idnits 2.17.1 draft-ietf-lisp-predictive-rlocs-06.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 7, 2020) is 1449 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- == Unused Reference: 'I-D.ietf-lisp-eid-anonymity' is defined on line 578, but no explicit reference was found in the text ** Obsolete normative reference: RFC 6830 (Obsoleted by RFC 9300, RFC 9301) == Outdated reference: A later version (-15) exists of draft-farinacci-lisp-geo-09 == Outdated reference: A later version (-16) exists of draft-ietf-lisp-eid-anonymity-08 == Outdated reference: A later version (-13) exists of draft-ietf-lisp-eid-mobility-05 == Outdated reference: A later version (-29) exists of draft-ietf-lisp-sec-20 Summary: 1 error (**), 0 flaws (~~), 6 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group D. Farinacci 3 Internet-Draft lispers.net 4 Intended status: Experimental P. Pillay-Esnault 5 Expires: November 8, 2020 Independent 6 May 7, 2020 8 LISP Predictive RLOCs 9 draft-ietf-lisp-predictive-rlocs-06 11 Abstract 13 This specification describes a method to achieve near-zero packet 14 loss when an EID is roaming quickly across RLOCs. 16 Requirements Language 18 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 19 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 20 document are to be interpreted as described in [RFC2119]. 22 Status of This Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at https://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on November 8, 2020. 39 Copyright Notice 41 Copyright (c) 2020 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents 46 (https://trustee.ietf.org/license-info) in effect on the date of 47 publication of this document. Please review these documents 48 carefully, as they describe your rights and restrictions with respect 49 to this document. Code Components extracted from this document must 50 include Simplified BSD License text as described in Section 4.e of 51 the Trust Legal Provisions and are provided without warranty as 52 described in the Simplified BSD License. 54 Table of Contents 56 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 57 2. Definition of Terms . . . . . . . . . . . . . . . . . . . . . 3 58 3. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 3 59 4. Design Details . . . . . . . . . . . . . . . . . . . . . . . 5 60 4.1. RLE Encoding . . . . . . . . . . . . . . . . . . . . . . 6 61 4.2. Packet Delivery Optimizations . . . . . . . . . . . . . . 7 62 4.3. Trading Off Replication Cost . . . . . . . . . . . . . . 8 63 5. Directional Paths with Intersections . . . . . . . . . . . . 9 64 6. Multicast Considerations . . . . . . . . . . . . . . . . . . 10 65 7. Multiple Address-Family Considerations . . . . . . . . . . . 11 66 8. Scaling Considerations . . . . . . . . . . . . . . . . . . . 11 67 9. Security Considerations . . . . . . . . . . . . . . . . . . . 12 68 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 12 69 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 12 70 11.1. Normative References . . . . . . . . . . . . . . . . . . 12 71 11.2. Informative References . . . . . . . . . . . . . . . . . 13 72 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . 14 73 Appendix B. Document Change Log . . . . . . . . . . . . . . . . 14 74 B.1. Changes to draft-ietf-lisp-predictive-rlocs-05 . . . . . 14 75 B.2. Changes to draft-ietf-lisp-predictive-rlocs-05 . . . . . 14 76 B.3. Changes to draft-ietf-lisp-predictive-rlocs-04 . . . . . 14 77 B.4. Changes to draft-ietf-lisp-predictive-rlocs-03 . . . . . 14 78 B.5. Changes to draft-ietf-lisp-predictive-rlocs-02 . . . . . 14 79 B.6. Changes to draft-ietf-lisp-predictive-rlocs-01 . . . . . 14 80 B.7. Changes to draft-ietf-lisp-predictive-rlocs-00 . . . . . 15 81 B.8. Changes to draft-farinacci-lisp-predictive-rlocs-02 . . . 15 82 B.9. Changes to draft-farinacci-lisp-predictive-rlocs-01 . . . 15 83 B.10. Changes to draft-farinacci-lisp-predictive-rlocs-00 . . . 15 84 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 15 86 1. Introduction 88 The LISP architecture [RFC6830] specifies two namespaces, End-Point 89 IDs (EIDs) and Routing Locators (RLOCs). An EID identifies a node in 90 the network and the RLOC indicates the EID's topological location. 