idnits 2.17.1 draft-ietf-lisp-deployment-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 : ---------------------------------------------------------------------------- ** The document seems to lack a both a reference to RFC 2119 and the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. RFC 2119 keyword, line 509: '... operators, who SHOULD check ownershi...' Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 486 has weird spacing: '...ructure x ...' -- The document date (January 26, 2013) is 4107 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 6830 (Obsoleted by RFC 9300, RFC 9301) ** Obsolete normative reference: RFC 6833 (Obsoleted by RFC 9301) == Outdated reference: A later version (-13) exists of draft-ietf-lisp-eid-block-03 == Outdated reference: A later version (-29) exists of draft-ietf-lisp-sec-04 == Outdated reference: A later version (-15) exists of draft-ietf-lisp-threats-03 -- Obsolete informational reference (is this intentional?): RFC 6834 (Obsoleted by RFC 9302) Summary: 3 errors (**), 0 flaws (~~), 5 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group L. Jakab 3 Internet-Draft Cisco Systems 4 Intended status: Informational A. Cabellos-Aparicio 5 Expires: July 30, 2013 F. Coras 6 J. Domingo-Pascual 7 Technical University of 8 Catalonia 9 D. Lewis 10 Cisco Systems 11 January 26, 2013 13 LISP Network Element Deployment Considerations 14 draft-ietf-lisp-deployment-06.txt 16 Abstract 18 This document discusses the different scenarios for the deployment of 19 the new network elements introduced by the Locator/Identifier 20 Separation Protocol (LISP). 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 http://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 July 30, 2013. 39 Copyright Notice 41 Copyright (c) 2013 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 (http://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 . . . . . . . . . . . . . . . . . . . . . . . . . 3 57 2. Tunnel Routers . . . . . . . . . . . . . . . . . . . . . . . . 4 58 2.1. Customer Edge . . . . . . . . . . . . . . . . . . . . . . 4 59 2.2. Provider Edge . . . . . . . . . . . . . . . . . . . . . . 5 60 2.3. Split ITR/ETR . . . . . . . . . . . . . . . . . . . . . . 7 61 2.4. Inter-Service Provider Traffic Engineering . . . . . . . . 8 62 2.5. Tunnel Routers Behind NAT . . . . . . . . . . . . . . . . 10 63 2.5.1. ITR . . . . . . . . . . . . . . . . . . . . . . . . . 10 64 2.5.2. ETR . . . . . . . . . . . . . . . . . . . . . . . . . 11 65 2.6. Summary and Feature Matrix . . . . . . . . . . . . . . . . 11 66 3. Map Resolvers and Map Servers . . . . . . . . . . . . . . . . 11 67 3.1. Map Servers . . . . . . . . . . . . . . . . . . . . . . . 11 68 3.2. Map Resolvers . . . . . . . . . . . . . . . . . . . . . . 12 69 4. Proxy Tunnel Routers . . . . . . . . . . . . . . . . . . . . . 13 70 4.1. P-ITR . . . . . . . . . . . . . . . . . . . . . . . . . . 13 71 4.2. P-ETR . . . . . . . . . . . . . . . . . . . . . . . . . . 14 72 5. Migration to LISP . . . . . . . . . . . . . . . . . . . . . . 15 73 5.1. LISP+BGP . . . . . . . . . . . . . . . . . . . . . . . . . 15 74 5.2. Mapping Service Provider (MSP) P-ITR Service . . . . . . . 16 75 5.3. Proxy-ITR Route Distribution (PITR-RD) . . . . . . . . . . 16 76 5.4. Migration Summary . . . . . . . . . . . . . . . . . . . . 19 77 6. Step-by-Step Example BGP to LISP Migration Procedure . . . . . 19 78 6.1. Customer Pre-Install and Pre-Turn-up Checklist . . . . . . 19 79 6.2. Customer Activating LISP Service . . . . . . . . . . . . . 21 80 6.3. Cut-Over Provider Preparation and Changes . . . . . . . . 21 81 7. Security Considerations . . . . . . . . . . . . . . . . . . . 22 82 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 22 83 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 22 84 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 23 85 10.1. Normative References . . . . . . . . . . . . . . . . . . . 23 86 10.2. Informative References . . . . . . . . . . . . . . . . . . 23 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 24 89 1. Introduction 91 The Locator/Identifier Separation Protocol (LISP) addresses the 92 scaling issues of the global Internet routing system by separating 93 the current addressing scheme into Endpoint IDentifiers (EIDs) and 94 Routing LOCators (RLOCs). The main protocol specification [RFC6830] 95 describes how the separation is achieved, which new network elements 96 are introduced, and details the packet formats for the data and 97 control planes. 99 LISP assumes that such separation is between the edge and core and 100 uses a map-and-encap scheme for forwarding. While the boundary 101 between both is not strictly defined, one widely accepted definition 102 places it at the border routers of stub autonomous systems, which may 103 carry a partial or complete default-free zone (DFZ) routing table. 104 The initial design of LISP took this location as a baseline for 105 protocol development. However, the applications of LISP go beyond of 106 just decreasing the size of the DFZ routing table, and include 107 improved multihoming and ingress traffic engineering (TE) support for 108 edge networks, and even individual hosts. Throughout the draft we 109 will use the term LISP site to refer to these networks/hosts behind a 110 LISP Tunnel Router. We formally define it as: 112 LISP site: A single host or a set of network elements in an edge 113 network under the administrative control of a single organization, 114 delimited from other networks by LISP Tunnel Router(s). 116 Network element: Active or passive device that is connected 117 connected to other active or passive devices for transporting 118 packet switched data. 120 Since LISP is a protocol which can be used for different purposes, it 121 is important to identify possible deployment scenarios and the 122 additional requirements they may impose on the protocol specification 123 and other protocols. Additionally, this document is intended as a 124 guide for the operational community for LISP deployments in their 125 networks. It is expected to evolve as LISP deployment progresses, 126 and the described scenarios are better understood or new scenarios 127 are discovered. 129 Each subsection considers an element type, discussing the impact of 130 deployment scenarios on the protocol specification. For definition 131 of terms, please refer to the appropriate documents (as cited in the 132 respective sections). 