idnits 2.17.1 draft-templin-atn-bgp-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (May 12, 2017) is 2542 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'RFC0791' is defined on line 517, but no explicit reference was found in the text == Unused Reference: 'RFC2460' is defined on line 521, but no explicit reference was found in the text == Unused Reference: 'RFC4861' is defined on line 545, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) == Outdated reference: A later version (-38) exists of draft-ietf-lisp-rfc6830bis-03 == Outdated reference: A later version (-31) exists of draft-ietf-lisp-rfc6833bis-05 == Outdated reference: A later version (-82) exists of draft-templin-aerolink-74 -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) Summary: 1 error (**), 0 flaws (~~), 7 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group F. Templin, Ed. 3 Internet-Draft Boeing Research & Technology 4 Intended status: Informational May 12, 2017 5 Expires: November 13, 2017 7 A Simple BGP-based Mobile Routing System for the Aeronautical 8 Telecommunications Network 9 draft-templin-atn-bgp-03.txt 11 Abstract 13 The International Civil Aviation Organization (ICAO) is investigating 14 mobile routing solutions for a worldwide Aeronautical 15 Telecommunications Network with Internet Protocol Services (ATN/IPS). 16 The ATN/IPS will eventually replace existing communication services 17 with an IPv6-based service supporting pervasive Air Traffic 18 Management (ATM) for Air Traffic Controllers (ATC), Airline 19 Operations Controllers (AOC), and all commercial aircraft worldwide. 20 This informational document describes a simple mobile routing service 21 based on mature industry standards to address the ATN/IPS 22 requirements. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on November 13, 2017. 41 Copyright Notice 43 Copyright (c) 2017 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 59 2. Proposed BGP-based ATN/IPS Routing System . . . . . . . . . . 4 60 3. Route Optimization . . . . . . . . . . . . . . . . . . . . . 8 61 4. Route Availability . . . . . . . . . . . . . . . . . . . . . 10 62 5. BGP Protocol Considerations . . . . . . . . . . . . . . . . . 11 63 6. Implementation Status . . . . . . . . . . . . . . . . . . . . 12 64 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 65 8. Security Considerations . . . . . . . . . . . . . . . . . . . 12 66 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 12 67 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 68 10.1. Normative References . . . . . . . . . . . . . . . . . . 13 69 10.2. Informative References . . . . . . . . . . . . . . . . . 13 70 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 15 72 1. Introduction 74 The International Civil Aviation Organization [ICAO] is investigating 75 mobile routing solutions for a worldwide Aeronautical 76 Telecommunications Network with Internet Protocol Services (ATN/IPS). 77 The ATN/IPS will eventually replace existing communication services 78 with an IPv6-based service supporting pervasive Air Traffic 79 Management (ATM) for Air Traffic Controllers (ATC), Airline 80 Operations Controllers (AOC), and all commercial aircraft worldwide. 81 This informational document describes a simple mobile routing service 82 based on mature industry standards to address the ATN/IPS 83 requirements. 85 Aircraft communicate via wireless aviation data links that typically 86 support much lower data rates than terrestrial wireless and wired- 87 line communications. For example, VHF-based data links only support 88 data rates on the order of 32Kbps and an emerging L-Band data link 89 that is expected to play a key role in future aeronautical 90 communications only supports rates on the order of 1Mbps. Although 91 satellite data links can provide much higher data rates during 92 optimal conditions, they (like all other aviation data links) are 93 subject to errors, delay, disruption, signal intermittence, 94 degradation due to atmospheric conditions, etc. The well-connected 95 ground domain ATN/IPS network should therefore treat each safety-of- 96 flight critical packet produced by (or destined to) an aircraft as a 97 precious commodity and strive for a "better-than-best-effort" service 98 that provides the highest possible degree of reliability. 