91 When an node roams in the network, its EID remains fixed and 92 unchanged but the RLOCs associated with it change to reflect its new 93 topological attachment point. This specification will focus EIDs and 94 RLOCs residing in separate nodes. An EID is assigned to a host node 95 that roams while the RLOCs are assigned to network nodes that stay 96 stationary and are part of the network topology. For example, a set 97 of devices on an aircraft are assigned EIDs, and base stations on the 98 ground attached to the Internet infrastructure are configured as LISP 99 xTRs where their RLOCs are used for the bindings of the EIDs on the 100 aircraft up in the air. 102 The scope of this specification will not emphasize general physical 103 roaming as an aircraft would do in the sky but in a direction that is 104 more predictable such as a train traveling on a track or vehicle that 105 travels along a road. 107 2. Definition of Terms 109 Roaming-EID - is a network node that moves from one topological 110 location in the network to another. The network node uses the 111 same EID when it is roaming. That is, the EID address does not 112 change for reasons of mobility. A roaming-EID can also be a 113 roaming EID-prefix where a set of EIDs covered by the prefix are 114 all roaming and fate-sharing the same set of RLOCs at the same 115 time. 117 Predictive RLOCs - is a set of ordered RLOCs in a list each assigned 118 to LISP xTRs where the next RLOC in the list has high probability 119 it will be the next LISP xTR in a physical path going in a single 120 predictable direction. 122 Road-Side-Units (RSUs) - is a network node that acts as a router, 123 more specifically as a LISP xTR. The xTR automatically discovers 124 roaming-EIDs that come into network connectivity range and relays 125 packets to and from the roaming-EID. RSUs are typically deployed 126 along a directional path like a train track or road and are in 127 connectivity range of devices that travel along the directional 128 path. 130 3. Overview 132 The goal of this specification is to describe a make-before-break 133 EID-mobility mechanism that offers near-zero packet loss. Offering 134 minimal packet loss, not only allows transport layers to operate more 135 efficiently, but because an EID does not change while moving, 136 transport layer session continuity is maintained. To achieve these 137 requirements, a mechanism that reacts to the mobility event is 138 necessary but not sufficient. So the question is not that there 139 isn't a reaction but when it happens. By using some predictive 140 algorithms, we can guess with high probability where the EID will 141 roam to next. We can achieve this to a point where packet data will 142 be at the new location when the EID arrives. 144 First we should examine both the send and receive directions with 145 respect to the roaming-EID. Refer to Figure 1 for discussion. We 146 show a network node with a fixed EID address assigned to a roaming- 147 EID moving along a train track. And there are LISP xTRs deployed as 148 Road-Side-Units to support the connectivity between the roaming-EID 149 and the infrastructure or to another roaming-EID. 151 Roaming-EID ----> 153 ====//====//====//====//====//====//====//====//===//====//====//==== 154 // // // // // // // // // // // 155 ====//====//====//====//====//====//====//====//===//====//====//==== 157 xTR xTR xTR xTR xTR xTR 158 A B C D E F 160 Figure 1: Directional Mobility 162 For the send direction from roaming-EID to any destination can be 163 accomplish as a local decision. As long as the roaming-EID is in 164 signal range to any xTR along the path, it can use it to forward 165 packets. The LISP xTR, acting as an ITR, can forward packets to 166 destinations in non-LISP sites as well as to stationary and roaming 167 EIDs in LISP sites. This is accomplished by using the LISP overlay 168 via dynamic packet encapsulation. When the roaming-EID sends 169 packets, the LISP xTR must discover the EID and MAY register the EID 170 with a set of RLOCs to the mapping system 171 [I-D.ietf-lisp-eid-mobility]. The discovery process is important 172 because the LISP xTR, acting as an ETR for decapsulating packets that 173 arrive, needs to know what local ports or radios to send packets to 174 the roaming-EID. 176 Much of the focus of this design is on the packet direction to the 177 roaming-EID. And how remote LISP ITRs find the current location 178 (RLOCs) quickly when the roaming-EID is moving at high speed. This 179 specification solves the fast roaming with the introduction of the 180 Predictive-RLOCs algorithm. 