134 2. Tunnel Routers 136 The device that is the gateway between the edge and the core is 137 called Tunnel Router (xTR), performing one or both of two separate 138 functions: 140 1. Encapsulating packets originating from an end host to be 141 transported over intermediary (transit) networks towards the 142 other end-point of the communication 144 2. Decapsulating packets entering from intermediary (transit) 145 networks, originated at a remote end host. 147 The first function is performed by an Ingress Tunnel Router (ITR), 148 the second by an Egress Tunnel Router (ETR). 150 Section 8 of the main LISP specification [RFC6830] has a short 151 discussion of where Tunnel Routers can be deployed and some of the 152 associated advantages and disadvantages. This section adds more 153 detail to the scenarios presented there, and provides additional 154 scenarios as well. 156 2.1. Customer Edge 158 The first scenario we discuss is customer edge, when xTR 159 functionality is placed on the router(s) that connect the LISP site 160 to its upstream(s), but are under its control. As such, this is the 161 most common expected scenario for xTRs, and this document considers 162 it the reference location, comparing the other scenarios to this one. 164 ISP1 ISP2 165 | | 166 | | 167 +----+ +----+ 168 +--|xTR1|--|xTR2|--+ 169 | +----+ +----+ | 170 | | 171 | LISP site | 172 +------------------+ 174 Figure 1: xTRs at the customer edge 176 From the LISP site perspective the main advantage of this type of 177 deployment (compared to the one described in the next section) is 178 having direct control over its ingress traffic engineering. This 179 makes it easy to set up and maintain active/active, active/backup, or 180 more complex TE policies, without involving third parties. 182 Being under the same administrative control, reachability information 183 of all ETRs is easier to synchronize, because the necessary control 184 traffic can be allowed between the locators of the ETRs. A correct 185 synchronous global view of the reachability status is thus available, 186 and the Loc-Status-Bits can be set correctly in the LISP data header 187 of outgoing packets. 189 By placing the tunnel router at the edge of the site, existing 190 internal network configuration does not need to be modified. 191 Firewall rules, router configurations and address assignments inside 192 the LISP site remain unchanged. This helps with incremental 193 deployment and allows a quick upgrade path to LISP. For larger sites 194 with many external connections, distributed in geographically diverse 195 PoPs, and complex internal topology, it may however make more sense 196 to both encapsulate and decapsulate as soon as possible, to benefit 197 from the information in the IGP to choose the best path (see 198 Section 2.3 for a discussion of this scenario). 200 Another thing to consider when placing tunnel routers are MTU issues. 201 Since encapsulating packets increases overhead, the MTU of the end- 202 to-end path may decrease, when encapsulated packets need to travel 203 over segments having close to minimum MTU. Some transit networks are 204 known to provide larger MTU than the typical value of 1500 bytes of 205 popular access technologies used at end hosts (e.g., IEEE 802.3 and 206 802.11). However, placing the LISP router connecting to such a 207 network at the customer edge could possibly bring up MTU issues, 208 depending on the link type to the provider as opposed to the 209 following scenario. See [RFC4459] for MTU considerations of 210 tunneling protocols. 212 2.2. Provider Edge 214 The other location at the core-edge boundary for deploying LISP 215 routers is at the Internet service provider edge. The main incentive 216 for this case is that the customer does not have to upgrade the CE 217 router(s), or change the configuration of any equipment. 218 Encapsulation/decapsulation happens in the provider's network, which 219 may be able to serve several customers with a single device. For 220 large ISPs with many residential/business customers asking for LISP 221 this can lead to important savings, since there is no need to upgrade 222 the software (or hardware, if it's the case) at each client's 223 location. Instead, they can upgrade the software (or hardware) on a 224 few PE routers serving the customers. This scenario is depicted in 225 Figure 2. 227 +----------+ +------------------+ 228 | ISP1 | | ISP2 | 229 | | | | 230 | +----+ | | +----+ +----+ | 231 +--|xTR1|--+ +--|xTR2|--|xTR3|--+ 232 +----+ +----+ +----+ 233 | | | 234 | | | 235 +--<[LISP site]>---+-------+ 237 Figure 2: xTR at the PE 239 While this approach can make transition easy for customers and may be 240 cheaper for providers, the LISP site looses one of the main benefits 241 of LISP: ingress traffic engineering. Since the provider controls 242 the ETRs, additional complexity would be needed to allow customers to 243 modify their mapping entries. 245 The problem is aggravated when the LISP site is multihomed. Consider 246 the scenario in Figure 2: whenever a change to TE policies is 247 required, the customer contacts both ISP1 and ISP2 to make the 248 necessary changes on the routers (if they provide this possibility). 249 It is however unlikely, that both ISPs will apply changes 250 simultaneously, which may lead to inconsistent state for the mappings 251 of the LISP site. Since the different upstream ISPs are usually 252 competing business entities, the ETRs may even be configured to 253 compete, either to attract all the traffic or to get no traffic. The 254 former will happen if the customer pays per volume, the latter if the 255 connectivity has a fixed price. A solution could be to have the 256 mappings in the Map Server(s), and have their operator give control 257 over the entries to customer, much like in the Domain Name System at 258 the time of this writing. 260 Additionally, since xTR1, xTR2, and xTR3 are in different 261 administrative domains, locator reachability information is unlikely 262 to be exchanged among them, making it difficult to set Loc-Status- 263 Bits (LSB) correctly on encapsulated packets. Because of this, and 264 due to the security concerns about LSB described in 265 [I-D.ietf-lisp-threats] their use is discouraged without verifying 266 ETR reachability through the mapping system or other means. Mapping 267 versioning is another alternative [RFC6834]. 