100 The ATN/IPS assumes a worldwide connected Internetwork for carrying 101 ATM communications. The Internetwork could be manifested as a 102 private collection of long-haul backbone links (e.g., fiberoptics, 103 copper, SATCOM, etc.) interconnected by high-performance networking 104 gear such as bridges, switches and routers. Such a private 105 Internetwork would need to connect all ATN/IPS participants 106 worldwide, and could therefore present a considerable cost for a 107 large-scale deployment of new infrastructure. Alternatively, the 108 ATN/IPS could be deployed as an overlay over the existing global 109 public Internet itself as long as sufficient security and reliability 110 provisions are met. For example, ATN/IPS nodes could be deployed as 111 part of an SD-WAN or an MPLS-WAN that rides over the public Internet 112 via secured tunnels. 114 The ATN/IPS further assumes that each aircraft will receive an IPv6 115 Mobile Network Prefix (MNP) that accompanies the aircraft wherever it 116 travels. ATCs and AOCs will likewise receive IPv6 prefixes, but they 117 would typically appear in static (not mobile) deployments. 118 Throughout the rest of this document, we therefore use the term "MNP" 119 when discussing an IPv6 prefix that is delegated to any ATN/IPS end 120 system, including ATCs, AOCs and aircraft. We also use the term 121 Mobility Service Prefix (MSP) to refer to an aggregated prefix 122 assigned to the ATN/IPS by an Internet assigned numbers authority, 123 and from which all MNPs are delegated (e.g., up to 2**32 IPv6 /64 124 MNPs could be delegated from the MSP 2001:db8::/32). 126 [CBB] describes an aviation mobile routing service based on dynamic 127 updates in the global public Internet Border Gateway Protocol (BGP) 128 [RFC4271] routing system. Practical experience with the approach has 129 shown that frequent injections and withdrawals of MNPs in the 130 Internet routing system results in excessive BGP update messaging, 131 slow routing table convergence times, and extended outages when no 132 route is available. This is due to both conservative default BGP 133 protocol timing parameters (see Section 5) and the complex peering 134 interconnections of BGP routers within the global Internet 135 infrastructure. The situation is further exacerbated by frequent 136 aircraft mobility events that each result in BGP updates that must be 137 propagated to all BGP routers in the Internet that carry a full 138 routing table. 140 We therefore consider an approach using a BGP overlay network routing 141 system where a private BGP routing protocol instance is maintained 142 between ATN/IPS Autonomous System (AS) Border Routers (ASBRs). The 143 private BGP instance does not interact with the Internetwork BGP 144 routing system, and BGP updates are unidirectional from "stub" ASBRs 145 (s-ASBRs) to a very small set of "core" ASBRs (c-ASBRs) in a hub-and- 146 spokes arrangement. For the AERO proposal [I-D.templin-aerolink], 147 the s-ASBRs correspond to AERO Servers. For the LISP proposal 148 [I-D.ietf-lisp-rfc6830bis], the s-ASBRs correspond to xTRs that 149 connect directly to the BGP system instead of via Map Servers and 150 Resolvers. No non-standard extensions of the BGP protocol are 151 necessary. 153 The s-ASBRs for each stub AS connect to a small number of c-ASBRs via 154 dedicated high speed links and/or tunnels across the Internetwork 155 using industry-standard encapsulations (e.g., Generic Routing 156 Encapsulation (GRE) [RFC2784], IPsec [RFC4301] etc.). The s-ASBRs 157 engage in external BGP (eBGP) peerings with their respective c-ASBRs, 158 and only maintain routing table entries for the MNPs currently active 159 within the stub AS. A stub AS may connect to the core via multiple 160 s-ASBRs, in which case the s-ASBRs would engage in an Interior 161 Gateway Protocol (IGP) among themselves to maintain a common view of 162 the stub AS MNPs. (The s-ASBRs need not engage in internal BGP 163 (iBGP) peerings, since they do not receive any BGP updates from 164 c-ASBRs and therefore have no BGP information to share with each 165 other.) Finally, the s-ASBRs also maintain default routes with their 166 c-ASBRs as the next hop, and therefore hold only partial topology 167 information. 