idnits 2.17.1 draft-templin-intarea-6706bis-95.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 (March 17, 2021) is 1136 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 1052, but not defined == Missing Reference: 'ULA0' is mentioned on line 1052, but not defined == Missing Reference: 'ULA1' is mentioned on line 1052, but not defined == Missing Reference: 'ULA2' is mentioned on line 1052, but not defined == Unused Reference: 'RFC3971' is defined on line 3432, but no explicit reference was found in the text == Unused Reference: 'RFC3972' is defined on line 3437, but no explicit reference was found in the text == Unused Reference: 'RFC4193' is defined on line 3445, but no explicit reference was found in the text == Unused Reference: 'BGP' is defined on line 3490, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-comp-rtg-hdr' is defined on line 3493, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-crh-helper-opt' is defined on line 3499, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3504, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3527, but no explicit reference was found in the text == Unused Reference: 'RFC2003' is defined on line 3566, but no explicit reference was found in the text == Unused Reference: 'RFC4122' is defined on line 3605, but no explicit reference was found in the text == Unused Reference: 'RFC4982' is defined on line 3654, but no explicit reference was found in the text == Unused Reference: 'RFC6273' is defined on line 3710, but no explicit reference was found in the text == Unused Reference: 'RFC6355' is defined on line 3719, 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 (~~), 25 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, March 17, 2021 5 rfc6179, rfc6706 (if 6 approved) 7 Intended status: Standards Track 8 Expires: September 18, 2021 10 Automatic Extended Route Optimization (AERO) 11 draft-templin-intarea-6706bis-95 13 Abstract 15 This document specifies an Automatic Extended Route Optimization 16 (AERO) service for IP internetworking over Overlay Multilink Network 17 (OMNI) interfaces. AERO/OMNI use an IPv6 link-local address format 18 that supports operation of the IPv6 Neighbor Discovery (ND) protocol 19 and links ND to IP forwarding. Prefix delegation/registration 20 services are employed for network admission and to manage the routing 21 system. Multilink operation, mobility management, multicast, quality 22 of service (QoS) signaling and route optimization are naturally 23 supported through dynamic neighbor cache updates. AERO is a widely- 24 applicable mobile internetworking service especially well-suited to 25 aviation services, intelligent transportation systems, mobile Virtual 26 Private Networks (VPNs) and many other applications. 28 Status of This Memo 30 This Internet-Draft is submitted in full conformance with the 31 provisions of BCP 78 and BCP 79. 33 Internet-Drafts are working documents of the Internet Engineering 34 Task Force (IETF). Note that other groups may also distribute 35 working documents as Internet-Drafts. The list of current Internet- 36 Drafts is at https://datatracker.ietf.org/drafts/current/. 38 Internet-Drafts are draft documents valid for a maximum of six months 39 and may be updated, replaced, or obsoleted by other documents at any 40 time. It is inappropriate to use Internet-Drafts as reference 41 material or to cite them other than as "work in progress." 43 This Internet-Draft will expire on September 18, 2021. 45 Copyright Notice 47 Copyright (c) 2021 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (https://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 63 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 64 3. Automatic Extended Route Optimization (AERO) . . . . . . . . 11 65 3.1. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 12 66 3.2. The AERO Service over OMNI Links . . . . . . . . . . . . 13 67 3.2.1. AERO/OMNI Reference Model . . . . . . . . . . . . . . 13 68 3.2.2. Addressing and Node Identification . . . . . . . . . 15 69 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 16 70 3.2.4. OMNI Link Segment Routing . . . . . . . . . . . . . . 18 71 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 22 72 3.2.6. Segment Routing For OMNI Link Selection . . . . . . . 23 73 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 23 74 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 24 75 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 26 76 3.4.1. AERO Server/Relay Behavior . . . . . . . . . . . . . 26 77 3.4.2. AERO Proxy Behavior . . . . . . . . . . . . . . . . . 26 78 3.4.3. AERO Client Behavior . . . . . . . . . . . . . . . . 27 79 3.4.4. AERO Bridge Behavior . . . . . . . . . . . . . . . . 27 80 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 27 81 3.5.1. OMNI Neighbor Interface Attributes . . . . . . . . . 29 82 3.5.2. OMNI Neighbor Advertisement Message Flags . . . . . . 29 83 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 30 84 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 30 85 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 31 86 3.9. OMNI Interface MTU . . . . . . . . . . . . . . . . . . . 31 87 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 32 88 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 33 89 3.10.2. Proxy/Server Forwarding Algorithm . . . . . . . . . 34 90 3.10.3. Bridge Forwarding Algorithm . . . . . . . . . . . . 36 91 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 37 92 3.12. AERO Router Discovery, Prefix Delegation and 93 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 39 94 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 40 95 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 40 96 3.12.3. AERO Server Behavior . . . . . . . . . . . . . . . . 42 97 3.13. The AERO Proxy . . . . . . . . . . . . . . . . . . . . . 45 98 3.13.1. Combined Proxy/Servers . . . . . . . . . . . . . . . 48 99 3.13.2. Detecting and Responding to Server Failures . . . . 48 100 3.13.3. Point-to-Multipoint Server Coordination . . . . . . 49 101 3.14. AERO Address Resolution . . . . . . . . . . . . . . . . . 49 102 3.14.1. Route Optimization Initiation . . . . . . . . . . . 50 103 3.14.2. Relaying the NS(AR) *NET Packet(s) . . . . . . . . . 51 104 3.14.3. Processing the NS(AR) and Sending the NA(AR) . . . . 51 105 3.14.4. Relaying the NA(AR) . . . . . . . . . . . . . . . . 52 106 3.14.5. Processing the NA(AR) . . . . . . . . . . . . . . . 52 107 3.14.6. Route Optimization Maintenance . . . . . . . . . . . 53 108 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 54 109 3.16. Mobility Management and Quality of Service (QoS) . . . . 55 110 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 56 111 3.16.2. Announcing Link-Layer Address and/or QoS Preference 112 Changes . . . . . . . . . . . . . . . . . . . . . . 57 113 3.16.3. Bringing New Links Into Service . . . . . . . . . . 57 114 3.16.4. Deactivating Existing Links . . . . . . . . . . . . 57 115 3.16.5. Moving Between Servers . . . . . . . . . . . . . . . 58 116 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 59 117 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 59 118 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 61 119 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 62 120 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 62 121 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 63 122 3.20. Transition Considerations . . . . . . . . . . . . . . . . 63 123 3.21. Detecting and Reacting to Server and Bridge Failures . . 64 124 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 64 125 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 68 126 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 68 127 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 68 128 6. Security Considerations . . . . . . . . . . . . . . . . . . . 69 129 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 71 130 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 72 131 8.1. Normative References . . . . . . . . . . . . . . . . . . 72 132 8.2. Informative References . . . . . . . . . . . . . . . . . 74 133 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 80 134 A.1. Implementation Strategies for Route Optimization . . . . 80 135 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 81 136 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 81 137 A.4. AERO Critical Infrastructure Considerations . . . . . . . 81 138 A.5. AERO Server Failure Implications . . . . . . . . . . . . 82 139 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 83 140 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 85 141 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 86 143 1. Introduction 145 Automatic Extended Route Optimization (AERO) fulfills the 146 requirements of Distributed Mobility Management (DMM) [RFC7333] and 147 route optimization [RFC5522] for aeronautical networking and other 148 network mobility use cases such as intelligent transportation 149 systems. AERO is an internetworking and mobility management service 150 that employs the Overlay Multilink Network Interface (OMNI) 151 [I-D.templin-6man-omni-interface] Non-Broadcast, Multiple Access 152 (NBMA) virtual link model. The OMNI link is a virtual overlay 153 configured over one or more underlying Internetworks, and nodes on 154 the link can exchange IP packets as single-hop neighbors via 155 encapsulation. The OMNI Adaptation Layer (OAL) supports multilink 156 operation for increased reliability, bandwidth optimization and 157 traffic path selection while accommodating Maximum Transmission Unit 158 (MTU) diversity. 160 The AERO service comprises Clients, Proxys, Servers and Relays that 161 are seen as OMNI link neighbors as well as Bridges that interconnect 162 OMNI link segments through OAL forwarding at a layer below IP. Each 163 node's OMNI interface uses an IPv6 link-local address format that 164 supports operation of the IPv6 Neighbor Discovery (ND) protocol 165 [RFC4861] and links ND to IP forwarding. A node's OMNI interface can 166 be configured over multiple underlying interfaces, and therefore 167 appears as a single interface with multiple link-layer addresses. 168 Each link-layer address is subject to change due to mobility and/or 169 QoS fluctuations, and link-layer address changes are signaled by ND 170 messaging the same as for any IPv6 link. 172 AERO provides a cloud-based service where mobile node Clients may use 173 any Server acting as a Mobility Anchor Point (MAP) and fixed nodes 174 may use any Relay on the link for efficient communications. Fixed 175 nodes forward packets destined to other AERO nodes via the nearest 176 Relay, which forwards them through the cloud. A mobile node's 177 initial packets are forwarded through the Server, while direct 178 routing is supported through automatic extended route optimization 179 while data packets are flowing. Both unicast and multicast 180 communications are supported, and mobile nodes may efficiently move 181 between locations while maintaining continuous communications with 182 correspondents and without changing their IP Address. 184 AERO Bridges are interconnected in a secured private BGP overlay 185 routing instance to provide an OAL routing/bridging service that 186 joins the underlying Internetworks of multiple disjoint 187 administrative domains into a single unified OMNI link at a layer 188 below IP. Each OMNI link instance is characterized by the set of 189 Mobility Service Prefixes (MSPs) common to all mobile nodes. The 190 link extends to the point where a Relay/Server is on the optimal 191 route from any correspondent node on the link, and provides a conduit 192 between the underlying Internetwork and the OMNI link. To the 193 underlying Internetwork, the Relay/Server is the source of a route to 194 the MSP, and hence uplink traffic to the mobile node is naturally 195 routed to the nearest Relay/Server. 197 AERO can be used with OMNI links that span private-use Internetworks 198 and/or public Internetworks such as the global Internet. In the 199 latter case, some end systems may be located behind global Internet 200 Network Address Translators (NATs). A means for robust traversal of 201 NATs while avoiding "triangle routing" is therefore provided. 203 AERO assumes the use of PIM Sparse Mode in support of multicast 204 communication. In support of Source Specific Multicast (SSM) when a 205 Mobile Node is the source, AERO route optimization ensures that a 206 shortest-path multicast tree is established with provisions for 207 mobility and multilink operation. In all other multicast scenarios 208 there are no AERO dependencies. 210 AERO was designed for aeronautical networking for both manned and 211 unmanned aircraft, where the aircraft is treated as a mobile node 212 that can connect an Internet of Things (IoT). AERO is also 213 applicable to a wide variety of other use cases. For example, it can 214 be used to coordinate the links of mobile nodes (e.g., cellphones, 215 tablets, laptop computers, etc.) that connect into a home enterprise 216 network via public access networks using tunneling software such as 217 OpenVPN [OVPN] with VPN or non-VPN services enabled according to the 218 appropriate security model. AERO can also be used to facilitate 219 terrestrial vehicular and urban air mobility (as well as pedestrian 220 communication services) for future intelligent transportation systems 221 [I-D.ietf-ipwave-vehicular-networking][I-D.templin-ipwave-uam-its]. 222 Other applicable use cases are also in scope. 224 The following numbered sections present the AERO specification. The 225 appendices at the end of the document are non-normative. 227 2. Terminology 229 The terminology in the normative references applies; especially, the 230 terminology in the OMNI specification 231 [I-D.templin-6man-omni-interface] is used extensively throughout. 232 The following terms are defined within the scope of this document: 234 IPv6 Neighbor Discovery (ND) 235 an IPv6 control message service for coordinating neighbor 236 relationships between nodes connected to a common link. AERO uses 237 the ND service specified in [RFC4861]. 239 IPv6 Prefix Delegation 240 a networking service for delegating IPv6 prefixes to nodes on the 241 link. The nominal service is DHCPv6 [RFC8415], however alternate 242 services (e.g., based on ND messaging) are also in scope. Most 243 notably, a minimal form of prefix delegation known as "prefix 244 registration" can be used if the Client knows its prefix in 245 advance and can represent it in the IPv6 source address of an ND 246 message. 248 Access Network (ANET) 249 a node's first-hop data link service network (e.g., a radio access 250 network, cellular service provider network, corporate enterprise 251 network, etc.) that often provides link-layer security services 252 such as IEEE 802.1X and physical-layer security (e.g., "protected 253 spectrum") to prevent unauthorized access internally and with 254 border network-layer security services such as firewalls and 255 proxys that prevent unauthorized outside access. 257 ANET interface 258 a node's attachment to a link in an ANET. 260 Internetwork (INET) 261 a connected IP network topology with a coherent routing and 262 addressing plan and that provides a transit backbone service for 263 ANET end systems. INETs also provide an underlay service over 264 which the AERO virtual link is configured. Example INETs include 265 corporate enterprise networks, aviation networks, and the public 266 Internet itself. When there is no administrative boundary between 267 an ANET and the INET, the ANET and INET are one and the same. 269 INET interface 270 a node's attachment to a link in an INET. 272 *NET 273 a "wildcard" term referring to either ANET or INET when it is not 274 necessary to draw a distinction between the two. 276 *NET interface 277 a node's attachment to a link in a *NET. 279 *NET Partition 280 frequently, *NETs such as large corporate enterprise networks are 281 sub-divided internally into separate isolated partitions. Each 282 partition is fully connected internally but disconnected from 283 other partitions, and there is no requirement that separate 284 partitions maintain consistent Internet Protocol and/or addressing 285 plans. (Each *NET partition is seen as a separate OMNI link 286 segment as discussed below.) 288 *NET address 289 an IP address assigned to a node's interface connection to a *NET. 291 *NET encapsulation 292 the encapsulation of a packet in an outer header or headers that 293 can be routed within the scope of the local *NET partition. 295 OMNI link 296 the same as defined in [I-D.templin-6man-omni-interface], and 297 manifested by IPv6 encapsulation [RFC2473]. The OMNI link spans 298 underlying INET segments joined by virtual bridges in a spanning 299 tree the same as a bridged campus LAN. AERO nodes on the OMNI 300 link appear as single-hop neighbors even though they may be 301 separated by multiple underlying INET hops, and can use Segment 302 Routing [RFC8402] to cause packets to visit selected waypoints on 303 the link. 305 OMNI Interface 306 a node's attachment to an OMNI link. Since the addresses assigned 307 to an OMNI interface are managed for uniqueness, OMNI interfaces 308 do not require Duplicate Address Detection (DAD) and therefore set 309 the administrative variable 'DupAddrDetectTransmits' to zero 310 [RFC4862]. 312 OMNI Adaptation Layer (OAL) 313 an OMNI interface process whereby original IP packets admitted 314 into the interface are wrapped in a mid-layer IPv6 header and 315 subject to fragmentation and reassembly. The OAL is also 316 responsible for generating MTU-related control messages as 317 necessary, and for providing addressing context for spanning 318 multiple segments of a bridged OMNI link. 320 original (or, "inner") IP packet 321 a whole IP packet or fragment admitted into the OMNI interface by 322 the network layer prior to OAL encapsulation and fragmentation, or 323 an IP packet delivered to the network layer by the OMNI interface 324 following OAL decapsulation and reassembly. 326 OAL packet 327 an original IP packet encapsulated in OAL headers and trailers 328 before OAL fragmentation, or following OAL reassembly. 330 OAL fragment 331 a portion of an OAL packet following fragmentation but prior to 332 *NET encapsulation, or following *NET encapsulation but prior to 333 OAL reassembly. 335 *NET packet 336 an OAL fragment following *NET encapsulation or prior to *NET 337 decapsulation. OAL sources and destinations exchange *NET packets 338 over underlying interfaces, and may be separated by one or more 339 OAL intermediate nodes. 341 OAL source 342 an OMNI interface acts as an OAL source when it encapsulates 343 original IP packets to form OAL packets, then performs OAL 344 fragmentation and *NET encapsulation. 346 OAL destination 347 an OMNI interface acts as an OAL destination when it decapsulates 348 *NET packets, then performs OAL reassembly and decapsulation to 349 derive the original IP packet. 351 underlying interface 352 a *NET interface over which an OMNI interface is configured. 354 Mobility Service Prefix (MSP) 355 an aggregated IP Global Unicast Address (GUA) prefix (e.g., 356 2001:db8::/32, 192.0.2.0/24, etc.) assigned to the OMNI link and 357 from which more-specific Mobile Network Prefixes (MNPs) are 358 delegated. OMNI link administrators typically obtain MSPs from an 359 Internet address registry, however private-use prefixes can 360 alternatively be used subject to certain limitations (see: 361 [I-D.templin-6man-omni-interface]). OMNI links that connect to 362 the global Internet advertise their MSPs to their interdomain 363 routing peers. 365 Mobile Network Prefix (MNP) 366 a longer IP prefix delegated from an MSP (e.g., 367 2001:db8:1000:2000::/56, 192.0.2.8/30, etc.) and delegated to an 368 AERO Client or Relay. 370 Mobile Network Prefix Link Local Address (MNP-LLA) 371 an IPv6 Link Local Address that embeds the most significant 64 372 bits of an MNP in the lower 64 bits of fe80::/64, as specified in 373 [I-D.templin-6man-omni-interface]. 375 Mobile Network Prefix Unique Local Address (MNP-ULA) 376 an IPv6 Unique-Local Address derived from an MNP-LLA. 378 Administrative Link Local Address (ADM-LLA) 379 an IPv6 Link Local Address that embeds a 32-bit administratively- 380 assigned identification value in the lower 32 bits of fe80::/96, 381 as specified in [I-D.templin-6man-omni-interface]. 383 Administrative Unique Local Address (ADM-ULA) 384 an IPv6 Unique-Local Address derived from an ADM-LLA. 386 AERO node 387 a node that is connected to an OMNI link and participates in the 388 AERO internetworking and mobility service. 390 AERO Client ("Client") 391 an AERO node that connects over one or more underlying interfaces 392 and requests MNP delegation/registration service from AERO 393 Servers. The Client assigns an MNP-LLA to the OMNI interface for 394 use in ND exchanges with other AERO nodes and forwards packets to 395 correspondents according to OMNI interface neighbor cache state. 397 AERO Server ("Server") 398 an INET node that configures an OMNI interface to provide default 399 forwarding and Mobility Anchor Point (MAP) services for AERO 400 Clients. The Server assigns an ADM-LLA to its OMNI interface to 401 support the operation of the ND services, and advertises all of 402 its associated MNPs via BGP peerings with Bridges. 404 AERO Relay ("Relay") 405 an AERO Server that also provides forwarding services between 406 nodes reached via the OMNI link and correspondents on other links. 407 AERO Relays are provisioned with MNPs (i.e., the same as for an 408 AERO Client) and run a dynamic routing protocol to discover any 409 non-MNP IP GUA routes in service on its connected INET links. In 410 both cases, the Relay advertises the MSP(s) to its downstream 411 networks, and distributes all of its associated MNPs and non-MNP 412 IP GUA routes via BGP peerings with Bridges (i.e., the same as for 413 an AERO Server). 415 AERO Bridge ("Bridge") 416 a node that provides hybrid routing/bridging services (as well as 417 a security trust anchor) for nodes on an OMNI link. The Bridge 418 forwards packets between OMNI link segments at the OAL layer 419 without decrementing the network layer IP TTL/Hop Limit. AERO 420 Bridges peer with Servers and other Bridges to discover the full 421 set of MNPs for the link as well as any non-MNP IP GUA routes that 422 are reachable via Relays. 424 AERO Proxy ("Proxy") 425 a node that provides proxying services between Clients in an ANET 426 and Servers in external INETs. The AERO Proxy is a conduit 427 between the ANET and external INETs in the same manner as for 428 common web proxys, and behaves in a similar fashion as for ND 429 proxys [RFC4389]. A node may be configured to act as either a 430 Proxy and/or a Server, depending on Client Server selection 431 criteria. 433 ingress tunnel endpoint (ITE) 434 an OMNI interface endpoint that injects encapsulated packets into 435 an OMNI link. 437 egress tunnel endpoint (ETE) 438 an OMNI interface endpoint that receives encapsulated packets from 439 an OMNI link. 441 link-layer address 442 an IP address used as an encapsulation header source or 443 destination address from the perspective of the OMNI interface. 444 When an upper layer protocol (e.g., UDP) is used as part of the 445 encapsulation, the port number is also considered as part of the 446 link-layer address. 448 network layer address 449 the source or destination address of an encapsulated IP packet 450 presented to the OMNI interface. 452 end user network (EUN) 453 an internal virtual or external edge IP network that an AERO 454 Client or Relay connects to the rest of the network via the OMNI 455 interface. The Client/Relay sees each EUN as a "downstream" 456 network, and sees the OMNI interface as the point of attachment to 457 the "upstream" network. 459 Mobile Node (MN) 460 an AERO Client and all of its downstream-attached networks that 461 move together as a single unit, i.e., an end system that connects 462 an Internet of Things. 464 Mobile Router (MR) 465 a MN's on-board router that forwards packets between any 466 downstream-attached networks and the OMNI link. The MR is the MN 467 entity that hosts the AERO Client. 469 Route Optimization Source (ROS) 470 the AERO node nearest the source that initiates route 471 optimization. The ROS may be a Server or Proxy acting on behalf 472 of the source Client, or may be the Client itself if the Client is 473 connected to the INET either directly or through a NAT. 475 Route Optimization responder (ROR) 476 the AERO node nearest the target destination that responds to 477 route optimization requests. The ROR may be a Proxy/Server acting 478 on behalf of a target MNP Client, a Relay for a non-MNP 479 destination or may be the target Client itself. 481 MAP List 482 a geographically and/or topologically referenced list of addresses 483 of all Servers within the same OMNI link. There is a single MAP 484 list for the entire OMNI link. 486 Distributed Mobility Management (DMM) 487 a BGP-based overlay routing service coordinated by Servers and 488 Bridges that tracks all Server-to-Client associations. 490 Mobility Service (MS) 491 the collective set of all Servers, Proxys, Bridges and Relays that 492 provide the AERO Service to Clients. 494 Mobility Service Endpoint MSE) 495 an individual Server, Proxy, Bridge or Relay in the Mobility 496 Service. 498 Throughout the document, the simple terms "Client", "Server", 499 "Bridge", "Proxy" and "Relay" refer to "AERO Client", "AERO Server", 500 "AERO Bridge", "AERO Proxy" and "AERO Relay", respectively. 501 Capitalization is used to distinguish these terms from other common 502 Internetworking uses in which they appear without capitalization. 504 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 505 the names of node variables, messages and protocol constants) is used 506 throughout this document. The terms "All-Routers multicast", "All- 507 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 508 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 509 generically refer to either Internet Protocol version, i.e., IPv4 510 [RFC0791] or IPv6 [RFC8200]. 