idnits 2.17.1 draft-templin-6man-aero-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == The 'Obsoletes: ' line in the draft header should list only the _numbers_ of the RFCs which will be obsoleted by this document (if approved); it should not include the word 'RFC' in the list. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (April 16, 2021) is 1104 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 1096, but not defined == Missing Reference: 'ULA0' is mentioned on line 1096, but not defined == Missing Reference: 'ULA1' is mentioned on line 1096, but not defined == Missing Reference: 'ULA2' is mentioned on line 1096, but not defined == Unused Reference: 'RFC3971' is defined on line 3590, but no explicit reference was found in the text == Unused Reference: 'RFC3972' is defined on line 3595, but no explicit reference was found in the text == Unused Reference: 'RFC4193' is defined on line 3603, but no explicit reference was found in the text == Unused Reference: 'RFC7401' is defined on line 3626, but no explicit reference was found in the text == Unused Reference: 'BGP' is defined on line 3652, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-comp-rtg-hdr' is defined on line 3655, but no explicit reference was found in the text == Unused Reference: 'I-D.bonica-6man-crh-helper-opt' is defined on line 3661, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-frag-fragile' is defined on line 3666, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-6man-dhcpv6-ndopt' is defined on line 3689, but no explicit reference was found in the text == Unused Reference: 'RFC2003' is defined on line 3728, but no explicit reference was found in the text == Unused Reference: 'RFC2983' is defined on line 3749, but no explicit reference was found in the text == Unused Reference: 'RFC3168' is defined on line 3753, but no explicit reference was found in the text == Unused Reference: 'RFC4122' is defined on line 3767, but no explicit reference was found in the text == Unused Reference: 'RFC4982' is defined on line 3816, but no explicit reference was found in the text == Unused Reference: 'RFC6139' is defined on line 3858, but no explicit reference was found in the text == Unused Reference: 'RFC6273' is defined on line 3878, but no explicit reference was found in the text == Unused Reference: 'RFC6355' is defined on line 3887, but no explicit reference was found in the text == Unused Reference: 'RFC6438' is defined on line 3892, but no explicit reference was found in the text == Unused Reference: 'RFC6935' is defined on line 3901, but no explicit reference was found in the text == Unused Reference: 'RFC6936' is defined on line 3906, 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-02 == Outdated reference: A later version (-13) exists of draft-ietf-intarea-tunnels-10 == Outdated reference: A later version (-30) exists of draft-ietf-ipwave-vehicular-networking-19 == Outdated reference: A later version (-26) exists of draft-ietf-rtgwg-atn-bgp-10 -- Obsolete informational reference (is this intentional?): RFC 3330 (Obsoleted by RFC 5735) -- Obsolete informational reference (is this intentional?): RFC 6106 (Obsoleted by RFC 8106) -- Obsolete informational reference (is this intentional?): RFC 6347 (Obsoleted by RFC 9147) Summary: 0 errors (**), 0 flaws (~~), 32 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 Obsoletes: rfc5320, rfc5558, rfc5720, April 16, 2021 5 rfc6139, rfc6179, rfc6706 (if 6 approved) 7 Intended status: Informational 8 Expires: October 18, 2021 10 Asymmetric Extended Route Optimization (AERO) 11 draft-templin-6man-aero-02 13 Abstract 15 This document specifies an Asymmetric Extended Route Optimization 16 (AERO) service for IP internetworking over Overlay Multilink Network 17 (OMNI) interfaces. AERO/OMNI use an IPv6 link-local address format 18 that supports operation of the IPv6 Neighbor Discovery (ND) protocol 19 and links ND to IP forwarding. Prefix delegation/registration 20 services are employed for network admission and to manage the routing 21 system. Secure multilink operation, mobility management, multicast, 22 quality of service (QoS) signaling and route optimization are 23 naturally supported through dynamic neighbor cache updates. AERO is 24 a widely-applicable mobile internetworking service especially well- 25 suited to aviation services, intelligent transportation systems, 26 mobile Virtual Private Networks (VPNs) and many other applications. 28 Status of This Memo 30 This Internet-Draft is submitted in full conformance with the 31 provisions of BCP 78 and BCP 79. 33 Internet-Drafts are working documents of the Internet Engineering 34 Task Force (IETF). Note that other groups may also distribute 35 working documents as Internet-Drafts. The list of current Internet- 36 Drafts is at https://datatracker.ietf.org/drafts/current/. 38 Internet-Drafts are draft documents valid for a maximum of six months 39 and may be updated, replaced, or obsoleted by other documents at any 40 time. It is inappropriate to use Internet-Drafts as reference 41 material or to cite them other than as "work in progress." 43 This Internet-Draft will expire on October 18, 2021. 45 Copyright Notice 47 Copyright (c) 2021 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (https://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 63 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6 64 3. Asymmetric Extended Route Optimization (AERO) . . . . . . . . 12 65 3.1. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 12 66 3.2. The AERO Service over OMNI Links . . . . . . . . . . . . 14 67 3.2.1. AERO/OMNI Reference Model . . . . . . . . . . . . . . 14 68 3.2.2. Addressing and Node Identification . . . . . . . . . 16 69 3.2.3. AERO Routing System . . . . . . . . . . . . . . . . . 17 70 3.2.4. OMNI Link Segment Routing . . . . . . . . . . . . . . 19 71 3.2.5. Segment Routing Topologies (SRTs) . . . . . . . . . . 23 72 3.2.6. Segment Routing For OMNI Link Selection . . . . . . . 24 73 3.2.7. Segment Routing Within the OMNI Link . . . . . . . . 24 74 3.3. OMNI Interface Characteristics . . . . . . . . . . . . . 25 75 3.4. OMNI Interface Initialization . . . . . . . . . . . . . . 27 76 3.4.1. AERO Proxy/Server and Relay Behavior . . . . . . . . 27 77 3.4.2. AERO Client Behavior . . . . . . . . . . . . . . . . 28 78 3.4.3. AERO Bridge Behavior . . . . . . . . . . . . . . . . 28 79 3.5. OMNI Interface Neighbor Cache Maintenance . . . . . . . . 28 80 3.5.1. OMNI Neighbor Interface Attributes . . . . . . . . . 30 81 3.5.2. OMNI Neighbor Advertisement Message Flags . . . . . . 30 82 3.6. OMNI Interface Encapsulation and Re-encapsulation . . . . 31 83 3.7. OMNI Interface Decapsulation . . . . . . . . . . . . . . 31 84 3.8. OMNI Interface Data Origin Authentication . . . . . . . . 31 85 3.9. OMNI Interface MTU . . . . . . . . . . . . . . . . . . . 32 86 3.10. OMNI Interface Forwarding Algorithm . . . . . . . . . . . 33 87 3.10.1. Client Forwarding Algorithm . . . . . . . . . . . . 34 88 3.10.2. Proxy/Server and Relay Forwarding Algorithm . . . . 35 89 3.10.3. Bridge Forwarding Algorithm . . . . . . . . . . . . 37 90 3.11. OMNI Interface Error Handling . . . . . . . . . . . . . . 39 91 3.12. AERO Router Discovery, Prefix Delegation and 92 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 41 94 3.12.1. AERO Service Model . . . . . . . . . . . . . . . . . 41 95 3.12.2. AERO Client Behavior . . . . . . . . . . . . . . . . 42 96 3.12.3. AERO Proxy/Server Behavior . . . . . . . . . . . . . 44 97 3.13. The AERO Proxy Function . . . . . . . . . . . . . . . . . 47 98 3.13.1. Detecting and Responding to Proxy/Server Failures . 50 99 3.13.2. Point-to-Multipoint Proxy/Server Coordination . . . 51 100 3.14. AERO Route Optimization . . . . . . . . . . . . . . . . . 52 101 3.14.1. Route Optimization Initiation . . . . . . . . . . . 52 102 3.14.2. Relaying the NS(AR) *NET Packet(s) . . . . . . . . . 53 103 3.14.3. Processing the NS(AR) and Sending the NA(AR) . . . . 54 104 3.14.4. Relaying the NA(AR) . . . . . . . . . . . . . . . . 55 105 3.14.5. Processing the NA(AR) . . . . . . . . . . . . . . . 55 106 3.14.6. Route Optimization Maintenance . . . . . . . . . . . 56 107 3.15. Neighbor Unreachability Detection (NUD) . . . . . . . . . 57 108 3.16. Mobility Management and Quality of Service (QoS) . . . . 59 109 3.16.1. Mobility Update Messaging . . . . . . . . . . . . . 60 110 3.16.2. Announcing Link-Layer Address and/or QoS Preference 111 Changes . . . . . . . . . . . . . . . . . . . . . . 61 112 3.16.3. Bringing New Links Into Service . . . . . . . . . . 61 113 3.16.4. Deactivating Existing Links . . . . . . . . . . . . 61 114 3.16.5. Moving Between Proxy/Servers . . . . . . . . . . . . 62 115 3.17. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 63 116 3.17.1. Source-Specific Multicast (SSM) . . . . . . . . . . 63 117 3.17.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 65 118 3.17.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 66 119 3.18. Operation over Multiple OMNI Links . . . . . . . . . . . 66 120 3.19. DNS Considerations . . . . . . . . . . . . . . . . . . . 66 121 3.20. Transition/Coexistence Considerations . . . . . . . . . . 67 122 3.21. Detecting and Reacting to Proxy/Server and Bridge 123 Failures . . . . . . . . . . . . . . . . . . . . . . . . 67 124 3.22. AERO Clients on the Open Internet . . . . . . . . . . . . 68 125 3.23. Time-Varying MNPs . . . . . . . . . . . . . . . . . . . . 71 126 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 72 127 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 72 128 6. Security Considerations . . . . . . . . . . . . . . . . . . . 72 129 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 74 130 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 76 131 8.1. Normative References . . . . . . . . . . . . . . . . . . 76 132 8.2. Informative References . . . . . . . . . . . . . . . . . 77 133 Appendix A. Non-Normative Considerations . . . . . . . . . . . . 84 134 A.1. Implementation Strategies for Route Optimization . . . . 84 135 A.2. Implicit Mobility Management . . . . . . . . . . . . . . 84 136 A.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 85 137 A.4. AERO Critical Infrastructure Considerations . . . . . . . 85 138 A.5. AERO Server Failure Implications . . . . . . . . . . . . 86 139 A.6. AERO Client / Server Architecture . . . . . . . . . . . . 86 140 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 88 141 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 89 143 1. Introduction 145 Asymmetric Extended Route Optimization (AERO) fulfills the 146 requirements of Distributed Mobility Management (DMM) [RFC7333] and 147 route optimization [RFC5522] for aeronautical networking and other 148 network mobility use cases including intelligent transportation 149 systems and enterprise mobile device users. AERO is a secure 150 internetworking and mobility management service that employs the 151 Overlay Multilink Network Interface (OMNI) [I-D.templin-6man-omni] 152 Non-Broadcast, Multiple Access (NBMA) virtual link model. The OMNI 153 link is a virtual overlay configured over one or more underlying 154 Internetworks, and nodes on the link can exchange original IP packets 155 as single-hop neighbors. The OMNI Adaptation Layer (OAL) supports 156 end system multilink operation for increased reliability, bandwidth 157 optimization and traffic path selection while performing 158 fragmentation and reassembly to support Internetwork segment routing 159 and Maximum Transmission Unit (MTU) diversity. 161 The AERO service comprises Clients, Proxy/Servers and Relays that are 162 seen as OMNI link neighbors as well as Bridges that interconnect 163 diverse Internetworks as OMNI link segments through OAL forwarding at 164 a layer below IP. Each node's OMNI interface uses an IPv6 link-local 165 address format that supports operation of the IPv6 Neighbor Discovery 166 (ND) protocol [RFC4861] and links ND to IP forwarding. A node's OMNI 167 interface can be configured over multiple underlying interfaces, and 168 therefore appears as a single interface with multiple link-layer 169 addresses. Each link-layer address is subject to change due to 170 mobility and/or QoS fluctuations, and link-layer address changes are 171 signaled by ND messaging the same as for any IPv6 link. 173 AERO provides a secure cloud-based service where mobile node Clients 174 may use any Proxy/Server acting as a Mobility Anchor Point (MAP) and 175 fixed nodes may use any Relay on the link for efficient 176 communications. Fixed nodes forward original IP packets destined to 177 other AERO nodes via the nearest Relay, which forwards them through 178 the cloud. A mobile node's initial packets are forwarded through the 179 Proxy/Server, and direct routing is supported through route 180 optimization while packets are flowing. Both unicast and multicast 181 communications are supported, and mobile nodes may efficiently move 182 between locations while maintaining continuous communications with 183 correspondents and without changing their IP Address. 185 AERO Bridges are interconnected in a secured private BGP overlay 186 routing instance to provide an OAL routing/bridging service that 187 joins the underlying Internetworks of multiple disjoint 188 administrative domains into a single unified OMNI link at a layer 189 below IP. Each OMNI link instance is characterized by the set of 190 Mobility Service Prefixes (MSPs) common to all mobile nodes. Relays 191 provide an optimal route from correspondent nodes on the underlying 192 Internetwork to nodes on the OMNI link. To the underlying 193 Internetwork, the Relay is the source of a route to the MSP, and 194 hence uplink traffic to the mobile node is naturally routed to the 195 nearest Relay. 197 AERO can be used with OMNI links that span private-use Internetworks 198 and/or public Internetworks such as the global Internet. In the 199 latter case, some end systems may be located behind global Internet 200 Network Address Translators (NATs). A means for robust traversal of 201 NATs while avoiding "triangle routing" is therefore provided. 203 AERO assumes the use of PIM Sparse Mode in support of multicast 204 communication. In support of Source Specific Multicast (SSM) when a 205 Mobile Node is the source, AERO route optimization ensures that a 206 shortest-path multicast tree is established with provisions for 207 mobility and multilink operation. In all other multicast scenarios 208 there are no AERO dependencies. 210 AERO was designed as a secure aeronautical internetworking service 211 for both manned and unmanned aircraft, where the aircraft is treated 212 as a mobile node that can connect an Internet of Things (IoT). AERO 213 is also applicable to a wide variety of other use cases. For 214 example, it can be used to coordinate the links of mobile nodes 215 (e.g., cellphones, tablets, laptop computers, etc.) that connect into 216 a home enterprise network via public access networks using tunneling 217 software such as OpenVPN [OVPN] with VPN or non-VPN services enabled 218 according to the appropriate security model. AERO can also be used 219 to facilitate terrestrial vehicular and urban air mobility (as well 220 as pedestrian communication services) for future intelligent 221 transportation systems 222 [I-D.ietf-ipwave-vehicular-networking][I-D.templin-ipwave-uam-its]. 223 Other applicable use cases are also in scope. 225 Along with OMNI, AERO provides secured optimal routing support for 226 the "6M's" of modern Internetworking, including: 228 1. Multilink - a mobile node's ability to coordinate multiple 229 diverse underlying data links as a single logical unit (i.e., the 230 OMNI interface) to achieve the required communications 231 performance and reliability objectives. 233 2. Multinet - the ability to span the OMNI link across multiple 234 diverse network administrative segments while maintaining 235 seamless end-to-end communications between mobile nodes and 236 correspondents such as air traffic controllers, fleet 237 administrators, etc. 239 3. Mobility - a mobile node's ability to change network points of 240 attachment (e.g., moving between wireless base stations) which 241 may result in an underlying interface address change, but without 242 disruptions to ongoing communication sessions with peers over the 243 OMNI link. 245 4. Multicast - the ability to send a single network transmission 246 that reaches multiple nodes belonging to the same interest group, 247 but without disturbing other nodes not subscribed to the interest 248 group. 250 5. Multihop - a mobile node vehicle-to-vehicle relaying capability 251 useful when multiple forwarding hops between vehicles may be 252 necessary to "reach back" to an infrastructure access point 253 connection to the OMNI link. 255 6. MTU assurance - the ability to deliver packets of various robust 256 sizes between peers without loss due to a link size restriction, 257 and to dynamically adjust packets sizes to achieve the optimal 258 performance for each independent traffic flow. 260 The following numbered sections present the AERO specification. The 261 appendices at the end of the document are non-normative. 263 2. Terminology 265 The terminology in the normative references applies; especially, the 266 terminology in the OMNI specification [I-D.templin-6man-omni] is used 267 extensively throughout. The following terms are defined within the 268 scope of this document: 270 IPv6 Neighbor Discovery (ND) 271 a control message service for coordinating neighbor relationships 272 between nodes connected to a common link. AERO uses the IPv6 ND 273 messaging service specified in [RFC4861]. 275 IPv6 Prefix Delegation 276 a networking service for delegating IPv6 prefixes to nodes on the 277 link. The nominal service is DHCPv6 [RFC8415], however alternate 278 services (e.g., based on ND messaging) are also in scope. Most 279 notably, a minimal form of prefix delegation known as "prefix 280 registration" can be used if the Client knows its prefix in 281 advance and can represent it in the IPv6 source address of an ND 282 message. 284 Access Network (ANET) 285 a node's first-hop data link service network (e.g., a radio access 286 network, cellular service provider network, corporate enterprise 287 network, etc.) that often provides link-layer security services 288 such as IEEE 802.1X and physical-layer security (e.g., "protected 289 spectrum") to prevent unauthorized access internally and with 290 border network-layer security services such as firewalls and 291 proxys that prevent unauthorized outside access. 293 ANET interface 294 a node's attachment to a link in an ANET. 296 Internetwork (INET) 297 a connected IP network topology with a coherent routing and 298 addressing plan and that provides a transit backbone service for 299 ANET end systems. INETs also provide an underlay service over 300 which the AERO virtual link is configured. Example INETs include 301 corporate enterprise networks, aviation networks, and the public 302 Internet itself. When there is no administrative boundary between 303 an ANET and the INET, the ANET and INET are one and the same. 305 INET interface 306 a node's attachment to a link in an INET. 308 *NET 309 a "wildcard" term referring to either ANET or INET when it is not 310 necessary to draw a distinction between the two. 312 *NET interface 313 a node's attachment to a link in a *NET. 315 *NET Partition 316 frequently, *NETs such as large corporate enterprise networks are 317 sub-divided internally into separate isolated partitions (a 318 technique also known as "network segmentation"). Each partition 319 is fully connected internally but disconnected from other 320 partitions, and there is no requirement that separate partitions 321 maintain consistent Internet Protocol and/or addressing plans. 322 (Each *NET partition is seen as a separate OMNI link segment as 323 discussed below.) 325 *NET address 326 an IP address assigned to a node's interface connection to a *NET. 328 *NET encapsulation 329 the encapsulation of a packet in an outer header or headers that 330 can be routed within the scope of the local *NET partition. 332 OMNI link 333 the same as defined in [I-D.templin-6man-omni], and manifested by 334 IPv6 encapsulation [RFC2473]. The OMNI link spans underlying *NET 335 segments joined by virtual bridges in a spanning tree the same as 336 a bridged campus LAN. AERO nodes on the OMNI link appear as 337 single-hop neighbors even though they may be separated by multiple 338 underlying *NET hops, and can use Segment Routing [RFC8402] to 339 cause packets to visit selected waypoints on the link. 341 OMNI Interface 342 a node's attachment to an OMNI link. Since the addresses assigned 343 to an OMNI interface are managed for uniqueness, OMNI interfaces 344 do not require Duplicate Address Detection (DAD) and therefore set 345 the administrative variable 'DupAddrDetectTransmits' to zero 346 [RFC4862]. 348 OMNI Adaptation Layer (OAL) 349 an OMNI interface process whereby original IP packets admitted 350 into the interface are wrapped in a mid-layer IPv6 header and 351 subject to fragmentation and reassembly. The OAL is also 352 responsible for generating MTU-related control messages as 353 necessary, and for providing addressing context for spanning 354 multiple segments of a bridged OMNI link. 356 original IP packet 357 a whole IP packet or fragment admitted into the OMNI interface by 358 the network layer prior to OAL encapsulation and fragmentation, or 359 an IP packet delivered to the network layer by the OMNI interface 360 following OAL decapsulation and reassembly. 362 OAL packet 363 an original IP packet encapsulated in OAL headers and trailers 364 before OAL fragmentation, or following OAL reassembly. 366 OAL fragment 367 a portion of an OAL packet following fragmentation but prior to 368 *NET encapsulation, or following *NET encapsulation but prior to 369 OAL reassembly. 371 (OAL) atomic fragment 372 an OAL packet that does not require fragmentation is always 373 encapsulated as an "atomic fragment" with a Fragment Header with 374 Fragment Offset and More Fragments both set to 0, but with a valid 375 Identification value. 377 (OAL) carrier packet 378 an encapsulated OAL fragment following *NET encapsulation or prior 379 to *NET decapsulation. OAL sources and destinations exchange 380 carrier packets over underlying interfaces, and may be separated 381 by one or more OAL intermediate nodes. OAL intermediate nodes may 382 perform re-encapsulation on carrier packets by removing the *NET 383 headers of the first hop network and replacing them with new *NET 384 headers for the next hop network. 386 OAL source 387 an OMNI interface acts as an OAL source when it encapsulates 388 original IP packets to form OAL packets, then performs OAL 389 fragmentation and *NET encapsulation to create carrier packets. 391 OAL destination 392 an OMNI interface acts as an OAL destination when it decapsulates 393 carrier packets, then performs OAL reassembly and decapsulation to 394 derive the original IP packet. 396 OAL intermediate node 397 an OMNI interface acts as an OAL intermediate node when it removes 398 the *NET headers of carrier packets received on a first segment, 399 then re-encapsulates the carrier packets in new *NET headers and 400 forwards them into the next segment. OAL intermediate nodes 401 decrement the Hop Limit of the OAL IPv6 header during re- 402 encapsulation, and discard the packet if the Hop Limit reaches 0. 404 underlying interface 405 a *NET interface over which an OMNI interface is configured. 407 Mobility Service Prefix (MSP) 408 an aggregated IP Global Unicast Address (GUA) prefix (e.g., 409 2001:db8::/32, 192.0.2.0/24, etc.) assigned to the OMNI link and 410 from which more-specific Mobile Network Prefixes (MNPs) are 411 delegated. OMNI link administrators typically obtain MSPs from an 412 Internet address registry, however private-use prefixes can 413 alternatively be used subject to certain limitations (see: 414 [I-D.templin-6man-omni]). OMNI links that connect to the global 415 Internet advertise their MSPs to their interdomain routing peers. 417 Mobile Network Prefix (MNP) 418 a longer IP prefix delegated from an MSP (e.g., 419 2001:db8:1000:2000::/56, 192.0.2.8/30, etc.) and delegated to an 420 AERO Client or Relay. 422 Mobile Network Prefix Link Local Address (MNP-LLA) 423 an IPv6 Link Local Address that embeds the most significant 64 424 bits of an MNP in the lower 64 bits of fe80::/64, as specified in 425 [I-D.templin-6man-omni]. 427 Mobile Network Prefix Unique Local Address (MNP-ULA) 428 an IPv6 Unique-Local Address derived from an MNP-LLA. 430 Administrative Link Local Address (ADM-LLA) 431 an IPv6 Link Local Address that embeds a 32-bit administratively- 432 assigned identification value in the lower 32 bits of fe80::/96, 433 as specified in [I-D.templin-6man-omni]. 435 Administrative Unique Local Address (ADM-ULA) 436 an IPv6 Unique-Local Address derived from an ADM-LLA. 438 AERO node 439 a node that is connected to an OMNI link and participates in the 440 AERO internetworking and mobility service. 442 AERO Client ("Client") 443 an AERO node that connects over one or more underlying interfaces 444 and requests MNP delegation/registration service from AERO Proxy/ 445 Servers. The Client assigns an MNP-LLA to the OMNI interface for 446 use in ND exchanges with other AERO nodes and forwards original IP 447 packets to correspondents according to OMNI interface neighbor 448 cache state. 450 AERO Proxy/Server ("Proxy/Server") 451 a dual-function node that provides a proxying service between AERO 452 Clients and external peers on its Client-facing ANET interfaces 453 (i.e., in the same fashion as for an enterprise network proxy) as 454 well as default forwarding and Mobility Anchor Point (MAP) 455 services for coordination with correspondents on its INET-facing 456 interfaces (Proxy/Servers in the open Internetwork instead have a 457 single INET interface). The Proxy/Server configures an OMNI 458 interface and assigns an ADM-LLA to support the operation of IPv6 459 ND services, while advertising all of its associated MNPs via BGP 460 peerings with Bridges. Note that the Proxy and Server functions 461 can be considered logically separable, but since each Proxy/Server 462 must be informed of all of the Client's other multilink Proxy/ 463 Server affiliations the AERO service is best supported when the 464 two functions are coresident on the same physical or logical 465 platform. 467 AERO Relay ("Relay") 468 a Proxy/Server that provides forwarding services between nodes 469 reached via the OMNI link and correspondents on connected 470 downstream links. AERO Relays configure an OMNI interface and 471 assign an ADM-LLA the same as Proxy/Servers. AERO Relays also run 472 a dynamic routing protocol to discover any non-MNP IP GUA routes 473 in service on its connected downstream network links. In both 474 cases, the Relay advertises the MSP(s) to its downstream networks, 475 and distributes all of its associated non-MNP IP GUA routes via 476 BGP peerings with Bridges (i.e., the same as for Proxy/Servers). 478 AERO Bridge ("Bridge") 479 a node that provides hybrid routing/bridging services (as well as 480 a security trust anchor) for nodes on an OMNI link. The Bridge 481 forwards carrier packets between OMNI link segments as OAL 482 intermediate nodes while decrementing the OAL IPv6 header Hop 483 Limit but without decrementing the network layer IP TTL/Hop Limit. 484 AERO Bridges peer with Proxy/Servers and other Bridges over 485 secured tunnels to discover the full set of MNPs for the link as 486 well as any non-MNP IP GUA routes that are reachable via Relays. 488 link-layer address 489 an IP address used as an encapsulation header source or 490 destination address from the perspective of the OMNI interface. 491 When an upper layer protocol (e.g., UDP) is used as part of the 492 encapsulation, the port number is also considered as part of the 493 link-layer address. 495 network layer address 496 the source or destination address of an original IP packet 497 presented to the OMNI interface. 