169 The c-ASBRs connect to other c-ASBRs using iBGP peerings over which 170 they collaboratively maintain a full routing table for all active 171 MNPs currently in service. Therefore, only the c-ASBRs maintain a 172 full BGP routing table and never send any BGP updates to s-ASBRs. 173 This simple arrangement therefore greatly reduces the number of BGP 174 updates that need to be synchronized among peers, and the number is 175 reduced further still when localized mobility events within stub ASes 176 (i.e., "intradomain" mobility events) are mitigated within the AS 177 instead of being propagated to the core. 179 The following section provides a detailed discussion of the proposed 180 BGP-based ATN/IPS routing system. 182 2. Proposed BGP-based ATN/IPS Routing System 184 The proposed ATN/IPS routing system comprises a private BGP instance 185 coordinated between ASBRs in an overlay network. The overlay does 186 not interact with the native Internetwork BGP routing system, and 187 each c-ASBR advertises only a small and unchanging set of MSPs into 188 the Internetwork instead of the full dynamically changing set of 189 MNPs. The system corresponds to the framework first specified by the 190 LISP+ALT proposal [RFC6836] and later also adopted by the AERO 191 proposal [I-D.templin-aerolink]. The system differs from the LISP 192 Delegated Database Tree (DDT) proposal [I-D.ietf-lisp-ddt] that is 193 designed with scalability as the primary consideration. 195 In a reference deployment, one or more s-ASBRs connect each stub AS 196 to the overlay using a shared stub AS Number (ASN). Each s-ASBR 197 further uses eBGP to peer with one or more c-ASBRs. All c-ASBRs are 198 members of the same core AS, and use a shared core ASN. The c-ASBRs 199 further use iBGP to maintain a synchronized consistent view of all 200 active MNPs currently in service. Figure 1 below represents the 201 reference deployment. Note that in the figure only two s-ASBRs show 202 detail, but similar arrangements are implied for all other s-ASBRs. 203 Note also that each stub AS shows only a single s-ASBR with a single 204 c-ASBR connection, but in practical deployments each stub AS may have 205 multiple s-ASBRs that peer with multiple c-ASBRs via eBGP, e.g., for 206 fault tolerance. 208 ........................................................... 209 . . 210 . (:::)-. <- Stub ASes -> (:::)-. . 211 . MNPs-> .-(:::::::::) .-(:::::::::) <-MNPs . 212 . `-(::::)-' `-(::::)-' . 213 . +-------+ +-------+ . 214 . |s-ASBR1| |s-ASBR2| . 215 . +--+----+ +-----+-+ . 216 . \ / . 217 . \eBGP /eBGP . 218 . \ / . 219 . +-------+ +-------+ . 220 . eBGP+-----+c-ASBR1| +c-ASBR2+-----+eBGP . 221 . +-------+ / +--+----+ +-----+-+ \ +-------+ . 222 . |s-ASBRn+/ iBGP\ (:::)-. /iBGP \+s-ASBR3| . 223 . +-------+ .-(::::::::) +-------+ . 224 . . .-(::::::::::::::)-. . 225 . . (:::: Core AS :::) . 226 . +-------+ `-(:::::::::::::)-' +-------+ . 227 . |s-ASBR7+\ iBGP/`-(:::::::-'\iBGP /+s-ASBR4| . 228 . +-------+ \ +-+-----+ +----+--+ / +-------+ . 229 . eBGP+-----+c-ASBRn| |c-ASBR3+-----+eBGP . 230 . +-------+ +-------+ . 231 . / \ . 232 . /eBGP \eBGP . 233 . / \ . 234 . +---+---+ +-----+-+ . 235 . |s-ASBR6| |s-ASBR5| . 236 . +-------+ +-------+ . 237 . . 238 . . 239 . <------------------- Internetwork --------------------> . 240 ............................................................ 242 Figure 1: Reference Deployment 244 In the reference deployment, each s-ASBR maintains routes for active 245 MNPs that currently belong to its stub AS, and dynamically announces 246 new MNPs and withdraws departed MNPs in its eBGP updates to c-ASBRs 247 in response to "interdomain" mobility events. Since ATN/IPS end 248 systems are expected to remain within the same stub AS for extended 249 timeframes, however, intradomain mobility events (such as an aircraft 250 handing off between cell towers) are handled locally within the stub 251 AS instead of being propagated as interdomain eBGP updates. 