idnits 2.17.1 draft-templin-intarea-6706bis-20.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 (January 27, 2020) is 1544 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Looks like a reference, but probably isn't: '1' on line 1010 -- Looks like a reference, but probably isn't: '2' on line 1014 == Missing Reference: 'N' is mentioned on line 1024, but not defined == Unused Reference: 'RFC2474' is defined on line 3010, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-intarea-gue-extensions' is defined on line 3076, but no explicit reference was found in the text == Unused Reference: 'RFC2764' is defined on line 3158, but no explicit reference was found in the text == Unused Reference: 'RFC4607' is defined on line 3245, but no explicit reference was found in the text == Unused Reference: 'RFC7269' is defined on line 3332, but no explicit reference was found in the text == Unused Reference: 'RFC8086' is defined on line 3348, but no explicit reference was found in the text == Outdated reference: A later version (-15) exists of draft-ietf-dmm-distributed-mobility-anchoring-14 == Outdated reference: A later version (-13) exists of draft-ietf-intarea-tunnels-10 == Outdated reference: A later version (-26) exists of draft-ietf-rtgwg-atn-bgp-05 == Outdated reference: A later version (-11) exists of draft-templin-6man-dhcpv6-ndopt-09 == Outdated reference: A later version (-21) exists of draft-templin-atn-aero-interface-12 == Outdated reference: A later version (-27) exists of draft-templin-v6ops-pdhost-25 -- 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 (~~), 15 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, January 27, 2020 5 rfc6179, rfc6706 (if 6 approved) 7 Intended status: Standards Track 8 Expires: July 30, 2020 10 Asymmetric Extended Route Optimization (AERO) 11 draft-templin-intarea-6706bis-20 13 Abstract 15 This document specifies the operation of IP over tunnel virtual links 16 using Asymmetric Extended Route Optimization (AERO). AERO interfaces 17 use an IPv6 link-local address format that supports operation of the 18 IPv6 Neighbor Discovery (ND) protocol and links ND to IP forwarding. 19 Prefix delegation/registration services are employed for network 20 admission and to manage the routing system. Multilink operation, 21 mobility management, quality of service (QoS) signaling and route 22 optimization are naturally supported through dynamic neighbor cache 23 updates. Standard IP multicasting services are also supported. AERO 24 is a widely-applicable mobile internetworking service especially 25 well-suited to aviation services, mobile Virtual Private Networks 26 (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 July 30, 2020. 45 Copyright Notice 47 Copyright (c) 2020 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (https://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 63 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 64 3. Asymmetric Extended Route Optimization (AERO) . . . . . . . . 10 65 3.1. AERO Link Reference Model . . . . . . . . . . . . . . . . 10 66 3.2. AERO Node Types . . . . . . . . . . . . . . . . . . . . . 12 67 3.3. AERO Routing System . . . . . . . . . . . . . . . . . . . 13 68 3.3.1. IPv4 Compatibility Routing . . . . . . . . . . . . . 15 69 3.4. AERO Addresses . . . . . . . . . . . . . . . . . . . . . 15 70 3.5. Spanning Partitioned AERO Networks (SPAN) . . . . . . . . 17 71 3.5.1. SPAN Compatibility Addressing . . . . . . . . . . . . 21 72 3.6. AERO Interface Characteristics . . . . . . . . . . . . . 21 73 3.7. AERO Interface Initialization . . . . . . . . . . . . . . 25 74 3.7.1. AERO Server/Gateway Behavior . . . . . . . . . . . . 25 75 3.7.2. AERO Proxy Behavior . . . . . . . . . . . . . . . . . 26 76 3.7.3. AERO Client Behavior . . . . . . . . . . . . . . . . 26 77 3.7.4. AERO Relay Behavior . . . . . . . . . . . . . . . . . 26 78 3.8. AERO Interface Neighbor Cache Maintenance . . . . . . . . 26 79 3.9. AERO Interface Encapsulation and Re-encapsulation . . . . 28 80 3.10. AERO Interface Decapsulation . . . . . . . . . . . . . . 29 81 3.11. AERO Interface Data Origin Authentication . . . . . . . . 30 82 3.12. AERO Interface Forwarding Algorithm . . . . . . . . . . . 30 83 3.12.1. Client Forwarding Algorithm . . . . . . . . . . . . 31 84 3.12.2. Proxy Forwarding Algorithm . . . . . . . . . . . . . 31 85 3.12.3. Server/Gateway Forwarding Algorithm . . . . . . . . 32 86 3.12.4. Relay Forwarding Algorithm . . . . . . . . . . . . . 34 87 3.13. AERO Interface MTU and Fragmentation . . . . . . . . . . 34 88 3.13.1. AERO MTU Requirements . . . . . . . . . . . . . . . 37 89 3.14. AERO Interface Error Handling . . . . . . . . . . . . . . 37 90 3.15. AERO Router Discovery, Prefix Delegation and 91 Autoconfiguration . . . . . . . . . . . . . . . . . . . . 40 92 3.15.1. AERO ND/PD Service Model . . . . . . . . . . . . . . 40 93 3.15.2. AERO Client Behavior . . . . . . . . . . . . . . . . 41 94 3.15.3. AERO Server Behavior . . . . . . . . . . . . . . . . 43 95 3.16. The AERO Proxy . . . . . . . . . . . . . . . . . . . . . 45 96 3.16.1. Detecting and Responding to Server Failures . . . . 48 97 3.17. AERO Route Optimization . . . . . . . . . . . . . . . . . 48 98 3.17.1. Route Optimization Initiation . . . . . . . . . . . 49 99 3.17.2. Relaying the NS . . . . . . . . . . . . . . . . . . 49 100 3.17.3. Processing the NS and Sending the NA . . . . . . . . 49 101 3.17.4. Relaying the NA . . . . . . . . . . . . . . . . . . 50 102 3.17.5. Processing the NA . . . . . . . . . . . . . . . . . 50 103 3.17.6. Route Optimization Maintenance . . . . . . . . . . . 51 104 3.18. Neighbor Unreachability Detection (NUD) . . . . . . . . . 52 105 3.19. Mobility Management and Quality of Service (QoS) . . . . 52 106 3.19.1. Mobility Update Messaging . . . . . . . . . . . . . 53 107 3.19.2. Announcing Link-Layer Address and/or QoS Preference 108 Changes . . . . . . . . . . . . . . . . . . . . . . 54 109 3.19.3. Bringing New Links Into Service . . . . . . . . . . 54 110 3.19.4. Removing Existing Links from Service . . . . . . . . 54 111 3.19.5. Moving to a New Server . . . . . . . . . . . . . . . 55 112 3.20. Multicast . . . . . . . . . . . . . . . . . . . . . . . . 56 113 3.20.1. Source-Specific Multicast (SSM) . . . . . . . . . . 56 114 3.20.2. Any-Source Multicast (ASM) . . . . . . . . . . . . . 58 115 3.20.3. Bi-Directional PIM (BIDIR-PIM) . . . . . . . . . . . 58 116 3.21. Operation over Multiple AERO Links (VLANs) . . . . . . . 59 117 3.22. DNS Considerations . . . . . . . . . . . . . . . . . . . 60 118 3.23. Transition Considerations . . . . . . . . . . . . . . . . 60 119 3.24. Detecting and Reacting to Server and Relay Failures . . . 61 120 4. Implementation Status . . . . . . . . . . . . . . . . . . . . 61 121 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 61 122 6. Security Considerations . . . . . . . . . . . . . . . . . . . 62 123 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 63 124 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 65 125 8.1. Normative References . . . . . . . . . . . . . . . . . . 65 126 8.2. Informative References . . . . . . . . . . . . . . . . . 66 127 Appendix A. AERO Alternate Encapsulations . . . . . . . . . . . 73 128 Appendix B. Non-Normative Considerations . . . . . . . . . . . . 75 129 B.1. Implementation Strategies for Route Optimization . . . . 75 130 B.2. Implicit Mobility Management . . . . . . . . . . . . . . 76 131 B.3. Direct Underlying Interfaces . . . . . . . . . . . . . . 76 132 B.4. AERO Clients on the Open Internetwork . . . . . . . . . . 76 133 B.5. Operation on AERO Links with /64 ASPs . . . . . . . . . . 77 134 B.6. AERO Adaptations for SEcure Neighbor Discovery (SEND) . . 77 135 B.7. AERO Critical Infrastructure Considerations . . . . . . . 78 136 B.8. AERO Server Failure Implications . . . . . . . . . . . . 79 137 B.9. AERO Client / Server Architecture . . . . . . . . . . . . 79 138 Appendix C. Change Log . . . . . . . . . . . . . . . . . . . . . 81 139 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 87 141 1. Introduction 143 Asymmetric Extended Route Optimization (AERO) fulfills the 144 requirements of Distributed Mobility Management (DMM) [RFC7333] and 145 route optimization [RFC5522] for aeronautical networking and other 146 network mobility use cases. AERO is based on a Non-Broadcast, 147 Multiple Access (NBMA) virtual link model known as the AERO link. 148 The AERO link is a virtual overlay configured over one or more 149 underlying Internetworks, and nodes on the link can exchange IP 150 packets via tunneling. Multilink operation allows for increased 151 reliability, bandwidth optimization and traffic path diversity. 153 The AERO service comprises Clients, Proxys, Servers, and Gateways 154 that are seen as AERO link neighbors. Each node's AERO interface 155 uses an IPv6 link-local address format (known as the AERO address) 156 that supports operation of the IPv6 Neighbor Discovery (ND) protocol 157 [RFC4861] and links ND to IP forwarding. A node's AERO interface can 158 be configured over multiple underlying interfaces, and may therefore 159 may appear as a single interface with multiple link-layer addresses. 160 Each link-layer address is subject to change due to mobility and/or 161 QoS fluctuations, and link-layer address changes are signaled by ND 162 messaging the same as for any IPv6 link. 164 AERO links provide a cloud-based service where mobile nodes may use 165 any Server acting as a Mobility Anchor Point (MAP) and fixed nodes 166 may use any Gateway on the link for efficient communications. Fixed 167 nodes forward packets destined to other AERO nodes to the nearest 168 Gateway, which forwards them through the cloud. A mobile node's 169 initial packets are forwarded through the MAP, while direct routing 170 is supported through asymmetric extended route optimization while 171 data packets are flowing. Both unicast and multicast communications 172 are supported, and mobile nodes may efficiently move between 173 locations while maintaining continuous communications with 174 correspondents and without changing their IP Address. 176 AERO Relays are interconnected in a secured private BGP overlay 177 routing instance known as the "SPAN". The SPAN provides a hybrid 178 routing/bridging service to join the underlying Internetworks of 179 multiple disjoint administrative domains into a single unified AERO 180 link. Each AERO link instance is characterized by the set of 181 Mobility Service Prefixes (MSPs) common to all mobile nodes. The 182 link extends to the point where a Gateway/MAP is on the optimal route 183 from any correspondent node on the link, and provides a gateway 184 between the underlying Internetwork and the SPAN. To the underlying 185 Internetwork, the Gateway/MAP is the source of a route to its MSP, 186 and hence uplink traffic to the mobile node is naturally routed to 187 the nearest Gateway/MAP. 189 AERO assumes the use of PIM Sparse Mode in support of multicast 190 communication. In support of Source Specific Multicast (SSM) when a 191 Mobile Node is the source, AERO route optimization ensures that a 192 shortest-path multicast tree is established with provisions for 193 mobility and multilink operation. In all other multicast scenarios 194 there are no AERO dependencies. 196 AERO was designed for aeronautical networking for both manned and 197 unmanned aircraft, where the aircraft is treated as a mobile node 198 that can connect an Internet of Things (IoT). AERO is also 199 applicable to a wide variety of other use cases. For example, it can 200 be used to coordinate the Virtual Private Network (VPN) links of 201 mobile nodes (e.g., cellphones, tablets, laptop computers, etc.) that 202 connect into a home enterprise network via public access networks 203 using services such as OpenVPN [OVPN]. Other applicable use cases 204 are also in scope. 206 The following numbered sections present the AERO specification. The 207 appendices at the end of the document are non-normative. 209 2. Terminology 211 The terminology in the normative references applies; the following 212 terms are defined within the scope of this document: 214 IPv6 Neighbor Discovery (ND) 215 an IPv6 control message service for coordinating neighbor 216 relationships between nodes connected to a common link. AERO 217 interfaces use the ND service specified in [RFC4861]. 219 IPv6 Prefix Delegation (PD) 220 a networking service for delegating IPv6 prefixes to nodes on the 221 link. The nominal PD service is DHCPv6 [RFC8415], however 222 alternate services (e.g., based on ND messaging) are also in scope 223 [I-D.templin-v6ops-pdhost][I-D.templin-6man-dhcpv6-ndopt]. Most 224 notably, a form of PD known as "prefix registration" can be used 225 if the Client knows its prefix in advance and can represent it in 226 the IPv6 source address of an ND message. 228 Access Network (ANET) 229 a node's first-hop data link service network, e.g., a radio access 230 network, cellular service provider network, corporate enterprise 231 network, or the public Internet itself. For secured ANETs, link- 232 layer security services such as IEEE 802.1X and physical-layer 233 security prevent unauthorized access internally while border 234 network-layer security services such as firewalls and proxies 235 prevent unauthorized outside access. 237 ANET interface 238 a node's attachment to a link in an ANET. 240 ANET address 241 an IP address assigned to a node's interface connection to an 242 ANET. 244 Internetwork (INET) 245 a connected IP network topology with a coherent routing and 246 addressing plan and that provides a transit backbone service for 247 ANET end systems. INETs also provide an underlay service over 248 which the AERO virtual link is configured. Example INETs include 249 corporate enterprise networks, aviation networks, and the public 250 Internet itself. When there is no administrative boundary between 251 an ANET and the INET, the ANET and INET are one and the same. 253 INET Partition 254 frequently, INETs such as large corporate enterprise networks are 255 sub-divided internally into separate isolated partitions. Each 256 partition is fully connected internally but disconnected from 257 other partitions, and there is no requirement that separate 258 partitions maintain consistent Internet Protocol and/or addressing 259 plans. (An INET partition is the same as a SPAN segment discussed 260 below.) 262 INET interface 263 a node's attachment to a link in an INET. 265 INET address 266 an IP address assigned to a node's interface connection to an 267 INET. 269 AERO link 270 a Non-Broadcast, Multiple Access (NBMA) tunnel virtual overlay 271 configured over one or more underlying INETs. Nodes on the AERO 272 link appear as single-hop neighbors from the perspective of the 273 virtual overlay even though they may be separated by many 274 underlying INET hops. AERO links may be configured over multiple 275 underlying SPAN segments (see below). 277 AERO interface 278 a node's attachment to an AERO link. Since the addresses assigned 279 to an AERO interface are managed for uniqueness, AERO interfaces 280 do not require Duplicate Address Detection (DAD) and therefore set 281 the administrative variable 'DupAddrDetectTransmits' to zero 282 [RFC4862]. 284 underlying interface 285 an ANET or INET interface over which an AERO interface is 286 configured. 288 AERO address 289 an IPv6 link-local address assigned to an AERO interface and 290 constructed as specified in Section 3.4. 292 base AERO address 293 the lowest-numbered AERO address aggregated by the MNP (see 294 Section 3.4). 296 Mobility Service Prefix (MSP) 297 an IP prefix assigned to the AERO link and from which more- 298 specific Mobile Network Prefixes (MNPs) are derived. 300 Mobile Network Prefix (MNP) 301 an IP prefix allocated from an MSP and delegated to an AERO Client 302 or Gateway. 304 AERO node 305 a node that is connected to an AERO link, or that provides 306 services to other nodes on an AERO link. 308 AERO Client ("Client") 309 an AERO node that connects to one or more ANETs and requests MNP 310 PDs from AERO Servers. The Client assigns a Client AERO address 311 to the AERO interface for use in ND exchanges with other AERO 312 nodes and forwards packets to correspondents according to AERO 313 interface neighbor cache state. 315 AERO Server ("Server") 316 an INET node that configures an AERO interface to provide default 317 forwarding services and a Mobility Anchor Point (MAP) for AERO 318 Clients. The Server assigns an administratively-provisioned AERO 319 address to its AERO interface to support the operation of the ND/ 320 PD services, and advertises all of its associated MNPs via BGP 321 peerings with Relays. 323 AERO Gateway ("Gateway") 324 an AERO Server that also provides forwarding services between 325 nodes reached via the AERO link and correspondents on other links. 326 AERO Gateways are provisioned with MNPs (i.e., the same as for an 327 AERO Client) and run a dynamic routing protocol to discover any 328 non-MNP IP routes. In both cases, the Gateway advertises the 329 MSP(s) over INET interfaces, and distributes all of its associated 330 MNPs and non-MNP IP routes via BGP peerings with Relays (i.e., the 331 same as for an AERO Server). 333 AERO Relay ("Relay") 334 a node that provides hybrid routing/bridging services (as well as 335 a security trust anchor) for nodes on an AERO link. As a router, 336 the Relay forwards packets using standard IP forwarding. As a 337 bridge, the Relay forwards packets over the AERO link without 338 decrementing the IPv6 Hop Limit. AERO Relays peer with Servers 339 and other Relays to discover the full set of MNPs for the link as 340 well as any non-MNPs that are reachable via Gateways. 342 AERO Proxy ("Proxy") 343 a node that provides proxying services between Clients in an ANET 344 and Servers in external INETs. The AERO Proxy is a conduit 345 between the ANET and external INETs in the same manner as for 346 common web proxies, and behaves in a similar fashion as for ND 347 proxies [RFC4389]. 349 Spanning Partitioned AERO Networks (SPAN) 350 a means for bridging disjoint INET partitions as segments of a 351 unified AERO link the same as for a bridged campus LAN. The SPAN 352 is a mid-layer IPv6 encapsulation service in the AERO routing 353 system that supports a unified AERO link view for all segments. 354 Each segment in the SPAN is a distinct INET partition. 356 SPAN Service Prefix (SSP) 357 a global or unique local /96 IPv6 prefix assigned to the AERO link 358 to support SPAN services. 360 SPAN Partition Prefix (SPP) 361 a sub-prefix of the SPAN Service Prefix uniquely assigned to a 362 single SPAN segment. 364 SPAN Address 365 a global or unique local IPv6 address taken from a SPAN Partition 366 Prefix and constructed as specified in Section 3.5. SPAN 367 addresses are statelessly derived from AERO addresses, and vice- 368 versa. 370 ingress tunnel endpoint (ITE) 371 an AERO interface endpoint that injects encapsulated packets into 372 an AERO link. 374 egress tunnel endpoint (ETE) 375 an AERO interface endpoint that receives encapsulated packets from 376 an AERO link. 378 link-layer address 379 an IP address used as an encapsulation header source or 380 destination address from the perspective of the AERO interface. 382 When UDP encapsulation is used, the UDP port number is also 383 considered as part of the link-layer address. From the 384 perspective of the AERO interface, the link-layer address is 385 either an INET address for intra-segment encapsulation or a SPAN 386 address for inter-segment encapsulation. 388 network layer address 389 the source or destination address of an encapsulated IP packet 390 presented to the AERO interface. 392 end user network (EUN) 393 an internal virtual or external edge IP network that an AERO 394 Client or Gateway connects to the rest of the network via the AERO 395 interface. The Client/Gateway sees each EUN as a "downstream" 396 network, and sees the AERO interface as the point of attachment to 397 the "upstream" network. 399 Mobile Node (MN) 400 an AERO Client and all of its downstream-attached networks that 401 move together as a single unit, i.e., an end system that connects 402 an Internet of Things. 404 Mobile Router (MR) 405 a MN's on-board router that forwards packets between any 406 downstream-attached networks and the AERO link. 408 Mobility Anchor Point (MAP) 409 an AERO Server that is currently tracking and reporting the 410 mobility events of its associated Mobile Node Clients. 412 Route Optimization Source (ROS) 413 the AERO node nearest the source that initiates route 414 optimization. The ROS may be a Server or Proxy acting on behalf 415 of the source Client. 417 Route Optimization responder (ROR) 418 the AERO node nearest the target destination that responds to 419 route optimization requests. The ROR may be a Server acting as a 420 MAP on behalf of a target MNP Client, or a Gateway for a non-MNP 421 destination. 423 MAP List 424 a geographically and/or topologically referenced list of AERO 425 addresses of all MAPs within the same AERO link. There is a 426 single MAP list for the entire AERO link. 428 ROS List 429 a list of AERO/SPAN-to-INET address mappings of all ROSes within 430 the same SPAN segment. There is a distinct ROS list for each 431 segment. 433 Distributed Mobility Management (DMM) 434 a BGP-based overlay routing service coordinated by Servers and 435 Relays that tracks all MAP-to-Client associations. 437 Throughout the document, the simple terms "Client", "Server", 438 "Relay", "Proxy" and "Gateway" refer to "AERO Client", "AERO Server", 439 "AERO Relay", "AERO Proxy" and "AERO Gateway", respectively. 440 Capitalization is used to distinguish these terms from other common 441 Internetworking uses in which they appear without capitalization. 443 The terminology of DHCPv6 [RFC8415] and IPv6 ND [RFC4861] (including 444 the names of node variables, messages and protocol constants) is used 445 throughout this document. Also, the term "IP" is used to generically 446 refer to either Internet Protocol version, i.e., IPv4 [RFC0791] or 447 IPv6 [RFC8200]. 449 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 450 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 451 "OPTIONAL" in this document are to be interpreted as described in BCP 452 14 [RFC2119][RFC8174] when, and only when, they appear in all 453 capitals, as shown here. 455 3. Asymmetric Extended Route Optimization (AERO) 457 The following sections specify the operation of IP over Asymmetric 458 Extended Route Optimization (AERO) links: 460 3.1. AERO Link Reference Model 461 +----------------+ 462 | AERO Relay R1 | 463 | Nbr: S1, S2, P1| 464 |(X1->S1; X2->S2)| 465 | MSP M1 | 466 +-+---------+--+-+ 467 +--------------+ | Secured | | +--------------+ 468 |AERO Server S1| | tunnels | | |AERO Server S2| 469 | Nbr: C1, R1 +-----+ | +-----+ Nbr: C2, R1 | 470 | default->R1 | | | default->R1 | 471 | X1->C1 | | | X2->C2 | 472 +-------+------+ | +------+-------+ 473 | AERO Link | | 474 X===+===+===================+==)===============+===+===X 475 | | | | 476 +-----+--------+ +--------+--+-----+ +--------+-----+ 477 |AERO Client C1| | AERO Proxy P1 | |AERO Client C2| 478 | Nbr: S1 | |(Proxy Nbr Cache)| | Nbr: S2 | 479 | default->S1 | +--------+--------+ | default->S2 | 480 | MNP X1 | | | MNP X2 | 481 +------+-------+ .--------+------. +-----+--------+ 482 | (- Proxyed Clients -) | 483 .-. `---------------' .-. 484 ,-( _)-. ,-( _)-. 485 .-(_ IP )-. +-------+ +-------+ .-(_ IP )-. 486 (__ EUN )--|Host H1| |Host H2|--(__ EUN ) 487 `-(______)-' +-------+ +-------+ `-(______)-' 489 Figure 1: AERO Link Reference Model 491 Figure 1 presents the AERO link reference model. In this model: 493 o the AERO link is an overlay network service configured over one or 494 more underlying INET partitions which may be managed by different 495 administrative authorities and have incompatible protocols and/or 496 addressing plans. 498 o AERO Relay R1 aggregates Mobility Service Prefix (MSP) M1, 499 discovers Mobile Network Prefixes (MNPs) X* and advertises the MSP 500 via BGP peerings over secured tunnels to Servers (S1, S2). Relays 501 use the SPAN service to bridge disjoint segments of a partitioned 502 AERO link. 504 o AERO Servers S1 and S2 configure secured tunnels with Relay R1 and 505 also act as Mobility Anchor Points (MAPs) and default routers for 506 their associated Clients C1 and C2. 508 o AERO Clients C1 and C2 associate with Servers S1 and S2, 509 respectively. They receive Mobile Network Prefix (MNP) 510 delegations X1 and X2, and also act as default routers for their 511 associated physical or internal virtual EUNs. Simple hosts H1 and 512 H2 attach to the EUNs served by Clients C1 and C2, respectively. 514 o AERO Proxy P1 configures a secured tunnel with Relay R1 and 515 provides proxy services for AERO Clients in secured enclaves that 516 cannot associate directly with other AERO link neighbors. 518 Each node on the AERO link maintains an AERO interface neighbor cache 519 and an IP forwarding table the same as for any link. Although the 520 figure shows a limited deployment, in common operational practice 521 there will normally be many additional Relays, Servers, Clients and 522 Proxys. 