idnits 2.17.1 draft-templin-6man-aero-07.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (May 27, 2021) is 1063 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'ULA' is mentioned on line 1094, but not defined == Missing Reference: 'ULA0' is mentioned on line 1094, but not defined == Missing Reference: 'ULA1' is mentioned on line 1094, but not defined == Missing Reference: 'ULA2' is mentioned on line 1094, but not defined == Unused Reference: 'RFC3971' is defined on line 3673, but no explicit reference was found in the text == Unused Reference: 'RFC3972' is defined on line 3678, but no explicit reference was found in the text == Unused Reference: 'RFC4193' is defined on line 3686, but no explicit reference was found in the text == Unused Reference: 'RFC7739' is defined on line 3714, but no explicit reference was found in the text == Unused Reference: 'BGP' is defined on line 3735, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-comp-rtg-hdr' is defined on line 3738, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-crh-helper-opt' is defined on line 3744, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3749, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3772, but no explicit reference was found in the text == Unused Reference: 'RFC2003' is defined on line 3811, but no explicit reference was found in the text == Unused Reference: 'RFC2983' is defined on line 3832, but no explicit reference was found in the text == Unused Reference: 'RFC3168' is defined on line 3836, but no explicit reference was found in the text == Unused Reference: 'RFC4122' is defined on line 3850, but no explicit reference was found in the text == Unused Reference: 'RFC4982' is defined on line 3899, but no explicit reference was found in the text == Unused Reference: 'RFC6139' is defined on line 3941, but no explicit reference was found in the text == Unused Reference: 'RFC6273' is defined on line 3961, but no explicit reference was found in the text == Unused Reference: 'RFC6355' is defined on line 3970, but no explicit reference was found in the text == Unused Reference: 'RFC6438' is defined on line 3975, but no explicit reference was found in the text == Unused Reference: 'RFC6935' is defined on line 3984, but no explicit reference was found in the text == Unused Reference: 'RFC6936' is defined on line 3989, but no explicit reference was found in the text == Outdated reference: A later version (-74) exists of draft-templin-6man-omni-03 == 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-03 == 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-20 == 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 (~~), 31 warnings (==), 5 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 Intended status: Informational May 27, 2021 5 Expires: November 28, 2021 7 Asymmetric Extended Route Optimization (AERO) 8 draft-templin-6man-aero-07 10 Abstract 12 This document specifies an Asymmetric Extended Route Optimization 13 (AERO) service for IP internetworking over Overlay Multilink Network 14 (OMNI) interfaces. AERO/OMNI use an IPv6 link-local address format 15 that supports operation of the IPv6 Neighbor Discovery (ND) protocol 16 and links ND to IP forwarding. Prefix delegation/registration 17 services are employed for network admission and to manage the routing 18 system. Secure multilink operation, mobility management, multicast, 19 traffic selector signaling and route optimization are naturally 20 supported through dynamic neighbor cache updates. AERO is a widely- 21 applicable mobile internetworking service especially well-suited to 22 aviation services, intelligent transportation systems, mobile Virtual 23 Private Networks (VPNs) and many other applications. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at https://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on November 28, 2021. 42 Copyright Notice 44 Copyright (c) 2021 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (https://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 60 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6 61 3. Asymmetric Extended Route Optimization (AERO) . . . . . . . . 12 62 3.1. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 12 63 3.2. The AERO Service over OMNI Links . . . . . . . . . . . . 13 64 3.2.1. AERO/OMNI Reference Model . . . . . . . . . . . . . . 13 65 3.2.2. Addressing and Node Identification . . . . . . . . . 16 66 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 17 67 3.2.4. OMNI Link Segment Routing . . . . . . . . . . . . . . 19 68 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 23 69 3.2.6. Segment Routing For OMNI Link Selection . . . . . . . 24 70 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 24 71 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 25 72 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 27 73 3.4.1. AERO Proxy/Server and Relay Behavior . . . . . . . . 27 74 3.4.2. AERO Client Behavior . . . . . . . . . . . . . . . . 28 75 3.4.3. AERO Bridge Behavior . . . . . . . . . . . . . . . . 28 76 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 28 77 3.5.1. OMNI ND Messages . . . . . . . . . . . . . . . . . . 30 78 3.5.2. OMNI Neighbor Advertisement Message Flags . . . . . . 30 79 3.5.3. OMNI Neighbor Window Synchronization . . . . . . . . 31 80 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 31 81 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 32 82 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 32 83 3.9. OMNI Interface MTU . . . . . . . . . . . . . . . . . . . 32 84 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 33 85 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 34 86 3.10.2. Proxy/Server and Relay Forwarding Algorithm . . . . 36 87 3.10.3. Bridge Forwarding Algorithm . . . . . . . . . . . . 38 88 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 40 89 3.12. AERO Router Discovery, Prefix Delegation and 90 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 42 91 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 43 92 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 43 93 3.12.3. AERO Proxy/Server Behavior . . . . . . . . . . . . . 45 94 3.13. The AERO Proxy Function . . . . . . . . . . . . . . . . . 48 95 3.13.1. Detecting and Responding to Proxy/Server Failures . 51 96 3.13.2. Point-to-Multipoint Proxy/Server Coordination . . . 52 98 3.14. AERO Route Optimization . . . . . . . . . . . . . . . . . 52 99 3.14.1. Route Optimization Initiation . . . . . . . . . . . 53 100 3.14.2. Relaying the NS(AR) *NET Packet(s) . . . . . . . . . 54 101 3.14.3. Processing the NS(AR) and Sending the NA(AR) . . . . 54 102 3.14.4. Relaying the NA(AR) . . . . . . . . . . . . . . . . 56 103 3.14.5. Processing the NA(AR) . . . . . . . . . . . . . . . 56 104 3.14.6. Forwarding Packets to Route Optimized Targets . . . 56 105 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 59 106 3.16. Mobility Management and Quality of Service (QoS) . . . . 61 107 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 61 108 3.16.2. Announcing Link-Layer Address and/or QoS Preference 109 Changes . . . . . . . . . . . . . . . . . . . . . . 63 110 3.16.3. Bringing New Links Into Service . . . . . . . . . . 63 111 3.16.4. Deactivating Existing Links . . . . . . . . . . . . 63 112 3.16.5. Moving Between Proxy/Servers . . . . . . . . . . . . 64 113 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 65 114 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 65 115 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 67 116 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 67 117 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 67 118 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 68 119 3.20. Transition/Coexistence Considerations . . . . . . . . . . 69 120 3.21. Detecting and Reacting to Proxy/Server and Bridge 121 Failures . . . . . . . . . . . . . . . . . . . . . . . . 69 122 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 70 123 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 73 124 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 73 125 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 73 126 6. Security Considerations . . . . . . . . . . . . . . . . . . . 74 127 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 76 128 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 77 129 8.1. Normative References . . . . . . . . . . . . . . . . . . 77 130 8.2. Informative References . . . . . . . . . . . . . . . . . 79 131 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 85 132 A.1. Implementation Strategies for Route Optimization . . . . 85 133 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 86 134 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 86 135 A.4. AERO Critical Infrastructure Considerations . . . . . . . 87 136 A.5. AERO Server Failure Implications . . . . . . . . . . . . 87 137 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 88 138 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 90 139 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 91 141 1. Introduction 143 Asymmetric Extended Route Optimization (AERO) fulfills the 144 requirements of Distributed Mobility Management (DMM) [RFC7333] and 145 route optimization [RFC5522] for aeronautical networking and other 146 network mobility use cases including intelligent transportation 147 systems and enterprise mobile device users. AERO is a secure 148 internetworking and mobility management service that employs the 149 Overlay Multilink Network Interface (OMNI) [I-D.templin-6man-omni] 150 Non-Broadcast, Multiple Access (NBMA) virtual link model. The OMNI 151 link is a virtual overlay configured over one or more underlying 152 Internetworks, and nodes on the link can exchange original IP packets 153 as single-hop neighbors. The OMNI Adaptation Layer (OAL) supports 154 end system multilink operation for increased reliability, bandwidth 155 optimization and traffic path selection while performing 156 fragmentation and reassembly to support Internetwork segment routing 157 and Maximum Transmission Unit (MTU) diversity. 159 The AERO service comprises Clients, Proxy/Servers and Relays that are 160 seen as OMNI link neighbors as well as Bridges that interconnect 161 diverse Internetworks as OMNI link segments through OAL forwarding at 162 a layer below IP. Each node's OMNI interface uses an IPv6 link-local 163 address format that supports operation of the IPv6 Neighbor Discovery 164 (ND) protocol [RFC4861] and links ND to IP forwarding. A node's OMNI 165 interface can be configured over multiple underlying interfaces, and 166 therefore appears as a single interface with multiple link-layer 167 addresses. Each link-layer address is subject to change due to 168 mobility and/or multilink fluctuations, and link-layer address 169 changes are signaled by ND messaging the same as for any IPv6 link. 171 AERO provides a secure cloud-based service where mobile node Clients 172 may use any Proxy/Server acting as a Mobility Anchor Point (MAP) and 173 fixed nodes may use any Relay on the link for efficient 174 communications. Fixed nodes forward original IP packets destined to 175 other AERO nodes via the nearest Relay, which forwards them through 176 the cloud. A mobile node's initial packets are forwarded through the 177 Proxy/Server, and direct routing is supported through route 178 optimization while packets are flowing. Both unicast and multicast 179 communications are supported, and mobile nodes may efficiently move 180 between locations while maintaining continuous communications with 181 correspondents and without changing their IP Address. 183 AERO Bridges are interconnected in a secured private BGP overlay 184 routing instance to provide an OAL routing/bridging service that 185 joins the underlying Internetworks of multiple disjoint 186 administrative domains into a single unified OMNI link at a layer 187 below IP. Each OMNI link instance is characterized by the set of 188 Mobility Service Prefixes (MSPs) common to all mobile nodes. Relays 189 provide an optimal route from correspondent nodes on the underlying 190 Internetwork to nodes on the OMNI link. To the underlying 191 Internetwork, the Relay is the source of a route to the MSP, and 192 hence uplink traffic to the mobile node is naturally routed to the 193 nearest Relay. 195 AERO can be used with OMNI links that span private-use Internetworks 196 and/or public Internetworks such as the global Internet. In the 197 latter case, some end systems may be located behind global Internet 198 Network Address Translators (NATs). A means for robust traversal of 199 NATs while avoiding "triangle routing" is therefore provided. 201 AERO assumes the use of PIM Sparse Mode in support of multicast 202 communication. In support of Source Specific Multicast (SSM) when a 203 Mobile Node is the source, AERO route optimization ensures that a 204 shortest-path multicast tree is established with provisions for 205 mobility and multilink operation. In all other multicast scenarios 206 there are no AERO dependencies. 208 AERO was designed as a secure aeronautical internetworking service 209 for both manned and unmanned aircraft, where the aircraft is treated 210 as a mobile node that can connect an Internet of Things (IoT). AERO 211 is also applicable to a wide variety of other use cases. For 212 example, it can be used to coordinate the links of mobile nodes 213 (e.g., cellphones, tablets, laptop computers, etc.) that connect into 214 a home enterprise network via public access networks using tunneling 215 software such as OpenVPN [OVPN] with VPN or non-VPN services enabled 216 according to the appropriate security model. AERO can also be used 217 to facilitate terrestrial vehicular and urban air mobility (as well 218 as pedestrian communication services) for future intelligent 219 transportation systems 220 [I-D.ietf-ipwave-vehicular-networking][I-D.templin-ipwave-uam-its]. 221 Other applicable use cases are also in scope. 223 Along with OMNI, AERO provides secured optimal routing support for 224 the "6M's" of modern Internetworking, including: 226 1. Multilink - a mobile node's ability to coordinate multiple 227 diverse underlying data links as a single logical unit (i.e., the 228 OMNI interface) to achieve the required communications 229 performance and reliability objectives. 231 2. Multinet - the ability to span the OMNI link across multiple 232 diverse network administrative segments while maintaining 233 seamless end-to-end communications between mobile nodes and 234 correspondents such as air traffic controllers, fleet 235 administrators, etc. 237 3. Mobility - a mobile node's ability to change network points of 238 attachment (e.g., moving between wireless base stations) which 239 may result in an underlying interface address change, but without 240 disruptions to ongoing communication sessions with peers over the 241 OMNI link. 243 4. Multicast - the ability to send a single network transmission 244 that reaches multiple nodes belonging to the same interest group, 245 but without disturbing other nodes not subscribed to the interest 246 group. 248 5. Multihop - a mobile node vehicle-to-vehicle relaying capability 249 useful when multiple forwarding hops between vehicles may be 250 necessary to "reach back" to an infrastructure access point 251 connection to the OMNI link. 253 6. MTU assurance - the ability to deliver packets of various robust 254 sizes between peers without loss due to a link size restriction, 255 and to dynamically adjust packets sizes to achieve the optimal 256 performance for each independent traffic flow. 258 The following numbered sections present the AERO specification. The 259 appendices at the end of the document are non-normative. 261 2. Terminology 263 The terminology in the normative references applies; especially, the 264 terminology in the OMNI specification [I-D.templin-6man-omni] is used 265 extensively throughout. The following terms are defined within the 266 scope of this document: 268 IPv6 Neighbor Discovery (ND) 269 a control message service for coordinating neighbor relationships 270 between nodes connected to a common link. AERO uses the IPv6 ND 271 messaging service specified in [RFC4861]. 273 IPv6 Prefix Delegation 274 a networking service for delegating IPv6 prefixes to nodes on the 275 link. The nominal service is DHCPv6 [RFC8415], however alternate 276 services (e.g., based on ND messaging) are also in scope. Most 277 notably, a minimal form of prefix delegation known as "prefix 278 registration" can be used if the Client knows its prefix in 279 advance and can represent it in the IPv6 source address of an ND 280 message. 282 Access Network (ANET) 283 a node's first-hop data link service network (e.g., a radio access 284 network, cellular service provider network, corporate enterprise 285 network, etc.) that often provides link-layer security services 286 such as IEEE 802.1X and physical-layer security (e.g., "protected 287 spectrum") to prevent unauthorized access internally and with 288 border network-layer security services such as firewalls and 289 proxys that prevent unauthorized outside access. 291 ANET interface 292 a node's attachment to a link in an ANET. 294 Internetwork (INET) 295 a connected IP network topology with a coherent routing and 296 addressing plan and that provides a transit backbone service for 297 ANET end systems. INETs also provide an underlay service over 298 which the AERO virtual link is configured. Example INETs include 299 corporate enterprise networks, aviation networks, and the public 300 Internet itself. When there is no administrative boundary between 301 an ANET and the INET, the ANET and INET are one and the same. 303 INET interface 304 a node's attachment to a link in an INET. 306 *NET 307 a "wildcard" term referring to either ANET or INET when it is not 308 necessary to draw a distinction between the two. 310 *NET interface 311 a node's attachment to a link in a *NET. 313 *NET Partition 314 frequently, *NETs such as large corporate enterprise networks are 315 sub-divided internally into separate isolated partitions (a 316 technique also known as "network segmentation"). Each partition 317 is fully connected internally but disconnected from other 318 partitions, and there is no requirement that separate partitions 319 maintain consistent Internet Protocol and/or addressing plans. 320 (Each *NET partition is seen as a separate OMNI link segment as 321 discussed below.) 323 *NET address 324 an IP address assigned to a node's interface connection to a *NET. 326 *NET encapsulation 327 the encapsulation of a packet in an outer header or headers that 328 can be routed within the scope of the local *NET partition. 330 OMNI link 331 the same as defined in [I-D.templin-6man-omni], and manifested by 332 IPv6 encapsulation [RFC2473]. The OMNI link spans underlying *NET 333 segments joined by virtual bridges in a spanning tree the same as 334 a bridged campus LAN. AERO nodes on the OMNI link appear as 335 single-hop neighbors even though they may be separated by multiple 336 underlying *NET hops, and can use Segment Routing [RFC8402] to 337 cause packets to visit selected waypoints on the link. 339 OMNI Interface 340 a node's attachment to an OMNI link. Since the addresses assigned 341 to an OMNI interface are managed for uniqueness, OMNI interfaces 342 do not require Duplicate Address Detection (DAD) and therefore set 343 the administrative variable 'DupAddrDetectTransmits' to zero 344 [RFC4862]. 346 OMNI Adaptation Layer (OAL) 347 an OMNI interface process whereby original IP packets admitted 348 into the interface are wrapped in a mid-layer IPv6 header and 349 subject to fragmentation and reassembly. The OAL is also 350 responsible for generating MTU-related control messages as 351 necessary, and for providing addressing context for spanning 352 multiple segments of a bridged OMNI link. 354 original IP packet 355 a whole IP packet or fragment admitted into the OMNI interface by 356 the network layer prior to OAL encapsulation and fragmentation, or 357 an IP packet delivered to the network layer by the OMNI interface 358 following OAL decapsulation and reassembly. 360 OAL packet 361 an original IP packet encapsulated in OAL headers and trailers 362 before OAL fragmentation, or following OAL reassembly. 364 OAL fragment 365 a portion of an OAL packet following fragmentation but prior to 366 *NET encapsulation, or following *NET encapsulation but prior to 367 OAL reassembly. 369 (OAL) atomic fragment 370 an OAL packet that does not require fragmentation is always 371 encapsulated as an "atomic fragment" with a Fragment Header with 372 Fragment Offset and More Fragments both set to 0, but with a valid 373 Identification value. 375 (OAL) carrier packet 376 an encapsulated OAL fragment following *NET encapsulation or prior 377 to *NET decapsulation. OAL sources and destinations exchange 378 carrier packets over underlying interfaces, and may be separated 379 by one or more OAL intermediate nodes. OAL intermediate nodes may 380 perform re-encapsulation on carrier packets by removing the *NET 381 headers of the first hop network and replacing them with new *NET 382 headers for the next hop network. 384 OAL source 385 an OMNI interface acts as an OAL source when it encapsulates 386 original IP packets to form OAL packets, then performs OAL 387 fragmentation and *NET encapsulation to create carrier packets. 389 OAL destination 390 an OMNI interface acts as an OAL destination when it decapsulates 391 carrier packets, then performs OAL reassembly and decapsulation to 392 derive the original IP packet. 394 OAL intermediate node 395 an OMNI interface acts as an OAL intermediate node when it removes 396 the *NET headers of carrier packets received on a first segment, 397 then re-encapsulates the carrier packets in new *NET headers and 398 forwards them into the next segment. OAL intermediate nodes 399 decrement the Hop Limit of the OAL IPv6 header during re- 400 encapsulation, and discard the packet if the Hop Limit reaches 0. 402 underlying interface 403 a *NET interface over which an OMNI interface is configured. 405 Mobility Service Prefix (MSP) 406 an aggregated IP Global Unicast Address (GUA) prefix (e.g., 407 2001:db8::/32, 192.0.2.0/24, etc.) assigned to the OMNI link and 408 from which more-specific Mobile Network Prefixes (MNPs) are 409 delegated. OMNI link administrators typically obtain MSPs from an 410 Internet address registry, however private-use prefixes can 411 alternatively be used subject to certain limitations (see: 412 [I-D.templin-6man-omni]). OMNI links that connect to the global 413 Internet advertise their MSPs to their interdomain routing peers. 415 Mobile Network Prefix (MNP) 416 a longer IP prefix delegated from an MSP (e.g., 417 2001:db8:1000:2000::/56, 192.0.2.8/30, etc.) and delegated to an 418 AERO Client or Relay. 420 Mobile Network Prefix Link Local Address (MNP-LLA) 421 an IPv6 Link Local Address that embeds the most significant 64 422 bits of an MNP in the lower 64 bits of fe80::/64, as specified in 423 [I-D.templin-6man-omni]. 425 Mobile Network Prefix Unique Local Address (MNP-ULA) 426 an IPv6 Unique-Local Address derived from an MNP-LLA. 428 Administrative Link Local Address (ADM-LLA) 429 an IPv6 Link Local Address that embeds a 32-bit administratively- 430 assigned identification value in the lower 32 bits of fe80::/96, 431 as specified in [I-D.templin-6man-omni]. 433 Administrative Unique Local Address (ADM-ULA) 434 an IPv6 Unique-Local Address derived from an ADM-LLA. 436 AERO node 437 a node that is connected to an OMNI link and participates in the 438 AERO internetworking and mobility service. 440 AERO Client ("Client") 441 an AERO node that connects over one or more underlying interfaces 442 and requests MNP delegation/registration service from AERO Proxy/ 443 Servers. The Client assigns an MNP-LLA to the OMNI interface for 444 use in ND exchanges with other AERO nodes and forwards original IP 445 packets to correspondents according to OMNI interface neighbor 446 cache state. 448 AERO Proxy/Server ("Proxy/Server") 449 a dual-function node that provides a proxying service between AERO 450 Clients and external peers on its Client-facing ANET interfaces 451 (i.e., in the same fashion as for an enterprise network proxy) as 452 well as default forwarding and Mobility Anchor Point (MAP) 453 services for coordination with correspondents on its INET-facing 454 interfaces. (Proxy/Servers in the open INET instead configure 455 only an INET interface and no ANET interfaces.) The Proxy/Server 456 configures an OMNI interface and assigns an ADM-LLA to support the 457 operation of IPv6 ND services, while advertising all of its 458 associated MNPs via BGP peerings with Bridges. Note that the 459 Proxy and Server functions can be considered logically separable, 460 but since each Proxy/Server must be informed of all of the 461 Client's other multilink Proxy/Server affiliations the AERO 462 service is best supported when the two functions are coresident on 463 the same physical or logical platform. 465 AERO Relay ("Relay") 466 a Proxy/Server that provides forwarding services between nodes 467 reached via the OMNI link and correspondents on connected 468 downstream links. AERO Relays configure an OMNI interface and 469 assign an ADM-LLA the same as Proxy/Servers. AERO Relays also run 470 a dynamic routing protocol to discover any non-MNP IP GUA routes 471 in service on its connected downstream network links. In both 472 cases, the Relay advertises the MSP(s) to its downstream networks, 473 and distributes all of its associated non-MNP IP GUA routes via 474 BGP peerings with Bridges (i.e., the same as for Proxy/Servers). 476 AERO Bridge ("Bridge") 477 a node that provides hybrid routing/bridging services (as well as 478 a security trust anchor) for nodes on an OMNI link. The Bridge 479 forwards carrier packets between OMNI link segments as OAL 480 intermediate nodes while decrementing the OAL IPv6 header Hop 481 Limit but without decrementing the network layer IP TTL/Hop Limit. 482 AERO Bridges peer with Proxy/Servers and other Bridges over 483 secured tunnels to discover the full set of MNPs for the link as 484 well as any non-MNP IP GUA routes that are reachable via Relays. 486 link-layer address 487 an IP address used as an encapsulation header source or 488 destination address from the perspective of the OMNI interface. 489 When an upper layer protocol (e.g., UDP) is used as part of the 490 encapsulation, the port number is also considered as part of the 491 link-layer address. 493 network layer address 494 the source or destination address of an original IP packet 495 presented to the OMNI interface. 497 end user network (EUN) 498 an internal virtual or external edge IP network that an AERO 499 Client or Relay connects to the rest of the network via the OMNI 500 interface. The Client/Relay sees each EUN as a "downstream" 501 network, and sees the OMNI interface as the point of attachment to 502 the "upstream" network. 504 Mobile Node (MN) 505 an AERO Client and all of its downstream-attached networks that 506 move together as a single unit, i.e., an end system that connects 507 an Internet of Things. 509 Mobile Router (MR) 510 a MN's on-board router that forwards original IP packets between 511 any downstream-attached networks and the OMNI link. The MR is the 512 MN entity that hosts the AERO Client. 514 Route Optimization Source (ROS) 515 the AERO node nearest the source that initiates route 516 optimization. The ROS may be a Proxy/Server or Relay acting on 517 behalf of the source, or may be the source Client itself. 519 Route Optimization responder (ROR) 520 the AERO node nearest the target destination that responds to 521 route optimization requests. The ROR may be a Proxy/Server acting 522 on behalf of a target MNP Client or a Relay for a non-MNP 523 destination. 525 MAP List 526 a geographically and/or topologically referenced list of addresses 527 of all Proxy/Servers within the same OMNI link. Each OMNI link 528 has its own MAP list. 530 Distributed Mobility Management (DMM) 531 a BGP-based overlay routing service coordinated by Proxy/Servers 532 and Bridges that tracks all Proxy/Server-to-Client associations. 534 Mobility Service (MS) 535 the collective set of all Proxy/Servers, Bridges and Relays that 536 provide the AERO Service to Clients. 538 Mobility Service Endpoint MSE) 539 an individual Proxy/Server, Bridge or Relay in the Mobility 540 Service. 542 Throughout the document, the simple terms "Client", "Proxy/Server", 543 "Bridge" and "Relay" refer to "AERO Client", "AERO Proxy/Server", 544 "AERO Bridge" and "AERO Relay", respectively. Capitalization is used 545 to distinguish these terms from other common Internetworking uses in 546 which they appear without capitalization. 