269 Compared to the customer edge scenario, deploying LISP at the 270 provider edge might have the advantage of diminishing potential MTU 271 issues, because the tunnel router is closer to the core, where links 272 typically have higher MTUs than edge network links. 274 2.3. Split ITR/ETR 276 In a simple LISP deployment, xTRs are located at the border of the 277 LISP site (see Section 2.1). In this scenario packets are routed 278 inside the domain according to the EID. However, more complex 279 networks may want to route packets according to the destination RLOC. 280 This would enable them to choose the best egress point. 282 The LISP specification separates the ITR and ETR functionality and 283 considers that both entities can be deployed in separated network 284 equipment. ITRs can be deployed closer to the host (i.e., access 285 routers). This way packets are encapsulated as soon as possible, and 286 packets exit the network through the best egress point in terms of 287 BGP policy. In turn, ETRs can be deployed at the border routers of 288 the network, and packets are decapsulated as soon as possible. Once 289 decapsulated, packets are routed based on destination EID, according 290 to internal routing policy. 292 In the following figure we can see an example. The Source (S) 293 transmits packets using its EID and in this particular case packets 294 are encapsulated at ITR_1. The encapsulated packets are routed 295 inside the domain according to the destination RLOC, and can egress 296 the network through the best point (i.e., closer to the RLOC's AS). 297 On the other hand, inbound packets are received by ETR_1 which 298 decapsulates them. Then packets are routed towards S according to 299 the EID, again following the best path. 301 +---------------------------------------+ 302 | | 303 | +-------+ +-------+ +-------+ 304 | | ITR_1 |---------+ | ETR_1 |-RLOC_A--| ISP_A | 305 | +-------+ | +-------+ +-------+ 306 | +-+ | | | 307 | |S| | IGP | | 308 | +-+ | | | 309 | +-------+ | +-------+ +-------+ 310 | | ITR_2 |---------+ | ETR_2 |-RLOC_B--| ISP_B | 311 | +-------+ +-------+ +-------+ 312 | | 313 +---------------------------------------+ 315 Figure 3: Split ITR/ETR Scenario 317 This scenario has a set of implications: 319 o The site must carry at least partial BGP routes in order to choose 320 the best egress point, increasing the complexity of the network. 321 However, this is usually already the case for LISP sites that 322 would benefit from this scenario. 324 o If the site is multihomed to different ISPs and any of the 325 upstream ISPs is doing uRPF filtering, this scenario may become 326 impractical. ITRs need to determine the exit ETR, for setting the 327 correct source RLOC in the encapsulation header. This adds 328 complexity and reliability concerns. 330 o In LISP, ITRs set the reachability bits when encapsulating data 331 packets. Hence, ITRs need a mechanism to be aware of the liveness 332 of all ETRs serving their site. 334 o MTU within the site network must be large enough to accommodate 335 encapsulated packets. 337 o In this scenario, each ITR is serving fewer hosts than in the case 338 when it is deployed at the border of the network. It has been 339 shown that cache hit ratio grows logarithmically with the amount 340 of users [cache]. Taking this into account, when ITRs are 341 deployed closer to the host the effectiveness of the mapping cache 342 may be lower (i.e., the miss ratio is higher). Another 343 consequence of this is that the site may transmit a higher amount 344 of Map-Requests, increasing the load on the distributed mapping 345 database. To lower the impact, the site could use a local caching 346 Map Resolver. 348 o By placing the ITRs inside the site, they will still need global 349 RLOCs, and this may add complexity to intra-site routing 350 configuration, and further intra-site issues when there is a 351 change of providers. 353 2.4. Inter-Service Provider Traffic Engineering 355 With LISP, two LISP sites can route packets among them and control 356 their ingress TE policies. Typically, LISP is seen as applicable to 357 stub networks, however the LISP protocol can also be applied to 358 transit networks recursively. 360 Consider the scenario depicted in Figure 4. Packets originating from 361 the LISP site Stub1, client of ISP_A, with destination Stub4, client 362 of ISP_B, are LISP encapsulated at their entry point into the ISP_A's 363 network. The external IP header now has as the source RLOC an IP 364 from ISP_A's address space and destination RLOC from ISP_B's address 365 space. One or more ASes separate ISP_A from ISP_B. With a single 366 level of LISP encapsulation, Stub4 has control over its ingress 367 traffic. However, at the time of this writing, ISP_B has only BGP 368 tools (such as prefix deaggregation) to control on which of his own 369 upstream or peering links should packets enter. This is either not 370 feasible (if fine-grained per-customer control is required, the very 371 specific prefixes may not be propagated) or increases DFZ table size. 373 _.--. 374 Stub1 ... +-------+ ,-'' `--. +-------+ ... Stub3 375 \ | R_A1|----,' `. ---|R_B1 | / 376 --| R_A2|---( Transit ) | |-- 377 Stub2 .../ | R_A3|-----. ,' ---|R_B2 | \... Stub4 378 +-------+ `--. _.-' +-------+ 379 ... ISP_A `--'' ISP_B ... 381 Figure 4: Inter-Service provider TE scenario 383 A solution for this is to apply LISP recursively. ISP_A and ISP_B 384 may reach a bilateral agreement to deploy their own private mapping 385 system. ISP_A then encapsulates packets destined for the prefixes of 386 ISP_B, which are listed in the shared mapping system. Note that in 387 this case the packet is double-encapsulated (using R_A1, R_A2 or R_A3 388 as source and R_B1 or R_B2 as destination in the example above). 389 ISP_B's ETR removes the outer, second layer of LISP encapsulation 390 from the incoming packet, and routes it towards the original RLOC, 391 the ETR of Stub4, which does the final decapsulation. 393 If ISP_A and ISP_B agree to share a private distributed mapping 394 database, both can control their ingress TE without the need of 395 deaggregating prefixes. In this scenario the private database 396 contains RLOC-to-RLOC bindings. The convergence time on the TE 397 policies updates is expected to be fast, since ISPs only have to 398 update/query a mapping to/from the database. 