time Exceeded" and "Parameter Problem" 1787 [RFC0792][RFC4443]. (OMNI interfaces ignore all link-layer IPv4 1788 "Fragmentation Needed" and IPv6 "Packet Too Big" messages since they 1789 only emit packets that are guaranteed to be no larger than the IP 1790 minimum link MTU as discussed in Section 3.9.) 1792 The ICMP header is followed by the leading portion of the packet that 1793 generated the error, also known as the "packet-in-error". For 1794 ICMPv6, [RFC4443] specifies that the packet-in-error includes: "As 1795 much of invoking packet as possible without the ICMPv6 packet 1796 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1797 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1798 "Internet Header + 64 bits of Original Data Datagram", however 1799 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1800 ICMP datagram SHOULD contain as much of the original datagram as 1801 possible without the length of the ICMP datagram exceeding 576 1802 bytes". 1804 The link-layer error message format is shown in Figure 5 (where, "L2" 1805 and "L3" refer to link-layer and network-layer, respectively): 1807 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1808 ~ ~ 1809 | L2 IP Header of | 1810 | error message | 1811 ~ ~ 1812 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1813 | L2 ICMP Header | 1814 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1815 ~ ~ P 1816 | IP and other encapsulation | a 1817 | headers of original L3 packet | c 1818 ~ ~ k 1819 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1820 ~ ~ t 1821 | IP header of | 1822 | original L3 packet | i 1823 ~ ~ n 1824 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1825 ~ ~ e 1826 | Upper layer headers and | r 1827 | leading portion of body | r 1828 | of the original L3 packet | o 1829 ~ ~ r 1830 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1832 Figure 5: OMNI Interface Link-Layer Error Message Format 1834 The AERO node rules for processing these link-layer error messages 1835 are as follows: 1837 o When an AERO node receives a link-layer Parameter Problem message, 1838 it processes the message the same as described as for ordinary 1839 ICMP errors in the normative references [RFC0792][RFC4443]. 1841 o When an AERO node receives persistent link-layer Time Exceeded 1842 messages, the IP ID field may be wrapping before earlier fragments 1843 awaiting reassembly have been processed. In that case, the node 1844 should begin including integrity checks and/or institute rate 1845 limits for subsequent packets. 1847 o When an AERO node receives persistent link-layer Destination 1848 Unreachable messages in response to encapsulated packets that it 1849 sends to one of its asymmetric neighbor correspondents, the node 1850 should process the message as an indication that a path may be 1851 failing, and optionally initiate NUD over that path. If it 1852 receives Destination Unreachable messages over multiple paths, the 1853 node should allow future packets destined to the correspondent to 1854 flow through a default route and re-initiate route optimization. 1856 o When an AERO Client receives persistent link-layer Destination 1857 Unreachable messages in response to encapsulated packets that it 1858 sends to one of its symmetric neighbor Servers, the Client should 1859 mark the path as unusable and use another path. If it receives 1860 Destination Unreachable messages on many or all paths, the Client 1861 should associate with a new Server and release its association 1862 with the old Server as specified in Section 3.16.5. 1864 o When an AERO Server receives persistent link-layer Destination 1865 Unreachable messages in response to encapsulated packets that it 1866 sends to one of its symmetric neighbor Clients, the Server should 1867 mark the underlying path as unusable and use another underlying 1868 path. 1870 o When an AERO Server or Proxy receives link-layer Destination 1871 Unreachable messages in response to an encapsulated packet that it 1872 sends to one of its permanent neighbors, it treats the messages as 1873 an indication that the path to the neighbor may be failing. 1874 However, the dynamic routing protocol should soon reconverge and 1875 correct the temporary outage. 1877 When an AERO Bridge receives a packet for which the network-layer 1878 destination address is covered by an MSP, if there is no more- 1879 specific routing information for the destination the Bridge drops the 1880 packet and returns a network-layer Destination Unreachable message 1881 subject to rate limiting. The Bridge writes the network-layer source 1882 address of the original packet as the destination address and uses 1883 one of its non link-local addresses as the source address of the 1884 message. 1886 When an AERO node receives an encapsulated packet for which the 1887 reassembly buffer it too small, it drops the packet and returns a 1888 network-layer Packet Too Big (PTB) message. The node first writes 1889 the MRU value into the PTB message MTU field, writes the network- 1890 layer source address of the original packet as the destination 1891 address and writes one of its non link-local addresses as the source 1892 address. 1894 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1896 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1897 coordinated as discussed in the following Sections. 1899 3.12.1. AERO Service Model 1901 Each AERO Server on the OMNI link is configured to facilitate Client 1902 prefix delegation/registration requests. Each Server is provisioned 1903 with a database of MNP-to-Client ID mappings for all Clients enrolled 1904 in the AERO service, as well as any information necessary to 1905 authenticate each Client. The Client database is maintained by a 1906 central administrative authority for the OMNI link and securely 1907 distributed to all Servers, e.g., via the Lightweight Directory 1908 Access Protocol (LDAP) [RFC4511], via static configuration, etc. 1909 Clients receive the same service regardless of the Servers they 1910 select. 1912 AERO Clients and Servers use ND messages to maintain neighbor cache 1913 entries. AERO Servers configure their OMNI interfaces as advertising 1914 NBMA interfaces, and therefore send unicast RA messages with a short 1915 Router Lifetime value (e.g., ReachableTime seconds) in response to a 1916 Client's RS message. Thereafter, Clients send additional RS messages 1917 to keep Server state alive. 1919 AERO Clients and Servers include prefix delegation and/or 1920 registration parameters in RS/RA messages (see 1921 [I-D.templin-6man-omni-interface]). The ND messages are exchanged 1922 between Client and Server according to the prefix management schedule 1923 required by the service. If the Client knows its MNP in advance, it 1924 can employ prefix registration by including its MNP-LLA as the source 1925 address of an RS message and with an OMNI option with valid prefix 1926 registration information for the MNP. If the Server (and Proxy) 1927 accept the Client's MNP assertion, they inject the prefix into the 1928 routing system and establish the necessary neighbor cache state. 1930 The following sections specify the Client and Server behavior. 1932 3.12.2. AERO Client Behavior 1934 AERO Clients discover the addresses of Servers in a similar manner as 1935 described in [RFC5214]. Discovery methods include static 1936 configuration (e.g., from a flat-file map of Server addresses and 1937 locations), or through an automated means such as Domain Name System 1938 (DNS) name resolution [RFC1035]. Alternatively, the Client can 1939 discover Server addresses through a layer 2 data link login exchange, 1940 or through a unicast RA response to a multicast/anycast RS as 1941 described below. In the absence of other information, the Client can 1942 resolve the DNS Fully-Qualified Domain Name (FQDN) 1943 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1944 text string and "[domainname]" is a DNS suffix for the OMNI link 1945 (e.g., "example.com"). 1947 To associate with a Server, the Client acts as a requesting router to 1948 request MNPs. The Client prepares an RS message with prefix 1949 management parameters and includes a Nonce and Timestamp option if 1950 the Client needs to correlate RA replies. If the Client already 1951 knows the Server's ADM-LLA, it includes the LLA as the network-layer 1952 destination address; otherwise, it includes (link-local) All-Routers 1953 multicast as the network-layer destination. If the Client already 1954 knows its own MNP-LLA, it uses the LLA as the network-layer source 1955 address; otherwise, it uses a Temporary LLA as the network-layer 1956 source address and includes a DHCP Unique Identifier (DUID) sub- 1957 option in the OMNI option (see: [I-D.templin-6man-omni-interface]). 1959 The Client next includes an OMNI option in the RS message to register 1960 its link-layer information with the Server. The Client sets the OMNI 1961 option prefix registration information according to the MNP, and 1962 includes Interface Attributes corresponding to the underlying 1963 interface over which the Client will send the RS message. The Client 1964 MAY include additional Interface Attributes specific to other 1965 underlying interfaces. 1967 The Client then sends the RS message (either directly via Direct 1968 interfaces, via a VPN for VPNed interfaces, via a Proxy for ANET 1969 interfaces or via INET encapsulation for INET interfaces) and waits 1970 for an RA message reply (see Section 3.12.3). The Client retries up 1971 to MAX_RTR_SOLICITATIONS times until an RA is received. If the 1972 Client receives no RAs, or if it receives an RA with Router Lifetime 1973 set to 0, the Client SHOULD abandon this Server and try another 1974 Server. Otherwise, the Client processes the prefix information found 1975 in the RA message. 1977 Next, the Client creates a symmetric neighbor cache entry with the 1978 Server's ADM-LLA as the network-layer address and the Server's 1979 encapsulation and/or link-layer addresses as the link-layer address. 1980 The Client records the RA Router Lifetime field value in the neighbor 1981 cache entry as the time for which the Server has committed to 1982 maintaining the MNP in the routing system via this underlying 1983 interface, and caches the other RA configuration information 1984 including Cur Hop Limit, M and O flags, Reachable Time and Retrans 1985 Timer. The Client then autoconfigures MNP-LLAs for each of the 1986 delegated MNPs and assigns them to the OMNI interface. The Client 1987 also caches any MSPs included in Route Information Options (RIOs) 1988 [RFC4191] as MSPs to associate with the OMNI link, and assigns the 1989 MTU value in the MTU option to the underlying interface. 1991 The Client then registers additional underlying interfaces with the 1992 Server by sending RS messages via each additional interface. The RS 1993 messages include the same parameters as for the initial RS/RA 1994 exchange, but with destination address set to the Server's ADM-LLA. 1996 Following autoconfiguration, the Client sub-delegates the MNPs to its 1997 attached EUNs and/or the Client's own internal virtual interfaces as 1998 described in [I-D.templin-v6ops-pdhost] to support the Client's 1999 downstream attached "Internet of Things (IoT)". The Client 2000 subsequently sends additional RS messages over each underlying 2001 interface before the Router Lifetime received for that interface 2002 expires. 2004 After the Client registers its underlying interfaces, it may wish to 2005 change one or more registrations, e.g., if an interface changes 2006 address or becomes unavailable, if QoS preferences change, etc. To 2007 do so, the Client prepares an RS message to send over any available 2008 underlying interface. The RS includes an OMNI option with prefix 2009 registration information specific to its MNP, with Interface 2010 Attributes specific to the selected underlying interface, and with 2011 any additional Interface Attributes specific to other underlying 2012 interfaces. When the Client receives the Server's RA response, it 2013 has assurance that the Server has been updated with the new 2014 information. 2016 If the Client wishes to discontinue use of a Server it issues an RS 2017 message over any underlying interface with an OMNI option with a 2018 prefix release indication. When the Server processes the message, it 2019 releases the MNP, sets the symmetric neighbor cache entry state for 2020 the Client to DEPARTED and returns an RA reply with Router Lifetime 2021 set to 0. After a short delay (e.g., 2 seconds), the Server 2022 withdraws the MNP from the routing system. 2024 3.12.3. AERO Server Behavior 2026 AERO Servers act as IP routers and support a prefix delegation/ 2027 registration service for Clients. Servers arrange to add their ADM- 2028 LLAs to a static map of Server addresses for the link and/or the DNS 2029 resource records for the FQDN "linkupnetworks.[domainname]" before 2030 entering service. Server addresses should be geographically and/or 2031 topologically referenced, and made available for discovery by Clients 2032 on the OMNI link. 2034 When a Server receives a prospective Client's RS message on its OMNI 2035 interface, it SHOULD return an immediate RA reply with Router 2036 Lifetime set to 0 if it is currently too busy or otherwise unable to 2037 service the Client. Otherwise, the Server authenticates the RS 2038 message and processes the prefix delegation/registration parameters. 2039 The Server first determines the correct MNPs to provide to the Client 2040 by searching the Client database. When the Server returns the MNPs, 2041 it also creates a forwarding table entry for the MNP-ULA 2042 corresponding to each MNP so that the MNPs are propagated into the 2043 routing system (see: Section 3.2.3). For IPv6, the Server creates an 2044 IPv6 forwarding table entry for each MNP. For IPv4, the Server 2045 creates an IPv6 forwarding table entry with the IPv4-compatibility 2046 MNP-ULA prefix corresponding to the IPv4 address. 2048 The Server next creates a symmetric neighbor cache entry for the 2049 Client using the base MNP-LLA as the network-layer address and with 2050 lifetime set to no more than the smallest prefix lifetime. Next, the 2051 Server updates the neighbor cache entry by recording the information 2052 in each Interface Attributes sub-option in the RS OMNI option. The 2053 Server also records the actual OAL/INET addresses in the neighbor 2054 cache entry. 