253 Each c-ASBR configures a black-hole route for each of its MSPs. By 254 black-holing the MSPs, the c-ASBR will maintain forwarding table 255 entries only for the MNPs that are currently active, and packets 256 destined to all other MNPs will correctly incur ICMPv6 Destination 257 Unreachable messages [RFC4443] due to the black hole route. The 258 c-ASBRs do not send eBGP updates for MNPs to s-ASBRs, but instead 259 originate a default route. In this way, s-ASBRs have only partial 260 topology knowledge (i.e., they know only about the active MNPs 261 currently within their stub ASes) and they forward all other packets 262 to c-ASBRs which have full topology knowledge. 264 Scaling properties of this ATN/IPS routing system are limited by the 265 number of BGP routes that can be carried by the c-ASBRs. A 2015 266 study showed that BGP routers in the global public Internet at that 267 time carried more than 500K routes with linear growth and no signs of 268 router resource exhaustion [BGP]. A more recent network emulation 269 study also showed that a single c-ASBR can accommodate at least 1M 270 dynamically changing BGP routes even on a lightweight virtual 271 machine, with the expectation that high-performance dedicated router 272 hardware can support even more. 274 Therefore, assuming each c-ASBR can carry 1M or more routes, this 275 means that at least 1M ATN/IPS end system MNPs can be serviced by a 276 single set of c-ASBRs. A means of increasing scaling would be to 277 assign a different set of c-ASBRs for each set of MSPs. In that 278 case, each s-ASBR still peers with one or more c-ASBRs from each set 279 of c-ASBRs, but the s-ASBR institutes route filters so that it only 280 sends BGP updates to the specific set of c-ASBRs that aggregate the 281 MSP. For example, if the MSP for the ATN/IPS deployment is 282 2001:db8::/32, a first set of c-ASBRs could service the MSP segment 283 2001:db8::/40, a second set could service 2001:db8:0100::/40, a third 284 set could service 2001:db8:0200::/40, etc. 286 Assuming a sufficient number of c-ASBR sets, the ATN/IPS routing 287 system can then accommodate 1B or more MNPs. In this way, each set 288 of c-ASBRs services a specific set of MSPs that they advertise to the 289 native Internetwork routing system, and each s-ASBR configures MSP- 290 specific routes that list the correct set of c-ASBRs as next hops. 291 This arrangement also allows for natural incremental deployment, and 292 can support small scale initial deployments followed by dynamic 293 deployment of additional ATN/IPS infrastructure elements without 294 disturbing the already-deployed base. 296 Finally, c-ASBRs may have multiple routing table entries for a single 297 MNP advertised by multiple s-ASBRs. Each s-ASBR can be assigned a 298 MULTI_EXIT_DISC (MED) metric for routes that it originates in its 299 eBGP peering configurations [RFC4451] so that c-ASBRs can determine 300 preferences for MNPs learned from multiple s-ASBRs. In this way, 301 c-ASBRs can select the neighboring s-ASBR with the lowest MED value, 302 i.e., even if it is not on the shortest path. The c-ASBR can then 303 fail over to a s-ASBR with a larger MED value in case of MNP 304 withdrawal or s-ASBR failure. Such an event could correspond to an 305 aviation data link handover, e.g., when an aircraft switches over 306 from a satellite link to an L-Band link. 308 3. Route Optimization 310 ATN/IPS end systems will frequently need to communicate with 311 correspondents located in other stub ASes. In the ASBR peering 312 arrangement discussed in Section 2, this can initially only be 313 accommodated by having the source s-ASBR forward packets to a c-ASBR 314 which then forwards the packets toward the destination s-ASBR where 315 the destination ATN/IPS end system resides. In many cases, it would 316 be desirable to eliminate c-ASBRs from this "dogleg" route so that 317 the source s-ASBR can send packets directly to the destination s-ASBR 318 through tunneling across the Internetwork. This can be accomplished 319 using a mapping resolution service such as proposed in AERO 320 [I-D.templin-aerolink] or LISP 321 [I-D.ietf-lisp-rfc6830bis][I-D.ietf-lisp-rfc6833bis]. Employment of 322 the mapping resolution service results in a condition known as route 323 optimization. 