idnits 2.17.1 draft-farinacci-lisp-rfc6830bis-00.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 (November 13, 2016) is 2720 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC5496' is defined on line 2152, but no explicit reference was found in the text == Unused Reference: 'RFC6115' is defined on line 2161, but no explicit reference was found in the text == Unused Reference: 'RFC6833' is defined on line 2169, but no explicit reference was found in the text == Unused Reference: 'RFC6836' is defined on line 2177, but no explicit reference was found in the text == Unused Reference: 'BGP-SEC' is defined on line 2191, but no explicit reference was found in the text == Unused Reference: 'CONS' is defined on line 2198, but no explicit reference was found in the text == Unused Reference: 'EMACS' is defined on line 2203, but no explicit reference was found in the text == Unused Reference: 'LISP-DDT' is defined on line 2223, but no explicit reference was found in the text == Unused Reference: 'RADIR' is defined on line 2257, but no explicit reference was found in the text == Unused Reference: 'RFC4107' is defined on line 2279, but no explicit reference was found in the text == Unused Reference: 'RFC6480' is defined on line 2298, but no explicit reference was found in the text == Unused Reference: 'RFC6518' is defined on line 2302, but no explicit reference was found in the text == Unused Reference: 'RFC6837' is defined on line 2318, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) ** Downref: Normative reference to an Informational RFC: RFC 3232 ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126) ** Downref: Normative reference to an Informational RFC: RFC 6115 ** Obsolete normative reference: RFC 6833 (Obsoleted by RFC 9301) ** Obsolete normative reference: RFC 6834 (Obsoleted by RFC 9302) ** Downref: Normative reference to an Experimental RFC: RFC 6836 -- Duplicate reference: RFC6834, mentioned in 'RFC7215', was also mentioned in 'RFC6834'. ** Obsolete normative reference: RFC 6834 (ref. 'RFC7215') (Obsoleted by RFC 9302) Summary: 8 errors (**), 0 flaws (~~), 14 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group D. Farinacci 3 Internet-Draft V. Fuller 4 Intended status: Standards Track D. Meyer 5 Expires: May 17, 2017 D. Lewis 6 Cisco Systems 7 A. Cabellos (Ed.) 8 UPC/BarcelonaTech 9 November 13, 2016 11 The Locator/ID Separation Protocol (LISP) 12 draft-farinacci-lisp-rfc6830bis-00 14 Abstract 16 This document describes the Locator/ID Separation Protocol (LISP) 17 data-plane encapsulation protocol. LISP defines two namespaces, End- 18 point Identifiers (EIDs) that identify end-hosts and Routing Locators 19 (RLOCs) that identify network attachment points. With this, LISP 20 effectively separates control from data, and allows routers to create 21 overlay networks. LISP-capable routers exchange encapsulated packets 22 according to EID-to-RLOC mappings stored in a local map-cache. The 23 map-cache is populated by the LISP Control-Plane protocol 24 [REF_TO_RFC6833bis]. 26 LISP requires no change to either host protocol stacks or to underlay 27 routers and offers Traffic Engineering, multihoming and mobility, 28 among other features. 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at http://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on May 17, 2017. 47 Copyright Notice 49 Copyright (c) 2016 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (http://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 65 2. Requirements Notation . . . . . . . . . . . . . . . . . . . . 4 66 3. Definition of Terms . . . . . . . . . . . . . . . . . . . . . 4 67 4. Basic Overview . . . . . . . . . . . . . . . . . . . . . . . 9 68 4.1. Packet Flow Sequence . . . . . . . . . . . . . . . . . . 11 69 5. LISP Encapsulation Details . . . . . . . . . . . . . . . . . 13 70 5.1. LISP IPv4-in-IPv4 Header Format . . . . . . . . . . . . . 14 71 5.2. LISP IPv6-in-IPv6 Header Format . . . . . . . . . . . . . 15 72 5.3. Tunnel Header Field Descriptions . . . . . . . . . . . . 16 73 6. LISP EID-to-RLOC Map-Cache . . . . . . . . . . . . . . . . . 20 74 7. Dealing with Large Encapsulated Packets . . . . . . . . . . . 20 75 7.1. A Stateless Solution to MTU Handling . . . . . . . . . . 21 76 7.2. A Stateful Solution to MTU Handling . . . . . . . . . . . 22 77 8. Using Virtualization and Segmentation with LISP . . . . . . . 22 78 9. Routing Locator Selection . . . . . . . . . . . . . . . . . . 23 79 10. Routing Locator Reachability . . . . . . . . . . . . . . . . 24 80 10.1. Echo Nonce Algorithm . . . . . . . . . . . . . . . . . . 27 81 10.2. RLOC-Probing Algorithm . . . . . . . . . . . . . . . . . 28 82 11. EID Reachability within a LISP Site . . . . . . . . . . . . . 29 83 12. Routing Locator Hashing . . . . . . . . . . . . . . . . . . . 29 84 13. Changing the Contents of EID-to-RLOC Mappings . . . . . . . . 30 85 13.1. Clock Sweep . . . . . . . . . . . . . . . . . . . . . . 31 86 13.2. Solicit-Map-Request (SMR) . . . . . . . . . . . . . . . 32 87 13.3. Database Map-Versioning . . . . . . . . . . . . . . . . 33 88 14. Multicast Considerations . . . . . . . . . . . . . . . . . . 34 89 15. Router Performance Considerations . . . . . . . . . . . . . . 35 90 16. Mobility Considerations . . . . . . . . . . . . . . . . . . . 36 91 16.1. Slow Mobility . . . . . . . . . . . . . . . . . . . . . 36 92 16.2. Fast Mobility . . . . . . . . . . . . . . . . . . . . . 36 93 16.3. LISP Mobile Node Mobility . . . . . . . . . . . . . . . 37 94 17. LISP xTR Placement and Encapsulation Methods . . . . . . . . 37 95 17.1. First-Hop/Last-Hop xTRs . . . . . . . . . . . . . . . . 39 96 17.2. Border/Edge xTRs . . . . . . . . . . . . . . . . . . . . 39 97 17.3. ISP Provider Edge (PE) xTRs . . . . . . . . . . . . . . 40 98 17.4. LISP Functionality with Conventional NATs . . . . . . . 40 99 17.5. Packets Egressing a LISP Site . . . . . . . . . . . . . 40 100 18. Traceroute Considerations . . . . . . . . . . . . . . . . . . 41 101 18.1. IPv6 Traceroute . . . . . . . . . . . . . . . . . . . . 42 102 18.2. IPv4 Traceroute . . . . . . . . . . . . . . . . . . . . 42 103 18.3. Traceroute Using Mixed Locators . . . . . . . . . . . . 43 104 19. Security Considerations . . . . . . . . . . . . . . . . . . . 43 105 20. Network Management Considerations . . . . . . . . . . . . . . 43 106 21. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 43 107 21.1. LISP ACT and Flag Fields . . . . . . . . . . . . . . . . 44 108 21.2. LISP Address Type Codes . . . . . . . . . . . . . . . . 44 109 21.3. LISP UDP Port Numbers . . . . . . . . . . . . . . . . . 44 110 21.4. LISP Key ID Numbers . . . . . . . . . . . . . . . . . . 44 111 22. References . . . . . . . . . . . . . . . . . . . . . . . . . 45 112 22.1. Normative References . . . . . . . . . . . . . . . . . . 45 113 22.2. Informative References . . . . . . . . . . . . . . . . . 47 114 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . 51 115 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 51 117 1. Introduction 119 This document describes the Locator/Identifier Separation Protocol 120 (LISP). LISP is an encapsulation protocol built around the 121 fundamental idea of separating the topological location of a network 122 attachment point from the node's identity [CHIAPPA]. As a result 123 LISP creates two namespaces: Endpoint Identifiers (EIDs), that are 124 used to identify end-hosts (e.g., nodes or Virtual Machines) and 125 routable Routing Locators (RLOCs), used to identify network 126 attachment points. LISP then defines functions for mapping between 127 the two numbering spaces and for encapsulating traffic originated by 128 devices using non-routable EIDs for transport across a network 129 infrastructure that routes and forwards using RLOCs. 