524 3.2. AERO Node Types 526 AERO Relays provide hybrid routing/bridging services (as well as a 527 security trust anchor) for nodes on an AERO link. Relays use 528 standard IPv6 routing to forward packets both within the same INET 529 partitions and between disjoint INET partitions based on a mid-layer 530 IPv6 encapsulation known as the SPAN header. The inner IP layer 531 experiences a virtual bridging service since the inner IP TTL/Hop 532 Limit is not decremented during forwarding. Each Relay also peers 533 with Servers and other Relays in a dynamic routing protocol instance 534 to provide a Distributed Mobility Management (DMM) service for the 535 list of active MNPs (see Section 3.3). Relays present the AERO link 536 as a set of one or more Mobility Service Prefixes (MSPs) but as link- 537 layer devices need not connect directly to the AERO link themselves 538 unless an administrative interface is desired. Relays configure 539 secured tunnels with Servers, Proxys and other Relays; they further 540 maintain IP forwarding table entries for each Mobile Network Prefix 541 (MNP) and any other reachable non-MNP prefixes. 543 AERO Servers provide default forwarding services and a Mobility 544 Anchor Point (MAP) for AERO Client Mobile Nodes (MNs). Each Server 545 also peers with Relays in a dynamic routing protocol instance to 546 advertise its list of associated MNPs (see Section 3.3). Servers 547 facilitate PD exchanges with Clients, where each delegated prefix 548 becomes an MNP taken from an MSP. Servers forward packets between 549 AERO interface neighbors and track each Client's mobility profiles. 551 AERO Clients register their MNPs through PD exchanges with AERO 552 Servers over the AERO link, and distribute the MNPs to nodes on EUNs. 553 A Client may also be co-resident on the same physical or virtual 554 platform as a Server; in that case, the Client and Server behave as a 555 single functional unit and without the need for any Client/Server 556 control messaging. 558 AERO Proxys provide a conduit for ANET AERO Clients to associate with 559 AERO Servers in external INETs. Client and Servers exchange control 560 plane messages via the Proxy acting as a bridge between the ANET/INET 561 boundary. The Proxy forwards data packets to and from Clients 562 according to forwarding information in the neighbor cache. The Proxy 563 function is specified in Section 3.16. 565 AERO Gateways are Servers that provide forwarding services between 566 the AERO interface and INET/EUN interfaces. Gateways are provisioned 567 with MNPs the same as for an AERO Client, and also run a dynamic 568 routing protocol to discover any non-MNP IP routes. The Gateway 569 advertises the MSP(s) to INETs, and distributes all of its associated 570 MNPs and non-MNP IP routes via BGP peerings with Relays. 572 AERO Relays, Servers, Proxys and Gateways are critical infrastructure 573 elements in fixed (i.e., non-mobile) INET deployments and hence have 574 permanent and unchanging INET addresses. AERO Clients are MNs that 575 connect via ANET interfaces, i.e., their ANET addresses may change 576 when the Client moves to a new ANET connection. 578 3.3. AERO Routing System 580 The AERO routing system comprises a private instance of the Border 581 Gateway Protocol (BGP) [RFC4271] that is coordinated between Relays 582 and Servers and does not interact with either the public Internet BGP 583 routing system or any underlying INET routing systems. 585 In a reference deployment, each Server is configured as an Autonomous 586 System Border Router (ASBR) for a stub Autonomous System (AS) using 587 an AS Number (ASN) that is unique within the BGP instance, and each 588 Server further uses eBGP to peer with one or more Relays but does not 589 peer with other Servers. Each INET of a multi-segment AERO link must 590 include one or more Relays, which peer with the Servers and Proxys 591 within that INET. All Relays within the same INET are members of the 592 same hub AS using a common ASN, and use iBGP to maintain a consistent 593 view of all active MNPs currently in service. The Relays of 594 different INETs peer with one another using eBGP. 596 Relays advertise the AERO link's MSPs and any non-MNP routes to each 597 of their Servers. This means that any aggregated non-MNPs (including 598 "default") are advertised to all Servers. Each Relay configures a 599 black-hole route for each of its MSPs. By black-holing the MSPs, the 600 Relay will maintain forwarding table entries only for the MNPs that 601 are currently active, and packets destined to all other MNPs will 602 correctly incur Destination Unreachable messages due to the black- 603 hole route. In this way, Servers have only partial topology 604 knowledge (i.e., they know only about the MNPs of their directly 605 associated Clients) and they forward all other packets to Relays 606 which have full topology knowledge. 608 Servers maintain a working set of associated MNPs, and dynamically 609 announce new MNPs and withdraw departed MNPs in eBGP updates to 610 Relays. Servers that are configured as Gateways also redistribute 611 non-MNP routes learned from non-AERO interfaces via their eBGP Relay 612 peerings. 614 Clients are expected to remain associated with their current Servers 615 for extended timeframes, however Servers SHOULD selectively suppress 616 updates for impatient Clients that repeatedly associate and 617 disassociate with them in order to dampen routing churn. Servers 618 that are configured as Gateways advertise the MSPs via INET/EUN 619 interfaces, and forward packets between INET/EUN interfaces and the 620 AERO interface using standard IP forwarding. 622 Scaling properties of the AERO routing system are limited by the 623 number of BGP routes that can be carried by Relays. As of 2015, the 624 global public Internet BGP routing system manages more than 500K 625 routes with linear growth and no signs of router resource exhaustion 626 [BGP]. More recent network emulation studies have also shown that a 627 single Relay can accommodate at least 1M dynamically changing BGP 628 routes even on a lightweight virtual machine, i.e., and without 629 requiring high-end dedicated router hardware. 631 Therefore, assuming each Relay can carry 1M or more routes, this 632 means that at least 1M Clients can be serviced by a single set of 633 Relays. A means of increasing scaling would be to assign a different 634 set of Relays for each set of MSPs. In that case, each Server still 635 peers with one or more Relays, but institutes route filters so that 636 BGP updates are only sent to the specific set of Relays that 637 aggregate the MSP. For example, if the MSP for the AERO link is 638 2001:db8::/32, a first set of Relays could service the MSP 639 2001:db8::/40, a second set of Relays could service 640 2001:db8:0100::/40, a third set could service 2001:db8:0200::/40, 641 etc. 643 Assuming up to 1K sets of Relays, the AERO routing system can then 644 accommodate 1B or more MNPs with no additional overhead (for example, 645 it should be possible to service 1B /64 MNPs taken from a /34 MSP and 646 even more for shorter prefixes). In this way, each set of Relays 647 services a specific set of MSPs that they advertise to the native 648 Internetwork routing system, and each Server configures MSP-specific 649 routes that list the correct set of Relays as next hops. This 650 arrangement also allows for natural incremental deployment, and can 651 support small scale initial deployments followed by dynamic 652 deployment of additional Clients, Servers and Relays without 653 disturbing the already-deployed base. 655 Server and Relays can use the Bidirectional Forwarding Detection 656 (BFD) protocol [RFC5880] to quickly detect link failures that don't 657 result in interface state changes, BGP peer failures, and 658 administrative state changes. BFD is important in environments where 659 rapid response to failures is required for routing reconvergence and, 660 hence, communications continuity. 662 A full discussion of the BGP-based routing system used by AERO is 663 found in [I-D.ietf-rtgwg-atn-bgp]. The system provides for 664 Distributed Mobility Management (DMM) per the distributed mobility 665 anchoring architecture [I-D.ietf-dmm-distributed-mobility-anchoring]. 667 3.3.1. IPv4 Compatibility Routing 669 For IPv6 MNPs, the AERO routing system includes ordinary IPv6 routes. 670 For IPv4 MNPs, the AERO routing system includes IPv6 routes based on 671 an IPv4-embedded IPv6 address format discussed in Section 3.5.1. 673 3.4. AERO Addresses 675 A Client's AERO address is an IPv6 link-local address with an 676 interface identifier based on the Client's delegated MNP. Relay, 677 Server and Proxy AERO addresses are assigned from the range fe80::/96 678 and include an administratively-provisioned value in the lower 32 679 bits. 681 For IPv6, Client AERO addresses begin with the prefix fe80::/64 and 682 include in the interface identifier (i.e., the lower 64 bits) a 683 64-bit prefix taken from one of the Client's IPv6 MNPs. For example, 684 if the AERO Client receives the IPv6 MNP: 686 2001:db8:1000:2000::/56 688 it constructs its corresponding AERO addresses as: 690 fe80::2001:db8:1000:2000 692 fe80::2001:db8:1000:2001 694 fe80::2001:db8:1000:2002 696 ... etc. ... 698 fe80::2001:db8:1000:20ff 700 For IPv4, Client AERO addresses are based on an IPv4-mapped IPv6 701 address [RFC4291] formed from an IPv4 MNP and with a Prefix Length of 702 96 plus the MNP prefix length. For example, for the IPv4 MNP 703 192.0.2.32/28 the IPv4-mapped IPv6 MNP is: 705 0:0:0:0:0:FFFF:192.0.2.16/124 (also written as 706 0:0:0:0:0:FFFF:c000:0210/124) 708 The Client then constructs its AERO addresses with the prefix 709 fe80::/64 and with the lower 64 bits of the IPv4-mapped IPv6 address 710 in the interface identifier as: 712 fe80::FFFF:192.0.2.16 714 fe80::FFFF:192.0.2.17 716 fe80::FFFF:192.0.2.18 718 ... etc. ... 720 fe80:FFFF:192.0.2.31 722 Relay, Server and Proxy AERO addresses are allocated from the range 723 fe80::/96, and MUST be managed for uniqueness. The lower 32 bits of 724 the AERO address includes a unique integer value between 1 and 0xfffe 725 (e.g., fe80::1, fe80::2, fe80::3, etc.) as assigned by the 726 administrative authority for the link. If the link spans multiple 727 SPAN segments, the AERO addresses are assigned to each segment in 1x1 728 correspondence with SPAN addresses (see: Section 3.5). The address 729 fe80:: is the IPv6 link-local Subnet Router Anycast address, and the 730 address fe80::ffff:ffff is reserved as the unspecified AERO address. 732 The lowest-numbered AERO address from a Client's MNP delegation 733 serves as the "base" AERO address (for example, for the MNP 734 2001:db8:1000:2000::/56 the base AERO address is 735 fe80::2001:db8:1000:2000). The Client then assigns the base AERO 736 address to the AERO interface and uses it for the purpose of 737 maintaining the neighbor cache entry. The Server likewise uses the 738 AERO address as its index into the neighbor cache for this Client. 740 If the Client has multiple AERO addresses (i.e., when there are 741 multiple MNPs and/or MNPs with prefix lengths shorter than /64), the 742 Client originates ND messages using the base AERO address as the 743 source address and accepts and responds to ND messages destined to 744 any of its AERO addresses as equivalent to the base AERO address. In 745 this way, the Client maintains a single neighbor cache entry that may 746 be indexed by multiple AERO addresses. 748 The Client's Subnet Router Anycast address can be statelessly 749 determined from its AERO address by simply transposing the AERO 750 address into the upper N bits of the Anycast address followed by 751 128-N bits of zeroes. For example, for the AERO address 752 fe80::2001:db8:1:2 the subnet router anycast address is 753 2001:db8:1:2::. 755 AERO addresses for mobile node Clients embed a MNP as discussed 756 above, while AERO addresses for non-MNP destinations are constructed 757 in exactly the same way. A Client AERO address therefore encodes 758 either an MNP if the prefix is reached via the SPAN or a non-MNP if 759 the prefix is reached via a Gateway. 761 3.5. Spanning Partitioned AERO Networks (SPAN) 763 An AERO link configured over a single INET appears as a single 764 unified link with a consistent underlying network addressing plan. 765 In that case, all nodes on the link can exchange packets via 766 encapsulation with INET addresses, since the underlying INET is 767 connected. In common practice, however, an AERO link may be 768 partitioned into multiple "segments", where each segment is a 769 distinct INET potentially managed under a different administrative 770 authority (e.g., as for worldwide aviation service providers such as 771 ARINC, SITA, Inmarsat, etc.). Individual INETs may themselves be 772 partitioned internally, in which case each internal partition is seen 773 as a separate segment. 775 The addressing plan of each segment is consistent internally but will 776 often bear no relation to the addressing plans of other segments. 777 Each segment is also likely to be separated from others by network 778 security devices (e.g., firewalls, proxies, packet filtering 779 gateways, etc.), and in many cases disjoint segments may not even 780 have any common physical link connections at all. Therefore, nodes 781 can only be assured of exchanging packets directly with 782 correspondents in the same segment, and not with those in other 783 segments. The only means for joining the segments therefore is 784 through inter-domain peerings between AERO Relays. 786 The same as for traditional campus LANs, multiple AERO link segments 787 can be joined into a single unified link via a virtual bridging 788 service termed the "SPAN". The SPAN performs link-layer packet 789 forwarding between segments (i.e., bridging) without decrementing the 790 network-layer TTL/Hop Limit. The SPAN model is depicted in Figure 2: 792 . . . . . . . . . . . . . . . . . . . . . . . 793 . . 794 . .-(::::::::) . 795 . .-(::::::::::::)-. +-+ . 796 . (:::: Segment A :::)--|R|---+ . 797 . `-(::::::::::::)-' +-+ | . 798 . `-(::::::)-' | . 799 . | . 800 . .-(::::::::) | . 801 . .-(::::::::::::)-. +-+ | . 802 . (:::: Segment B :::)--|R|---+ . 803 . `-(::::::::::::)-' +-+ | . 804 . `-(::::::)-' | . 805 . | . 806 . .-(::::::::) | . 807 . .-(::::::::::::)-. +-+ | . 808 . (:::: Segment C :::)--|R|---+ . 809 . `-(::::::::::::)-' +-+ | . 810 . `-(::::::)-' | . 811 . | . 812 . ..(etc).. x . 813 . . 814 . . 815 . <- AERO Link Bridged by the SPAN -> . 816 . . . . . . . . . . . . . .. . . . . . . . . 818 Figure 2: The SPAN 820 To support the SPAN, AERO links require a reserved /64 IPv6 "SPAN 821 Service Prefix (SSP)". Although any routable IPv6 prefix can be 822 used, a Unique Local Address (ULA) prefix (e.g., fd00::/64) [RFC4389] 823 is recommended since border routers are commonly configured to 824 prevent packets with ULAs from being injected into the AERO link by 825 an external IPv6 node and from leaking out of the AERO link to the 826 outside world. 828 Each segment in the SPAN assigns a unique sub-prefix of SSP::/96 829 termed a "SPAN Partition Prefix (SPP)". For example, a first segment 830 could assign fd00::1000/116, a second could assign fd00::2000/116, a 831 third could assign fd00::3000/116, etc. The administrative 832 authorities for each segment must therefore coordinate to assure 833 mutually-exclusive SPP assignments, but internal provisioning of the 834 SPP is an independent local consideration for each administrative 835 authority. 837 A "SPAN address" is an address taken from a SPP and assigned to a 838 Relay, Server or Proxy interface. SPAN addresses are formed by 839 simply replacing the upper portion of an administratively-assigned 840 AERO address with the SPP. For example, if the SPP is 841 fd00::1000/116, the SPAN address formed from the AERO address 842 fe80::1001 is simply fd00::1001. 844 An "INET address" is an address of a node's interface connection to 845 an INET. AERO/SPAN/INET address mappings are maintained as permanent 846 neighbor cache entires as discussed in Section 3.8. 848 AERO Relays serve as bridges to join multiple segments into a unified 849 AERO link over multiple diverse administrative domains. They support 850 the bridging function by first establishing forwarding table entries 851 for their SPPs either via standard BGP routing or static routes. For 852 example, if three Relays ('A', 'B' and 'C') from different segments 853 serviced the SPPs fd00::1000/116, fd00::2000/116 and fd00::3000/116 854 respectively, then the forwarding tables in each Relay are as 855 follows: 857 A: fd00::1000/116->local, fd00::2000/116->B, fd00::3000/116->C 859 B: fd00::1000/116->A, fd00::2000/116->local, fd00::3000/116->C 861 C: fd00::1000/116->A, fd00::2000/116->B, fd00::3000/116->local 863 These forwarding table entries are permanent and never change, since 864 they correspond to fixed infrastructure elements in their respective 865 segments. This provides the basis for a link-layer forwarding 866 service that cannot be disrupted by routing updates due to node 867 mobility. 869 With the SPPs in place in each Relay's forwarding table, control and 870 data packets sent between AERO nodes in different segments can 871 therefore be carried over the SPAN via encapsulation. For example, 872 when a source node in segment A forwards a packet with IPv6 address 873 2001:db8:1:2::1 to a destination node in segment C with IPv6 address 874 2001:db8:1000:2000::1, it first encapsulates the packet in a SPAN 875 header with source SPAN address taken from fd00::1000/116 (e.g., 876 fd00::1001) and destination SPAN address taken from fd00::3000/116 877 (e.g., fd00::3001). Next, it encapsulates the SPAN message in an 878 INET header with source address set to its own INET address (e.g., 879 192.0.2.100) and destination set to the INET address of a Relay 880 (e.g., 192.0.2.1). 882 SPAN encapsulation is based on Generic Packet Tunneling in IPv6 883 [RFC2473]; the encapsulation format in the above example is shown in 884 Figure 3: 886 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 887 | INET Header | 888 | src = 192.0.2.100 | 889 | dst = 192.0.2.1 | 890 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 891 | SPAN Header | 892 | src = fd00::1001 | 893 | dst = fd00::3001 | 894 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 895 | Inner IP Header | 896 | src = 2001:db8:1:2::1 | 897 | dst = 2001:db8:1000:2000::1 | 898 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 899 | | 900 ~ ~ 901 ~ Inner Packet Body ~ 902 ~ ~ 903 | | 904 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 906 Figure 3: SPAN Encapsulation 908 In this format, the inner IP header and packet body are the original 909 IP packet, the SPAN header is an IPv6 header prepared according to 910 [RFC2473], and the INET header is prepared according to Section 3.9. 911 A packet is said to be "forwarded/sent into the SPAN" when it is 912 encapsulated as described above then forwarded via a secured tunnel 913 to a neighboring Relay. 915 This gives rise to a routing system that contains both MNP routes 916 that may change dynamically due to regional node mobility and SPAN 917 routes that never change. The Relays can therefore provide link- 918 layer bridging by sending packets into the SPAN instead of network- 919 layer routing according to MNP routes. As a result, opportunities 920 for packet loss due to node mobility between different segments are 921 mitigated. 923 With reference to Figure 3, for a Client's AERO address the SPAN 924 address is simply set to the Subnet Router Anycast address. For non- 925 link-local addresses, the destination SPAN address may not be known 926 in advance for the first few packets of a flow sent via the SPAN. In 927 that case, the SPAN destination address is set to the original 928 packet's destination, and the SPAN routing system will direct the 929 packet to the correct SPAN egress node. (In the above example, the 930 SPAN destination address is simply 2001:db8:1000:2000::1.) 932 3.5.1. SPAN Compatibility Addressing 934 For IPv4 MNPs, Servers injects a "SPAN Compatibility Prefix (SCP)" 935 that embeds the MNP into the BGP routing system. The SCP begins with 936 the upper 64 bits of the SSP, followed by the constant string 937 "0000:FFFF" followed by the IPv4 MNP. For example, if the SSP is 938 fd00::/64 and the MNP is 192.0.2.0/24 then the SCP is 939 fd00::FFFF:192.0.2.0/120. 941 This allows for encapsulation of IPv4 packets in IPv6 headers with 942 "SPAN Compatibility Addresses (SCAs)". In this example, the SCA 943 corresponding to the SCP is simply fd00::FFFF:192.0.2.0, and can be 944 used as the SPAN destination address for packets forwarded via the 945 SPAN. This allows for forwarding of initial IPv4 packets over IPv6 946 SPAN routes, followed by route optimization for direct 947 communications. 949 3.6. AERO Interface Characteristics 951 AERO interfaces are virtual interfaces configured over one or more 952 underlying interfaces classified as follows: 954 o Native interfaces have global IP addresses that are reachable from 955 any INET correspondent. All Server and Relay interfaces are 956 native interfaces, as are INET-facing interfaces of Proxys. 958 o NATed interfaces connect to a private network behind a Network 959 Address Translator (NAT). The NAT does not participate in any 960 AERO control message signaling, but the Server can issue control 961 messages on behalf of the Client. Clients that are behind a NAT 962 are required to send periodic keepalive messages to keep NAT state 963 alive when there are no data packets flowing. If no other 964 periodic messaging service is available, the Client can send RS 965 messages to receive RA replies from its Server(s). 967 o VPNed interfaces use security encapsulation to a Virtual Private 968 Network (VPN) server that also acts as an AERO Server. As with 969 NATed links, the Server can issue control messages on behalf of 970 the Client, but the Client need not send periodic keepalives in 971 addition to those already used to maintain the VPN connection. 973 o Proxyed interfaces connect to an ANET that is separated from the 974 open INET by an AERO Proxy. Unlike NATed and VPNed interfaces, 975 the Proxy can actively issue control messages on behalf of the 976 Client. 978 o Direct interfaces connect a Client directly to a neighbor without 979 crossing any ANET/INET paths. An example is a line-of-sight link 980 between a remote pilot and an unmanned aircraft. 982 AERO interfaces use encapsulation (see: Section 3.9) to exchange 983 packets with AERO link neighbors over Native, NATed or VPNed 984 interfaces. AERO interfaces do not use encapsulation over Proxyed 985 and Direct underlying interfaces. 987 AERO interfaces maintain a neighbor cache for tracking per-neighbor 988 state the same as for any interface. AERO interfaces use ND messages 989 including Router Solicitation (RS), Router Advertisement (RA), 990 Neighbor Solicitation (NS) and Neighbor Advertisement (NA) for 991 neighbor cache management. 993 AERO interfaces send ND messages with an Overlay Multilink Network 994 Interface (OMNI) option formatted as specified in 995 [I-D.templin-atn-aero-interface]. The OMNI option includes prefix 996 registration information and "ifIndex-tuples" containing link quality 997 information for the AERO interface's underlying interfaces. 999 When encapsulation is used, AERO interface ND messages MAY also 1000 include an AERO Source/Target Link-Layer Address Option (S/TLLAO) 1001 formatted as shown in Figure 4: 1003 0 1 2 3 1004 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 1005 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1006 | Type | Length | ifIndex[1] |V| Reserved[1] | 1007 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1008 ~ Link Layer Address [1] ~ 1009 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1010 | Port Number [1] | ifIndex[2] |V| Reserved[2] | 1011 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1012 ~ Link Layer Address [2] ~ 1013 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1014 | Port Number [2] | ~ 1015 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ ~ 1016 ~ ~ 1017 ~ ... ~ 1018 ~ ~ 1019 ~ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1020 ~ | ifIndex[N] |V| Reserved[N] | 1021 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1022 ~ Link Layer Address [N] ~ 1023 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1024 | Port Number [N] | Trailing zero padding | 1025 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1026 | Trailing zero padding (if necessary) | 1027 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1029 Figure 4: AERO Source/Target Link-Layer Address Option (S/TLLAO) 1030 Format 1032 In this format, Type and Length are set the same as specified for S/ 1033 TLLAOs in [RFC4861], with trailing zero padding octets added as 1034 necessary to produce an integral number of 8 octet blocks. The S/ 1035 TLLAO includes N ifIndex-tuples in correspondence to ifIndex-tuples 1036 that appear in the OMNI option. Each ifIndex-tuple includes the 1037 folllowing information: 1039 o ifIndex[i] - the same value as in the corresponding ifIndex-tuple 1040 included in the OMNI option. 