325 A route optimization example is shown in Figure 2 and Figure 3 below. 326 In the first figure, the dogleg route between correspondents in the 327 stub ASes traverses the path from s-ASBR1 to c-ASBR1 to c-ASBR2 to 328 S-ASBR2. In the second figure, the optimized route goes directly 329 from s-ASBR1 to s-ASBR2, i.e., the c-ASBRs are not included in the 330 path. 332 ........................................................... 333 . . 334 . (:::)-. <- Stub ASes -> (:::)-. . 335 . MNPs-> .-(:::::::::) .-(:::::::::) <-MNPs . 336 . `-(::::)-' `-(::::)-' . 337 . +-------+ +-------+ . 338 . |s-ASBR1| |s-ASBR2| . 339 . +--+--^^+ +^^---+-+ . 340 . \ \\ Dogleg // / . 341 . eBGP\ \\ Route // /eBGP . 342 . \ \\============// / . 343 . +-------+ +-------+ . 344 . +c-ASBR1| +c-ASBR2+ . 345 . +--+----+ +-----+-+ . 346 . +--------------+ . 347 . iBGP . 348 . . 349 . <------------------- Internetwork --------------------> . 350 ............................................................ 352 Figure 2: Dogleg Route Before Optimization 354 ........................................................... 355 . . 356 . (:::)-. <- Stub ASes -> (:::)-. . 357 . MNPs-> .-(:::::::::) .-(:::::::::) <-MNPs . 358 . `-(::::)-' `-(::::)-' . 359 . +-------+ Direct +-------+ . 360 . |s-ASBR1<================>s-ASBR2| . 361 . +--+----+ Route +-----+-+ . 362 . \ / . 363 . eBGP\ /eBGP . 364 . \ / . 365 . +-------+ +-------+ . 366 . +c-ASBR1| +c-ASBR2+ . 367 . +--+----+ +-----+-+ . 368 . +--------------+ . 369 . iBGP . 370 . . 371 . <------------------- Internetwork --------------------> . 372 ............................................................ 374 Figure 3: Direct Route Following Optimization 376 Note that route optimization can fail if the source s-ASBR cannot 377 tunnel packets directly to the destination s-ASBR due to some form of 378 Internetwork blockage such as filtering middleboxes. It is also 379 necessary for the source s-ASBR to quickly detect and adjust to 380 failure of the destination s-ASBR and/or movement of the destination 381 MNP. In both of these cases, significant packet loss could occur 382 before the source s-ASBR can detect that the destination MNP is no 383 longer reachable via the route-optimized path. This implies that 384 route optimized paths may not always be the best choice for carrying 385 safety-of-flight critical packets with high reliability requirements. 387 In all cases, s-ASBRs do not advertise MNPs discovered via route 388 optimization to c-ASBRs. Instead, s-ASBRs keep MNPs discovered via 389 route optimization in a local table that is kept separate from the 390 MNPs of ATN/IPS end systems within their own stub AS. 392 4. Route Availability 394 In the ATN/IPS BGP-based routing system proposed in this document, 395 each s-ASBR always has a default route and can therefore always send 396 packets via the dogleg route through a c-ASBR even if a route 397 optimized path has been established. The direct paths between 398 s-ASBRs and c-ASBRs are maintained by BGP peering session keepalives 399 such that, if a link or an ASBR goes down, BGP will detect the 400 failure and readjust the routing tables. However, ASBRs and the 401 links that interconnect them are expected to be secured as highly- 402 available and fault tolerant critical infrastructure such that 403 peering session failures should be extremely rare. 405 This represents a distinct architectural difference from other 406 approaches that only operate over route optimized paths. With the 407 approach described herein the source s-ASBR will always have a 408 working route, even if only via a dogleg path through a c-ASBR. This 409 gives rise to the possibility of sending {high-priority, low-data- 410 rate} packets via the assured dogleg route and {low-priority, high- 411 data-rate} packets via the optimized route, e.g., based on per-packet 412 quality of service indications. This could also give rise to a fair 413 pricing model that would charge more for use of the high-assurance 414 dogleg