182 Since a safe assumption is that the roaming-EID is going in one 183 direction and cannot deviate from it allows us to know a priori the 184 next set of RLOCs the roaming-EID will pass by. Referring to 185 Figure 1, if the roaming-EID is in range near xTR-A, then as it 186 moves, it will at some point pass by xTR-B and xTR-C, and so on. As 187 the roaming-EID moves, one could time when the EID is mapped to RLOC 188 A, and when it should change to RLOC B and so on. However, the speed 189 of movement of the roaming-EID won't be constant and the variables 190 involved in consistent timing cannot be relied on. Furthermore, 191 timing the move is not a make-before-break algorithm, meaning the 192 reaction of the binding happens at the time the roaming-EID is 193 discovered by an xTR. One cannot achieve fast hand-offs when message 194 signaling will be required to inform remote ITRs of the new binding. 196 The Predictive RLOCs algorithm allows a set of RLOCs, in an ordered 197 list, to be provided to remote ITRs so they have the information 198 available and local for when they need to use it. Therefore, no 199 control-plane message signaling occurs when the roaming-EID is 200 discovered by LISP xTRs. 202 4. Design Details 204 Predictive RLOCs accommodates for encapsulated packets to be 205 delivered to Road-Side-Unit LISP xTRs regardless where the roaming- 206 EID is currently positioned. 208 Referring to Figure 1, the following sequence is performed: 210 1. The Predictive RLOCs are registered to the mapping system as a 211 LCAF encoded Replication List Entry (RLE) Type [RFC8060]. The 212 registration can happen by one or more RSUs or by a third-party. 213 When registered by an RSU, and when no coordination is desired, 214 they each register their own RLOC with merge-semantics so the 215 list can be created and maintained in the LISP Map-Server. When 216 registered by a third-party, the complete list of RLOCs can be 217 included in the RLE. 219 2. There can be multiple RLEs present each as different RLOC- 220 records so a remote ITR can select one RLOC-record versus the 221 other based in priority and weight policy [RFC6830]. 223 3. When a remote ITR receives a packet destined for a roaming-EID, 224 it encapsulates and replicates to each RLOC in the RLE thereby 225 delivering the packet to the locations the roaming-EID is about 226 to appear. There are some cases where packets will go to 227 locations where the roaming-EID has already been, but see 228 Section 4.2 for packet delivery optimizations. 230 4. When the ETR resident RSU receives an encapsulated packet, it 231 decapsulates the packet and then determines if the roaming-EID 232 had been previously discovered. If the EID has not been 233 discovered, the ETR drops the packet. Otherwise, the ETR 234 delivers the decapsulated packet on the port interface the 235 roaming-EID was discovered on. 237 4.1. RLE Encoding 239 The LCAF [RFC8060] Replication List Entry (RLE) will be used to 240 encode the Predictive RLOCs in an RLOC-record for Map-Registers, Map- 241 Reply, and Map-Notify messages [RFC6830]. 243 0 1 2 3 244 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 1 245 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 246 | AFI = 16387 | Rsvd1 | Flags | 247 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 248 | Type = 13 | Rsvd2 | 4 + n | 249 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 250 | Rsvd3 | Rsvd4 | Level Value | 251 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 252 | AFI = x | RTR/ETR #1 ... | 253 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 254 | Rsvd3 | Rsvd4 | Level Value | 255 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 256 | AFI = x | RTR/ETR #n ... | 257 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 259 When the RLOC-record contains an RLE with RLOC entries all with the 260 same level value, it means the physical order listed is the 261 directional path of the RSUs. This will typically be the result of a 262 third-party doing the registration where it knows ahead of time the 263 RSU deployment. 