1042 o V[i] - a bit that identifies the IP protocol version of the 1043 address found in the Link Layer Address [i] field. The bit is set 1044 to 0 for IPv4 or 1 for IPv6. 1046 o Reserved[i] - MUST encode the value 0 on transmission, and ignored 1047 on reception. 1049 o Link Layer Address [i] - the IPv4 or IPv6 address used as the 1050 encapsulation source address. The field is 4 bytes in length for 1051 IPv4 or 16 bytes in length for IPv6. 1053 o Port Number [i] - the upper layer protocol port number used as the 1054 encapsulation source port, or 0 when no upper layer protocol 1055 encapsulation is used. The field is 2 bytes in length. 1057 If present, the first S/TLLAO ifIndex-tuple MUST correspond to the 1058 first OMNI option ifIndex-tuple, and any additional S/TLLAO ifIndex- 1059 tuples MUST correspond to a proper subset of the remaining OMNI 1060 option ifIndex-tuples. 1062 A Client's AERO interface may be configured over multiple underlying 1063 interface connections. For example, common mobile handheld devices 1064 have both wireless local area network ("WLAN") and cellular wireless 1065 links. These links are typically used "one at a time" with low-cost 1066 WLAN preferred and highly-available cellular wireless as a standby. 1067 In a more complex example, aircraft frequently have many wireless 1068 data link types (e.g. satellite-based, cellular, terrestrial, air-to- 1069 air directional, etc.) with diverse performance and cost properties. 1071 If a Client's multiple underlying interfaces are used "one at a time" 1072 (i.e., all other interfaces are in standby mode while one interface 1073 is active), then ND message OMNI options include only a single 1074 ifIndex-tuple and set to a constant value. In that case, the Client 1075 would appear to have a single interface but with a dynamically 1076 changing link-layer address. 1078 If the Client has multiple active underlying interfaces, then from 1079 the perspective of ND it would appear to have multiple link-layer 1080 addresses. In that case, ND message OMNI options MAY include 1081 multiple ifIndex-tuples - each with a value that corresponds to a 1082 specific interface. The OMNI option MUST include a first ifIndex- 1083 tuple that corresponds to the interface over which the ND message is 1084 sent. Every ND message need not include all OMNI and/or S/TLLAO 1085 ifIndex-tuples; for any ifIndex-tuple not included, the neighbor 1086 considers the status as unchanged. 1088 Relay, Server and Proxy AERO interfaces may be configured over one or 1089 more secured tunnel interfaces. The AERO interface configures both 1090 an AERO address and its corresponding SPAN address, while the 1091 underlying secured tunnel interfaces are either unnumbered or 1092 configure the same SPAN address. The AERO interface encapsulates 1093 each IP packet in a SPAN header and presents the packet to the 1094 underlying secured tunnel interface. For Relays that do not 1095 configure an AERO interface, the secured tunnel interfaces themselves 1096 are exposed to the IP layer with each interface configuring the 1097 Relay's SPAN address. Routing protocols such as BGP therefore run 1098 directly over the Relay's secured tunnel interfaces. For nodes that 1099 configure an AERO interface, routing protocols such as BGP run over 1100 the AERO interface but do not employ SPAN encapsulation. Instead, 1101 the AERO interface presents the routing protocol messages directly to 1102 the underlying secured tunnels without applying encapsulation and 1103 while using the SPAN address as the source address. This distinction 1104 must be honored consistently according to each node's configuration 1105 so that the IP forwarding table will associate discovered IP routes 1106 with the correct interface. 1108 3.7. AERO Interface Initialization 1110 AERO Servers, Proxys and Clients configure AERO interfaces as their 1111 point of attachment to the AERO link. AERO nodes assign the MSPs for 1112 the link to their AERO interfaces (i.e., as a "route-to-interface") 1113 to ensure that packets with destination addresses covered by an MNP 1114 not explicitly assigned to a non-AERO interface are directed to the 1115 AERO interface. 1117 AERO interface initialization procedures for Servers, Proxys, Clients 1118 and Relays are discussed in the following sections. 1120 3.7.1. AERO Server/Gateway Behavior 1122 When a Server enables an AERO interface, it assigns AERO/SPAN 1123 addresses and configures permanent neighbor cache entries for 1124 neighbors in the same SPAN segment by consulting the ROS list for the 1125 segment. The Server also configures secured tunnels with one or more 1126 neighboring Relays and engages in a BGP routing protocol session with 1127 each Relay. 1129 The AERO interface provides a single interface abstraction to the IP 1130 layer, but internally comprises multiple secured tunnels as well as 1131 an NBMA nexus for sending encapsulated data packets to AERO interface 1132 neighbors. The Server further configures a service to facilitate ND/ 1133 PD exchanges with AERO Clients and manages per-Client neighbor cache 1134 entries and IP forwarding table entries based on control message 1135 exchanges. 1137 Gateways are simply Servers that run a dynamic routing protocol 1138 between the AERO interface and INET/EUN interfaces (see: 1139 Section 3.3). The Gateway provisions MNPs to networks on the INET/ 1140 EUN interfaces (i.e., the same as a Client would do) and advertises 1141 the MSP(s) for the AERO link over the INET/EUN interfaces. The 1142 Gateway further provides an attachment point of the AERO link to the 1143 non-MNP-based global topology. 1145 3.7.2. AERO Proxy Behavior 1147 When a Proxy enables an AERO interface, it assigns AERO/SPAN 1148 addresses and configures permanent neighbor cache entries the same as 1149 for Servers. The Proxy also configures secured tunnels with one or 1150 more neighboring Relays and maintains per-Client neighbor cache 1151 entries based on control message exchanges. 1153 3.7.3. AERO Client Behavior 1155 When a Client enables an AERO interface, it sends an RS message with 1156 ND/PD parameters over an ANET interface to a Server in the MAP list, 1157 which returns an RA message with corresponding PD parameters. (The 1158 RS/RA messages may pass through a Proxy in the case of a Client's 1159 Proxyed interface.) 1161 After the initial ND/PD message exchange, the Client assigns AERO 1162 addresses to the AERO interface based on the delegated prefix(es). 1163 The Client can then register additional ANET interfaces with the 1164 Server by sending an RS message over each ANET interface. 1166 3.7.4. AERO Relay Behavior 1168 AERO Relays need not connect directly to the AERO link, since they 1169 operate as link-layer forwarding devices instead of network layer 1170 routers. Configuration of AERO interfaces on Relays is therefore 1171 OPTIONAL, e.g., if an administrative interface is needed. Relays 1172 configure secured tunnels with Servers, Proxys and other Relays; they 1173 also configure AERO/SPAN addresses and permanent neighbor cache 1174 entries the same as Servers. Relays engage in a BGP routing protocol 1175 session with a subset of the Servers on the local SPAN segment, and 1176 with other Relays on the SPAN (see: Section 3.3). 1178 3.8. AERO Interface Neighbor Cache Maintenance 1180 Each AERO interface maintains a conceptual neighbor cache that 1181 includes an entry for each neighbor it communicates with on the AERO 1182 link per [RFC4861]. AERO interface neighbor cache entries are said 1183 to be one of "permanent", "symmetric", "asymmetric" or "proxy". 1185 Permanent neighbor cache entries are created through explicit 1186 administrative action; they have no timeout values and remain in 1187 place until explicitly deleted. AERO Servers and Proxys maintain 1188 permanent neighbor cache entries for all other Servers and Proxys 1189 within the same SPAN segment. Each entry maintains the mapping 1190 between the neighbor's network-layer AERO address and corresponding 1191 INET address. The list of all permanent neighbor cache entries for 1192 the SPAN segment is maintained in the segment's ROS list. 1194 Symmetric neighbor cache entries are created and maintained through 1195 RS/RA exchanges as specified in Section 3.15, and remain in place for 1196 durations bounded by ND/PD lifetimes. AERO Servers maintain 1197 symmetric neighbor cache entries for each of their associated 1198 Clients, and AERO Clients maintain symmetric neighbor cache entries 1199 for each of their associated Servers. The list of all Servers on the 1200 AERO link is maintained in the link's MAP list. 1202 Asymmetric neighbor cache entries are created or updated based on 1203 route optimization messaging as specified in Section 3.17, and are 1204 garbage-collected when keepalive timers expire. AERO route 1205 optimization sources (ROSs) maintain asymmetric neighbor cache 1206 entries for active targets with lifetimes based on ND messaging 1207 constants. Asymmetric neighbor cache entries are unidirectional 1208 since only the ROS and not the target (e.g., a Client's MAP) creates 1209 an entry. 1211 Proxy neighbor cache entries are created and maintained by AERO 1212 Proxys when they process Client/Server ND/PD exchanges, and remain in 1213 place for durations bounded by ND/PD lifetimes. AERO Proxys maintain 1214 proxy neighbor cache entries for each of their associated Clients. 1215 Proxy neighbor cache entries track the Client state and the address 1216 of the Client's associated Server. 1218 To the list of neighbor cache entry states in Section 7.3.2 of 1219 [RFC4861], AERO interfaces add an additional state DEPARTED that 1220 applies to symmetric and proxy neighbor cache entries for Clients 1221 that have recently departed. The interface sets a "DepartTime" 1222 variable for the neighbor cache entry to "DEPARTTIME" seconds. 1223 DepartTime is decremented unless a new ND message causes the state to 1224 return to REACHABLE. While a neighbor cache entry is in the DEPARTED 1225 state, packets destined to the target Client are forwarded to the 1226 Client's new location instead of being dropped. When DepartTime 1227 decrements to 0, the neighbor cache entry is deleted. It is 1228 RECOMMENDED that DEPARTTIME be set to the default constant value 40 1229 seconds to allow for packets in flight to be delivered while stale 1230 route optimization state may be present. 1232 When a target Server (acting as a Mobility Anchor Point (MAP)) 1233 receives a valid NS message used for route optimization, it searches 1234 for a symmetric neighbor cache entry for the target Client. The MAP 1235 then returns a solicited NA message without creating a neighbor cache 1236 entry for the ROS, but creates or updates a target Client "Report 1237 List" entry for the ROS and sets a "ReportTime" variable for the 1238 entry to REPORTTIME seconds. The MAP resets ReportTime when it 1239 receives a new authentic NS message, and otherwise decrements 1240 ReportTime while no NS messages have been received. It is 1241 RECOMMENDED that REPORTTIME be set to the default constant value 40 1242 seconds to allow a 10 second window so that route optimization can 1243 converge before ReportTime decrements below REACHABLETIME. 1245 When the ROS receives a solicited NA message response to its NS 1246 message, it creates or updates an asymmetric neighbor cache entry for 1247 the target network-layer and link-layer addresses. The ROS then 1248 (re)sets ReachableTime for the neighbor cache entry to REACHABLETIME 1249 seconds and uses this value to determine whether packets can be 1250 forwarded directly to the target, i.e., instead of via a default 1251 route. The ROS otherwise decrements ReachableTime while no further 1252 solicited NA messages arrive. It is RECOMMENDED that REACHABLETIME 1253 be set to the default constant value 30 seconds as specified in 1254 [RFC4861]. 1256 The ROS also uses the value MAX_UNICAST_SOLICIT to limit the number 1257 of NS keepalives sent when a correspondent may have gone unreachable, 1258 the value MAX_RTR_SOLICITATIONS to limit the number of RS messages 1259 sent without receiving an RA and the value MAX_NEIGHBOR_ADVERTISEMENT 1260 to limit the number of unsolicited NAs that can be sent based on a 1261 single event. It is RECOMMENDED that MAX_UNICAST_SOLICIT, 1262 MAX_RTR_SOLICITATIONS and MAX_NEIGHBOR_ADVERTISEMENT be set to 3 the 1263 same as specified in [RFC4861]. 1265 Different values for DEPARTTIME, REPORTTIME, REACHABLETIME, 1266 MAX_UNICAST_SOLICIT, MAX_RTR_SOLCITATIONS and 1267 MAX_NEIGHBOR_ADVERTISEMENT MAY be administratively set; however, if 1268 different values are chosen, all nodes on the link MUST consistently 1269 configure the same values. Most importantly, DEPARTTIME and 1270 REPORTTIME SHOULD be set to a value that is sufficiently longer than 1271 REACHABLETIME to avoid packet loss due to stale route optimization 1272 state. 1274 3.9. AERO Interface Encapsulation and Re-encapsulation 1276 AERO interfaces encapsulate packets according to whether they are 1277 entering the AERO interface from the network layer or if they are 1278 being re-admitted into the same AERO link they arrived on. This 1279 latter form of encapsulation is known as "re-encapsulation". 1281 For packets entering the AERO interface from the network layer, the 1282 AERO interface copies the "TTL/Hop Limit", "Type of Service/Traffic 1283 Class" [RFC2983], "Flow Label"[RFC6438] (for IPv6) and "Congestion 1284 Experienced" [RFC3168] values in the packet's IP header into the 1285 corresponding fields in the encapsulation header(s). 1287 For packets undergoing re-encapsulation, the AERO interface instead 1288 copies these values from the original encapsulation header into the 1289 new encapsulation header, i.e., the values are transferred between 1290 encapsulation headers and *not* copied from the encapsulated packet's 1291 network-layer header. (Note especially that by copying the TTL/Hop 1292 Limit between encapsulation headers the value will eventually 1293 decrement to 0 if there is a (temporary) routing loop.) For IPv4 1294 encapsulation/re-encapsulation, the AERO interface sets the DF bit as 1295 discussed in Section 3.13. 1297 AERO interfaces configured over INET underlying interfaces 1298 encapsulate each packet in a SPAN header, then encapsulate the 1299 resulting SPAN packet in an INET header according to the next hop 1300 determined in the forwarding algorithm in Section 3.12. If the next 1301 hop is reached via a secured tunnel, the AERO interface uses an INET 1302 encapsulation format specific to the secured tunnel type (see: 1303 Section 6). If the next hop is reached via an unsecured underlying 1304 interface, the AERO interface instead uses Generic UDP Encapsulation 1305 (GUE) [I-D.ietf-intarea-gue] or an alternate minimal encapsulation 1306 format Appendix A. 1308 When GUE encapsulation is used, the AERO interface next sets the UDP 1309 source port to a constant value that it will use in each successive 1310 packet it sends, and sets the UDP length field to the length of the 1311 SPAN packet plus 8 bytes for the UDP header itself plus the length of 1312 the GUE header (or 0 if GUE direct IP encapsulation is used). For 1313 packets sent to a Server or Relay, the AERO interface sets the UDP 1314 destination port to 8060, i.e., the IANA-registered port number for 1315 AERO. For packets sent to a Client, the AERO interface sets the UDP 1316 destination port to the port value stored in the neighbor cache entry 1317 for this Client. The AERO interface then either includes or omits 1318 the UDP checksum according to the GUE specification. 1320 Client AERO interfaces can avoid encapsulation over Direct underlying 1321 interface and Proxyed underlying interfaces for which the first-hop 1322 access router is AERO-aware. 1324 AERO interfaces observes the packet sizing and fragmentation 1325 considerations found in Section 3.13. 1327 3.10. AERO Interface Decapsulation 1329 AERO interfaces decapsulate packets destined either to the AERO node 1330 itself or to a destination reached via an interface other than the 1331 AERO interface the packet was received on. When the encapsulated 1332 packet arrives in multiple fragments, the AERO interface reassembles 1333 as discussed in Section 3.13. Further decapsulation steps are 1334 performed according to the appropriate encapsulation format 1335 specification. 1337 3.11. AERO Interface Data Origin Authentication 1339 AERO nodes employ simple data origin authentication procedures. In 1340 particular: 1342 o AERO Relays, Servers and Proxys accept encapsulated data packets 1343 and control messages received from secured tunnels. 1345 o AERO Servers and Proxys accept encapsulated data packets and NS 1346 messages used for Neighbor Unreachability Detection (NUD) received 1347 from a source found in the ROS list. 1349 o AERO Proxys and Clients accept packets that originate from within 1350 the same secured ANET. 1352 o AERO Clients and Gateways accept packets from downstream network 1353 correspondents based on ingress filtering. 1355 AERO nodes silently drop any packets that do not satisfy the above 1356 data origin authentication procedures. Further security 1357 considerations are discussed Section 6. 1359 3.12. AERO Interface Forwarding Algorithm 1361 IP packets enter a node's AERO interface either from the network 1362 layer (i.e., from a local application or the IP forwarding system) or 1363 from the link layer (i.e., from an AERO interface neighbor). All 1364 packets entering a node's AERO interface first undergo data origin 1365 authentication as discussed in Section 3.11. Packets that satisfy 1366 data origin authentication are processed further, while all others 1367 are dropped silently. 1369 Packets that enter the AERO interface from the network layer are 1370 forwarded to an AERO interface neighbor. Packets that enter the AERO 1371 interface from the link layer are either re-admitted into the AERO 1372 link or forwarded to the network layer where they are subject to 1373 either local delivery or IP forwarding. In all cases, the AERO 1374 interface itself MUST NOT decrement the network layer TTL/Hop-count 1375 since its forwarding actions occur below the network layer. 1377 AERO interfaces may have multiple underlying interfaces and/or 1378 neighbor cache entries for neighbors with multiple ifIndex-tuple 1379 registrations (see Section 3.6). The AERO interface uses each 1380 packet's DSCP value (and/or port number) to select an outgoing 1381 underlying interface based on the node's own QoS preferences, and 1382 also to select a destination link-layer address based on the 1383 neighbor's underlying interface with the highest preference. AERO 1384 implementations SHOULD allow for QoS preference values to be modified 1385 at runtime through network management. 1387 If multiple outgoing interfaces and/or neighbor interfaces have a 1388 preference of "high", the AERO node replicates the packet and sends 1389 one copy via each of the (outgoing / neighbor) interface pairs; 1390 otherwise, the node sends a single copy of the packet via the 1391 interface with the highest preference. AERO nodes keep track of 1392 which underlying interfaces are currently "reachable" or 1393 "unreachable", and only use "reachable" interfaces for forwarding 1394 purposes. 1396 The following sections discuss the AERO interface forwarding 1397 algorithms for Clients, Proxys, Servers and Relays. In the following 1398 discussion, a packet's destination address is said to "match" if it 1399 is the same as a cached address, or if it is covered by a cached 1400 prefix (which may be encoded in an AERO address). 1402 3.12.1. Client Forwarding Algorithm 1404 When an IP packet enters a Client's AERO interface from the network 1405 layer the Client searches for an asymmetric neighbor cache entry that 1406 matches the destination. If there is a match, the Client uses one or 1407 more "reachable" neighbor interfaces in the entry for packet 1408 forwarding. If there is no asymmetric neighbor cache entry, the 1409 Client instead forwards the packet toward a Server (the packet is 1410 intercepted by a Proxy if there is a Proxy on the path). 1412 When an IP packet enters a Client's AERO interface from the link- 1413 layer, if the destination matches one of the Client's MNPs or link- 1414 local addresses the Client decapsulates the packet (if necessary) and 1415 delivers it to the network layer. Otherwise, the Client drops the 1416 packet and MAY return a network-layer ICMP Destination Unreachable 1417 message subject to rate limiting (see: Section 3.14). 1419 3.12.2. Proxy Forwarding Algorithm 1421 For control messages originating from or destined to a Client, the 1422 Proxy intercepts the message and updates its proxy neighbor cache 1423 entry for the Client. The Proxy then forwards a (proxyed) copy of 1424 the control message. (For example, the Proxy forwards a proxied 1425 version of a Client's NS/RS message to the target neighbor, and 1426 forwards a proxied version of the NA/RA reply to the Client.) 1428 When the Proxy receives a data packet from a Client within the ANET, 1429 the Proxy searches for an asymmetric neighbor cache entry that 1430 matches the destination and forwards the packet as follows: 1432 o if the destination matches an asymmetric neighbor cache entry, the 1433 Proxy uses one or more "reachable" neighbor interfaces in the 1434 entry for packet forwarding via encapsulation. If the neighbor 1435 interface is in the same SPAN segment, the Proxy forwards the 1436 packet directly to the neighbor; otherwise, it forwards the packet 1437 to a Relay. 1439 o else, the Proxy encapsulates and forwards the packet to a Relay 1440 while using the packet's destination address as the SPAN 1441 destination address. (If the destination is an AERO address, the 1442 Proxy instead uses the corresponding Subnet Router Anycast address 1443 for Client AERO addresses and the SPAN address for 1444 administratively-provisioned AERO addresses.). 1446 When the Proxy receives an encapsulated data packet from an INET 1447 neighbor or from a secured tunnel, it accepts the packet only if data 1448 origin authentication succeeds and the SPAN destination address is 1449 its own address. If the packet is a SPAN fragment, the Proxy then 1450 adds the fragment to the reassembly buffer and returns if the 1451 reassembly is still incomplete. Otherwise, the Proxy reassembles the 1452 packet (if necessary) and continues processing. 1454 Next, the Proxy searches for a proxy neighbor cache entry that 1455 matches the destination. If there is a proxy neighbor cache entry in 1456 the REACHABLE state, the Proxy decapsulates and forwards the packet 1457 to the Client. If the neighbor cache entry is in the DEPARTED state, 1458 the Proxy instead re-encapsulates the message and forwards it to a 1459 Relay. If there is no neighbor cache entry, the Proxy instead 1460 discards the packet. 1462 3.12.3. Server/Gateway Forwarding Algorithm 1464 For control messages destined to a target Client's AERO address that 1465 are received from a secured tunnel, the Server (acting as a MAP) 1466 intercepts the message and sends an appropriate response on behalf of 1467 the Client. (For example, the Server sends an NA message reply in 1468 response to an NS message directed to one of its associated Clients.) 1469 If the Client's neighbor cache entry is in the DEPARTED state, 1470 however, the Server instead forwards the packet to the Client's new 1471 Server as discussed in Section 3.19. 1473 When the Server receives an encapsulated data packet from an INET 1474 neighbor or from a secured tunnel, it accepts the packet only if data 1475 origin authentication succeeds. If the SPAN destination address is 1476 its own address, the Server reassembles if necessary and discards the 1477 SPAN header (if the reassembly is incomplete, the Server instead adds 1478 the fragment to the reassembly buffer and returns). The Server then 1479 continues processing as follows: 1481 o if the destination matches a symmetric neighbor cache entry in the 1482 REACHABLE state the Server prepares the packet for forwarding to 1483 the destination Client. If the current header is a SPAN header, 1484 the Server reassembles if necessary and discards the SPAN header 1485 (if the reassembly is incomplete, the Server instead adds the 1486 fragment to the reassembly buffer and returns). The Server then 1487 forwards the packet according to the cached link-layer 1488 information, while using SPAN encapsulation for the Client's 1489 Proxyed/Native interfaces, simple INET encapsulation for NATed/ 1490 VPNed interfaces, or no encapsulation for Direct interfaces. 1492 o else, if the destination matches a symmetric neighbor cache entry 1493 in the DEPARETED state the Server encapsulates the packet in a new 1494 SPAN header and forwards it to the Client's new Server (noting 1495 that the encapsulation may result in the addition of a second SPAN 1496 header). The Server uses its own SPAN address as the source and 1497 the SPAN address of the new Server as the destination. 