499 end user network (EUN) 500 an internal virtual or external edge IP network that an AERO 501 Client or Relay connects to the rest of the network via the OMNI 502 interface. The Client/Relay sees each EUN as a "downstream" 503 network, and sees the OMNI interface as the point of attachment to 504 the "upstream" network. 506 Mobile Node (MN) 507 an AERO Client and all of its downstream-attached networks that 508 move together as a single unit, i.e., an end system that connects 509 an Internet of Things. 511 Mobile Router (MR) 512 a MN's on-board router that forwards original IP packets between 513 any downstream-attached networks and the OMNI link. The MR is the 514 MN entity that hosts the AERO Client. 516 Route Optimization Source (ROS) 517 the AERO node nearest the source that initiates route 518 optimization. The ROS may be a Proxy/Server or Relay acting on 519 behalf of the source, or may be the source Client itself. 521 Route Optimization responder (ROR) 522 the AERO node nearest the target destination that responds to 523 route optimization requests. The ROR may be a Proxy/Server acting 524 on behalf of a target MNP Client, a Relay for a non-MNP 525 destination or may be the target Client itself. 527 MAP List 528 a geographically and/or topologically referenced list of addresses 529 of all Proxy/Servers within the same OMNI link. There is a single 530 MAP list for the entire OMNI link. 532 Distributed Mobility Management (DMM) 533 a BGP-based overlay routing service coordinated by Proxy/Servers 534 and Bridges that tracks all Proxy/Server-to-Client associations. 536 Mobility Service (MS) 537 the collective set of all Proxy/Servers, Bridges and Relays that 538 provide the AERO Service to Clients. 540 Mobility Service Endpoint MSE) 541 an individual Proxy/Server, Bridge or Relay in the Mobility 542 Service. 544 Throughout the document, the simple terms "Client", "Proxy/Server", 545 "Bridge" and "Relay" refer to "AERO Client", "AERO Proxy/Server", 546 "AERO Bridge" and "AERO Relay", respectively. Capitalization is used 547 to distinguish these terms from other common Internetworking uses in 548 which they appear without capitalization. 550 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 551 the names of node variables, messages and protocol constants) is used 552 throughout this document. The terms "All-Routers multicast", "All- 553 Nodes multicast", "Solicited-Node multicast" and "Subnet-Router 554 anycast" are defined in [RFC4291]. Also, the term "IP" is used to 555 generically refer to either Internet Protocol version, i.e., IPv4 556 [RFC0791] or IPv6 [RFC8200]. 558 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 559 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 560 "OPTIONAL" in this document are to be interpreted as described in BCP 561 14 [RFC2119][RFC8174] when, and only when, they appear in all 562 capitals, as shown here. 564 3. Asymmetric Extended Route Optimization (AERO) 566 The following sections specify the operation of IP over OMNI links 567 using the AERO service: 569 3.1. AERO Node Types 571 AERO Clients are Mobile Nodes (MNs) that configure OMNI interfaces 572 over underlying interfaces with addresses that may change when the 573 Client moves to a new network connection point. AERO Clients 574 register their Mobile Network Prefixes (MNPs) with the AERO service, 575 and distribute the MNPs to nodes on EUNs. AERO Bridges, Proxy/ 576 Servers and Relays are critical infrastructure elements in fixed 577 (i.e., non-mobile) INET deployments and hence have permanent and 578 unchanging INET addresses. Together, they constitute the AERO 579 service which provides an OMNI link virtual overlay for connecting 580 AERO Clients. 582 AERO Bridges provide hybrid routing/bridging services (as well as a 583 security trust anchor) for nodes on an OMNI link. Bridges use 584 standard IPv6 routing to forward carrier packets both within the same 585 *NET partition and between disjoint *NET partitions based on an IPv6 586 encapsulation mid-layer known as the OMNI Adaptation Layer (OAL) 587 [I-D.templin-6man-omni]. During forwarding, the inner IP layer 588 experiences a virtual bridging service since the inner IP TTL/Hop 589 Limit is not decremented. Each Bridge also peers with Proxy/Servers 590 and other Bridges in a dynamic routing protocol instance to provide a 591 Distributed Mobility Management (DMM) service for the list of active 592 MNPs (see Section 3.2.3). Bridges present the OMNI link as a set of 593 one or more Mobility Service Prefixes (MSPs) and configure secured 594 tunnels with Proxy/Servers, Relays and other Bridges; they further 595 maintain IP forwarding table entries for each MNP and any other 596 reachable non-MNP prefixes. 598 AERO Proxy/Servers in distributed *NET locations provide default 599 forwarding and mobility/multilink services for AERO Client Mobile 600 Nodes (MNs). Each Proxy/Server also peers with Bridges in a dynamic 601 routing protocol instance to advertise its list of associated MNPs 602 (see Section 3.2.3). Proxy/Servers facilitate prefix delegation/ 603 registration exchanges with Clients, where each delegated prefix 604 becomes an MNP taken from an MSP. Proxy/Servers forward carrier 605 packets between OMNI interface neighbors and track each Client's 606 mobility profiles. Proxy/Servers at ANET/INET boundaries provide a 607 conduit for ANET Clients to associate with peers reached through 608 external INETs. Proxy/Servers in the open INET support INET Clients 609 through authenticated IPv6 ND message exchanges. 611 AERO Relays are Proxy/Servers that provide forwarding services to 612 exchange original IP packets between the OMNI interface and INET/EUN 613 interfaces. Relays are provisioned with MNPs the same as for an AERO 614 Client, and also run a dynamic routing protocol to discover any non- 615 MNP IP routes. The Relay advertises the MSP(s) to its connected 616 networks, and distributes all of its associated MNP and non-MNP 617 routes via BGP peerings with Bridges 619 3.2. The AERO Service over OMNI Links 621 3.2.1. AERO/OMNI Reference Model 623 Figure 1 presents the basic OMNI link reference model: 625 +----------------+ 626 | AERO Bridge B1 | 627 | Nbr: S1, S2, P1| 628 |(X1->S1; X2->S2)| 629 | MSP M1 | 630 +-+------------+-+ 631 +--------------+ | Secured | +--------------+ 632 | AERO P/S S1 | | tunnels | | AERO P/S S2 | 633 | Nbr: C1, B1 +-----+ +-----+ Nbr: C2, B1 | 634 | default->B1 | | default->B1 | 635 | X1->C1 | | X2->C2 | 636 +-------+------+ +------+-------+ 637 | OMNI link | 638 X===+===+======================================+===+===X 639 | | 640 +-----+--------+ +--------+-----+ 641 |AERO Client C1| |AERO Client C2| 642 | Nbr: S1 | | Nbr: S2 | 643 | default->S1 | | default->S2 | 644 | MNP X1 | | MNP X2 | 645 +------+-------+ +-----+--------+ 646 | | 647 .-. .-. 648 ,-( _)-. ,-( _)-. 649 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 650 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 651 `-(______)-' +-------+ +-------+ `-(______)-' 653 Figure 1: AERO/OMNI Reference Model 655 In this model: 657 o the OMNI link is an overlay network service configured over one or 658 more underlying *NET partitions which may be managed by different 659 administrative authorities and have incompatible protocols and/or 660 addressing plans. 662 o AERO Bridge B1 aggregates Mobility Service Prefix (MSP) M1, 663 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 664 via BGP peerings over secured tunnels to Proxy/Servers (S1, S2). 665 Bridges connect the disjoint segments of a partitioned OMNI link. 667 o AERO Proxy/Servers S1 and S2 configure secured tunnels with Bridge 668 B1 and also provide mobility, multilink, multicast and default 669 router services for the MNPs of their associated Clients C1 and 670 C2. (AERO Proxy/Servers that act as Relays can also advertise 671 non-MNP routes for non-mobile correspondent nodes the same as for 672 MNP Clients.) 674 o AERO Clients C1 and C2 associate with Proxy/Servers S1 and S2, 675 respectively. They receive MNP delegations X1 and X2, and also 676 act as default routers for their associated physical or internal 677 virtual EUNs. Simple hosts H1 and H2 attach to the EUNs served by 678 Clients C1 and C2, respectively. 680 An OMNI link configured over a single *NET appears as a single 681 unified link with a consistent underlying network addressing plan. 682 In that case, all nodes on the link can exchange carrier packets via 683 simple *NET encapsulation, since the underlying *NET is connected. 684 In common practice, however, an OMNI link may be partitioned into 685 multiple "segments", where each segment is a distinct *NET 686 potentially managed under a different administrative authority (e.g., 687 as for worldwide aviation service providers such as ARINC, SITA, 688 Inmarsat, etc.). Individual *NETs may also themselves be partitioned 689 internally, in which case each internal partition is seen as a 690 separate segment. 692 The addressing plan of each segment is consistent internally but will 693 often bear no relation to the addressing plans of other segments. 694 Each segment is also likely to be separated from others by network 695 security devices (e.g., firewalls, proxys, packet filtering gateways, 696 etc.), and in many cases disjoint segments may not even have any 697 common physical link connections. Therefore, nodes can only be 698 assured of exchanging carrier packets directly with correspondents in 699 the same segment, and not with those in other segments. The only 700 means for joining the segments therefore is through inter-domain 701 peerings between AERO Bridges. 703 The same as for traditional campus LANs, multiple OMNI link segments 704 can be joined into a single unified link via a virtual bridging 705 service using the OMNI Adaptation Layer (OAL) [I-D.templin-6man-omni] 706 which inserts a mid-layer IPv6 encapsulation header that supports 707 inter-segment forwarding (i.e., bridging) without decrementing the 708 network-layer TTL/Hop Limit. This bridging of OMNI link segments is 709 shown in Figure 2: 711 . . . . . . . . . . . . . . . . . . . . . . . 712 . . 713 . .-(::::::::) . 714 . .-(::::::::::::)-. +-+ . 715 . (:::: Segment A :::)--|B|---+ . 716 . `-(::::::::::::)-' +-+ | . 717 . `-(::::::)-' | . 718 . | . 719 . .-(::::::::) | . 720 . .-(::::::::::::)-. +-+ | . 721 . (:::: Segment B :::)--|B|---+ . 722 . `-(::::::::::::)-' +-+ | . 723 . `-(::::::)-' | . 724 . | . 725 . .-(::::::::) | . 726 . .-(::::::::::::)-. +-+ | . 727 . (:::: Segment C :::)--|B|---+ . 728 . `-(::::::::::::)-' +-+ | . 729 . `-(::::::)-' | . 730 . | . 731 . ..(etc).. x . 732 . . 733 . . 734 . <- OMNI link Bridged by encapsulation -> . 735 . . . . . . . . . . . . . .. . . . . . . . . 737 Figure 2: Bridging OMNI Link Segments 739 Bridges, Proxy/Servers and Relays connect via secured INET tunnels 740 over their respective segments in a spanning tree topology rooted at 741 the Bridges. The secured spanning tree supports strong 742 authentication for IPv6 ND control messages and may also be used to 743 convey the initial carrier packets in a flow. Route optimization can 744 then be employed to cause carrier packets to take more direct paths 745 between OMNI link neighbors without having to strictly follow the 746 spanning tree. 748 3.2.2. Addressing and Node Identification 750 AERO nodes on OMNI links use the Link-Local Address (LLA) prefix 751 fe80::/64 [RFC4291] to assign LLAs used for network-layer addresses 752 in link-scoped IPv6 ND and data messages. AERO Clients use LLAs 753 constructed from MNPs (i.e., "MNP-LLAs") while other AERO nodes use 754 LLAs constructed from administrative identification values ("ADM- 755 LLAs") as specified in [I-D.templin-6man-omni]. Non-MNP routes are 756 also represented the same as for MNP-LLAs, but may include a prefix 757 that is not properly covered by the MSP. 759 AERO nodes also use the Unique Local Address (ULA) prefix fd00::/8 760 followed by a pseudo-random 40-bit OMNI domain identifier to form the 761 prefix [ULA]::/48, then include a 16-bit OMNI link identifier '*' to 762 form the prefix [ULA*]::/64 [RFC4291]. The AERO node then uses the 763 prefix [ULA*]::/64 to form "MNP-ULAs" or "ADM-ULA"s as specified in 764 [I-D.templin-6man-omni] to support OAL addressing. (The prefix 765 [ULA*]::/64 appearing alone and with no suffix represents "default".) 766 AERO Clients also use Temporary ULAs constructed per 767 [I-D.templin-6man-omni], where the addresses are typically used only 768 in initial control message exchanges until a stable MNP-LLA/ULA is 769 assigned. 771 AERO MSPs, MNPs and non-MNP routes are typically based on Global 772 Unicast Addresses (GUAs), but in some cases may be based on private- 773 use addresses. See [I-D.templin-6man-omni] for a full specification 774 of LLAs, ULAs and GUAs used by AERO nodes on OMNI links. 776 Finally, AERO Clients and Proxy/Servers configure node identification 777 values as specified in [I-D.templin-6man-omni]. 779 3.2.3. AERO Routing System 781 The AERO routing system comprises a private instance of the Border 782 Gateway Protocol (BGP) [RFC4271] that is coordinated between Bridges 783 and Proxy/Servers and does not interact with either the public 784 Internet BGP routing system or any underlying INET routing systems. 786 In a reference deployment, each Proxy/Server is configured as an 787 Autonomous System Border Router (ASBR) for a stub Autonomous System 788 (AS) using a 32-bit AS Number (ASN) [RFC4271] that is unique within 789 the BGP instance, and each Proxy/Server further uses eBGP to peer 790 with one or more Bridges but does not peer with other Proxy/Servers. 791 Each *NET of a multi-segment OMNI link must include one or more 792 Bridges, which peer with the Proxy/Servers within that *NET. All 793 Bridges within the same *NET are members of the same hub AS, and use 794 iBGP to maintain a consistent view of all active routes currently in 795 service. The Bridges of different *NETs peer with one another using 796 eBGP. 798 Bridges maintain forwarding table entries only for the MNP-ULAs 799 corresponding to MNP and non-MNP routes that are currently active, 800 and carrier packets destined to all other MNP-ULAs will correctly 801 incur Destination Unreachable messages due to the black-hole route. 802 In this way, Proxy/Servers and Relays have only partial topology 803 knowledge (i.e., they know only about the routes their directly 804 associated Clients and non-AERO links) and they forward all other 805 carrier packets to Bridges which have full topology knowledge. 807 Each OMNI link segment assigns a unique ADM-ULA sub-prefix of 808 [ULA*]::/96. For example, a first segment could assign 809 [ULA*]::1000/116, a second could assign [ULA*]::2000/116, a third 810 could assign [ULA*]::3000/116, etc. Within each segment, each Proxy/ 811 Server configures an ADM-ULA within the segment's prefix, e.g., the 812 Proxy/Servers within [ULA*]::2000/116 could assign the ADM-ULAs 813 [ULA*]::2011/116, [ULA*]::2026/116, [ULA*]::2003/116, etc. 815 The administrative authorities for each segment must therefore 816 coordinate to assure mutually-exclusive ADM-ULA prefix assignments, 817 but internal provisioning of ADM-ULAs an independent local 818 consideration for each administrative authority. For each ADM-ULA 819 prefix, the Bridge(s) that connect that segment assign the all-zero's 820 address of the prefix as a Subnet Router Anycast address. For 821 example, the Subnet Router Anycast address for [ULA*]::1023/116 is 822 simply [ULA*]::1000. 824 ADM-ULA prefixes are statically represented in Bridge forwarding 825 tables. Bridges join multiple segments into a unified OMNI link over 826 multiple diverse administrative domains. They support a bridging 827 function by first establishing forwarding table entries for their 828 ADM-ULA prefixes either via standard BGP routing or static routes. 829 For example, if three Bridges ('A', 'B' and 'C') from different 830 segments serviced [ULA*]::1000/116, [ULA*]::2000/116 and 831 [ULA*]::3000/116 respectively, then the forwarding tables in each 832 Bridge are as follows: 834 A: [ULA*]::1000/116->local, [ULA*]::2000/116->B, [ULA*]::3000/116->C 836 B: [ULA*]::1000/116->A, [ULA*]::2000/116->local, [ULA*]::3000/116->C 838 C: [ULA*]::1000/116->A, [ULA*]::2000/116->B, [ULA*]::3000/116->local 840 These forwarding table entries are permanent and never change, since 841 they correspond to fixed infrastructure elements in their respective 842 segments. 844 MNP ULAs are instead dynamically advertised in the AERO routing 845 system by Proxy/Servers and Relays that provide service for their 846 corresponding MNPs. For example, if three Proxy/Servers ('D', 'E' 847 and 'F') service the MNPs 2001:db8:1000:2000::/56, 848 2001:db8:3000:4000::/56 and 2001:db8:5000:6000::/56 then the routing 849 system would include: 851 D: [ULA*]:2001:db8:1000:2000/120 853 E: [ULA*]:2001:db8:3000:4000/120 854 F: [ULA*]:2001:db8:5000:6000/120 856 A full discussion of the BGP-based routing system used by AERO is 857 found in [I-D.ietf-rtgwg-atn-bgp]. 859 3.2.4. OMNI Link Segment Routing 861 With the Client and partition prefixes in place in Bridge forwarding 862 tables, the OMNI interface sends control and data messages toward 863 AERO destination nodes located in different OMNI link segments over 864 the spanning tree. The OMNI interface uses the OMNI Adaptation Layer 865 (OAL) encapsulation service [I-D.templin-6man-omni], and includes an 866 OMNI Routing Header (ORH) as an extension to the OAL header if final 867 segment forwarding information is available, e.g., in the neighbor 868 cache. (For nodes located in the same OMNI link segment, or when no 869 final segment forwarding information is available, the ORH may be 870 omitted.) The ORH is formatted as shown in Figure 3: 872 0 1 2 3 873 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 1 874 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 875 | Next Header | Hdr Ext Len | Routing Type | SRT | FMT | 876 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 877 | Last Hop Segment-id (LHS) | 878 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 879 ~ Link Layer Address (L2ADDR) ~ 880 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 881 ~ Destination Suffix (if necessary) ~ 882 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 883 ~ Null Padding (if necessary) ~ 884 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 886 Figure 3: OMNI Routing Header (ORH) Format 888 In this format: 890 o Next Header identifies the type of header immediately following 891 the ORH. 893 o Hdr Ext Len is the length of the Routing header in 8-octet units 894 (not including the first 8 octets), with trailing padding added if 895 necessary to produce an integral number of 8-octet units. 897 o Routing Type is set to TBD1 (see IANA Considerations). 899 o Segments Left is omitted, and replaced by a 5-bit SRT and 3-bit 900 FMT field. 902 o SRT - a 5-bit Segment Routing Topology prefix length value that 903 (when added to 96) determines the prefix length to apply to the 904 ADM-ULA formed from concatenating [ULA*]::/96 with the 32 bit LHS 905 value that follows (for example, the value 16 corresponds to the 906 prefix length 112). 908 o FMT - a 3-bit "Framework/Mode/Type" code corresponding to the 909 included Link Layer Address as follows: 911 * When the most significant bit (i.e., "Framework") is set to 1, 912 L2ADDR is the *NET encapsulation address for the target Client 913 itself; otherwise L2ADDR is the address of the Proxy/Server 914 named in the LHS. 916 * When the next most significant bit (i.e., "Mode") is set to 1, 917 the Framework node is (likely) located behind a *NET Network 918 Address Translator (NAT); otherwise, it is on the open *NET. 920 * When the least significant bit (i.e., "Type") is set to 0, 921 L2ADDR includes a UDP Port Number followed by an IPv4 address; 922 otherwise, it includes a UDP Port Number followed by an IPv6 923 address. 925 o LHS - the 32 bit ID of a node in the Last Hop Segment that 926 services the target. When SRT and LHS are both set to 0, the LHS 927 is considered unspecified. When SRT is set to 0 and LHS is non- 928 zero, the prefix length is set to 128. SRT and LHS provide 929 guidance to the OMNI interface forwarding algorithm. 930 Specifically, if SRT/LHS is located in the local OMNI link 931 segment, the OAL source can omit the ORH and (following any 932 necessary NAT traversal messaging) send directly to the OAL 933 destination according to FMT/L2ADDR. Otherwise, it includes the 934 ORH and forwards according to the OMNI link spanning tree. 936 o Link Layer Address (L2ADDR) - Formatted according to FMT, and 937 identifies the link-layer address (i.e., the encapsulation 938 address) of the target. The UDP Port Number appears in the first 939 two octets and the IP address appears in the next 4 octets for 940 IPv4 or 16 octets for IPv6. The Port Number and IP address are 941 recorded in network byte order, and in ones-compliment 942 "obfuscated" form per [RFC4380]. The OMNI interface forwarding 943 algorithm uses FMT/L2ADDR to determine the *NET encapsulation 944 address for local forwarding when SRT/LHS is located in the same 945 OMNI link segment. Note that if the target is behind a NAT, 946 L2ADDR will contain the mapped *NET address stored in the NAT; 947 otherwise, L2ADDR will contain the native *NET information of the 948 target itself. 950 o Destination Suffix is a 64-bit field included only for OAL non- 951 first-fragments. Present only when Hdr Ext Len indicates that at 952 least 8 bytes follow L2ADDR. When present, encodes the 64-bit 953 MNP-ULA suffix for the target Client. 955 o Null Padding contains zero-valued octets as necessary to pad the 956 ORH to an integral number of 8-octet units. 958 AERO neighbors use OAL encapsulation and fragmentation to exchange 959 OAL packets as specified in [I-D.templin-6man-omni]. When an AERO 960 node's OMNI interface (acting as an OAL source) uses OAL 961 encapsulation for an original IP packet with source address 962 2001:db8:1:2::1 and destination address 2001:db8:1234:5678::1, it 963 sets the OAL header source address to its own ULA (e.g., 964 [ULA*]::2001:db8:1:2), sets the destination address to the MNP-ULA 965 corresponding to the IP destination address (e.g., 966 [ULA*]::2001:db8:1234:5678), sets the Traffic Class, Flow Label, Hop 967 Limit and Payload Length as discussed in [I-D.templin-6man-omni], 968 then finally selects an Identification and appends an OAL checksum. 970 If the neighbor cache information indicates that the target is in a 971 different segment, the OAL source next inserts an ORH immediately 972 following the OAL header while including the correct SRT, FMT, LHS, 973 L2ADDR and Destination Suffix if fragmentation if needed (in this 974 case, the Destination Suffix is 2001:db8:1234:5678). Next, the OAL 975 source overwrites the OAL header destination address with the LHS 976 Subnet Router Anycast address (for example, for LHS 3000:4567 with 977 SRT 16, the Subnet Router Anycast address is [ULA*]::3000:0000). 978 (Note: if the ADM-ULA of the last-hop Proxy/Server is known but the 979 SRT, FMT, LHS and L2ADDR are not (yet) known, the OAL source instead 980 overwrites the OAL header destination address with the ADM-ULA.) 982 The OAL source then fragments the OAL packet, with each resulting OAL 983 fragment including the OAL/ORH headers while only the first fragment 984 includes the original IPv6 header. (Note that if no actual 985 fragmentation is required the OAL packet is still prepared as an 986 "atomic" fragment that includes a Fragment Header with Offset and 987 More Fragments both set to 0.) The OAL source finally encapsulates 988 each resulting OAL fragment in an *NET header to form an OAL carrier 989 packet, with source address set to its own *NET address (e.g., 990 192.0.2.100) and destination set to the *NET address of a Bridge 991 (e.g., 192.0.2.1). 993 The carrier packet encapsulation format in the above example is shown 994 in Figure 4: 996 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 997 | *NET Header | 998 | src = 192.0.2.100 | 999 | dst = 192.0.2.1 | 1000 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1001 | OAL IPv6 Header | 1002 | src = [ULA*]::2001:db8:1:2 | 1003 | dst= [ULA*]::3000:0000 | 1004 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1005 | ORH (if necessary) | 1006 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1007 | OAL Fragment Header | 1008 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1009 | Original IP Header | 1010 | (first-fragment only) | 1011 | src = 2001:db8:1:2::1 | 1012 | dst = 2001:db8:1234:5678::1 | 1013 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1014 | | 1015 ~ ~ 1016 ~ Original Packet Body/Fragment ~ 1017 ~ ~ 1018 | | 1019 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1021 Figure 4: Carrier Packet Format 1023 In this format, the original IP header and packet body/fragment are 1024 from the original IP packet, the OAL header is an IPv6 header 1025 prepared according to [RFC2473], the ORH is a Routing Header 1026 extension of the OAL header, the Fragment Header identifies each 1027 fragment, and the INET header is prepared as discussed in 1028 Section 3.6. When the OAL source transmits the resulting carrier 1029 packets, they are forwarded over possibly multiple OAL intermediate 1030 nodes in the OMNI link spanning tree until they arrive at the OAL 1031 destination. 1033 This gives rise to a routing system that contains both Client MNP-ULA 1034 routes that may change dynamically due to regional node mobility and 1035 per-partition ADM-ULA routes that rarely if ever change. The Bridges 1036 can therefore provide link-layer bridging by sending carrier packets 1037 over the spanning tree instead of network-layer routing according to 1038 MNP routes. As a result, opportunities for loss due to node mobility 1039 between different segments are mitigated. 1041 In normal operations, IPv6 ND messages are conveyed over secured 1042 paths between OMNI link neighbors so that specific Proxy/Servers or 1043 Relays can be addressed without being subject to mobility events. 1045 Conversely, only the first few carrier packets destined to Clients 1046 need to traverse secured paths until route optimization can determine 1047 a more direct path. 1049 Note: When the OAL source and destination are on the same INET 1050 segment, the ORH is not needed and OAL header compression can be used 1051 to significantly reduce encapsulation overhead 1052 [I-D.templin-6man-omni]. 1054 Note: When the OAL source has multiple original IP packets to send to 1055 the same OAL destination, it can perform "packing" to generate a 1056 "super-packet" [I-D.templin-6man-omni]. In that case, the OAL/ORH 1057 super-packet may include up to N original IP packets as long as the 1058 total length of the super-packet does not exceed the OMNI interface 1059 MTU. 1061 Note: Use of an IPv6 "minimal encapsulation" format (i.e., an IPv6 1062 variant of [RFC2004]) based on extensions to the ORH was considered 1063 and abandoned. In the approach, the ORH would be inserted as an 1064 extension header to the original IPv6 packet header. The IPv6 1065 destination address would then be written into the ORH, and the ULA 1066 corresponding to the destination would be overwritten in the IPv6 1067 destination address. This would seemingly convey enough forwarding 1068 information so that OAL encapsulation could be avoided. However, 1069 this "minimal encapsulation" IPv6 packet would then have a non-ULA 1070 source address and ULA destination address, an incorrect value in 1071 upper layer protocol checksums, and a Hop Limit that is decremented 1072 within the spanning tree when it should not be. The insertion and 1073 removal of the ORH would also entail rewriting the Payload Length and 1074 Next Header fields - again, invalidating upper layer checksums. 1075 These irregularities would result in implementation challenges and 1076 the potential for operational issues, e.g., since actionable ICMPv6 1077 error reports could not be delivered to the original source. In 1078 order to address the issues, still more information such as the 1079 original IPv6 source address could be written into the ORH. However, 1080 with the additional information the benefit of the "minimal 1081 encapsulation" savings quickly diminishes, and becomes overshadowed 1082 by the implementation and operational irregularities. 