131 LISP is an overlay protocol that separates control from data-plane, 132 this document specifies the data-plane, how LISP-capable routers 133 (Tunnel Routers) exchange packets by encapsulating them to the 134 appropriate location. Tunnel routers are equipped with a cache, 135 called map-cache, that contains EID-to-RLOC mappings. The map-cache 136 is populated using the LISP Control-Plane protocol [REF_to_6833bis]. 138 LISP does not require changes to either host protocol stack or to 139 underlay routers. By separating the EID from the RLOC space, LISP 140 offers native Traffic Engineering, multihoming and mobility, among 141 other features. 143 Creation of LISP was initially motivated by discussions during the 144 IAB-sponsored Routing and Addressing Workshop held in Amsterdam in 145 October 2006 (see [RFC4984]) 147 This document specifies the LISP data-plane encapsulation and other 148 xTR functionality while [REF_to_6833bis] specifies the LISP control 149 plane. LISP deployment guidelines can be found in [RFC7215] and 150 [RFC6835] describes considerations for network operational 151 management. Finally, [LISP-INTRO] describes the LISP architecture. 153 2. Requirements Notation 155 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 156 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 157 document are to be interpreted as described in [RFC2119]. 159 3. Definition of Terms 161 Provider-Independent (PI) Addresses: PI addresses are an address 162 block assigned from a pool where blocks are not associated with 163 any particular location in the network (e.g., from a particular 164 service provider) and are therefore not topologically aggregatable 165 in the routing system. 167 Provider-Assigned (PA) Addresses: PA addresses are an address block 168 assigned to a site by each service provider to which a site 169 connects. Typically, each block is a sub-block of a service 170 provider Classless Inter-Domain Routing (CIDR) [RFC4632] block and 171 is aggregated into the larger block before being advertised into 172 the global Internet. Traditionally, IP multihoming has been 173 implemented by each multihomed site acquiring its own globally 174 visible prefix. LISP uses only topologically assigned and 175 aggregatable address blocks for RLOCs, eliminating this 176 demonstrably non-scalable practice. 178 Routing Locator (RLOC): An RLOC is an IPv4 [RFC0791] or IPv6 179 [RFC2460] address of an Egress Tunnel Router (ETR). An RLOC is 180 the output of an EID-to-RLOC mapping lookup. An EID maps to one 181 or more RLOCs. Typically, RLOCs are numbered from topologically 182 aggregatable blocks that are assigned to a site at each point to 183 which it attaches to the global Internet; where the topology is 184 defined by the connectivity of provider networks, RLOCs can be 185 thought of as PA addresses. Multiple RLOCs can be assigned to the 186 same ETR device or to multiple ETR devices at a site. 188 Endpoint ID (EID): An EID is a 32-bit (for IPv4) or 128-bit (for 189 IPv6) value used in the source and destination address fields of 190 the first (most inner) LISP header of a packet. The host obtains 191 a destination EID the same way it obtains a destination address 192 today, for example, through a Domain Name System (DNS) [RFC1034] 193 lookup or Session Initiation Protocol (SIP) [RFC3261] exchange. 194 The source EID is obtained via existing mechanisms used to set a 195 host's "local" IP address. An EID used on the public Internet 196 must have the same properties as any other IP address used in that 197 manner; this means, among other things, that it must be globally 198 unique. An EID is allocated to a host from an EID-Prefix block 199 associated with the site where the host is located. An EID can be 200 used by a host to refer to other hosts. EIDs MUST NOT be used as 201 LISP RLOCs. Note that EID blocks MAY be assigned in a 202 hierarchical manner, independent of the network topology, to 203 facilitate scaling of the mapping database. In addition, an EID 204 block assigned to a site may have site-local structure 205 (subnetting) for routing within the site; this structure is not 206 visible to the global routing system. In theory, the bit string 207 that represents an EID for one device can represent an RLOC for a 208 different device. As the architecture is realized, if a given bit 209 string is both an RLOC and an EID, it must refer to the same 210 entity in both cases. When used in discussions with other 211 Locator/ID separation proposals, a LISP EID will be called an 212 "LEID". Throughout this document, any references to "EID" refer 213 to an LEID. 215 EID-Prefix: An EID-Prefix is a power-of-two block of EIDs that are 216 allocated to a site by an address allocation authority. EID- 217 Prefixes are associated with a set of RLOC addresses that make up 218 a "database mapping". EID-Prefix allocations can be broken up 219 into smaller blocks when an RLOC set is to be associated with the 220 larger EID-Prefix block. A globally routed address block (whether 221 PI or PA) is not inherently an EID-Prefix. A globally routed 222 address block MAY be used by its assignee as an EID block. The 223 converse is not supported. That is, a site that receives an 224 explicitly allocated EID-Prefix may not use that EID-Prefix as a 225 globally routed prefix. This would require coordination and 226 cooperation with the entities managing the mapping infrastructure. 227 Once this has been done, that block could be removed from the 228 globally routed IP system, if other suitable transition and access 229 mechanisms are in place. Discussion of such transition and access 230 mechanisms can be found in [RFC6832] and [RFC7215]. 232 End-system: An end-system is an IPv4 or IPv6 device that originates 233 packets with a single IPv4 or IPv6 header. The end-system 234 supplies an EID value for the destination address field of the IP 235 header when communicating globally (i.e., outside of its routing 236 domain). An end-system can be a host computer, a switch or router 237 device, or any network appliance. 239 Ingress Tunnel Router (ITR): An ITR is a router that resides in a 240 LISP site. Packets sent by sources inside of the LISP site to 241 destinations outside of the site are candidates for encapsulation 242 by the ITR. The ITR treats the IP destination address as an EID 243 and performs an EID-to-RLOC mapping lookup. The router then 244 prepends an "outer" IP header with one of its globally routable 245 RLOCs in the source address field and the result of the mapping 246 lookup in the destination address field. Note that this 247 destination RLOC MAY be an intermediate, proxy device that has 248 better knowledge of the EID-to-RLOC mapping closer to the 249 destination EID. In general, an ITR receives IP packets from site 250 end-systems on one side and sends LISP-encapsulated IP packets 251 toward the Internet on the other side. 253 Specifically, when a service provider prepends a LISP header for 254 Traffic Engineering purposes, the router that does this is also 255 regarded as an ITR. The outer RLOC the ISP ITR uses can be based 256 on the outer destination address (the originating ITR's supplied 257 RLOC) or the inner destination address (the originating host's 258 supplied EID). 