512 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 513 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 514 "OPTIONAL" in this document are to be interpreted as described in BCP 515 14 [RFC2119][RFC8174] when, and only when, they appear in all 516 capitals, as shown here. 518 3. Automatic Extended Route Optimization (AERO) 520 The following sections specify the operation of IP over OMNI links 521 using the AERO service: 523 3.1. AERO Node Types 525 AERO Clients are Mobile Nodes (MNs) that configure OMNI interfaces 526 over underlying interfaces with addresses that may change when the 527 Client moves to a new network connection point. AERO Clients 528 register their Mobile Network Prefixes (MNPs) with the AERO service, 529 and distribute the MNPs to nodes on EUNs. AERO Bridges, Servers, 530 Proxys and Relays are critical infrastructure elements in fixed 531 (i.e., non-mobile) INET deployments and hence have permanent and 532 unchanging INET addresses. Together, they constitute the AERO 533 service which provides an OMNI link virtual overlay for connecting 534 AERO Clients. 536 AERO Bridges provide hybrid routing/bridging services (as well as a 537 security trust anchor) for nodes on an OMNI link. Bridges use 538 standard IPv6 routing to forward packets both within the same *NET 539 partition and between disjoint *NET partitions based on an IPv6 540 encapsulation mid-layer known as the OMNI Adaptation Layer (OAL) 541 [I-D.templin-6man-omni-interface]. During forwarding, the inner IP 542 layer experiences a virtual bridging service since the inner IP TTL/ 543 Hop Limit is not decremented. Each Bridge also peers with Servers 544 and other Bridges in a dynamic routing protocol instance to provide a 545 Distributed Mobility Management (DMM) service for the list of active 546 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 547 one or more Mobility Service Prefixes (MSPs) and configure secured 548 tunnels with Servers, Relays, Proxys and other Bridges; they further 549 maintain IP forwarding table entries for each MNP and any other 550 reachable non-MNP prefixes. 552 AERO Servers in distributed INET locations provide default forwarding 553 and mobility/multilink services for AERO Client Mobile Nodes (MNs). 554 Each Server also peers with Bridges in a dynamic routing protocol 555 instance to advertise its list of associated MNPs (see 556 Section 3.2.3). Servers facilitate prefix delegation/registration 557 exchanges with Clients, where each delegated prefix becomes an MNP 558 taken from an MSP. Servers forward packets between OMNI interface 559 neighbors and track each Client's mobility profiles. Servers may 560 further act as Servers for some sets of Clients and as Proxys for 561 others. 563 AERO Proxys provide a conduit for ANET Clients to associate with 564 Servers in external INETs. Client and Servers exchange control plane 565 messages via the Proxy acting as a bridge between the ANET/INET 566 boundary. The Proxy forwards data packets between Clients and the 567 OMNI link according to forwarding information in the neighbor cache. 568 The Proxy function is specified in Section 3.13. Proxys may further 569 act as Proxys for some sets of Clients and as Servers for others. 571 AERO Relays are Servers that provide forwarding services between the 572 OMNI interface and INET/EUN interfaces. Relays are provisioned with 573 MNPs the same as for an AERO Client, and also run a dynamic routing 574 protocol to discover any non-MNP IP routes. The Relay advertises the 575 MSP(s) to its connected networks, and distributes all of its 576 associated MNPs and non-MNP IP GUA routes via BGP peerings with 577 Bridges 579 3.2. The AERO Service over OMNI Links 581 3.2.1. AERO/OMNI Reference Model 583 Figure 1 presents the basic OMNI link reference model: 585 +----------------+ 586 | AERO Bridge B1 | 587 | Nbr: S1, S2, P1| 588 |(X1->S1; X2->S2)| 589 | MSP M1 | 590 +-+---------+--+-+ 591 +--------------+ | Secured | | +--------------+ 592 |AERO Server S1| | tunnels | | |AERO Server S2| 593 | Nbr: C1, B1 +-----+ | +-----+ Nbr: C2, B1 | 594 | default->B1 | | | default->B1 | 595 | X1->C1 | | | X2->C2 | 596 +-------+------+ | +------+-------+ 597 | OMNI link | | 598 X===+===+===================+==)===============+===+===X 599 | | | | 600 +-----+--------+ +--------+--+-----+ +--------+-----+ 601 |AERO Client C1| | AERO Proxy P1 | |AERO Client C2| 602 | Nbr: S1 | |(Proxy Nbr Cache)| | Nbr: S2 | 603 | default->S1 | +--------+--------+ | default->S2 | 604 | MNP X1 | | | MNP X2 | 605 +------+-------+ .--------+------. +-----+--------+ 606 | (- Proxyed Clients -) | 607 .-. `---------------' .-. 608 ,-( _)-. ,-( _)-. 609 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 610 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 611 `-(______)-' +-------+ +-------+ `-(______)-' 613 Figure 1: AERO/OMNI Reference Model 615 In this model: 617 o the OMNI link is an overlay network service configured over one or 618 more underlying *NET partitions which may be managed by different 619 administrative authorities and have incompatible protocols and/or 620 addressing plans. 622 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 623 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 624 via BGP peerings over secured tunnels to Servers (S1, S2). 625 Bridges connect the disjoint segments of a partitioned OMNI link. 627 o AERO Servers/Relays S1 and S2 configure secured tunnels with 628 Bridge B1 and also provide mobility, multilink and default router 629 services for their associated Clients C1 and C2. 631 o AERO Clients C1 and C2 associate with Servers S1 and S2, 632 respectively. They receive Mobile Network Prefix (MNP) 633 delegations X1 and X2, and also act as default routers for their 634 associated physical or internal virtual EUNs. Simple hosts H1 and 635 H2 attach to the EUNs served by Clients C1 and C2, respectively. 637 o AERO Proxy P1 configures a secured tunnel with Bridge B1 and 638 provides proxy services for AERO Clients in secured enclaves that 639 cannot associate directly with other OMNI link neighbors. 641 An OMNI link configured over a single *NET appears as a single 642 unified link with a consistent underlying network addressing plan. 643 In that case, all nodes on the link can exchange packets via simple 644 *NET encapsulation, since the underlying *NET is connected. In 645 common practice, however, an OMNI link may be partitioned into 646 multiple "segments", where each segment is a distinct *NET 647 potentially managed under a different administrative authority (e.g., 648 as for worldwide aviation service providers such as ARINC, SITA, 649 Inmarsat, etc.). Individual *NETs may also themselves be partitioned 650 internally, in which case each internal partition is seen as a 651 separate segment. 653 The addressing plan of each segment is consistent internally but will 654 often bear no relation to the addressing plans of other segments. 655 Each segment is also likely to be separated from others by network 656 security devices (e.g., firewalls, proxys, packet filtering gateways, 657 etc.), and in many cases disjoint segments may not even have any 658 common physical link connections. Therefore, nodes can only be 659 assured of exchanging packets directly with correspondents in the 660 same segment, and not with those in other segments. The only means 661 for joining the segments therefore is through inter-domain peerings 662 between AERO Bridges. 664 The same as for traditional campus LANs, multiple OMNI link segments 665 can be joined into a single unified link via a virtual bridging 666 service using the OMNI Adaptation Layer (OAL) 668 [I-D.templin-6man-omni-interface] which inserts a mid-layer IPv6 669 encapsulation header that supports inter-segment forwarding (i.e., 670 bridging) without decrementing the network-layer TTL/Hop Limit. This 671 bridging of OMNI link segments is shown in Figure 2: 673 . . . . . . . . . . . . . . . . . . . . . . . 674 . . 675 . .-(::::::::) . 676 . .-(::::::::::::)-. +-+ . 677 . (:::: Segment A :::)--|B|---+ . 678 . `-(::::::::::::)-' +-+ | . 679 . `-(::::::)-' | . 680 . | . 681 . .-(::::::::) | . 682 . .-(::::::::::::)-. +-+ | . 683 . (:::: Segment B :::)--|B|---+ . 684 . `-(::::::::::::)-' +-+ | . 685 . `-(::::::)-' | . 686 . | . 687 . .-(::::::::) | . 688 . .-(::::::::::::)-. +-+ | . 689 . (:::: Segment C :::)--|B|---+ . 690 . `-(::::::::::::)-' +-+ | . 691 . `-(::::::)-' | . 692 . | . 693 . ..(etc).. x . 694 . . 695 . . 696 . <- OMNI link Bridged by encapsulation -> . 697 . . . . . . . . . . . . . .. . . . . . . . . 699 Figure 2: Bridging OMNI Link Segments 701 Bridges, Servers, Relays and Proxys connect via secured INET tunnels 702 over their respective segments in a spanning tree topology rooted at 703 the Bridges. The secured spanning tree supports strong 704 authentication for IPv6 ND control messages and may also be used to 705 convey the initial data packets in a flow. Route optimization can 706 then be employed to cause data packets to take more direct paths 707 between OMNI link neighbors without having to strictly follow the 708 spanning tree. 710 3.2.2. Addressing and Node Identification 712 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 713 fe80::/64 [RFC4291] to assign LLAs used for network-layer addresses 714 in link-scoped IPv6 ND and data messages. AERO Clients use LLAs 715 constructed from MNPs (i.e., "MNP-LLAs") while other AERO nodes use 716 LLAs constructed from administrative identification values ("ADM- 717 LLAs") as specified in [I-D.templin-6man-omni-interface]. 719 AERO nodes also use the Unique Local Address (ULA) prefix fd00::/8 720 followed by a pseudo-random 40-bit OMNI domain identifier to form the 721 prefix [ULA]::/48, then include a 16-bit OMNI link identifier '*' to 722 form the prefix [ULA*]::/64 [RFC4291]. The AERO node then uses the 723 prefix [ULA*]::/64 to form "MNP-ULAs" or "ADM-ULA"s as specified in 724 [I-D.templin-6man-omni-interface] to support OAL addressing. AERO 725 Clients also use Temporary ULAs constructed per 726 [I-D.templin-6man-omni-interface], where the addresses are typically 727 used only in initial control message exchanges until a stable MNP- 728 LLA/ULA is assigned. 730 AERO MSPs and MNPs are typically based on Global Unicast Addresses 731 (GUAs), but in some cases may be based on private-use addresses. See 732 [I-D.templin-6man-omni-interface] for a full specification of LLAs, 733 ULAs and GUAs used by AERO nodes on OMNI links. 735 Finally, AERO Clients and Servers configure node identification 736 values as specified in [I-D.templin-6man-omni-interface]. 738 3.2.3. AERO Routing System 740 The AERO routing system comprises a private instance of the Border 741 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 742 and Servers and does not interact with either the public Internet BGP 743 routing system or any underlying INET routing systems. 745 In a reference deployment, each Server is configured as an Autonomous 746 System Border Router (ASBR) for a stub Autonomous System (AS) using a 747 32-bit AS Number (ASN) [RFC4271] that is unique within the BGP 748 instance, and each Server further uses eBGP to peer with one or more 749 Bridges but does not peer with other Servers. Each *NET of a multi- 750 segment OMNI link must include one or more Bridges, which peer with 751 the Servers and Proxys within that *NET. All Bridges within the same 752 *NET are members of the same hub AS, and use iBGP to maintain a 753 consistent view of all active MNPs currently in service. The Bridges 754 of different *NETs peer with one another using eBGP. 756 Bridges advertise the OMNI link's MSPs and any non-MNP routes to each 757 of their Servers. This means that any aggregated non-MNPs (including 758 "default") are advertised to all Servers. Each Bridge configures a 759 black-hole route for each of its MSPs. By black-holing the MSPs, the 760 Bridge will maintain forwarding table entries only for the MNPs that 761 are currently active, and packets destined to all other MNPs will 762 correctly incur Destination Unreachable messages due to the black- 763 hole route. In this way, Servers have only partial topology 764 knowledge (i.e., they know only about the MNPs of their directly 765 associated Clients) and they forward all other packets to Bridges 766 which have full topology knowledge. 768 Each OMNI link segment assigns a unique ADM-ULA sub-prefix of 769 [ULA*]::/96. For example, a first segment could assign 770 [ULA*]::1000/116, a second could assign [ULA*]::2000/116, a third 771 could assign [ULA*]::3000/116, etc. Within each segment, each Proxy/ 772 Server configures an ADM-ULA within the segment's prefix, e.g., the 773 Proxy/Servers within [ULA*]::2000/116 could assign the ADM-ULAs 774 [ULA*]::2011/116, [ULA*]::2026/116, [ULA*]::2003/116, etc. 776 The administrative authorities for each segment must therefore 777 coordinate to assure mutually-exclusive ADM-ULA prefix assignments, 778 but internal provisioning of ADM-ULAs an independent local 779 consideration for each administrative authority. For each ADM-ULA 780 prefix, the Bridge(s) that connect that segment assign the all-zero's 781 address of the prefix as a Subnet Router Anycast address. For 782 example, the Subnet Router Anycast address for [ULA*]::1023/116 is 783 simply [ULA*]::1000. 785 ADM-ULA prefixes are statically represented in Bridge forwarding 786 tables. Bridges join multiple segments into a unified OMNI link over 787 multiple diverse administrative domains. They support a bridging 788 function by first establishing forwarding table entries for their 789 ADM-ULA prefixes either via standard BGP routing or static routes. 790 For example, if three Bridges ('A', 'B' and 'C') from different 791 segments serviced [ULA*]::1000/116, [ULA*]::2000/116 and 792 [ULA*]::3000/116 respectively, then the forwarding tables in each 793 Bridge are as follows: 795 A: [ULA*]::1000/116->local, [ULA*]::2000/116->B, [ULA*]::3000/116->C 797 B: [ULA*]::1000/116->A, [ULA*]::2000/116->local, [ULA*]::3000/116->C 799 C: [ULA*]::1000/116->A, [ULA*]::2000/116->B, [ULA*]::3000/116->local 801 These forwarding table entries are permanent and never change, since 802 they correspond to fixed infrastructure elements in their respective 803 segments. 805 MNP ULAs are instead dynamically advertised in the AERO routing 806 system by Servers and Relays that provide service for their 807 corresponding MNPs. For example, if three Servers ('D', 'E' and 'F') 808 service the MNPs 2001:db8:1000:2000::/56, 2001:db8:3000:4000::/56 and 809 2001:db8:5000:6000::/56 then the routing system would include: 811 D: [ULA*]:2001:db8:1000:2000/120 812 E: [ULA*]:2001:db8:3000:4000/120 814 F: [ULA*]:2001:db8:5000:6000/120 816 A full discussion of the BGP-based routing system used by AERO is 817 found in [I-D.ietf-rtgwg-atn-bgp]. 819 3.2.4. OMNI Link Segment Routing 821 With the Client and partition prefixes in place in Bridge forwarding 822 tables, the OMNI interface sends control and data packets toward AERO 823 destination nodes located in different OMNI link segments over the 824 spanning tree. The OMNI interface uses the OMNI Adaptation Layer 825 (OAL) encapsulation service [I-D.templin-6man-omni-interface], and 826 includes an OMNI Routing Header (ORH) as an extension to the OAL 827 header if final segment forwarding information is available, e.g., in 828 the neighbor cache. (For nodes located in the same OMNI link 829 segment, or when no final segment forwarding information is 830 available, the ORH may be omitted.) The ORH is formatted as shown in 831 Figure 3: 833 0 1 2 3 834 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 1 835 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 836 | Next Header | Hdr Ext Len | Routing Type | SRT | FMT | 837 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 838 | Last Hop Segment-id (LHS) | 839 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 840 ~ Link Layer Address (L2ADDR) ~ 841 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 842 ~ Destination Suffix (if necessary) ~ 843 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 844 ~ Null Padding (if necessary) ~ 845 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 847 Figure 3: OMNI Routing Header (ORH) Format 849 In this format: 851 o Next Header identifies the type of header immediately following 852 the ORH. 854 o Hdr Ext Len is the length of the Routing header in 8-octet units 855 (not including the first 8 octets), with trailing padding added if 856 necessary to produce an integral number of 8-octet units. 858 o Routing Type is set to TBD1 (see IANA Considerations). 860 o Segments Left is omitted, and replaced by a 5-bit SRT and 3-bit 861 FMT field. 863 o SRT - a 5-bit Segment Routing Topology prefix length value that 864 (when added to 96) determines the prefix length to apply to the 865 ADM-ULA formed from concatenating [ULA*]::/96 with the 32 bit LHS 866 value that follows (for example, the value 16 corresponds to the 867 prefix length 112). 869 o FMT - a 3-bit "Framework/Mode/Type" code corresponding to the 870 included Link Layer Address as follows: 872 * When the most significant bit (i.e., "Framework") is set to 1, 873 L2ADDR is the *NET encapsulation address for the target Client 874 itself; otherwise L2ADDR is the address of the Server/Proxy 875 named in the LHS. 877 * When the next most significant bit (i.e., "Mode") is set to 1, 878 the Framework node is (likely) located behind a *NET Network 879 Address Translator (NAT); otherwise, it is on the open *NET. 881 * When the least significant bit (i.e., "Type") is set to 0, 882 L2ADDR includes a UDP Port Number followed by an IPv4 address; 883 otherwise, it includes a UDP Port Number followed by an IPv6 884 address. 886 o LHS - the 32 bit ID of a node in the Last Hop Segment that 887 services the target. When SRT and LHS are both set to 0, the LHS 888 is considered unspecified. When SRT is set to 0 and LHS is non- 889 zero, the prefix length is set to 128. SRT and LHS provide 890 guidance to the OMNI interface forwarding algorithm. 891 Specifically, if SRT/LHS is located in the local OMNI link 892 segment, the OAL source can omit the ORH and (following any 893 necessary NAT traversal messaging) send directly to the OAL 894 destination according to FMT/L2ADDR. Otherwise, it includes the 895 ORH and forwards according to the OMNI link spanning tree. 897 o Link Layer Address (L2ADDR) - Formatted according to FMT, and 898 identifies the link-layer address (i.e., the encapsulation 899 address) of the target. The UDP Port Number appears in the first 900 two octets and the IP address appears in the next 4 octets for 901 IPv4 or 16 octets for IPv6. The Port Number and IP address are 902 recorded in network byte order, and in ones-compliment 903 "obfuscated" form per [RFC4380]. The OMNI interface forwarding 904 algorithm uses FMT/L2ADDR to determine the *NET encapsulation 905 address for local forwarding when SRT/LHS is located in the same 906 OMNI link segment. Note that if the target is behind a NAT, 907 L2ADDR will contain the mapped *NET address stored in the NAT; 908 otherwise, L2ADDR will contain the native *NET information of the 909 target itself. 911 o Destination Suffix is a 64-bit field included only for OAL non- 912 first-fragments. Present only when Hdr Ext Len indicates that at 913 least 8 bytes follow L2ADDR. When present, encodes the 64-bit 914 MNP-ULA suffix for the target Client. 916 o Null Padding contains zero-valued octets as necessary to pad the 917 ORH to an integral number of 8-octet units. 919 AERO neighbors use OAL encapsulation and fragmentation to exchange 920 OAL packets as specified in [I-D.templin-6man-omni-interface]. When 921 an AERO node's OMNI interface (acting as an OAL source) uses OAL 922 encapsulation for an original IP packet with source address 923 2001:db8:1:2::1 and destination address 2001:db8:1234:5678::1, it 924 sets the OAL header source address to its own ULA (e.g., 925 [ULA*]::2001:db8:1:2), sets the destination address to the MNP-ULA 926 corresponding to the IP destination address (e.g., 927 [ULA*]::2001:db8:1234:5678), selects an Identification and appends an 928 OAL checksum. 930 If the neighbor cache information indicates that the target is in a 931 different segment, the OAL source next inserts an ORH immediately 932 following the OAL header while including the correct SRT, FMT, LHS, 933 L2ADDR and Destination Suffix if fragmentation if needed (in this 934 case, the Destination Suffix is 2001:db8:1234:5678). Next, the OAL 935 source overwrites the OAL header destination address with the LHS 936 Subnet Router Anycast address (for example, for LHS 3000:4567 with 937 SRT 16, the Subnet Router Anycast address is [ULA*]::3000:0000). 938 (Note: if the ADM-ULA of the last-hop Proxy/Server is known but the 939 SRT, FMT, LHS and L2ADDR are not (yet) known, the OAL source instead 940 overwrites the OAL header destination address with the ADM-ULA.) 942 The OAL source then fragments the OAL packet, with each resulting OAL 943 fragment including the OAL/ORH headers while only the first fragment 944 includes the original IPv6 header. (Note that the packet is prepared 945 as an "atomic fragment" even if no actual fragmentation was 946 required.) The OAL source finally encapsulates each resulting OAL 947 fragment in an *NET header to form a *NET packet, with source address 948 set to its own *NET address (e.g., 192.0.2.100) and destination set 949 to the *NET address of a Bridge (e.g., 192.0.2.1). 951 The *NET packet encapsulation format in the above example is shown in 952 Figure 4: 954 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 955 | *NET Header | 956 | src = 192.0.2.100 | 957 | dst = 192.0.2.1 | 958 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 959 | OAL IPv6 Header | 960 | src = [ULA*]::2001:db8:1:2 | 961 | dst= [ULA*]::3000:0000 | 962 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 963 | ORH (if necessary) 964 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 965 | OAL Fragment Header | 966 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 967 | Inner IP Header | 968 | (first-fragment only) ) 969 | src = 2001:db8:1:2::1 | 970 | dst = 2001:db8:1234:5678::1 | 971 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 972 | | 973 ~ ~ 974 ~ Inner Packet Body/Fragment ~ 975 ~ ~ 976 | | 977 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 979 Figure 4: *NET Packet 981 In this format, the inner IP header and packet body/fragment are from 982 the original IP packet, the OAL header is an IPv6 header prepared 983 according to [RFC2473], the ORH is a Routing Header extension of the 984 OAL header the Fragment Header identifies each fragment, and the INET 985 header is prepared as discussed in Section 3.6. When the OAL source 986 transmits the resulting *NET packets, they are forwarded over 987 possibly multiple OMNI link spanning tree hops until they arrive at 988 the OAL destination. 990 This gives rise to a routing system that contains both Client MNP-ULA 991 routes that may change dynamically due to regional node mobility and 992 per-partition ADM-ULA routes that rarely if ever change. The Bridges 993 can therefore provide link-layer bridging by sending packets over the 994 spanning tree instead of network-layer routing according to MNP 995 routes. As a result, opportunities for packet loss due to node 996 mobility between different segments are mitigated. 998 In normal operations, IPv6 ND messages are conveyed over secured 999 paths between OMNI link neighbors so that specific Proxys, Servers or 1000 Relays can be addressed without being subject to mobility events. 1001 Conversely, only the first few packets destined to Clients need to 1002 traverse secured paths until route optimization can determine a more 1003 direct path. 1005 Note: When the OAL source and destination are on the same INET 1006 segment, the ORH is not needed and OAL header compression can be used 1007 to significantly reduce encapsulation overhead 1008 [I-D.templin-6man-omni-interface]. 1010 Note: When the OAL source has multiple inner IP packets to send to 1011 the same OAL destination, it can perform "packing" to generate a 1012 "super-packet" [I-D.templin-6man-omni-interface]. In that case, the 1013 OAL/ORH super-packet may include up to N inner packets as long as the 1014 total length of the super-packet does not exceed the OMNI interface 1015 MTU. 1017 Note: Use of an IPv6 "minimal encapsulation" format (i.e., an IPv6 1018 variant of [RFC2004]) based on extensions to the ORH was considered 1019 and abandoned. In the approach, the ORH would be inserted as an 1020 extension header to the original IPv6 packet header. The IPv6 1021 destination address would then be written into the ORH, and the ULA 1022 corresponding to the destination would be overwritten in the IPv6 1023 destination address. This would seemingly convey enough forwarding 1024 information so that OAL encapsulation could be avoided. However, 1025 this "minimal encapsulation" IPv6 packet would then have a non-ULA 1026 source address and ULA destination address, an incorrect value in 1027 upper layer protocol checksums, and a Hop Limit that is decremented 1028 within the spanning tree when it should not be. The insertion and 1029 removal of the ORH would also entail rewriting the Payload Length and 1030 Next Header fields - again, invalidating upper layer checksums. 1031 These irregularities would result in implementation challenges and 1032 the potential for operational issues, e.g., since actionable ICMPv6 1033 error reports could not be delivered to the original source. In 1034 order to address the issues, still more information such as the 1035 original IPv6 source address could be written into the ORH. However, 1036 with the additional information the benefit of the "minimal 1037 encapsulation" savings quickly diminishes, and becomes overshadowed 1038 by the implementation and operational irregularities. 1040 3.2.5. Segment Routing Topologies (SRTs) 1042 The 64-bit sub-prefixes of [ULA]::/48 identify up to 2^16 distinct 1043 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 1044 OMNI link overlay instance using a distinct set of ULAs, and emulates 1045 a Virtual LAN (VLAN) service for the OMNI link. In some cases (e.g., 1046 when redundant topologies are needed for fault tolerance and 1047 reliability) it may be beneficial to deploy multiple SRTs that act as 1048 independent overlay instances. A communication failure in one 1049 instance therefore will not affect communications in other instances. 1051 Each SRT is identified by a distinct value in bits 48-63 of 1052 [ULA]::/48, i.e., as [ULA0]::/64, [ULA1]::/64, [ULA2]::/64, etc. 1053 Each OMNI interface is identified by a unique interface name (e.g., 1054 omni0, omni1, omni2, etc.) and assigns an anycast ADM-ULA 1055 corresponding to its SRT prefix length. The anycast ADM-ULA is used 1056 for OMNI interface determination in Safety-Based Multilink (SBM) as 1057 discussed in [I-D.templin-6man-omni-interface]. Each OMNI interface 1058 further applies Performance-Based Multilink (PBM) internally. 