1499 o else, if the destination matches an asymmetric neighbor cache 1500 entry, the Server uses one or more "reachable" neighbor interfaces 1501 in the entry for packet forwarding via the local INET if the 1502 neighbor is in the same SPAN segment or via a Relay otherwise. 1504 o else, if the destination is an AERO address that is not assigned 1505 on the AERO interface the Server drops the packet. 1507 o else, the Server (acting as a Gateway) releases the packet to the 1508 network layer for local delivery or IP forwarding. Based on the 1509 information in the forwarding table, the network layer may return 1510 the packet to the same AERO interface in which case further 1511 processing occurs as below. (Note that this arrangement 1512 accommodates common implementations in which the IP forwarding 1513 table is not accessible from within the AERO interface. If the 1514 AERO interface can directly access the IP forwarding table, the 1515 forwarding table lookup can instead be performed internally from 1516 within the AERO interface itself.) 1518 When the Server's AERO interface receives a data packet from the 1519 network layer or from a NATed/VPNed/Direct Client, it processes the 1520 packet according to the network-layer destination address as follows: 1522 o if the destination matches a symmetric or asymmetric neighbor 1523 cache entry the Server processes the packet as above. 1525 o else, the Server encapsulates the packet and forwards it to a 1526 Relay. For administratively-assigned AERO address destinations, 1527 the Server uses the SPAN address corresponding to the destination 1528 as the SPAN destination address. For Client AERO address 1529 destinations, the Server uses the Subnet Router Anycast address 1530 corresponding to the destination as the SPAN destination address. 1531 For all others, the Server uses the packet's destination IP 1532 address as the SPAN destination address. 1534 3.12.4. Relay Forwarding Algorithm 1536 Relays forward packets over secured tunnels the same as any IP 1537 router. When the Relay receives an encapsulated packet via a secured 1538 tunnel, it removes the INET header and searches for a forwarding 1539 table entry that matches the destination address in the next header. 1540 The Relay then processes the packet as follows: 1542 o if the destination matches one of the Relay's own addresses, the 1543 Relay submits the packet for local delivery. 1545 o else, if the destination matches a forwarding table entry the 1546 Relay forwards the packet via a secured tunnel to the next hop. 1547 If the destination matches an MSP without matching an MNP, 1548 however, the Relay instead drops the packet and returns an ICMP 1549 Destination Unreachable message subject to rate limiting (see: 1550 Section 3.14). 1552 o else, the Relay drops the packet and returns an ICMP Destination 1553 Unreachable as above. 1555 As for any IP router, the Relay decrements the TTL/Hop Limit when it 1556 forwards the packet. If the packet is encapsulated in a SPAN header, 1557 only the Hop Limit in the SPAN header is decremented, and not the 1558 TTL/Hop Limit in the inner packet header. 1560 3.13. AERO Interface MTU and Fragmentation 1562 The AERO interface is the node's attachment to the AERO link. For 1563 AERO link neighbor underlying interface paths that do not require 1564 encapsulation, the AERO interface sends unencapsulated IP packets. 1565 For other paths, the AERO interface acts as a tunnel ingress when it 1566 sends packets to the neighbor and as a tunnel egress when it receives 1567 packets from the neighbor. 1569 AERO interfaces configure an MTU the same as for any IP interface, 1570 however the MTU does not reflect the physical size of any links in 1571 the path but rather determines the maximum size for reassembly. AERO 1572 interfaces observe the packet sizing considerations for tunnels 1573 discussed in [I-D.ietf-intarea-tunnels] and as specified below. 1575 The Internet Protocol expects that IP packets will either be 1576 delivered to the destination or a suitable Packet Too Big (PTB) 1577 message returned to support the process known as IP Path MTU 1578 Discovery (PMTUD) [RFC1191][RFC8201]. However, PTB messages may be 1579 crafted for malicious purposes or lost in the network [RFC2923]. 1580 This can be especially problematic for tunnels, where a condition 1581 known as a PMTUD "black hole" can result. For these reasons, AERO 1582 interfaces employ operational procedures that avoid interactions with 1583 PMTUD, including the use of fragmentation when necessary. 1585 AERO interfaces observe three different types of fragmentation. 1586 Source fragmentation occurs when the AERO interface (acting as a 1587 tunnel ingress) fragments the encapsulated packet into multiple 1588 fragments before admitting each fragment into the tunnel. Network 1589 fragmentation occurs when an encapsulated packet admitted into the 1590 tunnel by the ingress is fragmented by an IPv4 router on the path to 1591 the egress. Finally, link-layer fragmentation (aka link adaptation) 1592 occurs at a layer below IP and is coordinated between underlying data 1593 link endpoints. 1595 IPv6 specifies a minimum link Maximum Transmission Unit (MTU) of 1280 1596 bytes [RFC8200]. Although IPv4 specifies a smaller minimum link MTU 1597 of 68 bytes [RFC0791], AERO interfaces also observe the IPv6 minimum 1598 for IPv4 even if encapsulated packets may incur network 1599 fragmentation. 1601 IPv6 specifies a minimum Maximum Reassembly Unit (MRU) of 1500 bytes 1602 [RFC8200], while the minimum MRU for IPv4 is only 576 bytes [RFC1122] 1603 (but, note that many standard IPv6 over IPv4 tunnel types already 1604 assume a larger MRU than the IPv4 minimum). 1606 AERO interfaces therefore configure an MTU that MUST NOT be smaller 1607 than 1280 bytes, MUST NOT be larger than the minimum MRU among all 1608 nodes on the AERO link minus the encapsulation overhead ("ENCAPS"), 1609 and SHOULD NOT be smaller than 1500 bytes. AERO interfaces also 1610 configure a Maximum Segment Unit (MSU) as the maximum-sized 1611 encapsulated packet that the ingress can inject into the tunnel 1612 without source fragmentation. The MSU value MUST NOT be larger than 1613 1280 bytes unless there is operational assurance that a larger size 1614 can traverse the link along all paths. 1616 All AERO interfaces on the link MUST configure the same MTU value for 1617 reasons cited in [RFC3819][RFC4861]; in particular, multicast support 1618 requires a common MTU value among all nodes on the link. All AERO 1619 interfaces MUST configure an MRU large enough to reassemble packets 1620 up to (MTU+ENCAPS) bytes in length; nodes that cannot configure a 1621 large-enough MRU MUST NOT enable an AERO interface. For example, for 1622 an MTU of 1500 bytes an appropriate MRU might be 2KB. 1624 The network layer proceeds as follows when it forwards an IP packet 1625 to the AERO interface. For each IPv4 packet that is larger than the 1626 AERO interface MTU and with DF set to 0, the network layer uses IPv4 1627 fragmentation to break the packet into a minimum number of non- 1628 overlapping fragments where the first fragment is no larger than the 1629 MTU and the remaining fragments are no larger than the first. For 1630 all other IP packets, if the packet is larger than the AERO interface 1631 MTU, the network layer drops the packet and returns a PTB message to 1632 the original source. Otherwise, the network layer admits each IP 1633 packet or fragment into the AERO interface. 1635 For each IP packet admitted into AERO interface, if the neighbor is 1636 reached via an underlying interface that does not require 1637 encapsulation the AERO interface proceeds according to the underlying 1638 interface MTU. If the packet is no larger than the underlying 1639 interface MTU, the AERO interface presents the packet to the 1640 underlying interface. Otherwise, for IPv4 packets with DF set to 0 1641 the AERO interface uses IPv4 fragmentation to break the packet into 1642 fragments no larger than the underlying interface MTU. For other 1643 packets, the AERO interface either performs link adaptation or drops 1644 the packet and returns a PTB message to the original source. (If the 1645 original source corresponds to a local application, the PTB would 1646 appear to have originated from a router on the path when in fact it 1647 was locally generated from within the AERO interface.) In the same 1648 way, when a packet that has been admitted into the AERO link reaches 1649 a target neighbor but is too large to be delivered over the final-hop 1650 underlying interface, the target either performs link adaptation or 1651 drops the packet and returns a PTB. Link adaptation is preferred in 1652 both cases when possible to avoid packet loss. 1654 For underlying interfaces that require encapsulation, the AERO 1655 interface (acting as a tunnel ingress) instead encapsulates the 1656 packet and performs path MTU procedures according to the specific 1657 encapsulation format. For INET interfaces, the ingress encapsulates 1658 the packet in a SPAN header. If the SPAN packet is larger than the 1659 MSU, the ingress source fragments the SPAN packet into a minimum 1660 number of non-overlapping fragments where the first fragment is no 1661 larger than the MSU and the remaining fragments are no larger than 1662 the first. The ingress then encapsulates each SPAN packet/fragment 1663 in an INET header and admits them into the tunnel. For IPv4, the 1664 ingress sets the DF bit to 0 in the INET header in case any network 1665 fragmentation is necessary. The encapsulated packets will be 1666 delivered to the egress, which reassembles them into a whole packet 1667 if necessary. 1669 By fragmenting at the SPAN layer instead of lower layers, standard 1670 IPv6 fragmentation and reassembly [RFC8200] ensures that IPv4 issues 1671 such as data corruption due to reassembly misassociations will not 1672 occur [RFC6864][RFC4963]. The ingress sends each fragment with 1673 minimal delay so that individual fragments are unlikely to be 1674 diverted to different destinations due to routing fluctuations. 1676 Since the SPAN header and IPv6 fragment extension header reduces the 1677 room available for packet data, but the original source has no way to 1678 control its insertion, the ingress MUST include their lengths in 1679 ENCAPS even for packets in which the header is absent. 1681 3.13.1. AERO MTU Requirements 1683 In light of the above considerations, AERO interfaces SHOULD 1684 configure an MTU of 9180 bytes. This means that the AERO interface 1685 MUST be capable of reassembling original IP packets up to 9180 bytes 1686 in length. When an IP packet is admitted into an AERO interface, the 1687 interface encapsulates the packet using SPAN encapsulation and 1688 fragments the encapsulated packet into fragments that are no larger 1689 than 1280 bytes. The fragments will be reassembled by the tunnel 1690 egress that services the final destination. 1692 AERO Clients behind Proxys MAY configure an MTU smaller than 9180 1693 (but no smaller than IP minimum link MTU). If Clients configure a 1694 diversity of MTUs (e.g., 1280, 1500, 4KB, 8KB, etc.) then neighbors 1695 on the link would appear to have multiple MTUs. IPv6 Path MTU 1696 Discovery [RFC8201] accounts for this possibility since MTU discovery 1697 must be performed even between nodes that appear to be connected to 1698 the same link. 1700 Applications that cannot tolerate loss in the network due to MTU 1701 restrictions should restrict themselves to sending packets no larger 1702 than the IP minimum link MTU, i.e., even if the current path MTU 1703 would appear to support a larger size. This is due to the fact that 1704 routing changes could cause the path to traverse links with smaller 1705 MTUs at any given point in time. 1707 3.14. AERO Interface Error Handling 1709 When an AERO node admits encapsulated packets into the AERO 1710 interface, it may receive link-layer or network-layer error 1711 indications. 1713 A link-layer error indication is an ICMP error message generated by a 1714 router in the INET on the path to the neighbor or by the neighbor 1715 itself. The message includes an IP header with the address of the 1716 node that generated the error as the source address and with the 1717 link-layer address of the AERO node as the destination address. 1719 The IP header is followed by an ICMP header that includes an error 1720 Type, Code and Checksum. Valid type values include "Destination 1721 Unreachable", "Time Exceeded" and "Parameter Problem" 1722 [RFC0792][RFC4443]. (AERO interfaces ignore all link-layer IPv4 1723 "Fragmentation Needed" and IPv6 "Packet Too Big" messages since they 1724 only emit packets that are guaranteed to be no larger than the IP 1725 minimum link MTU as discussed in Section 3.13.) 1727 The ICMP header is followed by the leading portion of the packet that 1728 generated the error, also known as the "packet-in-error". For 1729 ICMPv6, [RFC4443] specifies that the packet-in-error includes: "As 1730 much of invoking packet as possible without the ICMPv6 packet 1731 exceeding the minimum IPv6 MTU" (i.e., no more than 1280 bytes). For 1732 ICMPv4, [RFC0792] specifies that the packet-in-error includes: 1733 "Internet Header + 64 bits of Original Data Datagram", however 1734 [RFC1812] Section 4.3.2.3 updates this specification by stating: "the 1735 ICMP datagram SHOULD contain as much of the original datagram as 1736 possible without the length of the ICMP datagram exceeding 576 1737 bytes". 1739 The link-layer error message format is shown in Figure 5 (where, "L2" 1740 and "L3" refer to link-layer and network-layer, respectively): 1742 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1743 ~ ~ 1744 | L2 IP Header of | 1745 | error message | 1746 ~ ~ 1747 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1748 | L2 ICMP Header | 1749 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1750 ~ ~ P 1751 | IP and other encapsulation | a 1752 | headers of original L3 packet | c 1753 ~ ~ k 1754 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ e 1755 ~ ~ t 1756 | IP header of | 1757 | original L3 packet | i 1758 ~ ~ n 1759 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1760 ~ ~ e 1761 | Upper layer headers and | r 1762 | leading portion of body | r 1763 | of the original L3 packet | o 1764 ~ ~ r 1765 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ --- 1767 Figure 5: AERO Interface Link-Layer Error Message Format 1769 The AERO node rules for processing these link-layer error messages 1770 are as follows: 1772 o When an AERO node receives a link-layer Parameter Problem message, 1773 it processes the message the same as described as for ordinary 1774 ICMP errors in the normative references [RFC0792][RFC4443]. 1776 o When an AERO node receives persistent link-layer Time Exceeded 1777 messages, the IP ID field may be wrapping before earlier fragments 1778 awaiting reassembly have been processed. In that case, the node 1779 should begin including integrity checks and/or institute rate 1780 limits for subsequent packets. 1782 o When an AERO node receives persistent link-layer Destination 1783 Unreachable messages in response to encapsulated packets that it 1784 sends to one of its asymmetric neighbor correspondents, the node 1785 should process the message as an indication that a path may be 1786 failing, and optionally initiate NUD over that path. If it 1787 receives Destination Unreachable messages over multiple paths, the 1788 node should allow future packets destined to the correspondent to 1789 flow through a default route and re-initiate route optimization. 1791 o When an AERO Client receives persistent link-layer Destination 1792 Unreachable messages in response to encapsulated packets that it 1793 sends to one of its symmetric neighbor Servers, the Client should 1794 mark the path as unusable and use another path. If it receives 1795 Destination Unreachable messages on many or all paths, the Client 1796 should associate with a new Server and release its association 1797 with the old Server as specified in Section 3.19.5. 1799 o When an AERO Server receives persistent link-layer Destination 1800 Unreachable messages in response to encapsulated packets that it 1801 sends to one of its symmetric neighbor Clients, the Server should 1802 mark the underlying path as unusable and use another underlying 1803 path. 1805 o When an AERO Server or Proxy receives link-layer Destination 1806 Unreachable messages in response to an encapsulated packet that it 1807 sends to one of its permanent neighbors, it treats the messages as 1808 an indication that the path to the neighbor may be failing. 1809 However, the dynamic routing protocol should soon reconverge and 1810 correct the temporary outage. 1812 When an AERO Relay receives a packet for which the network-layer 1813 destination address is covered by an MSP, if there is no more- 1814 specific routing information for the destination the Relay drops the 1815 packet and returns a network-layer Destination Unreachable message 1816 subject to rate limiting. The Relay writes the network-layer source 1817 address of the original packet as the destination address and uses 1818 one of its non link-local addresses as the source address of the 1819 message. 1821 When an AERO node receives an encapsulated packet for which the 1822 reassembly buffer it too small, it drops the packet and returns a 1823 network-layer Packet Too Big (PTB) message. The node first writes 1824 the MRU value into the PTB message MTU field, writes the network- 1825 layer source address of the original packet as the destination 1826 address and writes one of its non link-local addresses as the source 1827 address. 1829 3.15. AERO Router Discovery, Prefix Delegation and Autoconfiguration 1831 AERO Router Discovery, Prefix Delegation and Autoconfiguration are 1832 coordinated as discussed in the following Sections. 1834 3.15.1. AERO ND/PD Service Model 1836 Each AERO Server on the link configures a PD service to facilitate 1837 Client requests. Each Server is provisioned with a database of MNP- 1838 to-Client ID mappings for all Clients enrolled in the AERO service, 1839 as well as any information necessary to authenticate each Client. 1840 The Client database is maintained by a central administrative 1841 authority for the AERO link and securely distributed to all Servers, 1842 e.g., via the Lightweight Directory Access Protocol (LDAP) [RFC4511], 1843 via static configuration, etc. Clients receive the same service 1844 regardless of the Servers they select. 1846 AERO Clients and Servers use ND messages to maintain neighbor cache 1847 entries. AERO Servers configure their AERO interfaces as advertising 1848 interfaces, and therefore send unicast RA messages with configuration 1849 information in response to a Client's RS message. Thereafter, 1850 Clients send additional RS messages to refresh prefix and/or router 1851 lifetimes. 1853 AERO Clients and Servers include PD parameters in RS/RA messages (see 1854 [I-D.templin-6man-dhcpv6-ndopt] for ND/PD alternatives). The unified 1855 ND/PD messages are exchanged between Client and Server according to 1856 the prefix management schedule required by the PD service. If the 1857 Client knows its MNP in advance, it can include its AERO address as 1858 the source address of an RS message and with an OMNI option with a 1859 valid Prefix Length for the MNP. If the Server (and Proxy) accept 1860 the Client's MNP assertion, they inject the prefix into the routing 1861 system and establish the necessary neighbor cache state. 1863 The following sections specify the Client and Server behavior. 1865 3.15.2. AERO Client Behavior 1867 AERO Clients discover the addresses of Servers in the same manner 1868 described in [RFC5214]. Discovery methods include static 1869 configuration (e.g., from a flat-file map of Server addresses and 1870 locations), or through an automated means such as Domain Name System 1871 (DNS) name resolution [RFC1035]. Alternatively, the Client can 1872 discover Server addresses through a layer 2 data link login exchange, 1873 or through a unicast RA response to a multicast/anycast RS as 1874 described below. In the absence of other information, the Client can 1875 resolve the DNS Fully-Qualified Domain Name (FQDN) 1876 "linkupnetworks.[domainname]" where "linkupnetworks" is a constant 1877 text string and "[domainname]" is a DNS suffix for the AERO link 1878 (e.g., "example.com"). 1880 To associate with a Server, the Client acts as a requesting router to 1881 request MNPs. The Client prepares an RS message with PD parameters 1882 and includes a Nonce and Timestamp option if the Client needs to 1883 correlate RA replies. If the Client already knows the Server's AERO 1884 address, it includes the AERO address as the network-layer 1885 destination address; otherwise, it includes all-routers multicast 1886 (ff02::2) or subnet routers anycast (fe80::) as the network-layer 1887 destination address. If the Client already knows its own AERO 1888 address, it uses the AERO address as the network-layer source 1889 address; otherwise, it uses the unspecified AERO address 1890 (fe80::ffff:ffff) as the network-layer source address. 1892 The Client next includes an OMNI option in the RS message to register 1893 its link-layer information with the Server. The first ifIndex-tuple 1894 MUST correspond to the underlying interface over which the Client 1895 will send the RS message. The Client MAY include additional ifIndex- 1896 tuples specific to other underlying interfaces. When encapsulation 1897 is used, the Client also includes an SLLAO with a single ifIndex- 1898 tuple corresponding to the first OMNI option ifIndex-tuple, then 1899 encapsulates the RS message in an ANET header with its own ANET 1900 address as the source address and the INET address of the Server as 1901 the destination. 1903 The Client then sends the RS message (either directly via Direct 1904 interfaces, via INET encapsulation for NATed interfaces, via a VPN 1905 for VPNed interfaces, via a Proxy for proxyed interfaces or via a 1906 Relay for native interfaces) and waits for an RA message reply (see 1907 Section 3.15.3). The Client retries up to MAX_RTR_SOLICITATIONS 1908 times until an RA is received. If the Client receives no RAs, or if 1909 it receives an RA with Router Lifetime set to 0, the Client SHOULD 1910 abandon this Server and try another Server. Otherwise, the Client 1911 processes the PD information found in the RA message. 1913 Next, the Client creates a symmetric neighbor cache entry with the 1914 Server's AERO address as the network-layer address and the Server's 1915 encapsulation and/or link-layer addresses as the link-layer address. 1916 The Client records the RA Router Lifetime field value in the neighbor 1917 cache entry as the time for which the Server has committed to 1918 maintaining the MNP in the routing system. The Client then 1919 autoconfigures AERO addresses for each of the delegated MNPs and 1920 assigns them to the AERO interface. The Client also caches any MSPs 1921 included in Route Information Options (RIOs) [RFC4191] as MSPs to 1922 associate with the AERO link, and assigns the MTU value in the MTU 1923 option to its AERO interface while configuring an appropriate MRU. 1925 The Client then registers additional underlying interfaces with the 1926 Server by sending RS messages via each additional interface. The RS 1927 messages include the same parameters as for the initial RS/RA 1928 exchange, but with destination address set to the Server's AERO 1929 address and with the initial OMNI option ifIndex-tuple corresponding 1930 to the underlying interface. 1932 The Client examines the P bit in the RA message flags field. If the 1933 P bit is set to 1, this indicates that the Server received an RS 1934 message with an SLLAO in which the first ifIndex-tuple addressing 1935 information did not match the information in the encapsulation 1936 headers. 1938 Following autoconfiguration, the Client sub-delegates the MNPs to its 1939 attached EUNs and/or the Client's own internal virtual interfaces as 1940 described in [I-D.templin-v6ops-pdhost] to support the Client's 1941 downstream attached "Internet of Things (IoT)". The Client 1942 subsequently maintains its MNP delegations through each of its 1943 Servers by sending additional RS messages before Router Lifetime 1944 expires. 1946 After the Client registers its underlying interfaces, it may wish to 1947 change one or more registrations, e.g., if an interface changes 1948 address or becomes unavailable, if QoS preferences change, etc. To 1949 do so, the Client prepares an RS message to send over any available 1950 underlying interface. The RS includes an OMNI option with a first 1951 ifIndex-tuple specific to the selected interface, and MAY include any 1952 additional ifIndex-tuples specific to other underlying interfaces. 