2056 Next, the Server prepares an RA message using its ADM-LLA as the 2057 network-layer source address and the network-layer source address of 2058 the RS message as the network-layer destination address. The Server 2059 sets the Router Lifetime to the time for which it will maintain both 2060 this underlying interface individually and the symmetric neighbor 2061 cache entry as a whole. The Server also sets Cur Hop Limit, M and O 2062 flags, Reachable Time and Retrans Timer to values appropriate for the 2063 OMNI link. The Server includes the MNPs, any other prefix management 2064 parameters and an OMNI option with no Interface Attributes. The 2065 Server then includes one or more RIOs that encode the MSPs for the 2066 OMNI link, plus an MTU option (see Section 3.9). The Server finally 2067 forwards the message to the Client using OAL/INET, INET, or NULL 2068 encapsulation as necessary. 2070 After the initial RS/RA exchange, the Server maintains a 2071 ReachableTime timer for each of the Client's underlying interfaces 2072 individually (and for the Client's symmetric neighbor cache entry 2073 collectively) set to expire after ReachableTime seconds. If the 2074 Client (or Proxy) issues additional RS messages, the Server sends an 2075 RA response and resets ReachableTime. If the Server receives an ND 2076 message with a prefix release indication it sets the Client's 2077 symmetric neighbor cache entry to the DEPARTED state and withdraws 2078 the MNP from the routing system after a short delay (e.g., 2 2079 seconds). If ReachableTime expires before a new RS is received on an 2080 individual underlying interface, the Server marks the interface as 2081 DOWN. If ReachableTime expires before any new RS is received on any 2082 individual underlying interface, the Server sets the symmetric 2083 neighbor cache entry state to STALE and sets a 10 second timer. If 2084 the Server has not received a new RS or ND message with a prefix 2085 release indication before the 10 second timer expires, it deletes the 2086 neighbor cache entry and withdraws the MNP from the routing system. 2088 The Server processes any ND messages pertaining to the Client and 2089 returns an NA/RA reply in response to solicitations. The Server may 2090 also issue unsolicited RA messages, e.g., with reconfigure parameters 2091 to cause the Client to renegotiate its prefix delegation/ 2092 registrations, with Router Lifetime set to 0 if it can no longer 2093 service this Client, etc. Finally, If the symmetric neighbor cache 2094 entry is in the DEPARTED state, the Server deletes the entry after 2095 DepartTime expires. 2097 Note: Clients SHOULD notify former Servers of their departures, but 2098 Servers are responsible for expiring neighbor cache entries and 2099 withdrawing routes even if no departure notification is received 2100 (e.g., if the Client leaves the network unexpectedly). Servers 2101 SHOULD therefore set Router Lifetime to ReachableTime seconds in 2102 solicited RA messages to minimize persistent stale cache information 2103 in the absence of Client departure notifications. A short Router 2104 Lifetime also ensures that proactive Client/Server RS/RA messaging 2105 will keep any NAT state alive (see above). 2107 Note: All Servers on an OMNI link MUST advertise consistent values in 2108 the RA Cur Hop Limit, M and O flags, Reachable Time and Retrans Timer 2109 fields the same as for any link, since unpredictable behavior could 2110 result if different Servers on the same link advertised different 2111 values. 2113 3.12.3.1. DHCPv6-Based Prefix Registration 2115 When a Client is not pre-provisioned with an MNP-LLA, it will need 2116 for the Server to select one or more MNPs on its behalf and set up 2117 the correct state in the AERO routing service. (A Client with a pre- 2118 provisioned MNP may also request the Server to select additional 2119 MNPs.) The DHCPv6 service [RFC8415] is used to support this 2120 requirement. 2122 When a Client needs to have the Server select MNPs, it sends an RS 2123 message with a Temporary LLA and with an OMNI option that includes a 2124 DHCPv6 message sub-option with DHCPv6 Prefix Delegation (DHCPv6-PD) 2125 parameters. When the Server receives the RS message, it extracts the 2126 DHCPv6-PD message from the OMNI option. 2128 The Server then acts as a "Proxy DHCPv6 Client" in a message exchange 2129 with the locally-resident DHCPv6 server, which delegates MNPs and 2130 returns a DHCPv6-PD Reply message. (If the Server wishes to defer 2131 creation of MN state until the DHCPv6-PD Reply is received, it can 2132 instead act as a Lightweight DHCPv6 Relay Agent per [RFC6221] by 2133 encapsulating the DHCPv6-PD message in a Relay-forward/reply exchange 2134 with Relay Message and Interface ID options.) 2136 When the Server receives the DHCPv6-PD Reply, it adds a route to the 2137 routing system and creates an MNP-LLA based on the delegated MNP. 2138 The Server then sends an RA back to the Client with the (newly- 2139 created) MNP-LLA as the destination address and with the DHCPv6-PD 2140 Reply message coded in the OMNI option. When the Client receives the 2141 RA, it creates a default route, assigns the Subnet Router Anycast 2142 address and sets its MNP-LLA based on the delegated MNP. 2144 3.13. The AERO Proxy 2146 Clients may connect to protected-spectrum ANETs that employ physical 2147 and/or link-layer security services to facilitate communications to 2148 Servers in outside INETs. In that case, the ANET can employ an AERO 2149 Proxy. The Proxy is located at the ANET/INET border and listens for 2150 RS messages originating from or RA messages destined to ANET Clients. 2151 The Proxy acts on these control messages as follows: 2153 o when the Proxy receives an RS message from a new ANET Client, it 2154 first authenticates the message then examines the network-layer 2155 destination address. If the destination address is a Server's 2156 ADM-LLA, the Proxy proceeds to the next step. Otherwise, if the 2157 destination is (link-local) All-Routers multicast, the Proxy 2158 selects a "nearby" Server that is likely to be a good candidate to 2159 serve the Client and replaces the destination address with the 2160 Server's ADM-LLA. Next, the Proxy creates a proxy neighbor cache 2161 entry and caches the Client and Server link-layer addresses along 2162 with the OMNI option information and any other identifying 2163 information including Transaction IDs, Client Identifiers, Nonce 2164 values, etc. The Proxy finally encapsulates the (proxyed) RS 2165 message in an OAL header with source set to the Proxy's ADM-ULA 2166 and destination set to the Server's ADM-ULA. The Proxy also 2167 includes an OMNI header with an Interface Attributes option that 2168 includes its own INET address plus a unique Port Number for this 2169 Client, then forwards the message into the OMNI link spanning 2170 tree. (Note: including a unique Port Number allows the Server to 2171 distinguish different Clients located behind the same Proxy at the 2172 link-layer, whereas the link-layer addresses would otherwise be 2173 indistinguishable) 2175 o when the Server receives the RS, it authenticates the message then 2176 creates or updates a symmetric neighbor cache entry for the Client 2177 with the Proxy's ADM-ULA, INET address and Port Number as the 2178 link-layer address information. The Server then sends an RA 2179 message back to the Proxy via the spanning tree. 2181 o when the Proxy receives the RA, it authenticates the message and 2182 matches it with the proxy neighbor cache entry created by the RS. 2183 The Proxy then caches the prefix information as a mapping from the 2184 Client's MNPs to the Client's link-layer address, caches the 2185 Server's advertised Router Lifetime and sets the neighbor cache 2186 entry state to REACHABLE. The Proxy then optionally rewrites the 2187 Router Lifetime and forwards the (proxyed) message to the Client. 2188 The Proxy finally includes an MTU option (if necessary) with an 2189 MTU to use for the underlying ANET interface. 2191 After the initial RS/RA exchange, the Proxy forwards any Client data 2192 packets for which there is no matching asymmetric neighbor cache 2193 entry to a Bridge using OAL encapsulation with its own ADM-ULA as the 2194 source and the MNP-ULA corresponding to the Client as the 2195 destination. The Proxy instead forwards any Client data destined to 2196 an asymmetric neighbor cache target directly to the target according 2197 to the OAL/link-layer information - the process of establishing 2198 asymmetric neighbor cache entries is specified in Section 3.14. 2200 While the Client is still attached to the ANET, the Proxy sends NS, 2201 RS and/or unsolicited NA messages to update the Server's symmetric 2202 neighbor cache entries on behalf of the Client and/or to convey QoS 2203 updates. This allows for higher-frequency Proxy-initiated RS/RA 2204 messaging over well-connected INET infrastructure supplemented by 2205 lower-frequency Client-initiated RS/RA messaging over constrained 2206 ANET data links. 2208 If the Server ceases to send solicited advertisements, the Proxy 2209 sends unsolicited RAs on the ANET interface with destination set to 2210 (link-local) All-Nodes multicast and with Router Lifetime set to zero 2211 to inform Clients that the Server has failed. Although the Proxy 2212 engages in ND exchanges on behalf of the Client, the Client can also 2213 send ND messages on its own behalf, e.g., if it is in a better 2214 position than the Proxy to convey QoS changes, etc. For this reason, 2215 the Proxy marks any Client-originated solicitation messages (e.g. by 2216 inserting a Nonce option) so that it can return the solicited 2217 advertisement to the Client instead of processing it locally. 2219 If the Client becomes unreachable, the Proxy sets the neighbor cache 2220 entry state to DEPARTED and retains the entry for DepartTime seconds. 2221 While the state is DEPARTED, the Proxy forwards any packets destined 2222 to the Client to a Bridge via OAL/ORH encapsulation. When DepartTime 2223 expires, the Proxy deletes the neighbor cache entry and discards any 2224 further packets destined to this (now forgotten) Client. 2226 In some ANETs that employ a Proxy, the Client's MNP can be injected 2227 into the ANET routing system. In that case, the Client can send data 2228 messages without encapsulation so that the ANET routing system 2229 transports the unencapsulated packets to the Proxy. This can be very 2230 beneficial, e.g., if the Client connects to the ANET via low-end data 2231 links such as some aviation wireless links. 2233 If the first-hop ANET access router is on the same underlying link 2234 and recognizes the AERO/OMNI protocol, the Client can avoid 2235 encapsulation for both its control and data messages. When the 2236 Client connects to the link, it can send an unencapsulated RS message 2237 with source address set to its own MNP-LLA (or to a Temporary LLA), 2238 and with destination address set to the ADM-LLA of the Client's 2239 selected Server or to (link-local) All-Routers multicast. The Client 2240 includes an OMNI option formatted as specified in 2241 [I-D.templin-6man-omni-interface]. 2243 The Client then sends the unencapsulated RS message, which will be 2244 intercepted by the AERO-Aware access router. The access router then 2245 encapsulates the RS message in an ANET header with its own address as 2246 the source address and the address of a Proxy as the destination 2247 address. The access router further remembers the address of the 2248 Proxy so that it can encapsulate future data packets from the Client 2249 via the same Proxy. If the access router needs to change to a new 2250 Proxy, it simply sends another RS message toward the Server via the 2251 new Proxy on behalf of the Client. 2253 In some cases, the access router and Proxy may be one and the same 2254 node. In that case, the node would be located on the same physical 2255 link as the Client, but its message exchanges with the Server would 2256 need to pass through a security gateway at the ANET/INET border. The 2257 method for deploying access routers and Proxys (i.e. as a single node 2258 or multiple nodes) is an ANET-local administrative consideration. 2260 Note: The Proxy can apply packing as discussed in Section 5.2 of 2261 [I-D.templin-6man-omni-interface] if an opportunity arises to 2262 concatenate multiple payload packets that will be destined to the 2263 same destination. 2265 3.13.1. Combined Proxy/Servers 2267 Clients may need to connect directly to Servers via INET, Direct and 2268 VPNed interfaces (i.e., non-ANET interfaces). If the Client's 2269 underlying interfaces all connect via the same INET partition, then 2270 it can connect to a single controlling Server via all interfaces. 2272 If some Client interfaces connect via different INET partitions, 2273 however, the Client still selects a set of controlling Servers and 2274 sends RS messages via their directly-connected Servers while using 2275 the ADM-LLA of the controlling Server as the destination. 2277 When a Server receives an RS with destination set to the ADM-LLA of a 2278 controlling Server, it acts as a Proxy to forward the message to the 2279 controlling Server while forwarding the corresponding RA reply to the 2280 Client. 2282 3.13.2. Detecting and Responding to Server Failures 2284 In environments where fast recovery from Server failure is required, 2285 Proxys SHOULD use proactive Neighbor Unreachability Detection (NUD) 2286 to track Server reachability in a similar fashion as for 2287 Bidirectional Forwarding Detection (BFD) [RFC5880]. Proxys can then 2288 quickly detect and react to failures so that cached information is 2289 re-established through alternate paths. The NUD control messaging is 2290 carried only over well-connected ground domain networks (i.e., and 2291 not low-end aeronautical radio links) and can therefore be tuned for 2292 rapid response. 2294 Proxys perform proactive NUD with Servers for which there are 2295 currently active ANET Clients by sending continuous NS messages in 2296 rapid succession, e.g., one message per second. The Proxy sends the 2297 NS message via the spanning tree with the Proxy's ADM-LLA as the 2298 source and the ADM-LLA of the Server as the destination. When the 2299 Proxy is also sending RS messages to the Server on behalf of ANET 2300 Clients, the resulting RA responses can be considered as equivalent 2301 hints of forward progress. This means that the Proxy need not also 2302 send a periodic NS if it has already sent an RS within the same 2303 period. If the Server fails (i.e., if the Proxy ceases to receive 2304 advertisements), the Proxy can quickly inform Clients by sending 2305 multicast RA messages on the ANET interface. 