1060 3.2.6. Segment Routing For OMNI Link Selection 1062 An original IPv6 source can direct an IPv6 packet to an AERO node by 1063 including a standard IPv6 Segment Routing Header (SRH) [RFC8754] with 1064 the anycast ADM-ULA for the selected SRT as either the IPv6 1065 destination or as an intermediate hop within the SRH. This allows 1066 the original source to determine the specific OMNI link topology a 1067 packet will traverse when there may be multiple alternatives. 1069 When the AERO node processes the SRH and forwards the packet to the 1070 correct OMNI interface, the OMNI interface writes the next IPv6 1071 address from the SRH into the IPv6 destination address and decrements 1072 Segments Left. If decrementing would cause Segments Left to become 1073 0, the OMNI interface deletes the SRH before forwarding. This form 1074 of Segment Routing supports Safety-Based Multilink (SBM). 1076 3.2.7. Segment Routing Within the OMNI Link 1078 OAL sources can insert an ORH for Segment Routing within the OMNI 1079 link to influence the paths of OAL packets sent to OAL destinations 1080 in remote segments without requiring all packets to traverse strict 1081 spanning tree paths. 1083 When an AERO node's OMNI interface has an original IP packet to send 1084 to a target discovered through route optimization located in the same 1085 OMNI link segment, it acts as an OAL source to perform OAL 1086 encapsulation and fragmentation. The node then uses the target's 1087 Link Layer Address (L2ADDR) information for *NET encapsulation. 1089 When an AERO node's OMNI interface has an original IP packet to send 1090 to a route optimization target located in a remote OMNI link segment, 1091 it acts as an OAL source the same as above but also includes an ORH 1092 while setting the OAL destination to the Subnet Router Anycast 1093 address for the final OMNI link segment, then forwards the resulting 1094 *NET packets to a Bridge. 1096 When a Bridge receives a *NET packet destined to its Subnet Router 1097 Anycast address with an ORH with SRT/LHS values corresponding to the 1098 local segment, it examines the L2ADDR according to FMT and removes 1099 the ORH from the packet. The Bridge then writes the MNP-ULA 1100 corresponding to the ORH Destination Suffix into the OAL destination 1101 address, re-encapsulates the *NET packet according to L2ADDR and 1102 forwards the *NET packet either to the LHS Server/Proxy or directly 1103 to the target Client itself. In this way, the Bridge participates in 1104 route optimization to reduce traffic load and suboptimal routing 1105 through strict spanning tree paths. 1107 3.3. OMNI Interface Characteristics 1109 OMNI interfaces are virtual interfaces configured over one or more 1110 underlying interfaces classified as follows: 1112 o INET interfaces connect to an INET either natively or through one 1113 or several NATs. Native INET interfaces have global IP addresses 1114 that are reachable from any INET correspondent. All Server, Relay 1115 and Bridge interfaces are native interfaces, as are INET-facing 1116 interfaces of Proxys. NATed INET interfaces connect to a private 1117 network behind one or more NATs that provide INET access. Clients 1118 that are behind a NAT are required to send periodic keepalive 1119 messages to keep NAT state alive when there are no data packets 1120 flowing. 1122 o ANET interfaces connect to an ANET that is separated from the open 1123 INET by a Proxy. Proxys can actively issue control messages over 1124 the INET on behalf of the Client to reduce ANET congestion. 1126 o VPNed interfaces use security encapsulation over the INET to a 1127 Virtual Private Network (VPN) server that also acts as a Server or 1128 Proxy. Other than the link-layer encapsulation format, VPNed 1129 interfaces behave the same as Direct interfaces. 1131 o Direct (i.e., single-hop point-to-point) interfaces connect a 1132 Client directly to a Server or Proxy without crossing any ANET/ 1133 INET paths. An example is a line-of-sight link between a remote 1134 pilot and an unmanned aircraft. The same Client considerations 1135 apply as for VPNed interfaces. 1137 OMNI interfaces use OAL encapsulation as discussed in Section 3.2.4. 1138 OMNI interfaces use link-layer encapsulation (see: Section 3.6) to 1139 exchange packets with OMNI link neighbors over INET or VPNed 1140 interfaces as well as over ANET interfaces for which the Client and 1141 Proxy may be multiple IP hops away. OMNI interfaces do not use link- 1142 layer encapsulation over Direct underlying interfaces or ANET 1143 interfaces when the Client and Proxy are known to be on the same 1144 underlying link. 1146 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 1147 state the same as for any interface. OMNI interfaces use ND messages 1148 including Router Solicitation (RS), Router Advertisement (RA), 1149 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 1150 neighbor cache management. 1152 OMNI interfaces send ND messages with an OMNI option formatted as 1153 specified in [I-D.templin-6man-omni-interface]. The OMNI option 1154 includes prefix registration information and Interface Attributes 1155 containing link information parameters for the OMNI interface's 1156 underlying interfaces. Each OMNI option may include multiple 1157 Interface Attributes sub-options, each identified by an ifIndex 1158 value. 1160 A Client's OMNI interface may be configured over multiple underlying 1161 interface connections. For example, common mobile handheld devices 1162 have both wireless local area network ("WLAN") and cellular wireless 1163 links. These links are often used "one at a time" with low-cost WLAN 1164 preferred and highly-available cellular wireless as a standby, but a 1165 simultaneous-use capability could provide benefits. In a more 1166 complex example, aircraft frequently have many wireless data link 1167 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1168 directional, etc.) with diverse performance and cost properties. 1170 If a Client's multiple underlying interfaces are used "one at a time" 1171 (i.e., all other interfaces are in standby mode while one interface 1172 is active), then ND message OMNI options include only a single 1173 Interface Attributes sub-option set to constant values. In that 1174 case, the Client would appear to have a single interface but with a 1175 dynamically changing link-layer address. 1177 If the Client has multiple active underlying interfaces, then from 1178 the perspective of ND it would appear to have multiple link-layer 1179 addresses. In that case, ND message OMNI options MAY include 1180 multiple Interface Attributes sub-options - each with values that 1181 correspond to a specific interface. Every ND message need not 1182 include Interface Attributes for all underlying interfaces; for any 1183 attributes not included, the neighbor considers the status as 1184 unchanged. 1186 Bridge, Server and Proxy OMNI interfaces may be configured over one 1187 or more secured tunnel interfaces. The OMNI interface configures 1188 both an ADM-LLA and its corresponding ADM-ULA, while the underlying 1189 secured tunnel interfaces are either unnumbered or configure the same 1190 ULA. The OMNI interface acting as an OAL source encapsulates and 1191 fragments each original IP packet, then and presents the resulting 1192 *NET packets to the underlying secured tunnel interface. Routing 1193 protocols such as BGP that run over the OMNI interface do not employ 1194 OAL encapsulation, but rather present the routing protocol messages 1195 directly to the underlying secured tunnels while using the ULA as the 1196 source address. This distinction must be honored consistently 1197 according to each node's configuration so that the IP forwarding 1198 table will associate discovered IP routes with the correct interface. 1200 3.4. OMNI Interface Initialization 1202 AERO Servers, Proxys and Clients configure OMNI interfaces as their 1203 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1204 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1205 to ensure that original IP packets with destination addresses covered 1206 by an MNP not explicitly assigned to a non-OMNI interface are 1207 directed to the OMNI interface. 1209 OMNI interface initialization procedures for Servers, Proxys, Clients 1210 and Bridges are discussed in the following sections. 1212 3.4.1. AERO Server/Relay Behavior 1214 When a Server enables an OMNI interface, it assigns an ADM-{LLA,ULA} 1215 appropriate for the given OMNI link segment. The Server also 1216 configures secured tunnels with one or more neighboring Bridges and 1217 engages in a BGP routing protocol session with each Bridge. 1219 The OMNI interface provides a single interface abstraction to the IP 1220 layer, but internally comprises multiple secured tunnels as well as 1221 an NBMA nexus for sending *NET packets to OMNI interface neighbors. 1222 The Server further configures a service to facilitate ND exchanges 1223 with AERO Clients and manages per-Client neighbor cache entries and 1224 IP forwarding table entries based on control message exchanges. 1226 Relays are simply Servers that run a dynamic routing protocol to 1227 redistribute routes between the OMNI interface and INET/EUN 1228 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1229 networks on the INET/EUN interfaces (i.e., the same as a Client would 1230 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1231 interfaces. The Relay further provides an attachment point of the 1232 OMNI link to a non-MNP-based global topology. 1234 3.4.2. AERO Proxy Behavior 1236 When a Proxy enables an OMNI interface, it assigns an ADM-{LLA, ULA} 1237 and configures secured tunnels with neighboring Bridges the same as 1238 for Servers. The Proxy also configures secured tunnels with one or 1239 more neighboring Bridges and maintains per-Client neighbor cache 1240 entries based on control message exchanges. Importantly Proxys are 1241 often configured to act as Servers, and vice-versa. 1243 3.4.3. AERO Client Behavior 1245 When a Client enables an OMNI interface, it assigns either an 1246 MNP-{LLA, ULA} or a Temporary ULA and sends RS messages with ND 1247 parameters over its underlying interfaces to a Server, which returns 1248 an RA message with corresponding parameters. The RS/RA messages may 1249 pass through a Proxy in the case of a Client's ANET interface, or 1250 through one or more NATs in the case of a Client's INET interface. 1251 (Note: if the Client used a Temporary ULA in its initial RS message, 1252 it will discover an MNP-{LLA, ULA} in the corresponding RA that it 1253 receives from the Server and begin using these new addresses. If the 1254 Client is operating outside the context of AERO infrastructure such 1255 as in a Mobile Ad-hoc Network (MANET), however, it may continue using 1256 Temporary ULAs for Client-to-Client communications until it 1257 encounters an infrastructure element that can provide an MNP.) 1259 3.4.4. AERO Bridge Behavior 1261 AERO Bridges configure an OMNI interface and assign the ADM-ULA 1262 Subnet Router Anycast address for each OMNI link segment they connect 1263 to. Bridges configure secured tunnels with Servers, Proxys and other 1264 Bridges, and engage in a BGP routing protocol session with neighbors 1265 on the spanning tree (see: Section 3.2.3). 1267 3.5. OMNI Interface Neighbor Cache Maintenance 1269 Each OMNI interface maintains a conceptual neighbor cache that 1270 includes an entry for each neighbor it communicates with on the OMNI 1271 link per [RFC4861]. Proxy neighbor cache entries are created and 1272 maintained by AERO Proxys when they process Client/Server ND 1273 exchanges, and remain in place for durations bounded by ND and prefix 1274 lifetimes. AERO Proxys maintain proxy neighbor cache entries for 1275 each of their associated Clients. Proxy neighbor cache entries track 1276 the Client state and the address of the Client's associated 1277 Server(s). 1279 To the list of neighbor cache entry states in Section 7.3.2 of 1280 [RFC4861], Proxy and Server OMNI interfaces add an additional state 1281 DEPARTED that applies to Clients that have recently departed. The 1282 interface sets a "DepartTime" variable for the neighbor cache entry 1283 to "DEPART_TIME" seconds. DepartTime is decremented unless a new ND 1284 message causes the state to return to REACHABLE. While a neighbor 1285 cache entry is in the DEPARTED state, packets destined to the target 1286 Client are forwarded to the Client's new location instead of being 1287 dropped. When DepartTime decrements to 0, the neighbor cache entry 1288 is deleted. It is RECOMMENDED that DEPART_TIME be set to the default 1289 constant value REACHABLE_TIME plus 10 seconds (40 seconds by default) 1290 to allow a window for packets in flight to be delivered while stale 1291 route optimization state may be present. 1293 Servers can act as RORs on behalf of multilink and/or disadvantaged 1294 Clients according to the Proxy Neighbor Advertisement specification 1295 in Section 7.2.8 of [RFC4861]. (Well-connected Clients can act as an 1296 ROR on their own behalf, and perform normal IPv6 ND operations the 1297 same as for any node.) When a Server acting as an ROR receives an 1298 authentic NS message used for route optimization, it searches for a 1299 neighbor cache entry for the target Client. The ROR then returns a 1300 solicited NA message without creating a neighbor cache entry for the 1301 ROS, but creates or updates a target Client "Report List" entry for 1302 the ROS and sets a "ReportTime" variable for the entry to REPORT_TIME 1303 seconds. The ROR resets ReportTime when it receives a new authentic 1304 NS message, and otherwise decrements ReportTime while no authentic NS 1305 messages have been received. It is RECOMMENDED that REPORT_TIME be 1306 set to the default constant value REACHABLE_TIME plus 10 seconds (40 1307 seconds by default) to allow a window for route optimization to 1308 converge before ReportTime decrements below REACHABLE_TIME. 1310 When the ROS receives a solicited NA message response to its NS 1311 message used for route optimization, it creates or updates a neighbor 1312 cache entry for the target network-layer and link-layer addresses. 1313 The ROS then (re)sets ReachableTime for the neighbor cache entry to 1314 REACHABLE_TIME seconds and uses this value to determine whether 1315 packets can be forwarded directly to the target, i.e., instead of via 1316 a default route. The ROS otherwise decrements ReachableTime while no 1317 further solicited NA messages arrive. It is RECOMMENDED that 1318 REACHABLE_TIME be set to the default constant value 30 seconds as 1319 specified in [RFC4861]. 1321 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1322 of NS keepalives sent when a correspondent may have gone unreachable, 1323 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1324 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1325 to limit the number of unsolicited NAs that can be sent based on a 1326 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1327 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1328 same as specified in [RFC4861]. 1330 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1331 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1332 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1333 different values are chosen, all nodes on the link MUST consistently 1334 configure the same values. Most importantly, DEPART_TIME and 1335 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1336 REACHABLE_TIME to avoid packet loss due to stale route optimization 1337 state. 1339 3.5.1. OMNI Neighbor Interface Attributes 1341 OMNI interface IPv6 ND messages include OMNI options 1342 [I-D.templin-6man-omni-interface] with Interface Attributes that 1343 provide Link-Layer Address and QoS Preference information for the 1344 neighbor's underlying interfaces. This information is stored in the 1345 neighbor cache and provides the basis for the forwarding algorithm 1346 specified in Section 3.10. The information is cumulative and 1347 reflects the union of the OMNI information from the most recent ND 1348 messages received from the neighbor; it is therefore not required 1349 that each ND message contain all neighbor information. 1351 The OMNI option Interface Attributes for each underlying interface 1352 includes a two-part "Link-Layer Address" consisting of a simple IP 1353 encapsulation address determined by the FMT and L2ADDR fields and an 1354 ADM-ULA determined by the SRT and LHS fields. Underlying interfaces 1355 are further selected based on their associated preference values 1356 "high", "medium", "low" or "disabled". 1358 Note: the OMNI option is distinct from any Source/Target Link-Layer 1359 Address Options (S/TLLAOs) that may appear in an ND message according 1360 to the appropriate IPv6 over specific link layer specification (e.g., 1361 [RFC2464]). If both an OMNI option and S/TLLAO appear, the former 1362 pertains to encapsulation addresses while the latter pertains to the 1363 native L2 address format of the underlying media. 1365 3.5.2. OMNI Neighbor Advertisement Message Flags 1367 As discussed in Section 4.4 of [RFC4861] NA messages include three 1368 flag bits R, S and O. OMNI interface NA messages treat the flags as 1369 follows: 1371 o R: The R ("Router") flag is set to 1 in the NA messages sent by 1372 all AERO/OMNI node types. Simple hosts that would set R to 0 do 1373 not occur on the OMNI link itself, but may occur on the downstream 1374 links of Clients and Relays. 1376 o S: The S ("Solicited") flag is set exactly as specified in 1377 Section 4.4. of [RFC4861], i.e., it is set to 1 for Solicited NAs 1378 and set to 0 for Unsolicited NAs (both unicast and multicast). 1380 o O: The O ("Override") flag is set to 0 for solicited proxy NAs and 1381 set to 1 for all other solicited and unsolicited NAs. For further 1382 study is whether solicited NAs for anycast targets apply for OMNI 1383 links. Since MNP-LLAs must be uniquely assigned to Clients to 1384 support correct ND protocol operation, however, no role is 1385 currently seen for assigning the same MNP-LLA to multiple Clients. 1387 3.6. OMNI Interface Encapsulation and Re-encapsulation 1389 The OMNI interface admits original IP packets then (acting as an OAL 1390 source) performs OAL encapsulation and fragmentation as specified in 1391 Section 3.2.4[I-D.templin-6man-omni-interface]. For original IP 1392 packets entering the OMNI interface from the network layer, the OMNI 1393 interface copies the "TTL/Hop Limit", "Type of Service/Traffic Class" 1394 [RFC2983], "Flow Label"[RFC6438] (for IPv6) and "Congestion 1395 Experienced" [RFC3168] values in the original packet's IP header into 1396 the corresponding fields in the OAL and *NET encapsulation header(s). 1397 For *NET packets undergoing re-encapsulation, the OMNI interface 1398 instead copies these values from the original *NET encapsulation 1399 header into the new *NET encapsulation header, i.e., the values are 1400 transferred between *NET encapsulation headers and *not* copied from 1401 the encapsulated packet's network-layer header. (Note especially 1402 that by copying the TTL/Hop Limit between encapsulation headers the 1403 value will eventually decrement to 0 if there is a (temporary) 1404 routing loop.) 1406 The OAL source sets the *NET header UDP source port to a constant 1407 value that it will use in each successive packet it sends. For 1408 packets sent to a Server, Relay or Bridge, the OAL source sets the 1409 UDP destination port to 8060, i.e., the IANA-registered port number 1410 for AERO. For packets sent to a Client, the OAL source sets the UDP 1411 destination port to the port value stored in the neighbor cache entry 1412 for this Client. The OAL source finally includes/omits the UDP 1413 checksum according to [RFC6935][RFC6936]. 1415 When a Proxy, Relay or Server re-encapsulates a packet received from 1416 a Client that includes an OAL but no ORH, it inserts an ORH 1417 immediately following the OAL header and adjusts the OAL payload 1418 length and destination address field. The inserted ORH will be 1419 removed by the final-hop Bridge, but its insertion and removal will 1420 not interfere with reassembly at the final destination. For this 1421 reason, Clients must reserve 40 bytes for a maximum-length ORH when 1422 they perform OAL encapsulation (see: Section 3.9). 1424 3.7. OMNI Interface Decapsulation 1426 OMNI interfaces (acting as OAL destinations) decapsulate and 1427 reassemble OAL packets into original IP packets destined either to 1428 the AERO node itself or to a destination reached via an interface 1429 other than the OMNI interface the packet was received on. When *NET 1430 packets containing OAL fragments arrive, the OMNI interface 1431 reassembles as discussed in Section 3.9. 1433 3.8. OMNI Interface Data Origin Authentication 1435 AERO nodes employ simple data origin authentication procedures. In 1436 particular: 1438 o AERO Bridges, Servers and Proxys accept encapsulated data packets 1439 and control messages received from the (secured) spanning tree. 1441 o AERO Proxys and Clients accept packets that originate from within 1442 the same secured ANET. 1444 o AERO Clients and Relays accept packets from downstream network 1445 correspondents based on ingress filtering. 1447 o AERO Clients, Relays and Servers verify the outer UDP/IP 1448 encapsulation addresses according to [RFC4380]. 1450 o AERO Clients, Relays and Servers as OAL destinations accept OAL 1451 packets/fragments with Identification values within the current 1452 window for the OAL source. 1454 AERO nodes silently drop any packets that do not satisfy the above 1455 data origin authentication procedures. Further security 1456 considerations are discussed in Section 6. 1458 3.9. OMNI Interface MTU 1460 The OMNI interface observes the link nature of tunnels, including the 1461 Maximum Transmission Unit (MTU), Maximum Reassembly Unit (MRU) and 1462 the role of fragmentation and reassembly [I-D.ietf-intarea-tunnels]. 1463 The OMNI interface employs an OMNI Adaptation Layer (OAL) that 1464 accommodates multiple underlying links with diverse MTUs while 1465 observing both a minimum and per-path Maximum Payload Size (MPS). 1466 The functions of the OAL and the OMNI interface MTU/MRU/MPS are 1467 specified in Section 5 of [I-D.templin-6man-omni-interface] with MTU/ 1468 MRU both set to the constant value 9180 bytes, with minimum MPS set 1469 to 400 bytes, and with path MPS set to a potentially larger value 1470 depending on the underlying path. 1472 When the network layer presents an original IP packet to the OMNI 1473 interface, the OAL source encapsulates and fragments the packet if 1474 necessary. When the network layer presents the OMNI interface with 1475 multiple IP packets bound to the same OAL destination, the OAL source 1476 can concatenate multiple IP packets together into a single OAL super- 1477 packet as discussed in [I-D.templin-6man-omni-interface]. The OAL 1478 source then fragments the encapsulated packet if necessary according 1479 to the minimum/path MPS such that the OAL headers appear in each 1480 fragment while the original IP packet header appears only in the 1481 first fragment. The OAL source then encapsulates each OAL fragment 1482 in *NET headers for transmission over an underlying interface 1483 connected to either a physical link such as Ethernet, WiFi and the 1484 like or a virtual link such as an Internet or higher-layer tunnel 1485 (see the definition of link in [RFC8200]). 1487 Note: A Client that does not (yet) have neighbor cache state for a 1488 target may omit the ORH in packets with the understanding that a 1489 Proxy/Server may insert an ORH on its behalf. For this reason, 1490 Clients reserve 40 bytes for the largest possible ORH in their OAL 1491 fragment size calculations. 1493 Note: Although the ORH may be removed by a Bridge on the path (see: 1494 Section 3.10.3), this does not interfere with the destination's 1495 ability to reassemble in the event that the packet was fragmented. 1496 This is due to the fact that the ORH is not included in the 1497 fragmentable part; therefore, its removal does not invalidate the 1498 offset values in any fragment headers. 1500 3.10. OMNI Interface Forwarding Algorithm 1502 Original IP packets enter a node's OMNI interface either from the 1503 network layer (i.e., from a local application or the IP forwarding 1504 system) while *NET packets enter from the link layer (i.e., from an 1505 OMNI interface neighbor). All packets entering a node's OMNI 1506 interface first undergo data origin authentication as discussed in 1507 Section 3.8. Those packets that satisfy data origin authentication 1508 are processed further, while all others are dropped silently. 1510 Original IP packets that enter the OMNI interface from the network 1511 layer are forwarded to an OMNI interface neighbor using OAL 1512 encapsulation and fragmentation to produce *NET packets for 1513 transmission over underlying interfaces. (If routing indicates that 1514 the packet should instead be forwarded back to the network layer, the 1515 packet is dropped to avoid looping). *NET-encapsulated OAL fragments 1516 that enter the OMNI interface from the link layer are either re- 1517 encapsulated and re-admitted into the OMNI link, or reassembled and 1518 forwarded to the network layer where they are subject to either local 1519 delivery or IP forwarding. In all cases, the OAL MUST NOT decrement 1520 the network layer TTL/Hop-count since its forwarding actions occur 1521 below the network layer. 1523 OMNI interfaces may have multiple underlying interfaces and/or 1524 neighbor cache entries for neighbors with multiple underlying 1525 interfaces (see Section 3.3). The OAL uses interface attributes and/ 1526 or traffic classifiers (e.g., DSCP value, port number, flow 1527 specification, etc.) to select an outgoing underlying interface for 1528 each OAL packet based on the node's own QoS preferences, and also to 1529 select a destination link-layer address based on the neighbor's 1530 underlying interface with the highest preference. AERO 1531 implementations SHOULD allow for QoS preference values to be modified 1532 at runtime through network management. 