548 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 549 the names of node variables, messages and protocol constants) is used 550 throughout this document. The terms "All-Routers multicast", "All- 551 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 552 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 553 generically refer to either Internet Protocol version, i.e., IPv4 554 [RFC0791] or IPv6 [RFC8200]. 556 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 557 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 558 "OPTIONAL" in this document are to be interpreted as described in BCP 559 14 [RFC2119][RFC8174] when, and only when, they appear in all 560 capitals, as shown here. 562 3. Asymmetric Extended Route Optimization (AERO) 564 The following sections specify the operation of IP over OMNI links 565 using the AERO service: 567 3.1. AERO Node Types 569 AERO Clients are Mobile Nodes (MNs) that configure OMNI interfaces 570 over underlying interfaces with addresses that may change when the 571 Client moves to a new network connection point. AERO Clients 572 register their Mobile Network Prefixes (MNPs) with the AERO service, 573 and distribute the MNPs to nodes on EUNs. AERO Bridges, Proxy/ 574 Servers and Relays are critical infrastructure elements in fixed 575 (i.e., non-mobile) INET deployments and hence have permanent and 576 unchanging INET addresses. Together, they constitute the AERO 577 service which provides an OMNI link virtual overlay for connecting 578 AERO Clients. 580 AERO Bridges provide hybrid routing/bridging services (as well as a 581 security trust anchor) for nodes on an OMNI link. Bridges use 582 standard IPv6 routing to forward carrier packets both within the same 583 *NET partition and between disjoint *NET partitions based on an IPv6 584 encapsulation mid-layer known as the OMNI Adaptation Layer (OAL) 585 [I-D.templin-6man-omni]. During forwarding, the inner IP layer 586 experiences a virtual bridging service since the inner IP TTL/Hop 587 Limit is not decremented. Each Bridge also peers with Proxy/Servers 588 and other Bridges in a dynamic routing protocol instance to provide a 589 Distributed Mobility Management (DMM) service for the list of active 590 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 591 one or more Mobility Service Prefixes (MSPs) and configure secured 592 tunnels with Proxy/Servers, Relays and other Bridges; they further 593 maintain IP forwarding table entries for each MNP and any other 594 reachable non-MNP prefixes. 596 AERO Proxy/Servers in distributed *NET locations provide default 597 forwarding and mobility/multilink services for AERO Client Mobile 598 Nodes (MNs). Each Proxy/Server also peers with Bridges in a dynamic 599 routing protocol instance to advertise its list of associated MNPs 600 (see Section 3.2.3). Proxy/Servers facilitate prefix delegation/ 601 registration exchanges with Clients, where each delegated prefix 602 becomes an MNP taken from an MSP. Proxy/Servers forward carrier 603 packets between OMNI interface neighbors and track each Client's 604 mobility profiles. Proxy/Servers at ANET/INET boundaries provide a 605 conduit for ANET Clients to associate with peers reached through 606 external INETs. Proxy/Servers in the open INET support INET Clients 607 through authenticated IPv6 ND message exchanges. 609 AERO Relays are Proxy/Servers that provide forwarding services to 610 exchange original IP packets between the OMNI interface and INET/EUN 611 interfaces. Relays are provisioned with MNPs the same as for an AERO 612 Client, and also run a dynamic routing protocol to discover any non- 613 MNP IP routes. The Relay advertises the MSP(s) to its connected 614 networks, and distributes all of its associated MNP and non-MNP 615 routes via BGP peerings with Bridges 617 3.2. The AERO Service over OMNI Links 619 3.2.1. AERO/OMNI Reference Model 621 Figure 1 presents the basic OMNI link reference model: 623 +----------------+ 624 | AERO Bridge B1 | 625 | Nbr: S1, S2, P1| 626 |(X1->S1; X2->S2)| 627 | MSP M1 | 628 +-+------------+-+ 629 +--------------+ | Secured | +--------------+ 630 | AERO P/S S1 | | tunnels | | AERO P/S S2 | 631 | Nbr: C1, B1 +-----+ +-----+ Nbr: C2, B1 | 632 | default->B1 | | default->B1 | 633 | X1->C1 | | X2->C2 | 634 +-------+------+ +------+-------+ 635 | OMNI link | 636 X===+===+======================================+===+===X 637 | | 638 +-----+--------+ +--------+-----+ 639 |AERO Client C1| |AERO Client C2| 640 | Nbr: S1 | | Nbr: S2 | 641 | default->S1 | | default->S2 | 642 | MNP X1 | | MNP X2 | 643 +------+-------+ +-----+--------+ 644 | | 645 .-. .-. 646 ,-( _)-. ,-( _)-. 647 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 648 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 649 `-(______)-' +-------+ +-------+ `-(______)-' 651 Figure 1: AERO/OMNI Reference Model 653 In this model: 655 o the OMNI link is an overlay network service configured over one or 656 more underlying *NET partitions which may be managed by different 657 administrative authorities and have incompatible protocols and/or 658 addressing plans. 660 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 661 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 662 via BGP peerings over secured tunnels to Proxy/Servers (S1, S2). 663 Bridges connect the disjoint segments of a partitioned OMNI link. 665 o AERO Proxy/Servers S1 and S2 configure secured tunnels with Bridge 666 B1 and also provide mobility, multilink, multicast and default 667 router services for the MNPs of their associated Clients C1 and 668 C2. (AERO Proxy/Servers that act as Relays can also advertise 669 non-MNP routes for non-mobile correspondent nodes the same as for 670 MNP Clients.) 672 o AERO Clients C1 and C2 associate with Proxy/Servers S1 and S2, 673 respectively. They receive MNP delegations X1 and X2, and also 674 act as default routers for their associated physical or internal 675 virtual EUNs. Simple hosts H1 and H2 attach to the EUNs served by 676 Clients C1 and C2, respectively. 678 An OMNI link configured over a single *NET appears as a single 679 unified link with a consistent underlying network addressing plan. 680 In that case, all nodes on the link can exchange carrier packets via 681 simple *NET encapsulation, since the underlying *NET is connected. 682 In common practice, however, an OMNI link may be partitioned into 683 multiple "segments", where each segment is a distinct *NET 684 potentially managed under a different administrative authority (e.g., 685 as for worldwide aviation service providers such as ARINC, SITA, 686 Inmarsat, etc.). Individual *NETs may also themselves be partitioned 687 internally, in which case each internal partition is seen as a 688 separate segment. 690 The addressing plan of each segment is consistent internally but will 691 often bear no relation to the addressing plans of other segments. 692 Each segment is also likely to be separated from others by network 693 security devices (e.g., firewalls, proxys, packet filtering gateways, 694 etc.), and in many cases disjoint segments may not even have any 695 common physical link connections. Therefore, nodes can only be 696 assured of exchanging carrier packets directly with correspondents in 697 the same segment, and not with those in other segments. The only 698 means for joining the segments therefore is through inter-domain 699 peerings between AERO Bridges. 701 The same as for traditional campus LANs, multiple OMNI link segments 702 can be joined into a single unified link via a virtual bridging 703 service using the OMNI Adaptation Layer (OAL) [I-D.templin-6man-omni] 704 which inserts a mid-layer IPv6 encapsulation header that supports 705 inter-segment forwarding (i.e., bridging) without decrementing the 706 network-layer TTL/Hop Limit. This bridging of OMNI link segments is 707 shown in Figure 2: 709 . . . . . . . . . . . . . . . . . . . . . . . 710 . . 711 . .-(::::::::) . 712 . .-(::::::::::::)-. +-+ . 713 . (:::: Segment A :::)--|B|---+ . 714 . `-(::::::::::::)-' +-+ | . 715 . `-(::::::)-' | . 716 . | . 717 . .-(::::::::) | . 718 . .-(::::::::::::)-. +-+ | . 719 . (:::: Segment B :::)--|B|---+ . 720 . `-(::::::::::::)-' +-+ | . 721 . `-(::::::)-' | . 722 . | . 723 . .-(::::::::) | . 724 . .-(::::::::::::)-. +-+ | . 725 . (:::: Segment C :::)--|B|---+ . 726 . `-(::::::::::::)-' +-+ | . 727 . `-(::::::)-' | . 728 . | . 729 . ..(etc).. x . 730 . . 731 . . 732 . <- OMNI link Bridged by encapsulation -> . 733 . . . . . . . . . . . . . .. . . . . . . . . 735 Figure 2: Bridging OMNI Link Segments 737 Bridges, Proxy/Servers and Relay OMNI interfaces are configured over 738 both secured tunnels and open INET underlying interfaces over their 739 respective segments in a spanning tree topology rooted at the 740 Bridges. The "secured" spanning tree supports strong authentication 741 for control plane messages and may also be used to convey the initial 742 carrier packets in a flow. The "unsecured" spanning tree conveys 743 ordinary carrier packets without security codes and that must be 744 treated by destinations according to data origin authentication 745 procedures. Route optimization can be employed to cause carrier 746 packets to take more direct paths between OMNI link neighbors without 747 having to follow strict spanning tree paths. 749 3.2.2. Addressing and Node Identification 751 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 752 fe80::/64 [RFC4291] to assign LLAs used for network-layer addresses 753 in link-scoped IPv6 ND and data messages. AERO Clients use LLAs 754 constructed from MNPs (i.e., "MNP-LLAs") while other AERO nodes use 755 LLAs constructed from administrative identification values ("ADM- 756 LLAs") as specified in [I-D.templin-6man-omni]. Non-MNP routes are 757 also represented the same as for MNP-LLAs, but may include a prefix 758 that is not properly covered by the MSP. 760 AERO nodes also use the Unique Local Address (ULA) prefix fd00::/8 761 followed by a pseudo-random 40-bit OMNI domain identifier to form the 762 prefix [ULA]::/48, then include a 16-bit OMNI link identifier '*' to 763 form the prefix [ULA*]::/64 [RFC4291]. The AERO node then uses the 764 prefix [ULA*]::/64 to form "MNP-ULAs" or "ADM-ULA"s as specified in 765 [I-D.templin-6man-omni] to support OAL addressing. (The prefix 766 [ULA*]::/64 appearing alone and with no suffix represents "default".) 767 AERO Clients also use Temporary ULAs constructed per 768 [I-D.templin-6man-omni], where the addresses are typically used only 769 in initial control message exchanges until a stable MNP-LLA/ULA is 770 assigned. 772 AERO MSPs, MNPs and non-MNP routes are typically based on Global 773 Unicast Addresses (GUAs), but in some cases may be based on private- 774 use addresses. See [I-D.templin-6man-omni] for a full specification 775 of LLAs, ULAs and GUAs used by AERO nodes on OMNI links. 777 Finally, AERO Clients and Proxy/Servers configure node identification 778 values as specified in [I-D.templin-6man-omni]. 780 3.2.3. AERO Routing System 782 The AERO routing system comprises a private instance of the Border 783 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 784 and Proxy/Servers and does not interact with either the public 785 Internet BGP routing system or any underlying INET routing systems. 787 In a reference deployment, each Proxy/Server is configured as an 788 Autonomous System Border Router (ASBR) for a stub Autonomous System 789 (AS) using a 32-bit AS Number (ASN) [RFC4271] that is unique within 790 the BGP instance, and each Proxy/Server further uses eBGP to peer 791 with one or more Bridges but does not peer with other Proxy/Servers. 792 Each *NET of a multi-segment OMNI link must include one or more 793 Bridges, which peer with the Proxy/Servers within that *NET. All 794 Bridges within the same *NET are members of the same hub AS, and use 795 iBGP to maintain a consistent view of all active routes currently in 796 service. The Bridges of different *NETs peer with one another using 797 eBGP. 799 Bridges maintain forwarding table entries only for the MNP-ULAs 800 corresponding to MNP and non-MNP routes that are currently active, 801 and carrier packets destined to all other MNP-ULAs will correctly 802 incur Destination Unreachable messages due to the black-hole route. 803 In this way, Proxy/Servers and Relays have only partial topology 804 knowledge (i.e., they know only about the routes their directly 805 associated Clients and non-AERO links) and they forward all other 806 carrier packets to Bridges which have full topology knowledge. 808 Each OMNI link segment assigns a unique ADM-ULA sub-prefix of 809 [ULA*]::/96. For example, a first segment could assign 810 [ULA*]::1000/116, a second could assign [ULA*]::2000/116, a third 811 could assign [ULA*]::3000/116, etc. Within each segment, each Proxy/ 812 Server configures an ADM-ULA within the segment's prefix, e.g., the 813 Proxy/Servers within [ULA*]::2000/116 could assign the ADM-ULAs 814 [ULA*]::2011/116, [ULA*]::2026/116, [ULA*]::2003/116, etc. 816 The administrative authorities for each segment must therefore 817 coordinate to assure mutually-exclusive ADM-ULA prefix assignments, 818 but internal provisioning of ADM-ULAs an independent local 819 consideration for each administrative authority. For each ADM-ULA 820 prefix, the Bridge(s) that connect that segment assign the all-zero's 821 address of the prefix as a Subnet Router Anycast address. For 822 example, the Subnet Router Anycast address for [ULA*]::1023/116 is 823 simply [ULA*]::1000. 825 ADM-ULA prefixes are statically represented in Bridge forwarding 826 tables. Bridges join multiple segments into a unified OMNI link over 827 multiple diverse administrative domains. They support a bridging 828 function by first establishing forwarding table entries for their 829 ADM-ULA prefixes either via standard BGP routing or static routes. 830 For example, if three Bridges ('A', 'B' and 'C') from different 831 segments serviced [ULA*]::1000/116, [ULA*]::2000/116 and 832 [ULA*]::3000/116 respectively, then the forwarding tables in each 833 Bridge are as follows: 835 A: [ULA*]::1000/116->local, [ULA*]::2000/116->B, [ULA*]::3000/116->C 837 B: [ULA*]::1000/116->A, [ULA*]::2000/116->local, [ULA*]::3000/116->C 839 C: [ULA*]::1000/116->A, [ULA*]::2000/116->B, [ULA*]::3000/116->local 841 These forwarding table entries are permanent and never change, since 842 they correspond to fixed infrastructure elements in their respective 843 segments. 845 MNP ULAs are instead dynamically advertised in the AERO routing 846 system by Proxy/Servers and Relays that provide service for their 847 corresponding MNPs. For example, if three Proxy/Servers ('D', 'E' 848 and 'F') service the MNPs 2001:db8:1000:2000::/56, 849 2001:db8:3000:4000::/56 and 2001:db8:5000:6000::/56 then the routing 850 system would include: 852 D: [ULA*]:2001:db8:1000:2000/120 853 E: [ULA*]:2001:db8:3000:4000/120 855 F: [ULA*]:2001:db8:5000:6000/120 857 A full discussion of the BGP-based routing system used by AERO is 858 found in [I-D.ietf-rtgwg-atn-bgp]. 860 3.2.4. OMNI Link Segment Routing 862 With the Client and partition prefixes in place in Bridge forwarding 863 tables, the OMNI interface sends control and data carrier packets 864 toward AERO destination nodes located in different OMNI link segments 865 over the spanning tree. The OMNI interface uses the OMNI Adaptation 866 Layer (OAL) encapsulation service [I-D.templin-6man-omni], and 867 includes an OMNI Routing Header (ORH) as an extension to the OAL 868 header if final segment forwarding information is available, e.g., in 869 the neighbor cache. (For nodes located in the same OMNI link segment 870 the ORH may instead be omitted.) The ORH is formatted as shown in 871 Figure 3: 873 0 1 2 3 874 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 1 875 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 876 | Next Header | Hdr Ext Len | Routing Type | FMT | SRT | 877 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 878 | Last Hop Segment-id (LHS) | 879 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 880 ~ Link Layer Address (L2ADDR) ~ 881 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 882 ~ Destination Suffix (non-first-fragments only) ~ 883 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 884 | omIndex | Null Padding (if necessary) ~ 885 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 887 Figure 3: OMNI Routing Header (ORH) Format 889 In this format: 891 o Next Header identifies the type of header immediately following 892 the ORH. 894 o Hdr Ext Len is the length of the Routing header in 8-octet units 895 (not including the first 8 octets), with trailing padding added if 896 necessary to produce an integral number of 8-octet units. 898 o Routing Type is set to TBD1 (see IANA Considerations). 900 o Segments Left is omitted, and replaced by a 3-bit FMT field and 901 5-bit SRT field. 903 o FMT - a 3-bit "Framework/Mode/Type" code corresponding to the 904 included Link Layer Address as follows: 906 * When the most significant bit (i.e., "Framework") is set to 1, 907 the Client is responsible for reassembly; otherwise, the Proxy/ 908 Server must reassemble before forwarding to the Client (while 909 re-fragmenting, if necessary). 911 * When the next most significant bit (i.e., "Mode") is set to 1, 912 L2ADDR is the address of the Proxy/Server and the Client must 913 be reached through the Proxy/Server; otherwise, the Client can 914 be reached on the open *NET where it may be located behind one 915 or more NATs and L2ADDR is either the address of the Proxy/ 916 Server or the native INET address of the Client itself. 918 * When the least significant bit (i.e., "Type") is set to 1, the 919 ORH includes the LHS and L2ADDR fields; otherwise, the fields 920 are omitted. 922 o SRT - a 5-bit Segment Routing Topology prefix length consulted 923 only when the FMT Type is 1. Encodes a value that (when added to 924 96) determines the prefix length to apply to the ADM-ULA formed 925 from concatenating [ULA*]::/96 with the 32 bit LHS value that 926 follows (for example, the value 16 corresponds to the prefix 927 length 112). 929 o LHS - a 4-octet field present only when the FMT Type is 1 that 930 encodes the ID of a Last Hop Segment Proxy/Server for the target . 931 When SRT and LHS are both set to 0, the LHS is considered 932 unspecified. When SRT is set to 0 and LHS is non-zero, the prefix 933 length is set to 128. SRT and LHS provide guidance to the OMNI 934 interface forwarding algorithm. 936 o Link Layer Address (L2ADDR) - Present only when FMT Type is 1, 937 with the IP address version determined by consulting the ORH 938 length. When present, provides the link-layer address (i.e., the 939 encapsulation address) of the Proxy/Server or the target Client 940 itself. The UDP Port Number appears in the first two octets and 941 the IP address appears in the next 4 octets for IPv4 or 16 octets 942 for IPv6. The Port Number and IP address are recorded in network 943 byte order, and in ones-compliment "obfuscated" form per 944 [RFC4380]. The OMNI interface forwarding algorithm uses FMT/ 945 L2ADDR to determine the *NET encapsulation address for local 946 forwarding when SRT/LHS is located in the same OMNI link segment. 948 o Destination Suffix - an 8-octet field that MUST be included for 949 OAL non-first-fragments and MUST be omitted for OAL first- 950 fragments. When present, encodes the 64-bit MNP-ULA suffix for 951 the target Client. 953 o omIndex - a 1-octet field that encodes the interface Index for the 954 Client's underlying interface that is the target for this carrier 955 packet. 957 o Null Padding - zero-valued octets added as necessary to pad the 958 ORH to an integral number of 8-octet units. 960 AERO neighbors use OAL encapsulation and fragmentation to exchange 961 OAL packets as specified in [I-D.templin-6man-omni]. When an AERO 962 node's OMNI interface (acting as an OAL source) uses OAL 963 encapsulation for an original IP packet with source address 964 2001:db8:1:2::1 and destination address 2001:db8:1234:5678::1, it 965 sets the OAL header source address to its own ULA (e.g., 966 [ULA*]::2001:db8:1:2), sets the destination address to the MNP-ULA 967 corresponding to the IP destination address (e.g., 968 [ULA*]::2001:db8:1234:5678), sets the Traffic Class, Flow Label, Hop 969 Limit and Payload Length as discussed in [I-D.templin-6man-omni], 970 then finally selects an Identification and appends an OAL checksum. 972 If the neighbor cache information indicates that the target is in a 973 different segment, the OAL source next inserts an ORH immediately 974 following the OAL header while including Destination Suffix for non- 975 first-fragments only (in this case, the Destination Suffix is 976 2001:db8:1234:5678). Next, to direct the packet to a Bridge the 977 source prepares the ORH using FMT type 1 and overwrites the OAL 978 header destination address with the LHS Subnet Router Anycast address 979 (for example, for LHS 3000:4567 with SRT 16, the Subnet Router 980 Anycast address is [ULA*]::3000:0000). To send the packet directly 981 to the LHS Proxy/server, the OAL source instead prepares the ORH 982 using FMT type 0 and overwrites the OAL header destination address 983 with the Proxy/Server ADM-ULA.) 985 The OAL source then fragments the OAL packet, with each resulting OAL 986 fragment including the OAL/ORH headers while only the first fragment 987 includes the original IPv6 header. If the FMT code indicates that 988 the Client must reassemble, the Identification used for fragmentation 989 must be within the window for the Client; otherwise the 990 Identification must be within the window for the Client's Proxy/ 991 Server. (Note that if no actual fragmentation is required the OAL 992 source still prepares the packet as an "atomic" fragment that 993 includes a Fragment Header with Offset and More Fragments both set to 994 0.) The OAL source finally encapsulates each resulting OAL fragment 995 in an *NET header to form an OAL carrier packet, with source address 996 set to its own *NET address (e.g., 192.0.2.100) and destination set 997 to the *NET address of a Bridge (e.g., 192.0.2.1). 999 The carrier packet encapsulation format in the above example is shown 1000 in Figure 4: 1002 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1003 | *NET Header | 1004 | src = 192.0.2.100 | 1005 | dst = 192.0.2.1 | 1006 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1007 | OAL IPv6 Header | 1008 | src = [ULA*]::2001:db8:1:2 | 1009 | dst= [ULA*]::3000:0000 | 1010 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1011 | ORH (if necessary) | 1012 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1013 | OAL Fragment Header | 1014 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1015 | Original IP Header | 1016 | (first-fragment only) | 1017 | src = 2001:db8:1:2::1 | 1018 | dst = 2001:db8:1234:5678::1 | 1019 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1020 | | 1021 ~ ~ 1022 ~ Original Packet Body/Fragment ~ 1023 ~ ~ 1024 | | 1025 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1027 Figure 4: Carrier Packet Format 1029 In this format, the original IP header and packet body/fragment are 1030 from the original IP packet, the OAL header is an IPv6 header 1031 prepared according to [RFC2473], the ORH is a Routing Header 1032 extension of the OAL header, the Fragment Header identifies each 1033 fragment, and the INET header is prepared as discussed in 1034 Section 3.6. When the OAL source transmits the resulting carrier 1035 packets, they are forwarded over possibly multiple OAL intermediate 1036 nodes in the OMNI link spanning tree until they arrive at the OAL 1037 destination. 1039 This gives rise to a routing system that contains both Client MNP-ULA 1040 routes that may change dynamically due to regional node mobility and 1041 per-partition ADM-ULA routes that rarely if ever change. The 1042 spanning tree can therefore provide link-layer bridging by sending 1043 carrier packets over the spanning tree instead of network-layer 1044 routing according to MNP routes. As a result, opportunities for loss 1045 due to node mobility between different segments are mitigated. 1047 Note: When the OAL source and destination are on the same INET 1048 segment, the ORH is not needed and OAL header compression can be used 1049 to significantly reduce encapsulation overhead 1050 [I-D.templin-6man-omni]. 1052 Note: When the OAL source has multiple original IP packets to send to 1053 the same OAL destination, it can perform "packing" to generate a 1054 "super-packet" [I-D.templin-6man-omni]. In that case, the OAL/ORH 1055 super-packet may include up to N original IP packets as long as the 1056 total length of the super-packet does not exceed the OMNI interface 1057 MTU. 1059 Note: Use of an IPv6 "minimal encapsulation" format (i.e., an IPv6 1060 variant of [RFC2004]) based on extensions to the ORH was considered 1061 and abandoned. In the approach, the ORH would be inserted as an 1062 extension header to the original IPv6 packet header. The IPv6 1063 destination address would then be written into the ORH, and the ULA 1064 corresponding to the destination would be overwritten in the IPv6 1065 destination address. This would seemingly convey enough forwarding 1066 information so that OAL encapsulation could be avoided. However, 1067 this "minimal encapsulation" IPv6 packet would then have a non-ULA 1068 source address and ULA destination address, an incorrect value in 1069 upper layer protocol checksums, and a Hop Limit that is decremented 1070 within the spanning tree when it should not be. The insertion and 1071 removal of the ORH would also entail rewriting the Payload Length and 1072 Next Header fields - again, invalidating upper layer checksums. 1073 These irregularities would result in implementation challenges and 1074 the potential for operational issues, e.g., since actionable ICMPv6 1075 error reports could not be delivered to the original source. In 1076 order to address the issues, still more information such as the 1077 original IPv6 source address could be written into the ORH. However, 1078 with the additional information the benefit of the "minimal 1079 encapsulation" savings quickly diminishes, and becomes overshadowed 1080 by the implementation and operational irregularities. 1082 3.2.5. Segment Routing Topologies (SRTs) 1084 The 64-bit sub-prefixes of [ULA]::/48 identify up to 2^16 distinct 1085 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 1086 OMNI link overlay instance using a distinct set of ULAs, and emulates 1087 a Virtual LAN (VLAN) service for the OMNI link. In some cases (e.g., 1088 when redundant topologies are needed for fault tolerance and 1089 reliability) it may be beneficial to deploy multiple SRTs that act as 1090 independent overlay instances. A communication failure in one 1091 instance therefore will not affect communications in other instances. 1093 Each SRT is identified by a distinct value in bits 48-63 of 1094 [ULA]::/48, i.e., as [ULA0]::/64, [ULA1]::/64, [ULA2]::/64, etc. 1095 Each OMNI interface is identified by a unique interface name (e.g., 1096 omni0, omni1, omni2, etc.) and assigns an anycast ADM-ULA 1097 corresponding to its SRT prefix length. The anycast ADM-ULA is used 1098 for OMNI interface determination in Safety-Based Multilink (SBM) as 1099 discussed in [I-D.templin-6man-omni]. Each OMNI interface further 1100 applies Performance-Based Multilink (PBM) internally. 1102 3.2.6. Segment Routing For OMNI Link Selection 1104 An original IPv6 source can direct an IPv6 packet to an AERO node by 1105 including a standard IPv6 Segment Routing Header (SRH) [RFC8754] with 1106 the anycast ADM-ULA for the selected SRT as either the IPv6 1107 destination or as an intermediate hop within the SRH. This allows 1108 the original source to determine the specific OMNI link topology an 1109 original IPv6 packet will traverse when there may be multiple 1110 alternatives. 1112 When the AERO node processes the SRH and forwards the original IPv6 1113 packet to the correct OMNI interface, the OMNI interface writes the 1114 next IPv6 address from the SRH into the IPv6 destination address and 1115 decrements Segments Left. If decrementing would cause Segments Left 1116 to become 0, the OMNI interface deletes the SRH before forwarding. 1117 This form of Segment Routing supports Safety-Based Multilink (SBM). 1119 3.2.7. Segment Routing Within the OMNI Link 1121 OAL sources can insert an ORH for Segment Routing within the OMNI 1122 link to influence the paths of OAL packets sent to OAL destinations 1123 in remote segments without requiring all carrier packets to traverse 1124 strict spanning tree paths. 