1084 3.2.5. Segment Routing Topologies (SRTs) 1086 The 64-bit sub-prefixes of [ULA]::/48 identify up to 2^16 distinct 1087 Segment Routing Topologies (SRTs). Each SRT is a mutually-exclusive 1088 OMNI link overlay instance using a distinct set of ULAs, and emulates 1089 a Virtual LAN (VLAN) service for the OMNI link. In some cases (e.g., 1090 when redundant topologies are needed for fault tolerance and 1091 reliability) it may be beneficial to deploy multiple SRTs that act as 1092 independent overlay instances. A communication failure in one 1093 instance therefore will not affect communications in other instances. 1095 Each SRT is identified by a distinct value in bits 48-63 of 1096 [ULA]::/48, i.e., as [ULA0]::/64, [ULA1]::/64, [ULA2]::/64, etc. 1097 Each OMNI interface is identified by a unique interface name (e.g., 1098 omni0, omni1, omni2, etc.) and assigns an anycast ADM-ULA 1099 corresponding to its SRT prefix length. The anycast ADM-ULA is used 1100 for OMNI interface determination in Safety-Based Multilink (SBM) as 1101 discussed in [I-D.templin-6man-omni]. Each OMNI interface further 1102 applies Performance-Based Multilink (PBM) internally. 1104 3.2.6. Segment Routing For OMNI Link Selection 1106 An original IPv6 source can direct an IPv6 packet to an AERO node by 1107 including a standard IPv6 Segment Routing Header (SRH) [RFC8754] with 1108 the anycast ADM-ULA for the selected SRT as either the IPv6 1109 destination or as an intermediate hop within the SRH. This allows 1110 the original source to determine the specific OMNI link topology an 1111 original IPv6 packet will traverse when there may be multiple 1112 alternatives. 1114 When the AERO node processes the SRH and forwards the original IPv6 1115 packet to the correct OMNI interface, the OMNI interface writes the 1116 next IPv6 address from the SRH into the IPv6 destination address and 1117 decrements Segments Left. If decrementing would cause Segments Left 1118 to become 0, the OMNI interface deletes the SRH before forwarding. 1119 This form of Segment Routing supports Safety-Based Multilink (SBM). 1121 3.2.7. Segment Routing Within the OMNI Link 1123 OAL sources can insert an ORH for Segment Routing within the OMNI 1124 link to influence the paths of OAL packets sent to OAL destinations 1125 in remote segments without requiring all carrier packets to traverse 1126 strict spanning tree paths. 1128 When an AERO node's OMNI interface has an original IP packet to send 1129 to a target discovered through route optimization located in the same 1130 OMNI link segment, it acts as an OAL source to perform OAL 1131 encapsulation and fragmentation. The node then uses the target's 1132 Link Layer Address (L2ADDR) information for *NET encapsulation. 1134 When an AERO node's OMNI interface has an original IP packet to send 1135 to a route optimization target located in a remote OMNI link segment, 1136 it acts as an OAL source the same as above but also includes an ORH 1137 while setting the OAL destination to the Subnet Router Anycast 1138 address for the final OMNI link segment, then forwards the resulting 1139 carrier packets to a Bridge. 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 L2ADDR according to FMT and removes 1144 the ORH from the carrier packet. The Bridge then writes the MNP-ULA 1145 corresponding to the ORH Destination Suffix into the OAL destination 1146 address, decrements the OAL IPv6 header Hop Limit (and discards the 1147 packet if the Hop Limit reaches 0), re-encapsulates the carrier 1148 packet according to L2ADDR and forwards the carrier packet either to 1149 the LHS Proxy/Server or directly to the target Client itself. In 1150 this way, the Bridge participates in route optimization to reduce 1151 traffic load and suboptimal routing through strict spanning tree 1152 paths. 1154 3.3. OMNI Interface Characteristics 1156 OMNI interfaces are virtual interfaces configured over one or more 1157 underlying interfaces classified as follows: 1159 o INET interfaces connect to an INET either natively or through one 1160 or more NATs. Native INET interfaces have global IP addresses 1161 that are reachable from any INET correspondent. The INET-facing 1162 interfaces of Proxy/Servers are native interfaces, as are Relay 1163 and Bridge interfaces. NATed INET interfaces connect to a private 1164 network behind one or more NATs that provide INET access. Clients 1165 that are behind a NAT are required to send periodic keepalive 1166 messages to keep NAT state alive when there are no carrier packets 1167 flowing. 1169 o ANET interfaces connect to an ANET that is separated from the open 1170 INET by a Proxy/Server. Proxy/Servers can actively issue control 1171 messages over the INET on behalf of the Client to reduce ANET 1172 congestion. 1174 o VPNed interfaces use security encapsulation over the INET to a 1175 Virtual Private Network (VPN) server that also acts as a Proxy/ 1176 Server. Other than the link-layer encapsulation format, VPNed 1177 interfaces behave the same as Direct interfaces. 1179 o Direct (i.e., single-hop point-to-point) interfaces connect a 1180 Client directly to a Proxy/Server without crossing any ANET/INET 1181 paths. An example is a line-of-sight link between a remote pilot 1182 and an unmanned aircraft. The same Client considerations apply as 1183 for VPNed interfaces. 1185 OMNI interfaces use OAL encapsulation and fragmentation as discussed 1186 in Section 3.2.4. OMNI interfaces use *NET encapsulation (see: 1187 Section 3.6) to exchange carrier packets with OMNI link neighbors 1188 over INET or VPNed interfaces as well as over ANET interfaces for 1189 which the Client and Proxy/Server may be multiple IP hops away. OMNI 1190 interfaces do not use link-layer encapsulation over Direct underlying 1191 interfaces or ANET interfaces when the Client and Proxy/Server are 1192 known to be on the same underlying link. 1194 OMNI interfaces maintain a neighbor cache for tracking per-neighbor 1195 state the same as for any interface. OMNI interfaces use ND messages 1196 including Router Solicitation (RS), Router Advertisement (RA), 1197 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 1198 neighbor cache management. 1200 OMNI interfaces send ND messages with an OMNI option formatted as 1201 specified in [I-D.templin-6man-omni]. The OMNI option includes 1202 prefix registration information and Interface Attributes containing 1203 link information parameters for the OMNI interface's underlying 1204 interfaces. Each OMNI option may include multiple Interface 1205 Attributes sub-options, each identified by an ifIndex value. 1207 A Client's OMNI interface may be configured over multiple underlying 1208 interface connections. For example, common mobile handheld devices 1209 have both wireless local area network ("WLAN") and cellular wireless 1210 links. These links are often used "one at a time" with low-cost WLAN 1211 preferred and highly-available cellular wireless as a standby, but a 1212 simultaneous-use capability could provide benefits. In a more 1213 complex example, aircraft frequently have many wireless data link 1214 types (e.g. satellite-based, cellular, terrestrial, air-to-air 1215 directional, etc.) with diverse performance and cost properties. 1217 If a Client's multiple underlying interfaces are used "one at a time" 1218 (i.e., all other interfaces are in standby mode while one interface 1219 is active), then ND message OMNI options include only a single 1220 Interface Attributes sub-option set to constant values. In that 1221 case, the Client would appear to have a single interface but with a 1222 dynamically changing link-layer address. 1224 If the Client has multiple active underlying interfaces, then from 1225 the perspective of ND it would appear to have multiple link-layer 1226 addresses. In that case, ND message OMNI options MAY include 1227 multiple Interface Attributes sub-options - each with values that 1228 correspond to a specific interface. Every ND message need not 1229 include Interface Attributes for all underlying interfaces; for any 1230 attributes not included, the neighbor considers the status as 1231 unchanged. 1233 Bridge and Proxy/Server OMNI interfaces may be configured over one or 1234 more secured tunnel interfaces. The OMNI interface configures both 1235 an ADM-LLA and its corresponding ADM-ULA, while the underlying 1236 secured tunnel interfaces are either unnumbered or configure the same 1237 ULA. The OMNI interface acting as an OAL source encapsulates and 1238 fragments each original IP packet, then and presents the resulting 1239 carrier packets to the underlying secured tunnel interface. Routing 1240 protocols such as BGP that run over the OMNI interface do not employ 1241 OAL encapsulation, but rather present the routing protocol messages 1242 directly to the underlying secured tunnels while using the ULA as the 1243 source address. This distinction must be honored consistently 1244 according to each node's configuration so that the IP forwarding 1245 table will associate discovered IP routes with the correct interface. 1247 3.4. OMNI Interface Initialization 1249 AERO Proxy/Servers and Clients configure OMNI interfaces as their 1250 point of attachment to the OMNI link. AERO nodes assign the MSPs for 1251 the link to their OMNI interfaces (i.e., as a "route-to-interface") 1252 to ensure that original IP packets with destination addresses covered 1253 by an MNP not explicitly assigned to a non-OMNI interface are 1254 directed to the OMNI interface. 1256 OMNI interface initialization procedures for Proxy/Servers, Clients 1257 and Bridges are discussed in the following sections. 1259 3.4.1. AERO Proxy/Server and Relay Behavior 1261 When a Proxy/Server enables an OMNI interface, it assigns an 1262 ADM-{LLA,ULA} appropriate for the given OMNI link segment. The 1263 Proxy/Server also configures secured tunnels with one or more 1264 neighboring Bridges and engages in a BGP routing protocol session 1265 with each Bridge. 1267 The OMNI interface provides a single interface abstraction to the IP 1268 layer, but internally includes one or more secured tunnels as well as 1269 an NBMA nexus as underlying interfaces for sending carrier packets to 1270 OMNI interface neighbors. The Proxy/Server further configures a 1271 service to facilitate ND exchanges with AERO Clients and manages per- 1272 Client neighbor cache entries and IP forwarding table entries based 1273 on control message exchanges. 1275 Relays are simply Proxy/Servers that run a dynamic routing protocol 1276 to redistribute routes between the OMNI interface and INET/EUN 1277 interfaces (see: Section 3.2.3). The Relay provisions MNPs to 1278 networks on the INET/EUN interfaces (i.e., the same as a Client would 1279 do) and advertises the MSP(s) for the OMNI link over the INET/EUN 1280 interfaces. The Relay further provides an attachment point of the 1281 OMNI link to a non-MNP-based global topology. 1283 3.4.2. AERO Client Behavior 1285 When a Client enables an OMNI interface, it assigns either an 1286 MNP-{LLA, ULA} or a Temporary ULA and sends RS messages with ND 1287 parameters over its underlying interfaces to a Proxy/Server, which 1288 returns an RA message with corresponding parameters. The RS/RA 1289 messages may pass through one or more NATs in the case of a Client's 1290 INET interface. (Note: if the Client used a Temporary ULA in its 1291 initial RS message, it will discover an MNP-{LLA, ULA} in the 1292 corresponding RA that it receives from the Proxy/Server and begin 1293 using these new addresses. If the Client is operating outside the 1294 context of AERO infrastructure such as in a Mobile Ad-hoc Network 1295 (MANET), however, it may continue using Temporary ULAs for Client-to- 1296 Client communications until it encounters an infrastructure element 1297 that can provide an MNP.) 1299 3.4.3. AERO Bridge Behavior 1301 AERO Bridges configure an OMNI interface and assign the ADM-ULA 1302 Subnet Router Anycast address for each OMNI link segment they connect 1303 to. Bridges configure secured tunnels with Proxy/Servers and other 1304 Bridges, and engage in a BGP routing protocol session with neighbors 1305 on the spanning tree (see: Section 3.2.3). 1307 3.5. OMNI Interface Neighbor Cache Maintenance 1309 Each OMNI interface maintains a conceptual neighbor cache that 1310 includes an entry for each neighbor it communicates with on the OMNI 1311 link per [RFC4861]. In addition to ordinary neighbor cache entries, 1312 proxy neighbor cache entries are created and maintained by AERO 1313 Proxy/Servers when they proxy Client ND message exchanges [RFC4389]. 1314 AERO Proxy/Servers maintain proxy neighbor cache entries for each of 1315 their associated Clients. 1317 To the list of neighbor cache entry states in Section 7.3.2 of 1318 [RFC4861], Proxy/Server OMNI interfaces add an additional state 1319 DEPARTED that applies to Clients that have recently departed. The 1320 interface sets a "DepartTime" variable for the neighbor cache entry 1321 to "DEPART_TIME" seconds. DepartTime is decremented unless a new ND 1322 message causes the state to return to REACHABLE. While a neighbor 1323 cache entry is in the DEPARTED state, the Proxy/Server forwards 1324 carrier packets destined to the target Client to the Client's new 1325 location instead. When DepartTime decrements to 0, the neighbor 1326 cache entry is deleted. It is RECOMMENDED that DEPART_TIME be set to 1327 the default constant value REACHABLE_TIME plus 10 seconds (40 seconds 1328 by default) to allow a window for carrier packets in flight to be 1329 delivered while stale route optimization state may be present. 1331 Proxy/Servers can act as RORs on behalf of disadvantaged Clients 1332 according to the Proxy Neighbor Advertisement specification in 1333 Section 7.2.8 of [RFC4861], while well-connected Clients can act as 1334 an ROR on their own behalf. When a Proxy/Server ROR receives an 1335 authentic NS message used for route optimization, it first searches 1336 for a proxy neighbor cache entry for the target Client and accepts 1337 the message only if there is an entry. The Proxy/Server (or the 1338 actual target Client acting as an ROR) then returns a solicited NA 1339 message while creating a neighbor cache entry for the ROS and caching 1340 the Identification value found in the NS message carrier packet as 1341 the starting window Identification value for this ROS. Proxy/Servers 1342 acting as proxy RORs also create or update a "Report List" entry for 1343 the ROS in the target Client's proxy neighbor cache entry with a 1344 "ReportTime" variable set to REPORT_TIME seconds. The ROR resets 1345 ReportTime when it receives a new authentic NS message, and otherwise 1346 decrements ReportTime while no authentic NS messages have been 1347 received. It is RECOMMENDED that REPORT_TIME be set to the default 1348 constant value REACHABLE_TIME plus 10 seconds (40 seconds by default) 1349 to allow a window for route optimization to converge before 1350 ReportTime decrements below REACHABLE_TIME. 1352 When the ROS receives a solicited NA message response to its NS 1353 message used for route optimization, it creates or updates a neighbor 1354 cache entry for the target network-layer and link-layer addresses. 1355 The ROS then (re)sets ReachableTime for the neighbor cache entry to 1356 REACHABLE_TIME seconds and uses this value to determine whether 1357 carrier packets can be forwarded directly to the target, i.e., 1358 instead of via a default route. The ROS also maintains a window 1359 start Identification value that is monotonically incremented for each 1360 OAL packet sent to this target, and sets new window start 1361 Identification values when it sends a new NS. The ROS otherwise 1362 decrements ReachableTime while no further solicited NA messages 1363 arrive. It is RECOMMENDED that REACHABLE_TIME be set to the default 1364 constant value 30 seconds as specified in [RFC4861]. 1366 AERO nodes also use the value MAX_UNICAST_SOLICIT to limit the number 1367 of NS keepalives sent when a correspondent may have gone unreachable, 1368 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1369 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1370 to limit the number of unsolicited NAs that can be sent based on a 1371 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1372 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1373 same as specified in [RFC4861]. 1375 Different values for DEPART_TIME, REPORT_TIME, REACHABLE_TIME, 1376 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1377 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1378 different values are chosen, all nodes on the link MUST consistently 1379 configure the same values. Most importantly, DEPART_TIME and 1380 REPORT_TIME SHOULD be set to a value that is sufficiently longer than 1381 REACHABLE_TIME to avoid packet loss due to stale route optimization 1382 state. 1384 3.5.1. OMNI Neighbor Interface Attributes 1386 OMNI interface IPv6 ND messages include OMNI options 1387 [I-D.templin-6man-omni] with Interface Attributes that provide Link- 1388 Layer Address and QoS Preference information for the neighbor's 1389 underlying interfaces. This information is stored in the neighbor 1390 cache and provides the basis for the forwarding algorithm specified 1391 in Section 3.10. The information is cumulative and reflects the 1392 union of the OMNI information from the most recent ND messages 1393 received from the neighbor; it is therefore not required that each ND 1394 message contain all neighbor information. 1396 The OMNI option Interface Attributes for each underlying interface 1397 includes a two-part "Link-Layer Address" consisting of a simple IP 1398 encapsulation address determined by the FMT and L2ADDR fields and an 1399 ADM-ULA determined by the SRT and LHS fields. Underlying interfaces 1400 are further selected based on their associated preference values 1401 "high", "medium", "low" or "disabled". 1403 Note: the OMNI option is distinct from any Source/Target Link-Layer 1404 Address Options (S/TLLAOs) that may appear in an ND message according 1405 to the appropriate IPv6 over specific link layer specification (e.g., 1406 [RFC2464]). If both an OMNI option and S/TLLAO appear, the former 1407 pertains to encapsulation addresses while the latter pertains to the 1408 native L2 address format of the underlying media. 1410 3.5.2. OMNI Neighbor Advertisement Message Flags 1412 As discussed in Section 4.4 of [RFC4861] NA messages include three 1413 flag bits R, S and O. OMNI interface NA messages treat the flags as 1414 follows: 1416 o R: The R ("Router") flag is set to 1 in the NA messages sent by 1417 all AERO/OMNI node types. Simple hosts that would set R to 0 do 1418 not occur on the OMNI link itself, but may occur on the downstream 1419 links of Clients and Relays. 1421 o S: The S ("Solicited") flag is set exactly as specified in 1422 Section 4.4. of [RFC4861], i.e., it is set to 1 for Solicited NAs 1423 and set to 0 for Unsolicited NAs (both unicast and multicast). 1425 o O: The O ("Override") flag is set to 0 for solicited proxy NAs 1426 returned by a Proxy/Server ROR and set to 1 for all other 1427 solicited and unsolicited NAs. For further study is whether 1428 solicited NAs for anycast targets apply for OMNI links. Since 1429 MNP-LLAs must be uniquely assigned to Clients to support correct 1430 ND protocol operation, however, no role is currently seen for 1431 assigning the same MNP-LLA to multiple Clients. 1433 3.6. OMNI Interface Encapsulation and Re-encapsulation 1435 The OMNI interface admits original IP packets then (acting as an OAL 1436 source) performs OAL encapsulation and fragmentation as specified in 1437 [I-D.templin-6man-omni] while including an ORH if necessary as 1438 specified in Section 3.2.4. OAL encapsulation produces OAL packets, 1439 while OAL fragmentation turns them into OAL fragments which are then 1440 encapsulated in *NET headers as carrier packets. 1442 For carrier packets undergoing re-encapsulation at an OAL 1443 intermediate node, the OMNI interface decrements the OAL IPv6 header 1444 Hop Limit and discards the carrier packet if the Hop Limit reaches 0. 1445 The intermediate node next removes the *NET encapsulation headers 1446 from the first segment and re-encapsulates the packet in new *NET 1447 encapsulation headers for the next segment. 1449 When a Proxy/Server or Relay re-encapsulates a carrier packet 1450 received from a Client that includes an OAL but no ORH, it inserts an 1451 ORH immediately following the OAL header and adjusts the OAL payload 1452 length and destination address field. The inserted ORH will be 1453 removed by the final-hop Bridge, but its insertion and removal will 1454 not interfere with reassembly at the final destination. For this 1455 reason, Clients must reserve 40 bytes for a maximum-length ORH when 1456 they perform OAL encapsulation (see: Section 3.9). 1458 3.7. OMNI Interface Decapsulation 1460 OMNI interfaces (acting as OAL destinations) decapsulate and 1461 reassemble OAL packets into original IP packets destined either to 1462 the AERO node itself or to a destination reached via an interface 1463 other than the OMNI interface the original IP packet was received on. 1464 When carrier packets containing OAL fragments arrive, the OMNI 1465 interface reassembles as discussed in Section 3.9. 1467 3.8. OMNI Interface Data Origin Authentication 1469 AERO nodes employ simple data origin authentication procedures. In 1470 particular: 1472 o AERO Bridges and Proxy/Servers accept carrier packets (including 1473 either data or control messages) received from the (secured) 1474 spanning tree. 1476 o AERO Proxy/Servers and Clients accept carrier packets and original 1477 IP packets that originate from within the same secured ANET. 1479 o AERO Clients and Relays accept original IP packets from downstream 1480 network correspondents based on ingress filtering. 1482 o AERO Clients, Relays and Proxy/Servers verify carrier packet UDP/ 1483 IP encapsulation addresses according to [RFC4380]. 1485 o AERO Clients (as well as Proxy/Servers and Relays when acting as 1486 OAL destinations) accept OAL packets/fragments with Identification 1487 values within the current window for the OAL source. 1489 AERO nodes silently drop any packets that do not satisfy the above 1490 data origin authentication procedures. Further security 1491 considerations are discussed in Section 6. 1493 3.9. OMNI Interface MTU 1495 The OMNI interface observes the link nature of tunnels, including the 1496 Maximum Transmission Unit (MTU), Maximum Reassembly Unit (MRU) and 1497 the role of fragmentation and reassembly [I-D.ietf-intarea-tunnels]. 1498 The OMNI interface employs an OMNI Adaptation Layer (OAL) that 1499 accommodates multiple underlying links with diverse MTUs while 1500 observing both a minimum and per-path Maximum Payload Size (MPS). 1501 The functions of the OAL and the OMNI interface MTU/MRU/MPS are 1502 specified in [I-D.templin-6man-omni] with MTU/MRU both set to the 1503 constant value 9180 bytes, with minimum MPS set to 400 bytes, and 1504 with per-path MPS set to potentially larger values depending on the 1505 underlying path. 1507 When the network layer presents an original IP packet to the OMNI 1508 interface, the OAL source encapsulates and fragments the original IP 1509 packet if necessary. When the network layer presents the OMNI 1510 interface with multiple original IP packets bound to the same OAL 1511 destination, the OAL source can concatenate them together into a 1512 single OAL super-packet as discussed in [I-D.templin-6man-omni]. The 1513 OAL source then fragments the OAL packet if necessary according to 1514 the minimum/path MPS such that the OAL headers appear in each 1515 fragment while the original IP packet header appears only in the 1516 first fragment. The OAL source then encapsulates each OAL fragment 1517 in *NET headers for transmission as carrier packets over an 1518 underlying interface connected to either a physical link such as 1519 Ethernet, WiFi and the like or a virtual link such as an Internet or 1520 higher-layer tunnel (see the definition of link in [RFC8200]). 1522 Note: A Client that does not (yet) have neighbor cache state for a 1523 target may omit the ORH in carrier packets with the understanding 1524 that a Proxy/Server may insert an ORH on its behalf. For this 1525 reason, Clients reserve 40 bytes for the largest possible ORH in 1526 their OAL fragment size calculations. 1528 Note: Although the ORH may be removed by a Bridge on the path (see: 1529 Section 3.10.3), this does not interfere with the destination's 1530 ability to reassemble. This is due to the fact that the ORH is not 1531 included in the fragmentable part; therefore, its removal does not 1532 invalidate the offset values in any fragment headers. 1534 3.10. OMNI Interface Forwarding Algorithm 1536 Original IP packets enter a node's OMNI interface either from the 1537 network layer (i.e., from a local application or the IP forwarding 1538 system) while carrier packets enter from the link layer (i.e., from 1539 an OMNI interface neighbor). All original IP packets and carrier 1540 packets entering a node's OMNI interface first undergo data origin 1541 authentication as discussed in Section 3.8. Those that satisfy data 1542 origin authentication are processed further, while all others are 1543 dropped silently. 1545 Original IP packets that enter the OMNI interface from the network 1546 layer are forwarded to an OMNI interface neighbor using OAL 1547 encapsulation and fragmentation to produce carrier packets for 1548 transmission over underlying interfaces. (If routing indicates that 1549 the original IP packet should instead be forwarded back to the 1550 network layer, the packet is dropped to avoid looping). Carrier 1551 packets that enter the OMNI interface from the link layer are either 1552 re-encapsulated and re-admitted into the OMNI link, or reassembled 1553 and forwarded to the network layer where they are subject to either 1554 local delivery or IP forwarding. In all cases, the OAL MUST NOT 1555 decrement the network layer TTL/Hop-count since its forwarding 1556 actions occur below the network layer. 1558 OMNI interfaces may have multiple underlying interfaces and/or 1559 neighbor cache entries for neighbors with multiple underlying 1560 interfaces (see Section 3.3). The OAL uses interface attributes and/ 1561 or traffic classifiers (e.g., DSCP value, port number, flow 1562 specification, etc.) to select an outgoing underlying interface for 1563 each OAL packet based on the node's own QoS preferences, and also to 1564 select a destination link-layer address based on the neighbor's 1565 underlying interface with the highest preference. AERO 1566 implementations SHOULD allow for QoS preference values to be modified 1567 at runtime through network management. 1569 If multiple outgoing interfaces and/or neighbor interfaces have a 1570 preference of "high", the AERO node replicates the OAL packet and 1571 sends one copy via each of the (outgoing / neighbor) interface pairs; 1572 otherwise, the node sends a single copy of the OAL packet via an 1573 interface with the highest preference. (While not strictly required, 1574 successful delivery may be more likely when all OAL fragments of the 1575 same OAL packet are sent over the same underlying interface.) AERO 1576 nodes keep track of which underlying interfaces are currently 1577 "reachable" or "unreachable", and only use "reachable" interfaces for 1578 forwarding purposes. 1580 The following sections discuss the OMNI interface forwarding 1581 algorithms for Clients, Proxy/Servers and Bridges. In the following 1582 discussion, an original IP packet's destination address is said to 1583 "match" if it is the same as a cached address, or if it is covered by 1584 a cached prefix (which may be encoded in an MNP-LLA). 