1534 If multiple outgoing interfaces and/or neighbor interfaces have a 1535 preference of "high", the AERO node replicates the OAL packet and 1536 sends one copy via each of the (outgoing / neighbor) interface pairs; 1537 otherwise, the node sends a single copy of the OAL packet via an 1538 interface with the highest preference. (While not strictly required, 1539 successful delivery may be more likely when all OAL fragments of the 1540 same OAL packet are sent over the same underlying interface.) AERO 1541 nodes keep track of which underlying interfaces are currently 1542 "reachable" or "unreachable", and only use "reachable" interfaces for 1543 forwarding purposes. 1545 The following sections discuss the OMNI interface forwarding 1546 algorithms for Clients, Proxys, Servers and Bridges. In the 1547 following discussion, an original IP packet's destination address is 1548 said to "match" if it is the same as a cached address, or if it is 1549 covered by a cached prefix (which may be encoded in an MNP-LLA). 1551 3.10.1. Client Forwarding Algorithm 1553 When an original IP packet enters a Client's OMNI interface from the 1554 network layer the Client searches for a neighbor cache entry that 1555 matches the destination. If there is a match, the Client selects one 1556 or more "reachable" neighbor interfaces in the entry for forwarding 1557 purposes. If there is no neighbor cache entry, the Client instead 1558 forwards the packet toward a Server (the packet is intercepted by a 1559 Proxy if there is a Proxy on the path). The Client (acting as an OAL 1560 source) performs OAL encapsulation and sets the OAL destination 1561 address to the MNP-ULA if there is a matching neighbor cache entry; 1562 otherwise, it sets the OAL destination to the ADM-ULA of the Proxy/ 1563 Server. If the Client has multiple original IP packets to send to 1564 the same neighbor, it can concatenate them in a single super-packet 1565 [I-D.templin-6man-omni-interface]. The OAL source then performs 1566 fragmentation to create OAL fragments (see: Section 3.9), appends any 1567 *NET encapsulation, and sends the resulting *NET packets over 1568 underlying interfaces to the neighbor acting as an OAL destination. 1570 If the neighbor interface selected for forwarding is located on the 1571 same OMNI link segment and not behind a NAT, the Client forwards the 1572 *NET packets directly according to the L2ADDR information for the 1573 neighbor. If the neighbor interface is behind a NAT on the same OMNI 1574 link segment, the Client instead forwards the initial *NET packets to 1575 its Server and initiates NAT traversal procedures. If the Client's 1576 intended source underlying interface is also behind a NAT and located 1577 on the same OMNI link segment, it sends a "direct bubble" over the 1578 interface per [RFC6081][RFC4380] to the L2ADDR found in the neighbor 1579 cache in order to establish state in its own NAT by generating 1580 traffic toward the neighbor (note that no response to the bubble is 1581 expected). 1583 The Client next sends an NS(NUD) message toward the MNP-ULA of the 1584 neighbor via its Server as discussed in Section 3.15. If the Client 1585 receives an NA(NUD) from the neighbor over the underlying interface, 1586 it marks the neighbor interface as "trusted" and sends future 1587 packets/fragments directly to the L2ADDR information for the neighbor 1588 instead of indirectly via the Server. The Client must honor the 1589 neighbor cache maintenance procedure by sending additional direct 1590 bubbles and/or NS/NA(NUD) messages as discussed in [RFC6081][RFC4380] 1591 in order to keep NAT state alive as long as packets are still 1592 flowing. 1594 When an *NET packet enters a Client's OMNI interface from the link- 1595 layer, if the OAL destination matches one of the Client's MNPs or 1596 LLAs the Client (acting as an OAL destination) reassembles and 1597 decapsulates as necessary and delivers the inner packet to the 1598 network layer. Otherwise, the Client drops the packet and MAY return 1599 a network-layer ICMP Destination Unreachable message subject to rate 1600 limiting (see: Section 3.11). 1602 3.10.2. Proxy/Server Forwarding Algorithm 1604 For IPv6 ND control messages originating from or destined to a 1605 Client, the Proxy/Server intercepts the message and updates its 1606 neighbor cache entry for the Client. If acting in Proxy-only mode, 1607 it then forwards a (proxyed) copy of the control message to one or 1608 more neighbors. (For example, the Proxy forwards a proxyed version 1609 of a Client's NS/RS message to the neighbor, and forwards a proxyed 1610 version of the neighbor's NA/RA reply to the Client.) 1612 When the Proxy/Server receives an original IP packet from the network 1613 layer, it drops the packet if routing indicates that it should be 1614 forwarded back to the network layer to avoid looping. Otherwise, the 1615 Proxy/Server regards the original IP packet the same as if it had 1616 arrived as *NET packets with OAL destination set to its own ADM-ULA. 1617 When the Proxy/Server receives *NET packets with OAL destination set 1618 to its own ADM-ULA, it performs OAL reassembly if necessary to obtain 1619 the original IP packet. 1621 The Proxy/Server next searches for a neighbor cache entry that 1622 matches the original IP destination and proceeds as follows: 1624 o if the original IP destination matches a neighbor cache entry, the 1625 Proxy/Sever uses one or more "reachable" neighbor interfaces in 1626 the entry for packet forwarding using OAL encapsulation and 1627 fragmentation according to the cached link-layer address 1628 information. If the neighbor interface is in a different OMNI 1629 link segment, the Proxy/Server forwards the resulting *NET packets 1630 to a Bridge; otherwise, it forwards the *NET packets directly to 1631 the neighbor. If the neighbor is behind a NAT, the Proxy/Server 1632 instead forwards initial *NET packets via a Bridge while sending 1633 an NS(NUD) to the neighbor. When the Proxy/Server receives the 1634 NA(NUD), it can begin forwarding *NET packets directly to the 1635 neighbor the same as discussed in Section 3.10.1 while sending 1636 additional NS(NUD) messages as necessary to maintain NAT state. 1637 Note that no direct bubbles are necessary since the Proxy/Server 1638 is by definition not located behind a NAT. 1640 o else, if the original IP destination matches a non-MNP route in 1641 the IP forwarding table or an ADM-LLA assigned to the Proxy/ 1642 Server's OMNI interface, the Proxy/Server decapsulates the packet 1643 and presents it to the network layer for local delivery or IP 1644 forwarding. 1646 o else, the Proxy/Server initiates address resolution as discussed 1647 in Section 3.14, while retaining initial original IP packets in a 1648 small queue awaiting address resolution completion. 1650 When the Proxy/Server receives a *NET packet with OAL destination set 1651 to a non-MNP ULA, it accepts the packet only if data origin 1652 authentication succeeds and if there is a network layer routing table 1653 for a GUA route that matches the non-MNP ULA. If there is no route, 1654 the Proxy/Server drops the packet; otherwise, it reassembles and 1655 decapsulates to obtain the original IP packet and presents it to the 1656 network layer where it will be delivered according to standard IP 1657 forwarding. 1659 When the Proxy/Server receives a *NET packet with OAL destination set 1660 to an MNP-ULA, it accepts the packet only if data origin 1661 authentication succeeds and if there is a neighbor cache entry that 1662 matches the OAL destination. If the neighbor cache entry state is 1663 DEPARTED, the Proxy/Server inserts an ORH that encodes the MNP-ULA 1664 destination suffix and changes the OAL destination address to the 1665 ADM-ULA of the new Server, then re-encapsulates the *NET packet and 1666 forwards it to a Bridge which will eventually deliver it to the new 1667 Server. 1669 If the neighbor cache state is REACHABLE, the Proxy/Server can 1670 instead either reassemble first and then re-encapsulate/re-fragment 1671 before forwarding to the Client or forward the raw fragments on to 1672 the Client which then must reassemble. In the former case, the 1673 Proxy/Server can re-fragment to a size that better matches the link 1674 MTU for the Client, which may be important for low-end links with 1675 large MTUs. In the latter case, the Client may receive fragments 1676 that are smaller than its link MTU but can still be reassembled; this 1677 case may provide an important performance benefit to Proxy/Servers by 1678 permitting them to avoid excessive reassembly and re-fragmentation 1679 overhead. In either case, the Proxy/Server can return a PTB if 1680 necessary (see: [I-D.templin-6man-omni-interface]) when it receives a 1681 *NET packet containing an OAL first fragment. 1683 Note: If the Proxy/Server has multiple original IP packets to send to 1684 the same neighbor, it can concatenate them in a single OAL super- 1685 packet [I-D.templin-6man-omni-interface]. 1687 3.10.3. Bridge Forwarding Algorithm 1689 Bridges forward OAL fragments over secured tunnels the same as any 1690 IPv6 router. When the Bridge receives an OAL fragment or an original 1691 IP packet via a secured tunnel, it removes the outer *NET header and 1692 searches for a forwarding table entry that matches the OAL 1693 destination address. The Bridge then processes the packet as 1694 follows: 1696 o if the destination matches its ADM-ULA Subnet Router Anycast 1697 address the Bridge processes the packet locally before forwarding. 1698 The Bridge drops the packet if the OAL fragment does not include 1699 an ORH; otherwise, for NA(NUD) messages the Bridge replaces the 1700 OMNI option Interface Attributes sub-option with information for 1701 its own interface while retaining the ifIndex value supplied by 1702 the NA(NUD) message source. For all packet types, the Bridge next 1703 examines the ORH FMT code. If the code indicates the destination 1704 is a Client on the open *NET (or, a Client behind a NAT for which 1705 NAT traversal procedures have already converged) the Bridge 1706 removes the ORH then writes the MNP-ULA formed from the ORH 1707 Destination Suffix into the OAL destination. The Bridge then re- 1708 encapsulates the *NET fragment and forwards it to the ORH L2ADDR. 1709 For all other destination cases, the Bridge instead writes the 1710 ADM-ULA formed from the ORH SRT/LHS into the OAL destination 1711 address and forwards the OAL/ORH encapsulated packet to the ADM- 1712 ULA Server while invoking NAT traversal procedures the same as for 1713 Proxys and Servers if necessary, noting that no direct bubbles are 1714 necessary since only the target Client and not the Bridge is 1715 behind a NAT. 1717 o else, if the destination matches one of the Bridge's own 1718 addresses, the Bridge submits the packet for local delivery to 1719 support local applications such as routing protocols. 1721 o else, if the destination matches a forwarding table entry the 1722 Bridge forwards the packet via a secured tunnel to the next hop. 1723 If the destination matches an MSP without matching an MNP, 1724 however, the Bridge instead drops the packet and returns an ICMP 1725 Destination Unreachable message subject to rate limiting (see: 1726 Section 3.11). 1728 o else, the Bridge drops the packet and returns an ICMP Destination 1729 Unreachable as above. 1731 As for any IP router, the Bridge decrements the TTL/Hop Limit when it 1732 forwards the packet. Therefore, when an OAL header is present only 1733 the Hop Limit in the OAL header is decremented and not the TTL/Hop 1734 Limit in the inner packet header. Bridges do not insert OAL/ORH 1735 headers themselves; instead, they act as IPv6 routers and forward 1736 packets based on the destination address found in the headers of 1737 packets they receive. 1739 3.11. OMNI Interface Error Handling 1741 When an AERO node admits a packet into the OMNI interface, it may 1742 receive link-layer or network-layer error indications. 1744 A link-layer error indication is an ICMP error message generated by a 1745 router in the INET on the path to the neighbor or by the neighbor 1746 itself. The message includes an IP header with the address of the 1747 node that generated the error as the source address and with the 1748 link-layer address of the AERO node as the destination address. 1750 The IP header is followed by an ICMP header that includes an error 1751 Type, Code and Checksum. Valid type values include "Destination 1752 Unreachable", "Time Exceeded" and "Parameter Problem" 1753 [RFC0792][RFC4443]. (OMNI interfaces ignore all link-layer IPv4 1754 "Fragmentation Needed" and IPv6 "Packet Too Big" messages since they 1755 only emit packets that are guaranteed to be no larger than the IP 1756 minimum link MTU as discussed in Section 3.9.) 1758 The ICMP header is followed by the leading portion of the packet that 1759 generated the error, also known as the "packet-in-error". For 1760 ICMPv6, [RFC4443] specifies that the packet-in-error includes: "As 1761 much of invoking packet as possible without the ICMPv6 packet 1762 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1763 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1764 "Internet Header + 64 bits of Original Data Datagram", however 1765 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1766 ICMP datagram SHOULD contain as much of the original datagram as 1767 possible without the length of the ICMP datagram exceeding 576 1768 bytes". 1770 The link-layer error message format is shown in Figure 5 (where, "L2" 1771 and "L3" refer to link-layer and network-layer, respectively): 1773 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1774 ~ ~ 1775 | L2 IP Header of | 1776 | error message | 1777 ~ ~ 1778 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1779 | L2 ICMP Header | 1780 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1781 ~ ~ P 1782 | IP and other encapsulation | a 1783 | headers of original L3 packet | c 1784 ~ ~ k 1785 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1786 ~ ~ t 1787 | IP header of | 1788 | original L3 packet | i 1789 ~ ~ n 1790 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1791 ~ ~ e 1792 | Upper layer headers and | r 1793 | leading portion of body | r 1794 | of the original L3 packet | o 1795 ~ ~ r 1796 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1798 Figure 5: OMNI Interface Link-Layer Error Message Format 1800 The AERO node rules for processing these link-layer error messages 1801 are as follows: 1803 o When an AERO node receives a link-layer Parameter Problem message, 1804 it processes the message the same as described as for ordinary 1805 ICMP errors in the normative references [RFC0792][RFC4443]. 1807 o When an AERO node receives persistent link-layer Time Exceeded 1808 messages, the IP ID field may be wrapping before earlier fragments 1809 awaiting reassembly have been processed. In that case, the node 1810 should begin including integrity checks and/or institute rate 1811 limits for subsequent packets. 1813 o When an AERO node receives persistent link-layer Destination 1814 Unreachable messages in response to encapsulated packets that it 1815 sends to one of its neighbor correspondents, the node should 1816 process the message as an indication that a path may be failing, 1817 and optionally initiate NUD over that path. If it receives 1818 Destination Unreachable messages over multiple paths, the node 1819 should allow future packets destined to the correspondent to flow 1820 through a default route and re-initiate route optimization. 1822 o When an AERO Client receives persistent link-layer Destination 1823 Unreachable messages in response to encapsulated packets that it 1824 sends to one of its neighbor Servers, the Client should mark the 1825 path as unusable and use another path. If it receives Destination 1826 Unreachable messages on many or all paths, the Client should 1827 associate with a new Server and release its association with the 1828 old Server as specified in Section 3.16.5. 1830 o When an AERO Server receives persistent link-layer Destination 1831 Unreachable messages in response to encapsulated packets that it 1832 sends to one of its neighbor Clients, the Server should mark the 1833 underlying path as unusable and use another underlying path. 1835 o When an AERO Server or Proxy receives link-layer Destination 1836 Unreachable messages in response to an encapsulated packet that it 1837 sends to one of its permanent neighbors, it treats the messages as 1838 an indication that the path to the neighbor may be failing. 1839 However, the dynamic routing protocol should soon reconverge and 1840 correct the temporary outage. 1842 When an AERO Bridge receives a packet for which the network-layer 1843 destination address is covered by an MSP, the Bridge drops the packet 1844 if there is no more-specific routing information for the destination 1845 and returns a network-layer Destination Unreachable message subject 1846 to rate limiting. The Bridge writes the network-layer source address 1847 of the original packet as the destination address and uses one of its 1848 non link-local addresses as the source address of the message. 1850 When an AERO node receives an encapsulated packet for which the 1851 reassembly buffer it too small, it drops the packet and returns a 1852 network-layer Packet Too Big (PTB) message. The node first writes 1853 the MRU value into the PTB message MTU field, writes the network- 1854 layer source address of the original packet as the destination 1855 address and writes one of its non link-local addresses as the source 1856 address. 1858 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1860 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1861 coordinated as discussed in the following Sections. 1863 3.12.1. AERO Service Model 1865 Each AERO Server on the OMNI link is configured to facilitate Client 1866 prefix delegation/registration requests. Each Server is provisioned 1867 with a database of MNP-to-Client ID mappings for all Clients enrolled 1868 in the AERO service, as well as any information necessary to 1869 authenticate each Client. The Client database is maintained by a 1870 central administrative authority for the OMNI link and securely 1871 distributed to all Servers, e.g., via the Lightweight Directory 1872 Access Protocol (LDAP) [RFC4511], via static configuration, etc. 1873 Clients receive the same service regardless of the Servers they 1874 select. 1876 AERO Clients and Servers use ND messages to maintain neighbor cache 1877 entries. AERO Servers configure their OMNI interfaces as advertising 1878 NBMA interfaces, and therefore send unicast RA messages with a short 1879 Router Lifetime value (e.g., ReachableTime seconds) in response to a 1880 Client's RS message. Thereafter, Clients send additional RS messages 1881 to keep Server state alive. 1883 AERO Clients and Servers include prefix delegation and/or 1884 registration parameters in RS/RA messages (see 1885 [I-D.templin-6man-omni-interface]). The ND messages are exchanged 1886 between Client and Server according to the prefix management schedule 1887 required by the service. If the Client knows its MNP in advance, it 1888 can employ prefix registration by including its MNP-LLA as the source 1889 address of an RS message and with an OMNI option with valid prefix 1890 registration information for the MNP. If the Server (and Proxy) 1891 accept the Client's MNP assertion, they inject the prefix into the 1892 routing system and establish the necessary neighbor cache state. If 1893 the Client does not have a pre-assigned MNP, it can instead employ 1894 prefix delegation by including the unspecified address (::) as the 1895 source address of an RS message and with an OMNI option with prefix 1896 delegation parameters to request an MNP. 1898 The following sections specify the Client and Server behavior. 1900 3.12.2. AERO Client Behavior 1902 AERO Clients discover the addresses of Servers in a similar manner as 1903 described in [RFC5214]. Discovery methods include static 1904 configuration (e.g., from a flat-file map of Server addresses and 1905 locations), or through an automated means such as Domain Name System 1906 (DNS) name resolution [RFC1035]. Alternatively, the Client can 1907 discover Server addresses through a layer 2 data link login exchange, 1908 or through a unicast RA response to a multicast/anycast RS as 1909 described below. In the absence of other information, the Client can 1910 resolve the DNS Fully-Qualified Domain Name (FQDN) 1911 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1912 text string and "[domainname]" is a DNS suffix for the OMNI link 1913 (e.g., "example.com"). 1915 To associate with a Server, the Client acts as a requesting router to 1916 request MNPs. The Client prepares an RS message with prefix 1917 management parameters and includes a Nonce and Timestamp option if 1918 the Client needs to correlate RA replies. If the Client already 1919 knows the Server's ADM-LLA, it includes the LLA as the network-layer 1920 destination address; otherwise, the Client includes the (link-local) 1921 All-Routers multicast as the network-layer destination. If the 1922 Client already knows its own MNP-LLA, it can use the MNP-LLA as the 1923 network-layer source address and include an OMNI option with prefix 1924 registration information. Otherwise, the Client uses the unspecified 1925 address (::) as the network-layer source address and includes prefix 1926 delegation parameters in the OMNI option (see: 1927 [I-D.templin-6man-omni-interface]). The Client includes Interface 1928 Attributes corresponding to the underlying interface over which it 1929 will send the RS message, and MAY include additional Interface 1930 Attributes specific to other underlying interfaces. 1932 The Client then sends the RS message (either directly via Direct 1933 interfaces, via a VPN for VPNed interfaces, via a Proxy for ANET 1934 interfaces or via INET encapsulation for INET interfaces) while using 1935 OAL encapsulation/fragmentation, then waits for an RA message reply 1936 (see Section 3.12.3). The Client retries up to MAX_RTR_SOLICITATIONS 1937 times until an RA is received. If the Client receives no RAs, or if 1938 it receives an RA with Router Lifetime set to 0, the Client SHOULD 1939 abandon this Server and try another Server. Otherwise, the Client 1940 processes the prefix information found in the RA message. 1942 When the Client processes an RA, it first performs OAL reassembly and 1943 decapsulation then creates a neighbor cache entry with the Server's 1944 ADM-LLA as the network-layer address and the Server's encapsulation 1945 and/or link-layer addresses as the link-layer address. The Client 1946 next records the RA Router Lifetime field value in the neighbor cache 1947 entry as the time for which the Server has committed to maintaining 1948 the MNP in the routing system via this underlying interface, and 1949 caches the other RA configuration information including Cur Hop 1950 Limit, M and O flags, Reachable Time and Retrans Timer. The Client 1951 then autoconfigures MNP-LLAs for any delegated MNPs and assigns them 1952 to the OMNI interface. The Client also caches any MSPs included in 1953 Route Information Options (RIOs) [RFC4191] as MSPs to associate with 1954 the OMNI link, and assigns the MTU value in the MTU option to the 1955 underlying interface. 1957 The Client then registers additional underlying interfaces with the 1958 Server by sending RS messages via each additional interface as 1959 described above. The RS messages include the same parameters as for 1960 the initial RS/RA exchange, but with destination address set to the 1961 Server's ADM-LLA. The Client finally sub-delegates the MNPs to its 1962 attached EUNs and/or the Client's own internal virtual interfaces as 1963 described in [I-D.templin-v6ops-pdhost] to support the Client's 1964 downstream attached "Internet of Things (IoT)". The Client then 1965 sends additional RS messages over each underlying interface before 1966 the Router Lifetime received for that interface expires. 1968 After the Client registers its underlying interfaces, it may wish to 1969 change one or more registrations, e.g., if an interface changes 1970 address or becomes unavailable, if QoS preferences change, etc. To 1971 do so, the Client prepares an RS message to send over any available 1972 underlying interface as above. The RS includes an OMNI option with 1973 prefix registration/delegation information, with Interface Attributes 1974 specific to the selected underlying interface, and with any 1975 additional Interface Attributes specific to other underlying 1976 interfaces. When the Client receives the Server's RA response, it 1977 has assurance that the Server has been updated with the new 1978 information. 1980 If the Client wishes to discontinue use of a Server it issues an RS 1981 message over any underlying interface with an OMNI option with a 1982 prefix release indication. When the Server processes the message, it 1983 releases the MNP, sets the neighbor cache entry state for the Client 1984 to DEPARTED and returns an RA reply with Router Lifetime set to 0. 1985 After a short delay (e.g., 2 seconds), the Server withdraws the MNP 1986 from the routing system. 1988 3.12.3. AERO Server Behavior 1990 AERO Servers act as IP routers and support a prefix delegation/ 1991 registration service for Clients. Servers arrange to add their ADM- 1992 LLAs to a static map of Server addresses for the link and/or the DNS 1993 resource records for the FQDN "linkupnetworks.[domainname]" before 1994 entering service. Server addresses should be geographically and/or 1995 topologically referenced, and made available for discovery by Clients 1996 on the OMNI link. 1998 When a Server receives a prospective Client's RS message on its OMNI 1999 interface, it SHOULD return an immediate RA reply with Router 2000 Lifetime set to 0 if it is currently too busy or otherwise unable to 2001 service the Client. Otherwise, the Server performs OAL reassembly 2002 and decapsulation, then authenticates the RS message and processes 2003 the prefix delegation/registration parameters. The Server first 2004 determines the correct MNPs to provide to the Client by processing 2005 the MNP-LLA prefix parameters and/or the DHCPv6 OMNI sub-option. 2006 When the Server returns the MNPs, it also creates a forwarding table 2007 entry for the MNP-ULA corresponding to each MNP so that the MNPs are 2008 propagated into the routing system (see: Section 3.2.3). For IPv6, 2009 the Server creates an IPv6 forwarding table entry for each MNP. For 2010 IPv4, the Server creates an IPv6 forwarding table entry with the 2011 IPv4-compatibility MNP-ULA prefix corresponding to the IPv4 address. 