path and less for use of the lesser-assured route-optimized 415 path. 417 This distinction is important to aviation networking, where isolated 418 safety-of-flight critical packets such as produced by the Controller 419 Pilot Data Link Communications (CPDLC) facility may not be eligible 420 for retransmission, e.g., if an aviation data link is failing. If 421 there is no route available, the packet can be dropped or delayed and 422 safety-of-flight parameters could be lost. Even when an optimized 423 route is discovered on-demand, the route may not work and again 424 safety-of-flight critical packets could be lost. 426 5. BGP Protocol Considerations 428 The number of eBGP peering sessions that each c-ASBR must service is 429 proportional to the number of s-ASBRs in the system. Network 430 emulations with lightweight virtual machines have shown that a single 431 c-ASBR can service at least 100 eBGP peerings from s-ASBRs that each 432 advertise 10K MNP routes (i.e., 1M total). It is expected that 433 robust c-ASBRs can service many more peerings than this - possibly by 434 multiple orders of magnitude. But even assuming a conservative 435 limit, the number of s-ASBRs could be increased by also increasing 436 the number of c-ASBRs. Since c-ASBRs also peer with each other using 437 iBGP, however, larger-scale c-ASBR deployments may need to employ an 438 adjunct facility such as BGP route reflectors [RFC4456]. 440 The number of aircraft in operation at a given time wordlwide is 441 likely to be significantly less than 1M, but we will assume this 442 number for a worst-case analysis. Assuming an average 1hour flight 443 profile from gate-to-gate, and 10 data link transitions per flight, 444 the entire system will need to service at most 10M BGP updates per 445 hour (2778 updates per second). This number is within the realm of 446 the peak BGP update messaging seen in the global public Internet 447 today [BGP2]. Assuming a BGP update message size of 100 bytes 448 (800bits), the total amount of BGP control message traffic to a 449 single c-ASBR will be less than 2.5Mbps which is a nominal rate for 450 modern data links. 452 Industry standard BGP routers provide configurable parameters with 453 conservative default values. For example, the default hold time is 454 90 seconds, the default keepalive time is 1/3 of the hold time, and 455 the default MinRouteAdvertisementinterval is 30 seconds for eBGP 456 peers and 5 seconds for iBGP peers (see Section 10 of [RFC4271]). 457 For the simple mobile routing system described herein, these 458 parameters can and should be set to more aggressive values to support 459 faster neighbor/link failure detection and faster routing protocol 460 convergence times. For example, a hold time of 3 seconds and a 461 MinRouteAdvertisementinterval of 0 seconds for both iBGP and eBGP. 463 By default, MED only compares metrics that originate from multiple 464 neighbors within the same AS [RFC4451]. In order to compare MED 465 metrics that come from different ASes, a router configuration file 466 entry may be needed (e.g., Cisco routers require the configuration 467 file entry "bgp always-compare-med"). Furthermore, in order for the 468 MED discriminator to be applied correctly, the AS_PATH phase in the 469 BGP route selection process must be disabled (e.g., Cisco routers use 470 the configuration file entry "bgp bestpath as-path ignore"). 472 6. Implementation Status 474 The BGP routing arrangement described in this document has been 475 modeled in realistic network emulations showing that the MED process 476 results in selection of the best peer when multiple peers advertise 477 the same MNP. Modeling has also shown that at least 1 million MNPs 478 can be propagated to each c-ASBR even on lightweight virtual 479 machines. 481 7. IANA Considerations 483 This document does not introduce any IANA considerations. 