1126 When an AERO node's OMNI interface has an original IP packet to send 1127 to a target discovered through route optimization located in the same 1128 OMNI link segment, it acts as an OAL source to perform OAL 1129 encapsulation and fragmentation. The node then uses the target's 1130 Link Layer Address (L2ADDR) information for *NET encapsulation. 1132 When an AERO node's OMNI interface has an original IP packet to send 1133 to a route optimization target located in a remote OMNI link segment, 1134 it acts as an OAL source the same as above but also includes an ORH 1135 while setting the OAL destination to either the Subnet Router Anycast 1136 address for the final OMNI link segment for ORH FMT type 1, or the 1137 ADM-ULA of the LHS Proxy/Server for ORH FMT type 0. The OMNI 1138 interface then forwards the resulting carrier packets into the 1139 spanning tree. 1141 When a Bridge receives a carrier packet destined to its Subnet Router 1142 Anycast address with an ORH with SRT/LHS values corresponding to the 1143 local segment, it examines the FMT code to determine whether it can 1144 send the packet directly to the target Client's L2ADDR without 1145 involving the LHS Proxy/Server. If the Client can be reached 1146 directly, the Bridge then writes the MNP-ULA corresponding to the ORH 1147 Destination Suffix into the OAL destination address, decrements the 1148 OAL IPv6 header Hop Limit (and discards the packet if the Hop Limit 1149 reaches 0), removes the ORH, re-encapsulates the carrier packet 1150 according to L2ADDR then forwards the carrier packet directly to the 1151 target Client. If the Client cannot be reached directly, the Bridge 1152 instead writes the ADM-ULA corresponding to the LHS into the OAL 1153 destination address, decrements the Hop Limit, then forwards the 1154 packet back into the spanning tree without removing the ORH. In this 1155 way, the Bridge participates in route optimization to reduce traffic 1156 load and suboptimal routing through strict spanning tree paths. 1158 3.3. OMNI Interface Characteristics 1160 OMNI interfaces are virtual interfaces configured over one or more 1161 underlying interfaces classified as follows: 1163 o INET interfaces connect to an INET either natively or through one 1164 or more NATs. Native INET interfaces have global IP addresses 1165 that are reachable from any INET correspondent. The INET-facing 1166 interfaces of Proxy/Servers are native interfaces, as are Relay 1167 and Bridge interfaces. NATed INET interfaces connect to a private 1168 network behind one or more NATs that provide INET access. Clients 1169 that are behind a NAT are required to send periodic keepalive 1170 messages to keep NAT state alive when there are no carrier packets 1171 flowing. 1173 o ANET interfaces connect to an ANET that is separated from the open 1174 INET by a Proxy/Server. Proxy/Servers can actively issue control 1175 messages over the INET on behalf of the Client to reduce ANET 1176 congestion. 1178 o VPNed interfaces use security encapsulation over the INET to a 1179 Virtual Private Network (VPN) server that also acts as a Proxy/ 1180 Server. Other than the link-layer encapsulation format, VPNed 1181 interfaces behave the same as Direct interfaces. 1183 o Direct (i.e., single-hop point-to-point) interfaces connect a 1184 Client directly to a Proxy/Server without crossing any ANET/INET 1185 paths. An example is a line-of-sight link between a remote pilot 1186 and an unmanned aircraft. The same Client considerations apply as 1187 for VPNed interfaces. 1189 OMNI interfaces use OAL encapsulation and fragmentation as discussed 1190 in Section 3.2.4. OMNI interfaces use *NET encapsulation (see: 1191 Section 3.6) to exchange carrier packets with OMNI link neighbors 1192 over INET or VPNed interfaces as well as over ANET interfaces for 1193 which the Client and Proxy/Server may be multiple IP hops away. OMNI 1194 interfaces do not use link-layer encapsulation over Direct underlying 1195 interfaces or ANET interfaces when the Client and Proxy/Server are 1196 known to be on the same underlying link. 1198 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 1199 state the same as for any interface. OMNI interfaces use ND messages 1200 including Router Solicitation (RS), Router Advertisement (RA), 1201 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 1202 neighbor cache management. In environments where spoofing may be a 1203 threat, OMNI neighbors should employ OAL Identification window 1204 synchronization in their ND message exchanges. 1206 OMNI interfaces send ND messages with an OMNI option formatted as 1207 specified in [I-D.templin-6man-omni]. The OMNI option includes 1208 prefix registration information and Interface Attributes containing 1209 link information parameters for the OMNI interface's underlying 1210 interfaces. Each OMNI option may include multiple Interface 1211 Attributes sub-options, each identified by an omIndex value. 1213 A Client's OMNI interface may be configured over multiple underlying 1214 interface connections. For example, common mobile handheld devices 1215 have both wireless local area network ("WLAN") and cellular wireless 1216 links. These links are often used "one at a time" with low-cost WLAN 1217 preferred and highly-available cellular wireless as a standby, but a 1218 simultaneous-use capability could provide benefits. In a more 1219 complex example, aircraft frequently have many wireless data link 1220 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1221 directional, etc.) with diverse performance and cost properties. 1223 If a Client's multiple underlying interfaces are used "one at a time" 1224 (i.e., all other interfaces are in standby mode while one interface 1225 is active), then ND message OMNI options include Interface Attributes 1226 sub-options with the same underlying interface index. In that case, 1227 the Client would appear to have a single interface but with a 1228 dynamically changing link-layer address. 1230 If the Client has multiple active underlying interfaces, then from 1231 the perspective of ND it would appear to have multiple link-layer 1232 addresses. In that case, ND message OMNI options MAY include 1233 Interface Attributes sub-options with different underlying interface 1234 indexes. Every ND message need not include Interface Attributes for 1235 all underlying interfaces; for any attributes not included, the 1236 neighbor considers the status as unchanged. 1238 Bridge and Proxy/Server OMNI interfaces are configured over secured 1239 tunnel underlying interfaces for carrying IPv6 ND and BGP protocol 1240 control plane messages, plus open INET underlying interfaces for 1241 carrying unsecured messages. The OMNI interface configures both an 1242 ADM-LLA and its corresponding ADM-ULA, and acts as an OAL source to 1243 encapsulate and fragment original IP packets while presenting the 1244 resulting carrier packets to a secured or unsecured underlying 1245 interface. Note that Bridge and Proxy/Server BGP protocol TCP 1246 sessions are run directly over the OMNI interface using ADM-ULA 1247 source and destination addresses. The OMNI interface will 1248 encapsulate these as carrier packets even though the OAL header may 1249 use the same ADM-ULAs as the original IP header; these carrier 1250 packets must then be sent over a secured underlying interface. 1252 3.4. OMNI Interface Initialization 1254 AERO Proxy/Servers and Clients configure OMNI interfaces as their 1255 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1256 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1257 to ensure that original IP packets with destination addresses covered 1258 by an MNP not explicitly assigned to a non-OMNI interface are 1259 directed to the OMNI interface. 1261 OMNI interface initialization procedures for Proxy/Servers, Clients 1262 and Bridges are discussed in the following sections. 1264 3.4.1. AERO Proxy/Server and Relay Behavior 1266 When a Proxy/Server enables an OMNI interface, it assigns an 1267 ADM-{LLA,ULA} appropriate for the given OMNI link segment. The 1268 Proxy/Server also configures secured tunnels with one or more 1269 neighboring Bridges and engages in a BGP routing protocol session 1270 with each Bridge. 1272 The OMNI interface provides a single interface abstraction to the IP 1273 layer, but internally includes an NBMA nexus for sending carrier 1274 packets to OMNI interface neighbors over underlying INET interfaces 1275 and secured tunnels. The Proxy/Server further configures a service 1276 to facilitate ND exchanges with AERO Clients and manages per-Client 1277 neighbor cache entries and IP forwarding table entries based on 1278 control message exchanges. 1280 Relays are simply Proxy/Servers that run a dynamic routing protocol 1281 to redistribute routes between the OMNI interface and INET/EUN 1282 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1283 networks on the INET/EUN interfaces (i.e., the same as a Client would 1284 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1285 interfaces. The Relay further provides an attachment point of the 1286 OMNI link to a non-MNP-based global topology. 1288 3.4.2. AERO Client Behavior 1290 When a Client enables an OMNI interface, it assigns either an 1291 MNP-{LLA, ULA} or a Temporary ULA and sends RS messages with ND 1292 parameters over its underlying interfaces to a Proxy/Server, which 1293 returns an RA message with corresponding parameters. The RS/RA 1294 messages may pass through one or more NATs in the case of a Client's 1295 INET interface. (Note: if the Client used a Temporary ULA in its 1296 initial RS message, it will discover an MNP-{LLA, ULA} in the 1297 corresponding RA that it receives from the Proxy/Server and begin 1298 using these new addresses. If the Client is operating outside the 1299 context of AERO infrastructure such as in a Mobile Ad-hoc Network 1300 (MANET), however, it may continue using Temporary ULAs for Client-to- 1301 Client communications until it encounters an infrastructure element 1302 that can provide an MNP.) 1304 3.4.3. AERO Bridge Behavior 1306 AERO Bridges configure an OMNI interface and assign the ADM-ULA 1307 Subnet Router Anycast address for each OMNI link segment they connect 1308 to. Bridges configure secured tunnels with Proxy/Servers and other 1309 Bridges, and engage in a BGP routing protocol session with neighbors 1310 on the spanning tree (see: Section 3.2.3). 1312 3.5. OMNI Interface Neighbor Cache Maintenance 1314 Each OMNI interface maintains a conceptual neighbor cache that 1315 includes a Neighbor Cache Entry (NCE) for each neighbor it 1316 communicates with on the OMNI link per [RFC4861]. Each route 1317 optimization source NCE is indexed by the LLA of the neighbor, which 1318 must match the ULA used during OAL encapsulation. In addition to 1319 ordinary neighbor cache entries, proxy neighbor cache entries are 1320 created and maintained by AERO Proxy/Servers when they proxy Client 1321 ND message exchanges [RFC4389]. AERO Proxy/Servers maintain proxy 1322 neighbor cache entries for each of their associated Clients. 1324 To the list of NCE states in Section 7.3.2 of [RFC4861], Proxy/Server 1325 OMNI interfaces add an additional state DEPARTED that applies to 1326 Clients that have recently departed. The interface sets a 1327 "DepartTime" variable for the NCE to "DEPART_TIME" seconds. 1328 DepartTime is decremented unless a new ND message causes the state to 1329 return to REACHABLE. While a NCE is in the DEPARTED state, the 1330 Proxy/Server forwards carrier packets destined to the target Client 1331 to the Client's new location instead. When DepartTime decrements to 1332 0, the NCE is deleted. It is RECOMMENDED that DEPART_TIME be set to 1333 the default constant value REACHABLE_TIME plus 10 seconds (40 seconds 1334 by default) to allow a window for carrier packets in flight to be 1335 delivered while stale route optimization state may be present. 1337 Proxy/Servers can act as RORs on behalf of dependent Clients 1338 according to the Proxy Neighbor Advertisement specification in 1339 Section 7.2.8 of [RFC4861], while well-connected Clients can act as 1340 an ROR on their own behalf. When a Proxy/Server ROR receives an 1341 authentic NS message used for route optimization, it first searches 1342 for a proxy NCE for the target Client and accepts the message only if 1343 there is an entry. The Proxy/Server then returns a solicited NA 1344 message while creating a NCE for the ROS. Proxy/Server RORs also 1345 create or update a "Report List" entry for the ROS in the target 1346 Client's NCE with a "ReportTime" variable set to REPORT_TIME seconds. 1347 The ROR resets ReportTime when it receives a new authentic NS 1348 message, and otherwise decrements ReportTime while no authentic NS 1349 messages have been received. It is RECOMMENDED that REPORT_TIME be 1350 set to the default constant value REACHABLE_TIME plus 10 seconds (40 1351 seconds by default) to allow a window for route optimization to 1352 converge before ReportTime decrements below REACHABLE_TIME. 1354 When the ROS receives a solicited NA message response to its NS 1355 message used for route optimization, it creates or updates a NCE for 1356 the target network-layer and link-layer addresses. The ROS then 1357 (re)sets ReachableTime for the NCE to REACHABLE_TIME seconds and uses 1358 this value to determine whether carrier packets can be forwarded 1359 directly to the target, i.e., instead of via a default route. The 1360 ROS otherwise decrements ReachableTime while no further solicited NA 1361 messages arrive. It is RECOMMENDED that REACHABLE_TIME be set to the 1362 default constant value 30 seconds as specified in [RFC4861]. 1364 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1365 of NS keepalives sent when a correspondent may have gone unreachable, 1366 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1367 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1368 to limit the number of unsolicited NAs that can be sent based on a 1369 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1370 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1371 same as specified in [RFC4861]. 1373 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1374 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1375 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1376 different values are chosen, all nodes on the link MUST consistently 1377 configure the same values. Most importantly, DEPART_TIME and 1378 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1379 REACHABLE_TIME to avoid packet loss due to stale route optimization 1380 state. 1382 3.5.1. OMNI ND Messages 1384 OMNI interface IPv6 ND messages include OMNI options 1385 [I-D.templin-6man-omni] with per-neighbor information including 1386 Interface Attributes that provide Link-Layer Address and traffic 1387 selector information for the neighbor's underlying interfaces. This 1388 information is stored in the neighbor cache and provides the basis 1389 for the forwarding algorithm specified in Section 3.10. The 1390 information is cumulative and reflects the union of the OMNI 1391 information from the most recent ND messages received from the 1392 neighbor; it is therefore not required that each ND message contain 1393 all neighbor information. 1395 The OMNI option Interface Attributes for each underlying interface 1396 includes a two-part "Link-Layer Address" consisting of a simple IP 1397 encapsulation address determined by the FMT and L2ADDR fields and an 1398 ADM-ULA determined by the SRT and LHS fields. Underlying interfaces 1399 are further selected based on their associated traffic selectors. 1401 The OMNI option is distinct from any Source/Target Link-Layer Address 1402 Options (S/TLLAOs) that may appear in an ND message according to the 1403 appropriate IPv6 over specific link layer specification (e.g., 1404 [RFC2464]). If both an OMNI option and S/TLLAO appear, the former 1405 pertains to encapsulation addresses while the latter pertains to the 1406 native L2 address format of the underlying media 1408 OMNI interface IPv6 ND messages may also include other IPv6 ND 1409 options. In particular, solicitation messages may include Nonce and/ 1410 or Timestamp options if required for verification of advertisement 1411 replies. If an OMNI ND solicitation message includes a Nonce option, 1412 the advertisement reply must echo the same Nonce. If an OMNI ND 1413 solicitation message includes a Timestamp option, the advertisement 1414 reply should also include a Timestamp option. 1416 3.5.2. OMNI Neighbor Advertisement Message Flags 1418 As discussed in Section 4.4 of [RFC4861] NA messages include three 1419 flag bits R, S and O. OMNI interface NA messages treat the flags as 1420 follows: 1422 o R: The R ("Router") flag is set to 1 in the NA messages sent by 1423 all AERO/OMNI node types. Simple hosts that would set R to 0 do 1424 not occur on the OMNI link itself, but may occur on the downstream 1425 links of Clients and Relays. 1427 o S: The S ("Solicited") flag is set exactly as specified in 1428 Section 4.4. of [RFC4861], i.e., it is set to 1 for Solicited NAs 1429 and set to 0 for Unsolicited NAs (both unicast and multicast). 1431 o O: The O ("Override") flag is set to 0 for solicited proxy NAs 1432 returned by a Proxy/Server ROR and set to 1 for all other 1433 solicited and unsolicited NAs. For further study is whether 1434 solicited NAs for anycast targets apply for OMNI links. Since 1435 MNP-LLAs must be uniquely assigned to Clients to support correct 1436 ND protocol operation, however, no role is currently seen for 1437 assigning the same MNP-LLA to multiple Clients. 1439 3.5.3. OMNI Neighbor Window Synchronization 1441 In secured environments (e.g., such as between nodes on the same 1442 secured ANET), OMNI interface neighbors can exchange OAL packets 1443 using randomly-initialized and monotonically-increasing 1444 Identification values (modulo 2*32) without window synchronization. 1445 In environments where spoofing is considered a threat, OMNI interface 1446 neighbors instead invoke window synchronization in ND message 1447 exchanges to maintain send/receive window state in their respective 1448 neighbor cache entries as specified in [I-D.templin-6man-omni]. 1450 In the asymmetric window synchronization case, the initial ND message 1451 exchange establishes only the initiator's send window and the 1452 responder's receive window such that a corresponding NS/NA exchange 1453 would be needed to establish the reverse direction. In the symmetric 1454 case, the initiator and responder engage in a three-way handshake to 1455 symmetrically establish the send/receive windows of both parties. 1457 3.6. OMNI Interface Encapsulation and Re-encapsulation 1459 The OMNI interface admits original IP packets then (acting as an OAL 1460 source) performs OAL encapsulation and fragmentation as specified in 1461 [I-D.templin-6man-omni] while including an ORH if necessary as 1462 specified in Section 3.2.4. OAL encapsulation produces OAL packets, 1463 while OAL fragmentation turns them into OAL fragments which are then 1464 encapsulated in *NET headers as carrier packets. 1466 For carrier packets undergoing re-encapsulation at an OAL 1467 intermediate node, the OMNI interface decrements the OAL IPv6 header 1468 Hop Limit and discards the carrier packet if the Hop Limit reaches 0. 1469 The intermediate node next removes the *NET encapsulation headers 1470 from the first segment and re-encapsulates the packet in new *NET 1471 encapsulation headers for the next segment. 1473 When a Proxy/Server or Relay re-encapsulates a carrier packet 1474 received from a Client that includes an OAL but no ORH, it inserts an 1475 ORH immediately following the OAL header and adjusts the OAL payload 1476 length and destination address field. The inserted ORH will be 1477 removed by the final-hop Bridge, but its insertion and removal will 1478 not interfere with reassembly at the final destination. For this 1479 reason, Clients must reserve 40 bytes for a maximum-length ORH when 1480 they perform OAL encapsulation (see: Section 3.9). 1482 3.7. OMNI Interface Decapsulation 1484 OMNI interfaces (acting as OAL destinations) decapsulate and 1485 reassemble OAL packets into original IP packets destined either to 1486 the AERO node itself or to a destination reached via an interface 1487 other than the OMNI interface the original IP packet was received on. 1488 When carrier packets containing OAL fragments arrive, the OMNI 1489 interface reassembles as discussed in Section 3.9. 1491 3.8. OMNI Interface Data Origin Authentication 1493 AERO nodes employ simple data origin authentication procedures. In 1494 particular: 1496 o AERO Bridges and Proxy/Servers accept carrier packets received 1497 from secured underlying interfaces. 1499 o AERO Proxy/Servers and Clients accept carrier packets and original 1500 IP packets that originate from within the same secured ANET. 1502 o AERO Clients and Relays accept original IP packets from downstream 1503 network correspondents based on ingress filtering. 1505 o AERO Clients, Relays and Proxy/Servers verify carrier packet UDP/ 1506 IP encapsulation addresses according to [I-D.templin-6man-omni]. 1508 o AERO Clients (as well as Proxy/Servers and Relays when acting as 1509 OAL destinations) accept only carrier packets with Identification 1510 values within the current window for the OAL source neighbor when 1511 window synchronization is employed. 1513 AERO nodes silently drop any packets that do not satisfy the above 1514 data origin authentication procedures. Further security 1515 considerations are discussed in Section 6. 1517 3.9. OMNI Interface MTU 1519 The OMNI interface observes the link nature of tunnels, including the 1520 Maximum Transmission Unit (MTU), Maximum Reassembly Unit (MRU) and 1521 the role of fragmentation and reassembly [I-D.ietf-intarea-tunnels]. 1522 The OMNI interface employs an OMNI Adaptation Layer (OAL) that 1523 accommodates multiple underlying links with diverse MTUs while 1524 observing both a minimum and per-path Maximum Payload Size (MPS). 1525 The functions of the OAL and the OMNI interface MTU/MRU/MPS are 1526 specified in [I-D.templin-6man-omni] with MTU/MRU both set to the 1527 constant value 9180 bytes, with minimum MPS set to 400 bytes, and 1528 with per-path MPS set to potentially larger values depending on the 1529 underlying path. 1531 When the network layer presents an original IP packet to the OMNI 1532 interface, the OAL source encapsulates and fragments the original IP 1533 packet if necessary. When the network layer presents the OMNI 1534 interface with multiple original IP packets bound to the same OAL 1535 destination, the OAL source can concatenate them together into a 1536 single OAL super-packet as discussed in [I-D.templin-6man-omni]. The 1537 OAL source then fragments the OAL packet if necessary according to 1538 the minimum/path MPS such that the OAL headers appear in each 1539 fragment while the original IP packet header appears only in the 1540 first fragment. The OAL source then encapsulates each OAL fragment 1541 in *NET headers for transmission as carrier packets over an 1542 underlying interface connected to either a physical link such as 1543 Ethernet, WiFi and the like or a virtual link such as an Internet or 1544 higher-layer tunnel (see the definition of link in [RFC8200]). 1546 Note: A Client that does not (yet) have neighbor cache state for a 1547 target may omit the ORH in carrier packets with the understanding 1548 that a Proxy/Server may insert an ORH on its behalf. For this 1549 reason, Clients reserve 40 bytes for the largest possible ORH in 1550 their OAL fragment size calculations. 1552 Note: Although the ORH may be removed by a Bridge on the path (see: 1553 Section 3.10.3), this does not interfere with the destination's 1554 ability to reassemble. This is due to the fact that the ORH is not 1555 included in the fragmentable part; therefore, its removal does not 1556 invalidate the offset values in any fragment headers. 1558 3.10. OMNI Interface Forwarding Algorithm 1560 Original IP packets enter a node's OMNI interface either from the 1561 network layer (i.e., from a local application or the IP forwarding 1562 system) while carrier packets enter from the link layer (i.e., from 1563 an OMNI interface neighbor). All original IP packets and carrier 1564 packets entering a node's OMNI interface first undergo data origin 1565 authentication as discussed in Section 3.8. Those that satisfy data 1566 origin authentication are processed further, while all others are 1567 dropped silently. 1569 Original IP packets that enter the OMNI interface from the network 1570 layer are forwarded to an OMNI interface neighbor using OAL 1571 encapsulation and fragmentation to produce carrier packets for 1572 transmission over underlying interfaces. (If routing indicates that 1573 the original IP packet should instead be forwarded back to the 1574 network layer, the packet is dropped to avoid looping). Carrier 1575 packets that enter the OMNI interface from the link layer are either 1576 re-encapsulated and re-admitted into the OMNI link, or reassembled 1577 and forwarded to the network layer where they are subject to either 1578 local delivery or IP forwarding. In all cases, the OAL MUST NOT 1579 decrement the network layer TTL/Hop-count since its forwarding 1580 actions occur below the network layer. 1582 OMNI interfaces may have multiple underlying interfaces and/or 1583 neighbor cache entries for neighbors with multiple underlying 1584 interfaces (see Section 3.3). The OAL uses Interface Attribute 1585 traffic selectors (e.g., port number, flow specification, etc.) to 1586 select an outgoing underlying interface for each OAL packet based on 1587 the node's own interface attributes, and also to select a destination 1588 link-layer address based on the neighbor's underlying interface 1589 attributes. AERO implementations SHOULD allow for traffic selector 1590 values to be modified at runtime through network management. 1592 If multiple outgoing interfaces and/or neighbor interfaces match the 1593 traffic selectors, the AERO node replicates the OAL packet and sends 1594 one copy via each of the (outgoing / neighbor) interface pairs; 1595 otherwise, the node sends a single copy of the OAL packet via an 1596 interface with a matching traffic selector. (While not strictly 1597 required, successful delivery may be more likely when all fragments 1598 of the same OAL packet are sent over the same underlying interface.) 1599 AERO nodes keep track of which underlying interfaces are currently 1600 "reachable" or "unreachable", and only use "reachable" interfaces for 1601 forwarding purposes. 1603 The following sections discuss the OMNI interface forwarding 1604 algorithms for Clients, Proxy/Servers and Bridges. In the following 1605 discussion, an original IP packet's destination address is said to 1606 "match" if it is the same as a cached address, or if it is covered by 1607 a cached prefix (which may be encoded in an MNP-LLA). 1609 3.10.1. Client Forwarding Algorithm 1611 When an original IP packet enters a Client's OMNI interface from the 1612 network layer the Client searches for a NCE that matches the 1613 destination. If there is a match, the Client selects one or more 1614 "reachable" neighbor interfaces in the entry for forwarding purposes. 1615 If there is no NCE, the Client instead either enqueues the original 1616 IP packet and invokes route optimization or forwards the original IP 1617 packet toward a Proxy/Server. The Client (acting as an OAL source) 1618 performs OAL encapsulation and sets the OAL destination address to 1619 the MNP-ULA if there is a matching NCE; otherwise, it sets the OAL 1620 destination to the ADM-ULA of the Proxy/Server. If the Client has 1621 multiple original IP packets to send to the same neighbor, it can 1622 concatenate them in a single super-packet [I-D.templin-6man-omni]. 1624 The OAL source then performs fragmentation to create OAL fragments 1625 (see: Section 3.9), appends any *NET encapsulation, and sends the 1626 resulting carrier packets over underlying interfaces to the neighbor 1627 acting as an OAL destination. 1629 If the neighbor interface selected for forwarding is located on the 1630 same OMNI link segment and not behind a NAT, the Client forwards the 1631 carrier packets directly according to the L2ADDR information for the 1632 neighbor. If the neighbor interface is behind a NAT on the same OMNI 1633 link segment, the Client instead forwards the initial carrier packets 1634 to its Proxy/Server and initiates NAT traversal procedures. If the 1635 Client's intended source underlying interface is also behind a NAT 1636 and located on the same OMNI link segment, it sends a "direct bubble" 1637 over the interface per [RFC6081][RFC4380] to the L2ADDR found in the 1638 neighbor cache in order to establish state in its own NAT by 1639 generating traffic toward the neighbor (note that no response to the 1640 bubble is expected). 