265 When each RSU is registering with merge-semantics on their own, the 266 level number is used to place them in an ordered list. Since the 267 registrations come at different times and therefore arrive in 268 different order than the physical RSU path, the level number creates 269 the necessary sequencing. Each RSU needs to know its position in the 270 path relative to other RSUs. For example, in xTR-B, it would 271 register with level 1 since it is after xTR-A (and before xTR-C). So 272 if the registration order was xTR-B with level 1, xTR-C with level 2, 273 and xTR-A with level 0, the RLE list stored in the mapping system 274 would be (xTR-A, xTR-B, xTR-C). It is recommended that level numbers 275 be assigned in increments of 10 so latter insertion is possible. 277 The use of Geo-Prefixes and Geo-Points [I-D.farinacci-lisp-geo] can 278 be used to compare the physical presence of each RSU with respect to 279 each other, so they can choose level numbers to sequence themselves. 280 Also if the xTRs register with a Geo-Point in an RLOC-record, then 281 perhaps the Map-Server could sequence the RLE list. 283 4.2. Packet Delivery Optimizations 285 Since the remote ITR will replicate to all RLOCs in the RLE, a 286 situation is created where packets go to RLOCs that don't need to. 287 For instance, if the roaming-EID is along side of xTR-B and the RLE 288 is (xTR-A, xTR-B, xTR-C), there is no reason to replicate to xTR-A 289 since the roaming-EID has passed it and the the signal range is weak 290 or lost. However, replicating to xTR-B and xTR-C is important to 291 deliver packets to where the roaming-EID resides and where it is 292 about to go to. 294 A simple data-plane option, which converges fairly quickly is to have 295 the remote xTR, acting as an ETR, when packets are sent from the 296 roaming-EID, examine the source RLOC in the outer header of the 297 encapsulated packet. If the source RLOC is xTR-B, the remote xTR can 298 determine that the roaming-EID has moved past xTR-A and no longer 299 needs to encapsulate packets to xTR-A's RLOC. 301 In addition, the remote ITR can use RLOC-probing to determine if each 302 RLOC in the RLE is reachable. And if not reachable, exclude from the 303 list of RLOCs to replicate to. 305 This solution also handles the case where xTR-A and xTR-B may overlap 306 in radio signal range, but the signal is weak from the roaming-EID to 307 xTR-A but stronger to xTR-B. In this case, the roaming-EID selects 308 xTR-B to send packets that inform the remote xTR that return packets 309 should not be encapsulated to xTR-A. 311 There are also situations where the RSUs are in signal range of each 312 other in which case they could report reachability status of each 313 other. The use of the Locator-Status-Bits of the LISP encapsulation 314 header could be used to convey this information to the remote xTR. 315 This would only occur when the roaming-EID was discovered by both 316 xTR-A and xTR-B so it was possible for either xTR to reach the 317 roaming-EID. Either an IGP like routing protocol would be required 318 to allow each xTR to know the other could reach the roaming-EID or a 319 path trace tool (i.e. traceroute) could be originated by one xTR 320 targeted for the roaming-EID but MAC-forwarded through the other xTR. 321 These and other roaming-EID reachability mechanisms are work in 322 progress and for further study. 324 When a remote ITR is doing "Directional Mobility" and replicating to 325 the last RLOC in the RLE list, it has a decision to guess where the 326 roaming-EID will move to next. Conservatively, an ITR can replicate 327 to the entire set of RLOCs in the RLE list and wait to see if the 328 roaming-EID moves to one of the RLOCs in the RLE list. 330 Or more liberally, the remote ITR can assume the new roaming 331 direction. For example, with an RLE list of (xTR-A, xTR-B, xTR-C, 332 xTR-D) and the roaming-EID is at D, the remote ITR can replicate to 333 all of A, B, C and D to determine where the roaming-EID will move to 334 next. If the roaming-EID moves to C after it was at D, it is 335 possible that the EID is moving in the opposite direction from C to B 336 to A. This would be known as "Back-n-Forth Mobility". If an 337 implementation is configured to support this for a particluar EID, 338 the remote ITR could replicate in this sequence as the roaming-EID 339 moves from A to D and back to A: (A, B, C, D), (B, C, D), (C, D), (D, 340 C, B, A), (C, B, A), (B, A), and again (A, B, C, D). 