1586 3.10.1. Client Forwarding Algorithm 1588 When an original IP packet enters a Client's OMNI interface from the 1589 network layer the Client searches for a neighbor cache entry that 1590 matches the destination. If there is a match, the Client selects one 1591 or more "reachable" neighbor interfaces in the entry for forwarding 1592 purposes. If there is no neighbor cache entry, the Client instead 1593 forwards the original IP packet toward a Proxy/Server. The Client 1594 (acting as an OAL source) performs OAL encapsulation and sets the OAL 1595 destination address to the MNP-ULA if there is a matching neighbor 1596 cache entry; otherwise, it sets the OAL destination to the ADM-ULA of 1597 the Proxy/Server. If the Client has multiple original IP packets to 1598 send to the same neighbor, it can concatenate them in a single super- 1599 packet [I-D.templin-6man-omni]. The OAL source then performs 1600 fragmentation to create OAL fragments (see: Section 3.9), appends any 1601 *NET encapsulation, and sends the resulting carrier packets over 1602 underlying interfaces to the neighbor acting as an OAL destination. 1604 If the neighbor interface selected for forwarding is located on the 1605 same OMNI link segment and not behind a NAT, the Client forwards the 1606 carrier packets directly according to the L2ADDR information for the 1607 neighbor. If the neighbor interface is behind a NAT on the same OMNI 1608 link segment, the Client instead forwards the initial carrier packets 1609 to its Proxy/Server and initiates NAT traversal procedures. If the 1610 Client's intended source underlying interface is also behind a NAT 1611 and located on the same OMNI link segment, it sends a "direct bubble" 1612 over the interface per [RFC6081][RFC4380] to the L2ADDR found in the 1613 neighbor cache in order to establish state in its own NAT by 1614 generating traffic toward the neighbor (note that no response to the 1615 bubble is expected). 1617 The Client next sends an NS(NUD) message toward the MNP-ULA of the 1618 neighbor via its Proxy/Server as discussed in Section 3.15. If the 1619 Client receives an NA(NUD) from the neighbor over the underlying 1620 interface, it marks the neighbor interface as "trusted" and sends 1621 future carrier packets directly to the L2ADDR information for the 1622 neighbor instead of indirectly via the Proxy/Server. The Client must 1623 honor the neighbor cache maintenance procedure by sending additional 1624 direct bubbles and/or NS/NA(NUD) messages as discussed in 1625 [RFC6081][RFC4380] in order to keep NAT state alive as long as 1626 carrier packets are still flowing. 1628 When an carrier packet enters a Client's OMNI interface from the 1629 link-layer, if the OAL destination matches one of the Client's MNPs 1630 or LLAs the Client (acting as an OAL destination) reassembles and 1631 decapsulates as necessary and delivers the original IP packet to the 1632 network layer. Otherwise, the Client drops the original IP packet 1633 and MAY return a network-layer ICMP Destination Unreachable message 1634 subject to rate limiting (see: Section 3.11). 1636 Note: Clients and their Proxy/Server (and other Client) peers can 1637 exchange original IP packets over ANET underlying interfaces without 1638 invoking the OAL, since the ANET is secured at the link and physical 1639 layers. By forwarding original IP packets without invoking the OAL, 1640 however, the ANET peers can engage only in classical path MTU 1641 discovery since the packets are subject to loss and/or corruption due 1642 to the various per-link MTU limitations that may occur within the 1643 ANET. Moreover, the original IP packets do not include per-packet 1644 Identification values that can be used for data origin authentication 1645 and link-layer retransmission purposes, nor the OAL integrity check. 1646 The tradeoff therefore involves an assessment of the per-packet 1647 encapsulation overhead saved by bypassing the OAL vs. inheritance of 1648 classical network "brittleness". 1650 3.10.2. Proxy/Server and Relay Forwarding Algorithm 1652 When the Proxy/Server receives an original IP packet from the network 1653 layer, it drops the packet if routing indicates that it should be 1654 forwarded back to the network layer to avoid looping. Otherwise, the 1655 Proxy/Server regards the original IP packet the same as if it had 1656 arrived as carrier packets with OAL destination set to its own ADM- 1657 ULA. When the Proxy/Server receives carrier packets on underlying 1658 interfaces with OAL destination set to its own ADM-ULA, it performs 1659 OAL reassembly if necessary to obtain the original IP packet. 1661 The Proxy/Server next searches for a neighbor cache entry that 1662 matches the original IP destination and proceeds as follows: 1664 o if the original IP packet destination matches a neighbor cache 1665 entry, the Proxy/Sever uses one or more "reachable" neighbor 1666 interfaces in the entry for packet forwarding using OAL 1667 encapsulation and fragmentation according to the cached link-layer 1668 address information. If the neighbor interface is in a different 1669 OMNI link segment, the Proxy/Server forwards the resulting carrier 1670 packets to a Bridge; otherwise, it forwards the carrier packets 1671 directly to the neighbor. If the neighbor is behind a NAT, the 1672 Proxy/Server instead forwards initial carrier packets via a Bridge 1673 while sending an NS(NUD) to the neighbor. When the Proxy/Server 1674 receives the NA(NUD), it can begin forwarding carrier packets 1675 directly to the neighbor the same as discussed in Section 3.10.1 1676 while sending additional NS(NUD) messages as necessary to maintain 1677 NAT state. Note that no direct bubbles are necessary since the 1678 Proxy/Server is by definition not located behind a NAT. 1680 o else, if the original IP destination matches a non-MNP route in 1681 the IP forwarding table or an ADM-LLA assigned to the Proxy/ 1682 Server's OMNI interface, the Proxy/Server acting as a Relay 1683 presents the original IP packet to the network layer for local 1684 delivery or IP forwarding. 1686 o else, the Proxy/Server initiates address resolution as discussed 1687 in Section 3.14, while retaining initial original IP packets in a 1688 small queue awaiting address resolution completion. 1690 When the Proxy/Server receives a carrier packet with OAL destination 1691 set to an MNP-ULA that does not match the MSP, it accepts the carrier 1692 packet only if data origin authentication succeeds and if there is a 1693 network layer routing table entry for a GUA route that matches the 1694 MNP-ULA. If there is no route, the Proxy/Server drops the carrier 1695 packet; otherwise, it reassembles and decapsulates to obtain the 1696 original IP packet and acts as a Relay to present it to the network 1697 layer where it will be delivered according to standard IP forwarding. 1699 When the Proxy/Server receives a carrier packet with OAL destination 1700 set to an MNP-ULA, it accepts the carrier packet only if data origin 1701 authentication succeeds and if there is a neighbor cache entry that 1702 matches the OAL destination. If the neighbor cache entry state is 1703 DEPARTED, the Proxy/Server inserts an ORH that encodes the MNP-ULA 1704 destination suffix and changes the OAL destination address to the 1705 ADM-ULA of the new Proxy/Server, then re-encapsulates the carrier 1706 packet and forwards it to a Bridge which will eventually deliver it 1707 to the new Proxy/Server. 1709 If the neighbor cache state for the MNP-ULA is REACHABLE, the Proxy/ 1710 Server forwards the carrier packets to the Client which then must 1711 reassemble. (Note that the Proxy/Server does not reassemble carrier 1712 packets not explicitly addressed to its own ADM-ULA, since routing 1713 could direct some of the carrier packet of the same original IP 1714 packet through a different Proxy/Server.) In that case, the Client 1715 may receive fragments that are smaller than its link MTU but can 1716 still be reassembled; if this proves inefficient, the Client can in 1717 the future elect to employ the Proxy/Server as a ROR instead of 1718 serving in that role on its own behalf. 1720 Note: Clients and their Proxy/Server peers can exchange original IP 1721 packets over ANET underlying interfaces without invoking the OAL, 1722 since the ANET is secured at the link and physical layers. By 1723 forwarding original IP packets without invoking the OAL, however, the 1724 Client and Proxy/Server can engage only in classical path MTU 1725 discovery since the packets are subject to loss and/or corruption due 1726 to the various per-link MTU limitations that may occur within the 1727 ANET. Moreover, the original IP packets do not include per-packet 1728 Identification values that can be used for data origin authentication 1729 and link-layer retransmission purposes, nor the OAL integrity check. 1730 The tradeoff therefore involves an assessment of the per-packet 1731 encapsulation overhead saved by bypassing the OAL vs. inheritance of 1732 classical network "brittleness". 1734 Note: When a Proxy/Server receives a (non-OAL) original IP packet 1735 from an ANET Client, or a carrier packet with OAL destination set to 1736 its own ADM-ULA from any Client, the Proxy/Server reassembles if 1737 necessary then performs ROS functions on behalf of the Client. The 1738 Client may at some later time begin sending carrier packets to the 1739 OAL address of the actual target instead of the Proxy/Server, at 1740 which point it may begin functioning as an ROS on its own behalf and 1741 thereby "override" the Proxy/Server's ROS role. 1743 Note: If the Proxy/Server has multiple original IP packets to send to 1744 the same neighbor, it can concatenate them in a single OAL super- 1745 packet [I-D.templin-6man-omni]. 1747 3.10.3. Bridge Forwarding Algorithm 1749 Bridges forward carrier packets the same as any IPv6 router. Bridges 1750 convey carrier packets and original IP packets that encapsulate IPv6 1751 ND control messages or routing protocol control messages using 1752 security encapsulations, and may convey packets that encapsulate 1753 ordinary data without including security encapsulations. When the 1754 Bridge receives a carrier packet or an original IP packet, it removes 1755 the outer *NET header and searches for a forwarding table entry that 1756 matches the OAL destination address. The Bridge then processes the 1757 packet as follows: 1759 o if the packet is a carrier packet with a destination that matches 1760 its ADM-ULA Subnet Router Anycast address the Bridge processes the 1761 carrier packet locally before forwarding. The Bridge drops the 1762 carrier packet if it does not include an ORH; otherwise, for 1763 NA(NUD) messages the Bridge replaces the OMNI option Interface 1764 Attributes sub-option with information for its own interface while 1765 retaining the ifIndex value supplied by the NA(NUD) message 1766 source. The Bridge next examines the ORH FMT code. If the code 1767 indicates the destination is a Client on the open *NET (or, a 1768 Client behind a NAT for which NAT traversal procedures have 1769 already converged) the Bridge removes the ORH then writes the MNP- 1770 ULA formed from the ORH Destination Suffix into the OAL 1771 destination. The Bridge then re-encapsulates the carrier packet 1772 and forwards it to the ORH L2ADDR. For all other destination 1773 cases, the Bridge instead writes the ADM-ULA formed from the ORH 1774 SRT/LHS into the OAL destination address and forwards the carrier 1775 packet to the ADM-ULA Proxy/Server while invoking NAT traversal 1776 procedures the same as for Proxy/Servers if necessary, noting that 1777 no direct bubbles are necessary since only the target Client and 1778 not the Bridge is behind a NAT. 1780 o else, if the packet is a carrier packet with a destination that 1781 matches a forwarding table entry the Bridge forwards the carrier 1782 packet via a secured tunnel to the next hop. (If the destination 1783 matches an MSP without matching an MNP, however, the Bridge 1784 instead drops the packet and returns an ICMP Destination 1785 Unreachable message subject to rate limiting - see: Section 3.11). 1787 o else, if the packet is an original IP packet with a destination 1788 that matches one of the Bridge's own addresses, the Bridge submits 1789 the original IP packet for local delivery to support local 1790 applications such as routing protocols. 1792 o else, the Bridge drops the packet and returns an ICMP Destination 1793 Unreachable as above. 1795 As for any IP router, the Bridge decrements the OAL IPv6 header Hop 1796 Limit when it forwards the carrier packet and drops the packet if the 1797 Hop Limit reaches 0. Therefore, when an OAL header is present only 1798 the Hop Limit in the OAL header is decremented and not the TTL/Hop 1799 Limit in the original IP packet header. Bridges do not insert OAL/ 1800 ORH headers themselves; instead, they act as IPv6 routers and forward 1801 carrier packets based on their destination addresses. 1803 Bridges forward packets received from a first segment without 1804 security encapsulations to the next segment also without including 1805 security encapsulations. Bridges forward packets received from a 1806 first segment with security encapsulations to the next segment also 1807 including security encapsulations. Bridges use a single IPv6 routing 1808 table that always determines the same next hop for a given OAL 1809 destination whether or not security encapsulation is included. 1811 3.11. OMNI Interface Error Handling 1813 When an AERO node admits an original IP packet into the OMNI 1814 interface, it may receive link-layer or network-layer error 1815 indications. 1817 A link-layer error indication is an ICMP error message generated by a 1818 router in the INET on the path to the neighbor or by the neighbor 1819 itself. The message includes an IP header with the address of the 1820 node that generated the error as the source address and with the 1821 link-layer address of the AERO node as the destination address. 1823 The IP header is followed by an ICMP header that includes an error 1824 Type, Code and Checksum. Valid type values include "Destination 1825 Unreachable", "Time Exceeded" and "Parameter Problem" 1826 [RFC0792][RFC4443]. (OMNI interfaces ignore link-layer IPv4 1827 "Fragmentation Needed" and IPv6 "Packet Too Big" messages for carrier 1828 packets that are no larger than the minimum/path MPS as discussed in 1829 Section 3.9, however these messages may provide useful hints of probe 1830 failures during path MPS probing.) 1832 The ICMP header is followed by the leading portion of the carrier 1833 packet that generated the error, also known as the "packet-in-error". 1834 For ICMPv6, [RFC4443] specifies that the packet-in-error includes: 1835 "As much of invoking packet as possible without the ICMPv6 packet 1836 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1837 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1838 "Internet Header + 64 bits of Original Data Datagram", however 1839 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1840 ICMP datagram SHOULD contain as much of the original datagram as 1841 possible without the length of the ICMP datagram exceeding 576 1842 bytes". 1844 The link-layer error message format is shown in Figure 5 (where, "L2" 1845 and "L3" refer to link-layer and network-layer, respectively): 1847 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1848 ~ ~ 1849 | L2 IP Header of | 1850 | error message | 1851 ~ ~ 1852 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1853 | L2 ICMP Header | 1854 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1855 ~ ~ P 1856 | carrier packet *NET and OAL | a 1857 | encapsulation headers | c 1858 ~ ~ k 1859 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1860 ~ ~ t 1861 | original IP packet headers | 1862 | (first-fragment only) | i 1863 ~ ~ n 1864 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1865 ~ ~ e 1866 | Portion of the body of | r 1867 | the original IP packet | r 1868 | (all fragments) | o 1869 ~ ~ r 1870 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1872 Figure 5: OMNI Interface Link-Layer Error Message Format 1874 The AERO node rules for processing these link-layer error messages 1875 are as follows: 1877 o When an AERO node receives a link-layer Parameter Problem message, 1878 it processes the message the same as described as for ordinary 1879 ICMP errors in the normative references [RFC0792][RFC4443]. 1881 o When an AERO node receives persistent link-layer Time Exceeded 1882 messages, the IP ID field may be wrapping before earlier fragments 1883 awaiting reassembly have been processed. In that case, the node 1884 should begin including integrity checks and/or institute rate 1885 limits for subsequent packets. 1887 o When an AERO node receives persistent link-layer Destination 1888 Unreachable messages in response to carrier packets that it sends 1889 to one of its neighbor correspondents, the node should process the 1890 message as an indication that a path may be failing, and 1891 optionally initiate NUD over that path. If it receives 1892 Destination Unreachable messages over multiple paths, the node 1893 should allow future carrier packets destined to the correspondent 1894 to flow through a default route and re-initiate route 1895 optimization. 1897 o When an AERO Client receives persistent link-layer Destination 1898 Unreachable messages in response to carrier packets that it sends 1899 to one of its neighbor Proxy/Servers, the Client should mark the 1900 path as unusable and use another path. If it receives Destination 1901 Unreachable messages on many or all paths, the Client should 1902 associate with a new Proxy/Server and release its association with 1903 the old Proxy/Server as specified in Section 3.16.5. 1905 o When an AERO Proxy/Server receives persistent link-layer 1906 Destination Unreachable messages in response to carrier packets 1907 that it sends to one of its neighbor Clients, the Proxy/Server 1908 should mark the underlying path as unusable and use another 1909 underlying path. 1911 o When an AERO Proxy/Server receives link-layer Destination 1912 Unreachable messages in response to a carrier packet that it sends 1913 to one of its permanent neighbors, it treats the messages as an 1914 indication that the path to the neighbor may be failing. However, 1915 the dynamic routing protocol should soon reconverge and correct 1916 the temporary outage. 1918 When an AERO Bridge receives a carrier packet for which the network- 1919 layer destination address is covered by an MSP, the Bridge drops the 1920 packet if there is no more-specific routing information for the 1921 destination and returns a network-layer Destination Unreachable 1922 message subject to rate limiting. The Bridge writes the network- 1923 layer source address of the original IP packet as the destination 1924 address and uses one of its non link-local addresses as the source 1925 address of the message. 1927 When an AERO node receives a carrier packet for which reassembly is 1928 currently congested, it returns a network-layer Packet Too Big (PTB) 1929 message as discussed in [I-D.templin-6man-omni] (note that the PTB 1930 messages could indicate either "hard" or "soft" errors). 1932 3.12. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1934 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1935 coordinated as discussed in the following Sections. 1937 3.12.1. AERO Service Model 1939 Each AERO Proxy/Server on the OMNI link is configured to facilitate 1940 Client prefix delegation/registration requests. Each Proxy/Server is 1941 provisioned with a database of MNP-to-Client ID mappings for all 1942 Clients enrolled in the AERO service, as well as any information 1943 necessary to authenticate each Client. The Client database is 1944 maintained by a central administrative authority for the OMNI link 1945 and securely distributed to all Proxy/Servers, e.g., via the 1946 Lightweight Directory Access Protocol (LDAP) [RFC4511], via static 1947 configuration, etc. Clients receive the same service regardless of 1948 the Proxy/Servers they select. 1950 AERO Clients and Proxy/Servers use ND messages to maintain neighbor 1951 cache entries. AERO Proxy/Servers configure their OMNI interfaces as 1952 advertising NBMA interfaces, and therefore send unicast RA messages 1953 with a short Router Lifetime value (e.g., ReachableTime seconds) in 1954 response to a Client's RS message. Thereafter, Clients send 1955 additional RS messages to keep Proxy/Server state alive. 1957 AERO Clients and Proxy/Servers include prefix delegation and/or 1958 registration parameters in RS/RA messages (see 1959 [I-D.templin-6man-omni]). The ND messages are exchanged between 1960 Client and Proxy/Server according to the prefix management schedule 1961 required by the service. If the Client knows its MNP in advance, it 1962 can employ prefix registration by including its MNP-LLA as the source 1963 address of an RS message and with an OMNI option with valid prefix 1964 registration information for the MNP. If the Proxy/Server accepts 1965 the Client's MNP assertion, it injects the MNP into the routing 1966 system and establishes the necessary neighbor cache state. If the 1967 Client does not have a pre-assigned MNP, it can instead employ prefix 1968 delegation by including the unspecified address (::) as the source 1969 address of an RS message and with an OMNI option with prefix 1970 delegation parameters to request an MNP. 1972 The following sections specify the Client and Proxy/Server behavior. 1974 3.12.2. AERO Client Behavior 1976 AERO Clients discover the addresses of Proxy/Servers in a similar 1977 manner as described in [RFC5214]. Discovery methods include static 1978 configuration (e.g., from a flat-file map of Proxy/Server addresses 1979 and locations), or through an automated means such as Domain Name 1980 System (DNS) name resolution [RFC1035]. Alternatively, the Client 1981 can discover Proxy/Server addresses through a layer 2 data link login 1982 exchange, or through a unicast RA response to a multicast/anycast RS 1983 as described below. In the absence of other information, the Client 1984 can resolve the DNS Fully-Qualified Domain Name (FQDN) 1985 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1986 text string and "[domainname]" is a DNS suffix for the OMNI link 1987 (e.g., "example.com"). 1989 To associate with a Proxy/Server, the Client acts as a requesting 1990 router to request MNPs. The Client prepares an RS message with 1991 prefix management parameters and includes a Nonce and Timestamp 1992 option if the Client needs them to correlate RA replies. If the 1993 Client already knows the Proxy/Server's ADM-LLA, it includes the LLA 1994 as the network-layer destination address; otherwise, the Client 1995 includes the (link-local) All-Routers multicast as the network-layer 1996 destination. If the Client already knows its own MNP-LLA, it can use 1997 the MNP-LLA as the network-layer source address and include an OMNI 1998 option with prefix registration information. Otherwise, the Client 1999 uses the unspecified address (::) as the network-layer source address 2000 and includes prefix delegation parameters in the OMNI option (see: 2001 [I-D.templin-6man-omni]). The Client includes Interface Attributes 2002 corresponding to the underlying interface over which it will send the 2003 RS message, and MAY include additional Interface Attributes specific 2004 to other underlying interfaces. 2006 For INET Clients, the Client must ensure that the RS message is no 2007 larger than the minimum/path MPS for the chosen Proxy/Server and must 2008 include a security signature that the Proxy/Server can verify. The 2009 Client next applies OAL encapsulation such that the entire RS message 2010 fits within an OAL First Fragment (i.e., as an atomic fragment) while 2011 including an unpredictable OAL Identification number selected per 2012 [RFC7739] that will serve as the window start Identification value 2013 for future packets it will send/accept with its own MNP-ULA and the 2014 Proxy/Server's ADM-ULA as the OAL addresses. (The Proxy/Server in 2015 turn caches the Identification number as start value for future 2016 packets it will accept/send with its own ADM-ULA and the Client's 2017 MNP-ULA as the OAL addresses.) 2019 The Client then sends the RS message (either directly via Direct 2020 interfaces, via a VPN for VPNed interfaces, via an access router for 2021 ANET interfaces or via INET encapsulation for INET interfaces) while 2022 using OAL encapsulation/fragmentation, then waits for an RA message 2023 reply (see Section 3.12.3). The Client retries up to 2024 MAX_RTR_SOLICITATIONS times until an RA is received. If the Client 2025 receives no RAs, or if it receives an RA with Router Lifetime set to 2026 0, the Client SHOULD abandon attempts through the first Proxy/Server 2027 and try another Proxy/Server. Otherwise, the Client processes the 2028 prefix information found in the RA message. 2030 When the Client processes an RA, it first performs OAL reassembly and 2031 decapsulation then creates a neighbor cache entry with the Proxy/ 2032 Server's ADM-LLA as the network-layer address and the Proxy/Server's 2033 encapsulation and/or link-layer addresses as the link-layer address. 2034 The Client next records the RA Router Lifetime field value in the 2035 neighbor cache entry as the time for which the Proxy/Server has 2036 committed to maintaining the MNP in the routing system via this 2037 underlying interface, and caches the other RA configuration 2038 information including Cur Hop Limit, M and O flags, Reachable Time 2039 and Retrans Timer. The Client then autoconfigures MNP-LLAs for any 2040 delegated MNPs and assigns them to the OMNI interface. The Client 2041 also caches any MSPs included in Route Information Options (RIOs) 2042 [RFC4191] as MSPs to associate with the OMNI link, and assigns the 2043 MTU value in the MTU option to the underlying interface. 2045 The Client then registers additional underlying interfaces with the 2046 Proxy/Server by sending RS messages via each additional interface as 2047 described above. The RS messages include the same parameters as for 2048 the initial RS/RA exchange, but with destination address set to the 2049 Proxy/Server's ADM-LLA. The Client finally sub-delegates the MNPs to 2050 its attached EUNs and/or the Client's own internal virtual interfaces 2051 as described in [I-D.templin-v6ops-pdhost] to support the Client's 2052 downstream attached "Internet of Things (IoT)". The Client then 2053 sends additional RS messages over each underlying interface before 2054 the Router Lifetime received for that interface expires. 2056 After the Client registers its underlying interfaces, it may wish to 2057 change one or more registrations, e.g., if an interface changes 2058 address or becomes unavailable, if QoS preferences change, etc. To 2059 do so, the Client prepares an RS message to send over any available 2060 underlying interface as above. The RS includes an OMNI option with 2061 prefix registration/delegation information, with Interface Attributes 2062 specific to the selected underlying interface, and with any 2063 additional Interface Attributes specific to other underlying 2064 interfaces. When the Client receives the Proxy/Server's RA response, 2065 it has assurance that the Proxy/Server has been updated with the new 2066 information. 2068 If the Client wishes to discontinue use of a Proxy/Server it issues 2069 an RS message over any underlying interface with an OMNI option with 2070 a prefix release indication. When the Proxy/Server processes the 2071 message, it releases the MNP, sets the neighbor cache entry state for 2072 the Client to DEPARTED and returns an RA reply with Router Lifetime 2073 set to 0. After a short delay (e.g., 2 seconds), the Proxy/Server 2074 withdraws the MNP from the routing system. 2076 3.12.3. AERO Proxy/Server Behavior 2078 AERO Proxy/Servers act as IP routers and support a prefix delegation/ 2079 registration service for Clients. Proxy/Servers arrange to add their 2080 ADM-LLAs to a static map of Proxy/Server addresses for the link and/ 2081 or the DNS resource records for the FQDN 2082 "linkupnetworks.