1642 The Client next sends an NS(NUD) message toward the MNP-ULA of the 1643 neighbor via its Proxy/Server as discussed in Section 3.15. If the 1644 Client receives an NA(NUD) from the neighbor over the underlying 1645 interface, it marks the neighbor interface as "trusted" and sends 1646 future carrier packets directly to the L2ADDR information for the 1647 neighbor instead of indirectly via the Proxy/Server. The Client must 1648 honor the neighbor cache maintenance procedure by sending additional 1649 direct bubbles and/or NS/NA(NUD) messages as discussed in 1650 [RFC6081][RFC4380] in order to keep NAT state alive as long as 1651 carrier packets are still flowing. 1653 When a carrier packet enters a Client's OMNI interface from the link- 1654 layer, if the OAL destination matches one of the Client's ULAs the 1655 Client (acting as an OAL destination) reassembles and decapsulates as 1656 necessary and delivers the original IP packet to the network layer. 1657 Otherwise, the Client drops the original IP packet and MAY return a 1658 network-layer ICMP Destination Unreachable message subject to rate 1659 limiting (see: Section 3.11). 1661 Note: Clients and their Proxy/Server (and other Client) peers can 1662 exchange original IP packets over ANET underlying interfaces without 1663 invoking the OAL, since the ANET is secured at the link and physical 1664 layers. By forwarding original IP packets without invoking the OAL, 1665 however, the ANET peers can engage only in classical path MTU 1666 discovery since the packets are subject to loss and/or corruption due 1667 to the various per-link MTU limitations that may occur within the 1668 ANET. Moreover, the original IP packets do not include either the 1669 OAL integrity check or per-packet Identification values that can be 1670 used for data origin authentication and link-layer retransmissions. 1671 The tradeoff therefore involves an assessment of the per-packet 1672 encapsulation overhead saved by bypassing the OAL vs. inheritance of 1673 classical network "brittleness". (Note however that ANET peers can 1674 send small original IP packets without invoking the OAL, while 1675 invoking the OAL for larger packets. This presents the beneficial 1676 aspects of both small packet efficiency and large packet robustness.) 1678 3.10.2. Proxy/Server and Relay Forwarding Algorithm 1680 When the Proxy/Server receives an original IP packet from the network 1681 layer, it drops the packet if routing indicates that it should be 1682 forwarded back to the network layer to avoid looping. Otherwise, the 1683 Proxy/Server regards the original IP packet the same as if it had 1684 arrived as carrier packets with OAL destination set to its own ADM- 1685 ULA. When the Proxy/Server receives carrier packets on underlying 1686 interfaces with OAL destination set to its own ADM-ULA, it performs 1687 OAL reassembly if necessary to obtain the original IP packet. 1689 The Proxy/Server next searches for a NCE that matches the original IP 1690 destination and proceeds as follows: 1692 o if the original IP packet destination matches a NCE, the Proxy/ 1693 Sever uses one or more "reachable" neighbor interfaces in the 1694 entry for packet forwarding using OAL encapsulation and 1695 fragmentation according to the cached link-layer address 1696 information. If the neighbor interface is in a different OMNI 1697 link segment, the Proxy/Server forwards the resulting carrier 1698 packets via the spanning tree to a Bridge; otherwise, it forwards 1699 the carrier packets directly to the neighbor. If the neighbor is 1700 behind a NAT, the Proxy/Server instead forwards initial carrier 1701 packets via a Bridge while sending an NS(NUD) to the neighbor. 1702 When the Proxy/Server receives the NA(NUD), it can begin 1703 forwarding carrier packets directly to the neighbor the same as 1704 discussed in Section 3.10.1 while sending additional NS(NUD) 1705 messages as necessary to maintain NAT state. Note that no direct 1706 bubbles are necessary since the Proxy/Server is by definition not 1707 located behind a NAT. 1709 o else, if the original IP destination matches a non-MNP route in 1710 the IP forwarding table or an ADM-LLA assigned to the Proxy/ 1711 Server's OMNI interface, the Proxy/Server acting as a Relay 1712 presents the original IP packet to the network layer for local 1713 delivery or IP forwarding. 1715 o else, the Proxy/Server initiates address resolution as discussed 1716 in Section 3.14, while retaining initial original IP packets in a 1717 small queue awaiting address resolution completion. 1719 When the Proxy/Server receives a carrier packet with OAL destination 1720 set to an MNP-ULA that does not match the MSP, it accepts the carrier 1721 packet only if data origin authentication succeeds and if there is a 1722 network layer routing table entry for a GUA route that matches the 1723 MNP-ULA. If there is no route, the Proxy/Server drops the carrier 1724 packet; otherwise, it reassembles and decapsulates to obtain the 1725 original IP packet and acts as a Relay to present it to the network 1726 layer where it will be delivered according to standard IP forwarding. 1728 When the Proxy/Server receives a carrier packet from one of its 1729 Client neighbors with OAL destination set to another node, it 1730 forwards the packets via a matching NCE or via the spanning tree if 1731 there is no matching entry. When the Proxy/Server receives a carrier 1732 packet with OAL destination set to the MNP-ULA of one of its Client 1733 neighbors established through RS/RA exchanges, it accepts the carrier 1734 packet only if data origin authentication succeeds. If the NCE state 1735 is DEPARTED, the Proxy/Server inserts an ORH that encodes the MNP-ULA 1736 destination suffix and changes the OAL destination address to the 1737 ADM-ULA of the new Proxy/Server, then re-encapsulates the carrier 1738 packet and forwards it to a Bridge which will eventually deliver it 1739 to the new Proxy/Server. 1741 If the neighbor cache state for the MNP-ULA is REACHABLE, the Proxy/ 1742 Server forwards the carrier packets to the Client which then must 1743 reassemble. (Note that the Proxy/Server does not reassemble carrier 1744 packets not explicitly addressed to its own ADM-ULA, since routing 1745 could direct some of the carrier packet of the same original IP 1746 packet through a different Proxy/Server.) In that case, the Client 1747 may receive fragments that are smaller than its link MTU but can 1748 still be reassembled. 1750 Note: Proxy/Servers may receive carrier packets with ORHs that 1751 include additional forwarding information. Proxy/Servers use the 1752 forwarding information to determine the correct interface for 1753 forwarding to the target destination Client, then remove the ORH and 1754 forward the carrier packet. If the ORH information instead indicates 1755 that the Proxy/Server is responsible for reassembly, the Proxy/Server 1756 reassembles first before re-encapsulating (and possibly also re- 1757 fragmenting) then forwards to the target Client. For a full 1758 discussion of cases when the Proxy/Server may receive carrier packets 1759 with ORHs, see: Section 3.14.6. 1761 Note: Clients and their Proxy/Server peers can exchange original IP 1762 packets over ANET underlying interfaces without invoking the OAL, 1763 since the ANET is secured at the link and physical layers. By 1764 forwarding original IP packets without invoking the OAL, however, the 1765 Client and Proxy/Server can engage only in classical path MTU 1766 discovery since the packets are subject to loss and/or corruption due 1767 to the various per-link MTU limitations that may occur within the 1768 ANET. Moreover, the original IP packets do not include either the 1769 OAL integrity check or per-packet Identification values that can be 1770 used for data origin authentication and link-layer retransmissions. 1771 The tradeoff therefore involves an assessment of the per-packet 1772 encapsulation overhead saved by bypassing the OAL vs. inheritance of 1773 classical network "brittleness". (Note however that ANET peers can 1774 send small original IP packets without invoking the OAL, while 1775 invoking the OAL for larger packets. This presents the beneficial 1776 aspects of both small packet efficiency and large packet robustness.) 1778 Note: When a Proxy/Server receives a (non-OAL) original IP packet 1779 from an ANET Client, or a carrier packet with OAL destination set to 1780 its own ADM-ULA from any Client, the Proxy/Server reassembles if 1781 necessary then performs ROS functions on behalf of the Client. The 1782 Client may at some later time begin sending carrier packets to the 1783 OAL address of the actual target instead of the Proxy/Server, at 1784 which point it may begin functioning as an ROS on its own behalf and 1785 thereby "override" the Proxy/Server's ROS role. 1787 Note: Proxy/Servers forward secure control plane carrier packets via 1788 the secured spanning tree and forwards other carrier packets via the 1789 unsecured spanning tree. When a Proxy/Server receives a carrier 1790 packet from the secured spanning tree, it considers the message as 1791 authentic without having to verify upper layer authentication 1792 signatures. When a Proxy/Server receives a carrier packet from the 1793 unsecured spanning tree, it verifies any upper layer authentication 1794 signatures and/or forwards the unsecured message toward the 1795 destination which must apply data origin authentication. 1797 Note: If the Proxy/Server has multiple original IP packets to send to 1798 the same neighbor, it can concatenate them in a single OAL super- 1799 packet [I-D.templin-6man-omni]. 1801 3.10.3. Bridge Forwarding Algorithm 1803 Bridges forward carrier packets the same as any IPv6 router. Bridges 1804 convey carrier packets that encapsulate IPv6 ND control messages or 1805 routing protocol control messages using security encapsulations, and 1806 may convey carrier packets that encapsulate ordinary data without 1807 including security encapsulations. When the Bridge receives a 1808 carrier packet, it removes the outer *NET header and searches for a 1809 forwarding table entry that matches the OAL destination address. The 1810 Bridge then processes the packet as follows: 1812 o if the carrier packet destination matches its ADM-ULA or the 1813 corresponding ADM-ULA Subnet Router Anycast address and the OAL 1814 header is followed by an ORH, the Bridge removes and sets aside 1815 the ORH then processes the carrier packet locally before 1816 forwarding. If the encapsulated packet is an atomic fragment 1817 containing an NA(NUD) message, the Bridge replaces the OMNI option 1818 Interface Attributes sub-option with information for its own 1819 interface while retaining the omIndex value supplied by the 1820 NA(NUD) message source. The Bridge next examines the ORH FMT 1821 code. If the code indicates the destination is a Client on the 1822 open *NET (or, a Client behind a NAT for which NAT traversal 1823 procedures have already converged) the Bridge writes the MNP-ULA 1824 formed from the ORH Destination Suffix into the OAL destination. 1825 The Bridge then re-encapsulates the carrier packet and forwards it 1826 to the ORH L2ADDR. For all other cases, the Bridge instead writes 1827 the ADM-ULA formed from the ORH SRT/LHS into the OAL destination 1828 address and forwards the carrier packet to the ADM-ULA Proxy/ 1829 Server while invoking NAT traversal procedures the same as for 1830 Proxy/Servers if necessary, noting that no direct bubbles are 1831 necessary since only the target Client and not the Bridge is 1832 behind a NAT 1834 o else, if the carrier packet destination matches its ADM-ULA or the 1835 corresponding ADM-ULA Subnet Router Anycast address and the OAL 1836 header is not followed by an ORH, the Bridge submits the packet 1837 for reassembly. When reassembly is complete, the Bridge submits 1838 the original packet to the IP layer to support local applications 1839 such as BGP routing protocol sessions. 1841 o else, if the carrier packet destination matches a forwarding table 1842 entry the Bridge forwards the carrier packet to the next hop. (If 1843 the destination matches an MSP without matching an MNP, however, 1844 the Bridge instead drops the packet and returns an ICMP 1845 Destination Unreachable message subject to rate limiting - see: 1846 Section 3.11). 1848 o else, the Bridge drops the packet and returns an ICMP Destination 1849 Unreachable as above. 1851 As for any IP router, the Bridge decrements the OAL IPv6 header Hop 1852 Limit when it forwards the carrier packet and drops the packet if the 1853 Hop Limit reaches 0. Therefore, only the Hop Limit in the OAL header 1854 is decremented and not the TTL/Hop Limit in the original IP packet 1855 header. Bridges do not insert OAL/ORH headers themselves; instead, 1856 they act as IPv6 routers and forward carrier packets based on their 1857 destination addresses. 1859 Bridges forward carrier packets received from a first segment via the 1860 unsecured spanning tree to the next segment also via the unsecured 1861 spanning tree. Bridges forward carrier packets received from a first 1862 segment via the secured spanning tree to the next segment also via 1863 the secured spanning tree. Bridges use a single IPv6 routing table 1864 that always determines the same next hop for a given OAL destination, 1865 where the secured/unsecured spanning tree is determined through the 1866 selection of the underlying interface to be used for transmission 1867 (i.e., a secured tunnel or an open INET interface). 1869 3.11. OMNI Interface Error Handling 1871 When an AERO node admits an original IP packet into the OMNI 1872 interface, it may receive link-layer or network-layer error 1873 indications. 1875 A link-layer error indication is an ICMP error message generated by a 1876 router in the INET on the path to the neighbor or by the neighbor 1877 itself. The message includes an IP header with the address of the 1878 node that generated the error as the source address and with the 1879 link-layer address of the AERO node as the destination address. 1881 The IP header is followed by an ICMP header that includes an error 1882 Type, Code and Checksum. Valid type values include "Destination 1883 Unreachable", "Time Exceeded" and "Parameter Problem" 1884 [RFC0792][RFC4443]. (OMNI interfaces ignore link-layer IPv4 1885 "Fragmentation Needed" and IPv6 "Packet Too Big" messages for carrier 1886 packets that are no larger than the minimum/path MPS as discussed in 1887 Section 3.9, however these messages may provide useful hints of probe 1888 failures during path MPS probing.) 1890 The ICMP header is followed by the leading portion of the carrier 1891 packet that generated the error, also known as the "packet-in-error". 1892 For ICMPv6, [RFC4443] specifies that the packet-in-error includes: 1893 "As much of invoking packet as possible without the ICMPv6 packet 1894 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1895 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1896 "Internet Header + 64 bits of Original Data Datagram", however 1897 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1898 ICMP datagram SHOULD contain as much of the original datagram as 1899 possible without the length of the ICMP datagram exceeding 576 1900 bytes". 1902 The link-layer error message format is shown in Figure 5 (where, "L2" 1903 and "L3" refer to link-layer and network-layer, respectively): 1905 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1906 ~ ~ 1907 | L2 IP Header of | 1908 | error message | 1909 ~ ~ 1910 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1911 | L2 ICMP Header | 1912 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1913 ~ ~ P 1914 | carrier packet *NET and OAL | a 1915 | encapsulation headers | c 1916 ~ ~ k 1917 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1918 ~ ~ t 1919 | original IP packet headers | 1920 | (first-fragment only) | i 1921 ~ ~ n 1922 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1923 ~ ~ e 1924 | Portion of the body of | r 1925 | the original IP packet | r 1926 | (all fragments) | o 1927 ~ ~ r 1928 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1930 Figure 5: OMNI Interface Link-Layer Error Message Format 1932 The AERO node rules for processing these link-layer error messages 1933 are as follows: 1935 o When an AERO node receives a link-layer Parameter Problem message, 1936 it processes the message the same as described as for ordinary 1937 ICMP errors in the normative references [RFC0792][RFC4443]. 1939 o When an AERO node receives persistent link-layer Time Exceeded 1940 messages, the IP ID field may be wrapping before earlier fragments 1941 awaiting reassembly have been processed. In that case, the node 1942 should begin including integrity checks and/or institute rate 1943 limits for subsequent packets. 1945 o When an AERO node receives persistent link-layer Destination 1946 Unreachable messages in response to carrier packets that it sends 1947 to one of its neighbor correspondents, the node should process the 1948 message as an indication that a path may be failing, and 1949 optionally initiate NUD over that path. If it receives 1950 Destination Unreachable messages over multiple paths, the node 1951 should allow future carrier packets destined to the correspondent 1952 to flow through a default route and re-initiate route 1953 optimization. 1955 o When an AERO Client receives persistent link-layer Destination 1956 Unreachable messages in response to carrier packets that it sends 1957 to one of its neighbor Proxy/Servers, the Client should mark the 1958 path as unusable and use another path. If it receives Destination 1959 Unreachable messages on many or all paths, the Client should 1960 associate with a new Proxy/Server and release its association with 1961 the old Proxy/Server as specified in Section 3.16.5. 1963 o When an AERO Proxy/Server receives persistent link-layer 1964 Destination Unreachable messages in response to carrier packets 1965 that it sends to one of its neighbor Clients, the Proxy/Server 1966 should mark the underlying path as unusable and use another 1967 underlying path. 1969 o When an AERO Proxy/Server receives link-layer Destination 1970 Unreachable messages in response to a carrier packet that it sends 1971 to one of its permanent neighbors, it treats the messages as an 1972 indication that the path to the neighbor may be failing. However, 1973 the dynamic routing protocol should soon reconverge and correct 1974 the temporary outage. 1976 When an AERO Bridge receives a carrier packet for which the network- 1977 layer destination address is covered by an MSP, the Bridge drops the 1978 packet if there is no more-specific routing information for the 1979 destination and returns a network-layer Destination Unreachable 1980 message subject to rate limiting. The Bridge writes the network- 1981 layer source address of the original IP packet as the destination 1982 address and uses one of its non link-local addresses as the source 1983 address of the message. 1985 When an AERO node receives a carrier packet for which reassembly is 1986 currently congested, it returns a network-layer Packet Too Big (PTB) 1987 message as discussed in [I-D.templin-6man-omni] (note that the PTB 1988 messages could indicate either "hard" or "soft" errors). 1990 AERO nodes that act as OAL destinations include ICMPv6 error messages 1991 intended for the OAL source as sub-options in the OMNI option of 1992 secured uNA messages. When the OAL source receives the uNA message, 1993 it can extract the ICMPv6 error message enclosed in the OMNI option. 1995 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1997 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1998 coordinated as discussed in the following Sections. 2000 3.12.1. AERO Service Model 2002 Each AERO Proxy/Server on the OMNI link is configured to facilitate 2003 Client prefix delegation/registration requests. Each Proxy/Server is 2004 provisioned with a database of MNP-to-Client ID mappings for all 2005 Clients enrolled in the AERO service, as well as any information 2006 necessary to authenticate each Client. The Client database is 2007 maintained by a central administrative authority for the OMNI link 2008 and securely distributed to all Proxy/Servers, e.g., via the 2009 Lightweight Directory Access Protocol (LDAP) [RFC4511], via static 2010 configuration, etc. Clients receive the same service regardless of 2011 the Proxy/Servers they select. 2013 AERO Clients and Proxy/Servers use ND messages to maintain neighbor 2014 cache entries. AERO Proxy/Servers configure their OMNI interfaces as 2015 advertising NBMA interfaces, and therefore send unicast RA messages 2016 with a short Router Lifetime value (e.g., ReachableTime seconds) in 2017 response to a Client's RS message. Thereafter, Clients send 2018 additional RS messages to keep Proxy/Server state alive. 2020 AERO Clients and Proxy/Servers include prefix delegation and/or 2021 registration parameters in RS/RA messages (see 2022 [I-D.templin-6man-omni]). The ND messages are exchanged between 2023 Client and Proxy/Server according to the prefix management schedule 2024 required by the service. If the Client knows its MNP in advance, it 2025 can employ prefix registration by including its MNP-LLA as the source 2026 address of an RS message and with an OMNI option with valid prefix 2027 registration information for the MNP. If the Proxy/Server accepts 2028 the Client's MNP assertion, it injects the MNP into the routing 2029 system and establishes the necessary neighbor cache state. If the 2030 Client does not have a pre-assigned MNP, it can instead employ prefix 2031 delegation by including the unspecified address (::) as the source 2032 address of an RS message and with an OMNI option with prefix 2033 delegation parameters to request an MNP. 2035 The following sections specify the Client and Proxy/Server behavior. 2037 3.12.2. AERO Client Behavior 2039 AERO Clients discover the addresses of Proxy/Servers in a similar 2040 manner as described in [RFC5214]. Discovery methods include static 2041 configuration (e.g., from a flat-file map of Proxy/Server addresses 2042 and locations), or through an automated means such as Domain Name 2043 System (DNS) name resolution [RFC1035]. Alternatively, the Client 2044 can discover Proxy/Server addresses through a layer 2 data link login 2045 exchange, or through a unicast RA response to a multicast/anycast RS 2046 as described below. In the absence of other information, the Client 2047 can resolve the DNS Fully-Qualified Domain Name (FQDN) 2048 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 2049 text string and "[domainname]" is a DNS suffix for the OMNI link 2050 (e.g., "example.com"). 2052 To associate with a Proxy/Server, the Client acts as a requesting 2053 router to request MNPs by preparing an RS message with prefix 2054 management parameters. If the Client already knows the Proxy/ 2055 Server's ADM-LLA, it includes the LLA as the network-layer 2056 destination address; otherwise, the Client includes the (link-local) 2057 All-Routers multicast as the network-layer destination. If the 2058 Client already knows its own MNP-LLA, it can use the MNP-LLA as the 2059 network-layer source address and include an OMNI option with prefix 2060 registration information. Otherwise, the Client uses the unspecified 2061 address (::) as the network-layer source address and includes prefix 2062 delegation parameters in the OMNI option (see: 2063 [I-D.templin-6man-omni]). 2065 The Client next includes Interface Attributes corresponding to the 2066 underlying interface over which it will send the RS message, and MAY 2067 include additional Interface Attributes specific to other underlying 2068 interfaces. Next, the Client submits the RS for OAL encapsulation 2069 and fragmentation if necessary with its own MNP-ULA and the Proxy/ 2070 Server's ADM-ULA as the OAL addresses while selecting an 2071 Identification value and invoking window synchronization as specified 2072 in [I-D.templin-6man-omni]. 2074 The Client then sends the RS (either directly via Direct interfaces, 2075 via a VPN for VPNed interfaces, via an access router for ANET 2076 interfaces or via INET encapsulation for INET interfaces) then waits 2077 up to RetransTimer milliseconds for an RA message reply (see 2078 Section 3.12.3) (retrying up to MAX_RTR_SOLICITATIONS). If the 2079 Client receives no RAs, or if it receives an RA with Router Lifetime 2080 set to 0, the Client SHOULD abandon attempts through the first Proxy/ 2081 Server and try another Proxy/Server. Otherwise, the Client processes 2082 the prefix information found in the RA message. 2084 When the Client processes an RA, it first performs OAL reassembly and 2085 decapsulation if necessary then creates a NCE with the Proxy/Server's 2086 ADM-LLA as the network-layer address and the Proxy/Server's 2087 encapsulation and/or link-layer addresses as the link-layer address. 2088 The Client next records the RA Router Lifetime field value in the NCE 2089 as the time for which the Proxy/Server has committed to maintaining 2090 the MNP in the routing system via this underlying interface, and 2091 caches the other RA configuration information including Cur Hop 2092 Limit, M and O flags, Reachable Time and Retrans Timer. The Client 2093 then autoconfigures MNP-LLAs for any delegated MNPs and assigns them 2094 to the OMNI interface. The Client also caches any MSPs included in 2095 Route Information Options (RIOs) [RFC4191] as MSPs to associate with 2096 the OMNI link, and assigns the MTU value in the MTU option to the 2097 underlying interface. 2099 The Client then registers additional underlying interfaces with the 2100 Proxy/Server by sending RS messages via each additional interface as 2101 described above. The RS messages include the same parameters as for 2102 the initial RS/RA exchange, but with destination address set to the 2103 Proxy/Server's ADM-LLA. The Client finally sub-delegates the MNPs to 2104 its attached EUNs and/or the Client's own internal virtual interfaces 2105 as described in [I-D.templin-v6ops-pdhost] to support the Client's 2106 downstream attached "Internet of Things (IoT)". The Client then 2107 sends additional RS messages over each underlying interface before 2108 the Router Lifetime received for that interface expires. 2110 After the Client registers its underlying interfaces, it may wish to 2111 change one or more registrations, e.g., if an interface changes 2112 address or becomes unavailable, if traffic selectors change, etc. To 2113 do so, the Client prepares an RS message to send over any available 2114 underlying interface as above. The RS includes an OMNI option with 2115 prefix registration/delegation information, with Interface Attributes 2116 specific to the selected underlying interface, and with any 2117 additional Interface Attributes specific to other underlying 2118 interfaces. When the Client receives the Proxy/Server's RA response, 2119 it has assurance that the Proxy/Server has been updated with the new 2120 information. 2122 If the Client wishes to discontinue use of a Proxy/Server it issues 2123 an RS message over any underlying interface with an OMNI option with 2124 a prefix release indication. When the Proxy/Server processes the 2125 message, it releases the MNP, sets the NCE state for the Client to 2126 DEPARTED and returns an RA reply with Router Lifetime set to 0. 2127 After a short delay (e.g., 2 seconds), the Proxy/Server withdraws the 2128 MNP from the routing system. 2130 3.12.3. AERO Proxy/Server Behavior 2132 AERO Proxy/Servers act as IP routers and support a prefix delegation/ 2133 registration service for Clients. Proxy/Servers arrange to add their 2134 ADM-LLAs to a static map of Proxy/Server addresses for the link and/ 2135 or the DNS resource records for the FQDN 2136 "linkupnetworks.[domainname]" before entering service. Proxy/Server 2137 addresses should be geographically and/or topologically referenced, 2138 and made available for discovery by Clients on the OMNI link. 2140 When a Proxy/Server receives a prospective Client's RS message on its 2141 OMNI interface, it SHOULD return an immediate RA reply with Router 2142 Lifetime set to 0 if it is currently too busy or otherwise unable to 2143 service the Client. Otherwise, the Proxy/Server performs OAL 2144 reassembly and decapsulation if necessary, then authenticates the RS 2145 message and processes the prefix delegation/registration parameters. 