1953 The Client includes fresh ifIndex-tuple values to update the Server's 1954 neighbor cache entry. When the Client receives the Server's RA 1955 response, it has assurance that the Server has been updated with the 1956 new information. 1958 If the Client wishes to discontinue use of a Server it issues an RS 1959 message over any underlying interface with an OMNI option with R set 1960 to 0. When the Server processes the message, it releases the MNP, 1961 sets the symmetric neighbor cache entry state for the Client to 1962 DEPARTED and returns an RA reply with Router Lifetime set to 0. 1963 After a short delay (e.g., 2 seconds), the Server withdraws the MNP 1964 from the routing system. 1966 3.15.3. AERO Server Behavior 1968 AERO Servers act as IP routers and support a PD service for Clients. 1969 Servers arrange to add their AERO and INET addresses to a static map 1970 of Server addresses for the link and/or the DNS resource records for 1971 the FQDN "linkupnetworks.[domainname]" before entering service. 1972 Server addresses should be geographically and/or topologically 1973 referenced, and made available for discovery by Clients on the AERO 1974 link. 1976 When a Server receives a prospective Client's RS message on its AERO 1977 interface, it SHOULD return an immediate RA reply with Router 1978 Lifetime set to 0 if it is currently too busy or otherwise unable to 1979 service the Client. Otherwise, the Server authenticates the RS 1980 message and processes the PD parameters. The Server first determines 1981 the correct MNPs to delegate to the Client by searching the Client 1982 database. When the Server delegates the MNPs, it also creates an IP 1983 forwarding table entry for each MNP so that the MNPs are propagated 1984 into the routing system (see: Section 3.3). For IPv6, the Server 1985 creates an IPv6 forwarding table entry for each MNP. For IPv4, the 1986 Server creates an IPv6 forwarding table entry with the IPv4-mapped 1987 IPv6 address corresponding to the IPv4 address. 1989 The Server next creates a symmetric neighbor cache entry for the 1990 Client using the base AERO address as the network-layer address and 1991 with lifetime set to no more than the smallest PD lifetime. Next, 1992 the Server updates the neighbor cache entry by recording the 1993 information in each ifIndex-tuple in the RS OMNI option. The Server 1994 also records the actual INET header source address and port number in 1995 the neighbor cache entry. If an SLLAO option was present, the Server 1996 also compares the SLLAO address information for the first ifIndex- 1997 tuple with the INET header information and sets the P bit in the 1998 flags field of the RA message if the information was different. 2000 Next, the Server prepares an RA message using its AERO address as the 2001 network-layer source address and the network-layer source address of 2002 the RS message as the network-layer destination address. The Server 2003 includes the delegated MNPs, any other PD parameters and an OMNI 2004 option with an ifIndex-tuple with ifIndex set to 0. The Server then 2005 includes one or more RIOs that encode the MSPs for the AERO link, 2006 plus an MTU option for the link MTU (see Section 3.13). The Server 2007 finally forwards the message to the Client using SPAN/INET, INET, or 2008 NULL encapsulation as necessary. 2010 After the initial RS/RA exchange, the Server maintains the symmetric 2011 neighbor cache entry for the Client. If the Client (or Proxy) issues 2012 additional NS/RS messages, the Server resets ReachableTime. If the 2013 Client (or Proxy) issues an RS with PD release indication (e.g., by 2014 including an OMNI option with a release indication), or if the Client 2015 becomes unreachable, the Server sets the Client's symmetric neighbor 2016 cache entry to the DEPARTED state. After a short delay (e.g., 2 2017 seconds), the Server withdraws the MNP from the routing system. 2019 The Server processes these and any other Client ND/PD messages, and 2020 returns an NA/RA reply. The Server may also issue unsolicited RA 2021 messages, e.g., with PD reconfigure parameters to cause the Client to 2022 renegotiate its PDs, with Router Lifetime set to 0 if it can no 2023 longer service this Client, etc. Finally, If the symmetric neighbor 2024 cache entry is in the DEPARTED state, the Server deletes the entry 2025 after DepartTime expires. 2027 3.15.3.1. Lightweight DHCPv6 Relay Agent (LDRA) 2029 When DHCPv6 is used as the ND/PD service back end, AERO Clients and 2030 Servers are always on the same link (i.e., the AERO link) from the 2031 perspective of DHCPv6. However, in some implementations the DHCPv6 2032 server and ND function may be located in separate modules. In that 2033 case, the Server's AERO interface module can act as a Lightweight 2034 DHCPv6 Relay Agent (LDRA)[RFC6221] to relay PD messages to and from 2035 the DHCPv6 server module. 2037 When the LDRA receives an authentic RS message, it extracts the PD 2038 message parameters and uses them to construct an IPv6/UDP/DHCPv6 2039 message. It sets the IPv6 source address to the source address of 2040 the RS message, sets the IPv6 destination address to 2041 'All_DHCP_Relay_Agents_and_Servers' and sets the UDP fields to values 2042 that will be understood by the DHCPv6 server. 2044 The LDRA then wraps the message in a DHCPv6 'Relay-Forward' message 2045 header and includes an 'Interface-Id' option that includes enough 2046 information to allow the LDRA to forward the resulting Reply message 2047 back to the Client (e.g., the Client's link-layer addresses, a 2048 security association identifier, etc.). The LDRA also wraps the OMNI 2049 option and SLLAO into the Interface-Id option, then forwards the 2050 message to the DHCPv6 server. 2052 When the DHCPv6 server prepares a Reply message, it wraps the message 2053 in a 'Relay-Reply' message and echoes the Interface-Id option. The 2054 DHCPv6 server then delivers the Relay-Reply message to the LDRA, 2055 which discards the Relay-Reply wrapper and IPv6/UDP headers, then 2056 uses the DHCPv6 message to construct an RA response to the Client. 2057 The Server uses the information in the Interface-Id option to prepare 2058 the RA message and to cache the link-layer addresses taken from the 2059 OMNI option and SLLAO echoed in the Interface-Id option. 2061 3.16. The AERO Proxy 2063 Clients may connect to ANETs that require a perimeter security 2064 gateway to enable communications to Servers in outside INETs. In 2065 that case, the ANET can employ an AERO Proxy. The Proxy is located 2066 at the ANET/INET border and listens for RS messages originating from 2067 or RA messages destined to ANET Clients. The Proxy acts on these 2068 control messages as follows: 2070 o when the Proxy receives an RS message from a new ANET Client, it 2071 first authenticates the message then examines the network-layer 2072 destination address. If the destination address is a Server's 2073 AERO address, the Proxy proceeds to the next step. Otherwise, if 2074 the destination is all-routers multicast or subnet routers 2075 anycast, the Proxy selects a "nearby" Server that is likely to be 2076 a good candidate to serve the Client and replaces the destination 2077 address with the Server's AERO address. Next, the Proxy creates a 2078 proxy neighbor cache entry and caches the Client and Server 2079 addresses along with any identifying information including 2080 Transaction IDs, Client Identifiers, Nonce values, etc. The Proxy 2081 then inserts an SLLAO in the RS message with a single ifIndex- 2082 tuple matching the first ifIndex-tuple in the OMNI option and with 2083 the Link Layer Address and Port Number fields set to 0. The Proxy 2084 finally encapsulates the (proxyed) RS message in a SPAN header 2085 with destination set to the Server's SPAN address then forwards 2086 the message into the SPAN. 2088 o when the Server receives the RS message, it authenticates the 2089 message then creates or updates a symmetric neighbor cache entry 2090 for the Client with the Proxy's SPAN address as the link-layer 2091 address. The Server then sends an RA message back to the Proxy 2092 via the SPAN. 2094 o when the Proxy receives the RA message, it matches the message 2095 with the RS that created the proxy neighbor cache entry. The 2096 Proxy then caches the PD route information as a mapping from the 2097 Client's MNPs to the Client's ANET address, and sets the neighbor 2098 cache entry state to REACHABLE. The Proxy then forwards the 2099 (proxyed) message to the Client. 2101 After the initial RS/RA exchange, the Proxy forwards any Client data 2102 packets for which there is no matching asymmetric neighbor cache 2103 entry to a Relay via the SPAN. The Proxy instead forwards any Client 2104 data destined to an asymmetric neighbor cache target directly to the 2105 target according to the link-layer information - the process of 2106 establishing asymmetric neighbor cache entries is specified in 2107 Section 3.17. 2109 While the Client is still attached to the ANET, the Proxy send RS or 2110 unsolicited NA messages to update the Server's symmetric neighbor 2111 cache entries on behalf of the Client and/or to convey QoS updates. 2112 If the Server ceases to send solicited RA responses, the Proxy marks 2113 the Server as unreachable and sends an unsolicited RA with Router 2114 Lifetime set to zero to inform Clients that this Server is no longer 2115 able to provide service. Although the Proxy engages in ND exchanges 2116 on behalf of the Client, the Client can also send ND messages on its 2117 own behalf, e.g., if it is in a better position than the Proxy to 2118 convey QoS changes, etc. 2120 If the Client becomes unreachable, the Proxy sets the neighbor cache 2121 entry state to DEPARTED and retains the entry for DEPARTTIME seconds. 2122 While the state is DEPARTED, the Proxy forwards any packets destined 2123 to the Client to a Relay. The Relay in turn forwards the packets to 2124 the Client's current Server. When DepartTime expires, the Proxy 2125 deletes the neighbor cache entry and discards any further packets 2126 destined to this (now forgotten) Client. 2128 When a neighbor cache entry transitions to DEPARTED, some of the 2129 fragments of a multiple fragment packet may have already arrived at 2130 the Proxy while others are en route to the Client's new location. 2131 However, no special attention in the reassembly algorithm is 2132 necessary when re-routed packets are simply treated as loss. Since 2133 the fragments of a multiple-fragment packet are sent in minimal 2134 inter-packet delay bursts, such occasions will be rare. 2136 In some ANETs that employ a Proxy, the Client's MNP can be injected 2137 into the ANET routing system. In that case, the Client can send data 2138 messages without encapsulation so that the ANET native routing system 2139 transports the unencapsulated packets to the Proxy. This can be very 2140 beneficial, e.g., if the Client connects to the ANET via low-end data 2141 links such as some aviation wireless links. 2143 If the first-hop ANET access router is AERO-aware, the Client can 2144 avoid encapsulation for both its control and data messages. When the 2145 Client connects to the link, it can send an unencapsulated RS message 2146 with source address set to its AERO address and with destination 2147 address set to the AERO address of the Client's selected Server or to 2148 all-routers multicast or subnet router anycast. The Client includes 2149 an OMNI option formatted as specified in 2150 [I-D.templin-atn-aero-interface]. 2152 The Client then sends the unencapsulated RS message, which will be 2153 intercepted by the AERO-Aware access router. The access router then 2154 encapsulates the RS message in an ANET header with its own address as 2155 the source address and the address of a Proxy as the destination 2156 address. The access router further remembers the address of the 2157 Proxy so that it can encapsulate future data packets from the Client 2158 via the same Proxy. If the access router needs to change to a new 2159 Proxy, it simply sends another RS message toward the Server via the 2160 new Proxy on behalf of the Client. 2162 In some cases, the access router and Proxy may be one and the same 2163 node. In that case, the node would be located on the same physical 2164 link as the Client, but its message exchanges with the Server would 2165 need to pass through a security gateway at the ANET/INET border. The 2166 method for deploying access routers and Proxys (i.e. as a single node 2167 or multiple nodes) is an ANET-local administrative consideration. 2169 3.16.1. Detecting and Responding to Server Failures 2171 In environments where fast recovery from Server failure is required, 2172 Proxys SHOULD use proactive Neighbor Unreachability Detection (NUD) 2173 to track Server reachability in a similar fashion as for 2174 Bidirectional Forwarding Detection (BFD) [RFC5880]. Proxys can then 2175 quickly detect and react to failures so that cached information is 2176 re-established through alternate paths. The NUD control messaging is 2177 carried only over well-connected ground domain networks (i.e., and 2178 not low-end aeronautical radio links) and can therefore be tuned for 2179 rapid response. 2181 Proxys perform proactive NUD with Servers for which there are 2182 currently active ANET Clients by sending continuous NS messages in 2183 rapid succession, e.g., one message per second. The Proxy sends the 2184 NS message via the SPAN with the Proxy's AERO address as the source 2185 and the AERO address of the Server as the destination. If the Server 2186 fails (i.e., if the Proxy ceases to receive solicited NA messages), 2187 the Proxy can quickly inform Clients by sending RA messages on the 2188 ANET interface. The Proxy sends RA messages with source address set 2189 to the Server's address, destination address set to all-nodes 2190 multicast, and Router Lifetime set to 0. 2192 The Proxy SHOULD send MAX_FINAL_RTR_ADVERTISEMENTS RA messages 2193 separated by small delays [RFC4861]. Any Clients on the ANET that 2194 have been using the (now defunct) Server will receive the RA messages 2195 and associate with a new Server. 2197 3.17. AERO Route Optimization 2199 While data packets are flowing between a source and target node, 2200 route optimization SHOULD be used. Route optimization is initiated 2201 by the first eligible Route Optimization Source (ROS) closest to the 2202 source as follows: 2204 o For Clients on VPNed, NATed and Direct interfaces, the Server is 2205 the ROS. 2207 o For Clients on Proxyed interfaces, the Proxy is the ROS. 2209 o For Clients on native interfaces, the Client itself is the ROS. 2211 o For correspondent nodes on INET/EUN interfaces serviced by a 2212 Gateway, the Gateway is the ROS. 2214 The route optimization procedure is conducted between the ROS and the 2215 target Server/Gateway acting as a Route Optimization Responder (ROR) 2216 in the same manner as for IPv6 ND Address Resolution and using the 2217 same NS/NA messaging. The target may either be a MNP Client serviced 2218 by a Server, or a non-MNP correspondent reachable via a Gateway. 2220 The procedures are specified in the following sections. 2222 3.17.1. Route Optimization Initiation 2224 While data packets are flowing from the source node toward a target 2225 node, the ROS performs address resolution by sending an NS message to 2226 receive a solicited NA message from the ROR. 2228 When the ROS sends an NS, it includes the AERO address of the ROS as 2229 the source address (e.g., fe80::1) and the AERO address corresponding 2230 to the data packet's destination address as the destination address 2231 (e.g., if the destination address is 2001:db8:1:2::1 then the 2232 corresponding AERO address is fe80::2001:db8:1:2). The NS message 2233 includes an OMNI option with a single ifIndex-tuple with ifIndex set 2234 to 0. The message includes a Nonce and Timestamp option if the ROS 2235 needs to correlate NA replies. 2237 The ROS then encapsulates the NS message in a SPAN header with source 2238 set to its own SPAN address and destination set to the data packet's 2239 destination address, then sends the message into the SPAN without 2240 decrementing the network-layer TTL/Hop Limit field. 2242 3.17.2. Relaying the NS 2244 When the Relay receives the NS message from the ROS, it discards the 2245 INET header and determines that the ROR is the next hop by consulting 2246 its standard IPv6 forwarding table for the SPAN header destination 2247 address. The Relay then forwards the SPAN message toward the ROR the 2248 same as for any IPv6 router. The final-hop Relay in the SPAN will 2249 deliver the message via a secured tunnel to the ROR. 2251 3.17.3. Processing the NS and Sending the NA 2253 When the ROR receives the NS message, it examines the AERO 2254 destination address to determine whether it has a neighbor cache 2255 entry and/or route that matches the target; if not, it drops the NS 2256 message. Next, if the target belongs to an MNP Client neighbor in 2257 the DEPARTED state the ROR changes the NS message SPAN destination 2258 address to the SPAN address of the Client's new Server, forwards the 2259 message into the SPAN and returns from processing. 2261 If the target belongs to an MNP Client neighbor in the REACHABLE 2262 state, the ROR instead adds the AERO source address to the target 2263 Client's Report List with time set to ReportTime. If the target 2264 belongs to a non-MNP route, the ROR continues processing without 2265 adding an entry to the Report List. The ROR then prepares a 2266 solicited NA message to send back to the ROS but does not create a 2267 neighbor cache entry. The ROR sets the NA source address to the 2268 destination AERO address of the NS, and includes the Nonce value 2269 received in the NS plus the current Timestamp. 2271 If the target belongs to an MNP Client, the ROR then includes an OMNI 2272 option with prefix information set according to the MNP prefix 2273 length; otherwise, it sets it to the maximum of the non-MNP prefix 2274 length and 64. (Note that a /64 limit is imposed to avoid causing 2275 the ROS to set short prefixes (e.g., "default") that would match 2276 destinations for which the routing system includes more-specific 2277 prefixes.) 2279 The ROR next includes a first ifIndex-tuple in the OMNI option with 2280 with ifIndex set to 0. If the target belongs to an MNP Client, the 2281 ROR next includes additional ifIndex-tuples in the OMNI option for 2282 the target Client's underlying interfaces with current information 2283 for each interface 2285 The ROR then includes a TLLAO option with ifIndex-tuples in one-to- 2286 one correspondence with the tuples that appear in the OMNI option. 2287 For NATed, VPNed and Direct interfaces, the link layer addresses are 2288 the SPAN address of the ROR. For Proxyed and native interfaces, the 2289 link-layer addresses are the SPAN addresses of the Proxys and the 2290 Client's native interfaces. 2292 The ROR finally encapsulates the NA message in a SPAN header with 2293 source set to its own SPAN address and destination set to the source 2294 SPAN address of the NS message, then forwards the message into the 2295 SPAN without decrementing the network-layer TTL/Hop Limit field. 2297 3.17.4. Relaying the NA 2299 When the Relay receives the NA message from the ROR, it discards the 2300 INET header and determines that the ROS is the next hop by consulting 2301 its standard IPv6 forwarding table for the SPAN header destination 2302 address. The Relay then forwards the SPAN-encapsulated NA message 2303 toward the ROS the same as for any IPv6 router. The final-hop Relay 2304 in the SPAN will deliver the message via a secured tunnel to the ROS. 2306 3.17.5. Processing the NA 2308 When the ROS receives the solicited NA message, it caches the source 2309 SPAN address then discards the INET and SPAN headers. The ROS next 2310 verifies the Nonce and Timestamp values (if present), then creates an 2311 asymmetric neighbor cache entry for the ROR and caches all 2312 information found in the solicited NA OMNI and TLLAO options. The 2313 ROS finally sets the asymmetric neighbor cache entry lifetime to 2314 ReachableTime seconds. 2316 3.17.6. Route Optimization Maintenance 2318 Following route optimization, the ROS forwards future data packets 2319 destined to the target via the addresses found in the cached link- 2320 layer information. The route optimization is shared by all sources 2321 that send packets to the target via the ROS, i.e., and not just the 2322 source on behalf of which the route optimization was initiated. 2324 While new data packets destined to the target are flowing through the 2325 ROS, it sends additional NS messages to the ROR before ReachableTime 2326 expires to receive a fresh solicited NA message the same as described 2327 in the previous sections. (Route optimization refreshment strategies 2328 are an implementation matter, with a non-normative example given in 2329 Appendix B.1). The ROS uses the cached SPAN address of the ROR as 2330 the NS SPAN destination address, and sends up to MAX_UNICAST_SOLICIT 2331 NS messages separated by 1 second until an NA is received. If no NA 2332 is received, the ROS assumes the current ROR has become unreachable 2333 and deletes the neighbor cache entry. Subsequent data packets will 2334 trigger a new route optimization per Section 3.17.1 to discover a new 2335 ROR while initial data packets travel over a suboptimal route. 2337 If an NA is received, the ROS then updates the asymmetric neighbor 2338 cache entry to refresh ReachableTime, while (for MNP destinations) 2339 the ROR adds or updates the ROS address to the target Client's Report 2340 List and with time set to ReportTime. While no data packets are 2341 flowing, the ROS instead allows ReachableTime for the asymmetric 2342 neighbor cache entry to expire. When ReachableTime expires, the ROS 2343 deletes the asymmetric neighbor cache entry. Future data packets 2344 flowing through the ROS will again trigger a new route optimization. 2346 The ROS may also receive unsolicited NA messages from the ROR at any 2347 time. If there is an asymmetric neighbor cache entry for the target, 2348 the ROS updates the link-layer information but does not update 2349 ReachableTime since the receipt of an unsolicited NA does not confirm 2350 that the forward path is still working. If there is no asymmetric 2351 neighbor cache entry, the ROS simply discards the unsolicited NA. 2352 Cases in which unsolicited NA messages are generated are specified in 2353 Section 3.19. 2355 In this arrangement, the ROS holds an asymmetric neighbor cache entry 2356 for the ROR, but the ROR does not hold an asymmetric neighbor cache 2357 entry for the ROS. The route optimization neighbor relationship is 2358 therefore asymmetric and unidirectional. If the target node also has 2359 packets to send back to the source node, then a separate route 2360 optimization procedure is performed in the reverse direction. But, 2361 there is no requirement that the forward and reverse paths be 2362 symmetric. 2364 3.18. Neighbor Unreachability Detection (NUD) 2366 AERO nodes perform Neighbor Unreachability Detection (NUD) per 2367 [RFC4861]. NUD is performed either reactively in response to 2368 persistent link-layer errors (see Section 3.14) or proactively to 2369 confirm reachability. The NUD algorithm may further be seeded by ND 2370 hints of forward progress, but care must be taken to avoid inferring 2371 reachability based on spoofed information. 2373 When an ROR directs an ROS to a neighbor with one or more target 2374 link-layer addresses, the ROS can proactively test each direct path 2375 by sending an initial NS message to elicit a solicited NA response. 2376 While testing the paths, the ROS can optionally continue sending 2377 packets via the SPAN, maintain a small queue of packets until target 2378 reachability is confirmed, or (optimistically) allow packets to flow 2379 via the direct paths. In any case, the ROS should only consider the 2380 neighbor unreachable if NUD fails over multiple target link-layer 2381 address paths. 2383 When a ROS sends an NS message used for NUD, it uses its AERO 2384 addresses as the IPv6 source address and the AERO address 2385 corresponding to a target link-layer address as the destination. For 2386 each target link-layer address, the source node encapsulates the NS 2387 message in SPAN/INET headers with its own SPAN address as the source 2388 and the SPAN address of the target as the destination, If the target 2389 is located within the same SPAN segment, the source sets the INET 2390 address of the target as the destination; otherwise, it sets the INET 2391 address of a Relay as the destination. The source then forwards the 2392 message into the SPAN. 2394 Paths that pass NUD tests are marked as "reachable", while those that 2395 do not are marked as "unreachable". These markings inform the AERO 2396 interface forwarding algorithm specified in Section 3.12. 2398 Proxys can perform NUD to verify Server reachability on behalf of 2399 their proxyed Clients so that the Clients need not engage in NUD 2400 messaging themselves. 