2013 The Server next creates a neighbor cache entry for the Client using 2014 the base MNP-LLA as the network-layer address and with lifetime set 2015 to no more than the smallest prefix lifetime. Next, the Server 2016 updates the neighbor cache entry by recording the information in each 2017 Interface Attributes sub-option in the RS OMNI option. The Server 2018 also records the actual OAL/INET addresses in the neighbor cache 2019 entry. 2021 Next, the Server prepares an RA message using its ADM-LLA as the 2022 network-layer source address and the network-layer source address of 2023 the RS message as the network-layer destination address. The Server 2024 sets the Router Lifetime to the time for which it will maintain both 2025 this underlying interface individually and the neighbor cache entry 2026 as a whole. The Server also sets Cur Hop Limit, M and O flags, 2027 Reachable Time and Retrans Timer to values appropriate for the OMNI 2028 link. The Server includes the MNPs, any other prefix management 2029 parameters and an OMNI option with no Interface Attributes. The 2030 Server then includes one or more RIOs that encode the MSPs for the 2031 OMNI link, plus an MTU option (see Section 3.9). The Server finally 2032 forwards the message to the Client using OAL encapsulation/ 2033 fragmentation as necessary. 2035 After the initial RS/RA exchange, the Server maintains a 2036 ReachableTime timer for each of the Client's underlying interfaces 2037 individually (and for the Client's neighbor cache entry collectively) 2038 set to expire after ReachableTime seconds. If the Client (or Proxy) 2039 issues additional RS messages, the Server sends an RA response and 2040 resets ReachableTime. If the Server receives an ND message with a 2041 prefix release indication it sets the Client's neighbor cache entry 2042 to the DEPARTED state and withdraws the MNP from the routing system 2043 after a short delay (e.g., 2 seconds). If ReachableTime expires 2044 before a new RS is received on an individual underlying interface, 2045 the Server marks the interface as DOWN. If ReachableTime expires 2046 before any new RS is received on any individual underlying interface, 2047 the Server sets the neighbor cache entry state to STALE and sets a 10 2048 second timer. If the Server has not received a new RS or ND message 2049 with a prefix release indication before the 10 second timer expires, 2050 it deletes the neighbor cache entry and withdraws the MNP from the 2051 routing system. 2053 The Server processes any ND messages pertaining to the Client and 2054 returns an NA/RA reply in response to solicitations. The Server may 2055 also issue unsolicited RA messages, e.g., with reconfigure parameters 2056 to cause the Client to renegotiate its prefix delegation/ 2057 registrations, with Router Lifetime set to 0 if it can no longer 2058 service this Client, etc. Finally, If the neighbor cache entry is in 2059 the DEPARTED state, the Server deletes the entry after DepartTime 2060 expires. 2062 Note: Clients SHOULD notify former Servers of their departures, but 2063 Servers are responsible for expiring neighbor cache entries and 2064 withdrawing routes even if no departure notification is received 2065 (e.g., if the Client leaves the network unexpectedly). Servers 2066 SHOULD therefore set Router Lifetime to ReachableTime seconds in 2067 solicited RA messages to minimize persistent stale cache information 2068 in the absence of Client departure notifications. A short Router 2069 Lifetime also ensures that proactive Client/Server RS/RA messaging 2070 will keep any NAT state alive (see above). 2072 Note: All Servers on an OMNI link MUST advertise consistent values in 2073 the RA Cur Hop Limit, M and O flags, Reachable Time and Retrans Timer 2074 fields the same as for any link, since unpredictable behavior could 2075 result if different Servers on the same link advertised different 2076 values. 2078 3.12.3.1. DHCPv6-Based Prefix Registration 2080 When a Client is not pre-provisioned with an MNP-LLA, it will need 2081 for the Server to select one or more MNPs on its behalf and set up 2082 the correct state in the AERO routing service. (A Client with a pre- 2083 provisioned MNP may also request the Server to select additional 2084 MNPs.) The DHCPv6 service [RFC8415] is used to support this 2085 requirement. 2087 When a Client needs to have the Server select MNPs, it sends an RS 2088 message with source address set to the unspecified address (::) and 2089 with an OMNI option that includes a DHCPv6 message sub-option with 2090 DHCPv6 Prefix Delegation (DHCPv6-PD) parameters. When the Server 2091 receives the RS message, it extracts the DHCPv6-PD message from the 2092 OMNI option. 2094 The Server then acts as a "Proxy DHCPv6 Client" in a message exchange 2095 with the locally-resident DHCPv6 server, which delegates MNPs and 2096 returns a DHCPv6-PD Reply message. (If the Server wishes to defer 2097 creation of MN state until the DHCPv6-PD Reply is received, it can 2098 instead act as a Lightweight DHCPv6 Relay Agent per [RFC6221] by 2099 encapsulating the DHCPv6-PD message in a Relay-forward/reply exchange 2100 with Relay Message and Interface ID options.) 2101 When the Server receives the DHCPv6-PD Reply, it adds a route to the 2102 routing system and creates an MNP-LLA based on the delegated MNP. 2103 The Server then sends an RA back to the Client with the (newly- 2104 created) MNP-LLA as the destination address and with the DHCPv6-PD 2105 Reply message coded in the OMNI option. When the Client receives the 2106 RA, it creates a default route, assigns the Subnet Router Anycast 2107 address and sets its MNP-LLA based on the delegated MNP. 2109 Note: See [I-D.templin-6man-omni-interface] for an MNP delegation 2110 alternative in which the Client can optionally avoid including a 2111 DHCPv6 message sub-option. Namely, when the Client requests a single 2112 MNP it can set the RS source to the unspecified address (::) and 2113 include a Node Identification sub-option and Preflen in the OMNI 2114 option (but with no DHCPv6 message sub-option). When the Server 2115 receives the RS message, it forwards a self-generated DHCPv6 Solicit 2116 message to the DHCPv6 server on behalf of the Client. When the 2117 Server receives the DHCPv6 Reply, it prepares an RA message with an 2118 OMNI option with Preflen information (but with no DHCPv6 message sub- 2119 option), then places the (newly-created) MNP-LLA in the RA 2120 destination address and returns the message to the Client. 2122 3.13. The AERO Proxy 2124 Clients may connect to protected-spectrum ANETs that employ physical 2125 and/or link-layer security services to facilitate communications to 2126 Servers in outside INETs. In that case, the ANET can employ an AERO 2127 Proxy. The Proxy is located at the ANET/INET border and listens for 2128 RS messages originating from or RA messages destined to ANET Clients. 2129 The Proxy acts on these control messages as follows: 2131 o when the Proxy receives an RS message from a new ANET Client, it 2132 first authenticates the message then examines the network-layer 2133 destination address. If the destination address is a Server's 2134 ADM-LLA, the Proxy proceeds to the next step. Otherwise, if the 2135 destination is (link-local) All-Routers multicast, the Proxy 2136 selects a "nearby" Server that is likely to be a good candidate to 2137 serve the Client and replaces the destination address with the 2138 Server's ADM-LLA. Next, the Proxy creates a proxy neighbor cache 2139 entry and caches the Client and Server link-layer addresses along 2140 with the OMNI option information and any other identifying 2141 information including Transaction IDs, Client Identifiers, Nonce 2142 values, etc. The Proxy finally encapsulates the (proxyed) RS 2143 message in an OAL header with source set to the Proxy's ADM-ULA 2144 and destination set to the Server's ADM-ULA. The Proxy also 2145 includes an OMNI header with an Interface Attributes option that 2146 includes its own INET address plus a unique Port Number for this 2147 Client, then forwards the message into the OMNI link spanning 2148 tree. (Note: including a unique Port Number allows the Server to 2149 distinguish different Clients located behind the same Proxy at the 2150 link-layer, whereas the link-layer addresses would otherwise be 2151 indistinguishable) 2153 o when the Server receives the RS, it authenticates the message then 2154 creates or updates a neighbor cache entry for the Client with the 2155 Proxy's ADM-ULA, INET address and Port Number as the link-layer 2156 address information. The Server then sends an RA message back to 2157 the Proxy via the spanning tree. 2159 o when the Proxy receives the RA, it authenticates the message and 2160 matches it with the proxy neighbor cache entry created by the RS. 2161 The Proxy then caches the prefix information as a mapping from the 2162 Client's MNPs to the Client's link-layer address, caches the 2163 Server's advertised Router Lifetime and sets the neighbor cache 2164 entry state to REACHABLE. The Proxy then optionally rewrites the 2165 Router Lifetime and forwards the (proxyed) message to the Client. 2166 The Proxy finally includes an MTU option (if necessary) with an 2167 MTU to use for the underlying ANET interface. 2169 After the initial RS/RA exchange, the Proxy forwards any Client data 2170 packets for which there is no matching neighbor cache entry to a 2171 Bridge using OAL encapsulation with its own ADM-ULA as the source and 2172 the MNP-ULA corresponding to the Client as the destination. The 2173 Proxy instead forwards any Client data destined to a neighbor cache 2174 target directly to the target according to the OAL/link-layer 2175 information - the process of establishing neighbor cache entries is 2176 specified in Section 3.14. 2178 While the Client is still attached to the ANET, the Proxy sends NS, 2179 RS and/or unsolicited NA messages to update the Server's neighbor 2180 cache entries on behalf of the Client and/or to convey QoS updates. 2181 This allows for higher-frequency Proxy-initiated RS/RA messaging over 2182 well-connected INET infrastructure supplemented by lower-frequency 2183 Client-initiated RS/RA messaging over constrained ANET data links. 2185 If the Server ceases to send solicited advertisements, the Proxy 2186 sends unsolicited RAs on the ANET interface with destination set to 2187 (link-local) All-Nodes multicast and with Router Lifetime set to zero 2188 to inform Clients that the Server has failed. Although the Proxy 2189 engages in ND exchanges on behalf of the Client, the Client can also 2190 send ND messages on its own behalf, e.g., if it is in a better 2191 position than the Proxy to convey QoS changes, etc. For this reason, 2192 the Proxy marks any Client-originated solicitation messages (e.g. by 2193 inserting a Nonce option) so that it can return the solicited 2194 advertisement to the Client instead of processing it locally. 2196 If the Client becomes unreachable, the Proxy sets the neighbor cache 2197 entry state to DEPARTED and retains the entry for DepartTime seconds. 2198 While the state is DEPARTED, the Proxy forwards any packets destined 2199 to the Client to a Bridge via OAL/ORH encapsulation. When DepartTime 2200 expires, the Proxy deletes the neighbor cache entry and discards any 2201 further packets destined to this (now forgotten) Client. 2203 In some ANETs that employ a Proxy, the Client's MNP can be injected 2204 into the ANET routing system. In that case, the Client can send data 2205 messages without encapsulation so that the ANET routing system 2206 transports the unencapsulated packets to the Proxy. This can be very 2207 beneficial, e.g., if the Client connects to the ANET via low-end data 2208 links such as some aviation wireless links. 2210 If the first-hop ANET access router is on the same underlying link 2211 and recognizes the AERO/OMNI protocol, the Client can avoid 2212 encapsulation for both its control and data messages. When the 2213 Client connects to the link, it can send an unencapsulated RS message 2214 with source address set to its own MNP-LLA (or to a Temporary LLA), 2215 and with destination address set to the ADM-LLA of the Client's 2216 selected Server or to (link-local) All-Routers multicast. The Client 2217 includes an OMNI option formatted as specified in 2218 [I-D.templin-6man-omni-interface]. 2220 The Client then sends the unencapsulated RS message, which will be 2221 intercepted by the AERO-Aware access router. The access router then 2222 encapsulates the RS message in an ANET header with its own address as 2223 the source address and the address of a Proxy as the destination 2224 address. The access router further remembers the address of the 2225 Proxy so that it can encapsulate future data packets from the Client 2226 via the same Proxy. If the access router needs to change to a new 2227 Proxy, it simply sends another RS message toward the Server via the 2228 new Proxy on behalf of the Client. 2230 In some cases, the access router and Proxy may be one and the same 2231 node. In that case, the node would be located on the same physical 2232 link as the Client, but its message exchanges with the Server would 2233 need to pass through a security gateway at the ANET/INET border. The 2234 method for deploying access routers and Proxys (i.e. as a single node 2235 or multiple nodes) is an ANET-local administrative consideration. 2237 Note: The Proxy can apply packing as discussed in Section 5.2 of 2238 [I-D.templin-6man-omni-interface] if an opportunity arises to 2239 concatenate multiple payload packets that will be destined to the 2240 same neighbor. 2242 3.13.1. Combined Proxy/Servers 2244 Clients may need to connect directly to Servers via INET, Direct and 2245 VPNed interfaces (i.e., non-ANET interfaces). If the Client's 2246 underlying interfaces all connect via the same INET partition, then 2247 it can connect to a single controlling Server via all interfaces. 2249 If some Client interfaces connect via different INET partitions, 2250 however, the Client still selects a set of controlling Servers and 2251 sends RS messages via their directly-connected Servers while using 2252 the ADM-LLA of the controlling Server as the destination. 2254 When a Server receives an RS with destination set to the ADM-LLA of a 2255 controlling Server, it acts as a Proxy to forward the message to the 2256 controlling Server while forwarding the corresponding RA reply to the 2257 Client. 2259 3.13.2. Detecting and Responding to Server Failures 2261 In environments where fast recovery from Server failure is required, 2262 Proxys SHOULD use proactive Neighbor Unreachability Detection (NUD) 2263 to track Server reachability in a similar fashion as for 2264 Bidirectional Forwarding Detection (BFD) [RFC5880]. Proxys can then 2265 quickly detect and react to failures so that cached information is 2266 re-established through alternate paths. The NUD control messaging is 2267 carried only over well-connected ground domain networks (i.e., and 2268 not low-end aeronautical radio links) and can therefore be tuned for 2269 rapid response. 2271 Proxys perform proactive NUD with Servers for which there are 2272 currently active ANET Clients by sending continuous NS messages in 2273 rapid succession, e.g., one message per second. The Proxy sends the 2274 NS message via the spanning tree with the Proxy's ADM-LLA as the 2275 source and the ADM-LLA of the Server as the destination. When the 2276 Proxy is also sending RS messages to the Server on behalf of ANET 2277 Clients, the resulting RA responses can be considered as equivalent 2278 hints of forward progress. This means that the Proxy need not also 2279 send a periodic NS if it has already sent an RS within the same 2280 period. If the Server fails (i.e., if the Proxy ceases to receive 2281 advertisements), the Proxy can quickly inform Clients by sending 2282 multicast RA messages on the ANET interface. 2284 The Proxy sends RA messages on the ANET interface with source address 2285 set to the Server's address, destination address set to (link-local) 2286 All-Nodes multicast, and Router Lifetime set to 0. The Proxy SHOULD 2287 send MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small 2288 delays [RFC4861]. Any Clients on the ANET that had been using the 2289 failed Server will receive the RA messages and associate with a new 2290 Server. 2292 3.13.3. Point-to-Multipoint Server Coordination 2294 In environments where Client messaging over ANETs is bandwidth- 2295 limited and/or expensive, Clients can enlist the services of the 2296 Proxy to coordinate with multiple Servers in a single RS/RA message 2297 exchange. The Client can send a single RS message to (link-local) 2298 All-Routers multicast that includes the ID's of multiple Servers in 2299 MS-Register sub-options of the OMNI option. 2301 When the Proxy receives the RS and processes the OMNI option, it 2302 sends a separate RS to each MS-Register Server ID. When the Proxy 2303 receives an RA, it can optionally return an immediate "singleton" RA 2304 to the Client or record the Server's ID for inclusion in a pending 2305 "aggregate" RA message. The Proxy can then return aggregate RA 2306 messages to the Client including multiple Server IDs in order to 2307 conserve bandwidth. Each RA includes a proper subset of the Server 2308 IDs from the original RS message, and the Proxy must ensure that the 2309 message contents of each RA are consistent with the information 2310 received from the (aggregated) Servers. 2312 Clients can thereafter employ efficient point-to-multipoint Server 2313 coordination under the assistance of the Proxy to reduce the number 2314 of messages sent over the ANET while enlisting the support of 2315 multiple Servers for fault tolerance. Clients can further include 2316 MS-Release sub-options in IPv6 ND messages to request the Proxy to 2317 release from former Servers via the procedures discussed in 2318 Section 3.16.5. 2320 The OMNI interface specification [I-D.templin-6man-omni-interface] 2321 provides further discussion of the Client/Proxy RS/RA messaging 2322 involved in point-to-multipoint coordination. 2324 3.14. AERO Address Resolution 2326 While data packets are flowing between a source and target node, 2327 route optimization SHOULD be used. Route optimization is initiated 2328 by the first eligible Route Optimization Source (ROS) closest to the 2329 source as follows: 2331 o For Clients on VPNed and Direct interfaces, the Server is the ROS. 2333 o For Clients on ANET interfaces, the Proxy is the ROS. 2335 o For Clients on INET interfaces, the Client itself is the ROS. 2337 o For correspondent nodes on INET/EUN interfaces serviced by a 2338 Relay, the Relay is the ROS. 2340 The route optimization procedure is conducted between the ROS and 2341 with the target Server/Relay or the target Client itself acting as a 2342 Route Optimization Responder (ROR) in the same manner as for IPv6 ND 2343 Address Resolution and using the same NS/NA messaging. The target 2344 may either be a MNP Client serviced by a Server, or a non-MNP 2345 correspondent reachable via a Relay. 2347 The procedures are specified in the following sections. 2349 3.14.1. Route Optimization Initiation 2351 While original IP packet from the source node destined a target node 2352 arrives, the ROS sends an NS message for Address Resolution (NS(AR)) 2353 to receive a solicited NA(AR) message from a ROR. The NS(AR) message 2354 must be no larger than the OAL minimum MPS so that its entire 2355 contents will fit in the first fragment (i.e., as an "atomic 2356 fragment"). When the ROS sends an NS(AR), it first creates a 2357 neighbor cache entry for the target in the INCOMPLETE state for the 2358 target and places the original IP packet on a short queue. The ROS 2359 then prepares an NS(AR) that includes: 2361 o the LLA of the ROS as the source address. 2363 o the LLA corresponding to the original IP packet's destination as 2364 the Target Address, e.g., for 2001:db8:1:2::10:2000 the LLA is 2365 fe80::2001:db8:1:2. 2367 o the Solicited-Node multicast address [RFC4291] formed from the 2368 lower 24 bits of the original IP packet's destination as the 2369 destination address, e.g., for 2001:db8:1:2::10:2000 the NS 2370 destination address is ff02:0:0:0:0:1:ff10:2000. 2372 The NS(AR) message also includes an OMNI option with no Interface 2373 Attributes, such that the target will not create a neighbor cache 2374 entry. The Prefix Length in the OMNI option is set to the Prefix 2375 Length associated with the ROS's LLA. 2377 The ROS then submits the NS(AR) message for OAL encapsulation and 2378 fragmentation, with OAL source set to its own ULA and OAL destination 2379 set to the MNP-ULA corresponding to the target, then sends the 2380 resulting *NET packet (i.e., the "atomic fragment") into the spanning 2381 tree without decrementing the network-layer TTL/Hop Limit field. 2382 (When the ROS is a Client, it instead securely sends the *NET packet 2383 to one of its current Servers as specified in Section 3.22. The 2384 Server then forwards the *NET packet into the spanning tree on behalf 2385 of the Client.) 2387 3.14.2. Relaying the NS(AR) *NET Packet(s) 2389 When the Bridge receives the *NET packet from the ROS, it discards 2390 the *NET header and determines the next hop by consulting its 2391 standard IPv6 forwarding table for the OAL header destination 2392 address. The Bridge then re-encapsulates and forwards the *NET 2393 packet via the spanning tree the same as for any IPv6 router, where 2394 it may traverse multiple OMNI link segments. The final-hop Bridge in 2395 the spanning tree will deliver the *NET packet via a secured tunnel 2396 to a Server or Relay that services the target. 2398 3.14.3. Processing the NS(AR) and Sending the NA(AR) 2400 When the Server or Relay receives the *NET packet, it examines the 2401 OAL atomic fragment to determine that it contains an NS(AR), then 2402 examines the NS(AR) destination to determine whether it has a 2403 matching neighbor cache entry and/or route. If there is no match, 2404 the ROR drops the message. Otherwise, the ROR continues processing 2405 as follows: 2407 o if the destination matches a Client neighbor in the DEPARTED 2408 state, the Server inserts an ORH with destination prefix set to 2409 the lower 64 bits of the MNP-ULA and sets the destination address 2410 to the ADM-ULA of the Client's new Server. The (old) Server then 2411 re-encapsulates the *NET packet, forwards it into the spanning 2412 tree and returns from processing. 2414 o If the NS(AR) destination matches an MNP Client neighbor cache 2415 entry in the REACHABLE state, the ROR instead proceeds according 2416 to whether the Client requires Proxy services. For 2417 "disadvantaged" Clients (e.g., those on low-end links) and Clients 2418 with complex multilink arrangements, the Server nominates itself 2419 as the ROR and adds the NS(AR) source address to the target 2420 Client's Report List with time set to ReportTime. Otherwise, the 2421 Server re-encapsulates the *NET packet and forwards it to the 2422 target Client which will act as a ROR on its own behalf. 2424 o If the NS(AR) belongs to a non-MNP route, the ROR continues 2425 processing without adding an entry to the Report List. 2427 The ROR then prepares a (solicited) NA(AR) message to send back to 2428 the ROS but does not create a neighbor cache entry. The ROR sets the 2429 NA(AR) source address to its own LLA, sets the destination address to 2430 the source of the solicitation and sets the Target Address to the 2431 target of the solicitation. The ROR then includes an OMNI option 2432 with Prefix Length set to the length associated with the MNP-LLA. 2434 If the target is an MNP Client, the ROR next includes Interface 2435 Attributes in the OMNI option for each of the target Client's 2436 underlying interfaces with current information for each interface and 2437 with the S/T-ifIndex field in the OMNI header set to 0. 2439 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2440 according to its own INET address for VPNed or Direct interfaces, to 2441 the *NET address of the Proxy or to the Client's *NET address for 2442 INET interfaces. The ROR then includes the lower 32 bits of its own 2443 ADM-ULA (or the ADM-ULA of the Proxy/Server) as the LHS, encodes the 2444 ADM-ULA prefix length code in the SRT field and sets the FMT code 2445 accordingly as specified in Section 3.3. 2447 The ROR then sets the NA(AR) message R flag to 1 (as a router) and S 2448 flag to 1 (as a response to a solicitation). If the ROR is the 2449 Client itself, it sets the O flag to 1; if the ROR is the Server, it 2450 instead sets the O flag to 0 (as a proxy). The ROR finally submits 2451 the NA(AR) for OAL encapsulation with source set to its own ULA and 2452 destination set to the source ULA of the NS(AR) message, then 2453 performs *NET fragmentation and forwards the resulting *NET packets 2454 into the spanning tree without decrementing the network-layer TTL/Hop 2455 Limit field. 2457 3.14.4. Relaying the NA(AR) 2459 When the Bridge receives the *NET packets from the ROR, it discards 2460 the *NET header and determines the next hop by consulting its 2461 standard IPv6 forwarding table for the OAL header destination 2462 address. The Bridge then re-encapsulates and forwards the *NET 2463 packet via the spanning tree the same as for any IPv6 router, where 2464 it may traverse multiple OMNI link segments. The final-hop Bridge in 2465 the spanning tree will deliver the *NET packet via a secured tunnel 2466 to a Server for the ROS. 2468 3.14.5. Processing the NA(AR) 2470 When the ROS receives the NA(AR) message, it processes the message 2471 the same as for standard IPv6 Address Resolution [RFC4861]. In the 2472 process, it caches the source ULA and all information found in the 2473 OMNI option in the neighbor cache entry for the target. The ROS 2474 finally sets the neighbor cache entry state to REACHABLE and sets its 2475 lifetime to ReachableTime seconds. (When the ROS is a Client, the 2476 solicited NA(AR) message will first be delivered via the spanning 2477 tree to one of its current Servers, which then securely forwards the 2478 message to the Client as discussed in Section 3.22.) 