2146 The Proxy/Server first determines the correct MNPs to provide to the 2147 Client by processing the MNP-LLA prefix parameters and/or the DHCPv6 2148 OMNI sub-option. When the Proxy/Server returns the MNPs, it also 2149 creates a forwarding table entry for the MNP-ULA corresponding to 2150 each MNP so that the MNPs are propagated into the routing system 2151 (see: Section 3.2.3). For IPv6, the Proxy/Server creates an IPv6 2152 forwarding table entry for each MNP. For IPv4, the Proxy/Server 2153 creates an IPv6 forwarding table entry with the IPv4-compatibility 2154 MNP-ULA prefix corresponding to the IPv4 address. 2156 The Proxy/Server next creates a NCE for the Client using the base 2157 MNP-LLA as the network-layer address and with lifetime set to no more 2158 than the smallest prefix lifetime. Next, the Proxy/Server updates 2159 the NCE by recording the information in each Interface Attributes 2160 sub-option in the RS OMNI option. The Proxy/Server also records the 2161 actual OAL/*NET addresses and RS message window synchronization 2162 parameters (if any) in the NCE. 2164 Next, the Proxy/Server prepares an RA message using its ADM-LLA as 2165 the network-layer source address and the network-layer source address 2166 of the RS message as the network-layer destination address. The 2167 Proxy/Server sets the Router Lifetime to the time for which it will 2168 maintain both this underlying interface individually and the NCE as a 2169 whole. The Proxy/Server also sets Cur Hop Limit, M and O flags, 2170 Reachable Time and Retrans Timer to values appropriate for the OMNI 2171 link. The Proxy/Server includes the MNPs, any other prefix 2172 management parameters and an OMNI option with no Interface Attributes 2173 but with an Origin Indication sub-option per [I-D.templin-6man-omni] 2174 with the mapped and obfuscated Port Number and IP address 2175 corresponding to the Client's own INET address in the case of INET 2176 Clients or to the Proxy/Server's INET-facing address for all other 2177 Clients. The Proxy/Server then includes one or more RIOs that encode 2178 the MSPs for the OMNI link, plus an MTU option (see Section 3.9). 2179 The Proxy/Server finally forwards the message to the Client using OAL 2180 encapsulation/fragmentation if necessary while including an 2181 acknowledgement if the RS invoked window synchronization. 2183 After the initial RS/RA exchange, the Proxy/Server maintains a 2184 ReachableTime timer for each of the Client's underlying interfaces 2185 individually (and for the Client's NCE collectively) set to expire 2186 after ReachableTime seconds. If the Client (or Proxy) issues 2187 additional RS messages, the Proxy/Server sends an RA response and 2188 resets ReachableTime. If the Proxy/Server receives an ND message 2189 with a prefix release indication it sets the Client's NCE to the 2190 DEPARTED state and withdraws the MNP from the routing system after a 2191 short delay (e.g., 2 seconds). If ReachableTime expires before a new 2192 RS is received on an individual underlying interface, the Proxy/ 2193 Server marks the interface as DOWN. If ReachableTime expires before 2194 any new RS is received on any individual underlying interface, the 2195 Proxy/Server sets the NCE state to STALE and sets a 10 second timer. 2196 If the Proxy/Server has not received a new RS or ND message with a 2197 prefix release indication before the 10 second timer expires, it 2198 deletes the NCE and withdraws the MNP from the routing system. 2200 The Proxy/Server processes any ND messages pertaining to the Client 2201 and returns an NA/RA reply in response to solicitations. The Proxy/ 2202 Server may also issue unsolicited RA messages, e.g., with reconfigure 2203 parameters to cause the Client to renegotiate its prefix delegation/ 2204 registrations, with Router Lifetime set to 0 if it can no longer 2205 service this Client, etc. Finally, If the NCE is in the DEPARTED 2206 state, the Proxy/Server deletes the entry after DepartTime expires. 2208 Note: Clients SHOULD notify former Proxy/Servers of their departures, 2209 but Proxy/Servers are responsible for expiring neighbor cache entries 2210 and withdrawing routes even if no departure notification is received 2211 (e.g., if the Client leaves the network unexpectedly). Proxy/Servers 2212 SHOULD therefore set Router Lifetime to ReachableTime seconds in 2213 solicited RA messages to minimize persistent stale cache information 2214 in the absence of Client departure notifications. A short Router 2215 Lifetime also ensures that proactive RS/RA messaging between Clients 2216 and Proxy/Servers will keep any NAT state alive (see above). 2218 Note: All Proxy/Servers on an OMNI link MUST advertise consistent 2219 values in the RA Cur Hop Limit, M and O flags, Reachable Time and 2220 Retrans Timer fields the same as for any link, since unpredictable 2221 behavior could result if different Proxy/Servers on the same link 2222 advertised different values. 2224 3.12.3.1. DHCPv6-Based Prefix Registration 2226 When a Client is not pre-provisioned with an MNP-LLA, it will need 2227 for the Proxy/Server to select one or more MNPs on its behalf and set 2228 up the correct state in the AERO routing service. (A Client with a 2229 pre-provisioned MNP may also request the Proxy/Server to select 2230 additional MNPs.) The DHCPv6 service [RFC8415] is used to support 2231 this requirement. 2233 When a Client needs to have the Proxy/Server select MNPs, it sends an 2234 RS message with source address set to the unspecified address (::) 2235 and with an OMNI option that includes a DHCPv6 message sub-option 2236 with DHCPv6 Prefix Delegation (DHCPv6-PD) parameters. When the 2237 Proxy/Server receives the RS message, it extracts the DHCPv6-PD 2238 message from the OMNI option. 2240 The Proxy/Server then acts as a "Proxy DHCPv6 Client" in a message 2241 exchange with the locally-resident DHCPv6 server, which delegates 2242 MNPs and returns a DHCPv6-PD Reply message. (If the Proxy/Server 2243 wishes to defer creation of MN state until the DHCPv6-PD Reply is 2244 received, it can instead act as a Lightweight DHCPv6 Relay Agent per 2245 [RFC6221] by encapsulating the DHCPv6-PD message in a Relay-forward/ 2246 reply exchange with Relay Message and Interface ID options.) 2248 When the Proxy/Server receives the DHCPv6-PD Reply, it adds a route 2249 to the routing system and creates an MNP-LLA based on the delegated 2250 MNP. The Proxy/Server then sends an RA back to the Client with the 2251 (newly-created) MNP-LLA as the destination address and with the 2252 DHCPv6-PD Reply message coded in the OMNI option. When the Client 2253 receives the RA, it creates a default route, assigns the Subnet 2254 Router Anycast address and sets its MNP-LLA based on the delegated 2255 MNP. 2257 Note: See [I-D.templin-6man-omni] for an MNP delegation alternative 2258 that avoids including a DHCPv6 message sub-option in the RS. Namely, 2259 when the Client requests a single MNP it can set the RS source to the 2260 unspecified address (::) and include a Node Identification sub-option 2261 and Preflen in the OMNI option (but with no DHCPv6 message sub- 2262 option). When the Proxy/Server receives the RS message, it forwards 2263 a self-generated DHCPv6 Solicit message to the DHCPv6 server on 2264 behalf of the Client. When the Proxy/Server receives the DHCPv6 2265 Reply, it prepares an RA message with an OMNI option with Preflen 2266 information (but with no DHCPv6 message sub-option), then places the 2267 (newly-created) MNP-LLA in the RA destination address and returns the 2268 message to the Client. 2270 3.13. The AERO Proxy Function 2272 Clients connect to the OMNI link via Proxy/Servers, with one Proxy/ 2273 Server for each underlying interface. Each of the Client's Proxy/ 2274 Servers must be informed of all of the Client's additional underlying 2275 interfaces. For Clients on Direct and VPNed underlying interfaces 2276 the Proxy/Server "A" for that interface is directly connected, for 2277 Clients on ANET underlying interfaces Proxy/Server "A" is located on 2278 the ANET/INET boundary, and for Clients on INET underlying interfaces 2279 Proxy/Server "A" is located somewhere in the connected Internetwork. 2280 When the Client registers with Proxy/Server "A", it must also report 2281 the registration to any other Proxy/Servers for other underlying 2282 interfaces "B", "C", "D", etc. for which an underlying interface 2283 relationship has already been established. The Proxy/Server 2284 satisfies these requirements as follows: 2286 o when Proxy/Server "A" receives an RS message from a new Client, it 2287 first authenticates the message then examines the network-layer 2288 destination address. If the destination address is Proxy/Server 2289 "A"'s ADM-LLA or (link-local) All-Routers multicast, Proxy/Server 2290 "A" creates a proxy NCE and caches the Client link-layer addresses 2291 along with the OMNI option information and any other identifying 2292 information including OAL Identification values, Client 2293 Identifiers, Nonce values, etc. If the RS message destination was 2294 the ADM-LLA of a different Proxy/Server "B" (or, if the OMNI 2295 option included an MS-Register sub-option with the ADM-LLA of a 2296 different Proxy/Server "B"), Proxy/Server "A" encapsulates a 2297 proxyed version of the RS message in an OAL header with source set 2298 to Proxy/Server "A"'s ADM-ULA and destination set to Proxy/Server 2299 "B"'s ADM-ULA. Proxy/Server "A" also includes an OMNI header with 2300 an Interface Attributes option that includes its own INET address 2301 plus a unique UDP Port Number for this Client, then forwards the 2302 message into the OMNI link secured spanning tree. (Note: 2303 including a unique Port Number allows Proxy/Server "B" to 2304 distinguish different Clients located behind the same Proxy/Server 2305 "A" at the link-layer, whereas the link-layer addresses would 2306 otherwise be indistinguishable.) 2308 o when the Proxy/Server "B" receives the RS, it authenticates the 2309 message then creates or updates a NCE for the Client with Proxy/ 2310 Server "A"'s ADM-ULA, INET address and UDP Port Number as the 2311 link-layer address information. Proxy/Server "B" then sends an RA 2312 message back to Proxy/Server "A" via the secured spanning tree. 2314 o when Proxy/Server "A" receives the RA, it authenticates the 2315 message and matches it with the proxy NCE created by the RS. 2316 Proxy/Server "A" then caches the prefix information as a mapping 2317 from the Client's MNPs to the Client's link-layer address, caches 2318 the Proxy/Server's advertised Router Lifetime and sets the NCE 2319 state to REACHABLE. The Proxy/Server then optionally rewrites the 2320 Router Lifetime and forwards the (proxyed) message to the Client. 2321 The Proxy/Server finally includes an MTU option (if necessary) 2322 with an MTU to use for the underlying ANET interface. 2324 o The Client repeats this process with each Proxy/Server "B", "C", 2325 "D" for each of its additional underlying interfaces. 2327 After the initial RS/RA exchanges each Proxy/Server forwards any of 2328 the Client's carrier packets for which there is no matching NCE to a 2329 Bridge using OAL encapsulation with its own ADM-ULA as the source and 2330 the MNP-ULA corresponding to the Client as the destination. The 2331 Proxy/Server instead forwards any carrier packets destined to a 2332 neighbor cache target directly to the target according to the OAL/ 2333 link-layer information - the process of establishing neighbor cache 2334 entries is specified in Section 3.14. 2336 While the Client is still associated with each Proxy/Server "A", "A" 2337 can send NS, RS and/or unsolicited NA messages to update the neighbor 2338 cache entries of other AERO nodes on behalf of the Client and/or to 2339 convey Interface Attribute updates. This allows for higher-frequency 2340 Proxy-initiated RS/RA messaging over well-connected INET 2341 infrastructure supplemented by lower-frequency Client-initiated RS/RA 2342 messaging over constrained ANET data links. 2344 If any Proxy/Server "B", "C", "D" ceases to send solicited 2345 advertisements, Proxy/Server "A" sends unsolicited RAs to the Client 2346 with destination set to (link-local) All-Nodes multicast and with 2347 Router Lifetime set to zero to inform Clients that a Proxy/Server has 2348 failed. Although Proxy/Server "A" can engage in ND exchanges on 2349 behalf of the Client, the Client can also send ND messages on its own 2350 behalf, e.g., if it is in a better position than "A" to convey 2351 Interface Attribute changes, etc. The ND messages sent by the Client 2352 include the Client's MNP-LLA as the source in order to differentiate 2353 them from the ND messages sent by Proxy/Server "A". 2355 If the Client becomes unreachable over an underlying interface, 2356 Proxy/Server "A" sets the NCE state to DEPARTED and retains the entry 2357 for DepartTime seconds. While the state is DEPARTED, Proxy/Server 2358 "A" forwards any carrier packets destined to the Client to a Bridge 2359 via OAL/ORH encapsulation. When DepartTime expires, Proxy/Server "A" 2360 deletes the NCE and discards any further carrier packets destined to 2361 the former Client. 2363 In some ANETs that employ a Proxy/Server, the Client's MNP can be 2364 injected into the ANET routing system. In that case, the Client can 2365 send original IP packets without invoking the OAL so that the ANET 2366 routing system transports the original IP packets to the Proxy. This 2367 can be very beneficial, e.g., if the Client connects to the ANET via 2368 low-end data links such as some aviation wireless links. 2370 If the ANET first-hop access router is on the same underlying link as 2371 the Client and recognizes the AERO/OMNI protocol, the Client can 2372 avoid OAL encapsulation for both its control and data messages. When 2373 the Client connects to the link, it can send an unencapsulated RS 2374 message with source address set to its own MNP-LLA (or to a Temporary 2375 LLA), and with destination address set to the ADM-LLA of the Client's 2376 selected Proxy/Server or to (link-local) All-Routers multicast. The 2377 Client includes an OMNI option formatted as specified in 2378 [I-D.templin-6man-omni]. The Client then sends the unencapsulated RS 2379 message, which will be intercepted by the AERO-Aware access router. 2381 The ANET access router then performs OAL encapsulation on the RS 2382 message and forwards it to a Proxy/Server at the ANET/INET boundary. 2383 When the access router and Proxy/Server are one and the same node, 2384 the Proxy/Server would share and underlying link with the Client but 2385 its message exchanges with outside correspondents would need to pass 2386 through a security gateway at the ANET/INET border. The method for 2387 deploying access routers and Proxys (i.e. as a single node or 2388 multiple nodes) is an ANET-local administrative consideration. 2390 Note: The Proxy/Server can apply packing as discussed in 2391 [I-D.templin-6man-omni] if an opportunity arises to concatenate 2392 multiple original IP packets that will be destined to the same 2393 neighbor. 2395 3.13.1. Detecting and Responding to Proxy/Server Failures 2397 In environments where fast recovery from Proxy/Server failure is 2398 required, Proxy/Server "A" SHOULD use proactive Neighbor 2399 Unreachability Detection (NUD) to track peer Proxy/Server "B" 2400 reachability in a similar fashion as for Bidirectional Forwarding 2401 Detection (BFD) [RFC5880]. Proxy/Server "A" can then quickly detect 2402 and react to failures so that cached information is re-established 2403 through alternate paths. The NUD control messaging is carried only 2404 over well-connected ground domain networks (i.e., and not low-end 2405 aeronautical radio links) and can therefore be tuned for rapid 2406 response. 2408 Proxy/Server "A" performs proactive NUD with peer Proxy/Server "B" 2409 for which there are currently active Clients by sending continuous NS 2410 messages in rapid succession, e.g., one message per second. Proxy/ 2411 Server "A" sends the NS message via the spanning tree with its own 2412 ADM-LLA as the source and the ADM-LLA of the peer Proxy/Server "B" as 2413 the destination. When Proxy/Server "A" is also sending RS messages 2414 to the peer Proxy/Server "B" on behalf of ANET Clients, the resulting 2415 RA responses can be considered as equivalent hints of forward 2416 progress. This means that Proxy/Server "B" need not also send a 2417 periodic NS if it has already sent an RS within the same period. If 2418 the peer Proxy/Server "B" fails (i.e., if "A" ceases to receive 2419 advertisements), Proxy/Server "A" can quickly inform Clients by 2420 sending multicast RA messages on the ANET interface. 2422 Proxy/Server "A" sends RA messages on the ANET interface with source 2423 address set to Proxy/Server "B"'s address, destination address set to 2424 (link-local) All-Nodes multicast, and Router Lifetime set to 0. 2425 Proxy/Server "A" SHOULD send MAX_FINAL_RTR_ADVERTISEMENTS RA messages 2426 separated by small delays [RFC4861]. Any Clients on the ANET that 2427 had been using the failed Proxy/Server "B" will receive the RA 2428 messages and associate with a new Proxy/Server. 2430 3.13.2. Point-to-Multipoint Proxy/Server Coordination 2432 In environments where Client messaging over ANETs is bandwidth- 2433 limited and/or expensive, Clients can enlist the services of Proxy/ 2434 Server "A" to coordinate with multiple Proxy/Servers "B", "C", "D" 2435 etc. in a single RS/RA message exchange. The Client can send a 2436 single RS message to (link-local) All-Routers multicast that includes 2437 the ID's of multiple Proxy/Servers in MS-Register sub-options of the 2438 OMNI option. 2440 When Proxy/Server "A" receives the RS and processes the OMNI option, 2441 it sends a separate RS to each MS-Register Proxy/Server ID. When 2442 Proxy/Server "A" receives an RA, it can optionally return an 2443 immediate "singleton" RA to the Client or record the Proxy/Server's 2444 ID for inclusion in a pending "aggregate" RA message. Proxy/Server 2445 "A" can then return aggregate RA messages to the Client including 2446 multiple Proxy/Server IDs in order to conserve bandwidth. Each RA 2447 includes a proper subset of the Proxy/Server IDs from the original RS 2448 message, and Proxy/Server "A" must ensure that the message contents 2449 of each RA are consistent with the information received from the 2450 (aggregated) additional Proxy/Servers. 2452 Clients can thereafter employ efficient point-to-multipoint Proxy/ 2453 Server coordination under the assistance of Proxy/Server "A" to 2454 reduce the number of messages sent over the ANET while enlisting the 2455 support of multiple Proxy/Servers for fault tolerance. Clients can 2456 further include MS-Release sub-options in IPv6 ND messages to request 2457 Proxy/Server "A" to release from former Proxy/Servers via the 2458 procedures discussed in Section 3.16.5. 2460 The OMNI interface specification [I-D.templin-6man-omni] provides 2461 further discussion of the RS/RA messaging involved in point-to- 2462 multipoint coordination. 2464 3.14. AERO Route Optimization 2466 AERO nodes invoke route optimization when they need to forward 2467 packets to new target destinations. Route optimization is based on 2468 IPv6 ND Address Resolution messaging between a Route Optimization 2469 Source (ROS) and Route Optimization Responder (ROR). Route 2470 optimization is initiated by the first eligible ROS closest to the 2471 source as follows: 2473 o For Clients on VPNed and Direct interfaces, the Proxy/Server is 2474 the ROS. 2476 o For Clients on ANET interfaces, either the Client or the Proxy/ 2477 Server may be the ROS. 2479 o For Clients on INET interfaces, the Client itself is the ROS. 2481 o For correspondent nodes on INET/EUN interfaces serviced by a 2482 Relay, the Relay is the ROS. 2484 The route optimization procedure is conducted between the ROS and the 2485 nearest Proxy/Server/Relay for the target selected by routing as the 2486 ROR. In this arrangement, the ROS is always the Client or 2487 Proxy/Server/Relay nearest the source over the selected source 2488 underlying interface, while the ROR is always a Proxy/Server/Relay 2489 that services the target regardless of the target underlying 2490 interface. 2492 The AERO routing system directs a route optimization solicitation 2493 sent by the ROS to the nearest available ROR, which returns a route 2494 optimization reply. The exact ROR selected is unimportant; all that 2495 matters is that the route optimization information returned must be 2496 current and authentic. The ROS is responsible for periodically 2497 refreshing the route optimization, and the ROR is responsible for 2498 quickly informing the ROS of any changes. 2500 The procedures are specified in the following sections. 2502 3.14.1. Route Optimization Initiation 2504 When an original IP packet from a source node destined to a target 2505 node arrives, the ROS checks for a NCE with an MNP-LLA that matches 2506 the target destination. If there is a NCE in the REACHABLE state, 2507 the ROS invokes the OAL and forwards the resulting carrier packets 2508 according to the cached state then returns from processing. 2509 Otherwise, if there is no NCE the ROS creates one in the INCOMPLETE 2510 state. 2512 The ROS next places the original IP packet on a short queue then 2513 sends an NS message for Address Resolution (NS(AR)) to receive a 2514 solicited NA(AR) message from an ROR. The NS(AR) message must be 2515 sent securely, and includes: 2517 o the LLA of the ROS as the source address. 2519 o the MNP-LLA corresponding to the original IP packet's destination 2520 as the Target Address, e.g., for 2001:db8:1:2::10:2000 the Target 2521 Address is fe80::2001:db8:1:2. 2523 o the Solicited-Node multicast address [RFC4291] formed from the 2524 lower 24 bits of the original IP packet's destination as the 2525 destination address, e.g., for 2001:db8:1:2::10:2000 the NS 2526 destination address is ff02:0:0:0:0:1:ff10:2000. 2528 The NS(AR) message also includes an OMNI option with an Interface 2529 Attributes entry for the sending interface, with S/T-omIndex set to 0 2530 and with Preflen set to the prefix length associated with the NS(AR) 2531 source. The ROS then selects an Identification value submits the 2532 NS(AR) message for OAL encapsulation with OAL source set to its own 2533 ULA and OAL destination set to the ULA corresponding to the target. 2534 (The ROS does not include any window synchronization parameters, 2535 since it will never forward OAL packets directly to the ROR). 2537 The ROS then sends the resulting carrier packet(s) into the secured 2538 spanning tree without decrementing the network-layer TTL/Hop Limit 2539 field. (When the ROS is an INET Client, it instead must ensure that 2540 the NS(AR) message fits within a single carrier packet, i.e., as an 2541 atomic fragment. The ROS then sends the resulting carrier packet to 2542 the ADM-ULA of one of its current Proxy/Servers which can verify the 2543 NS(AR) signature without having to reassemble then forward over the 2544 secured spanning tree on behalf of the Client.) 2546 3.14.2. Relaying the NS(AR) *NET Packet(s) 2548 When the Bridge receives the carrier packet(s) containing the RS from 2549 the ROS, it discards the *NET headers and determines the next hop by 2550 consulting its standard IPv6 forwarding table for the OAL header 2551 destination address. The Bridge then decrements the OAL header Hop- 2552 Limit, then re-encapsulates and forwards the carrier packet(s) via 2553 the secured spanning tree the same as for any IPv6 router, where it 2554 may traverse multiple OMNI link segments. The final-hop Bridge will 2555 deliver the carrier packet(s) via the secured spanning tree to a 2556 Proxy/Server or Relay that services the target. 2558 3.14.3. Processing the NS(AR) and Sending the NA(AR) 2560 When the target Proxy/Server (or Relay) receives the secured carrier 2561 packet(s), it reassembles if necessary then examines the NS(AR) 2562 target to determine whether it has a matching NCE and/or non-MNP 2563 route. If there is no match, the Proxy/Server drops the message. 2564 Otherwise, the Proxy/Server/Relay continues processing as follows: 2566 o if the NS(AR) target matches a Client NCE in the DEPARTED state, 2567 the Proxy/Server re-encapsulates while setting the OAL source to 2568 the ULA of the ROS and OAL destination address to the ADM-ULA of 2569 the Client's new Proxy/Server. The (old) Proxy/Server then 2570 fragments if necessary and forwards the resulting carrier 2571 packet(s) over the secured spanning tree then returns from 2572 processing. 2574 o If the NS(AR) target matches the MNP-LLA of a Client NCE in the 2575 REACHABLE state, the Proxy/Server makes note of whether the NS 2576 (AR) arrived from the secured or unsecured spanning tree then acts 2577 as an ROR to provide route optimization information on behalf of 2578 the Client. (Note that if the message arrived via the secured 2579 spanning tree the ROR need not perform further authentication, but 2580 if it arrived over an open INET underlying interface it must check 2581 for and verify the message authentication signature before 2582 accepting.) 2584 o If the NS(AR) target matches one of its non-MNP routes, the Relay 2585 acts as both an ROR and a route optimization target, since the 2586 Relay forwards IP packets toward the (fixed network) target at the 2587 network layer. 2589 The ROR next checks the target NCE for a Report List entry that 2590 matches the NS(AR) source LLA/ULA of the ROS. If there is a Report 2591 List entry, the ROR refreshes ReportTime for this ROR; otherwise, the 2592 ROR creates a new entry for the ROS. 2594 The ROR then prepares a (solicited) NA(AR) message to return to the 2595 ROS with the source address set to the target's MNP-LLA, the 2596 destination address set to the NS(AR) LLA source address and the 2597 Target Address set to the same value that appeared in the NS(AR). 2598 (If the NS(AR) included an MNP-LLA, the ROR must ensure that the 2599 NA(AR) can be sent as an atomic fragment.) The ROR then includes an 2600 OMNI option with Preflen set to the prefix length associated with the 2601 NA(AR) source address. The ROR next includes Interface Attributes in 2602 the OMNI option for all of the target's underlying interfaces with 2603 current information for each interface. 2605 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2606 according to the Proxy/Server's INET address for VPNed or Direct 2607 interfaces, to the INET address of the Proxy/Server for proxyed 2608 interfaces or to the Client's INET address for INET interfaces. The 2609 ROR then includes the lower 32 bits of the Proxy/Server's ADM-ULA as 2610 the LHS, encodes the ADM-ULA prefix length code in the SRT field and 2611 sets the FMT code as specified in Section 3.3. 2613 The ROR then sets the NA(AR) message R flag to 1 (as a router) and S 2614 flag to 1 (as a response to a solicitation) and sets the O flag to 0 2615 (as a proxy) and sets the S/T-omIndex field in the OMNI header to 0. 2616 The ROR finally submits the NA(AR) for OAL encapsulation with source 2617 set to its own ULA and destination set to the ULA of the ROS, then 2618 performs OAL encapsulation and fragmentation using the same 2619 Identification value that appeared in the NS(AR) and finally forwards 2620 the resulting (*NET-encapsulated) carrier packets via the secured 2621 spanning tree without decrementing the network-layer TTL/Hop Limit 2622 field. 2624 3.14.4. Relaying the NA(AR) 2626 When the Bridge receives NA(AR) carrier packets from the ROR, it 2627 discards the *NET header and determines the next hop by consulting 2628 its standard IPv6 forwarding table for the OAL header destination 2629 address. The Bridge then decrements the OAL header Hop-Limit, re- 2630 encapsulates the carrier packet and forwards it via the secured 2631 spanning tree the same as for any IPv6 router, where it may traverse 2632 multiple OMNI link segments. The final-hop Bridge will deliver the 2633 carrier packet via the secured spanning tree to a Proxy/Server for 2634 the ROS. 2636 3.14.5. Processing the NA(AR) 2638 When the ROS receives the NA(AR) message from the ROR, it first 2639 searches for a NCE that matches the NA(AR) LLA source address. The 2640 ROS then processes the message the same as for standard IPv6 Address 2641 Resolution [RFC4861]. In the process, it caches all OMNI option 2642 information in the target NCE (including all Interface Attributes). 