260 TE-ITR: A TE-ITR is an ITR that is deployed in a service provider 261 network that prepends an additional LISP header for Traffic 262 Engineering purposes. 264 Egress Tunnel Router (ETR): An ETR is a router that accepts an IP 265 packet where the destination address in the "outer" IP header is 266 one of its own RLOCs. The router strips the "outer" header and 267 forwards the packet based on the next IP header found. In 268 general, an ETR receives LISP-encapsulated IP packets from the 269 Internet on one side and sends decapsulated IP packets to site 270 end-systems on the other side. ETR functionality does not have to 271 be limited to a router device. A server host can be the endpoint 272 of a LISP tunnel as well. 274 TE-ETR: A TE-ETR is an ETR that is deployed in a service provider 275 network that strips an outer LISP header for Traffic Engineering 276 purposes. 278 xTR: An xTR is a reference to an ITR or ETR when direction of data 279 flow is not part of the context description. "xTR" refers to the 280 router that is the tunnel endpoint and is used synonymously with 281 the term "Tunnel Router". For example, "An xTR can be located at 282 the Customer Edge (CE) router" indicates both ITR and ETR 283 functionality at the CE router. 285 LISP Router: A LISP router is a router that performs the functions 286 of any or all of the following: ITR, ETR, Proxy-ITR (PITR), or 287 Proxy-ETR (PETR). 289 EID-to-RLOC Map-Cache: The EID-to-RLOC map-cache is a short-lived, 290 on-demand table in an ITR that stores, tracks, and is responsible 291 for timing out and otherwise validating EID-to-RLOC mappings. 292 This cache is distinct from the full "database" of EID-to-RLOC 293 mappings; it is dynamic, local to the ITR(s), and relatively 294 small, while the database is distributed, relatively static, and 295 much more global in scope. 297 EID-to-RLOC Database: The EID-to-RLOC Database is a global 298 distributed database that contains all known EID-Prefix-to-RLOC 299 mappings. Each potential ETR typically contains a small piece of 300 the database: the EID-to-RLOC mappings for the EID-Prefixes 301 "behind" the router. These map to one of the router's own 302 globally visible IP addresses. The same database mapping entries 303 MUST be configured on all ETRs for a given site. In a steady 304 state, the EID-Prefixes for the site and the Locator-Set for each 305 EID-Prefix MUST be the same on all ETRs. Procedures to enforce 306 and/or verify this are outside the scope of this document. Note 307 that there MAY be transient conditions when the EID-Prefix for the 308 site and Locator-Set for each EID-Prefix may not be the same on 309 all ETRs. This has no negative implications, since a partial set 310 of Locators can be used. 312 Recursive Tunneling: Recursive Tunneling occurs when a packet has 313 more than one LISP IP header. Additional layers of tunneling MAY 314 be employed to implement Traffic Engineering or other re-routing 315 as needed. When this is done, an additional "outer" LISP header 316 is added, and the original RLOCs are preserved in the "inner" 317 header. Any references to tunnels in this specification refer to 318 dynamic encapsulating tunnels; they are never statically 319 configured. 321 Re-encapsulating Tunnels: Re-encapsulating Tunneling occurs when an 322 ETR removes a LISP header, then acts as an ITR to prepend another 323 LISP header. Doing this allows a packet to be re-routed by the 324 re-encapsulating router without adding the overhead of additional 325 tunnel headers. Any references to tunnels in this specification 326 refer to dynamic encapsulating tunnels; they are never statically 327 configured. When using multiple mapping database systems, care 328 must be taken to not create re-encapsulation loops through 329 misconfiguration. 331 LISP Header: LISP header is a term used in this document to refer 332 to the outer IPv4 or IPv6 header, a UDP header, and a LISP- 333 specific 8-octet header that follow the UDP header and that an ITR 334 prepends or an ETR strips. 336 Address Family Identifier (AFI): AFI is a term used to describe an 337 address encoding in a packet. An address family currently 338 pertains to an IPv4 or IPv6 address. See [AFI] and [RFC3232] for 339 details. An AFI value of 0 used in this specification indicates 340 an unspecified encoded address where the length of the address is 341 0 octets following the 16-bit AFI value of 0. 343 Negative Mapping Entry: A negative mapping entry, also known as a 344 negative cache entry, is an EID-to-RLOC entry where an EID-Prefix 345 is advertised or stored with no RLOCs. That is, the Locator-Set 346 for the EID-to-RLOC entry is empty or has an encoded Locator count 347 of 0. This type of entry could be used to describe a prefix from 348 a non-LISP site, which is explicitly not in the mapping database. 349 There are a set of well-defined actions that are encoded in a 350 Negative Map-Reply. 352 Data-Probe: A Data-Probe is a LISP-encapsulated data packet where 353 the inner-header destination address equals the outer-header 354 destination address used to trigger a Map-Reply by a decapsulating 355 ETR. In addition, the original packet is decapsulated and 356 delivered to the destination host if the destination EID is in the 357 EID-Prefix range configured on the ETR. Otherwise, the packet is 358 discarded. A Data-Probe is used in some of the mapping database 359 designs to "probe" or request a Map-Reply from an ETR; in other 360 cases, Map-Requests are used. See each mapping database design 361 for details. When using Data-Probes, by sending Map-Requests on 362 the underlying routing system, EID-Prefixes must be advertised. 363 However, this is discouraged if the core is to scale by having 364 less EID-Prefixes stored in the core router's routing tables. 366 Proxy-ITR (PITR): A PITR is defined and described in [RFC6832]. A 367 PITR acts like an ITR but does so on behalf of non-LISP sites that 368 send packets to destinations at LISP sites. 370 Proxy-ETR (PETR): A PETR is defined and described in [RFC6832]. A 371 PETR acts like an ETR but does so on behalf of LISP sites that 372 send packets to destinations at non-LISP sites. 374 Route-returnability: Route-returnability is an assumption that the 375 underlying routing system will deliver packets to the destination. 376 When combined with a nonce that is provided by a sender and 377 returned by a receiver, this limits off-path data insertion. A 378 route-returnability check is verified when a message is sent with 379 a nonce, another message is returned with the same nonce, and the 380 destination of the original message appears as the source of the 381 returned message. 383 LISP site: LISP site is a set of routers in an edge network that are 384 under a single technical administration. LISP routers that reside 385 in the edge network are the demarcation points to separate the 386 edge network from the core network. 388 Client-side: Client-side is a term used in this document to indicate 389 a connection initiation attempt by an EID. The ITR(s) at the LISP 390 site are the first to get involved in obtaining database Map-Cache 391 entries by sending Map-Request messages. 393 Server-side: Server-side is a term used in this document to indicate 394 that a connection initiation attempt is being accepted for a 395 destination EID. The ETR(s) at the destination LISP site are the 396 first to send Map-Replies to the source site initiating the 397 connection. The ETR(s) at this destination site can obtain 398 mappings by gleaning information from Map-Requests, Data-Probes, 399 or encapsulated packets. 