[domainname]" before entering service. Proxy/Server 2083 addresses should be geographically and/or topologically referenced, 2084 and made available for discovery by Clients on the OMNI link. 2086 When a Proxy/Server receives a prospective Client's RS message on its 2087 OMNI interface, it SHOULD return an immediate RA reply with Router 2088 Lifetime set to 0 if it is currently too busy or otherwise unable to 2089 service the Client. Otherwise, the Proxy/Server performs OAL 2090 reassembly and decapsulation, then authenticates the RS message and 2091 processes the prefix delegation/registration parameters. The Proxy/ 2092 Server first determines the correct MNPs to provide to the Client by 2093 processing the MNP-LLA prefix parameters and/or the DHCPv6 OMNI sub- 2094 option. When the Proxy/Server returns the MNPs, it also creates a 2095 forwarding table entry for the MNP-ULA corresponding to each MNP so 2096 that the MNPs are propagated into the routing system (see: 2097 Section 3.2.3). For IPv6, the Proxy/Server creates an IPv6 2098 forwarding table entry for each MNP. For IPv4, the Proxy/Server 2099 creates an IPv6 forwarding table entry with the IPv4-compatibility 2100 MNP-ULA prefix corresponding to the IPv4 address. 2102 The Proxy/Server next creates a neighbor cache entry for the Client 2103 using the base MNP-LLA as the network-layer address and with lifetime 2104 set to no more than the smallest prefix lifetime. Next, the Proxy/ 2105 Server updates the neighbor cache entry by recording the information 2106 in each Interface Attributes sub-option in the RS OMNI option. The 2107 Proxy/Server also records the actual OAL/*NET addresses in the 2108 neighbor cache entry. For RS messages encapsulated as carrier 2109 packets, the Proxy/Server also records the OAL Identification number 2110 as the starting value for the window of future packets it will send/ 2111 accept with its own ADM-ULA and the Client's MNP-ULA as the OAL 2112 addresses. (The Client in turn caches the Identification number as 2113 start value for future packets it will accept/send with its own MNP- 2114 ULA and the Proxy/Server's ADM-ULA as the OAL addresses.) 2116 Next, the Proxy/Server prepares an RA message using its ADM-LLA as 2117 the network-layer source address and the network-layer source address 2118 of the RS message as the network-layer destination address. The 2119 Proxy/Server sets the Router Lifetime to the time for which it will 2120 maintain both this underlying interface individually and the neighbor 2121 cache entry as a whole. The Proxy/Server also sets Cur Hop Limit, M 2122 and O flags, Reachable Time and Retrans Timer to values appropriate 2123 for the OMNI link. The Proxy/Server includes the MNPs, any other 2124 prefix management parameters and an OMNI option with no Interface 2125 Attributes but with an Origin Indication sub-option per 2126 [I-D.templin-6man-omni] with the mapped and obfuscated Port Number 2127 and IP address corresponding to the Client's own INET address in the 2128 case of INET Clients or to the Proxy/Server's INET-facing address for 2129 all other Clients. The Proxy/Server then includes one or more RIOs 2130 that encode the MSPs for the OMNI link, plus an MTU option (see 2131 Section 3.9). The Proxy/Server finally forwards the message to the 2132 Client using OAL encapsulation/fragmentation as necessary with an OAL 2133 Identification value that matches the RS. 2135 After the initial RS/RA exchange, the Proxy/Server maintains a 2136 ReachableTime timer for each of the Client's underlying interfaces 2137 individually (and for the Client's neighbor cache entry collectively) 2138 set to expire after ReachableTime seconds. If the Client (or Proxy) 2139 issues additional RS messages, the Proxy/Server sends an RA response 2140 and resets ReachableTime. If the Proxy/Server receives an ND message 2141 with a prefix release indication it sets the Client's neighbor cache 2142 entry to the DEPARTED state and withdraws the MNP from the routing 2143 system after a short delay (e.g., 2 seconds). If ReachableTime 2144 expires before a new RS is received on an individual underlying 2145 interface, the Proxy/Server marks the interface as DOWN. If 2146 ReachableTime expires before any new RS is received on any individual 2147 underlying interface, the Proxy/Server sets the neighbor cache entry 2148 state to STALE and sets a 10 second timer. If the Proxy/Server has 2149 not received a new RS or ND message with a prefix release indication 2150 before the 10 second timer expires, it deletes the neighbor cache 2151 entry and withdraws the MNP from the routing system. 2153 The Proxy/Server processes any ND messages pertaining to the Client 2154 and returns an NA/RA reply in response to solicitations. The Proxy/ 2155 Server may also issue unsolicited RA messages, e.g., with reconfigure 2156 parameters to cause the Client to renegotiate its prefix delegation/ 2157 registrations, with Router Lifetime set to 0 if it can no longer 2158 service this Client, etc. Finally, If the neighbor cache entry is in 2159 the DEPARTED state, the Proxy/Server deletes the entry after 2160 DepartTime expires. 2162 Note: Clients SHOULD notify former Proxy/Servers of their departures, 2163 but Proxy/Servers are responsible for expiring neighbor cache entries 2164 and withdrawing routes even if no departure notification is received 2165 (e.g., if the Client leaves the network unexpectedly). Proxy/Servers 2166 SHOULD therefore set Router Lifetime to ReachableTime seconds in 2167 solicited RA messages to minimize persistent stale cache information 2168 in the absence of Client departure notifications. A short Router 2169 Lifetime also ensures that proactive RS/RA messaging between Clients 2170 and Proxy/Servers will keep any NAT state alive (see above). 2172 Note: All Proxy/Servers on an OMNI link MUST advertise consistent 2173 values in the RA Cur Hop Limit, M and O flags, Reachable Time and 2174 Retrans Timer fields the same as for any link, since unpredictable 2175 behavior could result if different Proxy/Servers on the same link 2176 advertised different values. 2178 3.12.3.1. DHCPv6-Based Prefix Registration 2180 When a Client is not pre-provisioned with an MNP-LLA, it will need 2181 for the Proxy/Server to select one or more MNPs on its behalf and set 2182 up the correct state in the AERO routing service. (A Client with a 2183 pre-provisioned MNP may also request the Proxy/Server to select 2184 additional MNPs.) The DHCPv6 service [RFC8415] is used to support 2185 this requirement. 2187 When a Client needs to have the Proxy/Server select MNPs, it sends an 2188 RS message with source address set to the unspecified address (::) 2189 and with an OMNI option that includes a DHCPv6 message sub-option 2190 with DHCPv6 Prefix Delegation (DHCPv6-PD) parameters. When the 2191 Proxy/Server receives the RS message, it extracts the DHCPv6-PD 2192 message from the OMNI option. 2194 The Proxy/Server then acts as a "Proxy DHCPv6 Client" in a message 2195 exchange with the locally-resident DHCPv6 server, which delegates 2196 MNPs and returns a DHCPv6-PD Reply message. (If the Proxy/Server 2197 wishes to defer creation of MN state until the DHCPv6-PD Reply is 2198 received, it can instead act as a Lightweight DHCPv6 Relay Agent per 2199 [RFC6221] by encapsulating the DHCPv6-PD message in a Relay-forward/ 2200 reply exchange with Relay Message and Interface ID options.) 2202 When the Proxy/Server receives the DHCPv6-PD Reply, it adds a route 2203 to the routing system and creates an MNP-LLA based on the delegated 2204 MNP. The Proxy/Server then sends an RA back to the Client with the 2205 (newly-created) MNP-LLA as the destination address and with the 2206 DHCPv6-PD Reply message coded in the OMNI option. When the Client 2207 receives the RA, it creates a default route, assigns the Subnet 2208 Router Anycast address and sets its MNP-LLA based on the delegated 2209 MNP. 2211 Note: See [I-D.templin-6man-omni] for an MNP delegation alternative 2212 in which the Client can optionally avoid including a DHCPv6 message 2213 sub-option. Namely, when the Client requests a single MNP it can set 2214 the RS source to the unspecified address (::) and include a Node 2215 Identification sub-option and Preflen in the OMNI option (but with no 2216 DHCPv6 message sub-option). When the Proxy/Server receives the RS 2217 message, it forwards a self-generated DHCPv6 Solicit message to the 2218 DHCPv6 server on behalf of the Client. When the Proxy/Server 2219 receives the DHCPv6 Reply, it prepares an RA message with an OMNI 2220 option with Preflen information (but with no DHCPv6 message sub- 2221 option), then places the (newly-created) MNP-LLA in the RA 2222 destination address and returns the message to the Client. 2224 3.13. The AERO Proxy Function 2226 Clients connect to the OMNI link via Proxy/Servers, with one Proxy/ 2227 Server for each underlying interface. Each of the Client's Proxy/ 2228 Servers must be informed of all of the Client's additional underlying 2229 interfaces. For Clients on Direct and VPNed underlying interfaces 2230 the Proxy/Server "A" for that interface is directly connected, for 2231 Clients on ANET underlying interfaces Proxy/Server "A" is located on 2232 the ANET/INET boundary, and for Clients on INET underlying interfaces 2233 Proxy/Server "A" is located somewhere in the connected Internetwork. 2234 When the Client registers with Proxy/Server "A", it must also report 2235 the registration to any other Proxy/Servers for other underlying 2236 interfaces "B", "C", "D", etc. for which an underlying interface 2237 relationship has already been established. The Proxy/Server 2238 satisfies these requirements as follows: 2240 o when Proxy/Server "A" receives an RS message from a new Client, it 2241 first authenticates the message then examines the network-layer 2242 destination address. If the destination address is Proxy/Server 2243 "A"'s ADM-LLA or (link-local) All-Routers multicast, Proxy/Server 2244 "A" creates a proxy neighbor cache entry and caches the Client 2245 link-layer addresses along with the OMNI option information and 2246 any other identifying information including OAL Identification 2247 values, Client Identifiers, Nonce values, etc. If the RS message 2248 destination was the ADM-LLA of a different Proxy/Server "B" (or, 2249 if the OMNI option included an MS-Register sub-option with the 2250 ADM-LLA of a different Proxy/Server "B"), Proxy/Server "A" 2251 encapsulates a proxyed version of the RS message in an OAL header 2252 with source set to Proxy/Server "A"'s ADM-ULA and destination set 2253 to Proxy/Server "B"'s ADM-ULA. Proxy/Server "A" also includes an 2254 OMNI header with an Interface Attributes option that includes its 2255 own INET address plus a unique UDP Port Number for this Client, 2256 then forwards the message into the OMNI link spanning tree. 2257 (Note: including a unique Port Number allows Proxy/Server "B" to 2258 distinguish different Clients located behind the same Proxy/Server 2259 "A" at the link-layer, whereas the link-layer addresses would 2260 otherwise be indistinguishable.) 2262 o when the Proxy/Server "B" receives the RS, it authenticates the 2263 message then creates or updates a neighbor cache entry for the 2264 Client with Proxy/Server "A"'s ADM-ULA, INET address and UDP Port 2265 Number as the link-layer address information. Proxy/Server "B" 2266 then sends an RA message back to Proxy/Server "A" via the spanning 2267 tree. 2269 o when Proxy/Server "A" receives the RA, it authenticates the 2270 message and matches it with the proxy neighbor cache entry created 2271 by the RS. Proxy/Server "A" then caches the prefix information as 2272 a mapping from the Client's MNPs to the Client's link-layer 2273 address, caches the Proxy/Server's advertised Router Lifetime and 2274 sets the neighbor cache entry state to REACHABLE. The Proxy/ 2275 Server then optionally rewrites the Router Lifetime and forwards 2276 the (proxyed) message to the Client. The Proxy/Server finally 2277 includes an MTU option (if necessary) with an MTU to use for the 2278 underlying ANET interface. 2280 o The Client repeats this process with each Proxy/Server "B", "C", 2281 "D" for each of its additional underlying interfaces. 2283 After the initial RS/RA exchanges each Proxy/Server forwards any of 2284 the Client's carrier packets for which there is no matching neighbor 2285 cache entry to a Bridge using OAL encapsulation with its own ADM-ULA 2286 as the source and the MNP-ULA corresponding to the Client as the 2287 destination. The Proxy/Server instead forwards any carrier packets 2288 destined to a neighbor cache target directly to the target according 2289 to the OAL/link-layer information - the process of establishing 2290 neighbor cache entries is specified in Section 3.14. 2292 While the Client is still associated with each Proxy/Server "A", "A" 2293 can send NS, RS and/or unsolicited NA messages to update the neighbor 2294 cache entries of other AERO nodes on behalf of the Client and/or to 2295 convey QoS updates. This allows for higher-frequency Proxy-initiated 2296 RS/RA messaging over well-connected INET infrastructure supplemented 2297 by lower-frequency Client-initiated RS/RA messaging over constrained 2298 ANET data links. 2300 If any Proxy/Server "B", "C", "D" ceases to send solicited 2301 advertisements, Proxy/Server "A" sends unsolicited RAs to the Client 2302 with destination set to (link-local) All-Nodes multicast and with 2303 Router Lifetime set to zero to inform Clients that a Proxy/Server has 2304 failed. Although Proxy/Server "A" can engage in ND exchanges on 2305 behalf of the Client, the Client can also send ND messages on its own 2306 behalf, e.g., if it is in a better position than "A" to convey QoS 2307 changes, etc. The ND messages sent by the Client include the 2308 Client's MNP-LLA as the source in order to differentiate them from 2309 the ND messages sent by Proxy/Server "A". 2311 If the Client becomes unreachable over an underlying interface, 2312 Proxy/Server "A" sets the neighbor cache entry state to DEPARTED and 2313 retains the entry for DepartTime seconds. While the state is 2314 DEPARTED, Proxy/Server "A" forwards any carrier packets destined to 2315 the Client to a Bridge via OAL/ORH encapsulation. When DepartTime 2316 expires, Proxy/Server "A" deletes the neighbor cache entry and 2317 discards any further carrier packets destined to this (now forgotten) 2318 Client. 2320 In some ANETs that employ a Proxy/Server, the Client's MNP can be 2321 injected into the ANET routing system. In that case, the Client can 2322 send original IP packets without invoking the OAL so that the ANET 2323 routing system transports the original IP packets to the Proxy. This 2324 can be very beneficial, e.g., if the Client connects to the ANET via 2325 low-end data links such as some aviation wireless links. 2327 If the ANET first-hop access router is on the same underlying link as 2328 the Client and recognizes the AERO/OMNI protocol, the Client can 2329 avoid OAL encapsulation for both its control and data messages. When 2330 the Client connects to the link, it can send an unencapsulated RS 2331 message with source address set to its own MNP-LLA (or to a Temporary 2332 LLA), and with destination address set to the ADM-LLA of the Client's 2333 selected Proxy/Server or to (link-local) All-Routers multicast. The 2334 Client includes an OMNI option formatted as specified in 2335 [I-D.templin-6man-omni]. The Client then sends the unencapsulated RS 2336 message, which will be intercepted by the AERO-Aware access router. 2338 The ANET access router then performs OAL encapsulation on the RS 2339 message and forwards it to a Proxy/Server at the ANET/INET boundary. 2340 When the access router and Proxy/Server are one and the same node, 2341 the Proxy/Server would share and underlying link with the Client but 2342 its message exchanges with outside correspondents would need to pass 2343 through a security gateway at the ANET/INET border. The method for 2344 deploying access routers and Proxys (i.e. as a single node or 2345 multiple nodes) is an ANET-local administrative consideration. 2347 Note: The Proxy/Server can apply packing as discussed in 2348 [I-D.templin-6man-omni] if an opportunity arises to concatenate 2349 multiple original IP packets that will be destined to the same 2350 neighbor. 2352 3.13.1. Detecting and Responding to Proxy/Server Failures 2354 In environments where fast recovery from Proxy/Server failure is 2355 required, Proxy/Server "A" SHOULD use proactive Neighbor 2356 Unreachability Detection (NUD) to track peer Proxy/Server "B" 2357 reachability in a similar fashion as for Bidirectional Forwarding 2358 Detection (BFD) [RFC5880]. Proxy/Server "A" can then quickly detect 2359 and react to failures so that cached information is re-established 2360 through alternate paths. The NUD control messaging is carried only 2361 over well-connected ground domain networks (i.e., and not low-end 2362 aeronautical radio links) and can therefore be tuned for rapid 2363 response. 2365 Proxy/Server "A" performs proactive NUD with peer Proxy/Server "B" 2366 for which there are currently active Clients by sending continuous NS 2367 messages in rapid succession, e.g., one message per second. Proxy/ 2368 Server "A" sends the NS message via the spanning tree with its own 2369 ADM-LLA as the source and the ADM-LLA of the peer Proxy/Server "B" as 2370 the destination. When Proxy/Server "A" is also sending RS messages 2371 to the peer Proxy/Server "B" on behalf of ANET Clients, the resulting 2372 RA responses can be considered as equivalent hints of forward 2373 progress. This means that Proxy/Server "B" need not also send a 2374 periodic NS if it has already sent an RS within the same period. If 2375 the peer Proxy/Server "B" fails (i.e., if "A" ceases to receive 2376 advertisements), Proxy/Server "A" can quickly inform Clients by 2377 sending multicast RA messages on the ANET interface. 2379 Proxy/Server "A" sends RA messages on the ANET interface with source 2380 address set to Proxy/Server "B"'s address, destination address set to 2381 (link-local) All-Nodes multicast, and Router Lifetime set to 0. 2382 Proxy/Server "A" SHOULD send MAX_FINAL_RTR_ADVERTISEMENTS RA messages 2383 separated by small delays [RFC4861]. Any Clients on the ANET that 2384 had been using the failed Proxy/Server "B" will receive the RA 2385 messages and associate with a new Proxy/Server. 2387 3.13.2. Point-to-Multipoint Proxy/Server Coordination 2389 In environments where Client messaging over ANETs is bandwidth- 2390 limited and/or expensive, Clients can enlist the services of Proxy/ 2391 Server "A" to coordinate with multiple Proxy/Servers "B", "C", "D" 2392 etc. in a single RS/RA message exchange. The Client can send a 2393 single RS message to (link-local) All-Routers multicast that includes 2394 the ID's of multiple Proxy/Servers in MS-Register sub-options of the 2395 OMNI option. 2397 When Proxy/Server "A" receives the RS and processes the OMNI option, 2398 it sends a separate RS to each MS-Register Proxy/Server ID. When 2399 Proxy/Server "A" receives an RA, it can optionally return an 2400 immediate "singleton" RA to the Client or record the Proxy/Server's 2401 ID for inclusion in a pending "aggregate" RA message. Proxy/Server 2402 "A" can then return aggregate RA messages to the Client including 2403 multiple Proxy/Server IDs in order to conserve bandwidth. Each RA 2404 includes a proper subset of the Proxy/Server IDs from the original RS 2405 message, and Proxy/Server "A" must ensure that the message contents 2406 of each RA are consistent with the information received from the 2407 (aggregated) additional Proxy/Servers. 2409 Clients can thereafter employ efficient point-to-multipoint Proxy/ 2410 Server coordination under the assistance of Proxy/Server "A" to 2411 reduce the number of messages sent over the ANET while enlisting the 2412 support of multiple Proxy/Servers for fault tolerance. Clients can 2413 further include MS-Release sub-options in IPv6 ND messages to request 2414 Proxy/Server "A" to release from former Proxy/Servers via the 2415 procedures discussed in Section 3.16.5. 2417 The OMNI interface specification [I-D.templin-6man-omni] provides 2418 further discussion of the RS/RA messaging involved in point-to- 2419 multipoint coordination. 2421 3.14. AERO Route Optimization 2423 While AERO nodes can always send data packets over strict spanning 2424 tree paths, route optimization should be performed while carrier 2425 packets are flowing between a source and target node. Route 2426 optimization is based on asymmetric IPv6 ND Address Resolution 2427 messaging between a Route Optimization Source (ROS) and Route 2428 Optimization Responder (ROR), and later extended to the target using 2429 IPv6 ND Neighbor Unreachability Detection messaging. Route 2430 optimization is initiated by the first eligible ROS closest to the 2431 source as follows: 2433 o For Clients on VPNed and Direct interfaces, the Proxy/Server is 2434 the ROS. 2436 o For Clients on ANET interfaces, either the Client or the Proxy/ 2437 Server may be the ROS. 2439 o For Clients on INET interfaces, the Client itself is the ROS. 2441 o For correspondent nodes on INET/EUN interfaces serviced by a 2442 Relay, the Relay is the ROS. 2444 The route optimization procedure is conducted between the ROS and the 2445 target Proxy/Server/Relay acting as an ROR (the target may either be 2446 a MNP Client serviced by a Proxy/Server or a non-MNP correspondent 2447 reachable via a Relay). Note that in this arrangement the ROS is 2448 always the Client or Proxy/Server/Relay nearest the source over the 2449 selected source underlying interface, while the ROR is always a 2450 Proxy/Server/Relay that services the target regardless of the target 2451 underlying interface. 2453 The procedures are specified in the following sections. 2455 3.14.1. Route Optimization Initiation 2457 When an original IP packet from a source node destined to a target 2458 node arrives, the ROS checks for a neighbor cache entry for an MNP- 2459 LLA that matches the target destination. If there is a neighbor 2460 cache entry in the REACHABLE state, the ROS invokes the OAL and 2461 forwards the resulting carrier packets according to the cached state 2462 and returns from processing. Otherwise, if there is no neighbor 2463 cache entry the ROS creates one in the INCOMPLETE state. 2465 The ROS next places the original IP packet on a short queue then 2466 sends an NS message for Address Resolution (NS(AR)) to receive a 2467 solicited NA(AR) message from a ROR. The NS(AR) message must be no 2468 larger than the minimum/path MPS so that its entire contents will fit 2469 in an OAL first fragment (i.e., as an atomic fragment). The ROS 2470 prepares an NS(AR) that includes: 2472 o the LLA of the ROS as the source address. 2474 o the MNP-LLA corresponding to the original IP packet's destination 2475 as the Target Address, e.g., for 2001:db8:1:2::10:2000 the Target 2476 Address is fe80::2001:db8:1:2. 2478 o the Solicited-Node multicast address [RFC4291] formed from the 2479 lower 24 bits of the original IP packet's destination as the 2480 destination address, e.g., for 2001:db8:1:2::10:2000 the NS 2481 destination address is ff02:0:0:0:0:1:ff10:2000. 2483 The NS(AR) message also includes an OMNI option with an Interface 2484 Attributes entry for the sending interface, and with Preflen set to 2485 the prefix length associated with the NS(AR) source. The ROS then 2486 submits the NS(AR) message for OAL encapsulation and fragmentation, 2487 with OAL source set to its own ULA and OAL destination set to the ULA 2488 corresponding to the target, and with an unpredictable initial 2489 Identification value. The ROS caches the initial Identification 2490 value in the (newly-created) neighbor cache entry as the starting 2491 sequence number for the "send" window for future carrier packets sent 2492 to this target via the responding ROR. 2494 The ROS then sends the resulting carrier packet into the spanning 2495 tree without decrementing the network-layer TTL/Hop Limit field. 2496 (When the ROS is an INET Client, it instead must first sign the 2497 NS(AR) message and send the resulting carrier packet to the ADM-ULA 2498 of one of its current Proxy/Servers which then verifies the NS(AR) 2499 signature and forwards the carrier packet into the spanning tree on 2500 behalf of the Client.) 2502 3.14.2. Relaying the NS(AR) *NET Packet(s) 2504 When the Bridge receives the carrier packet containing the RS from 2505 the ROS, it discards the *NET headers and determines the next hop by 2506 consulting its standard IPv6 forwarding table for the OAL header 2507 destination address. The Bridge then decrements the OAL header Hop- 2508 Limit, re-encapsulates the carrier packet and forwards it via the 2509 spanning tree the same as for any IPv6 router, where it may traverse 2510 multiple OMNI link segments. The final-hop Bridge in the spanning 2511 tree will deliver the carrier packet via a secured tunnel to a Proxy/ 2512 Server or Relay that services the target. 2514 3.14.3. Processing the NS(AR) and Sending the NA(AR) 2516 When the target Proxy/Server (or Relay) receives the carrier packet, 2517 it examines the enclosed atomic fragment to determine that it 2518 contains an NS(AR) then examines the NS(AR) target to determine 2519 whether it has a matching neighbor cache entry and/or non-MNP route. 2520 If there is no match, the Proxy/Server drops the message. Otherwise, 2521 the Proxy/Server/Relay continues processing as follows: 2523 o if the NS(AR) target matches a Client neighbor cache entry in the 2524 DEPARTED state, the Proxy/Server inserts an ORH with destination 2525 prefix set to the lower 64 bits of the Client's MNP-ULA and sets 2526 the destination address to the ADM-ULA of the Client's new Proxy/ 2527 Server. The (old) Proxy/Server then re-encapsulates the carrier 2528 packet, forwards it into the spanning tree and returns from 2529 processing. 2531 o If the NS(AR) target matches the MNP-LLA of a Client neighbor 2532 cache entry in the REACHABLE state, the Proxy/Server acts as an 2533 ROR to provide route optimization information on behalf of the 2534 Client. 2536 o If the NS(AR) target matches one of its non-MNP routes, the Relay 2537 acts as an ROR since it serves as a router to forward IP packets 2538 toward the (fixed network) target at the network layer. 2540 The ROR next checks the target neighbor cache entry for a Report List 2541 entry that matches the NS(AR) source LLA/ULA of the ROS. If there is 2542 a Report List entry, the ROR accepts the NS(AR) only if the OAL 2543 Identification value is within the "accept" window for this ROS or if 2544 the NS(AR) was forwarded over the secured spanning tree. If the 2545 NS(AR) is authentic and the OAL Identification is outside of the 2546 current "accept" window for this ROS, the ROR resets the current 2547 "accept" window start to the new OAL Identification value. If the 2548 NS(AR) is authentic but the target neighbor cache entry does not 2549 already include a Report List entry for this ROS, the ROR creates a 2550 new entry and caches the ROS information. The Report List cache 2551 entry therefore includes the LLA and ULA of the ROS, the new "accept" 2552 Identification number for the ROS and the previous "accept" 2553 Identification number in case any packets sent under the previous 2554 window are still in flight. 2556 The ROR then prepares a (solicited) NA(AR) message to send back to 2557 the ROS using the same Identification value received in the NS(AR) 2558 (unlike the NS(AR), the NA(AR) need not fit in a single OAL 2559 fragment). The ROR sets the NA(AR) source address to the target's 2560 MNP-LLA, sets the destination address to the NS(AR) LLA source 2561 address and sets the Target Address to the same value that appeared 2562 in the NS(AR). The ROR then includes an OMNI option with Preflen set 2563 to the prefix length associated with the NA(AR) source address. If 2564 the NS(AR) target was an MNP Client, the ROR next includes Interface 2565 Attributes in the OMNI option for each of the target's underlying 2566 interfaces with current information for each interface and includes 2567 an authentication signature if necessary. The ROR then sets the S/ 2568 T-ifIndex field in the OMNI header to 0. 