2644 When the ROS is a Client, the solicited NA(AR) message will first be 2645 delivered via the secured spanning tree to one of its current Proxy/ 2646 Servers which then forwards the message to the Client. If the Client 2647 is on an ANET, ANET physical security and protected spectrum ensures 2648 security for the unmodified NA(AR); if the Client is on the open INET 2649 the Proxy/Server must instead insert an authentication signature. 2651 3.14.6. Forwarding Packets to Route Optimized Targets 2653 After the ROS receives the route optimization NA(AR) and updates the 2654 target NCE, it can begin forwarding packets along the best paths 2655 based on the target's Interface Attributes. The ROS selects target 2656 underlying interfaces according to traffic selectors and/or any other 2657 traffic discriminators, however each underlying interface selected 2658 must first establish window synchronization state if necessary. 2660 To establish window synchronization state, the ROS performs a secured 2661 unicast NS/NA exchange with window synchronization parameters 2662 according to Interface Attribute FMT codes. If the FMT code 2663 indicates that the target Client is responsible for reassembly, the 2664 ROS prepares an NS with its own LLA as the source and the target's 2665 MNP-LLA as the destination. The ROS then encapsulates the NS in an 2666 OAL header with the ADM-ULA of the ROS (or its Proxy/Server) as the 2667 source and the ADM-ULA of the target's LHS Proxy/Server as the 2668 destination and forwards the NS into the secured spanning tree. When 2669 the Proxy/Server receives the NS it changes the OAL destination to 2670 the MNP-ULA of the target Client and forwards the NS to the target 2671 Client over any of the Client's underlying interfaces. When the 2672 target Client receives the NS, it creates an NCE for the ROS, caches 2673 the window synchronization information and prepares an NA to return 2674 to the ROS with its MNP-LLA as the source and the LLA of the ROS as 2675 the destination. The Client then encapsulates the NA in an OAL 2676 header with its own MNP-ULA as the source and the ADM-ULA supplied by 2677 the ROS as the destination and forwards the NA into the secured 2678 spanning tree. When the destination ADM-ULA node receives the NA, it 2679 updates the target NCE based on the NA contents if the node is the 2680 ROS itself. Otherwise, the node acts as a Proxy/Server by changing 2681 the OAL destination to the MNP-ULA of the ROS and forwards the NA to 2682 the ROS which then updates its own NCE for the target Client. 2684 If the FMT code indicates that the LHS Proxy/Server (i.e., and not 2685 the target Client) is responsible for reassembly, the ROS instead 2686 sets the NS destination to the ADM-LLA of the LHS Proxy/Server and 2687 sets all other parameters the same as above. When the LHS Proxy/ 2688 Server receives the NS, it creates an NCE for the ROS, caches the 2689 window synchronization information, and returns an NA the same as 2690 above while using its own ADM-LLA as the source. When the ROS 2691 receives the NA, it must create and maintain a NCE for the LHS Proxy/ 2692 Server which it must regard as a companion to the existing NCE for 2693 the target Client. (The NCE for the LHS Proxy/Server can also be 2694 shared by multiple target Client NCEs if the ROS communicates with 2695 multiple active targets located behind the same Proxy/Server.) 2697 After window synchronization state has been established, the ROS can 2698 begin forwarding packets while performing additional NS/NA exchanges 2699 as above to update window state and/or test reachability. When the 2700 ROS has a packet to forward, it first checks the SRT, FMT and LHS 2701 codes for the selected underlying interface(s) and proceeds as 2702 follows: 2704 o If FMT indicates that the LHS Proxy/Server is responsible for 2705 reassembly, the ROS encapsulates the packet as an OAL packet with 2706 Identification set to the next send sequence number for this LHS 2707 Proxy/Server. The ROS sets the OAL source to its own ULA and set 2708 the OAL destination to the ADM-ULA of the LHS Proxy/Server. The 2709 ROS next submits the OAL packet for fragmentation and includes an 2710 ORH using FMT type 0 with each fragment as specified in 2711 Section 3.2.4 then sends each carrier packet into the unsecured 2712 spanning tree. When the LHS Proxy/Server receives the carrier 2713 packets, it examines the ORH to determine that it is responsible 2714 for reassembling and to determine the next hop target Client 2715 underlying interface identified by the omIndex. The LHS Proxy/ 2716 Server then removes the ORH, reassembles, and decapsulates to 2717 derive the original IP packet, then forwards the packet to the 2718 target Client over the underlying interface identified by omIndex 2719 while re-encapsulating and re-fragmenting if necessary. If the 2720 target Client is unreachable over the selected underlying 2721 interface, the LHS Proxy/Server instead drops the packet and 2722 returns a secured uNA message to the ROS with an ICMPv6 Error sub- 2723 option with a Destination Unreachable code. 2725 o If FMT indicates that the target Client is responsible for 2726 reassembly but that it must be reached via the LHS Proxy/Server, 2727 the ROS instead uses the next send sequence number for the Client 2728 as the Identification. The ROS then sets the OAL destination 2729 address to the ADM-ULA of the LHS Proxy/Server, fragments, and 2730 includes an ORH using FMT type 0 with each fragment the same as 2731 above then sends each carrier packet into the unsecured spanning 2732 tree. When the LHS Proxy/Server receives the carrier packets, it 2733 examines the ORH to determine that the Client is responsible for 2734 reassembling and to determine the next hop underlying interface 2735 based on omIndex. The LHS Proxy/Server then removes the ORH and 2736 forwards the carrier packets directly to the target Client without 2737 reassembling while returning a uNA Destination Unreachable the 2738 same as above if necessary. When the target Client receives the 2739 carrier packets, it first verifies that the Identifications are 2740 in-window for this ROS then reassembles. 2742 o If FMT indicates that the target Client is on the open 2743 Internetwork (where there may be NATs), the ROS proceeds the same 2744 as above except that it includes an ORH using FMT type 1 and sets 2745 the OAL destination address to the Subnet Router Anycast address 2746 for a Bridge on the final segment according to the SRT/LHS codes. 2747 When an LHS Bridge receives the resulting carrier packets, it 2748 forwards them directly to the target Client while bypassing the 2749 LHS Proxy/Server if possible; otherwise, it forwards the carrier 2750 packets to the LHS Proxy/Server. (Methods for the Bridge to 2751 determine whether direct or indirect forwarding must be used are 2752 discussed elsewhere throughout this document.) 2754 These forwarding procedures apply to the case where the selected 2755 target interface SRT/LHS codes indicate that the interface is located 2756 in a foreign OMNI link segment. In that case, the ROS must include 2757 ORHs and send the resulting carrier packets into the spanning tree. 2758 If the SRT/LHS codes indicate that the interface is in the local OMNI 2759 link segment, the ROS can instead forward carrier packets directly to 2760 the LHS Proxy/Server using the L2ADDR for encapsulation, or even to 2761 the target Client itself while invoking NAT traversal if necessary. 2762 When the ROS sends packets directly to the LHS Proxy/Server, it must 2763 still include an ORH using FMT type 0 so the Proxy/Server can 2764 determine whether or not to reassemble and the correct target Client 2765 interface for (re)forwarding. When the ROS sends packets directly to 2766 the target Client, it need not include an ORH. 2768 While the ROS continues to actively forward packets to the target 2769 Client, it is responsible for updating window synchronization state 2770 and per-interface reachability before expiration. Window 2771 synchronization state is shared by all underlying interfaces in the 2772 ROS' NCE that use the same destination LLA so that a single NS/ 2773 NA(NUD) exchange applies for all interfaces regardless of the 2774 (single) interface used to conduct the exchange. However, the window 2775 synchronization exchange only confirms target Client reachability 2776 over the specific interface used to conduct the exchange. 2777 Reachability for other underlying interfaces that share the same 2778 window synchronization state must be determined individually using 2779 NS/NA(NUD) messages which need not be secured as long as they use in- 2780 window Identifications and do not update other state information. 2782 3.15. Neighbor Unreachability Detection (NUD) 2784 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2785 [RFC4861] either reactively in response to persistent link-layer 2786 errors (see Section 3.11) or on-demand to confirm reachability. The 2787 NUD algorithm is based on periodic control message exchanges and may 2788 further be seeded by ND hints of forward progress, but care must be 2789 taken to avoid inferring reachability based on spoofed information. 2790 For example, IPv6 ND message exchanges that include authentication 2791 codes and/or in-window Identifications may be considered as 2792 acceptable hints of forward progress, while spurious random carrier 2793 packets should not be. 2795 AERO nodes can use standard NS/NA(NUD) exchanges sent over the OMNI 2796 link secured spanning tree to test reachability without risk of DoS 2797 attacks from nodes pretending to be a neighbor. These NS/NA(NUD) 2798 messages use the unicast LLAs and ULAs of the two parties involved in 2799 the NUD test the same as for standard IPv6 ND over the secured 2800 spanning tree. When only reachability information is required, 2801 unsecured NS/NA(NUD) messages may instead be exchanged directly 2802 between neighbors as long as they include in-window Identifications 2803 and do not update other state. 2805 When an ROR directs an ROS to a target neighbor with one or more 2806 link-layer addresses, the ROS probes each unsecured target underlying 2807 interface either proactively or on-demand of carrier packets directed 2808 to the path by multilink forwarding to maintain the interface's state 2809 as reachable. Probing is performed through NS(NUD) messages over 2810 either the secured or unsecured spanning tree, or NS(NUD) messages 2811 sent directly to an underlying interface of the target itself. While 2812 testing a target underlying interface, the ROS can optionally 2813 continue to send carrier packets via alternate interfaces and/or the 2814 ROR or maintain a small queue of carrier packets until target 2815 reachability is confirmed. 2817 When the ROS sends an NS(NUD) message, it sets the IPv6 source to its 2818 own LLA and sets both the destination and Target Address to the LLA 2819 of the target. The ROS also includes an OMNI option with S/T-omIndex 2820 set to the index of the target underlying interface and with a single 2821 Interface Attributes sub-option with the SRT, FMT, LHS and L2ADDR 2822 information for its own underlying interface. The ROS includes an 2823 Identification value within the current send window for this target 2824 neighbor, then encapsulates the message in OAL/INET headers with its 2825 own ULA as the source and the ULA formed from the SRT/LHS values for 2826 this target interface as the destination. The ROS then forwards the 2827 NS(NUD) message toward the ULA destination. 2829 When the ULA destination receives the NS(NUD) message, it examines 2830 the S/T-omIndex. If the corresponding underlying interface is a 2831 Direct or VPNed interface, the node acts as the target. If the 2832 underlying interface is a Proxyed interface, the node either acts as 2833 the target itself or forwards the message to the target Client. If 2834 the underlying interface is an INET interface, the node changes the 2835 OAL destination address to the ULA of the target Client and forwards 2836 the NS(NUD) over the underlying interface to the target Client while 2837 including an authentication signature. 2839 The target then creates a NCE for the ROS LLA address if necessary 2840 and caches receive window information for this ROS. The target then 2841 creates an NA(NUD) by reversing the OAL and IPv6 addresses from the 2842 NS(NUD) while copying the Identification value, and next including an 2843 Interface Attributes sub-option with attributes for its own interface 2844 identified by the NS(NUD) S/T-omIndex. The target sets the NA(NUD) 2845 S/T-omIndex to the index of the ROS, sets the Target Address to the 2846 same value that was in the NS(NUD), sets R flag to 1, the S flag to 1 2847 and the O flag to 1, and returns the message using its own underlying 2848 interface identified by NS(NUD) S/T-omIndex and destined to the ROS's 2849 interface identified by the original Interface Attributes sub-option. 2851 When the ROS receives the NA(NUD) message, it can determine from the 2852 Identification value and Target Address that the message matched its 2853 NS(NUD) and that it transited the direct path using the selected 2854 underlying interface pair. The ROS marks route optimization target 2855 paths that pass these NUD tests as "reachable", and those that do not 2856 as "unreachable". These markings inform the OMNI interface 2857 forwarding algorithm specified in Section 3.10. Following the 2858 initial NA(NUD), the ROS sends additional NS(NUDs) as necessary, but 2859 may send them via the cached SRT, FMT, LHS and L2ADDR for the target 2860 via each underlying interface to be tested. 2862 Note: If the target determines that the OMNI option Interface 2863 Attributes in the NS(NUD) is located in a different OMNI link segment 2864 than its own interface named in the S/T-omIndex, it instead returns 2865 the NA(NUD) via the secured or unsecured spanning tree while 2866 including an ORH and setting the OAL destination address to the 2867 Subnet Router Anycast address used by Bridges on the ROS segment. 2868 When a Bridge on the ROS segment receives the NA(NUD), it replaces 2869 the Interface Attributes with information for its own interface while 2870 using the omIndex value specific to the target. 2872 3.16. Mobility Management and Quality of Service (QoS) 2874 AERO is a Distributed Mobility Management (DMM) service. Each Proxy/ 2875 Server is responsible for only a subset of the Clients on the OMNI 2876 link, as opposed to a Centralized Mobility Management (CMM) service 2877 where there is a single network mobility collective entity for all 2878 Clients. Clients coordinate with their associated Proxy/Servers via 2879 RS/RA exchanges to maintain the DMM profile, and the AERO routing 2880 system tracks all current Client/Proxy/Server peering relationships. 2882 Proxy/Servers provide default routing and mobility/multilink services 2883 for their dependent Clients. Clients are responsible for maintaining 2884 neighbor relationships with their Proxy/Servers through periodic RS/ 2885 RA exchanges, which also serves to confirm neighbor reachability. 2886 When a Client's underlying Interface Attributes change, the Client is 2887 responsible for updating the Proxy/Server with this new information. 2888 Note that when there is a Proxy/Server in the path, the Proxy 2889 function can also perform some RS/RA exchanges on the Client's 2890 behalf. 2892 Mobility management messaging is based on the transmission and 2893 reception of unsolicited Neighbor Advertisement (uNA) messages. Each 2894 uNA message sets the IPv6 destination address to (link-local) All- 2895 Nodes multicast to convey a general update of Interface Attributes to 2896 (possibly) multiple recipients, or to a specific unicast LLA to 2897 announce a departure event to a specific recipient. Implementations 2898 must therefore examine the destination address to determine the 2899 nature of the mobility event (i.e., update vs departure). 2901 Mobility management considerations are specified in the following 2902 sections. 2904 3.16.1. Mobility Update Messaging 2906 RORs accommodate Client mobility and/or multilink change events by 2907 sending secured unsolicited NA (uNA) messages to each ROS in the 2908 target Client's Report List. When an ROR sends a uNA message, it 2909 sets the IPv6 source address to the its own LLA, sets the destination 2910 address to (link-local) All-Nodes multicast and sets the Target 2911 Address to the Client's MNP-LLA. The ROR also includes an OMNI 2912 option with Preflen set to the prefix length associated with the 2913 Client's MNP-LLA, with Interface Attributes for the target Client's 2914 underlying interfaces and with the OMNI header S/T-omIndex set to 0. 2915 The ROR then sets the NA R flag to 1, the S flag to 0 and the O flag 2916 to 1, then encapsulates the message in an OAL header with source set 2917 to its own ULA and destination set to the ULA of the ROS and sends 2918 the message into the secured spanning tree. 2920 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2921 reception of uNA messages is unreliable but provides a useful 2922 optimization. In well-connected Internetworks with robust data links 2923 uNA messages will be delivered with high probability, but in any case 2924 the Proxy/Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT 2925 uNAs to each ROS to increase the likelihood that at least one will be 2926 received. Alternatively, the Proxy/Server can set the PNG flag in 2927 the uNA OMNI option header to request a solicited NA acknowledgement 2928 as specified in [I-D.templin-6man-omni]. 2930 When the ROS receives a uNA message prepared as above, it ignores the 2931 message if there is no existing NCE for the target Client. 2932 Otherwise, it uses the included OMNI option information to update the 2933 NCE, but does not reset ReachableTime since the receipt of an 2934 unsolicited NA message from the ROR does not provide confirmation 2935 that any forward paths to the target Client are working. (If the uNA 2936 message PNG flag was set, the ROS then returns a solicited NA 2937 acknowledgement as specified in [I-D.templin-6man-omni]. 2939 In addition to sending uNA messages to the current set of ROSs for 2940 the Client, the ROR also sends uNAs to the MNP-ULA associated with 2941 the link-layer address for any underlying interface for which the 2942 link-layer address has changed. These uNA messages update an old 2943 Proxy/Server that cannot easily detect (e.g., without active probing) 2944 when a formerly-active Client has departed. When the ROR sends the 2945 uNA, it sets the IPv6 source address to its LLA, sets the destination 2946 address to the old Proxy/Server's ADM-LLA, and sets the Target 2947 Address to the Client's MNP-LLA. The ROR also includes an OMNI 2948 option with Preflen set to the prefix length associated with the 2949 Client's MNP-LLA, with Interface Attributes for the changed 2950 underlying interface, and with the OMNI header S/T-omIndex set to its 2951 own omIndex if the ROR is a Client or 0 if the ROR is a Proxy/Server. 2952 The ROR then sets the uNA R flag to 1, the S flag to 0 and the O flag 2953 to 1, then encapsulates the message in an OAL header with source set 2954 to its own ULA and destination set to the ADM-ULA of the old Proxy/ 2955 Server and sends the message into the secured spanning tree. 2957 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2959 When a Client needs to change its underlying Interface Attributes 2960 (e.g., due to a mobility event), the Client requests one of its 2961 Proxy/Servers to send RS messages to all of its other Proxy/Servers 2962 via the secured spanning tree with an OMNI option that includes 2963 Interface attributes with the new link quality and address 2964 information. 2966 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2967 sending carrier packets containing user data in case one or more RAs 2968 are lost. If all RAs are lost, the Client SHOULD re-associate with a 2969 new Proxy/Server. 2971 When the Proxy/Server receives the Client's changes, it sends uNA 2972 messages to all nodes in the Report List the same as described in the 2973 previous section. 2975 3.16.3. Bringing New Links Into Service 2977 When a Client needs to bring new underlying interfaces into service 2978 (e.g., when it activates a new data link), it sends an RS message to 2979 the Proxy/Server via the underlying interface with an OMNI option 2980 that includes Interface Attributes with appropriate link quality 2981 values and with link-layer address information for the new link. 2983 3.16.4. Deactivating Existing Links 2985 When a Client needs to deactivate an existing underlying interface, 2986 it sends an RS or uNA message to its Proxy/Server with an OMNI option 2987 with appropriate Interface Attribute values - in particular, the link 2988 quality value 0 assures that neighbors will cease to use the link. 2990 If the Client needs to send RS/uNA messages over an underlying 2991 interface other than the one being deactivated, it MUST include 2992 Interface Attributes with appropriate link quality values for any 2993 underlying interfaces being deactivated. 2995 Note that when a Client deactivates an underlying interface, 2996 neighbors that have received the RS/uNA messages need not purge all 2997 references for the underlying interface from their neighbor cache 2998 entries. The Client may reactivate or reuse the underlying interface 2999 and/or its omIndex at a later point in time, when it will send RS/uNA 3000 messages with fresh Interface Attributes to update any neighbors. 3002 3.16.5. Moving Between Proxy/Servers 3004 The Client performs the procedures specified in Section 3.12.2 when 3005 it first associates with a new Proxy/Server or renews its association 3006 with an existing Proxy/Server. The Client also includes MS-Release 3007 identifiers in the RS message OMNI option per [I-D.templin-6man-omni] 3008 if it wants the new Proxy/Server to notify any old Proxy/Servers from 3009 which the Client is departing. 3011 When the new Proxy/Server receives the Client's RS message, it 3012 returns an RA as specified in Section 3.12.3 and sends uNA messages 3013 to any old Proxy/Servers listed in OMNI option MS-Release 3014 identifiers. When the new Proxy/Server sends a uNA message, it sets 3015 the IPv6 source address to the Client's MNP-LLA, sets the destination 3016 address to the old Proxy/Server's ADM-LLA, and sets the Target 3017 Address to the Client's LLA. The new Proxy/Server also includes an 3018 OMNI option with Preflen set to the prefix length associated with the 3019 Client's MNP-LLA, with Interface Attributes for its own underlying 3020 interface, and with the OMNI header S/T-omIndex set to 0. The new 3021 Proxy/Server then sets the NA R flag to 1, the S flag to 0 and the O 3022 flag to 1, then encapsulates the message in an OAL header with source 3023 set to its own ADM-ULA and destination set to the ADM-ULA of the old 3024 Proxy/Server and sends the message into the secured spanning tree. 3026 When an old Proxy/Server receives the uNA, it changes the Client's 3027 NCE state to DEPARTED, sets the link-layer address of the Client to 3028 the new Proxy/Server's ADM-ULA, and resets DepartTime. After a short 3029 delay (e.g., 2 seconds) the old Proxy/Server withdraws the Client's 3030 MNP from the routing system. After DepartTime expires, the old 3031 Proxy/Server deletes the Client's NCE. 3033 The old Proxy/Server also iteratively forwards a copy of the uNA 3034 message to each ROS in the Client's Report List by changing the OAL 3035 destination address to the ULA of the ROS while leaving all other 3036 fields of the message unmodified. When the ROS receives the uNA, it 3037 examines the Target address to determine the correct NCE and verifies 3038 that the IPv6 destination address matches the old Proxy/Server. The 3039 ROS then caches the IPv6 source address as the new Proxy/Server for 3040 the existing NCE and marks the entry as STALE. While in the STALE 3041 state, the ROS allows new carrier packets to flow according to any 3042 existing cached link-layer information and sends new NS(AR) messages 3043 using its own ULA as the OAL source and the ADM-ULA of the new Proxy/ 3044 Server as the OAL destination address to elicit NA messages that 3045 reset the NCE state to REACHABLE. If no new NA message is received 3046 for 10 seconds while in the STALE state, the ROS deletes the NCE. 3048 Clients SHOULD NOT move rapidly between Proxy/Servers in order to 3049 avoid causing excessive oscillations in the AERO routing system. 3051 Examples of when a Client might wish to change to a different Proxy/ 3052 Server include a Proxy/Server that has gone unreachable, topological 3053 movements of significant distance, movement to a new geographic 3054 region, movement to a new OMNI link segment, etc. 3056 When a Client moves to a new Proxy/Server, some of the carrier 3057 packets of a multiple fragment OAL packet may have already arrived at 3058 the old Proxy/Server while others are en route to the new Proxy/ 3059 Server, however no special attention in the reassembly algorithm is 3060 necessary since all carrier packets will eventually arrive at the 3061 Client which can then reassemble. 3063 3.17. Multicast 3065 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 3066 [RFC3810] proxy service for its EUNs and/or hosted applications 3067 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 3068 underlying interfaces for which group membership is required. The 3069 IGMP/MLD messages may be further forwarded by a first-hop ANET access 3070 router acting as an IGMP/MLD-snooping switch [RFC4541], then 3071 ultimately delivered to an AERO Proxy/Server acting as a Protocol 3072 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 3073 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 3074 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 3075 INET/EUN networks. The behaviors identified in the following 3076 sections correspond to Source-Specific Multicast (SSM) and Any-Source 3077 Multicast (ASM) operational modes. 3079 3.17.1. Source-Specific Multicast (SSM) 3081 When an ROS "X" acting as PIM router receives a Join/Prune message 3082 from a node on its downstream interfaces containing one or more 3083 ((S)ource, (G)roup) pairs, it updates its Multicast Routing 3084 Information Base (MRIB) accordingly. For each S belonging to a 3085 prefix reachable via X's non-OMNI interfaces, X then forwards the (S, 3086 G) Join/Prune to any PIM routers on those interfaces per [RFC7761]. 3088 For each S belonging to a prefix reachable via X's OMNI interface, X 3089 includes a PIM Join/Prune for each (S,G) in the OMNI option of an 3090 NS(AR) message (see: Section 3.14) using its own LLA as the source 3091 address and ALL-PIM-ROUTERS as the destination address. X then 3092 encapsulates the NS(AR) in an OAL header with source address set to 3093 the ULA of X and destination address set to S then forwards the 3094 message into the secured spanning tree, which delivers it to ROR "Y" 3095 that services S. The resulting NA(AR) will return the LLA for the 3096 prefix that matches S as the network-layer source address and with an 3097 OMNI option with the ULA corresponding to any underlying interfaces 3098 that are currently servicing S. 3100 When Y processes the NS(AR) it examines the PIM Join/Prune message. 3101 If S is located behind any INET, Direct, or VPNed interfaces Y acts 3102 as a PIM router and updates its MRIB to list X as the next hop in the 3103 reverse path. If S is located behind any Proxys "Z"*, Y then sends 3104 an NS(NUD) message containing the PIM message to each Z* with 3105 addressing and encapsulation details the same as specified in 3106 Section 3.15. Each Z* then updates its MRIB accordingly and 3107 maintains the LLA of X as the next hop in the reverse path. Since 3108 the Bridges do not examine network layer control messages, this means 3109 that the (reverse) multicast tree path is simply from each Z* (and/or 3110 Y) to X with no other multicast-aware routers in the path. 3112 Following the initial combined Join/Prune and NS/NA messaging, X 3113 maintains a NCE for each S the same as if X was sending unicast data 3114 traffic to S. In particular, X performs additional NS/NA exchanges 3115 to keep the NCE alive for up to t_periodic seconds [RFC7761]. If no 3116 new Joins are received within t_periodic seconds, X allows the NCE to 3117 expire. Finally, if X receives any additional Join/Prune messages 3118 for (S,G) it forwards the messages in NS/NA exchanges with each Y and 3119 Z* in the NCE over the secured spanning tree. 3121 At some later time, Client C that holds an MNP for source S may 3122 depart from a first Proxy/Server Z1 and/or connect via a new Proxy/ 3123 Server Z2. In that case, Y sends a uNA message to X the same as 3124 specified for unicast mobility in Section 3.16. When X receives the 3125 uNA message, it updates its NCE for the LLA for source S and sends 3126 new Join messages to any new Proxys Z2. There is no requirement to 3127 send any Prune messages to old Proxy/Server Z1 since source S will no 3128 longer source any multicast data traffic via Z1. Instead, the 3129 multicast state for (S,G) in Proxy/Server Z1 will soon time out since 3130 no new Joins will arrive. 