400 This deployment scenario includes two important caveats. First, it 401 is intended to be deployed between only two ISPs (ISP_A and ISP_B in 402 Figure 4). If more than two ISPs use this approach, then the xTRs 403 deployed at the participating ISPs must either query multiple mapping 404 systems, or the ISPs must agree on a common shared mapping system. 405 Second, the scenario is only recommended for ISPs providing 406 connectivity to LISP sites, such that source RLOCs of packets to be 407 reencapsulated belong to said ISP. Otherwise the participating ISPs 408 must register prefixes they do not own in the above mentioned private 409 mapping system. Failure to follow these recommendations may lead to 410 operational and security issues when deploying this scenario. 412 Besides these recommendations, the main disadvantages of this 413 deployment case are: 415 o Extra LISP header is needed. This increases the packet size and 416 requires that the MTU between both ISPs accommodates double- 417 encapsulated packets. 419 o The ISP ITR must encapsulate packets and therefore must know the 420 RLOC-to-RLOC binding. These bindings are stored in a mapping 421 database and may be cached in the ITR's mapping cache. Cache 422 misses lead to an additional lookup latency, unless a push based 423 mapping system is used for the private mapping system. 425 o The operational overhead of maintaining the shared mapping 426 database. 428 o If an IPv6 address block is reserved for EID use, as specified in 429 [I-D.ietf-lisp-eid-block], the EID-to-RLOC encapsulation (first 430 level) can avoid LISP processing altogether for non-LISP 431 destinations. The ISP tunnel routers however will not be able to 432 take advantage of this optimization, all RLOC-to-RLOC mappings 433 need a lookup in the private database (or map-cache, once results 434 are cached). 436 2.5. Tunnel Routers Behind NAT 438 NAT in this section refers to IPv4 network address and port 439 translation. 441 2.5.1. ITR 443 Packets encapsulated by an ITR are just UDP packets from a NAT 444 device's point of view, and they are handled like any UDP packet, 445 there are no additional requirements for LISP data packets. 447 Map-Requests sent by an ITR, which create the state in the NAT table, 448 have a different 5-tuple in the IP header than the Map-Reply 449 generated by the authoritative ETR. Since the source address of this 450 packet is different from the destination address of the request 451 packet, no state will be matched in the NAT table and the packet will 452 be dropped. To avoid this, the NAT device has to do the following: 454 o Send all UDP packets with source port 4342, regardless of the 455 destination port, to the RLOC of the ITR. The most simple way to 456 achieve this is configuring 1:1 NAT mode from the external RLOC of 457 the NAT device to the ITR's RLOC (Called "DMZ" mode in consumer 458 broadband routers). 460 o Rewrite the ITR-AFI and "Originating ITR RLOC Address" fields in 461 the payload. 463 This setup supports only a single ITR behind the NAT device. 465 2.5.2. ETR 467 An ETR placed behind NAT is reachable from the outside by the 468 Internet-facing locator of the NAT device. It needs to know this 469 locator (and configure a loopback interface with it), so that it can 470 use it in Map-Reply and Map-Register messages. Thus support for 471 dynamic locators for the mapping database is needed in LISP 472 equipment. 474 Again, only one ETR behind the NAT device is supported. 476 An implication of the issues described above is that LISP sites with 477 xTRs can not be behind carrier based NATs, since two different sites 478 would collide on the port forwarding. 480 2.6. Summary and Feature Matrix 482 Feature CE PE Split Recursive 483 ------------------------------------------------------------- 484 Control of ingress TE x - x x 485 No modifications to existing 486 int. network infrastructure x x - - 487 Loc-Status-Bits sync x - x x 488 MTU/PMTUD issues minimized - x - x 490 3. Map Resolvers and Map Servers 492 3.1. Map Servers 494 The Map Server learns EID-to-RLOC mapping entries from an 495 authoritative source and publishes them in the distributed mapping 496 database. These entries are learned through authenticated Map- 497 Register messages sent by authoritative ETRs. Also, upon reception 498 of a Map-Request, the Map Server verifies that the destination EID 499 matches an EID-prefix for which it is authoritative for, and then re- 500 encapsulates and forwards it to a matching ETR. Map Server 501 functionality is described in detail in [RFC6833]. 503 The Map Server is provided by a Mapping Service Provider (MSP). The 504 MSP participates in the global distributed mapping database 505 infrastructure, by setting up connections to other participants, 506 according to the specific mapping system that is employed (e.g., ALT, 507 DDT). Participation in the mapping database, and the storing of EID- 508 to-RLOC mapping data is subject to the policies of the "root" 509 operators, who SHOULD check ownership rights for the EID prefixes 510 stored in the database by participants. These policies are out of 511 the scope of this document. 513 In all cases, the MSP configures its Map Server(s) to publish the 514 prefixes of its clients in the distributed mapping database and start 515 encapsulating and forwarding Map-Requests to the ETRs of the AS. 516 These ETRs register their prefix(es) with the Map Server(s) through 517 periodic authenticated Map-Register messages. In this context, for 518 some LISP end sites, there is a need for mechanisms to: 520 o Automatically distribute EID prefix(es) shared keys between the 521 ETRs and the EID-registrar Map Server. 