2570 For each Interface Attributes sub-option, the ROR sets the L2ADDR 2571 according to the Proxy/Server's INET address for VPNed or Direct 2572 interfaces, to the INET address of the Proxy/Server for proxyed 2573 interfaces or to the Client's INET address for INET interfaces. The 2574 ROR then includes the lower 32 bits of the Proxy/Server's ADM-ULA as 2575 the LHS, encodes the ADM-ULA prefix length code in the SRT field and 2576 sets the FMT code accordingly as specified in Section 3.3. 2578 The ROR then sets the NA(AR) message R flag to 1 (as a router) and S 2579 flag to 1 (as a response to a solicitation) and sets the O flag to 0 2580 (as a proxy). The ROR finally submits the NA(AR) for OAL 2581 encapsulation with source set to its own ULA and destination set to 2582 the ULA of the ROS, then performs OAL fragmentation using the same 2583 Identification value that appeared in the NS(AR) and forwards the 2584 resulting (*NET-encapsulated) carrier packets via the spanning tree 2585 without decrementing the network-layer TTL/Hop Limit field. 2587 3.14.4. Relaying the NA(AR) 2589 When the Bridge receives the carrier packets from the ROR, it 2590 discards the *NET header and determines the next hop by consulting 2591 its standard IPv6 forwarding table for the OAL header destination 2592 address. The Bridge then decrements the OAL header Hop-Limit, re- 2593 encapsulates the carrier packet and forwards it via the spanning tree 2594 the same as for any IPv6 router, where it may traverse multiple OMNI 2595 link segments. The final-hop Bridge in the spanning tree will 2596 deliver the carrier packet via a secured tunnel to a Proxy/Server for 2597 the ROS. 2599 3.14.5. Processing the NA(AR) 2601 When the ROS receives the NA(AR) message, it first searches for a 2602 neighbor cache entry that matches the NA(AR) LLA source address. If 2603 there is an entry in the INCOMPLETE or STALE state, the ROS matches 2604 the OAL Identification value with the value it had included in the 2605 corresponding NS(AR). If the values match, the ROS processes the 2606 message the same as for standard IPv6 Address Resolution [RFC4861]. 2607 In the process, it caches the NA(AR) LLA source address and all 2608 information found in the OMNI option in the neighbor cache entry for 2609 the target. The ROS finally sets the neighbor cache entry state to 2610 REACHABLE and sets its lifetime to ReachableTime seconds. (When the 2611 ROS is a Client, the solicited NA(AR) message will first be delivered 2612 via the spanning tree to one of its current Proxy/Servers, which then 2613 securely forwards the message to the Client. If the Client is on an 2614 ANET, ANET physical security and protected spectrum ensures security; 2615 if the Client is on the open ANET, the Proxy/Server must include an 2616 authentication signature.) 2618 3.14.6. Route Optimization Maintenance 2620 Following route optimization, the ROS forwards future carrier packets 2621 with user data destined to the target via the addresses found in the 2622 cached link-layer information and with a monotonically-incrementing 2623 Identification value for each OAL packet. The route optimization is 2624 shared by all sources that send original IP packets to the target via 2625 the ROS, i.e., and not just the source on behalf of which the route 2626 optimization was initiated. Note that route optimization is 2627 performed only for original IP packets that contain user data, and 2628 not for those that contain other IPv6 ND control messages. 2630 While the ROS continues to forward additional original IP packets 2631 destined to the target, it sends additional NS(AR) messages to the 2632 ROR before ReachableTime expires to receive a fresh NA(AR) message 2633 the same as described in the previous sections (route optimization 2634 refreshment strategies are an implementation matter, with a non- 2635 normative example given in Appendix A.1). The ROS may supply a new 2636 unpredictable Identification value if it wishes to reset the 2637 neighbor's "accept" Identification window. If the ROS is an INET 2638 Client, it must include an authentication signature with the NS(AR) 2639 message so that the Proxy/Server can authenticate. 2641 The ROS uses its own ULA as the NS(AR) OAL source address and the ULA 2642 of the ROR as the NS(AR) OAL destination address, and sends up to 2643 MAX_MULTICAST_SOLICIT NS(AR) messages separated by 1 second until an 2644 NA(AR) is received. If no NA(AR) is received, the ROS assumes that 2645 the current ROR has become unreachable and deletes the target 2646 neighbor cache entry. Subsequent original IP packets will trigger a 2647 new route optimization event (see: Section 3.14.1) which may discover 2648 a different ROR that services the same target. 2650 If an NA(AR) is received, the ROS then updates the neighbor cache 2651 entry for the target to refresh ReachableTime, while (for MNP 2652 targets) the ROR adds or updates the ROS address to the target's 2653 Report List and with time set to ReportTime. While no carrier 2654 packets are flowing, the ROS instead allows ReachableTime for the 2655 target neighbor cache entry to expire. When ReachableTime expires, 2656 the ROS places the target neighbor cache entry back in the STALE 2657 state. Any future carrier packets flowing through the ROS will again 2658 trigger a new route optimization. 2660 The ROS may also receive unsolicited NA (uNA) messages from the ROR 2661 at any time (see: Section 3.16). If there is a neighbor cache entry 2662 for the target and the carrier packet(s) containing the uNA is 2663 received securely, the ROS updates the link-layer information but 2664 does not update ReachableTime since the receipt of a uNA does not 2665 confirm that any forward paths are working. If there is no neighbor 2666 cache entry or the message cannot be authenticated, the ROS simply 2667 discards the uNA. 2669 In this arrangement, the ROS holds a neighbor cache entry for the 2670 target in the REACHABLE state with a "send" Identification window 2671 starting value, while the ROR's target neighbor cache entry holds a 2672 Report List entry for the ROS with an "accept" Identification window 2673 starting value for the ULA of the ROS. The route optimization 2674 neighbor relationship is therefore asymmetric and unidirectional. If 2675 the target node also has carrier packets to send back to the source 2676 node, then a separate route optimization procedure is performed in 2677 the reverse direction, but there is no requirement that the forward 2678 and reverse paths be symmetric. 2680 3.15. Neighbor Unreachability Detection (NUD) 2682 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2683 [RFC4861] either reactively in response to persistent link-layer 2684 errors (see Section 3.11) or on-demand to confirm reachability and/or 2685 initiate route optimizations from the ROS to the target via the ROR. 2686 The NUD algorithm is based on periodic control message exchanges. 2687 The algorithm may further be seeded by ND hints of forward progress, 2688 but care must be taken to avoid inferring reachability based on 2689 spoofed information. For example, authentic IPv6 ND message 2690 exchanges may be considered as acceptable hints of forward progress, 2691 while spurious random packets should not be. 2693 AERO nodes can use standard NS/NA(NUD) exchanges sent over the OMNI 2694 link spanning tree to securely test reachability without risk of DoS 2695 attacks from nodes pretending to be a neighbor. These NS/NA(NUD) 2696 messages use the unicast LLAs and ULAs of the two parties involved in 2697 the NUD test the same as for standard IPv6 ND over the secured 2698 spanning tree, however a means for an ROS to test the unsecured 2699 target route optimized paths is also necessary. 2701 When an ROR directs an ROS to a target neighbor with one or more 2702 link-layer addresses, the ROS performs extended route optimization 2703 for each unsecured target underlying interface either proactively or 2704 on-demand of carrier packets directed to the path by multilink 2705 forwarding. The route optimization is performed through secured 2706 NS(NUD) messages over the spanning tree in the forward path that 2707 invoke an unsecured NA(NUD) reply over the target underlying 2708 interface on return path. (The NS(NUD) messages must therefore 2709 include Identification values (and optionally Nonce and Timestamp 2710 options) that will be echoed in the unsecured NA(NUD) replies.) 2711 While testing a target underlying interface, the ROS can optionally 2712 continue to send carrier packets via the ROR or maintain a small 2713 queue of carrier packets until target reachability is confirmed. 2715 When the ROS sends an NS(NUD) message, it sets the IPv6 source to its 2716 own LLA and sets both the destination and Target Address to the LLA 2717 of the target. The ROS also includes an OMNI option with a single 2718 Interface Attributes sub-option with the SRT, FMT, LHS and L2ADDR 2719 information for its own underlying interface it wishes to test, but 2720 sets the S/T-ifIndex field to the index for target's underlying 2721 interface to be tested. The ROS includes an Identification value 2722 within the current "send" window for this ROR (and optionally Nonce 2723 and Timestamp options), then encapsulates the message in OAL/INET 2724 headers with its own ULA as the source and the ULA of the ROR as the 2725 destination. The ROS then forwards the NS(NUD) message toward the 2726 ROR (and ultimately the target itself) via the spanning tree. 2728 When the ROR receives the NS(NUD) message, it examines the S/ 2729 T-ifIndex field to determine the underlying interface target of the 2730 NS(NUD) test. If the underlying interface is a Direct or VPNed 2731 interface, the ROR acts as the target. If the underlying interface 2732 is a Proxyed interface, the ROR changes the OAL destination to the 2733 ULA of the Proxy and forwards the NS(NUD) to the Proxy which either 2734 acts as the target itself or forwards the message to the target 2735 Client. If the underlying interface is an INET interface, the ROR 2736 changes the OAL destination address to the ULA of the target Client 2737 and forwards the NS(NUD) over the underlying interface to the target 2738 Client while including an authentication signature. 2740 The target then creates a neighbor cache entry for the ROS LLA 2741 address if necessary and caches the Identification value as the start 2742 of the "accept" window for this ROS. The target then creates an 2743 NA(NUD) by reversing the OAL and IPv6 addresses from the NS(NUD) 2744 while copying the Identification value, and next including an 2745 Interface Attributes sub-option with attributes for its own interface 2746 identified by the NS(NUD) S/T-ifIndex. The target sets the NA(NUD) 2747 S/T-ifIndex to the index of the ROS, sets the Target Address to the 2748 same value that was in the NS(NUD), sets R flag to 1, the S flag to 0 2749 and the O flag to 1, and returns the message using its own underlying 2750 interface identified by NS(NUD) S/T-ifIndex and destined to the ROS's 2751 interface identified by the original Interface Attributes sub-option. 2753 When the ROS receives the NA(NUD) message, it can determine from the 2754 Identification value and Target Address (and optionally the Nonce and 2755 Timestamp) that the message matched its NS(NUD) and that it transited 2756 the direct path from the ROR using the selected underlying interface 2757 pair. The ROS marks route optimization target paths that pass these 2758 NUD tests as "reachable", and those that do not as "unreachable". 2759 These markings inform the OMNI interface forwarding algorithm 2760 specified in Section 3.10. 2762 Note: If the target determines that the OMNI option Interface 2763 Attributes in the NS(NUD) is located in a different OMNI link segment 2764 than its own interface named in the S/T-ifIndex, it instead returns 2765 the NA(NUD) via the spanning tree while including an ORH and setting 2766 the OAL destination address to the Subnet Router Anycast address used 2767 by Bridges on the ROS segment. When a Bridge on the ROS segment 2768 receives the NA(NUD), it replaces the Interface Attributes with 2769 information for its own interface while using the ifIndex value 2770 specific to the target. 2772 3.16. Mobility Management and Quality of Service (QoS) 2774 AERO is a Distributed Mobility Management (DMM) service. Each Proxy/ 2775 Server is responsible for only a subset of the Clients on the OMNI 2776 link, as opposed to a Centralized Mobility Management (CMM) service 2777 where there is a single network mobility collective entity for all 2778 Clients. Clients coordinate with their associated Proxy/Servers via 2779 RS/RA exchanges to maintain the DMM profile, and the AERO routing 2780 system tracks all current Client/Proxy/Server peering relationships. 2782 Proxy/Servers provide default routing and mobility/multilink services 2783 for their dependent Clients. Clients are responsible for maintaining 2784 neighbor relationships with their Proxy/Servers through periodic RS/ 2785 RA exchanges, which also serves to confirm neighbor reachability. 2786 When a Client's underlying interface address and/or QoS information 2787 changes, the Client is responsible for updating the Proxy/Server with 2788 this new information. Note that when there is a Proxy/Server in the 2789 path, the Proxy function can also perform some RS/RA exchanges on the 2790 Client's behalf. 2792 Mobility management messaging is based on the transmission and 2793 reception of unsolicited Neighbor Advertisement (uNA) messages. Each 2794 uNA message sets the IPv6 destination address to (link-local) All- 2795 Nodes multicast to convey a general update of Interface Attributes to 2796 (possibly) multiple recipients, or to a specific unicast LLA to 2797 announce a departure event to a specific recipient. Implementations 2798 must therefore examine the destination address to determine the 2799 nature of the mobility event (i.e., update vs departure). 2801 Mobility management considerations are specified in the following 2802 sections. 2804 3.16.1. Mobility Update Messaging 2806 RORs accommodate Client mobility, multilink and/or QoS change events 2807 by sending unsolicited NA (uNA) messages to each ROS in the target 2808 Client's Report List. When an ROR sends a uNA message, it sets the 2809 IPv6 source address to the its own LLA, sets the destination address 2810 to (link-local) All-Nodes multicast and sets the Target Address to 2811 the Client's MNP-LLA. The ROR also includes an OMNI option with 2812 Preflen set to the prefix length associated with the Client's MNP- 2813 LLA, with Interface Attributes for the target Client's underlying 2814 interfaces and with the OMNI header S/T-ifIndex set to 0. The ROR 2815 then sets the NA R flag to 1, the S flag to 0 and the O flag to 1, 2816 then encapsulates the message in an OAL header with source set to its 2817 own ULA and destination set to the ULA of the ROS and sends the 2818 message into the spanning tree. 2820 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2821 reception of uNA messages is unreliable but provides a useful 2822 optimization. In well-connected Internetworks with robust data links 2823 uNA messages will be delivered with high probability, but in any case 2824 the Proxy/Server can optionally send up to MAX_NEIGHBOR_ADVERTISEMENT 2825 uNAs to each ROS to increase the likelihood that at least one will be 2826 received. 2828 When the ROS receives a uNA message prepared as above, it ignores the 2829 message if there is no existing neighbor cache entry for the target 2830 Client. Otherwise, it uses the included OMNI option information to 2831 update the neighbor cache entry, but does not reset ReachableTime 2832 since the receipt of an unsolicited NA message from the ROR does not 2833 provide confirmation that any forward paths to the target Client are 2834 working. 2836 If uNA messages are lost, the ROS may be left with stale address and/ 2837 or QoS information for the Client for up to ReachableTime seconds. 2838 During this time, the ROS can continue sending carrier packets 2839 according to its stale neighbor cache information. When 2840 ReachableTime is close to expiring, the ROS will re-initiate route 2841 optimization and receive fresh link-layer address information. 2843 In addition to sending uNA messages to the current set of ROSs for 2844 the Client, the ROR also sends uNAs to the MNP-ULA associated with 2845 the link-layer address for any underlying interface for which the 2846 link-layer address has changed. These uNA messages update an old 2847 Proxy/Server that cannot easily detect (e.g., without active probing) 2848 when a formerly-active Client has departed. When the ROR sends the 2849 uNA, it sets the IPv6 source address to its LLA, sets the destination 2850 address to the old Proxy/Server's ADM-LLA, and sets the Target 2851 Address to the Client's MNP-LLA. The ROR also includes an OMNI 2852 option with Preflen set to the prefix length associated with the 2853 Client's MNP-LLA, with Interface Attributes for the changed 2854 underlying interface, and with the OMNI header S/T-ifIndex set to its 2855 own omIndex if the ROR is a Client or 0 if the ROR is a Proxy/Server. 2856 The ROR then sets the NA R flag to 1, the S flag to 0 and the O flag 2857 to 1, then encapsulates the message in an OAL header with source set 2858 to its own ULA and destination set to the ADM-ULA of the old Proxy/ 2859 Server and sends the message into the spanning tree. 2861 3.16.2. Announcing Link-Layer Address and/or QoS Preference Changes 2863 When a Client needs to change its underlying interface addresses and/ 2864 or QoS preferences (e.g., due to a mobility event), the Client 2865 requests one of its Proxy/Servers to send RS messages to all of its 2866 other Proxy/Servers via the spanning tree with an OMNI option that 2867 includes Interface attributes with the new link quality and address 2868 information. 2870 Up to MAX_RTR_SOLICITATIONS RS messages MAY be sent in parallel with 2871 sending carrier packets containing user data in case one or more RAs 2872 are lost. If all RAs are lost, the Client SHOULD re-associate with a 2873 new Proxy/Server. 2875 When the Proxy/Server receives the Client's changes, it sends uNA 2876 messages to all nodes in the Report List the same as described in the 2877 previous section. 2879 3.16.3. Bringing New Links Into Service 2881 When a Client needs to bring new underlying interfaces into service 2882 (e.g., when it activates a new data link), it sends an RS message to 2883 the Proxy/Server via the underlying interface with an OMNI option 2884 that includes Interface Attributes with appropriate link quality 2885 values and with link-layer address information for the new link. 2887 3.16.4. Deactivating Existing Links 2889 When a Client needs to deactivate an existing underlying interface, 2890 it sends an RS or uNA message to its Proxy/Server with an OMNI option 2891 with appropriate Interface Attribute values - in particular, the link 2892 quality value 0 assures that neighbors will cease to use the link. 2894 If the Client needs to send RS/uNA messages over an underlying 2895 interface other than the one being deactivated, it MUST include 2896 Interface Attributes with appropriate link quality values for any 2897 underlying interfaces being deactivated. 2899 Note that when a Client deactivates an underlying interface, 2900 neighbors that have received the RS/uNA messages need not purge all 2901 references for the underlying interface from their neighbor cache 2902 entries. The Client may reactivate or reuse the underlying interface 2903 and/or its ifIndex at a later point in time, when it will send RS/uNA 2904 messages with fresh Interface Attributes to update any neighbors. 2906 3.16.5. Moving Between Proxy/Servers 2908 The Client performs the procedures specified in Section 3.12.2 when 2909 it first associates with a new Proxy/Server or renews its association 2910 with an existing Proxy/Server. The Client also includes MS-Release 2911 identifiers in the RS message OMNI option per [I-D.templin-6man-omni] 2912 if it wants the new Proxy/Server to notify any old Proxy/Servers from 2913 which the Client is departing. 2915 When the new Proxy/Server receives the Client's RS message, it 2916 returns an RA as specified in Section 3.12.3 and sends up to 2917 MAX_NEIGHBOR_ADVERTIISEMENT uNA messages to any old Proxy/Servers 2918 listed in OMNI option MS-Release identifiers. When the new Proxy/ 2919 Server sends a uNA message, it sets the IPv6 source address to the 2920 Client's MNP-LLA, sets the destination address to the old Proxy/ 2921 Server's ADM-LLA, and sets the Target Address to the Client's LLA. 2922 The new Proxy/Server also includes an OMNI option with Preflen set to 2923 the prefix length associated with the Client's MNP-LLA, with 2924 Interface Attributes for its own underlying interface, and with the 2925 OMNI header S/T-ifIndex set to 0. The new Proxy/Server then sets the 2926 NA R flag to 1, the S flag to 0 and the O flag to 1, then 2927 encapsulates the message in an OAL header with source set to its own 2928 ADM-ULA and destination set to the ADM-ULA of the old Proxy/Server 2929 and sends the message into the spanning tree. 2931 When an old Proxy/Server receives the uNA, it changes the Client's 2932 neighbor cache entry state to DEPARTED, sets the link-layer address 2933 of the Client to the new Proxy/Server's ADM-ULA, and resets 2934 DepartTime. After a short delay (e.g., 2 seconds) the old Proxy/ 2935 Server withdraws the Client's MNP from the routing system. After 2936 DepartTime expires, the old Proxy/Server deletes the Client's 2937 neighbor cache entry. 2939 The old Proxy/Server also iteratively forwards a copy of the uNA 2940 message to each ROS in the Client's Report List by changing the OAL 2941 destination address to the ULA of the ROS while leaving all other 2942 fields of the message unmodified. When the ROS receives the uNA, it 2943 examines the Target address to determine the correct neighbor cache 2944 entry and verifies that the IPv6 destination address matches the old 2945 Proxy/Server. The ROS then caches the IPv6 source address as the new 2946 Proxy/Server for the existing neighbor cache entry and marks the 2947 entry as STALE. While in the STALE state, the ROS allows new carrier 2948 packets to flow according to any existing cached link-layer 2949 information and sends new NS(AR) messages using its own ULA as the 2950 OAL source and the ADM-ULA of the new Proxy/Server as the OAL 2951 destination address to elicit NA messages that reset the neighbor 2952 cache entry state to REACHABLE. If no new NA message is received for 2953 10 seconds while in the STALE state, the ROS deletes the neighbor 2954 cache entry. 2956 Clients SHOULD NOT move rapidly between Proxy/Servers in order to 2957 avoid causing excessive oscillations in the AERO routing system. 2958 Examples of when a Client might wish to change to a different Proxy/ 2959 Server include a Proxy/Server that has gone unreachable, topological 2960 movements of significant distance, movement to a new geographic 2961 region, movement to a new OMNI link segment, etc. 2963 When a Client moves to a new Proxy/Server, some of the fragments of a 2964 multiple fragment OAL packet may have already arrived at the old 2965 Proxy/Server while others are en route to the new Proxy/Server, 2966 however no special attention in the reassembly algorithm is necessary 2967 since all fragments will eventually be delivered to the Client which 2968 can then reassemble. 2970 3.17. Multicast 2972 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2973 [RFC3810] proxy service for its EUNs and/or hosted applications 2974 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2975 underlying interfaces for which group membership is required. The 2976 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2977 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2978 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2979 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2980 Designated Router (DR) [RFC7761]. AERO Relays also act as PIM 2981 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2982 INET/EUN networks. The behaviors identified in the following 2983 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2984 Multicast (ASM) operational modes. 2986 3.17.1. Source-Specific Multicast (SSM) 2988 When an ROS "X" acting as PIM router receives a Join/Prune message 2989 from a node on its downstream interfaces containing one or more 2990 ((S)ource, (G)roup) pairs, it updates its Multicast Routing 2991 Information Base (MRIB) accordingly. For each S belonging to a 2992 prefix reachable via X's non-OMNI interfaces, X then forwards the (S, 2993 G) Join/Prune to any PIM routers on those interfaces per [RFC7761]. 2995 For each S belonging to a prefix reachable via X's OMNI interface, X 2996 includes a PIM Join/Prune for each (S,G) in the OMNI option of an 2997 NS(AR) message (see: Section 3.14) using its own LLA as the source 2998 address and ALL-PIM-ROUTERS as the destination address. X then 2999 encapsulates the NS(AR) in an OAL header with source address set to 3000 the ULA of X and destination address set to S then forwards the 3001 message into the spanning tree, which delivers it to ROR "Y" that 3002 services S. The resulting NA(AR) will return the LLA for the prefix 3003 that matches S as the network-layer source address and with an OMNI 3004 option with the ULA corresponding to any underlying interfaces that 3005 are currently servicing S. 3007 When Y processes the NS(AR) it examines the PIM Join/Prune message. 3008 If S is located behind any INET, Direct, or VPNed interfaces Y acts 3009 as a PIM router and updates its MRIB to list X as the next hop in the 3010 reverse path. If S is located behind any Proxys "Z"*, X then sends 3011 an NS(NUD) message containing the PIM message to each Z* via Y with 3012 addressing and encapsulation details the same as specified in 3013 Section 3.15. Each Z* then updates its MRIB accordingly and 3014 maintains the LLA of X as the next hop in the reverse path. Since 3015 the Bridges do not examine network layer control messages, this means 3016 that the (reverse) multicast tree path is simply from each Z* (and/or 3017 Y) to X with no other multicast-aware routers in the path. 3019 Following the initial combined Join/Prune and NS/NA messaging, X 3020 maintains a neighbor cache entry for each S the same as if X was 3021 sending unicast data traffic to S. In particular, X performs 3022 additional NS/NA exchanges to keep the neighbor cache entry alive for 3023 up to t_periodic seconds [RFC7761]. If no new Joins are received 3024 within t_periodic seconds, X allows the neighbor cache entry to 3025 expire. Finally, if X receives any additional Join/Prune messages 3026 for (S,G) it forwards the messages in NS/NA exchanges with each Y and 3027 Z* in the neighbor cache entry over the spanning tree. 3029 At some later time, Client C that holds an MNP for source S may 3030 depart from a first Proxy/Server Z1 and/or connect via a new Proxy/ 3031 Server Z2. In that case, Y sends a uNA message to X the same as 3032 specified for unicast mobility in Section 3.16. When X receives the 3033 uNA message, it updates its neighbor cache entry for the LLA for 3034 source S and sends new Join messages to any new Proxys Z2. There is 3035 no requirement to send any Prune messages to old Proxy/Server Z1 3036 since source S will no longer source any multicast data traffic via 3037 Z1. Instead, the multicast state for (S,G) in Proxy/Server Z1 will 3038 soon time out since no new Joins will arrive. 