342 The roaming-EID could be doing "Circular Mobility" where it moves 343 from A to D directionally, next from D-to-A, and then back to A to D 344 directionally again. This form of mobility is just as predicatable 345 as "Back-n-Forth Mobility" since a consistent pattern can be relied 346 on. Both of these mobility forms can be achieved with near-zero 347 packet loss. 349 On the other hand, the roaming-EID can be roaming arbitrarily using 350 "Random Mobility" where it could roam in the following combinations: 351 A-to-B, A-to-C, A-to-D, B-to-A, B-to-C, B-to-D, C-to-A, C-to-B, C-to- 352 D, D-to-A, D-to-B, or D-to-C. In this situation, when a return 353 packet arrives at the ITR, it could then just replicate to where the 354 roaming-EID is at rest and do so for a short period of time before it 355 replciates to the entire RLE list again. Using the wrong time period 356 could lead to packet loss. All these types of mobility can be 357 supported by the remote ITR in a local manner without consulting or 358 depending on any other LISP system. It is left for further study, if 359 any of the mobility types above should be encoded in the mapping 360 system. 362 4.3. Trading Off Replication Cost 364 If RLE lists are large, packet replication can occur to locations 365 well before the roaming-EID arrives. Making RLE lists small is 366 useful without sacrificing hand-off issues or incurring packet loss 367 to the application. By having overlapping RLEs in separate RLOC- 368 records we a simple mechanism to solve this problem. Here is an 369 example mapping entry to illustrate the point: 371 EID = , RLOC-records: 372 RLOC = (RLE: xTR-A, xTR-B) 373 RLOC = (RLE: xTR-B, xTR-C, xTR-D, xTR-E) 374 RLOC = (RLE: xTR-E, xTR-F) 376 When the remote ITR is encapsulating to xTR-B as a decision to use 377 the first RLOC-record, it can decide to move to use the second RLOC- 378 record because xTR-B is the last entry in the first RLOC-record and 379 the first entry in the second RLOC-record. When there are 380 overlapping RLEs, the remote ITR can decide when it is more efficient 381 to switch over. For example, when the roaming-EID is in range of 382 xTR-A, the remote ITR uses the first RLOC-record so the wasted 383 replication cost is to xTR-B only versus a worse cost when using the 384 second RLOC-record. But when the roaming-EID is in range of xTR-B, 385 then replicating to the other xTRs in the second RLOC-record may be 386 crucial if the roaming-EID has increased speed. And when the 387 roaming-EID may be at rest in a parked mode, then the remote ITR 388 encapsulates to only xTR-F using the third RLOC-record since the 389 roaming-EID has moved past xTR-E. 391 In addition, to eliminate unnecessary replication to xTRs further 392 down a directional path, GEO-prefixes [I-D.farinacci-lisp-geo] can be 393 used so only nearby xTRs that the roaming-EID is about to come in 394 contact with are the only ones to receive encapsulated packets. 396 Even when replication lists are not large, we can reduce the cost of 397 replication that the entire network bears by moving the replicator 398 away from the the source (i.e. the ITR) and closer to the RSUs (i.e. 399 the ETRs). See the use of RTRs for Replication Engineering 400 techniques in [RFC8378]. 402 5. Directional Paths with Intersections 404 A roaming-EID could be registered to the mapping system with the 405 following nested RLE mapping: 407 EID = , RLOC-records: 408 RLOC = (RLE: xTR-A, xTR-B, xTR-C, (RLE: xTR-X, xTR-Y, xTR-Z), 409 (RLE: xTR-I, xTR-J, xTR-K), xTR-D, xTR-E) 411 The mapping entry above describes 3 directional paths where the 412 ordered list has encoded one-level of two nested RLEs to denote 413 intersections in a horizontal path. Which is drawn as: 415 | | xTR-K 416 | | 417 | | 418 | | xTR-J 419 | | 420 | | 421 Roaming | | xTR-I 422 EID ----> | | 423 --------------------------------------- ------------------------------ 424 --------------------------------------- ------------------------------ 425 xTR-A xTR-B xTR-C | | xTR-D xTR-E 426 | | 427 | | xTR-X 428 | | 429 | | 430 | | xTR-Y 431 | | 432 | | 433 | | xTR-Z 435 When the roaming-EID is on the horizontal path, the remote-ITRs 436 typically replicate to the rest the of the xTRs in the ordered list. 437 When a list has nested RLEs, the replication should occur to at least 438 the first RLOC in a nested RLE list. So if the remote-ITR is 439 replicating to xTR-C, xTR-D, and xTR-E, it should also replicate to 440 xTR-X and xTR-I anticipating a possible turn at the intersection. 