523 o Dynamically obtain the address of the Map Server in the ETR of the 524 AS. 526 The Map Server plays a key role in the reachability of the EID- 527 prefixes it is serving. On the one hand it is publishing these 528 prefixes into the distributed mapping database and on the other hand 529 it is encapsulating and forwarding Map-Requests to the authoritative 530 ETRs of these prefixes. ITRs encapsulating towards EIDs under the 531 responsibility of a failed Map Server will be unable to look up any 532 of their covering prefixes. The only exception are the ITRs that 533 already contain the mappings in their local cache. In this case ITRs 534 can reach ETRs until the entry expires (typically 24 hours). For 535 this reason, redundant Map Server deployments are desirable. A set 536 of Map Servers providing high-availability service to the same set of 537 prefixes is called a redundancy group. ETRs are configured to send 538 Map-Register messages to all Map Servers in the redundancy group. To 539 achieve fail-over (or load-balancing, if desired), known mapping 540 system specific best practices should be used. 542 Additionally, if a Map Server has no reachability for any ETR serving 543 a given EID block, it should not originate that block into the 544 mapping system. 546 3.2. Map Resolvers 548 A Map Resolver a is a network infrastructure component which accepts 549 LISP encapsulated Map-Requests, typically from an ITR, and finds the 550 appropriate EID-to-RLOC mapping by either consulting its local cache 551 or by consulting the distributed mapping database. Map Resolver 552 functionality is described in detail in [RFC6833]. 554 Anyone with access to the distributed mapping database can set up a 555 Map Resolver and provide EID-to-RLOC mapping lookup service. 556 Database access setup is mapping system specific. 558 For performance reasons, it is recommended that LISP sites use Map 559 Resolvers that are topologically close to their ITRs. ISPs 560 supporting LISP will provide this service to their customers, 561 possibly restricting access to their user base. LISP sites not in 562 this position can use open access Map Resolvers, if available. 563 However, regardless of the availability of open access resolvers, the 564 MSP providing the Map Server(s) for a LISP site should also make 565 available Map Resolver(s) for the use of that site. 567 In medium to large-size ASes, ITRs must be configured with the RLOC 568 of a Map Resolver, operation which can be done manually. However, in 569 Small Office Home Office (SOHO) scenarios a mechanism for 570 autoconfiguration should be provided. 572 One solution to avoid manual configuration in LISP sites of any size 573 is the use of anycast RLOCs for Map Resolvers similar to the DNS root 574 server infrastructure. Since LISP uses UDP encapsulation, the use of 575 anycast would not affect reliability. LISP routers are then shipped 576 with a preconfigured list of well know Map Resolver RLOCs, which can 577 be edited by the network administrator, if needed. 579 The use of anycast also helps improving mapping lookup performance. 580 Large MSPs can increase the number and geographical diversity of 581 their Map Resolver infrastructure, using a single anycasted RLOC. 582 Once LISP deployment is advanced enough, very large content providers 583 may also be interested running this kind of setup, to ensure minimal 584 connection setup latency for those connecting to their network from 585 LISP sites. 587 While Map Servers and Map Resolvers implement different 588 functionalities within the LISP mapping system, they can coexist on 589 the same device. For example, MSPs offering both services, can 590 deploy a single Map Resolver/Map Server in each PoP where they have a 591 presence. 593 4. Proxy Tunnel Routers 595 4.1. P-ITR 597 Proxy Ingress Tunnel Routers (P-ITRs) are part of the non-LISP/LISP 598 transition mechanism, allowing non-LISP sites to reach LISP sites. 599 They announce via BGP certain EID prefixes (aggregated, whenever 600 possible) to attract traffic from non-LISP sites towards EIDs in the 601 covered range. They do the mapping system lookup, and encapsulate 602 received packets towards the appropriate ETR. Note that for the 603 reverse path LISP sites can reach non-LISP sites simply by not 604 encapsulating traffic. See [RFC6832] for a detailed description of 605 P-ITR functionality. 607 The success of new protocols depends greatly on their ability to 608 maintain backwards compatibility and inter-operate with the 609 protocol(s) they intend to enhance or replace, and on the incentives 610 to deploy the necessary new software or equipment. A LISP site needs 611 an interworking mechanism to be reachable from non-LISP sites. A 612 P-ITR can fulfill this role, enabling early adopters to see the 613 benefits of LISP, similar to tunnel brokers helping the transition 614 from IPv4 to IPv6. A site benefits from new LISP functionality 615 (proportionally with existing global LISP deployment) when going 616 LISP, so it has the incentives to deploy the necessary tunnel 617 routers. In order to be reachable from non-LISP sites it has two 618 options: keep announcing its prefix(es) with BGP, or have a P-ITR 619 announce prefix(es) covering them. 621 If the goal of reducing the DFZ routing table size is to be reached, 622 the second option is preferred. Moreover, the second option allows 623 LISP-based ingress traffic engineering from all sites. However, the 624 placement of P-ITRs significantly influences performance and 625 deployment incentives. Section 5 is dedicated to the migration to a 626 LISP-enabled Internet, and includes deployment scenarios for P-ITRs. 628 4.2. P-ETR 630 In contrast to P-ITRs, P-ETRs are not required for the correct 631 functioning of all LISP sites. There are two cases, where they can 632 be of great help: 634 o LISP sites with unicast reverse path forwarding (uRPF) 635 restrictions, and 637 o Communication between sites using different address family RLOCs. 639 In the first case, uRPF filtering is applied at their upstream PE 640 router. When forwarding traffic to non-LISP sites, an ITR does not 641 encapsulate packets, leaving the original IP headers intact. As a 642 result, packets will have EIDs in their source address. Since we are 643 discussing the transition period, we can assume that a prefix 644 covering the EIDs belonging to the LISP site is advertised to the 645 global routing tables by a P-ITR, and the PE router has a route 646 towards it. However, the next hop will not be on the interface 647 towards the CE router, so non-encapsulated packets will fail uRPF 648 checks. 650 To avoid this filtering, the affected ITR encapsulates packets 651 towards the locator of the P-ETR for non-LISP destinations. Now the 652 source address of the packets, as seen by the PE router is the ITR's 653 locator, which will not fail the uRPF check. The P-ETR then 654 decapsulates and forwards the packets. 