401 Locator-Status-Bits (LSBs): Locator-Status-Bits are present in the 402 LISP header. They are used by ITRs to inform ETRs about the up/ 403 down status of all ETRs at the local site. These bits are used as 404 a hint to convey up/down router status and not path reachability 405 status. The LSBs can be verified by use of one of the Locator 406 reachability algorithms described in Section 10. 408 Anycast Address: Anycast Address is a term used in this document to 409 refer to the same IPv4 or IPv6 address configured and used on 410 multiple systems at the same time. An EID or RLOC can be an 411 anycast address in each of their own address spaces. 413 4. Basic Overview 415 One key concept of LISP is that end-systems operate the same way they 416 do today. The IP addresses that hosts use for tracking sockets and 417 connections, and for sending and receiving packets, do not change. 418 In LISP terminology, these IP addresses are called Endpoint 419 Identifiers (EIDs). 421 Routers continue to forward packets based on IP destination 422 addresses. When a packet is LISP encapsulated, these addresses are 423 referred to as Routing Locators (RLOCs). Most routers along a path 424 between two hosts will not change; they continue to perform routing/ 425 forwarding lookups on the destination addresses. For routers between 426 the source host and the ITR as well as routers from the ETR to the 427 destination host, the destination address is an EID. For the routers 428 between the ITR and the ETR, the destination address is an RLOC. 430 Another key LISP concept is the "Tunnel Router". A Tunnel Router 431 prepends LISP headers on host-originated packets and strips them 432 prior to final delivery to their destination. The IP addresses in 433 this "outer header" are RLOCs. During end-to-end packet exchange 434 between two Internet hosts, an ITR prepends a new LISP header to each 435 packet, and an ETR strips the new header. The ITR performs EID-to- 436 RLOC lookups to determine the routing path to the ETR, which has the 437 RLOC as one of its IP addresses. 439 Some basic rules governing LISP are: 441 o End-systems only send to addresses that are EIDs. They don't know 442 that addresses are EIDs versus RLOCs but assume that packets get 443 to their intended destinations. In a system where LISP is 444 deployed, LISP routers intercept EID-addressed packets and assist 445 in delivering them across the network core where EIDs cannot be 446 routed. The procedure a host uses to send IP packets does not 447 change. 449 o EIDs are typically IP addresses assigned to hosts. 451 o Other types of EID are supported by LISP, see [LCAF] for further 452 information. 454 o LISP routers mostly deal with Routing Locator addresses. See 455 details in Section 4.1 to clarify what is meant by "mostly". 457 o RLOCs are always IP addresses assigned to routers, preferably 458 topologically oriented addresses from provider CIDR (Classless 459 Inter-Domain Routing) blocks. 461 o When a router originates packets, it may use as a source address 462 either an EID or RLOC. When acting as a host (e.g., when 463 terminating a transport session such as Secure SHell (SSH), 464 TELNET, or the Simple Network Management Protocol (SNMP)), it may 465 use an EID that is explicitly assigned for that purpose. An EID 466 that identifies the router as a host MUST NOT be used as an RLOC; 467 an EID is only routable within the scope of a site. A typical BGP 468 configuration might demonstrate this "hybrid" EID/RLOC usage where 469 a router could use its "host-like" EID to terminate iBGP sessions 470 to other routers in a site while at the same time using RLOCs to 471 terminate eBGP sessions to routers outside the site. 473 o Packets with EIDs in them are not expected to be delivered end-to- 474 end in the absence of an EID-to-RLOC mapping operation. They are 475 expected to be used locally for intra-site communication or to be 476 encapsulated for inter-site communication. 478 o EID-Prefixes are likely to be hierarchically assigned in a manner 479 that is optimized for administrative convenience and to facilitate 480 scaling of the EID-to-RLOC mapping database. The hierarchy is 481 based on an address allocation hierarchy that is independent of 482 the network topology. 484 o EIDs may also be structured (subnetted) in a manner suitable for 485 local routing within an Autonomous System (AS). 487 An additional LISP header MAY be prepended to packets by a TE-ITR 488 when re-routing of the path for a packet is desired. A potential 489 use-case for this would be an ISP router that needs to perform 490 Traffic Engineering for packets flowing through its network. In such 491 a situation, termed "Recursive Tunneling", an ISP transit acts as an 492 additional ITR, and the RLOC it uses for the new prepended header 493 would be either a TE-ETR within the ISP (along an intra-ISP traffic 494 engineered path) or a TE-ETR within another ISP (an inter-ISP traffic 495 engineered path, where an agreement to build such a path exists). 497 In order to avoid excessive packet overhead as well as possible 498 encapsulation loops, this document mandates that a maximum of two 499 LISP headers can be prepended to a packet. For initial LISP 500 deployments, it is assumed that two headers is sufficient, where the 501 first prepended header is used at a site for Location/Identity 502 separation and the second prepended header is used inside a service 503 provider for Traffic Engineering purposes. 505 Tunnel Routers can be placed fairly flexibly in a multi-AS topology. 506 For example, the ITR for a particular end-to-end packet exchange 507 might be the first-hop or default router within a site for the source 508 host. Similarly, the ETR might be the last-hop router directly 509 connected to the destination host. Another example, perhaps for a 510 VPN service outsourced to an ISP by a site, the ITR could be the 511 site's border router at the service provider attachment point. 512 Mixing and matching of site-operated, ISP-operated, and other Tunnel 513 Routers is allowed for maximum flexibility. 515 4.1. Packet Flow Sequence 517 This section provides an example of the unicast packet flow, 518 including also control-plane information as specified in 519 [REF_TO_RFC6833bis]. The example also assumes the following 520 conditions: 522 o Source host "host1.abc.example.com" is sending a packet to 523 "host2.xyz.example.com", exactly what host1 would do if the site 524 was not using LISP. 526 o Each site is multihomed, so each Tunnel Router has an address 527 (RLOC) assigned from the service provider address block for each 528 provider to which that particular Tunnel Router is attached. 530 o The ITR(s) and ETR(s) are directly connected to the source and 531 destination, respectively, but the source and destination can be 532 located anywhere in the LISP site. 534 o Map-Requests are sent to the mapping database system by using the 535 LISP control-plane protocol documented in [REF_to_RFC6833bis]. A 536 Map-Request is sent for an external destination when the 537 destination is not found in the forwarding table or matches a 538 default route. 540 o Map-Replies are sent on the underlying routing system topology 541 using the [REF_TO_RFC6833bis] control-plane protocol. 