2402 3.19. Mobility Management and Quality of Service (QoS) 2404 AERO is a Distributed Mobility Management (DMM) service. Each Server 2405 is responsible for only a subset of the Clients on the AERO link, as 2406 opposed to a Centralized Mobility Management (CMM) service where 2407 there is a single network mobility service for all Clients. Clients 2408 coordinate with their associated Servers via RS/RA exchanges to 2409 maintain the DMM profile, and the AERO routing system tracks all 2410 current Client/Server peering relationships. 2412 Servers provide a Mobility Anchor Point (MAP) for their dependent 2413 Clients. Clients are responsible for maintaining neighbor 2414 relationships with their Servers through periodic RS/RA exchanges, 2415 which also serves to confirm neighbor reachability. When a Client's 2416 underlying interface address and/or QoS information changes, the 2417 Client is responsible for updating the Server with this new 2418 information. Note that for Proxyed interfaces, however, the Proxy 2419 can perform the RS/RA exchanges on the Client's behalf. 2421 Mobility management considerations are specified in the following 2422 sections. 2424 3.19.1. Mobility Update Messaging 2426 Servers acting as MAPs accommodate Client mobility and/or QoS change 2427 events by sending unsolicited NA messages to each ROS in the target 2428 Client's Report List. When a MAP sends an unsolicited NA message, it 2429 sets the IPv6 source address to the Client's AERO address and sets 2430 the IPv6 destination address to all-nodes multicast (ff02::1). The 2431 MAP also includes an OMNI option with a first ifIndex-tuple with 2432 ifIndex set to 0, and with additional ifIndex-tuples for the target 2433 Client's remaining interfaces. The MAP then includes a TLLAO with 2434 corresponding ifIndex-tuples, with the link layer address of the 2435 first tuple set to the MAP's SPAN address and with link layer 2436 addresses of the remaining tuples set to the corresponding target 2437 SPAN addresses. The MAP finally encapsulates the message in a SPAN 2438 header with source set to its own SPAN address and destination set to 2439 the SPAN address of the ROS, then sends the message to a Relay. 2441 As discussed in Section 7.2.6 of [RFC4861], the transmission and 2442 reception of unsolicited NA messages is unreliable but provides a 2443 useful optimization. In well-connected Internetworks with robust 2444 data links unsolicited NA messages will be delivered with high 2445 probability, but in any case the MAP can optionally send up to 2446 MAX_NEIGHBOR_ADVERTISEMENT unsolicited NAs to each ROS to increase 2447 the likelihood that at least one will be received. 2449 When the ROS receives an unsolicited NA message, it ignores the 2450 message if there is no existing neighbor cache entry for the Client. 2451 Otherwise, it uses the included OMNI option and TLLAO information to 2452 update the neighbor cache entry, but does not reset ReachableTime 2453 since the receipt of an unsolicited NA message from the target Server 2454 does not provide confirmation that any forward paths to the target 2455 Client are working. 2457 If unsolicited NA messages are lost, the ROS may be left with stale 2458 address and/or QoS information for the Client for up to REACHABLETIME 2459 seconds. During this time, the ROS can continue sending packets 2460 according to its stale neighbor cache information. When 2461 ReachableTime is close to expiring, the ROS will re-initiate route 2462 optimization and receive fresh state information. 2464 In addition to sending unsolicited NA messages to the current set of 2465 ROSs for the Client, the MAP also sends unsolicited NAs to the former 2466 link-layer address for any ifIndex-tuple for which the link-layer 2467 address has changed. The NA messages update Proxys or Servers that 2468 cannot easily detect (e.g., without active probing) when a formerly- 2469 active Client has departed. 2471 3.19.2. Announcing Link-Layer Address and/or QoS Preference Changes 2473 When a Client needs to change its ANET addresses and/or QoS 2474 preferences (e.g., due to a mobility event), either the Client or its 2475 Proxys send RS messages to the Server via the SPAN with an OMNI 2476 option and SLLAO that include an ifIndex-tuple with the new link 2477 quality and address information. 2479 Up to MAX_RTR_SOLICITATION RS messages MAY be sent in parallel with 2480 sending actual data packets in case one or more RAs are lost. If all 2481 RAs are lost, the Client SHOULD re-associate with a new Server. 2483 When the Server receives the Client's changes, it sends unsolicited 2484 NA messages to all nodes in the Report List the same as described in 2485 the previous section. 2487 3.19.3. Bringing New Links Into Service 2489 When a Client needs to bring new underlying interfaces into service 2490 (e.g., when it activates a new data link), it sends an RS message to 2491 the Server via the underlying interface with an OMNI option with 2492 appropriate link quality values and with an SLLAO (if necessary) with 2493 link-layer address information for the new link.. 2495 3.19.4. Removing Existing Links from Service 2497 When a Client needs to remove existing underlying interfaces from 2498 service (e.g., when it de-activates an existing data link), it sends 2499 an RS message to its Server with an OMNI option with appropriate link 2500 quality values. 2502 If the Client needs to send RS messages over an underlying interface 2503 other than the one being removed from service, it MUST include an 2504 ifIndex-tuple for the sending interface as the first tuple and 2505 include additional ifIndex-tuples with appropriate link quality 2506 values for any underlying interfaces being removed from service. 2508 3.19.5. Moving to a New Server 2510 When a Client associates with a new Server, it performs the Client 2511 procedures specified in Section 3.15.2. The Client also includes a 2512 notification identifier in the RS message OMNI option per 2513 [I-D.templin-atn-aero-interface] if it wants the new Server to notify 2514 the old Server. 2516 When the new Server receives the Client's RS message, it responds by 2517 returning an RA as specified in Section 3.15.3. If the Client's RS 2518 includes a notification identifier, the new Server also prepares an 2519 RS to send to the old Server. The RS message includes the Client's 2520 AERO address as the source address, the old Server's AERO address as 2521 the destination address, and an OMNI option and SLLAO with an 2522 ifIndex-tuple with ifIndex set to 0. The OMNI option includes a 2523 release indication, and the SLLAO includes the link-layer address of 2524 the new Server. The new Server retries up to MAX_RTR_SOLICITATIONS 2525 attempts until an RA is received. (Note that the Client can 2526 alternatively send RS messages with a release indication to the old 2527 Server on its own behalf, however, this additional Client messaging 2528 may be undesirable in some environments.) 2530 When the old Server processes the RS, it changes the symmetric 2531 neighbor cache entry state to DEPARTED, sets the link-layer address 2532 of the Client to the address found in the RS SLLAO, and sets 2533 DepartTime to DEPARTTIME seconds. The old Server then returns an 2534 immediate RA message with Router Lifetime set to 0. After a short 2535 delay (e.g., 2 seconds) the old Server withdraws the Client's MNP 2536 from the routing system. After DepartTime expires, the old Server 2537 deletes the symmetric neighbor cache entry. 2539 The old Server also sends unsolicited NA messages to all ROSs in the 2540 Client's Report List with an OMNI option and TLLAO with a single 2541 ifIndex-tuple with ifIndex set to 0 and with the link-layer address 2542 of the new Server. When the ROS receives the NA, it caches the 2543 address of the new Server in the existing asymmetric neighbor cache 2544 entry and marks the entry as STALE. Subsequent data packets will 2545 then flow according to any existing cached link-layer information and 2546 trigger a new NS/NA exchange via the new Server. 2548 Clients SHOULD NOT move rapidly between Servers in order to avoid 2549 causing excessive oscillations in the AERO routing system. Examples 2550 of when a Client might wish to change to a different Server include a 2551 Server that has gone unreachable, topological movements of 2552 significant distance, movement to a new geographic region, movement 2553 to a new SPAN segment, etc. 2555 When a Client moves to a new Server, some of the fragments of a 2556 multiple fragment packet may have already arrived at the old Server 2557 while others are en route to the new Server. However, no special 2558 attention in the reassembly algorithm is necessary when re-routed 2559 fragments are simply treated as loss. Since the fragments of a 2560 multiple-fragment packet are sent with minimal inter-packet delay, 2561 such occasions will be rare. 2563 3.20. Multicast 2565 The AERO Client provides an IGMP (IPv4) [RFC2236] or MLD (IPv6) 2566 [RFC3810] proxy service for its EUNs and/or hosted applications 2567 [RFC4605]. The Client forwards IGMP/MLD messages over any of its 2568 underlying interfaces for which group membership is required. The 2569 IGMP/MLD messages may be further forwarded by a first-hop ANET access 2570 router acting as an IGMP/MLD-snooping switch [RFC4541], then 2571 ultimately delivered to an AERO Proxy/Server acting as a Protocol 2572 Independent Multicast - Sparse-Mode (PIM-SM, or simply "PIM") 2573 Designated Router (DR) [RFC7761]. AERO Gateways also act as PIM 2574 routers (i.e., the same as AERO Proxys/Servers) on behalf of nodes on 2575 INET/EUN networks. The behaviors identified in the following 2576 sections correspond to Source-Specific Multicast (SSM) and Any-Source 2577 Multicast (ASM) operational modes. 2579 3.20.1. Source-Specific Multicast (SSM) 2581 When an ROS (i.e., an AERO Proxy/Server/Gateway) "X" acting as PIM 2582 router receives a Join/Prune message from a node on its downstream 2583 interfaces containing one or more ((S)ource, (G)roup) pairs, it 2584 updates its Multicast Routing Information Base (MRIB) accordingly. 2585 For each S belonging to a prefix reachable via X's non-AERO 2586 interfaces, X then forwards the (S, G) Join/Prune to any PIM routers 2587 on those interfaces per [RFC7761]. 2589 For each S belonging to a prefix reachable via X's AERO interface, X 2590 originates a separate copy of the Join/Prune for each (S,G) in the 2591 message using its own AERO address as the source address and ALL-PIM- 2592 ROUTERS as the destination address. X then encapsulates each message 2593 in a SPAN header with source address set to the SPAN address of X and 2594 destination address set to S then forwards the message into the SPAN. 2595 The SPAN in turn forwards the message to AERO Server/Gateway "Y" that 2596 services S. At the same time, if the message was a Join, X sends a 2597 route-optimization NS message toward each S the same as discussed in 2598 Section 3.17. The resulting NAs will return the AERO address for the 2599 prefix that matches S as the network-layer source address and TLLAOs 2600 with the SPAN addresses corresponding to any ifIndex-tuples that are 2601 currently servicing S. 2603 When Y processes the Join/Prune message, if S located behind any 2604 Native, Direct, VPNed or NATed interfaces Y acts as a PIM router and 2605 updates its MRIB to list X as the next hop in the reverse path. If S 2606 is located behind any Proxys "Z"*, Y also forwards the message to 2607 each Z* over the SPAN while continuing to use the AERO address of X 2608 as the source address. Each Z* then updates its MRIB accordingly and 2609 maintains the AERO address of X as the next hop in the reverse path. 2610 Since the Relays in the SPAN do not examine network layer control 2611 messages, this means that the (reverse) multicast tree path is simply 2612 from each Z* (and/or Y) to X with no other multicast-aware routers in 2613 the path. If any Z* (and/or Y) is located on the same SPAN segment 2614 as X, the multicast data traffic sent to X directly using SPAN/INET 2615 encapsulation instead of via a Relay. 2617 Following the initial Join/Prune and NS/NA messaging, X maintains an 2618 asymmetric neighbor cache entry for each S the same as if X was 2619 sending unicast data traffic to S. In particular, X performs 2620 additional NS/NA exchanges to keep the neighbor cache entry alive for 2621 up to t_periodic seconds [RFC7761]. If no new Joins are received 2622 within t_periodic seconds, X allows the neighbor cache entry to 2623 expire. Finally, if X receives any additional Join/Prune messages 2624 for (S,G) it forwards the messages to each Y and Z* in the neighbor 2625 cache entry over the SPAN. 2627 At some later time, Client C that holds an MNP for source S may 2628 depart from a first Proxy Z1 and/or connect via a new Proxy Z2. In 2629 that case, Y sends an unsolicited NA message to X the same as 2630 specified for unicast mobility in Section 3.19. When X receives the 2631 unsolicited NA message, it updates its asymmetric neighbor cache 2632 entry for the AERO address for source S and sends new Join messages 2633 to any new Proxys Z2. There is no requirement to send any Prune 2634 messages to old Proxys Z1 since source S will no longer source any 2635 multicast data traffic via Z1. Instead, the multicast state for 2636 (S,G) in Proxy Z1 will soon time out since no new Joins will arrive. 2638 After some later time, C may move to a new Server Y2 and depart from 2639 old Sever Y1. In that case, Y1 sends Join messages for any of C's 2640 active (S,G) groups to Y2 while including its own AERO address as the 2641 source address. This causes Y2 to include Y1 in the multicast 2642 forwarding tree during the interim time that Y1's symmetric neighbor 2643 cache entry for C is in the DEPARTED state. At the same time, Y1 2644 sends an unsolicited NA message to X with an OMNI option and TLLAO 2645 with ifIndex-tuple set to 0 and a release indication to cause X to 2646 release its asymmetric neighbor cache entry. X then sends a new Join 2647 message to S via the SPAN and re-initiates route optimization the 2648 same as if it were receiving a fresh Join message from a node on a 2649 downstream link. 2651 3.20.2. Any-Source Multicast (ASM) 2653 When an ROS X acting as a PIM router receives a Join/Prune from a 2654 node on its downstream interfaces containing one or more (*,G) pairs, 2655 it updates its Multicast Routing Information Base (MRIB) accordingly. 2656 X then forwards a copy of the message to the Rendezvous Point (RP) R 2657 for each G over the SPAN. X uses its own AERO address as the source 2658 address and ALL-PIM-ROUTERS as the destination address, then 2659 encapsulates each message in a SPAN header with source address set to 2660 the SPAN address of X and destination address set to R, then sends 2661 the message into the SPAN. At the same time, if the message was a 2662 Join X initiates NS/NA route optimization the same as for the SSM 2663 case discussed in Section 3.20.1. 2665 For each source S that sends multicast traffic to group G via R, the 2666 Proxy/Server Z* for the Client that aggregates S encapsulates the 2667 packets in PIM Register messages and forwards them to R via the SPAN. 2668 R may then elect to send a PIM Join to Z* over the SPAN. This will 2669 result in an (S,G) tree rooted at Z* with R as the next hop so that R 2670 will begin to receive two copies of the packet; one native copy from 2671 the (S, G) tree and a second copy from the pre-existing (*, G) tree 2672 that still uses PIM Register encapsulation. R can then issue a PIM 2673 Register-stop message to suppress the Register-encapsulated stream. 2674 At some later time, if C moves to a new Proxy/Server Z*, it resumes 2675 sending packets via PIM Register encapsulation via the new Z*. 2677 At the same time, as multicast listeners discover individual S's for 2678 a given G, they can initiate an (S,G) Join for each S under the same 2679 procedures discussed in Section 3.20.1. Once the (S,G) tree is 2680 established, the listeners can send (S, G) Prune messages to R so 2681 that multicast packets for group G sourced by S will only be 2682 delivered via the (S, G) tree and not from the (*, G) tree rooted at 2683 R. All mobility considerations discussed for SSM apply. 2685 3.20.3. Bi-Directional PIM (BIDIR-PIM) 2687 Bi-Directional PIM (BIDIR-PIM) [RFC5015] provides an alternate 2688 approach to ASM that treats the Rendezvous Point (RP) as a Designated 2689 Forwarder (DF). Further considerations for BIDIR-PIM are out of 2690 scope. 2692 3.21. Operation over Multiple AERO Links (VLANs) 2694 An AERO Client can connect to multiple AERO links the same as for any 2695 data link service. In that case, the Client maintains a distinct 2696 AERO interface for each link, e.g., 'aero0' for the first link, 2697 'aero1' for the second, 'aero2' for the third, etc. Each AERO link 2698 would include its own distinct set of Relays, Servers and Proxys, 2699 thereby providing redundancy in case of failures. 2701 The Relays, Servers and Proxys on each AERO link can assign AERO and 2702 SPAN addresses that use the same or different numberings from those 2703 on other links. Since the links are mutually independent there is no 2704 requirement for avoiding inter-link address duplication, e.g., the 2705 same AERO address such as fe80::1000 could be used to number distinct 2706 nodes that connect to different AERO links. 2708 Each AERO link could utilize the same or different ANET connections. 2709 The links can be distinguished at the link-layer via Virtual Local 2710 Area Network (VLAN) tagging (e.g., IEEE 802.1Q) and/or through 2711 assignment of distinct sets of MSPs on each link. This gives rise to 2712 the opportunity for supporting multiple redundant networked paths, 2713 where each VLAN is distinguished by a different label (e.g., colors 2714 such as Red, Green, Blue, etc.). In particular, the Client can tag 2715 its RS messages with the appropriate label to cause the network to 2716 select the desired VLAN. 2718 Clients that connect to multiple AERO interfaces can select the 2719 outgoing interface appropriate for a given Red/Blue/Green/etc. 2720 traffic profile while (in the reverse direction) correspondent nodes 2721 must have some way of steering their packets destined to a target via 2722 the correct AERO link. 2724 In a first alternative, if each AERO link services different MSPs, 2725 then the Client can receive a distinct MNP from each of the links. 2726 IP routing will therefore assure that the correct Red/Green/Blue/etc. 2727 network is used for both outbound and inbound traffic. This can be 2728 accomplished using existing technologies and approaches, and without 2729 requiring any special supporting code in correspondent nodes or 2730 Relays. 2732 In a second alternative, if each AERO link services the same MSP(s) 2733 then each link could assign a distinct "AERO Link Anycast" address 2734 that is configured by all Relays on the link. Correspondent nodes 2735 then include a "type 4" routing header with the Anycast address for 2736 the AERO link as the IPv6 destination and with the address of the 2737 target encoded as the "next segment" in the routing header 2738 [RFC8402][I-D.ietf-6man-segment-routing-header]. Standard IP routing 2739 will then direct the packet to the nearest Relay for the correct AERO 2740 link, which will replace the destination address with the target 2741 address then forward the packet to the target. 2743 3.22. DNS Considerations 2745 AERO Client MNs and INET correspondent nodes consult the Domain Name 2746 System (DNS) the same as for any Internetworking node. When 2747 correspondent nodes and Client MNs use different IP protocol versions 2748 (e.g., IPv4 correspondents and IPv6 MNs), the INET DNS must maintain 2749 A records for IPv4 address mappings to MNs which must then be 2750 populated in Gateway NAT64 mapping caches. In that way, an IPv4 2751 correspondent node can send packets to the IPv4 address mapping of 2752 the target MN, and the Gateway will translate the IPv4 header and 2753 destination address into an IPv6 header and IPv6 destination address 2754 of the MN. 2756 When an AERO Client registers with an AERO Server, the Server can 2757 return the address(es) of DNS servers in RDNSS options [RFC6106]. 2758 The DNS server provides the IP addresses of other MNs and 2759 correspondent nodes in AAAA records for IPv6 or A records for IPv4. 2761 3.23. Transition Considerations 2763 The SPAN ensures that dissimilar INET partitions can be joined into a 2764 single unified AERO link, even though the partitions themselves may 2765 have differing protocol versions and/or incompatible addressing 2766 plans. However, a commonality can be achieved by incrementally 2767 distributing globally routable (i.e., native) IP prefixes to 2768 eventually reach all nodes (both mobile and fixed) in all SPAN 2769 segments. This can be accomplished by incrementally deploying AERO 2770 Gateways on each INET partition, with each Gateway distributing its 2771 MNPs and/or discovering non-MNP prefixes on its INET links. 2773 This gives rise to the opportunity to eventually distribute native IP 2774 addresses to all nodes, and to present a unified AERO link view 2775 (bridged by the SPAN) even if the INET partitions remain in their 2776 current protocol and addressing plans. In that way, the AERO link 2777 can serve the dual purpose of providing a mobility service and a 2778 transition service. Or, if an INET partition is transitioned to a 2779 native IP protocol version and addressing scheme that is compatible 2780 with the AERO link MNP-based addressing scheme, the partition and 2781 AERO link can be joined by Gateways. 2783 Gateways that connect INETs/EUNs with dissimilar IP protocol versions 2784 must employ a network address and protocol translation function such 2785 as NAT64[RFC6146]. 2787 3.24. Detecting and Reacting to Server and Relay Failures 2789 In environments where rapid failure recovery is required, Servers and 2790 Relays SHOULD use Bidirectional Forwarding Detection (BFD) [RFC5880]. 2791 Nodes that use BFD can quickly detect and react to failures so that 2792 cached information is re-established through alternate nodes. BFD 2793 control messaging is carried only over well-connected ground domain 2794 networks (i.e., and not low-end radio links) and can therefore be 2795 tuned for rapid response. 2797 Servers and Relays maintain BFD sessions in parallel with their BGP 2798 peerings. If a Server or Relay fails, BGP peers will quickly re- 2799 establish routes through alternate paths the same as for common BGP 2800 deployments. 2802 Proxys SHOULD use proactive NUD for Servers for which there are 2803 currently active ANET Clients in a manner that parallels BFD, i.e., 2804 by sending unicast NS messages in rapid succession to receive 2805 solicited NA messages. If a Server fails, the Proxy will cease to 2806 receive NA messages and can quickly inform Clients of the outage by 2807 sending RA messages on the ANET interface. 2809 The Proxy sends RA messages with source address set to the Server's 2810 address, destination address set to all-nodes multicast, and Router 2811 Lifetime set to 0. The Proxy SHOULD send 2812 MAX_FINAL_RTR_ADVERTISEMENTS RA messages separated by small delays 2813 [RFC4861]. Any Clients on the ANET interface that have been using 2814 the (now defunct) Server will receive the RA messages and associate 2815 with a new Server. 2817 4. Implementation Status 2819 An AERO implementation based on OpenVPN (https://openvpn.net/) was 2820 announced on the v6ops mailing list on January 10, 2018 and an 2821 initial public release of the AERO proof-of-concept source code was 2822 announced on the intarea mailing list on August 21, 2015. The latest 2823 versions are available at: http://linkupnetworks.net/aero. 2825 5. IANA Considerations 2827 The IANA has assigned a 4-octet Private Enterprise Number "45282" for 2828 AERO in the "enterprise-numbers" registry. 2830 The IANA has assigned the UDP port number "8060" for an earlier 2831 experimental version of AERO [RFC6706]. This document obsoletes 2832 [RFC6706] and claims the UDP port number "8060" for all future use. 2834 No further IANA actions are required. 2836 6. Security Considerations 2838 AERO Relays configure secured tunnels with AERO Servers and Proxys 2839 within their local SPAN segments. Applicable secured tunnel 2840 alternatives include IPsec [RFC4301], TLS/SSL [RFC8446], DTLS 2841 [RFC6347], etc. The AERO Relays of all SPAN segments in turn 2842 configure secured tunnels for their neighboring AERO Relays across 2843 the SPAN. Therefore, packets that traverse the SPAN between any pair 2844 of AERO link neighbors are already secured. 2846 AERO Servers, Gateways and Proxys targeted by a route optimization 2847 may also receive packets directly from the INET partitions instead of 2848 via the SPAN. For INET partitions that apply effective ingress 2849 filtering to defeat source address spoofing, the simple data origin 2850 authentication procedures in Section 3.11 can be applied. This 2851 implies that the ROS list must be maintained consistently by all 2852 route optimization targets within the same INET partition, and that 2853 the ROS list must be securely managed by the partition administrative 2854 authority. 