2307 The Proxy sends RA messages on the ANET interface with source address 2308 set to the Server's address, destination address set to (link-local) 2309 All-Nodes multicast, and Router Lifetime set to 0. The Proxy SHOULD 2310 send MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small 2311 delays [RFC4861]. Any Clients on the ANET that had been using the 2312 failed Server will receive the RA messages and associate with a new 2313 Server. 2315 3.13.3. Point-to-Multipoint Server Coordination 2317 In environments where Client messaging over ANETs is bandwidth- 2318 limited and/or expensive, Clients can enlist the services of the 2319 Proxy to coordinate with multiple Servers in a single RS/RA message 2320 exchange. The Client can send a single RS message to (link-local) 2321 All-Routers multicast that includes the ID's of multiple Servers in 2322 MS-Register sub-options of the OMNI option. 2324 When the Proxy receives the RS and processes the OMNI option, it 2325 sends a separate RS to each MS-Register Server ID. When the Proxy 2326 receives an RA, it can optionally return an immediate "singleton" RA 2327 to the Client or record the Server's ID for inclusion in a pending 2328 "aggregate" RA message. The Proxy can then return aggregate RA 2329 messages to the Client including multiple Server IDs in order to 2330 conserve bandwidth. Each RA includes a proper subset of the Server 2331 IDs from the original RS message, and the Proxy must ensure that the 2332 message contents of each RA are consistent with the information 2333 received from the (aggregated) Servers. 2335 Clients can thereafter employ efficient point-to-multipoint Server 2336 coordination under the assistance of the Proxy to reduce the number 2337 of messages sent over the ANET while enlisting the support of 2338 multiple Servers for fault tolerance. Clients can further include 2339 MS-Release sub-options in IPv6 ND messages to request the Proxy to 2340 release from former Servers via the procedures discussed in 2341 Section 3.16.5. 2343 The OMNI interface specification [I-D.templin-6man-omni-interface] 2344 provides further discussion of the Client/Proxy RS/RA messaging 2345 involved in point-to-multipoint coordination. 2347 3.14. AERO Route Optimization / Address Resolution 2349 While data packets are flowing between a source and target node, 2350 route optimization SHOULD be used. Route optimization is initiated 2351 by the first eligible Route Optimization Source (ROS) closest to the 2352 source as follows: 2354 o For Clients on VPNed and Direct interfaces, the Server is the ROS. 2356 o For Clients on ANET interfaces, the Proxy is the ROS. 2358 o For Clients on INET interfaces, the Client itself is the ROS. 2360 o For correspondent nodes on INET/EUN interfaces serviced by a 2361 Relay, the Relay is the ROS. 2363 The route optimization procedure is conducted between the ROS and the 2364 target Server/Relay acting as a Route Optimization Responder (ROR) in 2365 the same manner as for IPv6 ND Address Resolution and using the same 2366 NS/NA messaging. The target may either be a MNP Client serviced by a 2367 Server, or a non-MNP correspondent reachable via a Relay. 2369 The procedures are specified in the following sections. 2371 3.14.1. Route Optimization Initiation 2373 While data packets are flowing from the source node toward a target 2374 node, the ROS performs address resolution by sending an NS message 2375 for Address Resolution (NS(AR)) to receive a solicited NA(AR) message 2376 from the ROR. When the ROS sends an NS(AR), it includes: 2378 o the LLA of the ROS as the source address. 2380 o the data packet's destination as the Target Address. 2382 o the Solicited-Node multicast address [RFC4291] formed from the 2383 lower 24 bits of the data packet's destination as the destination 2384 address, e.g., for 2001:db8:1:2::10:2000 the NS destination 2385 address is ff02:0:0:0:0:1:ff10:2000. 2387 The NS(AR) message includes an OMNI option with no Interface 2388 Attributes, such that the target will not create a neighbor cache 2389 entry. The Prefix Length in the OMNI option is set to the Prefix 2390 Length associated with the ROS's LLA. 2392 The ROS then encapsulates the NS(AR) message in an OAL header with 2393 source set to its own ULA and destination set to the MNP-ULA 2394 corresponding to the target, then sends the message into the spanning 2395 tree without decrementing the network-layer TTL/Hop Limit field. 2396 (When the ROS is a Client, it instead securely sends the NS(AR) to 2397 one of its current Servers as specified in Section 3.22. The Server 2398 then forwards an adapted version of the NS(AR) into the spanning tree 2399 on behalf of the Client.) 2401 Note: The ROS can apply packing as discussed in Section 5.2 of 2402 [I-D.templin-6man-omni-interface] to concatenate the NS(AR) onto the 2403 same OAL packet used to carry a data packet that triggered route 2404 optimization. 2406 3.14.2. Relaying the NS(AR) 2408 When the Bridge receives the NS(AR) message from the ROS, it discards 2409 the INET header and determines that the ROR is the next hop by 2410 consulting its standard IPv6 forwarding table for the OAL header 2411 destination address. The Bridge then forwards the message toward the 2412 ROR via the spanning tree the same as for any IPv6 router. The 2413 final-hop Bridge in the spanning tree will deliver the message via a 2414 secured tunnel to the ROR. 2416 3.14.3. Processing the NS(AR) and Sending the NA(AR) 2418 When the ROR receives the NS(AR) message, it examines the Target 2419 Address to determine whether it has a neighbor cache entry and/or 2420 route that matches the target. If there is no match, the ROR drops 2421 the message. Otherwise, the ROR continues processing as follows: 2423 o if the target belongs to an MNP Client neighbor in the DEPARTED 2424 state the ROR changes the NS(AR) message OAL destination address 2425 to the ADM-ULA of the Client's new Server, forwards the message 2426 into the spanning tree and returns from processing. 2428 o If the target belongs to an MNP Client neighbor in the REACHABLE 2429 state, the ROR instead adds the AERO source address to the target 2430 Client's Report List with time set to ReportTime. 2432 o If the target belongs to a non-MNP route, the ROR continues 2433 processing without adding an entry to the Report List. 2435 The ROR then prepares a (solicited) NA(AR) message to send back to 2436 the ROS but does not create a neighbor cache entry. The ROR sets the 2437 NA(AR) source address to the MNP-LLA corresponding to the target, 2438 sets the Target Address to the target of the solicitation, and sets 2439 the destination address to the source of the solicitation. The ROR 2440 then includes an OMNI option with Prefix Length set to the length 2441 associated with the MNP-LLA. 2443 If the target is an MNP Client, the ROR next includes Interface 2444 Attributes in the OMNI option for each of the target Client's 2445 underlying interfaces with current information for each interface and 2446 with the S/T-ifIndex field in the OMNI header set to 0 to indicate 2447 that the message originated from the ROR and not the Client. 2449 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2450 according to its own INET address for VPNed or Direct interfaces, to 2451 the INET address of the Proxy or to the Client's INET address for 2452 INET interfaces. The ROR then includes the lower 32 bits of its own 2453 ADM-ULA (or the ADM-ULA of the Proxy) as the LHS, encodes the ADM-ULA 2454 prefix length code in the SRT field and sets the FMT code accordingly 2455 as specified in Section 3.3. 2457 The ROR then sets the NA(AR) message R flag to 1 (as a router), S 2458 flag to 1 (as a response to a solicitation), and O flag to 0 (as a 2459 proxy). The ROR finally encapsulates the NA(AR) message in an OAL 2460 header with source set to its own ULA and destination set to the 2461 source ULA of the NS(AR) message, then forwards the message into the 2462 spanning tree without decrementing the network-layer TTL/Hop Limit 2463 field. 2465 3.14.4. Relaying the NA(AR) 2467 When the Bridge receives the NA(AR) message from the ROR, it discards 2468 the INET header and determines that the ROS is the next hop by 2469 consulting its standard IPv6 forwarding table for the OAL header 2470 destination address. The Bridge then forwards the OAL-encapsulated 2471 NA(AR) message toward the ROS the same as for any IPv6 router. The 2472 final-hop Bridge in the spanning tree will deliver the message via a 2473 secured tunnel to the ROS. 2475 3.14.5. Processing the NA(AR) 2477 When the ROS receives the NA(AR) message, it processes the message 2478 the same as for standard IPv6 Address Resolution [RFC4861]. In the 2479 process, it caches the source MNP-ULA then creates an asymmetric 2480 neighbor cache entry for the target and caches all information found 2481 in the OMNI option. The ROS finally sets the asymmetric neighbor 2482 cache entry lifetime to ReachableTime seconds. (When the ROS is a 2483 Client, the solicited NA(AR) message will first be delivered via the 2484 spanning tree to one of its current Servers, which then securely 2485 forwards the message to the Client as discussed in Section 3.22.) 2487 3.14.6. Route Optimization Maintenance 2489 Following route optimization, the ROS forwards future data packets 2490 destined to the target via the addresses found in the cached link- 2491 layer information. The route optimization is shared by all sources 2492 that send packets to the target via the ROS, i.e., and not just the 2493 source on behalf of which the route optimization was initiated. 2495 While new data packets destined to the target are flowing through the 2496 ROS, it sends additional NS(AR) messages to the ROR before 2497 ReachableTime expires to receive a fresh NA(AR) message the same as 2498 described in the previous sections (route optimization refreshment 2499 strategies are an implementation matter, with a non-normative example 2500 given in Appendix A.1). The ROS uses the cached ADM-ULA of the ROR 2501 as the NS(AR) OAL destination address (i.e., instead of using the ULA 2502 corresponding to the target as was the case for the initial NS(AR)), 2503 and sends up to MAX_MULTICAST_SOLICIT NS(AR) messages separated by 1 2504 second until an NA(AR) is received. If no NA(AR) is received, the 2505 ROS assumes that the current ROR has become unreachable and deletes 2506 the target neighbor cache entry. Subsequent data packets will 2507 trigger a new route optimization with an NS with OAL destination 2508 address set to the MNP-ULA corresponding to the target per 2509 Section 3.14.1 to discover a new ROR while initial data packets 2510 travel over a suboptimal route. 2512 If an NA(AR) is received, the ROS then updates the asymmetric 2513 neighbor cache entry to refresh ReachableTime, while (for MNP 2514 destinations) the ROR adds or updates the ROS address to the target's 2515 Report List and with time set to ReportTime. While no data packets 2516 are flowing, the ROS instead allows ReachableTime for the asymmetric 2517 neighbor cache entry to expire. When ReachableTime expires, the ROS 2518 deletes the asymmetric neighbor cache entry. Any future data packets 2519 flowing through the ROS will again trigger a new route optimization. 2521 The ROS may also receive unsolicited NA messages from the ROR at any 2522 time (see: Section 3.16). If there is an asymmetric neighbor cache 2523 entry for the target, the ROS updates the link-layer information but 2524 does not update ReachableTime since the receipt of an unsolicited NA 2525 does not confirm that any forward paths are working. If there is no 2526 asymmetric neighbor cache entry, the ROS simply discards the 2527 unsolicited NA. 2529 In this arrangement, the ROS holds an asymmetric neighbor cache entry 2530 for the target via the ROR, but the ROR does not hold an asymmetric 2531 neighbor cache entry for the ROS. The route optimization neighbor 2532 relationship is therefore asymmetric and unidirectional. If the 2533 target node also has packets to send back to the source node, then a 2534 separate route optimization procedure is performed in the reverse 2535 direction. But, there is no requirement that the forward and reverse 2536 paths be symmetric. 2538 3.15. Neighbor Unreachability Detection (NUD) 2540 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2541 [RFC4861] either reactively in response to persistent link-layer 2542 errors (see Section 3.11) or proactively to confirm reachability and/ 2543 or establish NAT state. The NUD algorithm is based on periodic 2544 control message exchanges. The algorithm may further be seeded by ND 2545 hints of forward progress, but care must be taken to avoid inferring 2546 reachability based on spoofed information. For example, authentic 2547 IPv6 ND message exchanges may be considered as acceptable hints of 2548 forward progress, while spurious data packets should not be. 2550 AERO nodes can use (OAL/ORH-encapsulated) standard NS/NA exchanges 2551 sent over the OMNI link spanning tree to securely test reachability 2552 without risk of DoS attacks from nodes pretending to be a neighbor 2553 (these NS/NA(NUD) messages use the unicast LLAs and ULAs of the two 2554 parties involved in the NUD test the same as for standard IPv6 ND, 2555 and both messages flow over the spanning tree). Proxys can further 2556 perform NUD to securely verify Server reachability on behalf of their 2557 proxyed Clients. However, a means for an ROS to test the unsecured 2558 target route optimized paths is also necessary. 2560 When an ROR directs an ROS to a target neighbor with one or more 2561 link-layer addresses, the ROS can proactively test each such 2562 unsecured route optimized path through secured NS(NUD) messages over 2563 the spanning tree that invoke an unsecured NA(NUD) reply that travels 2564 over the route optimized path.. (The NS(NUD) messages must therefore 2565 include Nonce and Timestamp options that will be echoed in the 2566 unsecured NA(NUD) replies.) While testing the paths, the ROS can 2567 optionally continue to send packets via the spanning tree, maintain a 2568 small queue of packets until target reachability is confirmed, or 2569 (optimistically) allow packets to flow via the route optimized paths. 