543 Client host1.abc.example.com wants to communicate with server 544 host2.xyz.example.com: 546 1. host1.abc.example.com wants to open a TCP connection to 547 host2.xyz.example.com. It does a DNS lookup on 548 host2.xyz.example.com. An A/AAAA record is returned. This 549 address is the destination EID. The locally assigned address of 550 host1.abc.example.com is used as the source EID. An IPv4 or IPv6 551 packet is built and forwarded through the LISP site as a normal 552 IP packet until it reaches a LISP ITR. 554 2. The LISP ITR must be able to map the destination EID to an RLOC 555 of one of the ETRs at the destination site. The specific method 556 used to do this is not described in this example. See 557 [REF_to_RFC6833bis] for further information. 559 3. The ITR sends a LISP Map-Request as specified in 560 [REF_TO_RFC6833bis]. Map-Requests SHOULD be rate-limited. 562 4. The mapping system helps forwarding the Map-Request to the 563 corresponding ETR. When the Map-Request arrives at one of the 564 ETRs at the destination site, it will process the packet as a 565 control message. 567 5. The ETR looks at the destination EID of the Map-Request and 568 matches it against the prefixes in the ETR's configured EID-to- 569 RLOC mapping database. This is the list of EID-Prefixes the ETR 570 is supporting for the site it resides in. If there is no match, 571 the Map-Request is dropped. Otherwise, a LISP Map-Reply is 572 returned to the ITR. 574 6. The ITR receives the Map-Reply message, parses the message (to 575 check for format validity), and stores the mapping information 576 from the packet. This information is stored in the ITR's EID-to- 577 RLOC map-cache. Note that the map-cache is an on-demand cache. 578 An ITR will manage its map-cache in such a way that optimizes for 579 its resource constraints. 581 7. Subsequent packets from host1.abc.example.com to 582 host2.xyz.example.com will have a LISP header prepended by the 583 ITR using the appropriate RLOC as the LISP header destination 584 address learned from the ETR. Note that the packet MAY be sent 585 to a different ETR than the one that returned the Map-Reply due 586 to the source site's hashing policy or the destination site's 587 Locator-Set policy. 589 8. The ETR receives these packets directly (since the destination 590 address is one of its assigned IP addresses), checks the validity 591 of the addresses, strips the LISP header, and forwards packets to 592 the attached destination host. 594 9. In order to defer the need for a mapping lookup in the reverse 595 direction, an ETR can OPTIONALLY create a cache entry that maps 596 the source EID (inner-header source IP address) to the source 597 RLOC (outer-header source IP address) in a received LISP packet. 598 Such a cache entry is termed a "gleaned" mapping and only 599 contains a single RLOC for the EID in question. More complete 600 information about additional RLOCs SHOULD be verified by sending 601 a LISP Map-Request for that EID. Both the ITR and the ETR may 602 also influence the decision the other makes in selecting an RLOC. 604 5. LISP Encapsulation Details 606 Since additional tunnel headers are prepended, the packet becomes 607 larger and can exceed the MTU of any link traversed from the ITR to 608 the ETR. It is RECOMMENDED in IPv4 that packets do not get 609 fragmented as they are encapsulated by the ITR. Instead, the packet 610 is dropped and an ICMP Too Big message is returned to the source. 612 This specification RECOMMENDS that implementations provide support 613 for one of the proposed fragmentation and reassembly schemes. Two 614 existing schemes are detailed in Section 7. 616 Since IPv4 or IPv6 addresses can be either EIDs or RLOCs, the LISP 617 architecture supports IPv4 EIDs with IPv6 RLOCs (where the inner 618 header is in IPv4 packet format and the outer header is in IPv6 619 packet format) or IPv6 EIDs with IPv4 RLOCs (where the inner header 620 is in IPv6 packet format and the outer header is in IPv4 packet 621 format). The next sub-sections illustrate packet formats for the 622 homogeneous case (IPv4-in-IPv4 and IPv6-in-IPv6), but all 4 623 combinations MUST be supported. Additional types of EIDs are defined 624 in [LCAF]. 626 5.1. LISP IPv4-in-IPv4 Header Format 628 0 1 2 3 629 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 1 630 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 631 / |Version| IHL |Type of Service| Total Length | 632 / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 633 | | Identification |Flags| Fragment Offset | 634 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 635 OH | Time to Live | Protocol = 17 | Header Checksum | 636 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 637 | | Source Routing Locator | 638 \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 639 \ | Destination Routing Locator | 640 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 641 / | Source Port = xxxx | Dest Port = 4341 | 642 UDP +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 643 \ | UDP Length | UDP Checksum | 644 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 645 L |N|L|E|V|I|R|K|K| Nonce/Map-Version | 646 I \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 647 S / | Instance ID/Locator-Status-Bits | 648 P +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 649 / |Version| IHL |Type of Service| Total Length | 650 / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 651 | | Identification |Flags| Fragment Offset | 652 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 653 IH | Time to Live | Protocol | Header Checksum | 654 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 655 | | Source EID | 656 \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 657 \ | Destination EID | 658 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 660 IHL = IP-Header-Length 662 5.2. LISP IPv6-in-IPv6 Header Format 664 0 1 2 3 665 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 1 666 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 667 / |Version| Traffic Class | Flow Label | 668 / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 669 | | Payload Length | Next Header=17| Hop Limit | 670 v +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 671 | | 672 O + + 673 u | | 674 t + Source Routing Locator + 675 e | | 676 r + + 677 | | 678 H +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 679 d | | 680 r + + 681 | | 682 ^ + Destination Routing Locator + 683 | | | 684 \ + + 685 \ | | 686 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 687 / | Source Port = xxxx | Dest Port = 4341 | 688 UDP +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 689 \ | UDP Length | UDP Checksum | 690 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 691 L |N|L|E|V|I|R|K|K| Nonce/Map-Version | 692 I \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 693 S / | Instance ID/Locator-Status-Bits | 694 P +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 695 / |Version| Traffic Class | Flow Label | 696 / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 697 / | Payload Length | Next Header | Hop Limit | 698 v +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 699 | | 700 I + + 701 n | | 702 n + Source EID + 703 e | | 704 r + + 705 | | 706 H +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 707 d | | 708 r + + 709 | | 711 ^ + Destination EID + 712 \ | | 713 \ + + 714 \ | | 715 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 717 5.3. Tunnel Header Field Descriptions 719 Inner Header (IH): The inner header is the header on the datagram 720 received from the originating host. The source and destination IP 721 addresses are EIDs [RFC0791] [RFC2460]. 