2856 For INET partitions that cannot apply effective ingress filtering, 2857 the two options for securing communications include 1) disable route 2858 optimization so that all traffic is conveyed over secured tunnels via 2859 the SPAN, or 2) enable on-demand secure tunnel creation between INET 2860 partition neighbors. Option 1) would result in longer routes than 2861 necessary and traffic concentration on critical infrastructure 2862 elements. Option 2) could be coordinated by establishing a secured 2863 tunnel on-demand instead of performing an NS/NA exchange in the route 2864 optimization procedures. Procedures for establishing on-demand 2865 secured tunnels are out of scope. 2867 AERO Clients that connect to secured enclaves need not apply security 2868 to their ND messages, since the messages will be intercepted by a 2869 perimeter Proxy that applies security on its outward-facing 2870 interface. AERO Clients located outside of secured enclaves SHOULD 2871 use symmetric network and/or transport layer security services, but 2872 when there are many prospective neighbors with dynamically changing 2873 connectivity an asymmetric security service such as SEND may be 2874 needed (see: Appendix B.6). 2876 Application endpoints SHOULD use application-layer security services 2877 such as TLS/SSL, DTLS or SSH [RFC4251] to assure the same level of 2878 protection as for critical secured Internet services. AERO Clients 2879 that require host-based VPN services SHOULD use symmetric network 2880 and/or transport layer security services such as IPsec, TLS/SSL, 2881 DTLS, etc. AERO Proxys and Servers can also provide a network-based 2882 VPN service on behalf of the Client, e.g., if the Client is located 2883 within a secured enclave and cannot establish a VPN on its own 2884 behalf. 2886 AERO Servers and Relays present targets for traffic amplification 2887 Denial of Service (DoS) attacks. This concern is no different than 2888 for widely-deployed VPN security gateways in the Internet, where 2889 attackers could send spoofed packets to the gateways at high data 2890 rates. This can be mitigated by connecting Servers and Relays over 2891 dedicated links with no connections to the Internet and/or when 2892 connections to the Internet are only permitted through well-managed 2893 firewalls. Traffic amplification DoS attacks can also target an AERO 2894 Client's low data rate links. This is a concern not only for Clients 2895 located on the open Internet but also for Clients in secured 2896 enclaves. AERO Servers and Proxys can institute rate limits that 2897 protect Clients from receiving packet floods that could DoS low data 2898 rate links. 2900 AERO Gateways must implement ingress filtering to avoid a spoofing 2901 attack in which spurious SPAN messages are injected into an AERO link 2902 from an outside attacker. AERO Clients MUST ensure that their 2903 connectivity is not used by unauthorized nodes on their EUNs to gain 2904 access to a protected network, i.e., AERO Clients that act as routers 2905 MUST NOT provide routing services for unauthorized nodes. (This 2906 concern is no different than for ordinary hosts that receive an IP 2907 address delegation but then "share" the address with other nodes via 2908 some form of Internet connection sharing such as tethering.) 2910 The MAP list and ROS lists MUST be well-managed and secured from 2911 unauthorized tampering, even though the list contains only public 2912 information. The MAP list can be conveyed to the Client in a similar 2913 fashion as in [RFC5214] (e.g., through layer 2 data link login 2914 messaging, secure upload of a static file, DNS lookups, etc.). The 2915 ROS list can be conveyed to Servers and Proxys through administrative 2916 action, secured file distribution, etc. 2918 Although public domain and commercial SEND implementations exist, 2919 concerns regarding the strength of the cryptographic hash algorithm 2920 have been documented [RFC6273] [RFC4982]. 2922 Security considerations for accepting link-layer ICMP messages and 2923 reflected packets are discussed throughout the document. 2925 7. Acknowledgements 2927 Discussions in the IETF, aviation standards communities and private 2928 exchanges helped shape some of the concepts in this work. 2929 Individuals who contributed insights include Mikael Abrahamsson, Mark 2930 Andrews, Fred Baker, Bob Braden, Stewart Bryant, Brian Carpenter, 2931 Wojciech Dec, Ralph Droms, Adrian Farrel, Nick Green, Sri Gundavelli, 2932 Brian Haberman, Bernhard Haindl, Joel Halpern, Tom Herbert, Sascha 2933 Hlusiak, Lee Howard, Andre Kostur, Hubert Kuenig, Ted Lemon, Andy 2934 Malis, Satoru Matsushima, Tomek Mrugalski, Madhu Niraula, Alexandru 2935 Petrescu, Behcet Saikaya, Michal Skorepa, Joe Touch, Bernie Volz, 2936 Ryuji Wakikawa, Tony Whyman, Lloyd Wood and James Woodyatt. Members 2937 of the IESG also provided valuable input during their review process 2938 that greatly improved the document. Special thanks go to Stewart 2939 Bryant, Joel Halpern and Brian Haberman for their shepherding 2940 guidance during the publication of the AERO first edition. 2942 This work has further been encouraged and supported by Boeing 2943 colleagues including Kyle Bae, M. Wayne Benson, Dave Bernhardt, Cam 2944 Brodie, Balaguruna Chidambaram, Irene Chin, Bruce Cornish, Claudiu 2945 Danilov, Don Dillenburg, Joe Dudkowski, Wen Fang, Samad Farooqui, 2946 Anthony Gregory, Jeff Holland, Seth Jahne, Brian Jaury, Greg 2947 Kimberly, Ed King, Madhuri Madhava Badgandi, Laurel Matthew, Gene 2948 MacLean III, Rob Muszkiewicz, Vijay Rajagopalan, Sean O'Sullivan, 2949 Greg Saccone, Rod Santiago, Kent Shuey, Brian Skeen, Mike Slane, 2950 Carrie Spiker, Katie Tran, Brendan Williams, Julie Wulff, Yueli Yang, 2951 Eric Yeh and other members of the BR&T and BIT mobile networking 2952 teams. Kyle Bae, Wayne Benson, Katie Tran and Eric Yeh are 2953 especially acknowledged for implementing the AERO functions as 2954 extensions to the public domain OpenVPN distribution. 2956 Earlier works on NBMA tunneling approaches are found in 2957 [RFC2529][RFC5214][RFC5569]. 2959 Many of the constructs presented in this second edition of AERO are 2960 based on the author's earlier works, including: 2962 o The Internet Routing Overlay Network (IRON) 2963 [RFC6179][I-D.templin-ironbis] 2965 o Virtual Enterprise Traversal (VET) 2966 [RFC5558][I-D.templin-intarea-vet] 2968 o The Subnetwork Encapsulation and Adaptation Layer (SEAL) 2969 [RFC5320][I-D.templin-intarea-seal] 2971 o AERO, First Edition [RFC6706] 2973 Note that these works cite numerous earlier efforts that are not also 2974 cited here due to space limitations. The authors of those earlier 2975 works are acknowledged for their insights. 2977 This work is aligned with the NASA Safe Autonomous Systems Operation 2978 (SASO) program under NASA contract number NNA16BD84C. 2980 This work is aligned with the FAA as per the SE2025 contract number 2981 DTFAWA-15-D-00030. 2983 This work is aligned with the Boeing Commercial Airplanes (BCA) 2984 Internet of Things (IoT) and autonomy programs. 2986 This work is aligned with the Boeing Information Technology (BIT) 2987 MobileNet program. 2989 8. References 2991 8.1. Normative References 2993 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 2994 DOI 10.17487/RFC0791, September 1981, 2995 . 2997 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 2998 RFC 792, DOI 10.17487/RFC0792, September 1981, 2999 . 3001 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 3002 Requirement Levels", BCP 14, RFC 2119, 3003 DOI 10.17487/RFC2119, March 1997, 3004 . 3006 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 3007 IPv6 Specification", RFC 2473, DOI 10.17487/RFC2473, 3008 December 1998, . 3010 [RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black, 3011 "Definition of the Differentiated Services Field (DS 3012 Field) in the IPv4 and IPv6 Headers", RFC 2474, 3013 DOI 10.17487/RFC2474, December 1998, 3014 . 3016 [RFC3971] Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander, 3017 "SEcure Neighbor Discovery (SEND)", RFC 3971, 3018 DOI 10.17487/RFC3971, March 2005, 3019 . 3021 [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", 3022 RFC 3972, DOI 10.17487/RFC3972, March 2005, 3023 . 3025 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 3026 More-Specific Routes", RFC 4191, DOI 10.17487/RFC4191, 3027 November 2005, . 3029 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 3030 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 3031 DOI 10.17487/RFC4861, September 2007, 3032 . 3034 [RFC4862] Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless 3035 Address Autoconfiguration", RFC 4862, 3036 DOI 10.17487/RFC4862, September 2007, 3037 . 3039 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 3040 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 3041 May 2017, . 3043 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 3044 (IPv6) Specification", STD 86, RFC 8200, 3045 DOI 10.17487/RFC8200, July 2017, 3046 . 3048 [RFC8415] Mrugalski, T., Siodelski, M., Volz, B., Yourtchenko, A., 3049 Richardson, M., Jiang, S., Lemon, T., and T. Winters, 3050 "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", 3051 RFC 8415, DOI 10.17487/RFC8415, November 2018, 3052 . 3054 8.2. Informative References 3056 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 3057 2016. 3059 [I-D.ietf-6man-segment-routing-header] 3060 Filsfils, C., Dukes, D., Previdi, S., Leddy, J., 3061 Matsushima, S., and D. Voyer, "IPv6 Segment Routing Header 3062 (SRH)", draft-ietf-6man-segment-routing-header-26 (work in 3063 progress), October 2019. 3065 [I-D.ietf-dmm-distributed-mobility-anchoring] 3066 Chan, A., Wei, X., Lee, J., Jeon, S., and C. Bernardos, 3067 "Distributed Mobility Anchoring", draft-ietf-dmm- 3068 distributed-mobility-anchoring-14 (work in progress), 3069 November 2019. 3071 [I-D.ietf-intarea-gue] 3072 Herbert, T., Yong, L., and O. Zia, "Generic UDP 3073 Encapsulation", draft-ietf-intarea-gue-09 (work in 3074 progress), October 2019. 3076 [I-D.ietf-intarea-gue-extensions] 3077 Herbert, T., Yong, L., and F. Templin, "Extensions for 3078 Generic UDP Encapsulation", draft-ietf-intarea-gue- 3079 extensions-06 (work in progress), March 2019. 3081 [I-D.ietf-intarea-tunnels] 3082 Touch, J. and M. Townsley, "IP Tunnels in the Internet 3083 Architecture", draft-ietf-intarea-tunnels-10 (work in 3084 progress), September 2019. 3086 [I-D.ietf-rtgwg-atn-bgp] 3087 Templin, F., Saccone, G., Dawra, G., Lindem, A., and V. 3088 Moreno, "A Simple BGP-based Mobile Routing System for the 3089 Aeronautical Telecommunications Network", draft-ietf- 3090 rtgwg-atn-bgp-05 (work in progress), January 2020. 3092 [I-D.templin-6man-dhcpv6-ndopt] 3093 Templin, F., "A Unified Stateful/Stateless Configuration 3094 Service for IPv6", draft-templin-6man-dhcpv6-ndopt-09 3095 (work in progress), January 2020. 3097 [I-D.templin-atn-aero-interface] 3098 Templin, F. and T. Whyman, "Transmission of IPv6 Packets 3099 over Overlay Multilink Network (OMNI) Interfaces", draft- 3100 templin-atn-aero-interface-12 (work in progress), January 3101 2020. 3103 [I-D.templin-intarea-grefrag] 3104 Templin, F., "GRE Tunnel Level Fragmentation", draft- 3105 templin-intarea-grefrag-04 (work in progress), July 2016. 3107 [I-D.templin-intarea-seal] 3108 Templin, F., "The Subnetwork Encapsulation and Adaptation 3109 Layer (SEAL)", draft-templin-intarea-seal-68 (work in 3110 progress), January 2014. 3112 [I-D.templin-intarea-vet] 3113 Templin, F., "Virtual Enterprise Traversal (VET)", draft- 3114 templin-intarea-vet-40 (work in progress), May 2013. 3116 [I-D.templin-ironbis] 3117 Templin, F., "The Interior Routing Overlay Network 3118 (IRON)", draft-templin-ironbis-16 (work in progress), 3119 March 2014. 3121 [I-D.templin-v6ops-pdhost] 3122 Templin, F., "IPv6 Prefix Delegation and Multi-Addressing 3123 Models", draft-templin-v6ops-pdhost-25 (work in progress), 3124 January 2020. 3126 [OVPN] OpenVPN, O., "http://openvpn.net", October 2016. 3128 [RFC1035] Mockapetris, P., "Domain names - implementation and 3129 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 3130 November 1987, . 3132 [RFC1122] Braden, R., Ed., "Requirements for Internet Hosts - 3133 Communication Layers", STD 3, RFC 1122, 3134 DOI 10.17487/RFC1122, October 1989, 3135 . 3137 [RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, 3138 DOI 10.17487/RFC1191, November 1990, 3139 . 3141 [RFC1812] Baker, F., Ed., "Requirements for IP Version 4 Routers", 3142 RFC 1812, DOI 10.17487/RFC1812, June 1995, 3143 . 3145 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 3146 DOI 10.17487/RFC2003, October 1996, 3147 . 3149 [RFC2236] Fenner, W., "Internet Group Management Protocol, Version 3150 2", RFC 2236, DOI 10.17487/RFC2236, November 1997, 3151 . 3153 [RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4 3154 Domains without Explicit Tunnels", RFC 2529, 3155 DOI 10.17487/RFC2529, March 1999, 3156 . 3158 [RFC2764] Gleeson, B., Lin, A., Heinanen, J., Armitage, G., and A. 3159 Malis, "A Framework for IP Based Virtual Private 3160 Networks", RFC 2764, DOI 10.17487/RFC2764, February 2000, 3161 . 3163 [RFC2784] Farinacci, D., Li, T., Hanks, S., Meyer, D., and P. 3164 Traina, "Generic Routing Encapsulation (GRE)", RFC 2784, 3165 DOI 10.17487/RFC2784, March 2000, 3166 . 3168 [RFC2890] Dommety, G., "Key and Sequence Number Extensions to GRE", 3169 RFC 2890, DOI 10.17487/RFC2890, September 2000, 3170 . 3172 [RFC2923] Lahey, K., "TCP Problems with Path MTU Discovery", 3173 RFC 2923, DOI 10.17487/RFC2923, September 2000, 3174 . 3176 [RFC2983] Black, D., "Differentiated Services and Tunnels", 3177 RFC 2983, DOI 10.17487/RFC2983, October 2000, 3178 . 3180 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 3181 of Explicit Congestion Notification (ECN) to IP", 3182 RFC 3168, DOI 10.17487/RFC3168, September 2001, 3183 . 3185 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 3186 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 3187 DOI 10.17487/RFC3810, June 2004, 3188 . 3190 [RFC3819] Karn, P., Ed., Bormann, C., Fairhurst, G., Grossman, D., 3191 Ludwig, R., Mahdavi, J., Montenegro, G., Touch, J., and L. 3192 Wood, "Advice for Internet Subnetwork Designers", BCP 89, 3193 RFC 3819, DOI 10.17487/RFC3819, July 2004, 3194 . 3196 [RFC4213] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms 3197 for IPv6 Hosts and Routers", RFC 4213, 3198 DOI 10.17487/RFC4213, October 2005, 3199 . 3201 [RFC4251] Ylonen, T. and C. Lonvick, Ed., "The Secure Shell (SSH) 3202 Protocol Architecture", RFC 4251, DOI 10.17487/RFC4251, 3203 January 2006, . 3205 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 3206 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 3207 DOI 10.17487/RFC4271, January 2006, 3208 . 3210 [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing 3211 Architecture", RFC 4291, DOI 10.17487/RFC4291, February 3212 2006, . 3214 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 3215 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 3216 December 2005, . 3218 [RFC4389] Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery 3219 Proxies (ND Proxy)", RFC 4389, DOI 10.17487/RFC4389, April 3220 2006, . 3222 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 3223 Control Message Protocol (ICMPv6) for the Internet 3224 Protocol Version 6 (IPv6) Specification", STD 89, 3225 RFC 4443, DOI 10.17487/RFC4443, March 2006, 3226 . 3228 [RFC4511] Sermersheim, J., Ed., "Lightweight Directory Access 3229 Protocol (LDAP): The Protocol", RFC 4511, 3230 DOI 10.17487/RFC4511, June 2006, 3231 . 3233 [RFC4541] Christensen, M., Kimball, K., and F. Solensky, 3234 "Considerations for Internet Group Management Protocol 3235 (IGMP) and Multicast Listener Discovery (MLD) Snooping 3236 Switches", RFC 4541, DOI 10.17487/RFC4541, May 2006, 3237 . 3239 [RFC4605] Fenner, B., He, H., Haberman, B., and H. Sandick, 3240 "Internet Group Management Protocol (IGMP) / Multicast 3241 Listener Discovery (MLD)-Based Multicast Forwarding 3242 ("IGMP/MLD Proxying")", RFC 4605, DOI 10.17487/RFC4605, 3243 August 2006, . 3245 [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for 3246 IP", RFC 4607, DOI 10.17487/RFC4607, August 2006, 3247 . 3249 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 3250 Errors at High Data Rates", RFC 4963, 3251 DOI 10.17487/RFC4963, July 2007, 3252 . 3254 [RFC4982] Bagnulo, M. and J. Arkko, "Support for Multiple Hash 3255 Algorithms in Cryptographically Generated Addresses 3256 (CGAs)", RFC 4982, DOI 10.17487/RFC4982, July 2007, 3257 . 3259 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 3260 "Bidirectional Protocol Independent Multicast (BIDIR- 3261 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 3262 . 3264 [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site 3265 Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, 3266 DOI 10.17487/RFC5214, March 2008, 3267 . 3269 [RFC5320] Templin, F., Ed., "The Subnetwork Encapsulation and 3270 Adaptation Layer (SEAL)", RFC 5320, DOI 10.17487/RFC5320, 3271 February 2010, . 3273 [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility 3274 Route Optimization Requirements for Operational Use in 3275 Aeronautics and Space Exploration Mobile Networks", 3276 RFC 5522, DOI 10.17487/RFC5522, October 2009, 3277 . 3279 [RFC5558] Templin, F., Ed., "Virtual Enterprise Traversal (VET)", 3280 RFC 5558, DOI 10.17487/RFC5558, February 2010, 3281 . 3283 [RFC5569] Despres, R., "IPv6 Rapid Deployment on IPv4 3284 Infrastructures (6rd)", RFC 5569, DOI 10.17487/RFC5569, 3285 January 2010, . 3287 [RFC5880] Katz, D. and D. Ward, "Bidirectional Forwarding Detection 3288 (BFD)", RFC 5880, DOI 10.17487/RFC5880, June 2010, 3289 . 3291 [RFC6106] Jeong, J., Park, S., Beloeil, L., and S. Madanapalli, 3292 "IPv6 Router Advertisement Options for DNS Configuration", 3293 RFC 6106, DOI 10.17487/RFC6106, November 2010, 3294 . 3296 [RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful 3297 NAT64: Network Address and Protocol Translation from IPv6 3298 Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146, 3299 April 2011, . 3301 [RFC6179] Templin, F., Ed., "The Internet Routing Overlay Network 3302 (IRON)", RFC 6179, DOI 10.17487/RFC6179, March 2011, 3303 . 3305 [RFC6221] Miles, D., Ed., Ooghe, S., Dec, W., Krishnan, S., and A. 3306 Kavanagh, "Lightweight DHCPv6 Relay Agent", RFC 6221, 3307 DOI 10.17487/RFC6221, May 2011, 3308 . 3310 [RFC6273] Kukec, A., Krishnan, S., and S. Jiang, "The Secure 3311 Neighbor Discovery (SEND) Hash Threat Analysis", RFC 6273, 3312 DOI 10.17487/RFC6273, June 2011, 3313 . 3315 [RFC6347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 3316 Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, 3317 January 2012, . 3319 [RFC6438] Carpenter, B. and S. Amante, "Using the IPv6 Flow Label 3320 for Equal Cost Multipath Routing and Link Aggregation in 3321 Tunnels", RFC 6438, DOI 10.17487/RFC6438, November 2011, 3322 . 3324 [RFC6706] Templin, F., Ed., "Asymmetric Extended Route Optimization 3325 (AERO)", RFC 6706, DOI 10.17487/RFC6706, August 2012, 3326 . 3328 [RFC6864] Touch, J., "Updated Specification of the IPv4 ID Field", 3329 RFC 6864, DOI 10.17487/RFC6864, February 2013, 3330 . 3332 [RFC7269] Chen, G., Cao, Z., Xie, C., and D. Binet, "NAT64 3333 Deployment Options and Experience", RFC 7269, 3334 DOI 10.17487/RFC7269, June 2014, 3335 . 3337 [RFC7333] Chan, H., Ed., Liu, D., Seite, P., Yokota, H., and J. 3338 Korhonen, "Requirements for Distributed Mobility 3339 Management", RFC 7333, DOI 10.17487/RFC7333, August 2014, 3340 . 3342 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 3343 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 3344 Multicast - Sparse Mode (PIM-SM): Protocol Specification 3345 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 3346 2016, . 3348 [RFC8086] Yong, L., Ed., Crabbe, E., Xu, X., and T. Herbert, "GRE- 3349 in-UDP Encapsulation", RFC 8086, DOI 10.17487/RFC8086, 3350 March 2017, . 3352 [RFC8201] McCann, J., Deering, S., Mogul, J., and R. Hinden, Ed., 3353 "Path MTU Discovery for IP version 6", STD 87, RFC 8201, 3354 DOI 10.17487/RFC8201, July 2017, 3355 . 3357 [RFC8402] Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L., 3358 Decraene, B., Litkowski, S., and R. Shakir, "Segment 3359 Routing Architecture", RFC 8402, DOI 10.17487/RFC8402, 3360 July 2018, . 3362 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 3363 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 3364 . 3366 Appendix A. AERO Alternate Encapsulations 3368 When GUE encapsulation is not needed, AERO can use common 3369 encapsulations such as IP-in-IP [RFC2003][RFC2473][RFC4213], Generic 3370 Routing Encapsulation (GRE) [RFC2784][RFC2890] and others. The 3371 encapsulation is therefore only differentiated from non-AERO tunnels 3372 through the application of AERO control messaging and not through, 3373 e.g., a well-known UDP port number. 3375 As for GUE encapsulation, alternate AERO encapsulation formats may 3376 require encapsulation layer fragmentation. For simple IP-in-IP 3377 encapsulation, an IPv6 fragment header is inserted directly between 3378 the inner and outer IP headers when needed, i.e., even if the outer 3379 header is IPv4. The IPv6 Fragment Header is identified to the outer 3380 IP layer by its IP protocol number, and the Next Header field in the 3381 IPv6 Fragment Header identifies the inner IP header version. For GRE 3382 encapsulation, a GRE fragment header is inserted within the GRE 3383 header [I-D.templin-intarea-grefrag]. 3385 Figure 6 shows the AERO IP-in-IP encapsulation format before any 3386 fragmentation is applied: 3388 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3389 | Outer IPv4 Header | | Outer IPv6 Header | 3390 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3391 |IPv6 Frag Header (optional)| |IPv6 Frag Header (optional)| 3392 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3393 | Inner IP Header | | Inner IP Header | 3394 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3395 | | | | 3396 ~ ~ ~ ~ 3397 ~ Inner Packet Body ~ ~ Inner Packet Body ~ 3398 ~ ~ ~ ~ 3399 | | | | 3400 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3402 Minimal Encapsulation in IPv4 Minimal Encapsulation in IPv6 3404 Figure 6: Minimal Encapsulation Format using IP-in-IP 3406 Figure 7 shows the AERO GRE encapsulation format before any 3407 fragmentation is applied: 3409 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3410 | Outer IP Header | 3411 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3412 | GRE Header | 3413 | (with checksum, key, etc..) | 3414 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3415 | GRE Fragment Header (optional)| 3416 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3417 | Inner IP Header | 3418 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3419 | | 3420 ~ ~ 3421 ~ Inner Packet Body ~ 3422 ~ ~ 3423 | | 3424 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3426 Figure 7: Minimal Encapsulation Using GRE 3428 Alternate encapsulation may be preferred in environments where GUE 3429 encapsulation would add unnecessary overhead. For example, certain 3430 low-bandwidth wireless data links may benefit from a reduced 3431 encapsulation overhead. 3433 GUE encapsulation can traverse network paths that are inaccessible to 3434 non-UDP encapsulations, e.g., for crossing Network Address 3435 Translators (NATs). More and more, network middleboxes are also 3436 being configured to discard packets that include anything other than 3437 a well-known IP protocol such as UDP and TCP. It may therefore be 3438 necessary to determine the potential for middlebox filtering before 3439 enabling alternate encapsulation in a given environment. 3441 In addition to IP-in-IP, GRE and GUE, AERO can also use security 3442 encapsulations such as IPsec, TLS/SSL, DTLS, etc. In that case, AERO 3443 control messaging and route determination occur before security 3444 encapsulation is applied for outgoing packets and after security 3445 decapsulation is applied for incoming packets. 3447 AERO is especially well suited for use with VPN system encapsulations 3448 such as OpenVPN [OVPN]. 3450 Appendix B. Non-Normative Considerations 3452 AERO can be applied to a multitude of Internetworking scenarios, with 3453 each having its own adaptations. The following considerations are 3454 provided as non-normative guidance: 3456 B.1. Implementation Strategies for Route Optimization 3458 Route optimization as discussed in Section 3.17 results in the route 3459 optimization source (ROS) creating an asymmetric neighbor cache entry 3460 for the target neighbor. The neighbor cache entry is maintained for 3461 at most REACHABLETIME seconds and then deleted unless updated. In 3462 order to refresh the neighbor cache entry lifetime before the 3463 ReachableTime timer expires, the specification requires 3464 implementations to issue a new NS/NA exchange to reset ReachableTime 3465 to REACHABLETIME seconds while data packets are still flowing. 3466 However, the decision of when to initiate a new NS/NA exchange and to 3467 perpetuate the process is left as an implementation detail. 