2480 3.14.6. Route Optimization Maintenance 2482 Following route optimization, the ROS forwards future data packets 2483 destined to the target via the addresses found in the cached link- 2484 layer information. The route optimization is shared by all sources 2485 that send packets to the target via the ROS, i.e., and not just the 2486 source on behalf of which the route optimization was initiated. 2488 While new data packets destined to the target are flowing through the 2489 ROS, it sends additional NS(AR) messages to the ROR before 2490 ReachableTime expires to receive a fresh NA(AR) message the same as 2491 described in the previous sections (route optimization refreshment 2492 strategies are an implementation matter, with a non-normative example 2493 given in Appendix A.1). The ROS uses the cached ULA of the ROR 2494 (i.e., either the ADM-ULA of the Server or the MNP-ULA of the Client 2495 itself) as the NS(AR) OAL destination address, and sends up to 2496 MAX_MULTICAST_SOLICIT NS(AR) messages separated by 1 second until an 2497 NA(AR) is received. If no NA(AR) is received, the ROS assumes that 2498 the current ROR has become unreachable and deletes the target 2499 neighbor cache entry. Subsequent data packets will trigger a new 2500 route optimization event (see: Section 3.14.1). 2502 If an NA(AR) is received, the ROS then updates the neighbor cache 2503 entry to refresh ReachableTime, while (for MNP destinations) the ROR 2504 adds or updates the ROS address to the target's Report List and with 2505 time set to ReportTime. While no data packets are flowing, the ROS 2506 instead allows ReachableTime for the neighbor cache entry to expire. 2507 When ReachableTime expires, the ROS deletes the neighbor cache entry. 2508 Any future data packets flowing through the ROS will again trigger a 2509 new route optimization. 2511 The ROS may also receive unsolicited NA messages from the ROR at any 2512 time (see: Section 3.16). If there is a neighbor cache entry for the 2513 target, the ROS updates the link-layer information but does not 2514 update ReachableTime since the receipt of an unsolicited NA does not 2515 confirm that any forward paths are working. If there is no neighbor 2516 cache entry, the ROS simply discards the unsolicited NA. 2518 In this arrangement, the ROS holds a neighbor cache entry for the 2519 target via the ROR, but the ROR does not hold a neighbor cache entry 2520 for the ROS. The route optimization neighbor relationship is 2521 therefore asymmetric and unidirectional. If the target node also has 2522 packets to send back to the source node, then a separate route 2523 optimization procedure is performed in the reverse direction. But, 2524 there is no requirement that the forward and reverse paths be 2525 symmetric. 2527 3.15. Neighbor Unreachability Detection (NUD) 2529 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2530 [RFC4861] either reactively in response to persistent link-layer 2531 errors (see Section 3.11) or proactively to confirm reachability and/ 2532 or establish NAT state. The NUD algorithm is based on periodic 2533 control message exchanges. The algorithm may further be seeded by ND 2534 hints of forward progress, but care must be taken to avoid inferring 2535 reachability based on spoofed information. For example, authentic 2536 IPv6 ND message exchanges may be considered as acceptable hints of 2537 forward progress, while spurious data packets should not be. 2539 AERO nodes can use standard NS/NA exchanges sent over the OMNI link 2540 spanning tree to securely test reachability without risk of DoS 2541 attacks from nodes pretending to be a neighbor (these NS/NA(NUD) 2542 messages use the unicast LLAs and ULAs of the two parties involved in 2543 the NUD test the same as for standard IPv6 ND, and both messages flow 2544 over the spanning tree). Proxys can further perform NUD to securely 2545 verify Server reachability on behalf of their proxyed Clients. 2546 However, a means for an ROS to test the unsecured target route 2547 optimized paths is also necessary. 2549 When an ROR directs an ROS to a target neighbor with one or more 2550 link-layer addresses, the ROS can proactively test each such 2551 unsecured route optimized path through secured NS(NUD) messages over 2552 the spanning tree that invoke an unsecured NA(NUD) reply that travels 2553 over the route optimized path.. (The NS(NUD) messages must therefore 2554 include Nonce and Timestamp options that will be echoed in the 2555 unsecured NA(NUD) replies.) While testing the paths, the ROS can 2556 optionally continue to send packets via the spanning tree, maintain a 2557 small queue of packets until target reachability is confirmed, or 2558 (optimistically) allow packets to flow via the route optimized paths. 2560 When the ROS sends an NS(NUD) message, it sets the IPv6 source to its 2561 own address, sets the destination to the MNP-LLA of the target, and 2562 sets the target's MNP Subnet-Router anycast address as the Target 2563 Address. The ROS also includes an OMNI option with a single 2564 Interface Attributes sub-option with the SRT, FMT, LHS and L2ADDR 2565 information for its own underlying interface it wishes to test, but 2566 sets the S/T-ifIndex field to the index for target's underlying 2567 interface to be tested. The ROS includes a Nonce and Timestamp 2568 option, then encapsulates the message in OAL/INET headers with its 2569 own ULA as the source and the ULA of the target as the destination. 2570 The ROS then forwards the NS(NUD) message toward the target via a 2571 Server or Bridge. 2573 When the target receives the NS(NUD) message, it creates an NA(NUD) 2574 by reversing the OAL and IPv6 addresses and including an Interface 2575 Attributes sub-option with attributes for its own interface 2576 identified by the NS(NUD) S/T-ifIndex. The target sets the NA(NUD) 2577 S/T-ifIndex to the index of the ROS, sets the Target Address to the 2578 same value that was in the NS(NUD), and returns the message using its 2579 own underlying interface identified by S/T-ifIndex and destined to 2580 the ROS's interface identified by the original Interface Attributes 2581 sub-option. 2583 When the ROS receives the NA(NUD) message, it can determine from the 2584 Nonce, Timestamp and Target Address that the message matched its 2585 NS(NUD) and that it transited the direct path from the ROR using the 2586 selected underlying interface pair. The ROS marks route optimization 2587 target paths that pass these NUD tests as "reachable", and those that 2588 do not as "unreachable". These markings inform the OMNI interface 2589 forwarding algorithm specified in Section 3.10. 2591 Note: If the target determines that the OMNI option Interface 2592 Attributes in the NS(NUD) is located in a different OMNI link segment 2593 than its own interface named in the S/T-ifIndex, it instead returns 2594 the NA(NUD) via the spanning tree while including an ORH and setting 2595 the OAL destination address to the Subnet Router Anycast address used 2596 by Bridges on the ROS segment. When a Bridge on the ROS segment 2597 receives the NA(NUD), it replaces the Interface Attributes with 2598 information for its own interface while using the ifIndex value 2599 specific to the target. 2601 3.16. Mobility Management and Quality of Service (QoS) 2603 AERO is a Distributed Mobility Management (DMM) service. Each Server 2604 is responsible for only a subset of the Clients on the OMNI link, as 2605 opposed to a Centralized Mobility Management (CMM) service where 2606 there is a single network mobility collective entity for all Clients. 2607 Clients coordinate with their associated Servers via RS/RA exchanges 2608 to maintain the DMM profile, and the AERO routing system tracks all 2609 current Client/Server peering relationships. 2611 Servers provide default routing and mobility/multilink services for 2612 their dependent Clients. Clients are responsible for maintaining 2613 neighbor relationships with their Servers through periodic RS/RA 2614 exchanges, which also serves to confirm neighbor reachability. When 2615 a Client's underlying interface address and/or QoS information 2616 changes, the Client is responsible for updating the Server with this 2617 new information. Note that when there is a Proxy in the path, the 2618 Proxy can also perform some RS/RA exchanges on the Client's behalf. 2620 Mobility management messaging is based on the transmission and 2621 reception of unsolicited Neighbor Advertisement (uNA) messages. Each 2622 uNA message sets the IPv6 destination address to (link-local) All- 2623 Nodes multicast to convey a general update of Interface Attributes to 2624 (possibly) multiple recipients, or to a specific unicast LLA to 2625 announce a departure event to a specific recipient. Implementations 2626 must therefore examine the destination address to determine the 2627 nature of the mobility event (i.e., update vs departure). 2629 Mobility management considerations are specified in the following 2630 sections. 2632 3.16.1. Mobility Update Messaging 2634 Servers accommodate Client mobility, multilink and/or QoS change 2635 events by sending unsolicited NA (uNA) messages to each ROS in the 2636 target Client's Report List. When a Server sends a uNA message, it 2637 sets the IPv6 source address to the Client's MNP-LLA, sets the 2638 destination address to (link-local) All-Nodes multicast and sets the 2639 Target Address to the Client's Subnet-Router anycast address. The 2640 Server also includes an OMNI option with Prefix Length set to the 2641 length associated with the Client's MNP-LLA, with Interface 2642 Attributes for the target Client's underlying interfaces and with the 2643 OMNI header S/T-ifIndex set to 0. The Server then sets the NA R flag 2644 to 1, the S flag to 0 and the O flag to 1, then encapsulates the 2645 message in an OAL header with source set to its own ADM-ULA and 2646 destination set to the ULA of the ROS and sends the message into the 2647 spanning tree. 2649 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2650 reception of uNA messages is unreliable but provides a useful 2651 optimization. In well-connected Internetworks with robust data links 2652 uNA messages will be delivered with high probability, but in any case 2653 the Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT uNAs 2654 to each ROS to increase the likelihood that at least one will be 2655 received. 2657 When the ROS receives a uNA message prepared as above, it ignores the 2658 message if there is no existing neighbor cache entry for the Client. 2659 Otherwise, it uses the included OMNI option information to update the 2660 neighbor cache entry, but does not reset ReachableTime since the 2661 receipt of an unsolicited NA message from the target Server does not 2662 provide confirmation that any forward paths to the target Client are 2663 working. 2665 If uNA messages are lost, the ROS may be left with stale address and/ 2666 or QoS information for the Client for up to ReachableTime seconds. 2667 During this time, the ROS can continue sending packets according to 2668 its stale neighbor cache information. When ReachableTime is close to 2669 expiring, the ROS will re-initiate route optimization and receive 2670 fresh link-layer address information. 2672 In addition to sending uNA messages to the current set of ROSs for 2673 the Client, the Server also sends uNAs to the MNP-ULA associated with 2674 the link-layer address for any underlying interface for which the 2675 link-layer address has changed. These uNA messages update an old 2676 Proxy/Server that cannot easily detect (e.g., without active probing) 2677 when a formerly-active Client has departed. When the Server sends 2678 the uNA, it sets the IPv6 source address to the Client's MNP-LLA, 2679 sets the destination address to the old Proxy/Server's ADM-LLA, and 2680 sets the Target Address to the Client's Subnet-Router anycast 2681 address. The Server also includes an OMNI option with Prefix Length 2682 set to the length associated with the Client's MNP-LLA, with 2683 Interface Attributes for the changed underlying interface, and with 2684 the OMNI header S/T-ifIndex set to 0. The Server then sets the NA R 2685 flag to 1, the S flag to 0 and the O flag to 1, then encapsulates the 2686 message in an OAL header with source set to its own ADM-ULA and 2687 destination set to the ADM-ULA of the old Proxy/Server and sends the 2688 message into the spanning tree. 2690 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2692 When a Client needs to change its underlying interface addresses and/ 2693 or QoS preferences (e.g., due to a mobility event), either the Client 2694 or its Proxys send RS messages to the Server via the spanning tree 2695 with an OMNI option that includes Interface attributes with the new 2696 link quality and address information. 2698 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2699 sending actual data packets in case one or more RAs are lost. If all 2700 RAs are lost, the Client SHOULD re-associate with a new Server. 2702 When the Server receives the Client's changes, it sends uNA messages 2703 to all nodes in the Report List the same as described in the previous 2704 section. 2706 3.16.3. Bringing New Links Into Service 2708 When a Client needs to bring new underlying interfaces into service 2709 (e.g., when it activates a new data link), it sends an RS message to 2710 the Server via the underlying interface with an OMNI option that 2711 includes Interface Attributes with appropriate link quality values 2712 and with link-layer address information for the new link. 2714 3.16.4. Deactivating Existing Links 2716 When a Client needs to deactivate an existing underlying interface, 2717 it sends an RS or uNA message to its Server with an OMNI option with 2718 appropriate Interface Attribute values - in particular, the link 2719 quality value 0 assures that neighbors will cease to use the link. 2721 If the Client needs to send RS/uNA messages over an underlying 2722 interface other than the one being deactivated, it MUST include 2723 Interface Attributes with appropriate link quality values for any 2724 underlying interfaces being deactivated. 2726 Note that when a Client deactivates an underlying interface, 2727 neighbors that have received the RS/uNA messages need not purge all 2728 references for the underlying interface from their neighbor cache 2729 entries. The Client may reactivate or reuse the underlying interface 2730 and/or its ifIndex at a later point in time, when it will send RS/uNA 2731 messages with fresh Interface Attributes to update any neighbors. 2733 3.16.5. Moving Between Servers 2735 The Client performs the procedures specified in Section 3.12.2 when 2736 it first associates with a new Server or renews its association with 2737 an existing Server. The Client also includes MS-Release identifiers 2738 in the RS message OMNI option per [I-D.templin-6man-omni-interface] 2739 if it wants the new Server to notify any old Servers from which the 2740 Client is departing. 2742 When the new Server receives the Client's RS message, it returns an 2743 RA as specified in Section 3.12.3 and sends up to 2744 MAX_NEIGHBOR_ADVERTIISEMENT uNA messages to any old Servers listed in 2745 OMNI option MS-Release identifiers. When the new Server sends a uNA 2746 message, it sets the IPv6 source address to the Client's MNP-LLA, 2747 sets the destination address to the old Server's ADM-LLA, and sets 2748 the Target Address to the Client's Subnet-Router anycast address. 2749 The new Server also includes an OMNI option with Prefix Length set to 2750 the length associated with the Client's MNP-LLA, with Interface 2751 Attributes for its own underlying interface, and with the OMNI header 2752 S/T-ifIndex set to 0. The new Server then sets the NA R flag to 1, 2753 the S flag to 0 and the O flag to 1, then encapsulates the message in 2754 an OAL header with source set to its own ADM-ULA and destination set 2755 to the ADM-ULA of the old Server and sends the message into the 2756 spanning tree. 2758 When an old Server receives the uNA, it changes the Client's neighbor 2759 cache entry state to DEPARTED, sets the link-layer address of the 2760 Client to the new Server's ADM-ULA, and resets DepartTime. After a 2761 short delay (e.g., 2 seconds) the old Server withdraws the Client's 2762 MNP from the routing system. After DepartTime expires, the old 2763 Server deletes the Client's neighbor cache entry. 2765 The old Server also iteratively forwards a copy of the uNA message to 2766 each ROS in the Client's Report List by changing the OAL destination 2767 address to the ULA of the ROS while leaving all other fields of the 2768 message unmodified. When the ROS receives the uNA, it examines the 2769 Target address to determine the correct neighbor cache entry and 2770 verifies that the IPv6 destination address matches the old Server. 2771 The ROS then caches the IPv6 source address as the new Server for the 2772 existing neighbor cache entry and marks the entry as STALE. While in 2773 the STALE state, the ROS allows new data packets to flow according to 2774 any existing cached link-layer information and sends new NS(AR) 2775 messages using its own ULA as the OAL source and the ADM-ULA of the 2776 new Server as the OAL destination address to elicit NA messages that 2777 reset the neighbor cache entry state to REACHABLE. If no new NA 2778 message is received for 10 seconds while in the STALE state, the ROS 2779 deletes the neighbor cache entry. 2781 Clients SHOULD NOT move rapidly between Servers in order to avoid 2782 causing excessive oscillations in the AERO routing system. Examples 2783 of when a Client might wish to change to a different Server include a 2784 Server that has gone unreachable, topological movements of 2785 significant distance, movement to a new geographic region, movement 2786 to a new OMNI link segment, etc. 2788 When a Client moves to a new Server, some of the fragments of a 2789 multiple fragment packet may have already arrived at the old Server 2790 while others are en route to the new Server, however no special 2791 attention in the reassembly algorithm is necessary when re-routed 2792 fragments are simply treated as loss. 2794 3.17. Multicast 2796 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2797 [RFC3810] proxy service for its EUNs and/or hosted applications 2798 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2799 underlying interfaces for which group membership is required. The 2800 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2801 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2802 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2803 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2804 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 2805 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2806 INET/EUN networks. The behaviors identified in the following 2807 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2808 Multicast (ASM) operational modes. 2810 3.17.1. Source-Specific Multicast (SSM) 2812 When an ROS (i.e., an AERO Proxy/Server/Relay) "X" acting as PIM 2813 router receives a Join/Prune message from a node on its downstream 2814 interfaces containing one or more ((S)ource, (G)roup) pairs, it 2815 updates its Multicast Routing Information Base (MRIB) accordingly. 2816 For each S belonging to a prefix reachable via X's non-OMNI 2817 interfaces, X then forwards the (S, G) Join/Prune to any PIM routers 2818 on those interfaces per [RFC7761]. 2820 For each S belonging to a prefix reachable via X's OMNI interface, X 2821 originates a separate copy of the Join/Prune for each (S,G) in the 2822 message using its own LLA as the source address and ALL-PIM-ROUTERS 2823 as the destination address. X then encapsulates each message in an 2824 OAL header with source address set to the ULA of X and destination 2825 address set to S then forwards the message into the spanning tree, 2826 which delivers it to AERO Server/Relay "Y" that services S. At the 2827 same time, if the message was a Join, X sends a route-optimization NS 2828 message toward each S the same as discussed in Section 3.14. The 2829 resulting NAs will return the LLA for the prefix that matches S as 2830 the network-layer source address and with an OMNI option with the ULA 2831 corresponding to any underlying interfaces that are currently 2832 servicing S. 2834 When Y processes the Join/Prune message, if S located behind any 2835 INET, Direct, or VPNed interfaces Y acts as a PIM router and updates 2836 its MRIB to list X as the next hop in the reverse path. If S is 2837 located behind any Proxys "Z"*, Y also forwards the message to each 2838 Z* over the spanning tree while continuing to use the LLA of X as the 2839 source address. Each Z* then updates its MRIB accordingly and 2840 maintains the LLA of X as the next hop in the reverse path. Since 2841 the Bridges do not examine network layer control messages, this means 2842 that the (reverse) multicast tree path is simply from each Z* (and/or 2843 Y) to X with no other multicast-aware routers in the path. If any Z* 2844 (and/or Y) is located on the same OMNI link segment as X, the 2845 multicast data traffic sent to X directly using OAL/INET 2846 encapsulation instead of via a Bridge. 2848 Following the initial Join/Prune and NS/NA messaging, X maintains a 2849 neighbor cache entry for each S the same as if X was sending unicast 2850 data traffic to S. In particular, X performs additional NS/NA 2851 exchanges to keep the neighbor cache entry alive for up to t_periodic 2852 seconds [RFC7761]. If no new Joins are received within t_periodic 2853 seconds, X allows the neighbor cache entry to expire. Finally, if X 2854 receives any additional Join/Prune messages for (S,G) it forwards the 2855 messages to each Y and Z* in the neighbor cache entry over the 2856 spanning tree. 2858 At some later time, Client C that holds an MNP for source S may 2859 depart from a first Proxy Z1 and/or connect via a new Proxy Z2. In 2860 that case, Y sends an unsolicited NA message to X the same as 2861 specified for unicast mobility in Section 3.16. When X receives the 2862 unsolicited NA message, it updates its neighbor cache entry for the 2863 LLA for source S and sends new Join messages to any new Proxys Z2. 2864 There is no requirement to send any Prune messages to old Proxys Z1 2865 since source S will no longer source any multicast data traffic via 2866 Z1. Instead, the multicast state for (S,G) in Proxy Z1 will soon 2867 time out since no new Joins will arrive. 2869 After some later time, C may move to a new Server Y2 and depart from 2870 old Sever Y1. In that case, Y1 sends Join messages for any of C's 2871 active (S,G) groups to Y2 while including its own LLA as the source 2872 address. This causes Y2 to include Y1 in the multicast forwarding 2873 tree during the interim time that Y1's neighbor cache entry for C is 2874 in the DEPARTED state. At the same time, Y1 sends an unsolicited NA 2875 message to X with an OMNI option with S/T-ifIndex in the header set 2876 to 0 and a release indication to cause X to release its neighbor 2877 cache entry. X then sends a new Join message to S via the spanning 2878 tree and re-initiates route optimization the same as if it were 2879 receiving a fresh Join message from a node on a downstream link. 2881 3.17.2. Any-Source Multicast (ASM) 2883 When an ROS X acting as a PIM router receives a Join/Prune from a 2884 node on its downstream interfaces containing one or more (*,G) pairs, 2885 it updates its Multicast Routing Information Base (MRIB) accordingly. 2886 X then forwards a copy of the message to the Rendezvous Point (RP) R 2887 for each G over the spanning tree. X uses its own LLA as the source 2888 address and ALL-PIM-ROUTERS as the destination address, then 2889 encapsulates each message in an OAL header with source address set to 2890 the ULA of X and destination address set to R, then sends the message 2891 into the spanning tree. At the same time, if the message was a Join 2892 X initiates NS/NA route optimization the same as for the SSM case 2893 discussed in Section 3.17.1. 2895 For each source S that sends multicast traffic to group G via R, the 2896 Proxy/Server Z* for the Client that aggregates S encapsulates the 2897 packets in PIM Register messages and forwards them to R via the 2898 spanning tree, which may then elect to send a PIM Join to Z*. This 2899 will result in an (S,G) tree rooted at Z* with R as the next hop so 2900 that R will begin to receive two copies of the packet; one native 2901 copy from the (S, G) tree and a second copy from the pre-existing (*, 2902 G) tree that still uses PIM Register encapsulation. R can then issue 2903 a PIM Register-stop message to suppress the Register-encapsulated 2904 stream. At some later time, if C moves to a new Proxy/Server Z*, it 2905 resumes sending packets via PIM Register encapsulation via the new 2906 Z*. 2908 At the same time, as multicast listeners discover individual S's for 2909 a given G, they can initiate an (S,G) Join for each S under the same 2910 procedures discussed in Section 3.17.1. Once the (S,G) tree is 2911 established, the listeners can send (S, G) Prune messages to R so 2912 that multicast packets for group G sourced by S will only be 2913 delivered via the (S, G) tree and not from the (*, G) tree rooted at 2914 R. All mobility considerations discussed for SSM apply. 2916 3.17.3. Bi-Directional PIM (BIDIR-PIM) 2918 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 2919 approach to ASM that treats the Rendezvous Point (RP) as a Designated 2920 Forwarder (DF). Further considerations for BIDIR-PIM are out of 2921 scope. 2923 3.18. Operation over Multiple OMNI Links 2925 An AERO Client can connect to multiple OMNI links the same as for any 2926 data link service. In that case, the Client maintains a distinct 2927 OMNI interface for each link, e.g., 'omni0' for the first link, 2928 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 2929 would include its own distinct set of Bridges, Servers and Proxys, 2930 thereby providing redundancy in case of failures. 2932 Each OMNI link could utilize the same or different ANET connections. 2933 The links can be distinguished at the link-layer via the SRT prefix 2934 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 2935 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 2936 MSPs on each link. This gives rise to the opportunity for supporting 2937 multiple redundant networked paths, with each VLAN distinguished by a 2938 different SRT "color" (see: Section 3.2.5). 