485 8. Security Considerations 487 ATN/IPS ASBRs on the open Internet are susceptible to the same attack 488 profiles as for any Internet nodes. For this reason, ASBRs should 489 employ physical security and/or IP securing mechanisms such as IPsec 490 [RFC4301], TLS [RFC5246], etc. 492 ATN/IPS ASBRs present targets for Distributed Denial of Service 493 (DDoS) attacks. This concern is no different than for any node on 494 the open Internet, where attackers could send spoofed packets to the 495 node at high data rates. This can be mitigated by connecting ATN/IPS 496 ASBRs over dedicated links with no connections to the Internet and/or 497 when ASBR connections to the Internet are only permitted through 498 well-managed firewalls. 500 ATN/IPS s-ASBRs should institute rate limits to protect low data rate 501 aviation data links from receiving DDoS packet floods. 503 9. Acknowledgements 505 This work is aligned with the FAA as per the SE2025 contract number 506 DTFAWA-15-D-00030. 508 This work is aligned with the NASA Safe Autonomous Systems Operation 509 (SASO) program under NASA contract number NNA16BD84C. 511 This work is aligned with the Boeing Information Technology (BIT) 512 MobileNet program. 514 10. References 515 10.1. Normative References 517 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 518 DOI 10.17487/RFC0791, September 1981, 519 . 521 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 522 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 523 December 1998, . 525 [RFC4271] Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A 526 Border Gateway Protocol 4 (BGP-4)", RFC 4271, 527 DOI 10.17487/RFC4271, January 2006, 528 . 530 [RFC4443] Conta, A., Deering, S., and M. Gupta, Ed., "Internet 531 Control Message Protocol (ICMPv6) for the Internet 532 Protocol Version 6 (IPv6) Specification", RFC 4443, 533 DOI 10.17487/RFC4443, March 2006, 534 . 536 [RFC4451] McPherson, D. and V. Gill, "BGP MULTI_EXIT_DISC (MED) 537 Considerations", RFC 4451, DOI 10.17487/RFC4451, March 538 2006, . 540 [RFC4456] Bates, T., Chen, E., and R. Chandra, "BGP Route 541 Reflection: An Alternative to Full Mesh Internal BGP 542 (IBGP)", RFC 4456, DOI 10.17487/RFC4456, April 2006, 543 . 545 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 546 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 547 DOI 10.17487/RFC4861, September 2007, 548 . 550 10.2. Informative References 552 [BGP] Huston, G., "BGP in 2015, http://potaroo.net", January 553 2016. 555 [BGP2] Huston, G., "BGP Instability Report, 556 http://bgpupdates.potaroo.net/instability/bgpupd.html", 557 May 2017. 559 [CBB] Dul, A., "Global IP Network Mobility using Border Gateway 560 Protocol (BGP), http://www.quark.net/docs/ 561 Global_IP_Network_Mobility_using_BGP.pdf", March 2006. 563 [I-D.ietf-lisp-ddt] 564 Fuller, V., Lewis, D., Ermagan, V., Jain, A., and A. 565 Smirnov, "LISP Delegated Database Tree", draft-ietf-lisp- 566 ddt-09 (work in progress), January 2017. 568 [I-D.ietf-lisp-rfc6830bis] 569 Farinacci, D., Fuller, V., Meyer, D., Lewis, D., and A. 570 Cabellos-Aparicio, "The Locator/ID Separation Protocol 571 (LISP)", draft-ietf-lisp-rfc6830bis-03 (work in progress), 572 May 2017. 574 [I-D.ietf-lisp-rfc6833bis] 575 Fuller, V., Farinacci, D., and A. Cabellos-Aparicio, 576 "Locator/ID Separation Protocol (LISP) Control-Plane", 577 draft-ietf-lisp-rfc6833bis-05 (work in progress), May 578 2017. 580 [I-D.templin-aerolink] 581 Templin, F., "Asymmetric Extended Route Optimization 582 (AERO)", draft-templin-aerolink-74 (work in progress), 583 November 2016. 585 [ICAO] ICAO, I., "http://www.icao.int/Pages/default.aspx", 586 February 2017. 588 [RFC2784] Farinacci, D., Li, T., Hanks, S., Meyer, D., and P. 589 Traina, "Generic Routing Encapsulation (GRE)", RFC 2784, 590 DOI 10.17487/RFC2784, March 2000, 591 . 593 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 594 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 595 December 2005, . 597 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 598 (TLS) Protocol Version 1.2", RFC 5246, 599 DOI 10.17487/RFC5246, August 2008, 600 . 602 [RFC6836] Fuller, V., Farinacci, D., Meyer, D., and D. Lewis, 603 "Locator/ID Separation Protocol Alternative Logical 604 Topology (LISP+ALT)", RFC 6836, DOI 10.17487/RFC6836, 605 January 2013, . 607 Author's Address 609 Fred L. Templin (editor) 610 Boeing Research & Technology 611 P.O. Box 3707 612 Seattle, WA 98124 613 USA 615 Email: fltemplin@acm.org