656 The second use case is IPv4-to-IPv6 transition. Service providers 657 using older access network hardware, which only supports IPv4 can 658 still offer IPv6 to their clients, by providing a CPE device running 659 LISP, and P-ETR(s) for accessing IPv6-only non-LISP sites and LISP 660 sites, with IPv6-only locators. Packets originating from the client 661 LISP site for these destinations would be encapsulated towards the 662 P-ETR's IPv4 locator. The P-ETR is in a native IPv6 network, 663 decapsulating and forwarding packets. For non-LISP destination, the 664 packet travels natively from the P-ETR. For LISP destinations with 665 IPv6-only locators, the packet will go through a P-ITR, in order to 666 reach its destination. 668 For more details on P-ETRs see the [RFC6832] draft. 670 P-ETRs can be deployed by ISPs wishing to offer value-added services 671 to their customers. As is the case with P-ITRs, P-ETRs too may 672 introduce path stretch. Because of this the ISP needs to consider 673 the tradeoff of using several devices, close to the customers, to 674 minimize it, or few devices, farther away from the customers, 675 minimizing cost instead. 677 Since the deployment incentives for P-ITRs and P-ETRs are different, 678 it is likely they will be deployed in separate devices, except for 679 the CDN case, which may deploy both in a single device. 681 In all cases, the existence of a P-ETR involves another step in the 682 configuration of a LISP router. CPE routers, which are typically 683 configured by DHCP, stand to benefit most from P-ETRs. 684 Autoconfiguration of the P-ETR locator could be achieved by a DHCP 685 option, or adding a P-ETR field to either Map-Notifys or Map-Replies. 687 5. Migration to LISP 689 This section discusses a deployment architecture to support the 690 migration to a LISP-enabled Internet. The loosely defined terms of 691 "early transition phase", "late transition phase", and "LISP Internet 692 phase" refer to time periods when LISP sites are a minority, a 693 majority, or represent all edge networks respectively. 695 5.1. LISP+BGP 697 For sites wishing to go LISP with their PI prefix the least 698 disruptive way is to upgrade their border routers to support LISP, 699 register the prefix into the LISP mapping system, but keep announcing 700 it with BGP as well. This way LISP sites will reach them over LISP, 701 while legacy sites will be unaffected by the change. The main 702 disadvantage of this approach is that no decrease in the DFZ routing 703 table size is achieved. Still, just increasing the number of LISP 704 sites is an important gain, as an increasing LISP/non-LISP site ratio 705 will slowly decrease the need for BGP-based traffic engineering that 706 leads to prefix deaggregation. That, in turn, may lead to a decrease 707 in the DFZ size and churn in the late transition phase. 709 This scenario is not limited to sites that already have their 710 prefixes announced with BGP. Newly allocated EID blocks could follow 711 this strategy as well during the early LISP deployment phase, 712 depending on the cost/benefit analysis of the individual networks. 713 Since this leads to an increase in the DFZ size, the following 714 architecture should be preferred for new allocations. 716 5.2. Mapping Service Provider (MSP) P-ITR Service 718 In addition to publishing their clients' registered prefixes in the 719 mapping system, MSPs with enough transit capacity can offer them 720 P-ITR service as a separate service. This service is especially 721 useful for new PI allocations, to sites without existing BGP 722 infrastructure, that wish to avoid BGP altogether. The MSP announces 723 the prefix into the DFZ, and the client benefits from ingress traffic 724 engineering without prefix deaggregation. The downside of this 725 scenario is adding path stretch. 727 Routing all non-LISP ingress traffic through a third party which is 728 not one of its ISPs is only feasible for sites with modest amounts of 729 traffic (like those using the IPv6 tunnel broker services today), 730 especially in the first stage of the transition to LISP, with a 731 significant number of legacy sites. This is because the handling of 732 said traffic is likely to result in additional costs, which would be 733 passed down to the client. When the LISP/non-LISP site ratio becomes 734 high enough, this approach can prove increasingly attractive. 736 Compared to LISP+BGP, this approach avoids DFZ bloat caused by prefix 737 deaggregation for traffic engineering purposes, resulting in slower 738 routing table increase in the case of new allocations and potential 739 decrease for existing ones. Moreover, MSPs serving different clients 740 with adjacent aggregatable prefixes may lead to additional decrease, 741 but quantifying this decrease is subject to future research study. 743 5.3. Proxy-ITR Route Distribution (PITR-RD) 745 Instead of a LISP site, or the MSP, announcing their EIDs with BGP to 746 the DFZ, this function can be outsourced to a third party, a P-ITR 747 Service Provider (PSP). This will result in a decrease of the 748 operational complexity both at the site and at the MSP. 750 The PSP manages a set of distributed P-ITR(s) that will advertise the 751 corresponding EID prefixes through BGP to the DFZ. These P-ITR(s) 752 will then encapsulate the traffic they receive for those EIDs towards 753 the RLOCs of the LISP site, ensuring their reachability from non-LISP 754 sites. Note that handling non-LISP-originated traffic may incur 755 additional costs for the PSP, which may be passed down to the client. 757 While it is possible for a PSP to manually configure each client's 758 EID routes to be announced, this approach offers little flexibility 759 and is not scalable. This section presents a scalable architecture 760 that offers automatic distribution of EID routes to LISP sites and 761 service providers. 763 The architecture requires no modification to existing LISP network 764 elements, but it introduces a new (conceptual) network element, the 765 EID Route Server, defined as a router that either propagates routes 766 learned from other EID Route Servers, or it originates EID Routes. 767 The EID-Routes that it originates are those that it is authoritative 768 for. It propagates these routes to Proxy-ITRs within the AS of the 769 EID Route Server. It is worth to note that a BGP capable router can 770 be also considered as an EID Route Server. 