441 But when the roaming-EID is known to be at xTR-D (a left or right 442 hand turn was not taken), replication should only occur to xTR-D and 443 xTR-E. Once either xTR-I or xTR-X is determined to be where the 444 roaming-EID resides, then the replication occurs on the respective 445 directional path only. 447 When nested RLEs are used it may be difficult to get merge-semantics 448 to work when each xTR registers itself. So it is suggested a third- 449 party registers nested RLEs. It is left to further study to 450 understand better how to automate this. 452 6. Multicast Considerations 454 In this design, the remote ITR is receiving a unicast packet from an 455 EID and replicating and encapsulating to each RLOC in an RLE list. 456 This form of replication is no different than a traditional multicast 457 replication function. So replicating multicast packets in the same 458 fashion is a fallout from this design. 460 If there are multiple roaming-EIDs joined to the same multicast group 461 but reside at different RSUs, a merge has to be done of any pruned 462 RLEs used for forwarding. So if roaming-EID-1 resides at xTR-A and 463 roaming-EID-2 resides at xTR-B and the RLE list is (xTR-A, xTR-B, 464 xTR-C), and they are joined to the same multicast group, then 465 replication occurs to all of xTR-A, xTR-B, and xTR-C. Even since 466 roaming-EID-2 is past xTR-A, packets need to be delivered to xTR-A 467 for roaming-EID-1. In addition, packets need to be delivered to 468 xTR-C because roaming-EID-1 and roaming-EID-2 will get to xTR-C (and 469 roaming-EID-1 may get there sooner if it is traveling faster than 470 roaming-EID-2). 472 When a roaming-EID is a multicast source, procedures from [RFC8378] 473 are used to deliver packets to multicast group members anywhere in 474 the network. The solution requires no signaling to the RSUs. When 475 RSUs receive multicast packets from a roaming-EID, they do a 476 (roaming-EID,G) mapping database lookup to find the replication list 477 of ETRs to encapsulate to. 479 7. Multiple Address-Family Considerations 481 Note that roaming-EIDs can be assigned IPv6 EID addresses while the 482 RSU xTRs could be using IPv4 RLOC addresses. Any combination of 483 address-families can be supported as well as for multicast packet 484 forwarding, where (S,G) are IPv6 addresses entries and replication is 485 done with IPv4 RLOCs in the outer header. 487 8. Scaling Considerations 489 One can imagine there will be a large number of roaming-EIDs. So 490 there is a strong desire to efficiently store state in the mapping 491 database and the in remote ITRs map-caches. It is likely, that 492 roaming-EIDs may share the same path and move at the same speed (EID 493 devices on a train) and therefore share the same Predictive RLOCs. 494 And since EIDs are not reassigned for mobility purposes or may be 495 temporal , they will not be topologically aggregatable, so they 496 cannot compress into a single EID-prefix mapping entry that share the 497 same RLOC-set. 499 By using a level of indirection with the mapping system this problem 500 can be solved. The following mapping entries could exist in the 501 mapping database: 503 EID = , RLOC-records: 504 RLOC = (afi=: "am-train-to-paris") 505 EID = , RLOC-records: 506 RLOC = (afi=: "am-train-to-paris") 507 EID = , RLOC-records: 508 RLOC = (afi=: "am-train-to-paris") 510 EID = "am-train-to-paris", RLOC-records: 511 RLOC = (afi=lcaf/RLE-type: xTR-A, xTR-B, xTR-C) 513 EID = "am-train-to-paris-passengers", RLOC-records: 514 RLOC = (afi=lcaf/afi-list-type: , , ) 516 Each passenger that boards a train has their EID registered to point 517 to the name of the train "am-train-to-paris". And then the train 518 with EID "am-train-to-paris" stores the Predictive RLOC-set. When a 519 remote-ITR wants to encapsulate packets for an EID, it looks up the 520 EID in the mapping database gets the name "am-train-to-paris" 521 returned. Then the remote-ITR does another lookup for the name "am- 522 train-to-paris" to get the RLE list returned. 