723 Outer Header: (OH) The outer header is a new header prepended by an 724 ITR. The address fields contain RLOCs obtained from the ingress 725 router's EID-to-RLOC Cache. The IP protocol number is "UDP (17)" 726 from [RFC0768]. The setting of the Don't Fragment (DF) bit 727 'Flags' field is according to rules listed in Sections 7.1 and 728 7.2. 730 UDP Header: The UDP header contains an ITR selected source port when 731 encapsulating a packet. See Section 12 for details on the hash 732 algorithm used to select a source port based on the 5-tuple of the 733 inner header. The destination port MUST be set to the well-known 734 IANA-assigned port value 4341. 736 UDP Checksum: The 'UDP Checksum' field SHOULD be transmitted as zero 737 by an ITR for either IPv4 [RFC0768] or IPv6 encapsulation 738 [UDP-TUNNELS] [UDP-ZERO]. When a packet with a zero UDP checksum 739 is received by an ETR, the ETR MUST accept the packet for 740 decapsulation. When an ITR transmits a non-zero value for the UDP 741 checksum, it MUST send a correctly computed value in this field. 742 When an ETR receives a packet with a non-zero UDP checksum, it MAY 743 choose to verify the checksum value. If it chooses to perform 744 such verification, and the verification fails, the packet MUST be 745 silently dropped. If the ETR chooses not to perform the 746 verification, or performs the verification successfully, the 747 packet MUST be accepted for decapsulation. The handling of UDP 748 checksums for all tunneling protocols, including LISP, is under 749 active discussion within the IETF. When that discussion 750 concludes, any necessary changes will be made to align LISP with 751 the outcome of the broader discussion. 753 UDP Length: The 'UDP Length' field is set for an IPv4-encapsulated 754 packet to be the sum of the inner-header IPv4 Total Length plus 755 the UDP and LISP header lengths. For an IPv6-encapsulated packet, 756 the 'UDP Length' field is the sum of the inner-header IPv6 Payload 757 Length, the size of the IPv6 header (40 octets), and the size of 758 the UDP and LISP headers. 760 N: The N-bit is the nonce-present bit. When this bit is set to 1, 761 the low-order 24 bits of the first 32 bits of the LISP header 762 contain a Nonce. See Section 10.1 for details. Both N- and 763 V-bits MUST NOT be set in the same packet. If they are, a 764 decapsulating ETR MUST treat the 'Nonce/Map-Version' field as 765 having a Nonce value present. 767 L: The L-bit is the 'Locator-Status-Bits' field enabled bit. When 768 this bit is set to 1, the Locator-Status-Bits in the second 769 32 bits of the LISP header are in use. 771 x 1 x x 0 x x x 772 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 773 |N|L|E|V|I|R|K|K| Nonce/Map-Version | 774 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 775 | Locator-Status-Bits | 776 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 778 E: The E-bit is the echo-nonce-request bit. This bit MUST be ignored 779 and has no meaning when the N-bit is set to 0. When the N-bit is 780 set to 1 and this bit is set to 1, an ITR is requesting that the 781 nonce value in the 'Nonce' field be echoed back in LISP- 782 encapsulated packets when the ITR is also an ETR. See 783 Section 10.1 for details. 785 V: The V-bit is the Map-Version present bit. When this bit is set to 786 1, the N-bit MUST be 0. Refer to Section 13.3 for more details. 787 This bit indicates that the LISP header is encoded in this 788 case as: 790 0 x 0 1 x x x x 791 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 792 |N|L|E|V|I|R|K|K| Source Map-Version | Dest Map-Version | 793 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 794 | Instance ID/Locator-Status-Bits | 795 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 797 I: The I-bit is the Instance ID bit. See Section 8 for more details. 798 When this bit is set to 1, the 'Locator-Status-Bits' field is 799 reduced to 8 bits and the high-order 24 bits are used as an 800 Instance ID. If the L-bit is set to 0, then the low-order 8 bits 801 are transmitted as zero and ignored on receipt. The format of the 802 LISP header would look like this: 804 x x x x 1 x x x 805 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 806 |N|L|E|V|I|R|K|K| Nonce/Map-Version | 807 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 808 | Instance ID | LSBs | 809 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 811 R: The R-bit is a Reserved bit for future use. It MUST be set to 0 812 on transmit and MUST be ignored on receipt. 814 KK: The KK-bits are a 2-bit field used when encapsualted packets are 815 encrypted. The field is set to 00 when the packet is not 816 encrypted. See [I.D-ietf-lisp-crypto] for further information. 818 LISP Nonce: The LISP 'Nonce' field is a 24-bit value that is 819 randomly generated by an ITR when the N-bit is set to 1. Nonce 820 generation algorithms are an implementation matter but are 821 required to generate different nonces when sending to different 822 destinations. However, the same nonce can be used for a period of 823 time to the same destination. The nonce is also used when the 824 E-bit is set to request the nonce value to be echoed by the other 825 side when packets are returned. When the E-bit is clear but the 826 N-bit is set, a remote ITR is either echoing a previously 827 requested echo-nonce or providing a random nonce. See 828 Section 10.1 for more details. 830 LISP Locator-Status-Bits (LSBs): When the L-bit is also set, the 831 'Locator-Status-Bits' field in the LISP header is set by an ITR to 832 indicate to an ETR the up/down status of the Locators in the 833 source site. Each RLOC in a Map-Reply is assigned an ordinal 834 value from 0 to n-1 (when there are n RLOCs in a mapping entry). 835 The Locator-Status-Bits are numbered from 0 to n-1 from the least 836 significant bit of the field. The field is 32 bits when the I-bit 837 is set to 0 and is 8 bits when the I-bit is set to 1. When a 838 Locator-Status-Bit is set to 1, the ITR is indicating to the ETR 839 that the RLOC associated with the bit ordinal has up status. See 840 Section 10 for details on how an ITR can determine the status of 841 the ETRs at the same site. When a site has multiple EID-Prefixes 842 that result in multiple mappings (where each could have a 843 different Locator-Set), the Locator-Status-Bits setting in an 844 encapsulated packet MUST reflect the mapping for the EID-Prefix 845 that the inner-header source EID address matches. If the LSB for 846 an anycast Locator is set to 1, then there is at least one RLOC 847 with that address, and the ETR is considered 'up'. 849 When doing ITR/PITR encapsulation: 851 o The outer-header 'Time to Live' field (or 'Hop Limit' field, in 852 the case of IPv6) SHOULD be copied from the inner-header 'Time to 853 Live' field. 855 o The outer-header 'Type of Service' field (or the 'Traffic Class' 856 field, in the case of IPv6) SHOULD be copied from the inner-header 857 'Type of Service' field (with one exception; see below). 859 When doing ETR/PETR decapsulation: 861 o The inner-header 'Time to Live' field (or 'Hop Limit' field, in 862 the case of IPv6) SHOULD be copied from the outer-header 'Time to 863 Live' field, when the Time to Live value of the outer header is 864 less than the Time to Live value of the inner header. Failing to 865 perform this check can cause the Time to Live of the inner header 866 to increment across encapsulation/decapsulation cycles. This 867 check is also performed when doing initial encapsulation, when a 868 packet comes to an ITR or PITR destined for a LISP site. 870 o The inner-header 'Type of Service' field (or the 'Traffic Class' 871 field, in the case of IPv6) SHOULD be copied from the outer-header 872 'Type of Service' field (with one exception; see below). 