2571 When the ROS sends an NS(NUD) message, it sets the IPv6 source to its 2572 own address, sets the destination to the MNP-LLA of the target, and 2573 sets the target's MNP Subnet-Router anycast address as the Target 2574 Address. The ROS also includes an OMNI option with a single 2575 Interface Attributes sub-option with the L2ADDR information for its 2576 own underlying interface it wishes to test, but sets the S/T-ifIndex 2577 field to the index for target's underlying interface to be tested. 2578 The ROS includes a Nonce and Timestamp option, then encapsulates the 2579 message in OAL/INET headers with its own ULA as the source and the 2580 ULA of the target as the destination. The ROS then forwards the 2581 NS(NUD) message toward the target via a Server or Bridge. 2583 When the target receives the NS(NUD) message, it creates an NA(NUD) 2584 by reversing the OAL and IPv6 addresses and including an Interface 2585 Attributes sub-option with attributes for its own interface 2586 identified by the NS(NUD) S/T-ifIndex. The target sets the NA(NUD) 2587 S/T-ifIndex to the index of the ROS, sets the Target Address to the 2588 same value that was in the NS(NUD), and returns the message using its 2589 own underlying interface identified by S/T-ifIndex and destined to 2590 the ROS's interface identified by the original Interface Attributes 2591 sub-option. 2593 When the ROS receives the NA(NUD) message, it can determine from the 2594 Nonce, Timestamp and Target Address that the message matched its 2595 NS(NUD) and that it transited the direct path from the ROR using the 2596 selected underlying interface pair. The ROS marks route optimization 2597 target paths that pass these NUD tests as "reachable", and those that 2598 do not as "unreachable". These markings inform the OMNI interface 2599 forwarding algorithm specified in Section 3.10. 2601 Note: If the target determines that the OMNI option Interface 2602 Attributes in the NS(NUD) is located in a different OMNI link segment 2603 than its own interface named in the S/T-ifIndex, it instead returns 2604 the NA(NUD) via the spanning tree while including an ORH and setting 2605 the OAL destination address to the Subnet Router Anycast address used 2606 by Bridges on the ROS segment. When a Bridge on the ROS segment 2607 receives the NA(NUD), it replaces the Interface Attributes with 2608 information for its own interface while using the ifIndex value 2609 specific to the target. 2611 3.16. Mobility Management and Quality of Service (QoS) 2613 AERO is a Distributed Mobility Management (DMM) service. Each Server 2614 is responsible for only a subset of the Clients on the OMNI link, as 2615 opposed to a Centralized Mobility Management (CMM) service where 2616 there is a single network mobility collective entity for all Clients. 2617 Clients coordinate with their associated Servers via RS/RA exchanges 2618 to maintain the DMM profile, and the AERO routing system tracks all 2619 current Client/Server peering relationships. 2621 Servers provide default routing and mobility/multilink services for 2622 their dependent Clients. Clients are responsible for maintaining 2623 neighbor relationships with their Servers through periodic RS/RA 2624 exchanges, which also serves to confirm neighbor reachability. When 2625 a Client's underlying interface address and/or QoS information 2626 changes, the Client is responsible for updating the Server with this 2627 new information. Note that when there is a Proxy in the path, the 2628 Proxy can also perform some RS/RA exchanges on the Client's behalf. 2630 Mobility management messaging is based on the transmission and 2631 reception of unsolicited Neighbor Advertisement (uNA) messages. Each 2632 uNA message sets the IPv6 destination address to (link-local) All- 2633 Nodes multicast to convey a general update of Interface Attributes to 2634 (possibly) multiple recipients, or to a specific unicast LLA to 2635 announce a departure event to a specific recipient. Implementations 2636 must therefore examine the destination address to determine the 2637 nature of the mobility event (i.e., update vs departure). 2639 Mobility management considerations are specified in the following 2640 sections. 2642 3.16.1. Mobility Update Messaging 2644 Servers accommodate Client mobility, multilink and/or QoS change 2645 events by sending unsolicited NA (uNA) messages to each ROS in the 2646 target Client's Report List. When a Server sends a uNA message, it 2647 sets the IPv6 source address to the Client's MNP-LLA, sets the 2648 destination address to (link-local) All-Nodes multicast and sets the 2649 Target Address to the Client's Subnet-Router anycast address. The 2650 Server also includes an OMNI option with Prefix Length set to the 2651 length associated with the Client's MNP-LLA, with Interface 2652 Attributes for the target Client's underlying interfaces and with the 2653 OMNI header S/T-ifIndex set to 0. The Server then sets the NA R flag 2654 to 1, the S flag to 0 and the O flag to 1, then encapsulates the 2655 message in an OAL header with source set to its own ADM-ULA and 2656 destination set to the ULA of the ROS and sends the message into the 2657 spanning tree. 2659 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2660 reception of uNA messages is unreliable but provides a useful 2661 optimization. In well-connected Internetworks with robust data links 2662 uNA messages will be delivered with high probability, but in any case 2663 the Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT uNAs 2664 to each ROS to increase the likelihood that at least one will be 2665 received. 2667 When the ROS receives a uNA message prepared as above, it ignores the 2668 message if there is no existing neighbor cache entry for the Client. 2669 Otherwise, it uses the included OMNI option information to update the 2670 neighbor cache entry, but does not reset ReachableTime since the 2671 receipt of an unsolicited NA message from the target Server does not 2672 provide confirmation that any forward paths to the target Client are 2673 working. 2675 If uNA messages are lost, the ROS may be left with stale address and/ 2676 or QoS information for the Client for up to ReachableTime seconds. 2677 During this time, the ROS can continue sending packets according to 2678 its stale neighbor cache information. When ReachableTime is close to 2679 expiring, the ROS will re-initiate route optimization and receive 2680 fresh link-layer address information. 2682 In addition to sending uNA messages to the current set of ROSs for 2683 the Client, the Server also sends uNAs to the MNP-ULA associated with 2684 the link-layer address for any underlying interface for which the 2685 link-layer address has changed. These uNA messages update an old 2686 Proxy/Server that cannot easily detect (e.g., without active probing) 2687 when a formerly-active Client has departed. When the Server sends 2688 the uNA, it sets the IPv6 source address to the Client's MNP-LLA, 2689 sets the destination address to the old Proxy/Server's ADM-LLA, and 2690 sets the Target Address to the Client's Subnet-Router anycast 2691 address. The Server also includes an OMNI option with Prefix Length 2692 set to the length associated with the Client's MNP-LLA, with 2693 Interface Attributes for the changed underlying interface, and with 2694 the OMNI header S/T-ifIndex set to 0. The Server then sets the NA R 2695 flag to 1, the S flag to 0 and the O flag to 1, then encapsulates the 2696 message in an OAL header with source set to its own ADM-ULA and 2697 destination set to the ADM-ULA of the old Proxy/Server and sends the 2698 message into the spanning tree. 2700 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2702 When a Client needs to change its underlying interface addresses and/ 2703 or QoS preferences (e.g., due to a mobility event), either the Client 2704 or its Proxys send RS messages to the Server via the spanning tree 2705 with an OMNI option that includes Interface attributes with the new 2706 link quality and address information. 2708 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2709 sending actual data packets in case one or more RAs are lost. If all 2710 RAs are lost, the Client SHOULD re-associate with a new Server. 2712 When the Server receives the Client's changes, it sends uNA messages 2713 to all nodes in the Report List the same as described in the previous 2714 section. 2716 3.16.3. Bringing New Links Into Service 2718 When a Client needs to bring new underlying interfaces into service 2719 (e.g., when it activates a new data link), it sends an RS message to 2720 the Server via the underlying interface with an OMNI option that 2721 includes Interface Attributes with appropriate link quality values 2722 and with link-layer address information for the new link. 2724 3.16.4. Deactivating Existing Links 2726 When a Client needs to deactivate an existing underlying interface, 2727 it sends an RS or uNA message to its Server with an OMNI option with 2728 appropriate Interface Attribute values - in particular, the link 2729 quality value 0 assures that neighbors will cease to use the link. 2731 If the Client needs to send RS/uNA messages over an underlying 2732 interface other than the one being deactivated, it MUST include 2733 Interface Attributes with appropriate link quality values for any 2734 underlying interfaces being deactivated. 2736 Note that when a Client deactivates an underlying interface, 2737 neighbors that have received the RS/uNA messages need not purge all 2738 references for the underlying interface from their neighbor cache 2739 entries. The Client may reactivate or reuse the underlying interface 2740 and/or its ifIndex at a later point in time, when it will send RS/uNA 2741 messages with fresh Interface Attributes to update any neighbors. 2743 3.16.5. Moving Between Servers 2745 The Client performs the procedures specified in Section 3.12.2 when 2746 it first associates with a new Server or renews its association with 2747 an existing Server. The Client also includes MS-Release identifiers 2748 in the RS message OMNI option per [I-D.templin-6man-omni-interface] 2749 if it wants the new Server to notify any old Servers from which the 2750 Client is departing. 2752 When the new Server receives the Client's RS message, it returns an 2753 RA as specified in Section 3.12.3 and sends up to 2754 MAX_NEIGHBOR_ADVERTIISEMENT uNA messages to any old Servers listed in 2755 OMNI option MS-Release identifiers. When the new Server sends a uNA 2756 message, it sets the IPv6 source address to the Client's MNP-LLA, 2757 sets the destination address to the old Server's ADM-LLA, and sets 2758 the Target Address to the Client's Subnet-Router anycast address. 2759 The new Server also includes an OMNI option with Prefix Length set to 2760 the length associated with the Client's MNP-LLA, with Interface 2761 Attributes for its own underlying interface, and with the OMNI header 2762 S/T-ifIndex set to 0. The new Server then sets the NA R flag to 1, 2763 the S flag to 0 and the O flag to 1, then encapsulates the message in 2764 an OAL header with source set to its own ADM-ULA and destination set 2765 to the ADM-ULA of the old Server and sends the message into the 2766 spanning tree. 2768 When an old Server receives the uNA, it changes the Client's neighbor 2769 cache entry state to DEPARTED, sets the link-layer address of the 2770 Client to the new Server's ADM-ULA, and resets DepartTime. After a 2771 short delay (e.g., 2 seconds) the old Server withdraws the Client's 2772 MNP from the routing system. After DepartTime expires, the old 2773 Server deletes the Client's neighbor cache entry. 2775 The old Server also iteratively forwards a copy of the uNA message to 2776 each ROS in the Client's Report List by changing the OAL destination 2777 address to the ULA of the ROS while leaving all other fields of the 2778 message unmodified. When the ROS receives the uNA, it examines the 2779 Target address to determine the correct asymmetric neighbor cache 2780 entry and verifies that the IPv6 destination address matches the old 2781 Server. The ROS then caches the IPv6 source address as the new 2782 Server for the existing asymmetric neighbor cache entry and marks the 2783 entry as STALE. While in the STALE state, the ROS allows new data 2784 packets to flow according to any existing cached link-layer 2785 information and sends new NS(AR) messages using its own ULA as the 2786 OAL source and the ADM-ULA of the new Server as the OAL destination 2787 address to elicit NA messages that reset the asymmetric neighbor 2788 cache entry state to REACHABLE. If no new NA message is received for 2789 10 seconds while in the STALE state, the ROS deletes the neighbor 2790 cache entry. 2792 Clients SHOULD NOT move rapidly between Servers in order to avoid 2793 causing excessive oscillations in the AERO routing system. Examples 2794 of when a Client might wish to change to a different Server include a 2795 Server that has gone unreachable, topological movements of 2796 significant distance, movement to a new geographic region, movement 2797 to a new OMNI link segment, etc. 2799 When a Client moves to a new Server, some of the fragments of a 2800 multiple fragment packet may have already arrived at the old Server 2801 while others are en route to the new Server, however no special 2802 attention in the reassembly algorithm is necessary when re-routed 2803 fragments are simply treated as loss. 2805 3.17. Multicast 2807 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2808 [RFC3810] proxy service for its EUNs and/or hosted applications 2809 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2810 underlying interfaces for which group membership is required. The 2811 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2812 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2813 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2814 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2815 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 2816 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2817 INET/EUN networks. The behaviors identified in the following 2818 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2819 Multicast (ASM) operational modes. 2821 3.17.1. Source-Specific Multicast (SSM) 2823 When an ROS (i.e., an AERO Proxy/Server/Relay) "X" acting as PIM 2824 router receives a Join/Prune message from a node on its downstream 2825 interfaces containing one or more ((S)ource, (G)roup) pairs, it 2826 updates its Multicast Routing Information Base (MRIB) accordingly. 