3132 After some later time, C may move to a new Proxy/Server Y2 and depart 3133 from old Sever Y1. In that case, Y1 sends Join messages for any of 3134 C's active (S,G) groups to Y2 while including its own LLA as the 3135 source address. This causes Y2 to include Y1 in the multicast 3136 forwarding tree during the interim time that Y1's NCE for C is in the 3137 DEPARTED state. At the same time, Y1 sends a uNA message to X with 3138 an OMNI option with S/T-omIndex in the header set to 0 and a release 3139 indication to cause X to release its NCE for S. X then sends a new 3140 Join message to S via the secured spanning tree and re-initiates 3141 route optimization the same as if it were receiving a fresh Join 3142 message from a node on a downstream link. 3144 3.17.2. Any-Source Multicast (ASM) 3146 When an ROS X acting as a PIM router receives a Join/Prune from a 3147 node on its downstream interfaces containing one or more (*,G) pairs, 3148 it updates its Multicast Routing Information Base (MRIB) accordingly. 3149 X then forwards a copy of the message within the OMNI option of an 3150 NS(AR) message to the Rendezvous Point (RP) R for each G over the 3151 secured spanning tree. X uses its own LLA as the source address and 3152 ALL-PIM-ROUTERS as the destination address, then encapsulates the 3153 NS(AR) message in an OAL header with source address set to the ULA of 3154 X and destination address set to R, then sends the message into the 3155 secured spanning tree. 3157 For each source S that sends multicast traffic to group G via R, the 3158 Proxy/Server Z* for the Client that aggregates S encapsulates the 3159 original IP packets in PIM Register messages and forwards them to R 3160 via the secured spanning tree, which may then elect to send a PIM 3161 Join to Z*. This will result in an (S,G) tree rooted at Z* with R as 3162 the next hop so that R will begin to receive two copies of the 3163 original IP packet; one native copy from the (S, G) tree and a second 3164 copy from the pre-existing (*, G) tree that still uses PIM Register 3165 encapsulation. R can then issue a PIM Register-stop message to 3166 suppress the Register-encapsulated stream. At some later time, if C 3167 moves to a new Proxy/Server Z*, it resumes sending original IP 3168 packets via PIM Register encapsulation via the new Z*. 3170 At the same time, as multicast listeners discover individual S's for 3171 a given G, they can initiate an (S,G) Join for each S under the same 3172 procedures discussed in Section 3.17.1. Once the (S,G) tree is 3173 established, the listeners can send (S, G) Prune messages to R so 3174 that multicast original IP packets for group G sourced by S will only 3175 be delivered via the (S, G) tree and not from the (*, G) tree rooted 3176 at R. All mobility considerations discussed for SSM apply. 3178 3.17.3. Bi-Directional PIM (BIDIR-PIM) 3180 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 3181 approach to ASM that treats the Rendezvous Point (RP) as a Designated 3182 Forwarder (DF). Further considerations for BIDIR-PIM are out of 3183 scope. 3185 3.18. Operation over Multiple OMNI Links 3187 An AERO Client can connect to multiple OMNI links the same as for any 3188 data link service. In that case, the Client maintains a distinct 3189 OMNI interface for each link, e.g., 'omni0' for the first link, 3190 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 3191 would include its own distinct set of Bridges and Proxy/Servers, 3192 thereby providing redundancy in case of failures. 3194 Each OMNI link could utilize the same or different ANET connections. 3195 The links can be distinguished at the link-layer via the SRT prefix 3196 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 3197 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 3198 MSPs on each link. This gives rise to the opportunity for supporting 3199 multiple redundant networked paths, with each VLAN distinguished by a 3200 different SRT "color" (see: Section 3.2.5). 3202 The Client's IP layer can select the outgoing OMNI interface 3203 appropriate for a given traffic profile while (in the reverse 3204 direction) correspondent nodes must have some way of steering their 3205 original IP packets destined to a target via the correct OMNI link. 3207 In a first alternative, if each OMNI link services different MSPs, 3208 then the Client can receive a distinct MNP from each of the links. 3209 IP routing will therefore assure that the correct OMNI link is used 3210 for both outbound and inbound traffic. This can be accomplished 3211 using existing technologies and approaches, and without requiring any 3212 special supporting code in correspondent nodes or Bridges. 3214 In a second alternative, if each OMNI link services the same MSP(s) 3215 then each link could assign a distinct "OMNI link Anycast" address 3216 that is configured by all Bridges on the link. Correspondent nodes 3217 can then perform Segment Routing to select the correct SRT, which 3218 will then direct the original IP packet over multiple hops to the 3219 target. 3221 3.19. DNS Considerations 3223 AERO Client MNs and INET correspondent nodes consult the Domain Name 3224 System (DNS) the same as for any Internetworking node. When 3225 correspondent nodes and Client MNs use different IP protocol versions 3226 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 3227 A records for IPv4 address mappings to MNs which must then be 3228 populated in Relay NAT64 mapping caches. In that way, an IPv4 3229 correspondent node can send original IPv4 packets to the IPv4 address 3230 mapping of the target MN, and the Relay will translate the IPv4 3231 header and destination address into an IPv6 header and IPv6 3232 destination address of the MN. 3234 When an AERO Client registers with an AERO Proxy/Server, the Proxy/ 3235 Server can return the address(es) of DNS servers in RDNSS options 3236 [RFC6106]. The DNS server provides the IP addresses of other MNs and 3237 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 3239 3.20. Transition/Coexistence Considerations 3241 OAL encapsulation ensures that dissimilar INET partitions can be 3242 joined into a single unified OMNI link, even though the partitions 3243 themselves may have differing protocol versions and/or incompatible 3244 addressing plans. However, a commonality can be achieved by 3245 incrementally distributing globally routable (i.e., native) IP 3246 prefixes to eventually reach all nodes (both mobile and fixed) in all 3247 OMNI link segments. This can be accomplished by incrementally 3248 deploying AERO Bridges on each INET partition, with each Bridge 3249 distributing its MNPs and/or discovering non-MNP IP GUA prefixes on 3250 its INET links. 3252 This gives rise to the opportunity to eventually distribute native IP 3253 addresses to all nodes, and to present a unified OMNI link view even 3254 if the INET partitions remain in their current protocol and 3255 addressing plans. In that way, the OMNI link can serve the dual 3256 purpose of providing a mobility/multilink service and a transition/ 3257 coexistence service. Or, if an INET partition is transitioned to a 3258 native IP protocol version and addressing scheme that is compatible 3259 with the OMNI link MNP-based addressing scheme, the partition and 3260 OMNI link can be joined by Bridges. 3262 Relays that connect INETs/EUNs with dissimilar IP protocol versions 3263 may need to employ a network address and protocol translation 3264 function such as NAT64 [RFC6146]. 3266 3.21. Detecting and Reacting to Proxy/Server and Bridge Failures 3268 In environments where rapid failure recovery is required, Proxy/ 3269 Servers and Bridges SHOULD use Bidirectional Forwarding Detection 3270 (BFD) [RFC5880]. Nodes that use BFD can quickly detect and react to 3271 failures so that cached information is re-established through 3272 alternate nodes. BFD control messaging is carried only over well- 3273 connected ground domain networks (i.e., and not low-end radio links) 3274 and can therefore be tuned for rapid response. 3276 Proxy/Servers and Bridges maintain BFD sessions in parallel with 3277 their BGP peerings. If a Proxy/Server or Bridge fails, BGP peers 3278 will quickly re-establish routes through alternate paths the same as 3279 for common BGP deployments. Similarly, Proxys maintain BFD sessions 3280 with their associated Bridges even though they do not establish BGP 3281 peerings with them. 3283 3.22. AERO Clients on the Open Internet 3285 AERO Clients that connect to the open Internet via INET interfaces 3286 can establish a VPN or direct link to securely connect to a Proxy/ 3287 Server in a "tethered" arrangement with all of the Client's traffic 3288 transiting the Proxy/Server which acts as a router. Alternatively, 3289 the Client can associate with an INET Proxy/Server using UDP/IP 3290 encapsulation and control message securing services as discussed in 3291 the following sections. 3293 When a Client's OMNI interface enables an INET underlying interface, 3294 it first determines whether the interface is likely to be behind a 3295 NAT. For IPv4, the Client assumes it is on the open Internet if the 3296 INET address is not a special-use IPv4 address per [RFC3330]. 3297 Similarly for IPv6, the Client assumes it is on the open Internet if 3298 the INET address is a Global Unicast Address (GUA) [RFC4291]. 3299 Otherwise, the Client assumes it may be behind one or several NATs. 3301 The Client then prepares an RS message with IPv6 source address set 3302 to its MNP-LLA, with IPv6 destination set to (link-local) All-Routers 3303 multicast and with an OMNI option with underlying interface 3304 attributes. If the Client believes that it is on the open Internet, 3305 it SHOULD include an L2ADDR in the Interface Attributes sub-option 3306 corresponding to the underlying interface; otherwise, it MAY set 3307 L2ADDR to 0. If the underlying address is IPv4, the Client includes 3308 the Port Number and IPv4 address written in obfuscated form [RFC4380] 3309 as discussed in Section 3.3. If the underlying interface address is 3310 IPv6, the Client instead includes the Port Number and IPv6 address in 3311 obfuscated form. The Client finally includes an authentication 3312 signature sub-option in the OMNI option [I-D.templin-6man-omni] to 3313 provide message authentication, selects an Identification value and 3314 window synchronization parameters, and submits the RS for OAL 3315 encapsulation. The Client then encapsulates the OAL fragment in UDP/ 3316 IP headers to form a carrier packet, sets the UDP/IP source to its 3317 INET address and UDP port, sets the UDP/IP destination to the Proxy/ 3318 Server's INET address and the AERO service port number (8060), then 3319 sends the carrier packet to the Proxy/Server. 3321 When the Proxy/Server receives the RS, it discards the OAL 3322 encapsulation, authenticates the RS message, creates a NCE and 3323 registers the Client's MNP, window synchronization state and INET 3324 interface information according to the OMNI option parameters. If 3325 the RS message OMNI option includes Interface Attributes with an 3326 L2ADDR, the Proxy/Server compares the encapsulation IP address and 3327 UDP port number with the (unobfuscated) values. If the values are 3328 the same, the Proxy/Server caches the Client's information as "INET" 3329 addresses meaning that the Client is likely to accept direct messages 3330 without requiring NAT traversal exchanges. If the values are 3331 different (or, if the OMNI option did not include an L2ADDR) the 3332 Proxy/Server instead caches the Client's information as "mapped" 3333 addresses meaning that NAT traversal exchanges may be necessary. 3335 The Proxy/Server then prepares an RA message with IPv6 source and 3336 destination set corresponding to the addresses in the RS, and with an 3337 OMNI option with an Origin Indication sub-option per 3338 [I-D.templin-6man-omni] with the mapped and obfuscated Port Number 3339 and IP address observed in the encapsulation headers. The Proxy/ 3340 Server also includes an authentication signature sub-option per 3341 [I-D.templin-6man-omni] and a symmetric window synchronization/ 3342 acknowledgement. The Proxy/Server then performs OAL encapsulation 3343 and fragmentation if necessary and encapsulates each fragment in UDP/ 3344 IP headers with addresses set per the L2ADDR information in the NCE 3345 for the Client. 3347 When the Client receives the RA, it authenticates the message then 3348 process the window synchronization/acknowledgement and compares the 3349 mapped Port Number and IP address from the Origin Indication sub- 3350 option with its own address. If the addresses are the same, the 3351 Client assumes the open Internet / Cone NAT principle; if the 3352 addresses are different, the Client instead assumes that further 3353 qualification procedures are necessary to detect the type of NAT and 3354 proceeds according to standard procedures [RFC6081][RFC4380]. The 3355 Client finally arranges to return an explicit/implicit 3356 acknowledgement, and sends periodic RS messages to receive fresh RA 3357 messages before the Router Lifetime received on each INET interface 3358 expires. 3360 When the Client sends messages to target IP addresses, it also 3361 invokes route optimization per Section 3.14. The Client first 3362 creates a NCE for the target in the INCOMPLETE state, then sends the 3363 NS(AR) message to the Proxy/Server with an OMNI option with an 3364 authentication signature sub-option. The Client sets the NS source 3365 address to its own MNP-LLA, destination address to the target 3366 solicited node multicast address and target address to the LLA of the 3367 target. The Client then includes window synchronization parameters 3368 then encapsulates the NS(AR) as an OAL atomic fragment with source 3369 address set to its own MNP-ULA and destination address set to the 3370 target's MNP-ULA. The Client then wraps the atomic fragment in a 3371 UDP/IP header and sends the resulting carrier packet to the Proxy/ 3372 Server. 3374 When the Client's Proxy/Server receives the OAL-encapsulated NS(AR), 3375 it authenticates the message by processing the authentication 3376 signature sub-option and forwards the message over the secured 3377 spanning tree on behalf of the Client. When the ROR receives the 3378 NS(AR), it adds an entry for the MNP-LLA to the target NCE Report 3379 List and caches the receive window information for the ROS Client. 3380 The ROR then returns an NA(AR) with an OMNI option including all of 3381 the target's Interface Attributes. 3383 When the Proxy/Sever for the ROS Client receives the resulting NA(AR) 3384 message via the secured spanning tree, it includes an authentication 3385 signature and forwards the NA(AR) to the ROS Client. The ROS Client 3386 then verifies the authentication signature and caches the NA(AR) 3387 information in the target NCE. Following route optimization, the ROS 3388 Client engages in NS/NA window synchronization and reachability 3389 detection, and forwards packet via the appropriate underlying 3390 interfaces. The ROR may return uNAs with updated information for the 3391 target via the secure spanning tree at any time. The uNAs will be 3392 received by the ROS Proxy/Server which inserts an authentication 3393 signature and forwards the uNA to the ROS Client. 3395 For route optimized targets in the same OMNI link segment, if the 3396 target's L2ADDR is on the open INET, the Client forwards carrier 3397 packets directly to the target INET address. If the target is behind 3398 a NAT, the Client first establishes NAT state for the L2ADDR using 3399 the "direct bubble" and NUD mechanisms discussed in Section 3.10.1. 3400 The Client continues to send carrier packets via its Proxy/Server 3401 until NAT state is populated, then begins forwarding carrier packets 3402 via the direct path through the NAT to the target. For targets in 3403 different OMNI link segments, the Client uses OAL/ORH encapsulation 3404 and forwards carrier packets to the Bridge that returned the NA(AR) 3405 message. 3407 The Client can send original IP packets to route-optimized neighbors 3408 in the same OMNI link segment no larger than the minimum/path MPS in 3409 one piece and with OAL encapsulation as atomic fragments. For larger 3410 original IP packets, the Client applies OAL encapsulation and 3411 fragmentation if necessary according to Section 3.9, with OAL header 3412 with source set to its own MNP-ULA and destination set to the MNP-ULA 3413 of the target. The Client then encapsulates each resulting carrier 3414 packet in UDP/IP *NET headers and sends them to the next hop. 3416 Note: The NAT traversal procedures specified in this document are 3417 applicable for Cone, Address-Restricted and Port-Restricted NATs 3418 only. While future updates to this document may specify procedures 3419 for other NAT variations (e.g., hairpinning and various forms of 3420 Symmetric NATs), it should be noted that continuous communications 3421 are always possible through forwarding via a Proxy/Server even if NAT 3422 traversal is not employed. 3424 3.23. Time-Varying MNPs 3426 In some use cases, it is desirable, beneficial and efficient for the 3427 Client to receive a constant MNP that travels with the Client 3428 wherever it moves. For example, this would allow air traffic 3429 controllers to easily track aircraft, etc. In other cases, however 3430 (e.g., intelligent transportation systems), the MN may be willing to 3431 sacrifice a modicum of efficiency in order to have time-varying MNPs 3432 that can be changed every so often to defeat adversarial tracking. 3434 The DHCPv6 service offers a way for Clients that desire time-varying 3435 MNPs to obtain short-lived prefixes (e.g., on the order of a small 3436 number of minutes). In that case, the identity of the Client would 3437 not be bound to the MNP but rather to a Node Identification value 3438 (see: [I-D.templin-6man-omni]) to be used as the Client ID seed for 3439 MNP prefix delegation. The Client would then be obligated to 3440 renumber its internal networks whenever its MNP (and therefore also 3441 its MNP-LLA) changes. This should not present a challenge for 3442 Clients with automated network renumbering services, however presents 3443 limits for the durations of ongoing sessions that would prefer to use 3444 a constant address. 3446 4. Implementation Status 3448 An early AERO implementation based on OpenVPN (https://openvpn.net/) 3449 was announced on the v6ops mailing list on January 10, 2018 and an 3450 initial public release of the AERO proof-of-concept source code was 3451 announced on the intarea mailing list on August 21, 2015. 3453 AERO Release-3.2 was tagged on March 30, 2021, and is undergoing 3454 internal testing. Additional internal releases expected within the 3455 coming months, with first public release expected end of 1H2021. 3457 5. IANA Considerations 3459 The IANA is instructed to assign a new type value TBD1 in the IPv6 3460 Routing Types registry (IANA registration procedure is IETF Review or 3461 IESG Approval). 3463 The IANA has assigned the UDP port number "8060" for an earlier 3464 experimental first version of AERO [RFC6706]. This document 3465 obsoletes [RFC6706], and together with [I-D.templin-6man-omni] 3466 reclaims the UDP port number "8060" for 'aero' as the service port 3467 for UDP/IP encapsulation. (Note that, although [RFC6706] was not 3468 widely implemented or deployed, any messages coded to that 3469 specification can be easily distinguished and ignored since they use 3470 the invalid ICMPv6 message type number '0'.) This document makes no 3471 request of IANA, since [I-D.templin-6man-omni] already provides 3472 instructions. 3474 No further IANA actions are required. 3476 6. Security Considerations 3478 AERO Bridges configure secured tunnels with AERO Proxy/Servers and 3479 Relays within their local OMNI link segments. Applicable secured 3480 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 3481 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 3482 segments in turn configure secured tunnels for their neighboring AERO 3483 Bridges in a secured spanning tree topology. Therefore, control 3484 messages exchanged between any pair of OMNI link neighbors over the 3485 secured spanning tree are already protected. 3487 AERO nodes acting as Route Optimization Responders (RORs) may also 3488 receive packets directly from arbitrary nodes in INET partitions 3489 instead of via the secured spanning tree. For INET partitions that 3490 apply effective ingress filtering to defeat source address spoofing, 3491 the simple data origin authentication procedures in Section 3.8 can 3492 be applied. 3494 For INET partitions that require strong security in the data plane, 3495 two options for securing communications include 1) disable route 3496 optimization so that all traffic is conveyed over secured tunnels, or 3497 2) enable on-demand secure tunnel creation between Client neighbors. 3498 Option 1) would result in longer routes than necessary and impose 3499 traffic concentration on critical infrastructure elements. Option 2) 3500 could be coordinated between Clients using NS/NA messages with OMNI 3501 Host Identity Protocol (HIP) "Initiator/Responder" message sub- 3502 options [RFC7401][I-D.templin-6man-omni] to create a secured tunnel 3503 on-demand. 3505 AERO Clients that connect to secured ANETs need not apply security to 3506 their ND messages, since the messages will be intercepted by a 3507 perimeter Proxy/Server that applies security on its INET-facing 3508 interface as part of the spanning tree (see above). AERO Clients 3509 connected to the open INET can use network and/or transport layer 3510 security services such as VPNs or can by some other means establish a 3511 direct link to a Proxy/Server. When a VPN or direct link may be 3512 impractical, however, INET Clients and Proxy/Servers SHOULD include 3513 and verify authentication signatures for their IPv6 ND messages as 3514 specified in [I-D.templin-6man-omni]. 3516 Application endpoints SHOULD use transport-layer (or higher-layer) 3517 security services such as TLS/SSL, DTLS or SSH [RFC4251] to assure 3518 the same level of protection as for critical secured Internet 3519 services. AERO Clients that require host-based VPN services SHOULD 3520 use network and/or transport layer security services such as IPsec, 3521 TLS/SSL, DTLS, etc. AERO Proxys and Proxy/Servers can also provide a 3522 network-based VPN service on behalf of the Client, e.g., if the 3523 Client is located within a secured enclave and cannot establish a VPN 3524 on its own behalf. 3526 AERO Proxy/Servers and Bridges present targets for traffic 3527 amplification Denial of Service (DoS) attacks. This concern is no 3528 different than for widely-deployed VPN security gateways in the 3529 Internet, where attackers could send spoofed packets to the gateways 3530 at high data rates. This can be mitigated through the AERO/OMNI data 3531 origin authentication procedures, as well as connecting Proxy/Servers 3532 and Bridges over dedicated links with no connections to the Internet 3533 and/or when connections to the Internet are only permitted through 3534 well-managed firewalls. Traffic amplification DoS attacks can also 3535 target an AERO Client's low data rate links. This is a concern not 3536 only for Clients located on the open Internet but also for Clients in 3537 secured enclaves. AERO Proxy/Servers and Proxys can institute rate 3538 limits that protect Clients from receiving packet floods that could 3539 DoS low data rate links. 3541 AERO Relays must implement ingress filtering to avoid a spoofing 3542 attack in which spurious messages with ULA addresses are injected 3543 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3544 that their connectivity is not used by unauthorized nodes on their 3545 EUNs to gain access to a protected network, i.e., AERO Clients that 3546 act as routers MUST NOT provide routing services for unauthorized 3547 nodes. (This concern is no different than for ordinary hosts that 3548 receive an IP address delegation but then "share" the address with 3549 other nodes via some form of Internet connection sharing such as 3550 tethering.) 3552 The MAP list MUST be well-managed and secured from unauthorized 3553 tampering, even though the list contains only public information. 3554 The MAP list can be conveyed to the Client in a similar fashion as in 3555 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3556 upload of a static file, DNS lookups, etc.). 3558 The AERO service for open INET Clients depends on a public key 3559 distribution service in which Client public keys and identities are 3560 maintained in a shared database accessible to all open INET Proxy/ 3561 Servers. Similarly, each Client must be able to determine the public 3562 key of each Proxy/Server, e.g. by consulting an online database. 3563 When AERO nodes register their public keys indexed by a unique Host 3564 Identity Tag (HIT) [RFC7401] in a distributed database such as the 3565 DNS, and use the HIT as an identity for applying IPv6 ND message 3566 authentication signatures, a means for determining public key 3567 attestation is available. 3569 Security considerations for IPv6 fragmentation and reassembly are 3570 discussed in [I-D.templin-6man-omni]. In environments where spoofing 3571 is considered a threat, OMNI nodes SHOULD employ Identification 3572 window synchronization and OAL destinations SHOULD configure an (end- 3573 system-based) firewall. 3575 SRH authentication facilities are specified in [RFC8754]. Security 3576 considerations for accepting link-layer ICMP messages and reflected 3577 packets are discussed throughout the document. 3579 7. Acknowledgements 3581 Discussions in the IETF, aviation standards communities and private 3582 exchanges helped shape some of the concepts in this work. 3583 Individuals who contributed insights include Mikael Abrahamsson, Mark 3584 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3585 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3586 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3587 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3588 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3589 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3590 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3591 Wood and James Woodyatt. Members of the IESG also provided valuable 3592 input during their review process that greatly improved the document. 3593 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3594 for their shepherding guidance during the publication of the AERO 3595 first edition. 3597 This work has further been encouraged and supported by Boeing 3598 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3599 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3600 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3601 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3602 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3603 Gene MacLean III, Kyle Mikos, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3604 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3605 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3606 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3607 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3608 Benson, Madhuri Madhava Badgandi, Vijayasarathy Rajagopalan, Katie 3609 Tran and Eric Yeh are especially acknowledged for implementing the 3610 AERO functions as extensions to the public domain OpenVPN 3611 distribution. Chuck Klabunde is honored and remembered for his early 3612 leadership, and we mourn his untimely loss. 3614 Earlier works on NBMA tunneling approaches are found in 3615 [RFC2529][RFC5214][RFC5569]. 3617 Many of the constructs presented in this second edition of AERO are 3618 based on the author's earlier works, including: 3620 o The Internet Routing Overlay Network (IRON) 3621 [RFC6179][I-D.templin-ironbis] 3623 o Virtual Enterprise Traversal (VET) 3624 [RFC5558][I-D.templin-intarea-vet] 3626 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3627 [RFC5320][I-D.templin-intarea-seal] 3629 o AERO, First Edition [RFC6706] 3631 Note that these works cite numerous earlier efforts that are not also 3632 cited here due to space limitations. The authors of those earlier 3633 works are acknowledged for their insights. 