874 Note that if an ETR/PETR is also an ITR/PITR and chooses to re- 875 encapsulate after decapsulating, the net effect of this is that the 876 new outer header will carry the same Time to Live as the old outer 877 header minus 1. 879 Copying the Time to Live (TTL) serves two purposes: first, it 880 preserves the distance the host intended the packet to travel; 881 second, and more importantly, it provides for suppression of looping 882 packets in the event there is a loop of concatenated tunnels due to 883 misconfiguration. See Section 18.3 for TTL exception handling for 884 traceroute packets. 886 The Explicit Congestion Notification ('ECN') field occupies bits 6 887 and 7 of both the IPv4 'Type of Service' field and the IPv6 'Traffic 888 Class' field [RFC3168]. The 'ECN' field requires special treatment 889 in order to avoid discarding indications of congestion [RFC3168]. 890 ITR encapsulation MUST copy the 2-bit 'ECN' field from the inner 891 header to the outer header. Re-encapsulation MUST copy the 2-bit 892 'ECN' field from the stripped outer header to the new outer header. 893 If the 'ECN' field contains a congestion indication codepoint (the 894 value is '11', the Congestion Experienced (CE) codepoint), then ETR 895 decapsulation MUST copy the 2-bit 'ECN' field from the stripped outer 896 header to the surviving inner header that is used to forward the 897 packet beyond the ETR. These requirements preserve CE indications 898 when a packet that uses ECN traverses a LISP tunnel and becomes 899 marked with a CE indication due to congestion between the tunnel 900 endpoints. 902 6. LISP EID-to-RLOC Map-Cache 904 ITRs and PITRs maintain an on-demand cache, referred as LISP EID-to- 905 RLOC Map-Cache, that contains mappings from EID-prefixes to locator 906 sets. The cache is used to encapsulate packets from the EID space to 907 the corresponding RLOC network attachment point. 909 When an ITR/PITR receives a packet from inside of the LISP site to 910 destinations outside of the site a longest-prefix match lookup of the 911 EID is done to the map-cache. 913 When the lookup succeeds, the locator-set retrieved from the map- 914 cache is used to send the packet to the EID's topological location. 916 If the lookup fails, the ITR/PITR needs to retrieve the mapping using 917 the LISP control-plane protocol [REF_TO_RFC6833bis]. The mapping is 918 then stored in the local map-cache to forward subsequent packets 919 addressed to the same EID-prefix. 921 The map-cache is a local cache of mappings, entries are expired based 922 on the associated Time to live. In addition, entries can be updated 923 with more current information, see Section 13 for further information 924 on this. Finally, the map-cache also contains reachability 925 information about EIDs and RLOCs, and uses LISP reachability 926 information mechanisms to determine the reachability of RLOCs, see 927 Section 10 for the specific mechanisms. 929 7. Dealing with Large Encapsulated Packets 931 This section proposes two mechanisms to deal with packets that exceed 932 the path MTU between the ITR and ETR. 934 It is left to the implementor to decide if the stateless or stateful 935 mechanism should be implemented. Both or neither can be used, since 936 it is a local decision in the ITR regarding how to deal with MTU 937 issues, and sites can interoperate with differing mechanisms. 939 Both stateless and stateful mechanisms also apply to Re-encapsulating 940 and Recursive Tunneling, so any actions below referring to an ITR 941 also apply to a TE-ITR. 943 7.1. A Stateless Solution to MTU Handling 945 An ITR stateless solution to handle MTU issues is described as 946 follows: 948 1. Define H to be the size, in octets, of the outer header an ITR 949 prepends to a packet. This includes the UDP and LISP header 950 lengths. 952 2. Define L to be the size, in octets, of the maximum-sized packet 953 an ITR can send to an ETR without the need for the ITR or any 954 intermediate routers to fragment the packet. 956 3. Define an architectural constant S for the maximum size of a 957 packet, in octets, an ITR must receive from the source so the 958 effective MTU can be met. That is, L = S + H. 960 When an ITR receives a packet from a site-facing interface and adds H 961 octets worth of encapsulation to yield a packet size greater than L 962 octets (meaning the received packet size was greater than S octets 963 from the source), it resolves the MTU issue by first splitting the 964 original packet into 2 equal-sized fragments. A LISP header is then 965 prepended to each fragment. The size of the encapsulated fragments 966 is then (S/2 + H), which is less than the ITR's estimate of the path 967 MTU between the ITR and its correspondent ETR. 969 When an ETR receives encapsulated fragments, it treats them as two 970 individually encapsulated packets. It strips the LISP headers and 971 then forwards each fragment to the destination host of the 972 destination site. The two fragments are reassembled at the 973 destination host into the single IP datagram that was originated by 974 the source host. Note that reassembly can happen at the ETR if the 975 encapsulated packet was fragmented at or after the ITR. 977 This behavior is performed by the ITR when the source host originates 978 a packet with the 'DF' field of the IP header set to 0. When the 979 'DF' field of the IP header is set to 1, or the packet is an IPv6 980 packet originated by the source host, the ITR will drop the packet 981 when the size is greater than L and send an ICMP Too Big message to 982 the source with a value of S, where S is (L - H). 984 When the outer-header encapsulation uses an IPv4 header, an 985 implementation SHOULD set the DF bit to 1 so ETR fragment reassembly 986 can be avoided. An implementation MAY set the DF bit in such headers 987 to 0 if it has good reason to believe there are unresolvable path MTU 988 issues between the sending ITR and the receiving ETR. 990 This specification RECOMMENDS that L be defined as 1500. 992 7.2. A Stateful Solution to MTU Handling 994 An ITR stateful solution to handle MTU issues is described as follows 995 and was first introduced in [OPENLISP]: 997 1. The ITR will keep state of the effective MTU for each Locator per 998 Map-Cache entry. The effective MTU is what the core network can 999 deliver along the path between the ITR and ETR. 1001 2. When an IPv6-encapsulated packet, or an IPv4-encapsulated packet 1002 with the DF bit set to 1, exceeds what the core network can 1003 deliver, one of the intermediate routers on the path will send an 1004 ICMP Too Big message to the ITR. The ITR will parse the ICMP 1005 message to determine which Locator is affected by the effective 1006 MTU change and then record the new effective MTU value in the 1007 Map-Cache entry. 1009 3. When a packet is received by the ITR from a source inside of the 1010 site and the size of the packet is greater than the effective MTU 1011 stored with the Map-Cache entry associated with the destination 1012 EID the packet is for, the ITR will send an ICMP Too Big message 1013 back to the source. The packet size advertised by the ITR in the 1014 ICMP Too Big message is the effective MTU minus the LISP 1015 encapsulation length. 1017 Even though this mechanism is stateful, it has advantages over the 1018 stateless IP fragmentation mechanism, by not involving the 1019 destination host with reassembly of ITR fragmented packets. 