3469 One possible strategy may be to monitor the neighbor cache entry 3470 watching for data packets for (REACHABLETIME - 5) seconds. If any 3471 data packets have been sent to the neighbor within this timeframe, 3472 then send an NS to receive a new NA. If no data packets have been 3473 sent, wait for 5 additional seconds and send an immediate NS if any 3474 data packets are sent within this "expiration pending" 5 second 3475 window. If no additional data packets are sent within the 5 second 3476 window, delete the neighbor cache entry. 3478 The monitoring of the neighbor data packet traffic therefore becomes 3479 an asymmetric ongoing process during the neighbor cache entry 3480 lifetime. If the neighbor cache entry expires, future data packets 3481 will trigger a new NS/NA exchange while the packets themselves are 3482 delivered over a longer path until route optimization state is re- 3483 established. 3485 B.2. Implicit Mobility Management 3487 AERO interface neighbors MAY provide a configuration option that 3488 allows them to perform implicit mobility management in which no ND 3489 messaging is used. In that case, the Client only transmits packets 3490 over a single interface at a time, and the neighbor always observes 3491 packets arriving from the Client from the same link-layer source 3492 address. 3494 If the Client's underlying interface address changes (either due to a 3495 readdressing of the original interface or switching to a new 3496 interface) the neighbor immediately updates the neighbor cache entry 3497 for the Client and begins accepting and sending packets according to 3498 the Client's new address. This implicit mobility method applies to 3499 use cases such as cellphones with both WiFi and Cellular interfaces 3500 where only one of the interfaces is active at a given time, and the 3501 Client automatically switches over to the backup interface if the 3502 primary interface fails. 3504 B.3. Direct Underlying Interfaces 3506 When a Client's AERO interface is configured over a Direct interface, 3507 the neighbor at the other end of the Direct link can receive packets 3508 without any encapsulation. In that case, the Client sends packets 3509 over the Direct link according to QoS preferences. If the Direct 3510 interface has the highest QoS preference, then the Client's IP 3511 packets are transmitted directly to the peer without going through an 3512 ANET/INET. If other interfaces have higher QoS preferences, then the 3513 Client's IP packets are transmitted via a different interface, which 3514 may result in the inclusion of Proxys, Servers and Relays in the 3515 communications path. Direct interfaces must be tested periodically 3516 for reachability, e.g., via NUD. 3518 B.4. AERO Clients on the Open Internetwork 3520 AERO Clients that connect to the open Internetwork via either a 3521 native or NATed interface can establish a VPN to securely connect to 3522 a Server. Alternatively, the Client can exchange ND messages 3523 directly with other AERO nodes on the same SPAN segment using INET 3524 encapsulation only and without joining the SPAN. In that case, 3525 however, the Client must apply asymmetric security for ND messages to 3526 ensure routing and neighbor cache integrity (see: Section 6). 3528 B.5. Operation on AERO Links with /64 ASPs 3530 IPv6 AERO links typically have MSPs that aggregate many candidate 3531 MNPs of length /64 or shorter. However, in some cases it may be 3532 desirable to use AERO over links that have only a /64 MSP. This can 3533 be accommodated by treating all Clients on the AERO link as simple 3534 hosts that receive /128 prefix delegations. 3536 In that case, the Client sends an RS message to the Server the same 3537 as for ordinary AERO links. The Server responds with an RA message 3538 that includes one or more /128 prefixes (i.e., singleton addresses) 3539 that include the /64 MSP prefix along with an interface identifier 3540 portion to be assigned to the Client. The Client and Server then 3541 configure their AERO addresses based on the interface identifier 3542 portions of the /128s (i.e., the lower 64 bits) and not based on the 3543 /64 prefix (i.e., the upper 64 bits). 3545 For example, if the MSP for the host-only IPv6 AERO link is 3546 2001:db8:1000:2000::/64, each Client will receive one or more /128 3547 IPv6 prefix delegations such as 2001:db8:1000:2000::1/128, 3548 2001:db8:1000:2000::2/128, etc. When the Client receives the prefix 3549 delegations, it assigns the AERO addresses fe80::1, fe80::2, etc. to 3550 the AERO interface, and assigns the global IPv6 addresses (i.e., the 3551 /128s) to either the AERO interface or an internal virtual interface 3552 such as a loopback. In this arrangement, the Client conducts route 3553 optimization in the same sense as discussed in Section 3.17. 3555 This specification has applicability for nodes that act as a Client 3556 on an "upstream" AERO link, but also act as a Server on "downstream" 3557 AERO links. More specifically, if the node acts as a Client to 3558 receive a /64 prefix from the upstream AERO link it can then act as a 3559 Server to provision /128s to Clients on downstream AERO links. 3561 B.6. AERO Adaptations for SEcure Neighbor Discovery (SEND) 3563 SEcure Neighbor Discovery (SEND) [RFC3971] and Cryptographically 3564 Generated Addresses (CGAs) [RFC3972] were designed to secure IPv6 ND 3565 messaging in environments where symmetric network and/or transport- 3566 layer security services are impractical (see: Section 6). AERO nodes 3567 that use SEND/CGA employ the following adaptations. 3569 When a source AERO node prepares a SEND-protected ND message, it uses 3570 a link-local CGA as the IPv6 source address and writes the prefix 3571 embedded in its AERO address (i.e., instead of fe80::/64) in the CGA 3572 parameters Subnet Prefix field. When the neighbor receives the ND 3573 message, it first verifies the message checksum and SEND/CGA 3574 parameters while using the link-local prefix fe80::/64 (i.e., instead 3575 of the value in the Subnet Prefix field) to match against the IPv6 3576 source address of the ND message. 3578 The neighbor then derives the AERO address of the source by using the 3579 value in the Subnet Prefix field as the interface identifier of an 3580 AERO address. For example, if the Subnet Prefix field contains 3581 2001:db8:1:2, the neighbor constructs the AERO address as 3582 fe80::2001:db8:1:2. The neighbor then caches the AERO address in the 3583 neighbor cache entry it creates for the source, and uses the AERO 3584 address as the IPv6 destination address of any ND message replies. 3586 B.7. AERO Critical Infrastructure Considerations 3588 AERO Relays can be either Commercial off-the Shelf (COTS) standard IP 3589 routers or virtual machines in the cloud. Relays must be 3590 provisioned, supported and managed by the INET administrative 3591 authority, and connected to the Relays of other INETs via inter- 3592 domain peerings. Cost for purchasing, configuring and managing 3593 Relays is nominal even for very large AERO links. 3595 AERO Servers can be standard dedicated server platforms, but most 3596 often will be deployed as virtual machines in the cloud. The only 3597 requirements for Servers are that they can run the AERO user-level 3598 code and have at least one network interface connection to the INET. 3599 As with Relays, Servers must be provisioned, supported and managed by 3600 the INET administrative authority. Cost for purchasing, configuring 3601 and managing Servers is nominal especially for virtual Servers hosted 3602 in the cloud. 3604 AERO Proxys are most often standard dedicated server platforms with 3605 one network interface connected to the ANET and a second interface 3606 connected to an INET. As with Servers, the only requirements are 3607 that they can run the AERO user-level code and have at least one 3608 interface connection to the INET. Proxys must be provisioned, 3609 supported and managed by the ANET administrative authority. Cost for 3610 purchasing, configuring and managing Proxys is nominal, and borne by 3611 the ANET administrative authority. 3613 AERO Gateways can be any dedicated server or COTS router platform 3614 connected to INETs and/or EUNs. The Gateway joins the SPAN and 3615 engages in eBGP peering with one or more Relays as a stub AS. The 3616 Gateway then injects its MNPs and/or non-MNP prefixes into the BGP 3617 routing system, and provisions the prefixes to its downstream- 3618 attached networks. The Gateway can perform ROS and MAP services the 3619 same as for any Server, and can route between the MNP and non-MNP 3620 address spaces. 3622 B.8. AERO Server Failure Implications 3624 AERO Servers may appear as a single point of failure in the 3625 architecture, but such is not the case since all Servers on the link 3626 provide identical services and loss of a Server does not imply 3627 immediate and/or comprehensive communication failures. Although 3628 Clients typically associate with a single Server at a time, Server 3629 failure is quickly detected and conveyed by Bidirectional Forward 3630 Detection (BFD) and/or proactive NUD allowing Clients to migrate to 3631 new Servers. 3633 If a Server fails, ongoing packet forwarding to Clients will continue 3634 by virtue of the asymmetric neighbor cache entries that have already 3635 been established in route optimization sources (ROSs). If a Client 3636 also experiences mobility events at roughly the same time the Server 3637 fails, unsolicited NA messages may be lost but proxy neighbor cache 3638 entries in the DEPARTED state will ensure that packet forwarding to 3639 the Client's new locations will continue for up to DEPARTTIME 3640 seconds. 3642 If a Client is left without a Server for an extended timeframe (e.g., 3643 greater than REACHABLETIIME seconds) then existing asymmetric 3644 neighbor cache entries will eventually expire and both ongoing and 3645 new communications will fail. The original source will continue to 3646 retransmit until the Client has established a new Server 3647 relationship, after which time continuous communications will resume. 3649 Therefore, providing many Servers on the link with high availability 3650 profiles provides resilience against loss of individual Servers and 3651 assurance that Clients can establish new Server relationships quickly 3652 in event of a Server failure. 3654 B.9. AERO Client / Server Architecture 3656 The AERO architectural model is client / server in the control plane, 3657 with route optimization in the data plane. The same as for common 3658 Internet services, the AERO Client discovers the addresses of AERO 3659 Servers and selects one Server to connect to. The AERO service is 3660 analogous to common Internet services such as google.com, yahoo.com, 3661 cnn.com, etc. However, there is only one AERO service for the link 3662 and all Servers provide identical services. 3664 Common Internet services provide differing strategies for advertising 3665 server addresses to clients. The strategy is conveyed through the 3666 DNS resource records returned in response to name resolution queries. 3667 As of January 2020 Internet-based 'nslookup' services were used to 3668 determine the following: 3670 o When a client resolves the domainname "google.com", the DNS always 3671 returns one A record (i.e., an IPv4 address) and one AAAA record 3672 (i.e., an IPv6 address). The client receives the same addresses 3673 each time it resolves the domainname via the same DNS resolver, 3674 but may receive different addresses when it resolves the 3675 domainname via different DNS resolvers. But, in each case, 3676 exactly one A and one AAAA record are returned. 3678 o When a client resolves the domainname "ietf.org", the DNS always 3679 returns one A record and one AAAA record with the same addresses 3680 regardless of which DNS resolver is used. 3682 o When a client resolves the domainname "yahoo.com", the DNS always 3683 returns a list of 4 A records and 4 AAAA records. Each time the 3684 client resolves the domainname via the same DNS resolver, the same 3685 list of addresses are returned but in randomized order (i.e., 3686 consistent with a DNS round-robin strategy). But, interestingly, 3687 the same addresses are returned (albeit in randomized order) when 3688 the domainname is resolved via different DNS resolvers. 3690 o When a client resolves the domainname "amazon.com", the DNS always 3691 returns a list of 3 A records and no AAAA records. As with 3692 "yahoo.com", the same three A records are returned from any 3693 worldwide Internet connection point in randomized order. 3695 The above example strategies show differing approaches to Internet 3696 resilience and service distribution offered by major Internet 3697 services. The Google approach exposes only a single IPv4 and a 3698 single IPv6 address to clients. Clients can then select whichever IP 3699 protocol version offers the best response, but will always use the 3700 same IP address according to the current Internet connection point. 3701 This means that the IP address offered by the network must lead to a 3702 highly-available server and/or service distribution point. In other 3703 words, resilience is predicated on high availability within the 3704 network and with no client-initiated failovers expected (i.e., it is 3705 all-or-nothing from the client's perspective). However, Google does 3706 provide for worldwide distributed service distribution by virtue of 3707 the fact that each Internet connection point responds with a 3708 different IPv6 and IPv4 address. The IETF approach is like google 3709 (all-or-nothing from the client's perspective), but provides only a 3710 single IPv4 or IPv6 address on a worldwide basis. This means that 3711 the addresses must be made highly-available at the network level with 3712 no client failover possibility, and if there is any worldwide service 3713 distribution it would need to be conducted by a network element that 3714 is reached via the IP address acting as a service distribution point. 3716 In contrast to the Google and IETF philosophies, Yahoo and Amazon 3717 both provide clients with a (short) list of IP addresses with Yahoo 3718 providing both IP protocol versions and Amazon as IPv4-only. The 3719 order of the list is randomized with each name service query 3720 response, with the effect of round-robin load balancing for service 3721 distribution. With a short list of addresses, there is still 3722 expectation that the network will implement high availability for 3723 each address but in case any single address fails the client can 3724 switch over to using a different address. The balance then becomes 3725 one of function in the network vs function in the end system. 3727 The same implications observed for common highly-available services 3728 in the Internet apply also to the AERO client/server architecture. 3729 When an AERO Client connects to one or more ANETs, it discovers one 3730 or more AERO Server addresses through the mechanisms discussed in 3731 earlier sections. Each Server address presumably leads to a fault- 3732 tolerant clustering arrangement such as supported by Linux-HA, 3733 Extended Virtual Synchrony or Paxos. Such an arrangement has 3734 precedence in common Internet service deployments in lightweight 3735 virtual machines without requiring expensive hardware deployment. 3736 Similarly, common Internet service deployments set service IP 3737 addresses on service distribution points that may relay requests to 3738 many different servers. 3740 For AERO, the expectation is that a combination of the Google/IETF 3741 and Yahoo/Amazon philosophies would be employed. The AERO Client 3742 connects to different ANET access points and can receive 1-2 Server 3743 AERO addresses at each point. It then selects one AERO Server 3744 address, and engages in RS/RA exchanges with the same Server from all 3745 ANET connections. The Client remains with this Server unless or 3746 until the Server fails, in which case it can switch over to an 3747 alternate Server. The Client can likewise switch over to a different 3748 Server at any time if there is some reason for it to do so. So, the 3749 AERO expectation is for a balance of function in the network and end 3750 system, with fault tolerance and resilience at both levels. 3752 Appendix C. Change Log 3754 << RFC Editor - remove prior to publication >> 3756 Changes from draft-templin-intarea-6706bis-19 to draft-templin- 3757 intrea-6706bis-20: 3759 o Included new route optimization source and destination addressing 3760 strategy. Now, route optimization maintenance uses the address of 3761 the existing Server instead of the data packet destination address 3762 so that less pressure is placed on the BGP routing system 3763 convergence time and Server constancy is supported. 3765 o Included new method for releasing from old MSE without requiring 3766 Client messaging. 3768 o Included references to new OMNI interface spec (including the OMNI 3769 option). 3771 o New appendix on AERO Client/Server architecture. 3773 Changes from draft-templin-intarea-6706bis-18 to draft-templin- 3774 intrea-6706bis-19: 3776 o Changed Proxy/Server keepalives to use "proactive NUD" in a manner 3777 tha paralles BFD 3779 Changes from draft-templin-intarea-6706bis-17 to draft-templin- 3780 intrea-6706bis-18: 3782 o Discuss how AERO option is used in relation to S/TLLAOs 3784 o New text on Bidirectional Forwarding Detection (BFD) 3786 o Cleaned up usage (and non-usage) of unsolicited NAs 3788 o New appendix on Server failures 3790 Changes from draft-templin-intarea-6706bis-15 to draft-templin- 3791 intrea-6706bis-17: 3793 o S/TLLAO now includes multiple link-layer addresses within a single 3794 option instead of requiring multiple options 3796 o New unsolicited NA message to inform the old link that a Client 3797 has moved to a new link 3799 Changes from draft-templin-intarea-6706bis-14 to draft-templin- 3800 intrea-6706bis-15: 3802 o MTU and fragmentation 3804 o New details in movement to new Server 3806 Changes from draft-templin-intarea-6706bis-13 to draft-templin- 3807 intrea-6706bis-14: 3809 o Security based on secured tunnels, ingress filtering, MAP list and 3810 ROS list 3812 Changes from draft-templin-intarea-6706bis-12 to draft-templin- 3813 intrea-6706bis-13: 3815 o New paragraph in Section 3.6 on AERO interface layering over 3816 secured tunnels 3818 o Removed extraneous text in Section 3.7 3820 o Added new detail to the forwarding algorithm in Section 3.9 3822 o Clarified use of fragmentation 3824 o Route optimization now supported for both MNP and non-MNP-based 3825 prefixes 3827 o Relays are now seen as link-layer elements in the architecture. 3829 o Built out multicast section in detail. 3831 o New Appendix on implementation considerations for route 3832 optimization. 3834 Changes from draft-templin-intarea-6706bis-11 to draft-templin- 3835 intrea-6706bis-12: 3837 o Introduced Gateways as a new AERO element for connecting 3838 Correspondent Nodes on INET links 3840 o Introduced terms "Access Network (ANET)" and "Internetwork (INET)" 3842 o Changed "ASP" to "MSP", and "ACP" to "MNP" 3844 o New figure on the relation of Segments to the SPAN and AERO link 3846 o New "S" bit in S/TLLAO to indicate the "Source" S/TLLAO as opposed 3847 to additional S/TLLAOs 3849 o Changed Interface ID for Servers from 255 to 0xffff 3851 o Significant updates to Route Optimization, NUD, and Mobility 3852 Management 3854 o New Section on Multicast 3856 o New Section on AERO Clients in the open Internetwork 3858 o New Section on Operation over multiple AERO links (VLANs over the 3859 SPAN) 3861 o New Sections on DNS considerations and Transition considerations 3863 o 3865 Changes from draft-templin-intarea-6706bis-10 to draft-templin- 3866 intrea-6706bis-11: 3868 o Added The SPAN 3870 Changes from draft-templin-intarea-6706bis-09 to draft-templin- 3871 intrea-6706bis-10: 3873 o Orphaned packets in flight (e.g., when a neighbor cache entry is 3874 in the DEPARTED state) are now forwarded at the link layer instead 3875 of at the network layer. Forwarding at the network layer can 3876 result in routing loops and/or excessive delays of forwarded 3877 packets while the routing system is still reconverging. 3879 o Update route optimization to clarify the unsecured nature of the 3880 first NS used for route discovery 3882 o Many cleanups and clarifications on ND messaging parameters 3884 Changes from draft-templin-intarea-6706bis-08 to draft-templin- 3885 intrea-6706bis-09: 3887 o Changed PRL to "MAP list" 3889 o For neighbor cache entries, changed "static" to "symmetric", and 3890 "dynamic" to "asymmetric" 3892 o Specified Proxy RS/RA exchanges with Servers on behalf of Clients 3894 o Added discussion of unsolicited NAs in Section 3.16, and included 3895 forward reference to Section 3.18 3897 o Added discussion of AERO Clients used as critical infrastructure 3898 elements to connect fixed networks. 3900 o Added network-based VPN under security considerations 3902 Changes from draft-templin-intarea-6706bis-07 to draft-templin- 3903 intrea-6706bis-08: 3905 o New section on AERO-Aware Access Router 3907 Changes from draft-templin-intarea-6706bis-06 to draft-templin- 3908 intrea-6706bis-07: 3910 o Added "R" bit for release of PDs. Now have a full RS/RA service 3911 that can do PD without requiring DHCPv6 messaging over-the-air 3913 o Clarifications on solicited vs unsolicited NAs 3915 o Clarified use of MAX_NEIGHBOR_ADVERTISEMENTS for the purpose of 3916 increase reliability 3918 Changes from draft-templin-intarea-6706bis-05 to draft-templin- 3919 intrea-6706bis-06: 3921 o Major re-work and simplification of Route Optimization function 3923 o Added Distributed Mobility Management (DMM) and Mobility Anchor 3924 Point (MAP) terminology 3926 o New section on "AERO Critical Infrastructure Element 3927 Considerations" demonstrating low overall cost for the service 3929 o minor text revisions and deletions 3931 o removed extraneous appendices 3933 Changes from draft-templin-intarea-6706bis-04 to draft-templin- 3934 intrea-6706bis-05: 3936 o New Appendix E on S/TLLAO Extensions for special-purpose links. 3937 Discussed ATN/IPS as example. 3939 o New sentence in introduction to declare appendices as non- 3940 normative. 3942 Changes from draft-templin-intarea-6706bis-03 to draft-templin- 3943 intrea-6706bis-04: 3945 o Added definitions for Potential Router List (PRL) and secure 3946 enclave 3948 o Included text on mapping transport layer port numbers to network 3949 layer DSCP values 3951 o Added reference to DTLS and DMM Distributed Mobility Anchoring 3952 working group document 3954 o Reworked Security Considerations 3956 o Updated references. 3958 Changes from draft-templin-intarea-6706bis-02 to draft-templin- 3959 intrea-6706bis-03: 3961 o Added new section on SEND. 3963 o Clarifications on "AERO Address" section. 3965 o Updated references and added new reference for RFC8086. 3967 o Security considerations updates. 3969 o General text clarifications and cleanup. 3971 Changes from draft-templin-intarea-6706bis-01 to draft-templin- 3972 intrea-6706bis-02: 3974 o Note on encapsulation avoidance in Section 4. 3976 Changes from draft-templin-intarea-6706bis-00 to draft-templin- 3977 intrea-6706bis-01: 3979 o Remove DHCPv6 Server Release procedures that leveraged the old way 3980 Relays used to "route" between Server link-local addresses 3982 o Remove all text relating to Relays needing to do any AERO-specific 3983 operations 3985 o Proxy sends RS and receives RA from Server using SEND. Use CGAs 3986 as source addresses, and destination address of RA reply is to the 3987 AERO address corresponding to the Client's ACP. 3989 o Proxy uses SEND to protect RS and authenticate RA (Client does not 3990 use SEND, but rather relies on subnetwork security. When the 3991 Proxy receives an RS from the Client, it creates a new RS using 3992 its own addresses as the source and uses SEND with CGAs to send a 3993 new RS to the Server. 3995 o Emphasize distributed mobility management 3997 o AERO address-based RS injection of ACP into underlying routing 3998 system. 4000 Changes from draft-templin-aerolink-82 to draft-templin-intarea- 4001 6706bis-00: 4003 o Document use of NUD (NS/NA) for reliable link-layer address 4004 updates as an alternative to unreliable unsolicited NA. 4005 Consistent with Section 7.2.6 of RFC4861. 4007 o Server adds additional layer of encapsulation between outer and 4008 inner headers of NS/NA messages for transmission through Relays 4009 that act as vanilla IPv6 routers. The messages include the AERO 4010 Server Subnet Router Anycast address as the source and the Subnet 4011 Router Anycast address corresponding to the Client's ACP as the 4012 destination. 4014 o Clients use Subnet Router Anycast address as the encapsulation 4015 source address when the access network does not provide a 4016 topologically-fixed address. 4018 Author's Address 4020 Fred L. Templin (editor) 4021 Boeing Research & Technology 4022 P.O. Box 3707 4023 Seattle, WA 98124 4024 USA 4026 Email: fltemplin@acm.org