772 Further, an EID-Route is defined as a prefix originated via the Route 773 Server of the mapping service provider, which should be aggregated if 774 the MSP has multiple customers inside a single large continuous 775 prefix. This prefix is propagated to other P-ITRs both within the 776 MSP and to other P-ITR operators it peers with. EID Route Servers 777 are operated either by the LISP site, MSPs or PSPs, and they may be 778 collocated with a Map Server or P-ITR, but are a functionally 779 discrete entity. They distribute EID-Routes, using BGP, to other 780 domains, according to policies set by participants. 782 MSP (AS64500) 783 RS ---> P-ITR 784 | / 785 | _.--./ 786 ,-'' /`--. 787 LISP site ---,' | v `. 788 ( | DFZ )----- Mapping system 789 non-LISP site ----. | ^ ,' 790 `--. / _.-' 791 | `--'' 792 v / 793 P-ITR 794 PSP (AS64501) 796 Figure 5: The P-ITR Route Distribution architecture 798 The architecture described above decouples EID origination from route 799 propagation, with the following benefits: 801 o Can accurately represent business relationships between P-ITR 802 operators 804 o More mapping system agnostic 806 o Minor changes to P-ITR implementation, no changes to other 807 components 809 In the example in the figure we have a MSP providing services to the 810 LISP site. The LISP site does not run BGP, and gets an EID 811 allocation directly from a RIR, or from the MSP, who may be a LIR. 812 Existing PI allocations can be migrated as well. The MSP ensures the 813 presence of the prefix in the mapping system, and runs an EID Route 814 Server to distribute it to P-ITR service providers. Since the LISP 815 site does not run BGP, the prefix will be originated with the AS 816 number of the MSP. 818 In the simple case depicted in Figure 5 the EID-Route of LISP Site 819 will be originated by the Route Server, and announced to the DFZ by 820 the PSP's P-ITRs with AS path 64501 64500. From that point on, the 821 usual BGP dynamics apply. This way, routes announced by P-ITR are 822 still originated by the authoritative Route Server. Note that the 823 peering relationships between MSP/PSPs and those in the underlying 824 forwarding plane may not be congruent, making the AS path to a P-ITR 825 shorter than it is in reality. 827 The non-LISP site will select the best path towards the EID-prefix, 828 according to its local BGP policies. Since AS-path length is usually 829 an important metric for selecting paths, a careful placement of P-ITR 830 could significantly reduce path-stretch between LISP and non-LISP 831 sites. 833 The architecture allows for flexible policies between MSP/PSPs. 834 Consider the EID Route Server networks as control plane overlays, 835 facilitating the implementation of policies necessary to reflect the 836 business relationships between participants. The results are then 837 injected to the common underlying forwarding plane. For example, 838 some MSP/PSPs may agree to exchange EID-Prefixes and only announce 839 them to each of their forwarding plane customers. Global 840 reachability of an EID-prefix depends on the MSP the LISP site buys 841 service from, and is also subject to agreement between the mentioned 842 parties. 844 In terms of impact on the DFZ, this architecture results in a slower 845 routing table increase for new allocations, since traffic engineering 846 will be done at the LISP level. For existing allocations migrating 847 to LISP, the DFZ may decrease since MSPs may be able to aggregate the 848 prefixes announced. 850 Compared to LISP+BGP, this approach avoids DFZ bloat caused by prefix 851 deaggregation for traffic engineering purposes, resulting in slower 852 routing table increase in the case of new allocations and potential 853 decrease for existing ones. Moreover, MSPs serving different clients 854 with adjacent aggregatable prefixes may lead to additional decrease, 855 but quantifying this decrease is subject to future research study. 857 The flexibility and scalability of this architecture does not come 858 without a cost however: A PSP operator has to establish either 859 transit or peering relationships to improve their connectivity. 861 5.4. Migration Summary 863 The following table presents the expected effects of the different 864 transition scenarios during a certain phase on the DFZ routing table 865 size: 867 Phase | LISP+BGP | MSP P-ITR | PITR-RD 868 -----------------+--------------+-----------------+---------------- 869 Early transition | no change | slower increase | slower increase 870 Late transition | may decrease | slower increase | slower increase 871 LISP Internet | considerable decrease 873 It is expected that PITR-RD will co-exist with LISP+BGP during the 874 migration, with the latter being more popular in the early transition 875 phase. As the transition progresses and the MSP P-ITR and PITR-RD 876 ecosystem gets more ubiquitous, LISP+BGP should become less 877 attractive, slowing down the increase of the number of routes in the 878 DFZ. 880 6. Step-by-Step Example BGP to LISP Migration Procedure 882 6.1. Customer Pre-Install and Pre-Turn-up Checklist 884 1. Determine how many current physical service provider connections 885 the customer has and their existing bandwidth and traffic 886 engineering requirements. 888 This information will determine the number of routing locators, 889 and the priorities and weights that should be configured on the 890 xTRs. 892 2. Make sure customer router has LISP capabilities. 894 * Check OS version of the CE router. If LISP is an add-on, 895 check if it is installed. 897 This information can be used to determine if the platform is 898 appropriate to support LISP, in order to determine if a 899 software and/or hardware upgrade is required. 901 * Have customer upgrade (if necessary, software and/or hardware) 902 to be LISP capable. 904 3. Obtain current running configuration of CE router. A suggested 905 LISP router configuration example can be customized to the 906 customer's existing environment. 908 4. Verify MTU Handling 910 * Request increase in MTU to 1556 on service provider 911 connections. Prior to MTU change verify that 1500 byte packet 912 from P-xTR to RLOC with do not fragment (DF-bit) bit set. 914 * Ensure they are not filtering ICMP unreachable or time- 915 exceeded on their firewall or router. 