2827 For each S belonging to a prefix reachable via X's non-OMNI 2828 interfaces, X then forwards the (S, G) Join/Prune to any PIM routers 2829 on those interfaces per [RFC7761]. 2831 For each S belonging to a prefix reachable via X's OMNI interface, X 2832 originates a separate copy of the Join/Prune for each (S,G) in the 2833 message using its own LLA as the source address and ALL-PIM-ROUTERS 2834 as the destination address. X then encapsulates each message in an 2835 OAL header with source address set to the ULA of X and destination 2836 address set to S then forwards the message into the spanning tree, 2837 which delivers it to AERO Server/Relay "Y" that services S. At the 2838 same time, if the message was a Join, X sends a route-optimization NS 2839 message toward each S the same as discussed in Section 3.14. The 2840 resulting NAs will return the LLA for the prefix that matches S as 2841 the network-layer source address and with an OMNI option with the ULA 2842 corresponding to any underlying interfaces that are currently 2843 servicing S. 2845 When Y processes the Join/Prune message, if S located behind any 2846 INET, Direct, or VPNed interfaces Y acts as a PIM router and updates 2847 its MRIB to list X as the next hop in the reverse path. If S is 2848 located behind any Proxys "Z"*, Y also forwards the message to each 2849 Z* over the spanning tree while continuing to use the LLA of X as the 2850 source address. Each Z* then updates its MRIB accordingly and 2851 maintains the LLA of X as the next hop in the reverse path. Since 2852 the Bridges do not examine network layer control messages, this means 2853 that the (reverse) multicast tree path is simply from each Z* (and/or 2854 Y) to X with no other multicast-aware routers in the path. If any Z* 2855 (and/or Y) is located on the same OMNI link segment as X, the 2856 multicast data traffic sent to X directly using OAL/INET 2857 encapsulation instead of via a Bridge. 2859 Following the initial Join/Prune and NS/NA messaging, X maintains an 2860 asymmetric neighbor cache entry for each S the same as if X was 2861 sending unicast data traffic to S. In particular, X performs 2862 additional NS/NA exchanges to keep the neighbor cache entry alive for 2863 up to t_periodic seconds [RFC7761]. If no new Joins are received 2864 within t_periodic seconds, X allows the neighbor cache entry to 2865 expire. Finally, if X receives any additional Join/Prune messages 2866 for (S,G) it forwards the messages to each Y and Z* in the neighbor 2867 cache entry over the spanning tree. 2869 At some later time, Client C that holds an MNP for source S may 2870 depart from a first Proxy Z1 and/or connect via a new Proxy Z2. In 2871 that case, Y sends an unsolicited NA message to X the same as 2872 specified for unicast mobility in Section 3.16. When X receives the 2873 unsolicited NA message, it updates its asymmetric neighbor cache 2874 entry for the LLA for source S and sends new Join messages to any new 2875 Proxys Z2. There is no requirement to send any Prune messages to old 2876 Proxys Z1 since source S will no longer source any multicast data 2877 traffic via Z1. Instead, the multicast state for (S,G) in Proxy Z1 2878 will soon time out since no new Joins will arrive. 2880 After some later time, C may move to a new Server Y2 and depart from 2881 old Sever Y1. In that case, Y1 sends Join messages for any of C's 2882 active (S,G) groups to Y2 while including its own LLA as the source 2883 address. This causes Y2 to include Y1 in the multicast forwarding 2884 tree during the interim time that Y1's symmetric neighbor cache entry 2885 for C is in the DEPARTED state. At the same time, Y1 sends an 2886 unsolicited NA message to X with an OMNI option with S/T-ifIndex in 2887 the header set to 0 and a release indication to cause X to release 2888 its asymmetric neighbor cache entry. X then sends a new Join message 2889 to S via the spanning tree and re-initiates route optimization the 2890 same as if it were receiving a fresh Join message from a node on a 2891 downstream link. 2893 3.17.2. Any-Source Multicast (ASM) 2895 When an ROS X acting as a PIM router receives a Join/Prune from a 2896 node on its downstream interfaces containing one or more (*,G) pairs, 2897 it updates its Multicast Routing Information Base (MRIB) accordingly. 2898 X then forwards a copy of the message to the Rendezvous Point (RP) R 2899 for each G over the spanning tree. X uses its own LLA as the source 2900 address and ALL-PIM-ROUTERS as the destination address, then 2901 encapsulates each message in an OAL header with source address set to 2902 the ULA of X and destination address set to R, then sends the message 2903 into the spanning tree. At the same time, if the message was a Join 2904 X initiates NS/NA route optimization the same as for the SSM case 2905 discussed in Section 3.17.1. 2907 For each source S that sends multicast traffic to group G via R, the 2908 Proxy/Server Z* for the Client that aggregates S encapsulates the 2909 packets in PIM Register messages and forwards them to R via the 2910 spanning tree, which may then elect to send a PIM Join to Z*. This 2911 will result in an (S,G) tree rooted at Z* with R as the next hop so 2912 that R will begin to receive two copies of the packet; one native 2913 copy from the (S, G) tree and a second copy from the pre-existing (*, 2914 G) tree that still uses PIM Register encapsulation. R can then issue 2915 a PIM Register-stop message to suppress the Register-encapsulated 2916 stream. At some later time, if C moves to a new Proxy/Server Z*, it 2917 resumes sending packets via PIM Register encapsulation via the new 2918 Z*. 2920 At the same time, as multicast listeners discover individual S's for 2921 a given G, they can initiate an (S,G) Join for each S under the same 2922 procedures discussed in Section 3.17.1. Once the (S,G) tree is 2923 established, the listeners can send (S, G) Prune messages to R so 2924 that multicast packets for group G sourced by S will only be 2925 delivered via the (S, G) tree and not from the (*, G) tree rooted at 2926 R. All mobility considerations discussed for SSM apply. 2928 3.17.3. Bi-Directional PIM (BIDIR-PIM) 2930 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 2931 approach to ASM that treats the Rendezvous Point (RP) as a Designated 2932 Forwarder (DF). Further considerations for BIDIR-PIM are out of 2933 scope. 2935 3.18. Operation over Multiple OMNI Links 2937 An AERO Client can connect to multiple OMNI links the same as for any 2938 data link service. In that case, the Client maintains a distinct 2939 OMNI interface for each link, e.g., 'omni0' for the first link, 2940 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 2941 would include its own distinct set of Bridges, Servers and Proxys, 2942 thereby providing redundancy in case of failures. 2944 Each OMNI link could utilize the same or different ANET connections. 2945 The links can be distinguished at the link-layer via the SRT prefix 2946 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 2947 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 2948 MSPs on each link. This gives rise to the opportunity for supporting 2949 multiple redundant networked paths, with each VLAN distinguished by a 2950 different SRT "color" (see: Section 3.2.5). 2952 The Client's IP layer can select the outgoing OMNI interface 2953 appropriate for a given traffic profile while (in the reverse 2954 direction) correspondent nodes must have some way of steering their 2955 packets destined to a target via the correct OMNI link. 2957 In a first alternative, if each OMNI link services different MSPs, 2958 then the Client can receive a distinct MNP from each of the links. 2959 IP routing will therefore assure that the correct OMNI link is used 2960 for both outbound and inbound traffic. This can be accomplished 2961 using existing technologies and approaches, and without requiring any 2962 special supporting code in correspondent nodes or Bridges. 2964 In a second alternative, if each OMNI link services the same MSP(s) 2965 then each link could assign a distinct "OMNI link Anycast" address 2966 that is configured by all Bridges on the link. Correspondent nodes 2967 can then perform Segment Routing to select the correct SRT, which 2968 will then direct the packet over multiple hops to the target. 2970 3.19. DNS Considerations 2972 AERO Client MNs and INET correspondent nodes consult the Domain Name 2973 System (DNS) the same as for any Internetworking node. When 2974 correspondent nodes and Client MNs use different IP protocol versions 2975 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 2976 A records for IPv4 address mappings to MNs which must then be 2977 populated in Relay NAT64 mapping caches. In that way, an IPv4 2978 correspondent node can send packets to the IPv4 address mapping of 2979 the target MN, and the Relay will translate the IPv4 header and 2980 destination address into an IPv6 header and IPv6 destination address 2981 of the MN. 2983 When an AERO Client registers with an AERO Server, the Server can 2984 return the address(es) of DNS servers in RDNSS options [RFC6106]. 2985 The DNS server provides the IP addresses of other MNs and 2986 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 2988 3.20. Transition Considerations 2990 OAL encapsulation ensures that dissimilar INET partitions can be 2991 joined into a single unified OMNI link, even though the partitions 2992 themselves may have differing protocol versions and/or incompatible 2993 addressing plans. However, a commonality can be achieved by 2994 incrementally distributing globally routable (i.e., native) IP 2995 prefixes to eventually reach all nodes (both mobile and fixed) in all 2996 OMNI link segments. This can be accomplished by incrementally 2997 deploying AERO Relays on each INET partition, with each Relay 2998 distributing its MNPs and/or discovering non-MNP IP GUA prefixes on 2999 its INET links. 3001 This gives rise to the opportunity to eventually distribute native IP 3002 addresses to all nodes, and to present a unified OMNI link view even 3003 if the INET partitions remain in their current protocol and 3004 addressing plans. In that way, the OMNI link can serve the dual 3005 purpose of providing a mobility/multilink service and a transition 3006 service. Or, if an INET partition is transitioned to a native IP 3007 protocol version and addressing scheme that is compatible with the 3008 OMNI link MNP-based addressing scheme, the partition and OMNI link 3009 can be joined by Relays. 3011 Relays that connect INETs/EUNs with dissimilar IP protocol versions 3012 may need to employ a network address and protocol translation 3013 function such as NAT64 [RFC6146]. 3015 3.21. Detecting and Reacting to Server and Bridge Failures 3017 In environments where rapid failure recovery is required, Servers and 3018 Bridges SHOULD use Bidirectional Forwarding Detection (BFD) 3019 [RFC5880]. Nodes that use BFD can quickly detect and react to 3020 failures so that cached information is re-established through 3021 alternate nodes. BFD control messaging is carried only over well- 3022 connected ground domain networks (i.e., and not low-end radio links) 3023 and can therefore be tuned for rapid response. 3025 Servers and Bridges maintain BFD sessions in parallel with their BGP 3026 peerings. If a Server or Bridge fails, BGP peers will quickly re- 3027 establish routes through alternate paths the same as for common BGP 3028 deployments. Similarly, Proxys maintain BFD sessions with their 3029 associated Bridges even though they do not establish BGP peerings 3030 with them. 3032 Proxys SHOULD use proactive NUD for Servers for which there are 3033 currently active ANET Clients in a manner that parallels BFD, i.e., 3034 by sending unicast NS messages in rapid succession to receive 3035 solicited NA messages. When the Proxy is also sending RS messages on 3036 behalf of ANET Clients, the RS/RA messaging can be considered as 3037 equivalent hints of forward progress. This means that the Proxy need 3038 not also send a periodic NS if it has already sent an RS within the 3039 same period. If a Server fails, the Proxy will cease to receive 3040 advertisements and can quickly inform Clients of the outage by 3041 sending multicast RA messages on the ANET interface. 3043 The Proxy sends multicast RA messages with source address set to the 3044 Server's address, destination address set to (link-local) All-Nodes 3045 multicast, and Router Lifetime set to 0. The Proxy SHOULD send 3046 MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small delays 3047 [RFC4861]. Any Clients on the ANET interface that have been using 3048 the (now defunct) Server will receive the RA messages and associate 3049 with a new Server. 3051 3.22. AERO Clients on the Open Internet 3053 AERO Clients that connect to the open Internet via INET interfaces 3054 can establish a VPN or direct link to securely connect to a Server in 3055 a "tethered" arrangement with all of the Client's traffic transiting 3056 the Server. Alternatively, the Client can associate with an INET 3057 Server using UDP/IP encapsulation and asymmetric securing services as 3058 discussed in the following sections. 3060 When a Client's OMNI interface enables an INET underlying interface, 3061 it first determines whether the interface is likely to be behind a 3062 NAT. For IPv4, the Client assumes it is on the open Internet if the 3063 INET address is not a special-use IPv4 address per [RFC3330]. 3064 Similarly for IPv6, the Client assumes it is on the open Internet if 3065 the INET address is not a link-local [RFC4291] or unique-local 3066 [RFC4193] IPv6 address. 3068 The Client then prepares a UDP/IP-encapsulated RS message with IPv6 3069 source address set to its MNP-LLA, with IPv6 destination set to 3070 (link-local) All-Routers multicast and with an OMNI option with 3071 underlying interface attributes. If the Client believes that it is 3072 on the open Internet, it SHOULD include Interface Attributes with the 3073 L2ADDR used for INET encapsulation (otherwise, it MAY omit L2ADDR). 3074 If the underlying address is IPv4, the Client includes the Port 3075 Number and IPv4 address written in obfuscated form [RFC4380] as 3076 discussed in Section 3.3. If the underlying interface address is 3077 IPv6, the Client instead includes the Port Number and IPv6 address in 3078 obfuscated form. The Client finally includes a HIP "Initiator" 3079 message sub-option in the OMNI option 3080 [I-D.templin-6man-omni-interface] to provide message authentication, 3081 sets the UDP/IP source to its INET address and UDP port, sets the 3082 UDP/IP destination to the Server's INET address and the AERO service 3083 port number (8060), then sends the message to the Server. 