3040 After some later time, C may move to a new Proxy/Server Y2 and depart 3041 from old Sever Y1. In that case, Y1 sends Join messages for any of 3042 C's active (S,G) groups to Y2 while including its own LLA as the 3043 source address. This causes Y2 to include Y1 in the multicast 3044 forwarding tree during the interim time that Y1's neighbor cache 3045 entry for C is in the DEPARTED state. At the same time, Y1 sends a 3046 uNA message to X with an OMNI option with S/T-ifIndex in the header 3047 set to 0 and a release indication to cause X to release its neighbor 3048 cache entry for S. X then sends a new Join message to S via the 3049 spanning tree and re-initiates route optimization the same as if it 3050 were receiving a fresh Join message from a node on a downstream link. 3052 3.17.2. Any-Source Multicast (ASM) 3054 When an ROS X acting as a PIM router receives a Join/Prune from a 3055 node on its downstream interfaces containing one or more (*,G) pairs, 3056 it updates its Multicast Routing Information Base (MRIB) accordingly. 3057 X then forwards a copy of the message within the OMNI option of an 3058 NS(AR) message to the Rendezvous Point (RP) R for each G over the 3059 spanning tree. X uses its own LLA as the source address and ALL-PIM- 3060 ROUTERS as the destination address, then encapsulates the NS(AR) 3061 message in an OAL header with source address set to the ULA of X and 3062 destination address set to R, then sends the message into the 3063 spanning tree. 3065 For each source S that sends multicast traffic to group G via R, the 3066 Proxy/Server Z* for the Client that aggregates S encapsulates the 3067 original IP packets in PIM Register messages and forwards them to R 3068 via the spanning tree, which may then elect to send a PIM Join to Z*. 3069 This will result in an (S,G) tree rooted at Z* with R as the next hop 3070 so that R will begin to receive two copies of the original IP packet; 3071 one native copy from the (S, G) tree and a second copy from the pre- 3072 existing (*, G) tree that still uses PIM Register encapsulation. R 3073 can then issue a PIM Register-stop message to suppress the Register- 3074 encapsulated stream. At some later time, if C moves to a new Proxy/ 3075 Server Z*, it resumes sending original IP packets via PIM Register 3076 encapsulation via the new Z*. 3078 At the same time, as multicast listeners discover individual S's for 3079 a given G, they can initiate an (S,G) Join for each S under the same 3080 procedures discussed in Section 3.17.1. Once the (S,G) tree is 3081 established, the listeners can send (S, G) Prune messages to R so 3082 that multicast original IP packets for group G sourced by S will only 3083 be delivered via the (S, G) tree and not from the (*, G) tree rooted 3084 at R. All mobility considerations discussed for SSM apply. 3086 3.17.3. Bi-Directional PIM (BIDIR-PIM) 3088 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 3089 approach to ASM that treats the Rendezvous Point (RP) as a Designated 3090 Forwarder (DF). Further considerations for BIDIR-PIM are out of 3091 scope. 3093 3.18. Operation over Multiple OMNI Links 3095 An AERO Client can connect to multiple OMNI links the same as for any 3096 data link service. In that case, the Client maintains a distinct 3097 OMNI interface for each link, e.g., 'omni0' for the first link, 3098 'omni1' for the second, 'omni2' for the third, etc. Each OMNI link 3099 would include its own distinct set of Bridges and Proxy/Servers, 3100 thereby providing redundancy in case of failures. 3102 Each OMNI link could utilize the same or different ANET connections. 3103 The links can be distinguished at the link-layer via the SRT prefix 3104 in a similar fashion as for Virtual Local Area Network (VLAN) tagging 3105 (e.g., IEEE 802.1Q) and/or through assignment of distinct sets of 3106 MSPs on each link. This gives rise to the opportunity for supporting 3107 multiple redundant networked paths, with each VLAN distinguished by a 3108 different SRT "color" (see: Section 3.2.5). 3110 The Client's IP layer can select the outgoing OMNI interface 3111 appropriate for a given traffic profile while (in the reverse 3112 direction) correspondent nodes must have some way of steering their 3113 original IP packets destined to a target via the correct OMNI link. 3115 In a first alternative, if each OMNI link services different MSPs, 3116 then the Client can receive a distinct MNP from each of the links. 3117 IP routing will therefore assure that the correct OMNI link is used 3118 for both outbound and inbound traffic. This can be accomplished 3119 using existing technologies and approaches, and without requiring any 3120 special supporting code in correspondent nodes or Bridges. 3122 In a second alternative, if each OMNI link services the same MSP(s) 3123 then each link could assign a distinct "OMNI link Anycast" address 3124 that is configured by all Bridges on the link. Correspondent nodes 3125 can then perform Segment Routing to select the correct SRT, which 3126 will then direct the original IP packet over multiple hops to the 3127 target. 3129 3.19. DNS Considerations 3131 AERO Client MNs and INET correspondent nodes consult the Domain Name 3132 System (DNS) the same as for any Internetworking node. When 3133 correspondent nodes and Client MNs use different IP protocol versions 3134 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 3135 A records for IPv4 address mappings to MNs which must then be 3136 populated in Relay NAT64 mapping caches. In that way, an IPv4 3137 correspondent node can send original IPv4 packets to the IPv4 address 3138 mapping of the target MN, and the Relay will translate the IPv4 3139 header and destination address into an IPv6 header and IPv6 3140 destination address of the MN. 3142 When an AERO Client registers with an AERO Proxy/Server, the Proxy/ 3143 Server can return the address(es) of DNS servers in RDNSS options 3144 [RFC6106]. The DNS server provides the IP addresses of other MNs and 3145 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 3147 3.20. Transition/Coexistence Considerations 3149 OAL encapsulation ensures that dissimilar INET partitions can be 3150 joined into a single unified OMNI link, even though the partitions 3151 themselves may have differing protocol versions and/or incompatible 3152 addressing plans. However, a commonality can be achieved by 3153 incrementally distributing globally routable (i.e., native) IP 3154 prefixes to eventually reach all nodes (both mobile and fixed) in all 3155 OMNI link segments. This can be accomplished by incrementally 3156 deploying AERO Bridges on each INET partition, with each Bridge 3157 distributing its MNPs and/or discovering non-MNP IP GUA prefixes on 3158 its INET links. 3160 This gives rise to the opportunity to eventually distribute native IP 3161 addresses to all nodes, and to present a unified OMNI link view even 3162 if the INET partitions remain in their current protocol and 3163 addressing plans. In that way, the OMNI link can serve the dual 3164 purpose of providing a mobility/multilink service and a transition/ 3165 coexistence service. Or, if an INET partition is transitioned to a 3166 native IP protocol version and addressing scheme that is compatible 3167 with the OMNI link MNP-based addressing scheme, the partition and 3168 OMNI link can be joined by Bridges. 3170 Relays that connect INETs/EUNs with dissimilar IP protocol versions 3171 may need to employ a network address and protocol translation 3172 function such as NAT64 [RFC6146]. 3174 3.21. Detecting and Reacting to Proxy/Server and Bridge Failures 3176 In environments where rapid failure recovery is required, Proxy/ 3177 Servers and Bridges SHOULD use Bidirectional Forwarding Detection 3178 (BFD) [RFC5880]. Nodes that use BFD can quickly detect and react to 3179 failures so that cached information is re-established through 3180 alternate nodes. BFD control messaging is carried only over well- 3181 connected ground domain networks (i.e., and not low-end radio links) 3182 and can therefore be tuned for rapid response. 3184 Proxy/Servers and Bridges maintain BFD sessions in parallel with 3185 their BGP peerings. If a Proxy/Server or Bridge fails, BGP peers 3186 will quickly re-establish routes through alternate paths the same as 3187 for common BGP deployments. Similarly, Proxys maintain BFD sessions 3188 with their associated Bridges even though they do not establish BGP 3189 peerings with them. 3191 3.22. AERO Clients on the Open Internet 3193 AERO Clients that connect to the open Internet via INET interfaces 3194 can establish a VPN or direct link to securely connect to a Proxy/ 3195 Server in a "tethered" arrangement with all of the Client's traffic 3196 transiting the Proxy/Server. Alternatively, the Client can associate 3197 with an INET Proxy/Server using UDP/IP encapsulation and control 3198 message securing services as discussed in the following sections. 3200 When a Client's OMNI interface enables an INET underlying interface, 3201 it first determines whether the interface is likely to be behind a 3202 NAT. For IPv4, the Client assumes it is on the open Internet if the 3203 INET address is not a special-use IPv4 address per [RFC3330]. 3204 Similarly for IPv6, the Client assumes it is on the open Internet if 3205 the INET address is a Global Unicast Address (GUA) [RFC4291]. 3206 Otherwise, the Client assumes it may be behind one or several NATs. 3208 The Client then prepares an RS message with IPv6 source address set 3209 to its MNP-LLA, with IPv6 destination set to (link-local) All-Routers 3210 multicast and with an OMNI option with underlying interface 3211 attributes. If the Client believes that it is on the open Internet, 3212 it SHOULD include an L2ADDR in the Interface Attributes sub-option 3213 corresponding to the underlying interface; otherwise, it MAY omit 3214 L2ADDR. If the underlying address is IPv4, the Client includes the 3215 Port Number and IPv4 address written in obfuscated form [RFC4380] as 3216 discussed in Section 3.3. If the underlying interface address is 3217 IPv6, the Client instead includes the Port Number and IPv6 address in 3218 obfuscated form. The Client finally includes an authentication 3219 signature sub-option in the OMNI option [I-D.templin-6man-omni] to 3220 provide message authentication and submits the RS for OAL 3221 encapsulation as an atomic fragment using an unpredictable 3222 Identification value to establish the start of the "send/accept" 3223 window for this Proxy/Server. The Client then encapsulates the OAL 3224 fragment in UDP/IP headers to form a carrier packet, sets the UDP/IP 3225 source to its INET address and UDP port, sets the UDP/IP destination 3226 to the Proxy/Server's INET address and the AERO service port number 3227 (8060), then sends the carrier packet to the Proxy/Server. 3229 When the Proxy/Server receives the RS, it discards the OAL 3230 encapsulation, authenticates the RS message, creates a neighbor cache 3231 entry and registers the Client's MNP, Identification and INET 3232 interface information according to the OMNI option parameters. If 3233 the RS message OMNI option includes Interface Attributes with an 3234 L2ADDR, the Proxy/Server compares the encapsulation IP address and 3235 UDP port number with the (unobfuscated) values. If the values are 3236 the same, the Proxy/Server caches the Client's information as "INET" 3237 addresses meaning that the Client is likely to accept direct messages 3238 without requiring NAT traversal exchanges. If the values are 3239 different (or, if the OMNI option did not include an L2ADDR) the 3240 Proxy/Server instead caches the Client's information as "mapped" 3241 addresses meaning that NAT traversal exchanges may be necessary. 3243 The Proxy/Server then prepares an RA message with IPv6 source and 3244 destination set corresponding to the addresses in the RS, and with an 3245 OMNI option with an Origin Indication sub-option per 3246 [I-D.templin-6man-omni] with the mapped and obfuscated Port Number 3247 and IP address observed in the encapsulation headers. The Proxy/ 3248 Server also includes an authentication signature sub-option per 3249 [I-D.templin-6man-omni] that contains an acknowledgement of the 3250 update sent by the Client. The Proxy/Server then performs OAL 3251 encapsulation and fragmentation if necessary using the same 3252 Identification value that appeared in the RS, and encapsulates each 3253 fragment in UDP/IP headers with addresses set per the L2ADDR 3254 information in the neighbor cache entry for the Client. 3256 When the Client receives the RA message, it verifies the OAL 3257 Identification value, performs OAL reassembly if necessary, 3258 authenticates the message, then compares the mapped Port Number and 3259 IP address from the Origin Indication sub-option with its own 3260 address. If the addresses are the same, the Client assumes the open 3261 Internet / Cone NAT principle; if the addresses are different, the 3262 Client instead assumes that further qualification procedures are 3263 necessary to detect the type of NAT and proceeds according to 3264 standard procedures [RFC6081][RFC4380]. 3266 After the Client has registered its INET interfaces in such RS/RA 3267 exchanges it sends periodic RS messages to receive fresh RA messages 3268 before the Router Lifetime received on each INET interface expires. 3269 The Client also maintains default routes via its Proxy/Servers, i.e., 3270 the same as described in earlier sections. 3272 When the Client sends messages to target IP addresses, it also 3273 invokes route optimization per Section 3.14 using IPv6 ND address 3274 resolution messaging. The Client first creates a neighbor cache 3275 entry for the target in the INCOMPLETE state, then sends the NS(AR) 3276 message to the Proxy/Server with an OMNI option with an 3277 authentication signature sub-option. The Client sets the NS source 3278 address to its own MNP-LLA, destination address to the target 3279 solicited node multicast address and target address to the LLA of the 3280 target. The Client then wraps the NS message in OAL headers (i.e., 3281 as an atomic fragment) with an unpredictable Identification value to 3282 establish the "send" window for this target, with source address set 3283 to its own MNP-ULA and destination address set to the target's MNP- 3284 ULA. The Client then wraps the atomic fragment in a UDP/IP header 3285 and sends the resulting carrier packet to the Proxy/Server. 3287 When the Client's Proxy/Server receives the OAL-encapsulated NS, it 3288 authenticates the message by processing the authentication signature 3289 sub-option and forwards the message over the spanning tree on behalf 3290 of the Client. When the ROR receives the NS(AR), it creates a 3291 neighbor cache entry for the ROS in the STALE state and caches the 3292 Identification value as the start of the "accept" window for packets 3293 originating from this ROS (if the ROR is a Proxy/Server, it also 3294 creates a Report List entry for this ROS in the target Client's 3295 neighbor cache entry). The ROR then returns an NA(AR) with OMNI 3296 option information for the target including all of the target's 3297 Interface Attributes. 3299 The ROR sets the NA(AR) source address to its own LLA, sets the 3300 destination address to the ROS LLA and sets the target address to the 3301 LLA of the target. The ROR then performs OAL encapsulation using the 3302 same Identification value that appeared in the NS(AR), then sets the 3303 OAL source address to the ROR's ULA and destination address to ULA 3304 source of the NS(AR). If the ROR is an INET Client, it includes an 3305 authentication signature and sends the NA(AR) to its Proxy/Sever 3306 which verifies the authentication signature and forwards the NA(AR) 3307 into the secured spanning tree. If the ROR is an ANET Client or a 3308 Proxy/Server, it simply forwards the NA(AR) into the secured spanning 3309 tree. 3311 When the Proxy/Sever for the ROS Client receives the NA(AR) message 3312 contained in one or more carrier packets, it verifies the OAL 3313 Identification matches the same value that was used in the NS(AR) 3314 then reassembles if necessary. When reassembly is complete, the 3315 Proxy/Server includes an authentication signature and forwards the 3316 NA(AR) to the ROS Client. The ROS Client then verifies the 3317 authentication signature and changes the neighbor cache entry state 3318 for this target to REACHABLE. 3320 Following route optimization for targets in the same OMNI link 3321 segment, if the target's L2ADDR is on the open INET, the Client 3322 forwards carrier packets directly to the target INET address. If the 3323 target is behind a NAT, the Client first establishes NAT state for 3324 the L2ADDR using the "direct bubble" and NUD mechanisms discussed in 3325 Section 3.10.1. The Client continues to send carrier packets via its 3326 Proxy/Server until NAT state is populated, then begins forwarding 3327 carrier packets via the direct path through the NAT to the target. 3328 For targets in different OMNI link segments, the Client uses OAL/ORH 3329 encapsulation and forwards carrier packets to the Bridge that 3330 returned the NA message. 3332 The ROR may return uNAs via the ROS Proxy/Server if the target moves, 3333 and the Proxy/Server will send corresponding uNAs to the Client with 3334 an OMNI authentication sub-option. The Client can also send NUD 3335 messages to test forward path reachability even though there is no 3336 security association between the Client and the target. 3338 The Client can send original IP packets to route-optimized neighbors 3339 in the same OMNI link segment no larger than the minimum/path MPS in 3340 one piece and with OAL encapsulation but without fragmentation. For 3341 larger original IP packets, the Client applies OAL encapsulation and 3342 fragmentation if necessary according to Section 3.9, with OAL header 3343 with source set to its own MNP-ULA and destination set to the MNP-ULA 3344 of the target. The Client then encapsulates each original IP packet 3345 or OAL fragment in UDP/IP *NET headers and sends them to the next 3346 hop. 3348 Note: The NAT traversal procedures specified in this document are 3349 applicable for Cone, Address-Restricted and Port-Restricted NATs 3350 only. While future updates to this document may specify procedures 3351 for other NAT variations (e.g., hairpinning and various forms of 3352 Symmetric NATs), it should be noted that continuous communications 3353 are always possible through forwarding via a Proxy/Server even if NAT 3354 traversal is not employed. 3356 3.23. Time-Varying MNPs 3358 In some use cases, it is desirable, beneficial and efficient for the 3359 Client to receive a constant MNP that travels with the Client 3360 wherever it moves. For example, this would allow air traffic 3361 controllers to easily track aircraft, etc. In other cases, however 3362 (e.g., intelligent transportation systems), the MN may be willing to 3363 sacrifice a modicum of efficiency in order to have time-varying MNPs 3364 that can be changed every so often to defeat adversarial tracking. 3366 The DHCPv6 service offers a way for Clients that desire time-varying 3367 MNPs to obtain short-lived prefixes (e.g., on the order of a small 3368 number of minutes). In that case, the identity of the Client would 3369 not be bound to the MNP but rather to a Node Identification value 3370 (see: [I-D.templin-6man-omni]) to be used as the Client ID seed for 3371 MNP prefix delegation. The Client would then be obligated to 3372 renumber its internal networks whenever its MNP (and therefore also 3373 its MNP-LLA) changes. This should not present a challenge for 3374 Clients with automated network renumbering services, however presents 3375 limits for the durations of ongoing sessions that would prefer to use 3376 a constant address. 3378 4. Implementation Status 3380 An early AERO implementation based on OpenVPN (https://openvpn.net/) 3381 was announced on the v6ops mailing list on January 10, 2018 and an 3382 initial public release of the AERO proof-of-concept source code was 3383 announced on the intarea mailing list on August 21, 2015. 3385 AERO Release-3.2 was tagged on March 30, 2021, and is undergoing 3386 internal testing. Additional internal releases expected within the 3387 coming months, with first public release expected end of 1H2021. 3389 5. IANA Considerations 3391 The IANA is instructed to assign a new type value TBD1 in the IPv6 3392 Routing Types registry (IANA registration procedure is IETF Review or 3393 IESG Approval). 3395 The IANA has assigned the UDP port number "8060" for an earlier 3396 experimental first version of AERO [RFC6706]. This document 3397 obsoletes [RFC6706], and together with [I-D.templin-6man-omni] 3398 reclaims the UDP port number "8060" for 'aero' as the service port 3399 for UDP/IP encapsulation. (Note that, although [RFC6706] was not 3400 widely implemented or deployed, any messages coded to that 3401 specification can be easily distinguished and ignored since they use 3402 the invalid ICMPv6 message type number '0'.) This document makes no 3403 request of IANA, since [I-D.templin-6man-omni] already provides 3404 instructions. 3406 No further IANA actions are required. 3408 6. Security Considerations 3410 AERO Bridges configure secured tunnels with AERO Proxy/Servers and 3411 Relays within their local OMNI link segments. Applicable secured 3412 tunnel alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 3413 [RFC6347], WireGuard [WG], etc. The AERO Bridges of all OMNI link 3414 segments in turn configure secured tunnels for their neighboring AERO 3415 Bridges in a spanning tree topology. Therefore, control messages 3416 exchanged between any pair of OMNI link neighbors on the spanning 3417 tree are already secured. 3419 AERO nodes acting as Route Optimization Responders (RORs) may also 3420 receive packets directly from arbitrary nodes in INET partitions 3421 instead of via the secured spanning tree. For INET partitions that 3422 apply effective ingress filtering to defeat source address spoofing, 3423 the simple data origin authentication procedures in Section 3.8 can 3424 be applied. 3426 For INET partitions that require strong security in the data plane, 3427 two options for securing communications include 1) disable route 3428 optimization so that all traffic is conveyed over secured tunnels, or 3429 2) enable on-demand secure tunnel creation between INET partition 3430 neighbors. Option 1) would result in longer routes than necessary 3431 and traffic concentration on critical infrastructure elements. 3432 Option 2) could be coordinated by establishing a secured tunnel on- 3433 demand instead of performing an NS/NA exchange in the route 3434 optimization procedures. 3436 AERO Clients that connect to secured ANETs need not apply security to 3437 their ND messages, since the messages will be intercepted by a 3438 perimeter Proxy/Server that applies security on its INET-facing 3439 interface as part of the spanning tree (see above). AERO Clients 3440 connected to the open INET can use network and/or transport layer 3441 security services such as VPNs or can by some other means establish a 3442 direct link to a Proxy/Server. When a VPN or direct link may be 3443 impractical, however, INET Clients and Proxy/Servers SHOULD include 3444 and verify authentication signatures for their IPv6 ND messages as 3445 specified in [I-D.templin-6man-omni]. 3447 Application endpoints SHOULD use application-layer security services 3448 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 3449 protection as for critical secured Internet services. AERO Clients 3450 that require host-based VPN services SHOULD use network and/or 3451 transport layer security services such as IPsec, TLS/SSL, DTLS, etc. 3452 AERO Proxys and Proxy/Servers can also provide a network-based VPN 3453 service on behalf of the Client, e.g., if the Client is located 3454 within a secured enclave and cannot establish a VPN on its own 3455 behalf. 3457 AERO Proxy/Servers and Bridges present targets for traffic 3458 amplification Denial of Service (DoS) attacks. This concern is no 3459 different than for widely-deployed VPN security gateways in the 3460 Internet, where attackers could send spoofed packets to the gateways 3461 at high data rates. This can be mitigated by connecting Proxy/ 3462 Servers and Bridges over dedicated links with no connections to the 3463 Internet and/or when connections to the Internet are only permitted 3464 through well-managed firewalls. Traffic amplification DoS attacks 3465 can also target an AERO Client's low data rate links. This is a 3466 concern not only for Clients located on the open Internet but also 3467 for Clients in secured enclaves. AERO Proxy/Servers and Proxys can 3468 institute rate limits that protect Clients from receiving packet 3469 floods that could DoS low data rate links. 3471 AERO Relays must implement ingress filtering to avoid a spoofing 3472 attack in which spurious messages with ULA addresses are injected 3473 into an OMNI link from an outside attacker. AERO Clients MUST ensure 3474 that their connectivity is not used by unauthorized nodes on their 3475 EUNs to gain access to a protected network, i.e., AERO Clients that 3476 act as routers MUST NOT provide routing services for unauthorized 3477 nodes. (This concern is no different than for ordinary hosts that 3478 receive an IP address delegation but then "share" the address with 3479 other nodes via some form of Internet connection sharing such as 3480 tethering.) 3482 The MAP list MUST be well-managed and secured from unauthorized 3483 tampering, even though the list contains only public information. 3484 The MAP list can be conveyed to the Client in a similar fashion as in 3485 [RFC5214] (e.g., through layer 2 data link login messaging, secure 3486 upload of a static file, DNS lookups, etc.). 3488 SRH authentication facilities are specified in [RFC8754]. 3490 Security considerations for accepting link-layer ICMP messages and 3491 reflected packets are discussed throughout the document. 3493 Security considerations for IPv6 fragmentation and reassembly are 3494 discussed in [I-D.templin-6man-omni]. 3496 7. Acknowledgements 3498 Discussions in the IETF, aviation standards communities and private 3499 exchanges helped shape some of the concepts in this work. 3500 Individuals who contributed insights include Mikael Abrahamsson, Mark 3501 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 3502 Wojciech Dec, Pavel Drasil, Ralph Droms, Adrian Farrel, Nick Green, 3503 Sri Gundavelli, Brian Haberman, Bernhard Haindl, Joel Halpern, Tom 3504 Herbert, Sascha Hlusiak, Lee Howard, Zdenek Jaron, Andre Kostur, 3505 Hubert Kuenig, Ted Lemon, Andy Malis, Satoru Matsushima, Tomek 3506 Mrugalski, Madhu Niraula, Alexandru Petrescu, Behcet Saikaya, Michal 3507 Skorepa, Joe Touch, Bernie Volz, Ryuji Wakikawa, Tony Whyman, Lloyd 3508 Wood and James Woodyatt. Members of the IESG also provided valuable 3509 input during their review process that greatly improved the document. 3510 Special thanks go to Stewart Bryant, Joel Halpern and Brian Haberman 3511 for their shepherding guidance during the publication of the AERO 3512 first edition. 3514 This work has further been encouraged and supported by Boeing 3515 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 3516 Brodie, John Bush, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, 3517 Claudiu Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad 3518 Farooqui, Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, 3519 Greg Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, 3520 Gene MacLean III, Kyle Mikos, Rob Muszkiewicz, Sean O'Sullivan, Vijay 3521 Rajagopalan, Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, 3522 Mike Slane, Carrie Spiker, Katie Tran, Brendan Williams, Amelia 3523 Wilson, Julie Wulff, Yueli Yang, Eric Yeh and other members of the 3524 Boeing mobility, networking and autonomy teams. Kyle Bae, Wayne 3525 Benson, Madhuri Madhava Badgandi, Vijayasarathy Rajagopalan, Katie 3526 Tran and Eric Yeh are especially acknowledged for implementing the 3527 AERO functions as extensions to the public domain OpenVPN 3528 distribution. Chuck Klabunde is honored and remembered for his early 3529 leadership, and we mourn his untimely loss. 3531 Earlier works on NBMA tunneling approaches are found in 3532 [RFC2529][RFC5214][RFC5569]. 