917 LISP, like any tunneling protocol, will increase the size of 918 packets when the LISP header is appended. If increasing the MTU 919 of the access links is not possible, care must be taken that ICMP 920 is not being filtered in order to allow for Path MTU Discovery to 921 take place. 923 5. Validate member prefix allocation. 925 This step is to check if the prefix used by the customer is a 926 direct (Provider Independent), or if it is a prefix assigned by a 927 physical service provider (Provider Aggregatable). If the 928 prefixes are assigned by other service provivers then a Letter of 929 Agreement is required to announce prefixes through the Proxy 930 Service Provider. 932 6. Verify the member RLOCs and their reachability. 934 This step ensures that the RLOCs configured on the CE router are 935 in fact reachable and working. 937 7. Prepare for cut-over. 939 * If possible, have a host outside of all security and filtering 940 policies connected to the console port of the edge router or 941 switch. 943 * Make sure customer has access to the router in order to 944 configure it. 946 6.2. Customer Activating LISP Service 948 1. Customer configures LISP on CE router(s) from service provider 949 recommended configuration. 951 The LISP configuration consists of the EID prefix, the locators, 952 and the weights and priorities of the mapping between the two 953 values. In addition, the xTR must be configured with Map 954 Resolver(s), Map Server(s) and the shared key for registering to 955 Map Server(s). If required, Proxy-ETR(s) may be configured as 956 well. 958 In addition to the LISP configuration, the following: 960 * Ensure default route(s) to next-hop external neighbors are 961 included and RLOCs are present in configuration. 963 * If two or more routers are used, ensure all RLOCs are included 964 in the LISP configuration on all routers. 966 * It will be necessary to redistribute default route via IGP 967 between the external routers. 969 2. When transition is ready perform a soft shutdown on existing eBGP 970 peer session(s) 972 * From CE router, use LIG to ensure registration is successful. 974 * To verify LISP connectivity, ping LISP connected sites. See 975 http://www.lisp4.net/ and/or http://www.lisp6.net/ for 976 potential candidates. If possible, find ping destinations 977 that are not covered by a prefix in the global BGP routing 978 system, because PITRs may deliver the packets even if LISP 979 connectivity is not working. Traceroutes may help discover if 980 this is the case. 982 * To verify connectivity to non-LISP sites, try accessing a 983 landmark (e.g., a major Internet site) via a web browser. 985 6.3. Cut-Over Provider Preparation and Changes 987 1. Verify site configuration and then active registration on Map 988 Server(s) 989 * Authentication key 991 * EID prefix 993 2. Add EID space to map-cache on proxies 995 3. Add networks to BGP advertisement on proxies 997 * Modify route-maps/policies on P-xTRs 999 * Modify route policies on core routers (if non-connected 1000 member) 1002 * Modify ingress policers on core routers 1004 * Ensure route announcement in looking glass servers, RouteViews 1006 4. Perform traffic verification test 1008 * Ensure MTU handling is as expected (PMTUD working) 1010 * Ensure proxy-ITR map-cache population 1012 * Ensure access from traceroute/ping servers around Internet 1014 * Use a looking glass, to check for external visibility of 1015 registration via several Map Resolvers (e.g., 1016 http://lispmon.net/). 1018 7. Security Considerations 1020 Security implications of LISP deployments are to be discussed in 1021 separate documents. [I-D.ietf-lisp-threats] gives an overview of 1022 LISP threat models, while securing mapping lookups is discussed in 1023 [I-D.ietf-lisp-sec]. 1025 8. IANA Considerations 1027 This memo includes no request to IANA. 1029 9. Acknowledgements 1031 Many thanks to Margaret Wasserman for her contribution to the IETF76 1032 presentation that kickstarted this work. The authors would also like 1033 to thank Damien Saucez, Luigi Iannone, Joel Halpern, Vince Fuller, 1034 Dino Farinacci, Terry Manderson, Noel Chiappa, Hannu Flinck, Paul 1035 Vinciguerra and everyone else who provided input. 1037 10. References 1039 10.1. Normative References 1041 [RFC6830] Farinacci, D., Fuller, V., Meyer, D., and D. Lewis, "The 1042 Locator/ID Separation Protocol (LISP)", RFC 6830, 1043 January 2013. 1045 [RFC6832] Lewis, D., Meyer, D., Farinacci, D., and V. Fuller, 1046 "Interworking between Locator/ID Separation Protocol 1047 (LISP) and Non-LISP Sites", RFC 6832, January 2013. 1049 [RFC6833] Fuller, V. and D. Farinacci, "Locator/ID Separation 1050 Protocol (LISP) Map-Server Interface", RFC 6833, 1051 January 2013. 1053 10.2. Informative References 1055 [I-D.ietf-lisp-eid-block] 1056 Iannone, L., Lewis, D., Meyer, D., and V. Fuller, "LISP 1057 EID Block", draft-ietf-lisp-eid-block-03 (work in 1058 progress), November 2012. 1060 [I-D.ietf-lisp-sec] 1061 Maino, F., Ermagan, V., Cabellos-Aparicio, A., Saucez, D., 1062 and O. Bonaventure, "LISP-Security (LISP-SEC)", 1063 draft-ietf-lisp-sec-04 (work in progress), October 2012. 1065 [I-D.ietf-lisp-threats] 1066 Saucez, D., Iannone, L., and O. Bonaventure, "LISP Threats 1067 Analysis", draft-ietf-lisp-threats-03 (work in progress), 1068 October 2012. 1070 [RFC4459] Savola, P., "MTU and Fragmentation Issues with In-the- 1071 Network Tunneling", RFC 4459, April 2006. 1073 [RFC6834] Iannone, L., Saucez, D., and O. Bonaventure, "Locator/ID 1074 Separation Protocol (LISP) Map-Versioning", RFC 6834, 1075 January 2013. 1077 [cache] Jung, J., Sit, E., Balakrishnan, H., and R. Morris, "DNS 1078 performance and the effectiveness of caching", 2002. 1080 Authors' Addresses 1082 Lorand Jakab 1083 Cisco Systems 1084 170 Tasman Drive 1085 San Jose, CA 95134 1086 USA 1088 Email: lojakab@cisco.com 1090 Albert Cabellos-Aparicio 1091 Technical University of Catalonia 1092 C/Jordi Girona, s/n 1093 BARCELONA 08034 1094 Spain 1096 Email: acabello@ac.upc.edu 1098 Florin Coras 1099 Technical University of Catalonia 1100 C/Jordi Girona, s/n 1101 BARCELONA 08034 1102 Spain 1104 Email: fcoras@ac.upc.edu 1106 Jordi Domingo-Pascual 1107 Technical University of Catalonia 1108 C/Jordi Girona, s/n 1109 BARCELONA 08034 1110 Spain 1112 Email: jordi.domingo@ac.upc.edu 1114 Darrel Lewis 1115 Cisco Systems 1116 170 Tasman Drive 1117 San Jose, CA 95134 1118 USA 1120 Email: darlewis@cisco.com