3085 When the Server receives the RS, it authenticates the message and 3086 registers the Client's MNP and INET interface information according 3087 to the OMNI option parameters. If the RS message OMNI option 3088 includes Interface Attributes with an L2ADDR, the Server compares the 3089 encapsulation IP address and UDP port number with the (unobfuscated) 3090 values. If the values are the same, the Server caches the Client's 3091 information as "INET" addresses meaning that the Client is likely to 3092 accept direct messages without requiring NAT traversal exchanges. If 3093 the values are different (or, if the OMNI option did not include an 3094 L2ADDR) the Server instead caches the Client's information as 3095 "mapped" addresses meaning that NAT traversal exchanges may be 3096 necessary. 3098 The Server then returns an RA message with IPv6 source and 3099 destination set corresponding to the addresses in the RS, and with an 3100 OMNI option with a HIP "Responder" message sub-option per 3101 [I-D.templin-6man-omni-interface] that contains an acknowledgement of 3102 the update sent by the Client. For IPv4, the Server also includes an 3103 Origin Indication sub-option per [I-D.templin-6man-omni-interface] 3104 with the mapped and obfuscated Port Number and IPv4 address observed 3105 in the encapsulation headers. 3107 When the Client receives the RA message, it compares the mapped Port 3108 Number and IP address from the Origin sub-option with its own 3109 address. If the addresses are the same, the Client assumes the open 3110 Internet / Cone NAT principle; if the addresses are different, the 3111 Client instead assumes that further qualification procedures are 3112 necessary to detect the type of NAT and proceeds according to 3113 standard [RFC4380] procedures. 3115 After the Client has registered its INET interfaces in such RS/RA 3116 exchanges it sends periodic RS messages to receive fresh RA messages 3117 before the Router Lifetime received on each INET interface expires. 3118 The Client also maintains default routes via its Servers, i.e., the 3119 same as described in earlier sections. 3121 When the Client sends messages to target IP addresses, it also 3122 invokes route optimization per Section 3.14 using IPv6 ND address 3123 resolution messaging. The Client sends the NS(AR) message to the 3124 Server with an OMNI option with a HIP "Update/Sequence" message sub- 3125 option. The Client sets the NS source address to the Client's MNP- 3126 LLA and destination address to the target solicited node multicast 3127 address. The Client wraps the NS message in an OAL header with 3128 source address set to its own MNP-ULA and destination address set to 3129 the Server's ADM-ULA. The Client then wraps the OAL message in a 3130 UDP/IP header and sends it to the Server. 3132 When the Server receives the OAL-encapsulated NS, it authenticates 3133 the message by processing the HIP message sub-option and sends a 3134 corresponding NS(AR) message over the spanning tree the same as if it 3135 were the ROS, but with the OAL source address set to the Server's 3136 ADM-ULA, with destination set to the MNP-ULA of the target, and with 3137 an OMNI option that includes no sub-options. When the ROR receives 3138 the NS(AR), it adds the Server's ADM-ULA and Client's MNP-LLA to the 3139 target's Report List, and returns an NA(AR) with OMNI option 3140 information for the target including all of the target's Interface 3141 Attributes. The ROR sets the NA(AR) source address to the target's 3142 MNP-LLA and sets the destination address to the Client's MNP-LLA, 3143 then sets the OAL source address to the ADM-ULA of the ROR and the 3144 destination to the ADM-ULA of the Server. When the Server receives 3145 the NA(AR) message, it rewrites the OAL source address to its own 3146 ADM-ULA and the destination address to the MNP-ULA of the Client, 3147 then includes a HIP "Update/Acknowledge" message sub-option in the 3148 OMNI option, wraps the message in UDP/IP headers, and sends it to the 3149 Client. 3151 Following route optimization for targets in the same OMNI link 3152 segment, if the target's L2ADDR is on the open INET, the Client 3153 forwards data packets directly to the target INET address. If the 3154 target is behind a NAT, the Client first establishes NAT state for 3155 the L2ADDR using the "direct bubble" and NUD mechanisms discussed in 3156 Section 3.10.1. The Client continues to send data packets via its 3157 Server until NAT state is populated, then begins forwarding packets 3158 via the direct path through the NAT to the target. For targets in 3159 different OMNI link segments, the Client uses OAL/ORH encapsulation 3160 and forwards data packets to the Bridge that returned the NA message. 3162 The ROR may return uNAs via the Server if the target moves, and the 3163 Server will send corresponding authentication-protected uNAs to the 3164 Client. The Client can also send NUD messages to test forward path 3165 reachability even though there is no security association between the 3166 Client and the target. 3168 The Client sends UDP/IP encapsulated IPv6 packets to route-optimized 3169 neighbors in the same OMNI link segment no larger than 576 bytes in 3170 one piece and without OAL encapsulation. Otherwise, the Client 3171 inserts an OAL header with source set to its own MNP-ULA and 3172 destination set to the MNP-ULA of the target and uses OAL 3173 fragmentation if necessary according to Section 3.9. The Client then 3174 encapsulates each fragment in a UDP/IP header and sends the fragments 3175 to the next hop. 3177 Note: The NAT traversal procedures specified in this document are 3178 applicable for Cone, Address-Restricted and Port-Restricted NATs 3179 only. While future updates to this document may specify procedures 3180 for other NAT variations (e.g., hairpinning and various forms of 3181 Symmetric NATs), it should be noted that continuous communications 3182 are always possible through forwarding via a Server even if NAT 3183 traversal is not employed. 3185 3.23. Time-Varying MNPs 3187 In some use cases, it is desirable, beneficial and efficient for the 3188 Client to receive a constant MNP that travels with the Client 3189 wherever it moves. For example, this would allow air traffic 3190 controllers to easily track aircraft, etc. In other cases, however 3191 (e.g., intelligent transportation systems), the MN may be willing to 3192 sacrifice a modicum of efficiency in order to have time-varying MNPs 3193 that can be changed every so often to defeat adversarial tracking. 3195 The DHCPv6 service offers a way for Clients that desire time-varying 3196 MNPs to obtain short-lived prefixes (e.g., on the order of a small 3197 number of minutes). In that case, the identity of the Client would 3198 not be bound to the MNP but rather the Client's identity would be 3199 bound to the DHCPv6 Device Unique Identifier (DUID) and used as the 3200 seed for Prefix Delegation. The Client would then be obligated to 3201 renumber its internal networks whenever its MNP (and therefore also 3202 its MNP-LLA) changes. This should not present a challenge for 3203 Clients with automated network renumbering services, however presents 3204 limits for the durations of ongoing sessions that would prefer to use 3205 a constant address. 3207 4. Implementation Status 3209 An early AERO implementation based on OpenVPN (https://openvpn.net/) 3210 was announced on the v6ops mailing list on January 10, 2018 and an 3211 initial public release of the AERO proof-of-concept source code was 3212 announced on the intarea mailing list on August 21, 2015. 3214 AERO Release-3.0.2 was tagged on October 15, 2020, and is undergoing 3215 internal testing. Additional releases expected Q42020, with first 3216 public release expected before year-end. 3218 5. IANA Considerations 3220 The IANA has assigned the UDP port number "8060" for an earlier 3221 experimental version of AERO [RFC6706]. This document obsoletes 3222 [RFC6706] and claims the UDP port number "8060" for all future use. 3224 The IANA is instructed to assign a new type value TBD in the IPv6 3225 Routing Types registry. 3227 No further IANA actions are required. 3229 6. Security Considerations 3231 AERO Bridges configure secured tunnels with AERO Servers, Relays and 3232 Proxys within their local OMNI link segments. Applicable secured 3233 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 3234 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 3235 segments in turn configure secured tunnels for their neighboring AERO 3236 Bridges in a spanning tree topology. Therefore, control messages 3237 exchanged between any pair of OMNI link neighbors on the spanning 3238 tree are already secured. 3240 AERO Servers, Relays and Proxys targeted by a route optimization may 3241 also receive data packets directly from arbitrary nodes in INET 3242 partitions instead of via the spanning tree. For INET partitions 3243 that apply effective ingress filtering to defeat source address 3244 spoofing, the simple data origin authentication procedures in 3245 Section 3.8 can be applied. 3247 For INET partitions that require strong security in the data plane, 3248 two options for securing communications include 1) disable route 3249 optimization so that all traffic is conveyed over secured tunnels, or 3250 2) enable on-demand secure tunnel creation between INET partition 3251 neighbors. Option 1) would result in longer routes than necessary 3252 and traffic concentration on critical infrastructure elements. 3253 Option 2) could be coordinated by establishing a secured tunnel on- 3254 demand instead of performing an NS/NA exchange in the route 3255 optimization procedures. Procedures for establishing on-demand 3256 secured tunnels are out of scope. 3258 AERO Clients that connect to secured ANETs need not apply security to 3259 their ND messages, since the messages will be intercepted by a 3260 perimeter Proxy that applies security on its INET-facing interface as 3261 part of the spanning tree (see above). AERO Clients connected to the 3262 open INET can use symmetric network and/or transport layer security 3263 services such as VPNs or can by some other means establish a direct 3264 link. When a VPN or direct link may be impractical, however, the 3265 HIP-based authentication services specified in [RFC7401] should be 3266 applied. 3268 Application endpoints SHOULD use application-layer security services 3269 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 3270 protection as for critical secured Internet services. AERO Clients 3271 that require host-based VPN services SHOULD use symmetric network 3272 and/or transport layer security services such as IPsec, TLS/SSL, 3273 DTLS, etc. AERO Proxys and Servers can also provide a network-based 3274 VPN service on behalf of the Client, e.g., if the Client is located 3275 within a secured enclave and cannot establish a VPN on its own 3276 behalf. 3278 AERO Servers and Bridges present targets for traffic amplification 3279 Denial of Service (DoS) attacks. This concern is no different than 3280 for widely-deployed VPN security gateways in the Internet, where 3281 attackers could send spoofed packets to the gateways at high data 3282 rates. This can be mitigated by connecting Servers and Bridges over 3283 dedicated links with no connections to the Internet and/or when 3284 connections to the Internet are only permitted through well-managed 3285 firewalls. Traffic amplification DoS attacks can also target an AERO 3286 Client's low data rate links. This is a concern not only for Clients 3287 located on the open Internet but also for Clients in secured 3288 enclaves. AERO Servers and Proxys can institute rate limits that 3289 protect Clients from receiving packet floods that could DoS low data 3290 rate links. 3292 AERO Relays must implement ingress filtering to avoid a spoofing 3293 attack in which spurious messages with ULA addresses are injected 3294 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3295 that their connectivity is not used by unauthorized nodes on their 3296 EUNs to gain access to a protected network, i.e., AERO Clients that 3297 act as routers MUST NOT provide routing services for unauthorized 3298 nodes. (This concern is no different than for ordinary hosts that 3299 receive an IP address delegation but then "share" the address with 3300 other nodes via some form of Internet connection sharing such as 3301 tethering.) 3303 The MAP list MUST be well-managed and secured from unauthorized 3304 tampering, even though the list contains only public information. 3305 The MAP list can be conveyed to the Client in a similar fashion as in 3306 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3307 upload of a static file, DNS lookups, etc.). 3309 SRH authentication facilities are specified in [RFC8754]. 3311 Security considerations for accepting link-layer ICMP messages and 3312 reflected packets are discussed throughout the document. 3314 Security considerations for IPv6 fragmentation and reassembly are 3315 discussed in [I-D.templin-6man-omni-interface]. 3317 7. Acknowledgements 3319 Discussions in the IETF, aviation standards communities and private 3320 exchanges helped shape some of the concepts in this work. 3321 Individuals who contributed insights include Mikael Abrahamsson, Mark 3322 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3323 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3324 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3325 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3326 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3327 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3328 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3329 Wood and James Woodyatt. Members of the IESG also provided valuable 3330 input during their review process that greatly improved the document. 3331 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3332 for their shepherding guidance during the publication of the AERO 3333 first edition. 