3635 This work is aligned with the NASA Safe Autonomous Systems Operation 3636 (SASO) program under NASA contract number NNA16BD84C. 3638 This work is aligned with the FAA as per the SE2025 contract number 3639 DTFAWA-15-D-00030. 3641 This work is aligned with the Boeing Commercial Airplanes (BCA) 3642 Internet of Things (IoT) and autonomy programs. 3644 This work is aligned with the Boeing Information Technology (BIT) 3645 MobileNet program. 3647 8. References 3649 8.1. Normative References 3651 [I-D.templin-6man-omni] 3652 Templin, F. L. and T. Whyman, "Transmission of IP Packets 3653 over Overlay Multilink Network (OMNI) Interfaces", draft- 3654 templin-6man-omni-03 (work in progress), April 2021. 3656 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3657 DOI 10.17487/RFC0791, September 1981, 3658 . 3660 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3661 RFC 792, DOI 10.17487/RFC0792, September 1981, 3662 . 3664 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3665 Requirement Levels", BCP 14, RFC 2119, 3666 DOI 10.17487/RFC2119, March 1997, 3667 . 3669 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3670 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3671 December 1998, . 3673 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3674 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3675 DOI 10.17487/RFC3971, March 2005, 3676 . 3678 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3679 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3680 . 3682 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3683 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3684 November 2005, . 3686 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3687 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3688 . 3690 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3691 Network Address Translations (NATs)", RFC 4380, 3692 DOI 10.17487/RFC4380, February 2006, 3693 . 3695 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3696 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3697 DOI 10.17487/RFC4861, September 2007, 3698 . 3700 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3701 Address Autoconfiguration", RFC 4862, 3702 DOI 10.17487/RFC4862, September 2007, 3703 . 3705 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3706 DOI 10.17487/RFC6081, January 2011, 3707 . 3709 [RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T. 3710 Henderson, "Host Identity Protocol Version 2 (HIPv2)", 3711 RFC 7401, DOI 10.17487/RFC7401, April 2015, 3712 . 3714 [RFC7739] Gont, F., "Security Implications of Predictable Fragment 3715 Identification Values", RFC 7739, DOI 10.17487/RFC7739, 3716 February 2016, . 3718 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3719 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3720 May 2017, . 3722 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3723 (IPv6) Specification", STD 86, RFC 8200, 3724 DOI 10.17487/RFC8200, July 2017, 3725 . 3727 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3728 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3729 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3730 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3731 . 3733 8.2. Informative References 3735 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3736 2016. 3738 [I-D.bonica-6man-comp-rtg-hdr] 3739 Bonica, R., Kamite, Y., Alston, A., Henriques, D., and L. 3740 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3741 bonica-6man-comp-rtg-hdr-24 (work in progress), January 3742 2021. 3744 [I-D.bonica-6man-crh-helper-opt] 3745 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3746 Routing Header (CRH) Helper Option", draft-bonica-6man- 3747 crh-helper-opt-03 (work in progress), April 2021. 3749 [I-D.ietf-intarea-frag-fragile] 3750 Bonica, R., Baker, F., Huston, G., Hinden, R. M., Troan, 3751 O., and F. Gont, "IP Fragmentation Considered Fragile", 3752 draft-ietf-intarea-frag-fragile-17 (work in progress), 3753 September 2019. 3755 [I-D.ietf-intarea-tunnels] 3756 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3757 Architecture", draft-ietf-intarea-tunnels-10 (work in 3758 progress), September 2019. 3760 [I-D.ietf-ipwave-vehicular-networking] 3761 (editor), J. (. J., "IPv6 Wireless Access in Vehicular 3762 Environments (IPWAVE): Problem Statement and Use Cases", 3763 draft-ietf-ipwave-vehicular-networking-20 (work in 3764 progress), March 2021. 3766 [I-D.ietf-rtgwg-atn-bgp] 3767 Templin, F. L., Saccone, G., Dawra, G., Lindem, A., and V. 3768 Moreno, "A Simple BGP-based Mobile Routing System for the 3769 Aeronautical Telecommunications Network", draft-ietf- 3770 rtgwg-atn-bgp-10 (work in progress), January 2021. 3772 [I-D.templin-6man-dhcpv6-ndopt] 3773 Templin, F. L., "A Unified Stateful/Stateless 3774 Configuration Service for IPv6", draft-templin-6man- 3775 dhcpv6-ndopt-11 (work in progress), January 2021. 3777 [I-D.templin-intarea-seal] 3778 Templin, F. L., "The Subnetwork Encapsulation and 3779 Adaptation Layer (SEAL)", draft-templin-intarea-seal-68 3780 (work in progress), January 2014. 3782 [I-D.templin-intarea-vet] 3783 Templin, F. L., "Virtual Enterprise Traversal (VET)", 3784 draft-templin-intarea-vet-40 (work in progress), May 2013. 3786 [I-D.templin-ipwave-uam-its] 3787 Templin, F. L., "Urban Air Mobility Implications for 3788 Intelligent Transportation Systems", draft-templin-ipwave- 3789 uam-its-04 (work in progress), January 2021. 3791 [I-D.templin-ironbis] 3792 Templin, F. L., "The Interior Routing Overlay Network 3793 (IRON)", draft-templin-ironbis-16 (work in progress), 3794 March 2014. 3796 [I-D.templin-v6ops-pdhost] 3797 Templin, F. L., "IPv6 Prefix Delegation and Multi- 3798 Addressing Models", draft-templin-v6ops-pdhost-27 (work in 3799 progress), January 2021. 3801 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3803 [RFC1035] Mockapetris, P., "Domain names - implementation and 3804 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3805 November 1987, . 3807 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3808 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3809 . 3811 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3812 DOI 10.17487/RFC2003, October 1996, 3813 . 3815 [RFC2004] Perkins, C., "Minimal Encapsulation within IP", RFC 2004, 3816 DOI 10.17487/RFC2004, October 1996, 3817 . 3819 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3820 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3821 . 3823 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 3824 Networks", RFC 2464, DOI 10.17487/RFC2464, December 1998, 3825 . 3827 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3828 Domains without Explicit Tunnels", RFC 2529, 3829 DOI 10.17487/RFC2529, March 1999, 3830 . 3832 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3833 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3834 . 3836 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3837 of Explicit Congestion Notification (ECN) to IP", 3838 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3839 . 3841 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3842 DOI 10.17487/RFC3330, September 2002, 3843 . 3845 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3846 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3847 DOI 10.17487/RFC3810, June 2004, 3848 . 3850 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 3851 Unique IDentifier (UUID) URN Namespace", RFC 4122, 3852 DOI 10.17487/RFC4122, July 2005, 3853 . 3855 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3856 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3857 January 2006, . 3859 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3860 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3861 DOI 10.17487/RFC4271, January 2006, 3862 . 3864 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3865 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3866 2006, . 3868 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3869 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3870 December 2005, . 3872 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3873 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3874 2006, . 3876 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3877 Control Message Protocol (ICMPv6) for the Internet 3878 Protocol Version 6 (IPv6) Specification", STD 89, 3879 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3880 . 3882 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3883 Protocol (LDAP): The Protocol", RFC 4511, 3884 DOI 10.17487/RFC4511, June 2006, 3885 . 3887 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3888 "Considerations for Internet Group Management Protocol 3889 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3890 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3891 . 3893 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3894 "Internet Group Management Protocol (IGMP) / Multicast 3895 Listener Discovery (MLD)-Based Multicast Forwarding 3896 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3897 August 2006, . 3899 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3900 Algorithms in Cryptographically Generated Addresses 3901 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3902 . 3904 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3905 "Bidirectional Protocol Independent Multicast (BIDIR- 3906 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3907 . 3909 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3910 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3911 DOI 10.17487/RFC5214, March 2008, 3912 . 3914 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3915 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3916 February 2010, . 3918 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3919 Route Optimization Requirements for Operational Use in 3920 Aeronautics and Space Exploration Mobile Networks", 3921 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3922 . 3924 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3925 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3926 . 3928 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3929 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3930 January 2010, . 3932 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3933 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3934 . 3936 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3937 "IPv6 Router Advertisement Options for DNS Configuration", 3938 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3939 . 3941 [RFC6139] Russert, S., Ed., Fleischman, E., Ed., and F. Templin, 3942 Ed., "Routing and Addressing in Networks with Global 3943 Enterprise Recursion (RANGER) Scenarios", RFC 6139, 3944 DOI 10.17487/RFC6139, February 2011, 3945 . 3947 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3948 NAT64: Network Address and Protocol Translation from IPv6 3949 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3950 April 2011, . 3952 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3953 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3954 . 3956 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3957 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3958 DOI 10.17487/RFC6221, May 2011, 3959 . 3961 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3962 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3963 DOI 10.17487/RFC6273, June 2011, 3964 . 3966 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3967 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3968 January 2012, . 3970 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 3971 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, 3972 DOI 10.17487/RFC6355, August 2011, 3973 . 3975 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3976 for Equal Cost Multipath Routing and Link Aggregation in 3977 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3978 . 3980 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3981 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3982 . 3984 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3985 UDP Checksums for Tunneled Packets", RFC 6935, 3986 DOI 10.17487/RFC6935, April 2013, 3987 . 3989 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3990 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3991 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3992 . 3994 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3995 Korhonen, "Requirements for Distributed Mobility 3996 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3997 . 3999 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 4000 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 4001 Multicast - Sparse Mode (PIM-SM): Protocol Specification 4002 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 4003 2016, . 4005 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 4006 Decraene, B., Litkowski, S., and R. Shakir, "Segment 4007 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 4008 July 2018, . 4010 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 4011 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 4012 . 4014 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 4015 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 4016 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 4017 . 4019 [WG] Wireguard, "WireGuard, https://www.wireguard.com", August 4020 2020. 4022 Appendix A. Non-Normative Considerations 4024 AERO can be applied to a multitude of Internetworking scenarios, with 4025 each having its own adaptations. The following considerations are 4026 provided as non-normative guidance: 4028 A.1. Implementation Strategies for Route Optimization 4030 Route optimization as discussed in Section 3.14 results in the route 4031 optimization source (ROS) creating a NCE for the target neighbor. 4032 The NCE state is set to REACHABLE for at most ReachableTime seconds. 4033 In order to refresh the NCE lifetime before the ReachableTime timer 4034 expires, the specification requires implementations to issue a new 4035 NS/NA exchange to reset ReachableTime while data packets are still 4036 flowing. However, the decision of when to initiate a new NS/NA 4037 exchange and to perpetuate the process is left as an implementation 4038 detail. 4040 One possible strategy may be to monitor the NCE watching for data 4041 packets for (ReachableTime - 5) seconds. If any data packets have 4042 been sent to the neighbor within this timeframe, then send an NS to 4043 receive a new NA. If no data packets have been sent, wait for 5 4044 additional seconds and send an immediate NS if any data packets are 4045 sent within this "expiration pending" 5 second window. If no 4046 additional data packets are sent within the 5 second window, reset 4047 the NCE state to STALE. 4049 The monitoring of the neighbor data packet traffic therefore becomes 4050 an ongoing process during the NCE lifetime. If the NCE expires, 4051 future data packets will trigger a new NS/NA exchange while the 4052 packets themselves are delivered over a longer path until route 4053 optimization state is re-established. 4055 A.2. Implicit Mobility Management 4057 OMNI interface neighbors MAY provide a configuration option that 4058 allows them to perform implicit mobility management in which no ND 4059 messaging is used. In that case, the Client only transmits packets 4060 over a single interface at a time, and the neighbor always observes 4061 packets arriving from the Client from the same link-layer source 4062 address. 4064 If the Client's underlying interface address changes (either due to a 4065 readdressing of the original interface or switching to a new 4066 interface) the neighbor immediately updates the NCE for the Client 4067 and begins accepting and sending packets according to the Client's 4068 new address. This implicit mobility method applies to use cases such 4069 as cellphones with both WiFi and Cellular interfaces where only one 4070 of the interfaces is active at a given time, and the Client 4071 automatically switches over to the backup interface if the primary 4072 interface fails. 4074 A.3. Direct Underlying Interfaces 4076 When a Client's OMNI interface is configured over a Direct interface, 4077 the neighbor at the other end of the Direct link can receive packets 4078 without any encapsulation. In that case, the Client sends packets 4079 over the Direct link according to traffic selectors. If the Direct 4080 interface is selected, then the Client's IP packets are transmitted 4081 directly to the peer without going through an ANET/INET. If other 4082 interfaces are selected, then the Client's IP packets are transmitted 4083 via a different interface, which may result in the inclusion of 4084 Proxy/Servers and Bridges in the communications path. Direct 4085 interfaces must be tested periodically for reachability, e.g., via 4086 NUD. 4088 A.4. AERO Critical Infrastructure Considerations 4090 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 4091 IP routers or virtual machines in the cloud. Bridges must be 4092 provisioned, supported and managed by the INET administrative 4093 authority, and connected to the Bridges of other INETs via inter- 4094 domain peerings. Cost for purchasing, configuring and managing 4095 Bridges is nominal even for very large OMNI links. 4097 AERO cloud Proxy/Servers can be standard dedicated server platforms, 4098 but most often will be deployed as virtual machines in the cloud. 4099 The only requirements for cloud Proxy/Servers are that they can run 4100 the AERO user-level code and have at least one network interface 4101 connection to the INET. Cloud Proxy/Servers must be provisioned, 4102 supported and managed by the INET administrative authority. Cost for 4103 purchasing, configuring and managing cloud Proxy/Servers is nominal 4104 especially for virtual machines. 4106 AERO ANET Proxy/Servers are most often standard dedicated server 4107 platforms with one underlying interface connected to the ANET and a 4108 second interface connected to an INET. As with cloud Proxy/Servers, 4109 the only requirements are that they can run the AERO user-level code 4110 and have at least one interface connection to the INET. ANET Proxy/ 4111 Servers must be provisioned, supported and managed by the ANET 4112 administrative authority. Cost for purchasing, configuring and 4113 managing Proxys is nominal, and borne by the ANET administrative 4114 authority. 4116 AERO Relays are simply Proxy/Servers connected to INETs and/or EUNs 4117 that provide forwarding services for non-MNP destinations. The Relay 4118 connects to the OMNI link and engages in eBGP peering with one or 4119 more Bridges as a stub AS. The Relay then injects its MNPs and/or 4120 non-MNP prefixes into the BGP routing system, and provisions the 4121 prefixes to its downstream-attached networks. The Relay can perform 4122 ROS/ROR services the same as for any Proxy/Server, and can route 4123 between the MNP and non-MNP address spaces. 4125 A.5. AERO Server Failure Implications 4127 AERO Proxy/Servers may appear as a single point of failure in the 4128 architecture, but such is not the case since all Proxy/Servers on the 4129 link provide identical services and loss of a Proxy/Server does not 4130 imply immediate and/or comprehensive communication failures. Proxy/ 4131 Server failure is quickly detected and conveyed by Bidirectional 4132 Forward Detection (BFD) and/or proactive NUD allowing Clients to 4133 migrate to new Proxy/Servers. 4135 If a Proxy/Server fails, ongoing packet forwarding to Clients will 4136 continue by virtue of the neighbor cache entries that have already 4137 been established in route optimization sources (ROSs). If a Client 4138 also experiences mobility events at roughly the same time the Proxy/ 4139 Server fails, unsolicited NA messages may be lost but neighbor cache 4140 entries in the DEPARTED state will ensure that packet forwarding to 4141 the Client's new locations will continue for up to DepartTime 4142 seconds. 4144 If a Client is left without a Proxy/Server for a considerable length 4145 of time (e.g., greater than ReachableTime seconds) then existing 4146 neighbor cache entries will eventually expire and both ongoing and 4147 new communications will fail. The original source will continue to 4148 retransmit until the Client has established a new Proxy/Server 4149 relationship, after which time continuous communications will resume. 4151 Therefore, providing many Proxy/Servers on the link with high 4152 availability profiles provides resilience against loss of individual 4153 Proxy/Servers and assurance that Clients can establish new Proxy/ 4154 Server relationships quickly in event of a Proxy/Server failure. 4156 A.6. AERO Client / Server Architecture 4158 The AERO architectural model is client / server in the control plane, 4159 with route optimization in the data plane. The same as for common 4160 Internet services, the AERO Client discovers the addresses of AERO 4161 Proxy/Servers and connects to one or more of them. The AERO service 4162 is analogous to common Internet services such as google.com, 4163 yahoo.com, cnn.com, etc. However, there is only one AERO service for 4164 the link and all Proxy/Servers provide identical services. 4166 Common Internet services provide differing strategies for advertising 4167 server addresses to clients. The strategy is conveyed through the 4168 DNS resource records returned in response to name resolution queries. 4169 As of January 2020 Internet-based 'nslookup' services were used to 4170 determine the following: 4172 o When a client resolves the domainname "google.com", the DNS always 4173 returns one A record (i.e., an IPv4 address) and one AAAA record 4174 (i.e., an IPv6 address). The client receives the same addresses 4175 each time it resolves the domainname via the same DNS resolver, 4176 but may receive different addresses when it resolves the 4177 domainname via different DNS resolvers. But, in each case, 4178 exactly one A and one AAAA record are returned. 4180 o When a client resolves the domainname "ietf.org", the DNS always 4181 returns one A record and one AAAA record with the same addresses 4182 regardless of which DNS resolver is used. 4184 o When a client resolves the domainname "yahoo.com", the DNS always 4185 returns a list of 4 A records and 4 AAAA records. Each time the 4186 client resolves the domainname via the same DNS resolver, the same 4187 list of addresses are returned but in randomized order (i.e., 4188 consistent with a DNS round-robin strategy). But, interestingly, 4189 the same addresses are returned (albeit in randomized order) when 4190 the domainname is resolved via different DNS resolvers. 4192 o When a client resolves the domainname "amazon.com", the DNS always 4193 returns a list of 3 A records and no AAAA records. As with 4194 "yahoo.com", the same three A records are returned from any 4195 worldwide Internet connection point in randomized order. 4197 The above example strategies show differing approaches to Internet 4198 resilience and service distribution offered by major Internet 4199 services. The Google approach exposes only a single IPv4 and a 4200 single IPv6 address to clients. Clients can then select whichever IP 4201 protocol version offers the best response, but will always use the 4202 same IP address according to the current Internet connection point. 4203 This means that the IP address offered by the network must lead to a 4204 highly-available server and/or service distribution point. In other 4205 words, resilience is predicated on high availability within the 4206 network and with no client-initiated failovers expected (i.e., it is 4207 all-or-nothing from the client's perspective). However, Google does 4208 provide for worldwide distributed service distribution by virtue of 4209 the fact that each Internet connection point responds with a 4210 different IPv6 and IPv4 address. The IETF approach is like google 4211 (all-or-nothing from the client's perspective), but provides only a 4212 single IPv4 or IPv6 address on a worldwide basis. This means that 4213 the addresses must be made highly-available at the network level with 4214 no client failover possibility, and if there is any worldwide service 4215 distribution it would need to be conducted by a network element that 4216 is reached via the IP address acting as a service distribution point. 4218 In contrast to the Google and IETF philosophies, Yahoo and Amazon 4219 both provide clients with a (short) list of IP addresses with Yahoo 4220 providing both IP protocol versions and Amazon as IPv4-only. The 4221 order of the list is randomized with each name service query 4222 response, with the effect of round-robin load balancing for service 4223 distribution. With a short list of addresses, there is still 4224 expectation that the network will implement high availability for 4225 each address but in case any single address fails the client can 4226 switch over to using a different address. The balance then becomes 4227 one of function in the network vs function in the end system. 4229 The same implications observed for common highly-available services 4230 in the Internet apply also to the AERO client/server architecture. 4231 When an AERO Client connects to one or more ANETs, it discovers one 4232 or more AERO Proxy/Server addresses through the mechanisms discussed 4233 in earlier sections. Each Proxy/Server address presumably leads to a 4234 fault-tolerant clustering arrangement such as supported by Linux-HA, 4235 Extended Virtual Synchrony or Paxos. Such an arrangement has 4236 precedence in common Internet service deployments in lightweight 4237 virtual machines without requiring expensive hardware deployment. 4238 Similarly, common Internet service deployments set service IP 4239 addresses on service distribution points that may relay requests to 4240 many different servers. 4242 For AERO, the expectation is that a combination of the Google/IETF 4243 and Yahoo/Amazon philosophies would be employed. The AERO Client 4244 connects to different ANET access points and can receive 1-2 Proxy/ 4245 Server ADM-LLAs at each point. It then selects one AERO Proxy/Server 4246 address, and engages in RS/RA exchanges with the same Proxy/Server 4247 from all ANET connections. The Client remains with this Proxy/Server 4248 unless or until the Proxy/Server fails, in which case it can switch 4249 over to an alternate Proxy/Server. The Client can likewise switch 4250 over to a different Proxy/Server at any time if there is some reason 4251 for it to do so. So, the AERO expectation is for a balance of 4252 function in the network and end system, with fault tolerance and 4253 resilience at both levels. 4255 Appendix B. Change Log 4257 << RFC Editor - remove prior to publication >> 4259 Changes from draft-templin-6man-aero-06 to draft-templin-6man-aero- 4260 07: 4262 o Final editorial review pass resulting in multiple changes. 4263 Document now submit for final approval (with reference to rfcdiff 4264 from previous version). 4266 Changes from draft-templin-6man-aero-05 to draft-templin-6man-aero- 4267 06: 4269 o Final editorial review pass resulting in multiple changes. 4270 Document now submit for final approval. 4272 Changes from draft-templin-6man-aero-04 to draft-templin-6man-aero- 4273 05: 4275 o Changed to use traffic selectors instead of the former multilink 4276 selection strategy. 4278 Changes from draft-templin-6man-aero-03 to draft-templin-6man-aero- 4279 04: 4281 o Removed documents from "Obsoletes" list. 4283 o Introduced the concept of "secured" and "unsecured" spanning tree. 4285 o Additional security considerations. 4287 o Additional route optimization considerations. 4289 Changes from draft-templin-6man-aero-02 to draft-templin-6man-aero- 4290 03: 4292 o Support for extended route optimization from ROR to target over 4293 target's underlying interfaces. 4295 Changes from draft-templin-6man-aero-01 to draft-templin-6man-aero- 4296 02: 4298 o Changed reference citations to "draft-templin-6man-omni". 4300 o Several important updates to IPv6 ND cache states and route 4301 optimization message addressing. 4303 o Included introductory description of the "6M's". 4305 o Updated Multicast specification. 4307 Changes from draft-templin-6man-aero-00 to draft-templin-6man-aero- 4308 01: 4310 o Changed category to "Informational". 4312 o Updated implementation status. 4314 Changes from earlier versions to draft-templin-6man-aero-00: 4316 o Established working baseline reference. 4318 Author's Address 4320 Fred L. Templin (editor) 4321 Boeing Research & Technology 4322 P.O. Box 3707 4323 Seattle, WA 98124 4324 USA 4326 Email: fltemplin@acm.org