2940 The Client's IP layer can select the outgoing OMNI interface 2941 appropriate for a given traffic profile while (in the reverse 2942 direction) correspondent nodes must have some way of steering their 2943 packets destined to a target via the correct OMNI link. 2945 In a first alternative, if each OMNI link services different MSPs, 2946 then the Client can receive a distinct MNP from each of the links. 2947 IP routing will therefore assure that the correct OMNI link is used 2948 for both outbound and inbound traffic. This can be accomplished 2949 using existing technologies and approaches, and without requiring any 2950 special supporting code in correspondent nodes or Bridges. 2952 In a second alternative, if each OMNI link services the same MSP(s) 2953 then each link could assign a distinct "OMNI link Anycast" address 2954 that is configured by all Bridges on the link. Correspondent nodes 2955 can then perform Segment Routing to select the correct SRT, which 2956 will then direct the packet over multiple hops to the target. 2958 3.19. DNS Considerations 2960 AERO Client MNs and INET correspondent nodes consult the Domain Name 2961 System (DNS) the same as for any Internetworking node. When 2962 correspondent nodes and Client MNs use different IP protocol versions 2963 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 2964 A records for IPv4 address mappings to MNs which must then be 2965 populated in Relay NAT64 mapping caches. In that way, an IPv4 2966 correspondent node can send packets to the IPv4 address mapping of 2967 the target MN, and the Relay will translate the IPv4 header and 2968 destination address into an IPv6 header and IPv6 destination address 2969 of the MN. 2971 When an AERO Client registers with an AERO Server, the Server can 2972 return the address(es) of DNS servers in RDNSS options [RFC6106]. 2973 The DNS server provides the IP addresses of other MNs and 2974 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 2976 3.20. Transition Considerations 2978 OAL encapsulation ensures that dissimilar INET partitions can be 2979 joined into a single unified OMNI link, even though the partitions 2980 themselves may have differing protocol versions and/or incompatible 2981 addressing plans. However, a commonality can be achieved by 2982 incrementally distributing globally routable (i.e., native) IP 2983 prefixes to eventually reach all nodes (both mobile and fixed) in all 2984 OMNI link segments. This can be accomplished by incrementally 2985 deploying AERO Relays on each INET partition, with each Relay 2986 distributing its MNPs and/or discovering non-MNP IP GUA prefixes on 2987 its INET links. 2989 This gives rise to the opportunity to eventually distribute native IP 2990 addresses to all nodes, and to present a unified OMNI link view even 2991 if the INET partitions remain in their current protocol and 2992 addressing plans. In that way, the OMNI link can serve the dual 2993 purpose of providing a mobility/multilink service and a transition 2994 service. Or, if an INET partition is transitioned to a native IP 2995 protocol version and addressing scheme that is compatible with the 2996 OMNI link MNP-based addressing scheme, the partition and OMNI link 2997 can be joined by Relays. 2999 Relays that connect INETs/EUNs with dissimilar IP protocol versions 3000 may need to employ a network address and protocol translation 3001 function such as NAT64 [RFC6146]. 3003 3.21. Detecting and Reacting to Server and Bridge Failures 3005 In environments where rapid failure recovery is required, Servers and 3006 Bridges SHOULD use Bidirectional Forwarding Detection (BFD) 3007 [RFC5880]. Nodes that use BFD can quickly detect and react to 3008 failures so that cached information is re-established through 3009 alternate nodes. BFD control messaging is carried only over well- 3010 connected ground domain networks (i.e., and not low-end radio links) 3011 and can therefore be tuned for rapid response. 3013 Servers and Bridges maintain BFD sessions in parallel with their BGP 3014 peerings. If a Server or Bridge fails, BGP peers will quickly re- 3015 establish routes through alternate paths the same as for common BGP 3016 deployments. Similarly, Proxys maintain BFD sessions with their 3017 associated Bridges even though they do not establish BGP peerings 3018 with them. 3020 Proxys SHOULD use proactive NUD for Servers for which there are 3021 currently active ANET Clients in a manner that parallels BFD, i.e., 3022 by sending unicast NS messages in rapid succession to receive 3023 solicited NA messages. When the Proxy is also sending RS messages on 3024 behalf of ANET Clients, the RS/RA messaging can be considered as 3025 equivalent hints of forward progress. This means that the Proxy need 3026 not also send a periodic NS if it has already sent an RS within the 3027 same period. If a Server fails, the Proxy will cease to receive 3028 advertisements and can quickly inform Clients of the outage by 3029 sending multicast RA messages on the ANET interface. 3031 The Proxy sends multicast RA messages with source address set to the 3032 Server's address, destination address set to (link-local) All-Nodes 3033 multicast, and Router Lifetime set to 0. The Proxy SHOULD send 3034 MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small delays 3035 [RFC4861]. Any Clients on the ANET interface that have been using 3036 the (now defunct) Server will receive the RA messages and associate 3037 with a new Server. 3039 3.22. AERO Clients on the Open Internet 3041 AERO Clients that connect to the open Internet via INET interfaces 3042 can establish a VPN or direct link to securely connect to a Server in 3043 a "tethered" arrangement with all of the Client's traffic transiting 3044 the Server. Alternatively, the Client can associate with an INET 3045 Server using UDP/IP encapsulation and control message securing 3046 services as discussed in the following sections. 3048 When a Client's OMNI interface enables an INET underlying interface, 3049 it first determines whether the interface is likely to be behind a 3050 NAT. For IPv4, the Client assumes it is on the open Internet if the 3051 INET address is not a special-use IPv4 address per [RFC3330]. 3052 Similarly for IPv6, the Client assumes it is on the open Internet if 3053 the INET address is a Global Unicast Address (GUA) [RFC4291]. 3054 Otherwise, the Client assumes it may be behind one or several NATs. 3056 The Client then prepares an RS message with IPv6 source address set 3057 to its MNP-LLA, with IPv6 destination set to (link-local) All-Routers 3058 multicast and with an OMNI option with underlying interface 3059 attributes. If the Client believes that it is on the open Internet, 3060 it SHOULD include an L2ADDR in the Interface Attributes sub-option 3061 corresponding to the underlying interface; otherwise, it MAY omit 3062 L2ADDR. If the underlying address is IPv4, the Client includes the 3063 Port Number and IPv4 address written in obfuscated form [RFC4380] as 3064 discussed in Section 3.3. If the underlying interface address is 3065 IPv6, the Client instead includes the Port Number and IPv6 address in 3066 obfuscated form. The Client finally includes a HIP "Initiator" 3067 message sub-option in the OMNI option 3068 [I-D.templin-6man-omni-interface] to provide message authentication 3069 and performs OAL encapsulation and fragmentation if necessary. The 3070 Client then encapsulates each fragment in UDP/IP headers, sets the 3071 UDP/IP source to its INET address and UDP port, sets the UDP/IP 3072 destination to the Server's INET address and the AERO service port 3073 number (8060), then sends the (whole or fragmented) message to the 3074 Server. 3076 When the Server receives the RS, it performs OAL reassembly if 3077 necessary then authenticates the message and registers the Client's 3078 MNP and INET interface information according to the OMNI option 3079 parameters. If the RS message OMNI option includes Interface 3080 Attributes with an L2ADDR, the Server compares the encapsulation IP 3081 address and UDP port number with the (unobfuscated) values. If the 3082 values are the same, the Server caches the Client's information as 3083 "INET" addresses meaning that the Client is likely to accept direct 3084 messages without requiring NAT traversal exchanges. If the values 3085 are different (or, if the OMNI option did not include an L2ADDR) the 3086 Server instead caches the Client's information as "mapped" addresses 3087 meaning that NAT traversal exchanges may be necessary. 3089 The Server then prepares an RA message with IPv6 source and 3090 destination set corresponding to the addresses in the RS, and with an 3091 OMNI option with an Origin Indication sub-option per 3092 [I-D.templin-6man-omni-interface] with the mapped and obfuscated Port 3093 Number and IP address observed in the encapsulation headers. The 3094 Server also includes a HIP "Responder" message sub-option per 3095 [I-D.templin-6man-omni-interface] that contains an acknowledgement of 3096 the update sent by the Client. The Server then performs OAL 3097 encapsulation and fragmentation if necessary, and encapsulates each 3098 fragment in UDP/IP headers with addresses set per the L2ADDR 3099 information in the neighbor cache entry for the Client. 3101 When the Client receives the RA message, it performs OAL reassembly 3102 if necessary, authenticates the HIP "Responder" message, then 3103 compares the mapped Port Number and IP address from the Origin 3104 Indication sub-option with its own address. If the addresses are the 3105 same, the Client assumes the open Internet / Cone NAT principle; if 3106 the addresses are different, the Client instead assumes that further 3107 qualification procedures are necessary to detect the type of NAT and 3108 proceeds according to standard procedures [RFC6081][RFC4380]. 3110 After the Client has registered its INET interfaces in such RS/RA 3111 exchanges it sends periodic RS messages to receive fresh RA messages 3112 before the Router Lifetime received on each INET interface expires. 3113 The Client also maintains default routes via its Servers, i.e., the 3114 same as described in earlier sections. 3116 When the Client sends messages to target IP addresses, it also 3117 invokes route optimization per Section 3.14 using IPv6 ND address 3118 resolution messaging. The Client sends the NS(AR) message to the 3119 Server with an OMNI option with a HIP "Update/Sequence" message sub- 3120 option. The Client sets the NS source address to the Client's MNP- 3121 LLA and destination address to the target solicited node multicast 3122 address. The Client wraps the NS message in an OAL header with 3123 source address set to its own MNP-ULA and destination address set to 3124 the Server's ADM-ULA. The Client then performs OAL fragmentation if 3125 necessary, wraps each fragment in a UDP/IP header and sends it to the 3126 Server. 3128 When the Server receives the OAL-encapsulated NS, it performs OAL 3129 reassembly if necessary. The Server then authenticates the message 3130 by processing the HIP message sub-option and sends a corresponding 3131 NS(AR) message over the spanning tree the same as if it were the ROS, 3132 but with the OAL source address set to the Server's ADM-ULA, with 3133 destination set to the MNP-ULA of the target, and with an OMNI option 3134 that includes no sub-options. When the ROR receives the NS(AR), it 3135 adds the Server's ADM-ULA and Client's MNP-LLA to the target's Report 3136 List, and returns an NA(AR) with OMNI option information for the 3137 target including all of the target's Interface Attributes. The ROR 3138 sets the NA(AR) source address to the target's MNP-LLA and sets the 3139 destination address to the Client's MNP-LLA, then sets the OAL source 3140 address to the ADM-ULA of the ROR and the destination to the ADM-ULA 3141 of the Server. When the Server receives the NA(AR) message, it 3142 rewrites the OAL source address to its own ADM-ULA and the 3143 destination address to the MNP-ULA of the Client, then includes a HIP 3144 "Update/Acknowledge" message sub-option in the OMNI option, performs 3145 OAL fragmentation if necessary, wraps each fragment in UDP/IP 3146 headers, and sends it to the Client. 3148 Following route optimization for targets in the same OMNI link 3149 segment, if the target's L2ADDR is on the open INET, the Client 3150 forwards data packets directly to the target INET address. If the 3151 target is behind a NAT, the Client first establishes NAT state for 3152 the L2ADDR using the "direct bubble" and NUD mechanisms discussed in 3153 Section 3.10.1. The Client continues to send data packets via its 3154 Server until NAT state is populated, then begins forwarding packets 3155 via the direct path through the NAT to the target. For targets in 3156 different OMNI link segments, the Client uses OAL/ORH encapsulation 3157 and forwards data packets to the Bridge that returned the NA message. 3159 The ROR may return uNAs via the Server if the target moves, and the 3160 Server will send corresponding uNAs to the Client with a HIP "Notify" 3161 authentication message. The Client can also send NUD messages to 3162 test forward path reachability even though there is no security 3163 association between the Client and the target. 3165 The Client sends IP packets to route-optimized neighbors in the same 3166 OMNI link segment no larger than the absolute/path MCS in one piece 3167 and without OAL encapsulation. For larger IP packets, the Client 3168 applies OAL encapsulation and fragmentation if necessary according to 3169 Section 3.9, with OAL header with source set to its own MNP-ULA and 3170 destination set to the MNP-ULA of the target. The Client then 3171 encapsulates each raw or OAL-encapsulated IP packet/fragment in a 3172 UDP/IP header and sends them to the next hop. 3174 Note: The NAT traversal procedures specified in this document are 3175 applicable for Cone, Address-Restricted and Port-Restricted NATs 3176 only. While future updates to this document may specify procedures 3177 for other NAT variations (e.g., hairpinning and various forms of 3178 Symmetric NATs), it should be noted that continuous communications 3179 are always possible through forwarding via a Server even if NAT 3180 traversal is not employed. 3182 Note: Following the initial HIP Initiator/Responder exchange, AERO 3183 Clients with OMNI interfaces configured over the open Internet 3184 maintain HIP associations through the transmission of IPv6 ND 3185 messages that include OMNI options with HIP "Update" and "Notify" 3186 messages. OMNI interfaces use the HIP "Update" message when an 3187 acknowledgement is required, and use the "Notify" message in 3188 unacknowledged isolated IPv6 ND messages (e.g., unsolicited NAs). 3190 Note: Servers on the open Internet that act as Proxys authenticate 3191 and remove OMNI option HIP message sub-options from RSes they forward 3192 from the MN to another Server, and insert and sign HIP message and 3193 Origin Indication sub-options in RAs they forward from another Server 3194 to the MN. Conversely, Servers that act as Proxys forward without 3195 processing any MNP registration/delegation information in RS/RA 3196 message exchanges between MNs and other Servers. The Server acting 3197 as a Proxy is therefore responsible for MN authentication, while the 3198 other Servers are responsible for registering/delegating MNPs (noting 3199 that the same node can act as both Proxy and Server). 3201 3.23. Time-Varying MNPs 3203 In some use cases, it is desirable, beneficial and efficient for the 3204 Client to receive a constant MNP that travels with the Client 3205 wherever it moves. For example, this would allow air traffic 3206 controllers to easily track aircraft, etc. In other cases, however 3207 (e.g., intelligent transportation systems), the MN may be willing to 3208 sacrifice a modicum of efficiency in order to have time-varying MNPs 3209 that can be changed every so often to defeat adversarial tracking. 3211 The DHCPv6 service offers a way for Clients that desire time-varying 3212 MNPs to obtain short-lived prefixes (e.g., on the order of a small 3213 number of minutes). In that case, the identity of the Client would 3214 not be bound to the MNP but rather to a Node Identification value 3215 (see: [I-D.templin-6man-omni-interface]) to be used as the Client ID 3216 seed for MNP prefix delegation. The Client would then be obligated 3217 to renumber its internal networks whenever its MNP (and therefore 3218 also its MNP-LLA) changes. This should not present a challenge for 3219 Clients with automated network renumbering services, however presents 3220 limits for the durations of ongoing sessions that would prefer to use 3221 a constant address. 3223 4. Implementation Status 3225 An early AERO implementation based on OpenVPN (https://openvpn.net/) 3226 was announced on the v6ops mailing list on January 10, 2018 and an 3227 initial public release of the AERO proof-of-concept source code was 3228 announced on the intarea mailing list on August 21, 2015. 3230 AERO Release-3.0.2 was tagged on October 15, 2020, and is undergoing 3231 internal testing. Additional internal releases expected within the 3232 coming months, with first public release expected end of 1H2021. 3234 5. IANA Considerations 3236 The IANA is instructed to assign a new type value TBD1 in the IPv6 3237 Routing Types registry. 3239 The IANA has assigned the UDP port number "8060" for an earlier 3240 experimental first version of AERO [RFC6706]. This document 3241 obsoletes [RFC6706], and together with 3242 [I-D.templin-6man-omni-interface] reclaims the UDP port number "8060" 3243 for 'aero' as the service port for UDP/IP encapsulation. (Note that, 3244 although [RFC6706] was not widely implemented or deployed, any 3245 messages coded to that specification can be easily distinguished and 3246 ignored since they use the invalid ICMPv6 message type number '0'.) 3247 This document makes no request of IANA, since 3248 [I-D.templin-6man-omni-interface] already provides instructions. 3250 No further IANA actions are required. 3252 6. Security Considerations 3254 AERO Bridges configure secured tunnels with AERO Servers, Relays and 3255 Proxys within their local OMNI link segments. Applicable secured 3256 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 3257 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 3258 segments in turn configure secured tunnels for their neighboring AERO 3259 Bridges in a spanning tree topology. Therefore, control messages 3260 exchanged between any pair of OMNI link neighbors on the spanning 3261 tree are already secured. 3263 AERO Servers, Relays and Proxys targeted by a route optimization may 3264 also receive data packets directly from arbitrary nodes in INET 3265 partitions instead of via the spanning tree. For INET partitions 3266 that apply effective ingress filtering to defeat source address 3267 spoofing, the simple data origin authentication procedures in 3268 Section 3.8 can be applied. 3270 For INET partitions that require strong security in the data plane, 3271 two options for securing communications include 1) disable route 3272 optimization so that all traffic is conveyed over secured tunnels, or 3273 2) enable on-demand secure tunnel creation between INET partition 3274 neighbors. Option 1) would result in longer routes than necessary 3275 and traffic concentration on critical infrastructure elements. 3276 Option 2) could be coordinated by establishing a secured tunnel on- 3277 demand instead of performing an NS/NA exchange in the route 3278 optimization procedures. Procedures for establishing on-demand 3279 secured tunnels are out of scope. 3281 AERO Clients that connect to secured ANETs need not apply security to 3282 their ND messages, since the messages will be intercepted by a 3283 perimeter Proxy that applies security on its INET-facing interface as 3284 part of the spanning tree (see above). AERO Clients connected to the 3285 open INET can use network and/or transport layer security services 3286 such as VPNs or can by some other means establish a direct link. 3287 When a VPN or direct link may be impractical, however, the 3288 authentication services specified in [RFC7401] and/or [RFC4380] 3289 should be applied. 3291 Application endpoints SHOULD use application-layer security services 3292 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 3293 protection as for critical secured Internet services. AERO Clients 3294 that require host-based VPN services SHOULD use network and/or 3295 transport layer security services such as IPsec, TLS/SSL, DTLS, etc. 3296 AERO Proxys and Servers can also provide a network-based VPN service 3297 on behalf of the Client, e.g., if the Client is located within a 3298 secured enclave and cannot establish a VPN on its own behalf. 3300 AERO Servers and Bridges present targets for traffic amplification 3301 Denial of Service (DoS) attacks. This concern is no different than 3302 for widely-deployed VPN security gateways in the Internet, where 3303 attackers could send spoofed packets to the gateways at high data 3304 rates. This can be mitigated by connecting Servers and Bridges over 3305 dedicated links with no connections to the Internet and/or when 3306 connections to the Internet are only permitted through well-managed 3307 firewalls. Traffic amplification DoS attacks can also target an AERO 3308 Client's low data rate links. This is a concern not only for Clients 3309 located on the open Internet but also for Clients in secured 3310 enclaves. AERO Servers and Proxys can institute rate limits that 3311 protect Clients from receiving packet floods that could DoS low data 3312 rate links. 3314 AERO Relays must implement ingress filtering to avoid a spoofing 3315 attack in which spurious messages with ULA addresses are injected 3316 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3317 that their connectivity is not used by unauthorized nodes on their 3318 EUNs to gain access to a protected network, i.e., AERO Clients that 3319 act as routers MUST NOT provide routing services for unauthorized 3320 nodes. (This concern is no different than for ordinary hosts that 3321 receive an IP address delegation but then "share" the address with 3322 other nodes via some form of Internet connection sharing such as 3323 tethering.) 3325 The MAP list MUST be well-managed and secured from unauthorized 3326 tampering, even though the list contains only public information. 3327 The MAP list can be conveyed to the Client in a similar fashion as in 3328 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3329 upload of a static file, DNS lookups, etc.). 3331 SRH authentication facilities are specified in [RFC8754]. 3333 Security considerations for accepting link-layer ICMP messages and 3334 reflected packets are discussed throughout the document. 3336 Security considerations for IPv6 fragmentation and reassembly are 3337 discussed in [I-D.templin-6man-omni-interface]. 3339 7. Acknowledgements 3341 Discussions in the IETF, aviation standards communities and private 3342 exchanges helped shape some of the concepts in this work. 3343 Individuals who contributed insights include Mikael Abrahamsson, Mark 3344 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3345 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3346 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3347 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3348 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3349 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3350 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3351 Wood and James Woodyatt. Members of the IESG also provided valuable 3352 input during their review process that greatly improved the document. 3353 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3354 for their shepherding guidance during the publication of the AERO 3355 first edition. 3357 This work has further been encouraged and supported by Boeing 3358 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3359 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3360 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3361 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3362 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3363 Gene MacLean III, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3364 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3365 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3366 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3367 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3368 Benson, Katie Tran and Eric Yeh are especially acknowledged for 3369 implementing the AERO functions as extensions to the public domain 3370 OpenVPN distribution. 3372 Earlier works on NBMA tunneling approaches are found in 3373 [RFC2529][RFC5214][RFC5569]. 3375 Many of the constructs presented in this second edition of AERO are 3376 based on the author's earlier works, including: 3378 o The Internet Routing Overlay Network (IRON) 3379 [RFC6179][I-D.templin-ironbis] 3381 o Virtual Enterprise Traversal (VET) 3382 [RFC5558][I-D.templin-intarea-vet] 3384 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3385 [RFC5320][I-D.templin-intarea-seal] 3387 o AERO, First Edition [RFC6706] 3389 Note that these works cite numerous earlier efforts that are not also 3390 cited here due to space limitations. The authors of those earlier 3391 works are acknowledged for their insights. 3393 This work is aligned with the NASA Safe Autonomous Systems Operation 3394 (SASO) program under NASA contract number NNA16BD84C. 3396 This work is aligned with the FAA as per the SE2025 contract number 3397 DTFAWA-15-D-00030. 3399 This work is aligned with the Boeing Commercial Airplanes (BCA) 3400 Internet of Things (IoT) and autonomy programs. 3402 This work is aligned with the Boeing Information Technology (BIT) 3403 MobileNet program. 3405 8. References 3407 8.1. Normative References 3409 [I-D.templin-6man-omni-interface] 3410 Templin, F. and T. Whyman, "Transmission of IP Packets 3411 over Overlay Multilink Network (OMNI) Interfaces", draft- 3412 templin-6man-omni-interface-69 (work in progress), January 3413 2021. 3415 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3416 DOI 10.17487/RFC0791, September 1981, 3417 . 3419 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3420 RFC 792, DOI 10.17487/RFC0792, September 1981, 3421 . 3423 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3424 Requirement Levels", BCP 14, RFC 2119, 3425 DOI 10.17487/RFC2119, March 1997, 3426 . 