3335 This work has further been encouraged and supported by Boeing 3336 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3337 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3338 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3339 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3340 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3341 Gene MacLean III, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3342 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3343 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3344 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3345 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3346 Benson, Katie Tran and Eric Yeh are especially acknowledged for 3347 implementing the AERO functions as extensions to the public domain 3348 OpenVPN distribution. 3350 Earlier works on NBMA tunneling approaches are found in 3351 [RFC2529][RFC5214][RFC5569]. 3353 Many of the constructs presented in this second edition of AERO are 3354 based on the author's earlier works, including: 3356 o The Internet Routing Overlay Network (IRON) 3357 [RFC6179][I-D.templin-ironbis] 3359 o Virtual Enterprise Traversal (VET) 3360 [RFC5558][I-D.templin-intarea-vet] 3362 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3363 [RFC5320][I-D.templin-intarea-seal] 3365 o AERO, First Edition [RFC6706] 3367 Note that these works cite numerous earlier efforts that are not also 3368 cited here due to space limitations. The authors of those earlier 3369 works are acknowledged for their insights. 3371 This work is aligned with the NASA Safe Autonomous Systems Operation 3372 (SASO) program under NASA contract number NNA16BD84C. 3374 This work is aligned with the FAA as per the SE2025 contract number 3375 DTFAWA-15-D-00030. 3377 This work is aligned with the Boeing Commercial Airplanes (BCA) 3378 Internet of Things (IoT) and autonomy programs. 3380 This work is aligned with the Boeing Information Technology (BIT) 3381 MobileNet program. 3383 8. References 3385 8.1. Normative References 3387 [I-D.templin-6man-omni-interface] 3388 Templin, F. and T. Whyman, "Transmission of IP Packets 3389 over Overlay Multilink Network (OMNI) Interfaces", draft- 3390 templin-6man-omni-interface-69 (work in progress), January 3391 2021. 3393 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3394 DOI 10.17487/RFC0791, September 1981, 3395 . 3397 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3398 RFC 792, DOI 10.17487/RFC0792, September 1981, 3399 . 3401 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3402 Requirement Levels", BCP 14, RFC 2119, 3403 DOI 10.17487/RFC2119, March 1997, 3404 . 3406 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3407 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3408 December 1998, . 3410 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3411 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3412 DOI 10.17487/RFC3971, March 2005, 3413 . 3415 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3416 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3417 . 3419 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3420 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3421 November 2005, . 3423 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3424 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3425 . 3427 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3428 Network Address Translations (NATs)", RFC 4380, 3429 DOI 10.17487/RFC4380, February 2006, 3430 . 3432 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3433 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3434 DOI 10.17487/RFC4861, September 2007, 3435 . 3437 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3438 Address Autoconfiguration", RFC 4862, 3439 DOI 10.17487/RFC4862, September 2007, 3440 . 3442 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3443 DOI 10.17487/RFC6081, January 2011, 3444 . 3446 [RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T. 3447 Henderson, "Host Identity Protocol Version 2 (HIPv2)", 3448 RFC 7401, DOI 10.17487/RFC7401, April 2015, 3449 . 3451 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3452 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3453 May 2017, . 3455 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3456 (IPv6) Specification", STD 86, RFC 8200, 3457 DOI 10.17487/RFC8200, July 2017, 3458 . 3460 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3461 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3462 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3463 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3464 . 3466 8.2. Informative References 3468 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3469 2016. 3471 [I-D.bonica-6man-comp-rtg-hdr] 3472 Bonica, R., Kamite, Y., Alston, A., Henriques, D., and L. 3473 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3474 bonica-6man-comp-rtg-hdr-24 (work in progress), January 3475 2021. 3477 [I-D.bonica-6man-crh-helper-opt] 3478 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3479 Routing Header (CRH) Helper Option", draft-bonica-6man- 3480 crh-helper-opt-02 (work in progress), October 2020. 3482 [I-D.ietf-intarea-frag-fragile] 3483 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 3484 and F. Gont, "IP Fragmentation Considered Fragile", draft- 3485 ietf-intarea-frag-fragile-17 (work in progress), September 3486 2019. 3488 [I-D.ietf-intarea-tunnels] 3489 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3490 Architecture", draft-ietf-intarea-tunnels-10 (work in 3491 progress), September 2019. 3493 [I-D.ietf-ipwave-vehicular-networking] 3494 Jeong, J., "IPv6 Wireless Access in Vehicular Environments 3495 (IPWAVE): Problem Statement and Use Cases", draft-ietf- 3496 ipwave-vehicular-networking-19 (work in progress), July 3497 2020. 3499 [I-D.ietf-rtgwg-atn-bgp] 3500 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3501 Moreno, "A Simple BGP-based Mobile Routing System for the 3502 Aeronautical Telecommunications Network", draft-ietf- 3503 rtgwg-atn-bgp-10 (work in progress), January 2021. 3505 [I-D.templin-6man-dhcpv6-ndopt] 3506 Templin, F., "A Unified Stateful/Stateless Configuration 3507 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-11 3508 (work in progress), January 2021. 3510 [I-D.templin-intarea-seal] 3511 Templin, F., "The Subnetwork Encapsulation and Adaptation 3512 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3513 progress), January 2014. 3515 [I-D.templin-intarea-vet] 3516 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3517 templin-intarea-vet-40 (work in progress), May 2013. 3519 [I-D.templin-ipwave-uam-its] 3520 Templin, F., "Urban Air Mobility Implications for 3521 Intelligent Transportation Systems", draft-templin-ipwave- 3522 uam-its-04 (work in progress), January 2021. 3524 [I-D.templin-ironbis] 3525 Templin, F., "The Interior Routing Overlay Network 3526 (IRON)", draft-templin-ironbis-16 (work in progress), 3527 March 2014. 3529 [I-D.templin-v6ops-pdhost] 3530 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3531 Models", draft-templin-v6ops-pdhost-27 (work in progress), 3532 January 2021. 3534 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3536 [RFC1035] Mockapetris, P., "Domain names - implementation and 3537 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3538 November 1987, . 3540 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3541 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3542 . 3544 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3545 DOI 10.17487/RFC2003, October 1996, 3546 . 3548 [RFC2004] Perkins, C., "Minimal Encapsulation within IP", RFC 2004, 3549 DOI 10.17487/RFC2004, October 1996, 3550 . 3552 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3553 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3554 . 3556 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 3557 Networks", RFC 2464, DOI 10.17487/RFC2464, December 1998, 3558 . 3560 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3561 Domains without Explicit Tunnels", RFC 2529, 3562 DOI 10.17487/RFC2529, March 1999, 3563 . 3565 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3566 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3567 . 3569 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3570 of Explicit Congestion Notification (ECN) to IP", 3571 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3572 . 3574 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3575 DOI 10.17487/RFC3330, September 2002, 3576 . 3578 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3579 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3580 DOI 10.17487/RFC3810, June 2004, 3581 . 3583 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 3584 Unique IDentifier (UUID) URN Namespace", RFC 4122, 3585 DOI 10.17487/RFC4122, July 2005, 3586 . 3588 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3589 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3590 January 2006, . 3592 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3593 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3594 DOI 10.17487/RFC4271, January 2006, 3595 . 3597 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3598 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3599 2006, . 3601 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3602 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3603 December 2005, . 3605 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3606 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3607 2006, . 3609 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3610 Control Message Protocol (ICMPv6) for the Internet 3611 Protocol Version 6 (IPv6) Specification", STD 89, 3612 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3613 . 3615 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3616 Protocol (LDAP): The Protocol", RFC 4511, 3617 DOI 10.17487/RFC4511, June 2006, 3618 . 3620 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3621 "Considerations for Internet Group Management Protocol 3622 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3623 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3624 . 3626 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3627 "Internet Group Management Protocol (IGMP) / Multicast 3628 Listener Discovery (MLD)-Based Multicast Forwarding 3629 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3630 August 2006, . 3632 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3633 Algorithms in Cryptographically Generated Addresses 3634 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3635 . 3637 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3638 "Bidirectional Protocol Independent Multicast (BIDIR- 3639 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3640 . 3642 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3643 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3644 DOI 10.17487/RFC5214, March 2008, 3645 . 3647 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3648 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3649 February 2010, . 3651 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3652 Route Optimization Requirements for Operational Use in 3653 Aeronautics and Space Exploration Mobile Networks", 3654 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3655 . 3657 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3658 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3659 . 3661 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3662 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3663 January 2010, . 3665 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3666 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3667 . 3669 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3670 "IPv6 Router Advertisement Options for DNS Configuration", 3671 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3672 . 3674 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3675 NAT64: Network Address and Protocol Translation from IPv6 3676 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3677 April 2011, . 3679 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3680 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3681 . 3683 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3684 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3685 DOI 10.17487/RFC6221, May 2011, 3686 . 3688 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3689 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3690 DOI 10.17487/RFC6273, June 2011, 3691 . 3693 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3694 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3695 January 2012, . 3697 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 3698 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, 3699 DOI 10.17487/RFC6355, August 2011, 3700 . 3702 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3703 for Equal Cost Multipath Routing and Link Aggregation in 3704 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3705 . 3707 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3708 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3709 . 3711 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3712 UDP Checksums for Tunneled Packets", RFC 6935, 3713 DOI 10.17487/RFC6935, April 2013, 3714 . 3716 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3717 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3718 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3719 . 3721 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3722 Korhonen, "Requirements for Distributed Mobility 3723 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3724 . 3726 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3727 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3728 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3729 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3730 2016, . 3732 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3733 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3734 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3735 July 2018, . 3737 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3738 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3739 . 3741 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 3742 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3743 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 3744 . 3746 [WG] Wireguard, "Wireguard, https://www.wireguard.com", August 3747 2020. 