524 When new EIDs board the train, the RLE mapping entry does not need to 525 be modified. Only an EID-to-name mapping is registered for the 526 specific new EID. Optionally, another name "am-train-to-paris- 527 passengers" can be registered as an EID to allow mapping to all 528 specific EIDs which are on the train. This can be used for 529 inventory, billing, or security purposes. 531 This optimization comes at a cost of a 2-stage lookup. However, if 532 both sets of mapping entries are registered to the same Map-Server, a 533 combined RLOC-set could be returned. This idea is for further study. 535 9. Security Considerations 537 LISP has procedures for supporting both control-plane security 538 [I-D.ietf-lisp-sec] and data-plane security [RFC8061]. 540 10. IANA Considerations 542 At this time there are no requests for IANA. 544 11. References 546 11.1. Normative References 548 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 549 Requirement Levels", BCP 14, RFC 2119, 550 DOI 10.17487/RFC2119, March 1997, 551 . 553 [RFC6830] Farinacci, D., Fuller, V., Meyer, D., and D. Lewis, "The 554 Locator/ID Separation Protocol (LISP)", RFC 6830, 555 DOI 10.17487/RFC6830, January 2013, 556 . 558 [RFC8060] Farinacci, D., Meyer, D., and J. Snijders, "LISP Canonical 559 Address Format (LCAF)", RFC 8060, DOI 10.17487/RFC8060, 560 February 2017, . 562 [RFC8061] Farinacci, D. and B. Weis, "Locator/ID Separation Protocol 563 (LISP) Data-Plane Confidentiality", RFC 8061, 564 DOI 10.17487/RFC8061, February 2017, 565 . 567 [RFC8378] Moreno, V. and D. Farinacci, "Signal-Free Locator/ID 568 Separation Protocol (LISP) Multicast", RFC 8378, 569 DOI 10.17487/RFC8378, May 2018, 570 . 572 11.2. Informative References 574 [I-D.farinacci-lisp-geo] 575 Farinacci, D., "LISP Geo-Coordinate Use-Cases", draft- 576 farinacci-lisp-geo-09 (work in progress), April 2020. 578 [I-D.ietf-lisp-eid-anonymity] 579 Farinacci, D., Pillay-Esnault, P., and W. Haddad, "LISP 580 EID Anonymity", draft-ietf-lisp-eid-anonymity-08 (work in 581 progress), April 2020. 583 [I-D.ietf-lisp-eid-mobility] 584 Portoles-Comeras, M., Ashtaputre, V., Moreno, V., Maino, 585 F., and D. Farinacci, "LISP L2/L3 EID Mobility Using a 586 Unified Control Plane", draft-ietf-lisp-eid-mobility-05 587 (work in progress), November 2019. 589 [I-D.ietf-lisp-sec] 590 Maino, F., Ermagan, V., Cabellos-Aparicio, A., and D. 591 Saucez, "LISP-Security (LISP-SEC)", draft-ietf-lisp-sec-20 592 (work in progress), January 2020. 594 Appendix A. Acknowledgments 596 The author would like to thank the LISP WG for their review and 597 acceptance of this draft. 599 Appendix B. Document Change Log 601 [RFC Editor: Please delete this section on publication as RFC.] 603 B.1. Changes to draft-ietf-lisp-predictive-rlocs-05 605 o Posted May 2020. 607 o Update document timer and references. 609 B.2. Changes to draft-ietf-lisp-predictive-rlocs-05 611 o Posted November 2019. 613 o Update document timer and references. 615 B.3. Changes to draft-ietf-lisp-predictive-rlocs-04 617 o Posted May 2019. 619 o Update document timer and references. 621 B.4. Changes to draft-ietf-lisp-predictive-rlocs-03 623 o Posted November 2018. 625 o Update document timer and references. 627 B.5. Changes to draft-ietf-lisp-predictive-rlocs-02 629 o Posted May 2018. 631 o Update document timer and references. 633 B.6. Changes to draft-ietf-lisp-predictive-rlocs-01 635 o Posted November 2017. 637 o Update document timer and references. 639 o Reflect comments from Prague meeting presentation. There is no 640 need for "RLE Usage Types" as suggested. The ITR can treat what 641 RLOCs it replicates to as a local matter via implementation 642 configuration. RLE Directional is default. Circular rotation, 643 back-n-forth, and random selection of RLOCs is up to the ITR. 645 B.7. Changes to draft-ietf-lisp-predictive-rlocs-00 647 o Posted June 2017. 649 o Make this specification a working group document. It is a copy of 650 draft-farinacci-lisp-predictive-rlocs-02. 652 B.8. Changes to draft-farinacci-lisp-predictive-rlocs-02 654 o Posted May 2017 to update document timer. 656 B.9. Changes to draft-farinacci-lisp-predictive-rlocs-01 658 o Posted November 2016 to update document timer. 660 B.10. Changes to draft-farinacci-lisp-predictive-rlocs-00 662 o Initial post April 2016. 664 Authors' Addresses 666 Dino Farinacci 667 lispers.net 668 San Jose, CA 669 USA 671 Email: farinacci@gmail.com 673 Padma Pillay-Esnault 674 Independent 675 San Clara, CA 676 USA 678 Email: padma.ietf@gmail.com