3428 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3429 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3430 December 1998, . 3432 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3433 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3434 DOI 10.17487/RFC3971, March 2005, 3435 . 3437 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3438 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3439 . 3441 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3442 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3443 November 2005, . 3445 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3446 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3447 . 3449 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3450 Network Address Translations (NATs)", RFC 4380, 3451 DOI 10.17487/RFC4380, February 2006, 3452 . 3454 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3455 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3456 DOI 10.17487/RFC4861, September 2007, 3457 . 3459 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3460 Address Autoconfiguration", RFC 4862, 3461 DOI 10.17487/RFC4862, September 2007, 3462 . 3464 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3465 DOI 10.17487/RFC6081, January 2011, 3466 . 3468 [RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T. 3469 Henderson, "Host Identity Protocol Version 2 (HIPv2)", 3470 RFC 7401, DOI 10.17487/RFC7401, April 2015, 3471 . 3473 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3474 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3475 May 2017, . 3477 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3478 (IPv6) Specification", STD 86, RFC 8200, 3479 DOI 10.17487/RFC8200, July 2017, 3480 . 3482 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3483 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3484 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3485 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3486 . 3488 8.2. Informative References 3490 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3491 2016. 3493 [I-D.bonica-6man-comp-rtg-hdr] 3494 Bonica, R., Kamite, Y., Alston, A., Henriques, D., and L. 3495 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3496 bonica-6man-comp-rtg-hdr-24 (work in progress), January 3497 2021. 3499 [I-D.bonica-6man-crh-helper-opt] 3500 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3501 Routing Header (CRH) Helper Option", draft-bonica-6man- 3502 crh-helper-opt-02 (work in progress), October 2020. 3504 [I-D.ietf-intarea-frag-fragile] 3505 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 3506 and F. Gont, "IP Fragmentation Considered Fragile", draft- 3507 ietf-intarea-frag-fragile-17 (work in progress), September 3508 2019. 3510 [I-D.ietf-intarea-tunnels] 3511 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3512 Architecture", draft-ietf-intarea-tunnels-10 (work in 3513 progress), September 2019. 3515 [I-D.ietf-ipwave-vehicular-networking] 3516 Jeong, J., "IPv6 Wireless Access in Vehicular Environments 3517 (IPWAVE): Problem Statement and Use Cases", draft-ietf- 3518 ipwave-vehicular-networking-19 (work in progress), July 3519 2020. 3521 [I-D.ietf-rtgwg-atn-bgp] 3522 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3523 Moreno, "A Simple BGP-based Mobile Routing System for the 3524 Aeronautical Telecommunications Network", draft-ietf- 3525 rtgwg-atn-bgp-10 (work in progress), January 2021. 3527 [I-D.templin-6man-dhcpv6-ndopt] 3528 Templin, F., "A Unified Stateful/Stateless Configuration 3529 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-11 3530 (work in progress), January 2021. 3532 [I-D.templin-intarea-seal] 3533 Templin, F., "The Subnetwork Encapsulation and Adaptation 3534 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3535 progress), January 2014. 3537 [I-D.templin-intarea-vet] 3538 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3539 templin-intarea-vet-40 (work in progress), May 2013. 3541 [I-D.templin-ipwave-uam-its] 3542 Templin, F., "Urban Air Mobility Implications for 3543 Intelligent Transportation Systems", draft-templin-ipwave- 3544 uam-its-04 (work in progress), January 2021. 3546 [I-D.templin-ironbis] 3547 Templin, F., "The Interior Routing Overlay Network 3548 (IRON)", draft-templin-ironbis-16 (work in progress), 3549 March 2014. 3551 [I-D.templin-v6ops-pdhost] 3552 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3553 Models", draft-templin-v6ops-pdhost-27 (work in progress), 3554 January 2021. 3556 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3558 [RFC1035] Mockapetris, P., "Domain names - implementation and 3559 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3560 November 1987, . 3562 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3563 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3564 . 3566 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3567 DOI 10.17487/RFC2003, October 1996, 3568 . 3570 [RFC2004] Perkins, C., "Minimal Encapsulation within IP", RFC 2004, 3571 DOI 10.17487/RFC2004, October 1996, 3572 . 3574 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3575 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3576 . 3578 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 3579 Networks", RFC 2464, DOI 10.17487/RFC2464, December 1998, 3580 . 3582 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3583 Domains without Explicit Tunnels", RFC 2529, 3584 DOI 10.17487/RFC2529, March 1999, 3585 . 3587 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3588 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3589 . 3591 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3592 of Explicit Congestion Notification (ECN) to IP", 3593 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3594 . 3596 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3597 DOI 10.17487/RFC3330, September 2002, 3598 . 3600 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3601 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3602 DOI 10.17487/RFC3810, June 2004, 3603 . 3605 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 3606 Unique IDentifier (UUID) URN Namespace", RFC 4122, 3607 DOI 10.17487/RFC4122, July 2005, 3608 . 3610 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3611 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3612 January 2006, . 3614 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3615 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3616 DOI 10.17487/RFC4271, January 2006, 3617 . 3619 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3620 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3621 2006, . 3623 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3624 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3625 December 2005, . 3627 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3628 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3629 2006, . 3631 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3632 Control Message Protocol (ICMPv6) for the Internet 3633 Protocol Version 6 (IPv6) Specification", STD 89, 3634 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3635 . 3637 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3638 Protocol (LDAP): The Protocol", RFC 4511, 3639 DOI 10.17487/RFC4511, June 2006, 3640 . 3642 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3643 "Considerations for Internet Group Management Protocol 3644 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3645 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3646 . 3648 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3649 "Internet Group Management Protocol (IGMP) / Multicast 3650 Listener Discovery (MLD)-Based Multicast Forwarding 3651 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3652 August 2006, . 3654 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3655 Algorithms in Cryptographically Generated Addresses 3656 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3657 . 3659 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3660 "Bidirectional Protocol Independent Multicast (BIDIR- 3661 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3662 . 3664 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3665 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3666 DOI 10.17487/RFC5214, March 2008, 3667 . 3669 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3670 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3671 February 2010, . 3673 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3674 Route Optimization Requirements for Operational Use in 3675 Aeronautics and Space Exploration Mobile Networks", 3676 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3677 . 3679 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3680 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3681 . 3683 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3684 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3685 January 2010, . 3687 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3688 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3689 . 3691 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3692 "IPv6 Router Advertisement Options for DNS Configuration", 3693 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3694 . 3696 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3697 NAT64: Network Address and Protocol Translation from IPv6 3698 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3699 April 2011, . 3701 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3702 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3703 . 3705 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3706 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3707 DOI 10.17487/RFC6221, May 2011, 3708 . 3710 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3711 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3712 DOI 10.17487/RFC6273, June 2011, 3713 . 3715 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3716 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3717 January 2012, . 3719 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 3720 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, 3721 DOI 10.17487/RFC6355, August 2011, 3722 . 3724 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3725 for Equal Cost Multipath Routing and Link Aggregation in 3726 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3727 . 3729 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3730 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3731 . 3733 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3734 UDP Checksums for Tunneled Packets", RFC 6935, 3735 DOI 10.17487/RFC6935, April 2013, 3736 . 3738 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3739 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3740 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3741 . 3743 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3744 Korhonen, "Requirements for Distributed Mobility 3745 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3746 . 3748 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3749 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3750 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3751 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3752 2016, . 3754 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3755 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3756 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3757 July 2018, . 3759 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3760 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3761 . 3763 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 3764 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3765 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 3766 . 3768 [WG] Wireguard, "Wireguard, https://www.wireguard.com", August 3769 2020. 3771 Appendix A. Non-Normative Considerations 3773 AERO can be applied to a multitude of Internetworking scenarios, with 3774 each having its own adaptations. The following considerations are 3775 provided as non-normative guidance: 3777 A.1. Implementation Strategies for Route Optimization 3779 Route optimization as discussed in Section 3.14 results in the route 3780 optimization source (ROS) creating a neighbor cache entry for the 3781 target neighbor. The neighbor cache entry is maintained for at most 3782 ReachableTime seconds and then deleted unless updated. In order to 3783 refresh the neighbor cache entry lifetime before the ReachableTime 3784 timer expires, the specification requires implementations to issue a 3785 new NS/NA exchange to reset ReachableTime while data packets are 3786 still flowing. However, the decision of when to initiate a new NS/NA 3787 exchange and to perpetuate the process is left as an implementation 3788 detail. 3790 One possible strategy may be to monitor the neighbor cache entry 3791 watching for data packets for (ReachableTime - 5) seconds. If any 3792 data packets have been sent to the neighbor within this timeframe, 3793 then send an NS to receive a new NA. If no data packets have been 3794 sent, wait for 5 additional seconds and send an immediate NS if any 3795 data packets are sent within this "expiration pending" 5 second 3796 window. If no additional data packets are sent within the 5 second 3797 window, delete the neighbor cache entry. 3799 The monitoring of the neighbor data packet traffic therefore becomes 3800 an asymmetric ongoing process during the neighbor cache entry 3801 lifetime. If the neighbor cache entry expires, future data packets 3802 will trigger a new NS/NA exchange while the packets themselves are 3803 delivered over a longer path until route optimization state is re- 3804 established. 3806 A.2. Implicit Mobility Management 3808 OMNI interface neighbors MAY provide a configuration option that 3809 allows them to perform implicit mobility management in which no ND 3810 messaging is used. In that case, the Client only transmits packets 3811 over a single interface at a time, and the neighbor always observes 3812 packets arriving from the Client from the same link-layer source 3813 address. 3815 If the Client's underlying interface address changes (either due to a 3816 readdressing of the original interface or switching to a new 3817 interface) the neighbor immediately updates the neighbor cache entry 3818 for the Client and begins accepting and sending packets according to 3819 the Client's new address. This implicit mobility method applies to 3820 use cases such as cellphones with both WiFi and Cellular interfaces 3821 where only one of the interfaces is active at a given time, and the 3822 Client automatically switches over to the backup interface if the 3823 primary interface fails. 3825 A.3. Direct Underlying Interfaces 3827 When a Client's OMNI interface is configured over a Direct interface, 3828 the neighbor at the other end of the Direct link can receive packets 3829 without any encapsulation. In that case, the Client sends packets 3830 over the Direct link according to QoS preferences. If the Direct 3831 interface has the highest QoS preference, then the Client's IP 3832 packets are transmitted directly to the peer without going through an 3833 ANET/INET. If other interfaces have higher QoS preferences, then the 3834 Client's IP packets are transmitted via a different interface, which 3835 may result in the inclusion of Proxys, Servers and Bridges in the 3836 communications path. Direct interfaces must be tested periodically 3837 for reachability, e.g., via NUD. 3839 A.4. AERO Critical Infrastructure Considerations 3841 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 3842 IP routers or virtual machines in the cloud. Bridges must be 3843 provisioned, supported and managed by the INET administrative 3844 authority, and connected to the Bridges of other INETs via inter- 3845 domain peerings. Cost for purchasing, configuring and managing 3846 Bridges is nominal even for very large OMNI links. 3848 AERO Servers can be standard dedicated server platforms, but most 3849 often will be deployed as virtual machines in the cloud. The only 3850 requirements for Servers are that they can run the AERO user-level 3851 code and have at least one network interface connection to the INET. 3852 As with Bridges, Servers must be provisioned, supported and managed 3853 by the INET administrative authority. Cost for purchasing, 3854 configuring and managing Servers is nominal especially for virtual 3855 Servers hosted in the cloud. 3857 AERO Proxys are most often standard dedicated server platforms with 3858 one network interface connected to the ANET and a second interface 3859 connected to an INET. As with Servers, the only requirements are 3860 that they can run the AERO user-level code and have at least one 3861 interface connection to the INET. Proxys must be provisioned, 3862 supported and managed by the ANET administrative authority. Cost for 3863 purchasing, configuring and managing Proxys is nominal, and borne by 3864 the ANET administrative authority. 3866 AERO Relays can be any dedicated server or COTS router platform 3867 connected to INETs and/or EUNs. The Relay connects to the OMNI link 3868 and engages in eBGP peering with one or more Bridges as a stub AS. 3869 The Relay then injects its MNPs and/or non-MNP prefixes into the BGP 3870 routing system, and provisions the prefixes to its downstream- 3871 attached networks. The Relay can perform ROS/ROR services the same 3872 as for any Server, and can route between the MNP and non-MNP address 3873 spaces. 3875 A.5. AERO Server Failure Implications 3877 AERO Servers may appear as a single point of failure in the 3878 architecture, but such is not the case since all Servers on the link 3879 provide identical services and loss of a Server does not imply 3880 immediate and/or comprehensive communication failures. Although 3881 Clients typically associate with a single Server at a time, Server 3882 failure is quickly detected and conveyed by Bidirectional Forward 3883 Detection (BFD) and/or proactive NUD allowing Clients to migrate to 3884 new Servers. 3886 If a Server fails, ongoing packet forwarding to Clients will continue 3887 by virtue of the neighbor cache entries that have already been 3888 established in route optimization sources (ROSs). If a Client also 3889 experiences mobility events at roughly the same time the Server 3890 fails, unsolicited NA messages may be lost but proxy neighbor cache 3891 entries in the DEPARTED state will ensure that packet forwarding to 3892 the Client's new locations will continue for up to DepartTime 3893 seconds. 3895 If a Client is left without a Server for an extended timeframe (e.g., 3896 greater than ReachableTime seconds) then existing neighbor cache 3897 entries will eventually expire and both ongoing and new 3898 communications will fail. The original source will continue to 3899 retransmit until the Client has established a new Server 3900 relationship, after which time continuous communications will resume. 3902 Therefore, providing many Servers on the link with high availability 3903 profiles provides resilience against loss of individual Servers and 3904 assurance that Clients can establish new Server relationships quickly 3905 in event of a Server failure. 3907 A.6. AERO Client / Server Architecture 3909 The AERO architectural model is client / server in the control plane, 3910 with route optimization in the data plane. The same as for common 3911 Internet services, the AERO Client discovers the addresses of AERO 3912 Servers and selects one Server to connect to. The AERO service is 3913 analogous to common Internet services such as google.com, yahoo.com, 3914 cnn.com, etc. However, there is only one AERO service for the link 3915 and all Servers provide identical services. 3917 Common Internet services provide differing strategies for advertising 3918 server addresses to clients. The strategy is conveyed through the 3919 DNS resource records returned in response to name resolution queries. 3920 As of January 2020 Internet-based 'nslookup' services were used to 3921 determine the following: 3923 o When a client resolves the domainname "google.com", the DNS always 3924 returns one A record (i.e., an IPv4 address) and one AAAA record 3925 (i.e., an IPv6 address). The client receives the same addresses 3926 each time it resolves the domainname via the same DNS resolver, 3927 but may receive different addresses when it resolves the 3928 domainname via different DNS resolvers. But, in each case, 3929 exactly one A and one AAAA record are returned. 3931 o When a client resolves the domainname "ietf.org", the DNS always 3932 returns one A record and one AAAA record with the same addresses 3933 regardless of which DNS resolver is used. 3935 o When a client resolves the domainname "yahoo.com", the DNS always 3936 returns a list of 4 A records and 4 AAAA records. Each time the 3937 client resolves the domainname via the same DNS resolver, the same 3938 list of addresses are returned but in randomized order (i.e., 3939 consistent with a DNS round-robin strategy). But, interestingly, 3940 the same addresses are returned (albeit in randomized order) when 3941 the domainname is resolved via different DNS resolvers. 3943 o When a client resolves the domainname "amazon.com", the DNS always 3944 returns a list of 3 A records and no AAAA records. As with 3945 "yahoo.com", the same three A records are returned from any 3946 worldwide Internet connection point in randomized order. 3948 The above example strategies show differing approaches to Internet 3949 resilience and service distribution offered by major Internet 3950 services. The Google approach exposes only a single IPv4 and a 3951 single IPv6 address to clients. Clients can then select whichever IP 3952 protocol version offers the best response, but will always use the 3953 same IP address according to the current Internet connection point. 3954 This means that the IP address offered by the network must lead to a 3955 highly-available server and/or service distribution point. In other 3956 words, resilience is predicated on high availability within the 3957 network and with no client-initiated failovers expected (i.e., it is 3958 all-or-nothing from the client's perspective). However, Google does 3959 provide for worldwide distributed service distribution by virtue of 3960 the fact that each Internet connection point responds with a 3961 different IPv6 and IPv4 address. The IETF approach is like google 3962 (all-or-nothing from the client's perspective), but provides only a 3963 single IPv4 or IPv6 address on a worldwide basis. This means that 3964 the addresses must be made highly-available at the network level with 3965 no client failover possibility, and if there is any worldwide service 3966 distribution it would need to be conducted by a network element that 3967 is reached via the IP address acting as a service distribution point. 3969 In contrast to the Google and IETF philosophies, Yahoo and Amazon 3970 both provide clients with a (short) list of IP addresses with Yahoo 3971 providing both IP protocol versions and Amazon as IPv4-only. The 3972 order of the list is randomized with each name service query 3973 response, with the effect of round-robin load balancing for service 3974 distribution. With a short list of addresses, there is still 3975 expectation that the network will implement high availability for 3976 each address but in case any single address fails the client can 3977 switch over to using a different address. The balance then becomes 3978 one of function in the network vs function in the end system. 3980 The same implications observed for common highly-available services 3981 in the Internet apply also to the AERO client/server architecture. 3982 When an AERO Client connects to one or more ANETs, it discovers one 3983 or more AERO Server addresses through the mechanisms discussed in 3984 earlier sections. Each Server address presumably leads to a fault- 3985 tolerant clustering arrangement such as supported by Linux-HA, 3986 Extended Virtual Synchrony or Paxos. Such an arrangement has 3987 precedence in common Internet service deployments in lightweight 3988 virtual machines without requiring expensive hardware deployment. 3989 Similarly, common Internet service deployments set service IP 3990 addresses on service distribution points that may relay requests to 3991 many different servers. 3993 For AERO, the expectation is that a combination of the Google/IETF 3994 and Yahoo/Amazon philosophies would be employed. The AERO Client 3995 connects to different ANET access points and can receive 1-2 Server 3996 ADM-LLAs at each point. It then selects one AERO Server address, and 3997 engages in RS/RA exchanges with the same Server from all ANET 3998 connections. The Client remains with this Server unless or until the 3999 Server fails, in which case it can switch over to an alternate 4000 Server. The Client can likewise switch over to a different Server at 4001 any time if there is some reason for it to do so. So, the AERO 4002 expectation is for a balance of function in the network and end 4003 system, with fault tolerance and resilience at both levels. 4005 Appendix B. Change Log 4007 << RFC Editor - remove prior to publication >> 4009 Changes from draft-templin-intarea-6706bis-61 to draft-templin- 4010 intrea-6706bis-62: 4012 o New sub-section on OMNI Neighbor Interface Attributes 4014 Changes from draft-templin-intarea-6706bis-59 to draft-templin- 4015 intrea-6706bis-60: 4017 o Removed all references to S/TLLAO - all Interface Attributes are 4018 now maintained completely in the OMNI option. 4020 Changes from draft-templin-intarea-6706bis-58 to draft-templin- 4021 intrea-6706bis-59: 4023 o The term "Relay" used in older draft versions is now "Bridge". 4024 "Relay" now refers to what was formally called: "Gateway". 4026 o Fine-grained cleanup of Forwarding Algorithm; IPv6 ND message 4027 addressing; OMNI Prefix Lengths, etc. 4029 Changes from draft-templin-intarea-6706bis-54 to draft-templin- 4030 intrea-6706bis-55: 4032 o Updates on Segment Routing and S/TLLAO contents. 4034 o Various editorials and addressing cleanups. 4036 Changes from draft-templin-intarea-6706bis-52 to draft-templin- 4037 intrea-6706bis-53: 4039 o Normative reference to the OMNI spec, and remove portions that are 4040 already specified in OMNI. 4042 o Renamed "AERO interface/link" to "OMIN interface/link" throughout 4043 the document. 4045 o Truncated obsolete back section matter. 4047 Author's Address 4049 Fred L. Templin (editor) 4050 Boeing Research & Technology 4051 P.O. Box 3707 4052 Seattle, WA 98124 4053 USA 4055 Email: fltemplin@acm.org