3749 Appendix A. Non-Normative Considerations 3751 AERO can be applied to a multitude of Internetworking scenarios, with 3752 each having its own adaptations. The following considerations are 3753 provided as non-normative guidance: 3755 A.1. Implementation Strategies for Route Optimization 3757 Route optimization as discussed in Section 3.14 results in the route 3758 optimization source (ROS) creating an asymmetric neighbor cache entry 3759 for the target neighbor. The neighbor cache entry is maintained for 3760 at most ReachableTime seconds and then deleted unless updated. In 3761 order to refresh the neighbor cache entry lifetime before the 3762 ReachableTime timer expires, the specification requires 3763 implementations to issue a new NS/NA exchange to reset ReachableTime 3764 while data packets are still flowing. However, the decision of when 3765 to initiate a new NS/NA exchange and to perpetuate the process is 3766 left as an implementation detail. 3768 One possible strategy may be to monitor the neighbor cache entry 3769 watching for data packets for (ReachableTime - 5) seconds. If any 3770 data packets have been sent to the neighbor within this timeframe, 3771 then send an NS to receive a new NA. If no data packets have been 3772 sent, wait for 5 additional seconds and send an immediate NS if any 3773 data packets are sent within this "expiration pending" 5 second 3774 window. If no additional data packets are sent within the 5 second 3775 window, delete the neighbor cache entry. 3777 The monitoring of the neighbor data packet traffic therefore becomes 3778 an asymmetric ongoing process during the neighbor cache entry 3779 lifetime. If the neighbor cache entry expires, future data packets 3780 will trigger a new NS/NA exchange while the packets themselves are 3781 delivered over a longer path until route optimization state is re- 3782 established. 3784 A.2. Implicit Mobility Management 3786 OMNI interface neighbors MAY provide a configuration option that 3787 allows them to perform implicit mobility management in which no ND 3788 messaging is used. In that case, the Client only transmits packets 3789 over a single interface at a time, and the neighbor always observes 3790 packets arriving from the Client from the same link-layer source 3791 address. 3793 If the Client's underlying interface address changes (either due to a 3794 readdressing of the original interface or switching to a new 3795 interface) the neighbor immediately updates the neighbor cache entry 3796 for the Client and begins accepting and sending packets according to 3797 the Client's new address. This implicit mobility method applies to 3798 use cases such as cellphones with both WiFi and Cellular interfaces 3799 where only one of the interfaces is active at a given time, and the 3800 Client automatically switches over to the backup interface if the 3801 primary interface fails. 3803 A.3. Direct Underlying Interfaces 3805 When a Client's OMNI interface is configured over a Direct interface, 3806 the neighbor at the other end of the Direct link can receive packets 3807 without any encapsulation. In that case, the Client sends packets 3808 over the Direct link according to QoS preferences. If the Direct 3809 interface has the highest QoS preference, then the Client's IP 3810 packets are transmitted directly to the peer without going through an 3811 ANET/INET. If other interfaces have higher QoS preferences, then the 3812 Client's IP packets are transmitted via a different interface, which 3813 may result in the inclusion of Proxys, Servers and Bridges in the 3814 communications path. Direct interfaces must be tested periodically 3815 for reachability, e.g., via NUD. 3817 A.4. AERO Critical Infrastructure Considerations 3819 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 3820 IP routers or virtual machines in the cloud. Bridges must be 3821 provisioned, supported and managed by the INET administrative 3822 authority, and connected to the Bridges of other INETs via inter- 3823 domain peerings. Cost for purchasing, configuring and managing 3824 Bridges is nominal even for very large OMNI links. 3826 AERO Servers can be standard dedicated server platforms, but most 3827 often will be deployed as virtual machines in the cloud. The only 3828 requirements for Servers are that they can run the AERO user-level 3829 code and have at least one network interface connection to the INET. 3830 As with Bridges, Servers must be provisioned, supported and managed 3831 by the INET administrative authority. Cost for purchasing, 3832 configuring and managing Servers is nominal especially for virtual 3833 Servers hosted in the cloud. 3835 AERO Proxys are most often standard dedicated server platforms with 3836 one network interface connected to the ANET and a second interface 3837 connected to an INET. As with Servers, the only requirements are 3838 that they can run the AERO user-level code and have at least one 3839 interface connection to the INET. Proxys must be provisioned, 3840 supported and managed by the ANET administrative authority. Cost for 3841 purchasing, configuring and managing Proxys is nominal, and borne by 3842 the ANET administrative authority. 3844 AERO Relays can be any dedicated server or COTS router platform 3845 connected to INETs and/or EUNs. The Relay connects to the OMNI link 3846 and engages in eBGP peering with one or more Bridges as a stub AS. 3847 The Relay then injects its MNPs and/or non-MNP prefixes into the BGP 3848 routing system, and provisions the prefixes to its downstream- 3849 attached networks. The Relay can perform ROS/ROR services the same 3850 as for any Server, and can route between the MNP and non-MNP address 3851 spaces. 3853 A.5. AERO Server Failure Implications 3855 AERO Servers may appear as a single point of failure in the 3856 architecture, but such is not the case since all Servers on the link 3857 provide identical services and loss of a Server does not imply 3858 immediate and/or comprehensive communication failures. Although 3859 Clients typically associate with a single Server at a time, Server 3860 failure is quickly detected and conveyed by Bidirectional Forward 3861 Detection (BFD) and/or proactive NUD allowing Clients to migrate to 3862 new Servers. 3864 If a Server fails, ongoing packet forwarding to Clients will continue 3865 by virtue of the asymmetric neighbor cache entries that have already 3866 been established in route optimization sources (ROSs). If a Client 3867 also experiences mobility events at roughly the same time the Server 3868 fails, unsolicited NA messages may be lost but proxy neighbor cache 3869 entries in the DEPARTED state will ensure that packet forwarding to 3870 the Client's new locations will continue for up to DepartTime 3871 seconds. 3873 If a Client is left without a Server for an extended timeframe (e.g., 3874 greater than ReachableTime seconds) then existing asymmetric neighbor 3875 cache entries will eventually expire and both ongoing and new 3876 communications will fail. The original source will continue to 3877 retransmit until the Client has established a new Server 3878 relationship, after which time continuous communications will resume. 3880 Therefore, providing many Servers on the link with high availability 3881 profiles provides resilience against loss of individual Servers and 3882 assurance that Clients can establish new Server relationships quickly 3883 in event of a Server failure. 3885 A.6. AERO Client / Server Architecture 3887 The AERO architectural model is client / server in the control plane, 3888 with route optimization in the data plane. The same as for common 3889 Internet services, the AERO Client discovers the addresses of AERO 3890 Servers and selects one Server to connect to. The AERO service is 3891 analogous to common Internet services such as google.com, yahoo.com, 3892 cnn.com, etc. However, there is only one AERO service for the link 3893 and all Servers provide identical services. 3895 Common Internet services provide differing strategies for advertising 3896 server addresses to clients. The strategy is conveyed through the 3897 DNS resource records returned in response to name resolution queries. 3899 As of January 2020 Internet-based 'nslookup' services were used to 3900 determine the following: 3902 o When a client resolves the domainname "google.com", the DNS always 3903 returns one A record (i.e., an IPv4 address) and one AAAA record 3904 (i.e., an IPv6 address). The client receives the same addresses 3905 each time it resolves the domainname via the same DNS resolver, 3906 but may receive different addresses when it resolves the 3907 domainname via different DNS resolvers. But, in each case, 3908 exactly one A and one AAAA record are returned. 3910 o When a client resolves the domainname "ietf.org", the DNS always 3911 returns one A record and one AAAA record with the same addresses 3912 regardless of which DNS resolver is used. 3914 o When a client resolves the domainname "yahoo.com", the DNS always 3915 returns a list of 4 A records and 4 AAAA records. Each time the 3916 client resolves the domainname via the same DNS resolver, the same 3917 list of addresses are returned but in randomized order (i.e., 3918 consistent with a DNS round-robin strategy). But, interestingly, 3919 the same addresses are returned (albeit in randomized order) when 3920 the domainname is resolved via different DNS resolvers. 3922 o When a client resolves the domainname "amazon.com", the DNS always 3923 returns a list of 3 A records and no AAAA records. As with 3924 "yahoo.com", the same three A records are returned from any 3925 worldwide Internet connection point in randomized order. 3927 The above example strategies show differing approaches to Internet 3928 resilience and service distribution offered by major Internet 3929 services. The Google approach exposes only a single IPv4 and a 3930 single IPv6 address to clients. Clients can then select whichever IP 3931 protocol version offers the best response, but will always use the 3932 same IP address according to the current Internet connection point. 3933 This means that the IP address offered by the network must lead to a 3934 highly-available server and/or service distribution point. In other 3935 words, resilience is predicated on high availability within the 3936 network and with no client-initiated failovers expected (i.e., it is 3937 all-or-nothing from the client's perspective). However, Google does 3938 provide for worldwide distributed service distribution by virtue of 3939 the fact that each Internet connection point responds with a 3940 different IPv6 and IPv4 address. The IETF approach is like google 3941 (all-or-nothing from the client's perspective), but provides only a 3942 single IPv4 or IPv6 address on a worldwide basis. This means that 3943 the addresses must be made highly-available at the network level with 3944 no client failover possibility, and if there is any worldwide service 3945 distribution it would need to be conducted by a network element that 3946 is reached via the IP address acting as a service distribution point. 3948 In contrast to the Google and IETF philosophies, Yahoo and Amazon 3949 both provide clients with a (short) list of IP addresses with Yahoo 3950 providing both IP protocol versions and Amazon as IPv4-only. The 3951 order of the list is randomized with each name service query 3952 response, with the effect of round-robin load balancing for service 3953 distribution. With a short list of addresses, there is still 3954 expectation that the network will implement high availability for 3955 each address but in case any single address fails the client can 3956 switch over to using a different address. The balance then becomes 3957 one of function in the network vs function in the end system. 3959 The same implications observed for common highly-available services 3960 in the Internet apply also to the AERO client/server architecture. 3961 When an AERO Client connects to one or more ANETs, it discovers one 3962 or more AERO Server addresses through the mechanisms discussed in 3963 earlier sections. Each Server address presumably leads to a fault- 3964 tolerant clustering arrangement such as supported by Linux-HA, 3965 Extended Virtual Synchrony or Paxos. Such an arrangement has 3966 precedence in common Internet service deployments in lightweight 3967 virtual machines without requiring expensive hardware deployment. 3968 Similarly, common Internet service deployments set service IP 3969 addresses on service distribution points that may relay requests to 3970 many different servers. 3972 For AERO, the expectation is that a combination of the Google/IETF 3973 and Yahoo/Amazon philosophies would be employed. The AERO Client 3974 connects to different ANET access points and can receive 1-2 Server 3975 ADM-LLAs at each point. It then selects one AERO Server address, and 3976 engages in RS/RA exchanges with the same Server from all ANET 3977 connections. The Client remains with this Server unless or until the 3978 Server fails, in which case it can switch over to an alternate 3979 Server. The Client can likewise switch over to a different Server at 3980 any time if there is some reason for it to do so. So, the AERO 3981 expectation is for a balance of function in the network and end 3982 system, with fault tolerance and resilience at both levels. 3984 Appendix B. Change Log 3986 << RFC Editor - remove prior to publication >> 3988 Changes from draft-templin-intarea-6706bis-61 to draft-templin- 3989 intrea-6706bis-62: 3991 o New sub-section on OMNI Neighbor Interface Attributes 3993 Changes from draft-templin-intarea-6706bis-59 to draft-templin- 3994 intrea-6706bis-60: 3996 o Removed all references to S/TLLAO - all Interface Attributes are 3997 now maintained completely in the OMNI option. 3999 Changes from draft-templin-intarea-6706bis-58 to draft-templin- 4000 intrea-6706bis-59: 4002 o The term "Relay" used in older draft versions is now "Bridge". 4003 "Relay" now refers to what was formally called: "Gateway". 4005 o Fine-grained cleanup of Forwarding Algorithm; IPv6 ND message 4006 addressing; OMNI Prefix Lengths, etc. 4008 Changes from draft-templin-intarea-6706bis-54 to draft-templin- 4009 intrea-6706bis-55: 4011 o Updates on Segment Routing and S/TLLAO contents. 4013 o Various editorials and addressing cleanups. 4015 Changes from draft-templin-intarea-6706bis-52 to draft-templin- 4016 intrea-6706bis-53: 4018 o Normative reference to the OMNI spec, and remove portions that are 4019 already specified in OMNI. 4021 o Renamed "AERO interface/link" to "OMIN interface/link" throughout 4022 the document. 4024 o Truncated obsolete back section matter. 4026 Author's Address 4028 Fred L. Templin (editor) 4029 Boeing Research & Technology 4030 P.O. Box 3707 4031 Seattle, WA 98124 4032 USA 4034 Email: fltemplin@acm.org