3534 Many of the constructs presented in this second edition of AERO are 3535 based on the author's earlier works, including: 3537 o The Internet Routing Overlay Network (IRON) 3538 [RFC6179][I-D.templin-ironbis] 3540 o Virtual Enterprise Traversal (VET) 3541 [RFC5558][I-D.templin-intarea-vet] 3543 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 3544 [RFC5320][I-D.templin-intarea-seal] 3546 o AERO, First Edition [RFC6706] 3548 Note that these works cite numerous earlier efforts that are not also 3549 cited here due to space limitations. The authors of those earlier 3550 works are acknowledged for their insights. 3552 This work is aligned with the NASA Safe Autonomous Systems Operation 3553 (SASO) program under NASA contract number NNA16BD84C. 3555 This work is aligned with the FAA as per the SE2025 contract number 3556 DTFAWA-15-D-00030. 3558 This work is aligned with the Boeing Commercial Airplanes (BCA) 3559 Internet of Things (IoT) and autonomy programs. 3561 This work is aligned with the Boeing Information Technology (BIT) 3562 MobileNet program. 3564 8. References 3566 8.1. Normative References 3568 [I-D.templin-6man-omni] 3569 Templin, F. L. and T. Whyman, "Transmission of IP Packets 3570 over Overlay Multilink Network (OMNI) Interfaces", draft- 3571 templin-6man-omni-03 (work in progress), April 2021. 3573 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 3574 DOI 10.17487/RFC0791, September 1981, 3575 . 3577 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 3578 RFC 792, DOI 10.17487/RFC0792, September 1981, 3579 . 3581 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3582 Requirement Levels", BCP 14, RFC 2119, 3583 DOI 10.17487/RFC2119, March 1997, 3584 . 3586 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3587 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3588 December 1998, . 3590 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3591 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3592 DOI 10.17487/RFC3971, March 2005, 3593 . 3595 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3596 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3597 . 3599 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3600 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3601 November 2005, . 3603 [RFC4193] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast 3604 Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005, 3605 . 3607 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 3608 Network Address Translations (NATs)", RFC 4380, 3609 DOI 10.17487/RFC4380, February 2006, 3610 . 3612 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3613 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3614 DOI 10.17487/RFC4861, September 2007, 3615 . 3617 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3618 Address Autoconfiguration", RFC 4862, 3619 DOI 10.17487/RFC4862, September 2007, 3620 . 3622 [RFC6081] Thaler, D., "Teredo Extensions", RFC 6081, 3623 DOI 10.17487/RFC6081, January 2011, 3624 . 3626 [RFC7401] Moskowitz, R., Ed., Heer, T., Jokela, P., and T. 3627 Henderson, "Host Identity Protocol Version 2 (HIPv2)", 3628 RFC 7401, DOI 10.17487/RFC7401, April 2015, 3629 . 3631 [RFC7739] Gont, F., "Security Implications of Predictable Fragment 3632 Identification Values", RFC 7739, DOI 10.17487/RFC7739, 3633 February 2016, . 3635 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3636 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3637 May 2017, . 3639 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3640 (IPv6) Specification", STD 86, RFC 8200, 3641 DOI 10.17487/RFC8200, July 2017, 3642 . 3644 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3645 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3646 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3647 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3648 . 3650 8.2. Informative References 3652 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3653 2016. 3655 [I-D.bonica-6man-comp-rtg-hdr] 3656 Bonica, R., Kamite, Y., Alston, A., Henriques, D., and L. 3657 Jalil, "The IPv6 Compact Routing Header (CRH)", draft- 3658 bonica-6man-comp-rtg-hdr-24 (work in progress), January 3659 2021. 3661 [I-D.bonica-6man-crh-helper-opt] 3662 Li, X., Bao, C., Ruan, E., and R. Bonica, "Compressed 3663 Routing Header (CRH) Helper Option", draft-bonica-6man- 3664 crh-helper-opt-02 (work in progress), October 2020. 3666 [I-D.ietf-intarea-frag-fragile] 3667 Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O., 3668 and F. Gont, "IP Fragmentation Considered Fragile", draft- 3669 ietf-intarea-frag-fragile-17 (work in progress), September 3670 2019. 3672 [I-D.ietf-intarea-tunnels] 3673 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3674 Architecture", draft-ietf-intarea-tunnels-10 (work in 3675 progress), September 2019. 3677 [I-D.ietf-ipwave-vehicular-networking] 3678 Jeong, J., "IPv6 Wireless Access in Vehicular Environments 3679 (IPWAVE): Problem Statement and Use Cases", draft-ietf- 3680 ipwave-vehicular-networking-19 (work in progress), July 3681 2020. 3683 [I-D.ietf-rtgwg-atn-bgp] 3684 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3685 Moreno, "A Simple BGP-based Mobile Routing System for the 3686 Aeronautical Telecommunications Network", draft-ietf- 3687 rtgwg-atn-bgp-10 (work in progress), January 2021. 3689 [I-D.templin-6man-dhcpv6-ndopt] 3690 Templin, F., "A Unified Stateful/Stateless Configuration 3691 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-11 3692 (work in progress), January 2021. 3694 [I-D.templin-intarea-seal] 3695 Templin, F., "The Subnetwork Encapsulation and Adaptation 3696 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3697 progress), January 2014. 3699 [I-D.templin-intarea-vet] 3700 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3701 templin-intarea-vet-40 (work in progress), May 2013. 3703 [I-D.templin-ipwave-uam-its] 3704 Templin, F., "Urban Air Mobility Implications for 3705 Intelligent Transportation Systems", draft-templin-ipwave- 3706 uam-its-04 (work in progress), January 2021. 3708 [I-D.templin-ironbis] 3709 Templin, F., "The Interior Routing Overlay Network 3710 (IRON)", draft-templin-ironbis-16 (work in progress), 3711 March 2014. 3713 [I-D.templin-v6ops-pdhost] 3714 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3715 Models", draft-templin-v6ops-pdhost-27 (work in progress), 3716 January 2021. 3718 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3720 [RFC1035] Mockapetris, P., "Domain names - implementation and 3721 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3722 November 1987, . 3724 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3725 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3726 . 3728 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3729 DOI 10.17487/RFC2003, October 1996, 3730 . 3732 [RFC2004] Perkins, C., "Minimal Encapsulation within IP", RFC 2004, 3733 DOI 10.17487/RFC2004, October 1996, 3734 . 3736 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3737 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3738 . 3740 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 3741 Networks", RFC 2464, DOI 10.17487/RFC2464, December 1998, 3742 . 3744 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3745 Domains without Explicit Tunnels", RFC 2529, 3746 DOI 10.17487/RFC2529, March 1999, 3747 . 3749 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3750 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3751 . 3753 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3754 of Explicit Congestion Notification (ECN) to IP", 3755 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3756 . 3758 [RFC3330] IANA, "Special-Use IPv4 Addresses", RFC 3330, 3759 DOI 10.17487/RFC3330, September 2002, 3760 . 3762 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3763 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3764 DOI 10.17487/RFC3810, June 2004, 3765 . 3767 [RFC4122] Leach, P., Mealling, M., and R. Salz, "A Universally 3768 Unique IDentifier (UUID) URN Namespace", RFC 4122, 3769 DOI 10.17487/RFC4122, July 2005, 3770 . 3772 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3773 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3774 January 2006, . 3776 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3777 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3778 DOI 10.17487/RFC4271, January 2006, 3779 . 3781 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3782 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3783 2006, . 3785 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3786 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3787 December 2005, . 3789 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3790 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3791 2006, . 3793 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3794 Control Message Protocol (ICMPv6) for the Internet 3795 Protocol Version 6 (IPv6) Specification", STD 89, 3796 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3797 . 3799 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3800 Protocol (LDAP): The Protocol", RFC 4511, 3801 DOI 10.17487/RFC4511, June 2006, 3802 . 3804 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3805 "Considerations for Internet Group Management Protocol 3806 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3807 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3808 . 3810 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3811 "Internet Group Management Protocol (IGMP) / Multicast 3812 Listener Discovery (MLD)-Based Multicast Forwarding 3813 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3814 August 2006, . 3816 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3817 Algorithms in Cryptographically Generated Addresses 3818 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3819 . 3821 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3822 "Bidirectional Protocol Independent Multicast (BIDIR- 3823 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3824 . 3826 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3827 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3828 DOI 10.17487/RFC5214, March 2008, 3829 . 3831 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3832 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3833 February 2010, . 3835 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3836 Route Optimization Requirements for Operational Use in 3837 Aeronautics and Space Exploration Mobile Networks", 3838 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3839 . 3841 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3842 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3843 . 3845 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3846 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3847 January 2010, . 3849 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3850 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3851 . 3853 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3854 "IPv6 Router Advertisement Options for DNS Configuration", 3855 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3856 . 3858 [RFC6139] Russert, S., Ed., Fleischman, E., Ed., and F. Templin, 3859 Ed., "Routing and Addressing in Networks with Global 3860 Enterprise Recursion (RANGER) Scenarios", RFC 6139, 3861 DOI 10.17487/RFC6139, February 2011, 3862 . 3864 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3865 NAT64: Network Address and Protocol Translation from IPv6 3866 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3867 April 2011, . 3869 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3870 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3871 . 3873 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3874 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3875 DOI 10.17487/RFC6221, May 2011, 3876 . 3878 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3879 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3880 DOI 10.17487/RFC6273, June 2011, 3881 . 3883 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3884 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3885 January 2012, . 3887 [RFC6355] Narten, T. and J. Johnson, "Definition of the UUID-Based 3888 DHCPv6 Unique Identifier (DUID-UUID)", RFC 6355, 3889 DOI 10.17487/RFC6355, August 2011, 3890 . 3892 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3893 for Equal Cost Multipath Routing and Link Aggregation in 3894 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3895 . 3897 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3898 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3899 . 3901 [RFC6935] Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 3902 UDP Checksums for Tunneled Packets", RFC 6935, 3903 DOI 10.17487/RFC6935, April 2013, 3904 . 3906 [RFC6936] Fairhurst, G. and M. Westerlund, "Applicability Statement 3907 for the Use of IPv6 UDP Datagrams with Zero Checksums", 3908 RFC 6936, DOI 10.17487/RFC6936, April 2013, 3909 . 3911 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3912 Korhonen, "Requirements for Distributed Mobility 3913 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3914 . 3916 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3917 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3918 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3919 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3920 2016, . 3922 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3923 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3924 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3925 July 2018, . 3927 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3928 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3929 . 3931 [RFC8754] Filsfils, C., Ed., Dukes, D., Ed., Previdi, S., Leddy, J., 3932 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3933 (SRH)", RFC 8754, DOI 10.17487/RFC8754, March 2020, 3934 . 3936 [WG] Wireguard, "Wireguard, https://www.wireguard.com", August 3937 2020. 3939 Appendix A. Non-Normative Considerations 3941 AERO can be applied to a multitude of Internetworking scenarios, with 3942 each having its own adaptations. The following considerations are 3943 provided as non-normative guidance: 3945 A.1. Implementation Strategies for Route Optimization 3947 Route optimization as discussed in Section 3.14 results in the route 3948 optimization source (ROS) creating a neighbor cache entry for the 3949 target neighbor. The neighbor cache entry state is set to REACHABLE 3950 for at most ReachableTime seconds. In order to refresh the neighbor 3951 cache entry lifetime before the ReachableTime timer expires, the 3952 specification requires implementations to issue a new NS/NA exchange 3953 to reset ReachableTime while data packets are still flowing. 3954 However, the decision of when to initiate a new NS/NA exchange and to 3955 perpetuate the process is left as an implementation detail. 3957 One possible strategy may be to monitor the neighbor cache entry 3958 watching for data packets for (ReachableTime - 5) seconds. If any 3959 data packets have been sent to the neighbor within this timeframe, 3960 then send an NS to receive a new NA. If no data packets have been 3961 sent, wait for 5 additional seconds and send an immediate NS if any 3962 data packets are sent within this "expiration pending" 5 second 3963 window. If no additional data packets are sent within the 5 second 3964 window, reset the neighbor cache entry state to STALE. 3966 The monitoring of the neighbor data packet traffic therefore becomes 3967 an ongoing process during the neighbor cache entry lifetime. If the 3968 neighbor cache entry expires, future data packets will trigger a new 3969 NS/NA exchange while the packets themselves are delivered over a 3970 longer path until route optimization state is re-established. 3972 A.2. Implicit Mobility Management 3974 OMNI interface neighbors MAY provide a configuration option that 3975 allows them to perform implicit mobility management in which no ND 3976 messaging is used. In that case, the Client only transmits packets 3977 over a single interface at a time, and the neighbor always observes 3978 packets arriving from the Client from the same link-layer source 3979 address. 3981 If the Client's underlying interface address changes (either due to a 3982 readdressing of the original interface or switching to a new 3983 interface) the neighbor immediately updates the neighbor cache entry 3984 for the Client and begins accepting and sending packets according to 3985 the Client's new address. This implicit mobility method applies to 3986 use cases such as cellphones with both WiFi and Cellular interfaces 3987 where only one of the interfaces is active at a given time, and the 3988 Client automatically switches over to the backup interface if the 3989 primary interface fails. 3991 A.3. Direct Underlying Interfaces 3993 When a Client's OMNI interface is configured over a Direct interface, 3994 the neighbor at the other end of the Direct link can receive packets 3995 without any encapsulation. In that case, the Client sends packets 3996 over the Direct link according to QoS preferences. If the Direct 3997 interface has the highest QoS preference, then the Client's IP 3998 packets are transmitted directly to the peer without going through an 3999 ANET/INET. If other interfaces have higher QoS preferences, then the 4000 Client's IP packets are transmitted via a different interface, which 4001 may result in the inclusion of Proxy/Servers and Bridges in the 4002 communications path. Direct interfaces must be tested periodically 4003 for reachability, e.g., via NUD. 4005 A.4. AERO Critical Infrastructure Considerations 4007 AERO Bridges can be either Commercial off-the Shelf (COTS) standard 4008 IP routers or virtual machines in the cloud. Bridges must be 4009 provisioned, supported and managed by the INET administrative 4010 authority, and connected to the Bridges of other INETs via inter- 4011 domain peerings. Cost for purchasing, configuring and managing 4012 Bridges is nominal even for very large OMNI links. 4014 AERO cloud Proxy/Servers can be standard dedicated server platforms, 4015 but most often will be deployed as virtual machines in the cloud. 4016 The only requirements for cloud Proxy/Servers are that they can run 4017 the AERO user-level code and have at least one network interface 4018 connection to the INET. Cloud Proxy/Servers must be provisioned, 4019 supported and managed by the INET administrative authority. Cost for 4020 purchasing, configuring and managing cloud Proxy/Servers is nominal 4021 especially for virtual machines. 4023 AERO ANET Proxy/Servers are most often standard dedicated server 4024 platforms with one underlying interface connected to the ANET and a 4025 second interface connected to an INET. As with cloud Proxy/Servers, 4026 the only requirements are that they can run the AERO user-level code 4027 and have at least one interface connection to the INET. ANET Proxy/ 4028 Servers must be provisioned, supported and managed by the ANET 4029 administrative authority. Cost for purchasing, configuring and 4030 managing Proxys is nominal, and borne by the ANET administrative 4031 authority. 4033 AERO Relays are simply Proxy/Servers connected to INETs and/or EUNs 4034 that provide forwarding services for non-MNP destinations. The Relay 4035 connects to the OMNI link and engages in eBGP peering with one or 4036 more Bridges as a stub AS. The Relay then injects its MNPs and/or 4037 non-MNP prefixes into the BGP routing system, and provisions the 4038 prefixes to its downstream-attached networks. The Relay can perform 4039 ROS/ROR services the same as for any Proxy/Server, and can route 4040 between the MNP and non-MNP address spaces. 4042 A.5. AERO Server Failure Implications 4044 AERO Proxy/Servers may appear as a single point of failure in the 4045 architecture, but such is not the case since all Proxy/Servers on the 4046 link provide identical services and loss of a Proxy/Server does not 4047 imply immediate and/or comprehensive communication failures. Proxy/ 4048 Server failure is quickly detected and conveyed by Bidirectional 4049 Forward Detection (BFD) and/or proactive NUD allowing Clients to 4050 migrate to new Proxy/Servers. 4052 If a Proxy/Server fails, ongoing packet forwarding to Clients will 4053 continue by virtue of the neighbor cache entries that have already 4054 been established in route optimization sources (ROSs). If a Client 4055 also experiences mobility events at roughly the same time the Proxy/ 4056 Server fails, unsolicited NA messages may be lost but neighbor cache 4057 entries in the DEPARTED state will ensure that packet forwarding to 4058 the Client's new locations will continue for up to DepartTime 4059 seconds. 4061 If a Client is left without a Proxy/Server for a considerable length 4062 of time (e.g., greater than ReachableTime seconds) then existing 4063 neighbor cache entries will eventually expire and both ongoing and 4064 new communications will fail. The original source will continue to 4065 retransmit until the Client has established a new Proxy/Server 4066 relationship, after which time continuous communications will resume. 4068 Therefore, providing many Proxy/Servers on the link with high 4069 availability profiles provides resilience against loss of individual 4070 Proxy/Servers and assurance that Clients can establish new Proxy/ 4071 Server relationships quickly in event of a Proxy/Server failure. 4073 A.6. AERO Client / Server Architecture 4075 The AERO architectural model is client / server in the control plane, 4076 with route optimization in the data plane. The same as for common 4077 Internet services, the AERO Client discovers the addresses of AERO 4078 Proxy/Servers and connects to one or more of them. The AERO service 4079 is analogous to common Internet services such as google.com, 4080 yahoo.com, cnn.com, etc. However, there is only one AERO service for 4081 the link and all Proxy/Servers provide identical services. 4083 Common Internet services provide differing strategies for advertising 4084 server addresses to clients. The strategy is conveyed through the 4085 DNS resource records returned in response to name resolution queries. 4086 As of January 2020 Internet-based 'nslookup' services were used to 4087 determine the following: 4089 o When a client resolves the domainname "google.com", the DNS always 4090 returns one A record (i.e., an IPv4 address) and one AAAA record 4091 (i.e., an IPv6 address). The client receives the same addresses 4092 each time it resolves the domainname via the same DNS resolver, 4093 but may receive different addresses when it resolves the 4094 domainname via different DNS resolvers. But, in each case, 4095 exactly one A and one AAAA record are returned. 4097 o When a client resolves the domainname "ietf.org", the DNS always 4098 returns one A record and one AAAA record with the same addresses 4099 regardless of which DNS resolver is used. 4101 o When a client resolves the domainname "yahoo.com", the DNS always 4102 returns a list of 4 A records and 4 AAAA records. Each time the 4103 client resolves the domainname via the same DNS resolver, the same 4104 list of addresses are returned but in randomized order (i.e., 4105 consistent with a DNS round-robin strategy). But, interestingly, 4106 the same addresses are returned (albeit in randomized order) when 4107 the domainname is resolved via different DNS resolvers. 4109 o When a client resolves the domainname "amazon.com", the DNS always 4110 returns a list of 3 A records and no AAAA records. As with 4111 "yahoo.com", the same three A records are returned from any 4112 worldwide Internet connection point in randomized order. 4114 The above example strategies show differing approaches to Internet 4115 resilience and service distribution offered by major Internet 4116 services. The Google approach exposes only a single IPv4 and a 4117 single IPv6 address to clients. Clients can then select whichever IP 4118 protocol version offers the best response, but will always use the 4119 same IP address according to the current Internet connection point. 4120 This means that the IP address offered by the network must lead to a 4121 highly-available server and/or service distribution point. In other 4122 words, resilience is predicated on high availability within the 4123 network and with no client-initiated failovers expected (i.e., it is 4124 all-or-nothing from the client's perspective). However, Google does 4125 provide for worldwide distributed service distribution by virtue of 4126 the fact that each Internet connection point responds with a 4127 different IPv6 and IPv4 address. The IETF approach is like google 4128 (all-or-nothing from the client's perspective), but provides only a 4129 single IPv4 or IPv6 address on a worldwide basis. This means that 4130 the addresses must be made highly-available at the network level with 4131 no client failover possibility, and if there is any worldwide service 4132 distribution it would need to be conducted by a network element that 4133 is reached via the IP address acting as a service distribution point. 4135 In contrast to the Google and IETF philosophies, Yahoo and Amazon 4136 both provide clients with a (short) list of IP addresses with Yahoo 4137 providing both IP protocol versions and Amazon as IPv4-only. The 4138 order of the list is randomized with each name service query 4139 response, with the effect of round-robin load balancing for service 4140 distribution. With a short list of addresses, there is still 4141 expectation that the network will implement high availability for 4142 each address but in case any single address fails the client can 4143 switch over to using a different address. The balance then becomes 4144 one of function in the network vs function in the end system. 4146 The same implications observed for common highly-available services 4147 in the Internet apply also to the AERO client/server architecture. 4148 When an AERO Client connects to one or more ANETs, it discovers one 4149 or more AERO Proxy/Server addresses through the mechanisms discussed 4150 in earlier sections. Each Proxy/Server address presumably leads to a 4151 fault-tolerant clustering arrangement such as supported by Linux-HA, 4152 Extended Virtual Synchrony or Paxos. Such an arrangement has 4153 precedence in common Internet service deployments in lightweight 4154 virtual machines without requiring expensive hardware deployment. 4155 Similarly, common Internet service deployments set service IP 4156 addresses on service distribution points that may relay requests to 4157 many different servers. 4159 For AERO, the expectation is that a combination of the Google/IETF 4160 and Yahoo/Amazon philosophies would be employed. The AERO Client 4161 connects to different ANET access points and can receive 1-2 Proxy/ 4162 Server ADM-LLAs at each point. It then selects one AERO Proxy/Server 4163 address, and engages in RS/RA exchanges with the same Proxy/Server 4164 from all ANET connections. The Client remains with this Proxy/Server 4165 unless or until the Proxy/Server fails, in which case it can switch 4166 over to an alternate Proxy/Server. The Client can likewise switch 4167 over to a different Proxy/Server at any time if there is some reason 4168 for it to do so. So, the AERO expectation is for a balance of 4169 function in the network and end system, with fault tolerance and 4170 resilience at both levels. 4172 Appendix B. Change Log 4174 << RFC Editor - remove prior to publication >> 4176 Changes from draft-templin-6man-aero-01 to draft-templin-6man-aero- 4177 02: 4179 o Changed reference citations to "draft-templin-6man-omni". 4181 o Several important updates to IPv6 ND cache states and route 4182 optimization message addressing. 4184 o Included introductory description of the "6M's". 4186 o Updated Multicast specification. 4188 Changes from draft-templin-6man-aero-00 to draft-templin-6man-aero- 4189 01: 4191 o Changed category to "Informational". 4193 o Updated implementation status. 4195 Changes from earlier versions to draft-templin-6man-aero-00: 4197 o Established working baseline reference. 4199 Author's Address 4201 Fred L. Templin (editor) 4202 Boeing Research & Technology 4203 P.O. Box 3707 4204 Seattle, WA 98124 4205 USA 4207 Email: fltemplin@acm.org