1021 8. Using Virtualization and Segmentation with LISP 1023 FIXME: Indicate how the instance-ID is 32 bits in control-plane and 1024 24-bits in the data-plane. 1026 When multiple organizations inside of a LISP site are using private 1027 addresses [RFC1918] as EID-Prefixes, their address spaces MUST remain 1028 segregated due to possible address duplication. An Instance ID in 1029 the address encoding can aid in making the entire AFI-based address 1030 unique. See IANA Considerations (Section 21.2) for details on 1031 possible address encodings. 1033 An Instance ID can be carried in a LISP-encapsulated packet. An ITR 1034 that prepends a LISP header will copy a 24-bit value used by the LISP 1035 router to uniquely identify the address space. The value is copied 1036 to the 'Instance ID' field of the LISP header, and the I-bit is set 1037 to 1. 1039 When an ETR decapsulates a packet, the Instance ID from the LISP 1040 header is used as a table identifier to locate the forwarding table 1041 to use for the inner destination EID lookup. 1043 For example, an 802.1Q VLAN tag or VPN identifier could be used as a 1044 24-bit Instance ID. 1046 9. Routing Locator Selection 1048 Both the client-side and server-side may need control over the 1049 selection of RLOCs for conversations between them. This control is 1050 achieved by manipulating the 'Priority' and 'Weight' fields in EID- 1051 to-RLOC Map-Reply messages. Alternatively, RLOC information MAY be 1052 gleaned from received tunneled packets or EID-to-RLOC Map-Request 1053 messages. 1055 The following are different scenarios for choosing RLOCs and the 1056 controls that are available: 1058 o The server-side returns one RLOC. The client-side can only use 1059 one RLOC. The server-side has complete control of the selection. 1061 o The server-side returns a list of RLOCs where a subset of the list 1062 has the same best Priority. The client can only use the subset 1063 list according to the weighting assigned by the server-side. In 1064 this case, the server-side controls both the subset list and load- 1065 splitting across its members. The client-side can use RLOCs 1066 outside of the subset list if it determines that the subset list 1067 is unreachable (unless RLOCs are set to a Priority of 255). Some 1068 sharing of control exists: the server-side determines the 1069 destination RLOC list and load distribution while the client-side 1070 has the option of using alternatives to this list if RLOCs in the 1071 list are unreachable. 1073 o The server-side sets a Weight of 0 for the RLOC subset list. In 1074 this case, the client-side can choose how the traffic load is 1075 spread across the subset list. Control is shared by the server- 1076 side determining the list and the client determining load 1077 distribution. Again, the client can use alternative RLOCs if the 1078 server-provided list of RLOCs is unreachable. 1080 o Either side (more likely the server-side ETR) decides not to send 1081 a Map-Request. For example, if the server-side ETR does not send 1082 Map-Requests, it gleans RLOCs from the client-side ITR, giving the 1083 client-side ITR responsibility for bidirectional RLOC reachability 1084 and preferability. Server-side ETR gleaning of the client-side 1085 ITR RLOC is done by caching the inner-header source EID and the 1086 outer-header source RLOC of received packets. The client-side ITR 1087 controls how traffic is returned and can alternate using an outer- 1088 header source RLOC, which then can be added to the list the 1089 server-side ETR uses to return traffic. Since no Priority or 1090 Weights are provided using this method, the server-side ETR MUST 1091 assume that each client-side ITR RLOC uses the same best Priority 1092 with a Weight of zero. In addition, since EID-Prefix encoding 1093 cannot be conveyed in data packets, the EID-to-RLOC Cache on 1094 Tunnel Routers can grow to be very large. 1096 o A "gleaned" Map-Cache entry, one learned from the source RLOC of a 1097 received encapsulated packet, is only stored and used for a few 1098 seconds, pending verification. Verification is performed by 1099 sending a Map-Request to the source EID (the inner-header IP 1100 source address) of the received encapsulated packet. A reply to 1101 this "verifying Map-Request" is used to fully populate the Map- 1102 Cache entry for the "gleaned" EID and is stored and used for the 1103 time indicated from the 'TTL' field of a received Map-Reply. When 1104 a verified Map-Cache entry is stored, data gleaning no longer 1105 occurs for subsequent packets that have a source EID that matches 1106 the EID-Prefix of the verified entry. This "gleaning" mechanism 1107 is OPTIONAL. 1109 RLOCs that appear in EID-to-RLOC Map-Reply messages are assumed to be 1110 reachable when the R-bit for the Locator record is set to 1. When 1111 the R-bit is set to 0, an ITR or PITR MUST NOT encapsulate to the 1112 RLOC. Neither the information contained in a Map-Reply nor that 1113 stored in the mapping database system provides reachability 1114 information for RLOCs. Note that reachability is not part of the 1115 mapping system and is determined using one or more of the Routing 1116 Locator reachability algorithms described in the next section. 1118 10. Routing Locator Reachability 1120 Several mechanisms for determining RLOC reachability are currently 1121 defined: 1123 1. An ETR may examine the Locator-Status-Bits in the LISP header of 1124 an encapsulated data packet received from an ITR. If the ETR is 1125 also acting as an ITR and has traffic to return to the original 1126 ITR site, it can use this status information to help select an 1127 RLOC. 1129 2. An ITR may receive an ICMP Network Unreachable or Host 1130 Unreachable message for an RLOC it is using. This indicates that 1131 the RLOC is likely down. Note that trusting ICMP messages may 1132 not be desirable, but neither is ignoring them completely. 1133 Implementations are encouraged to follow current best practices 1134 in treating these conditions. 1136 3. An ITR that participates in the global routing system can 1137 determine that an RLOC is down if no BGP Routing Information Base 1138 (RIB) route exists that matches the RLOC IP address. 1140 4. An ITR may receive an ICMP Port Unreachable message from a 1141 destination host. This occurs if an ITR attempts to use 1142 interworking [RFC6832] and LISP-encapsulated data is sent to a 1143 non-LISP-capable site. 1145 5. An ITR may receive a Map-Reply from an ETR in response to a 1146 previously sent Map-Request. The RLOC source of the Map-Reply is 1147 likely up, since the ETR was able to send the Map-Reply to the 1148 ITR. 1150 6. When an ETR receives an encapsulated packet from an ITR, the 1151 source RLOC from the outer header of the packet is likely up. 1153 7. An ITR/ETR pair can use the Locator reachability algorithms 1154 described in this section, namely Echo-Noncing or RLOC-Probing. 1156 When determining Locator up/down reachability by examining the 1157 Locator-Status-Bits from the LISP-encapsulated data packet, an ETR 1158 will receive up-to-date status from an encapsulating ITR about 1159 reachability for all ETRs at the site. CE-based ITRs at the source 1160 site can determine reachability relative to each other using the site 1161 IGP as follows: 1163 o Under normal circumstances, each ITR will advertise a default 1164 route into the site IGP. 1166 o If an ITR fails or if the upstream link to its PE fails, its 1167 default route will either time out or be withdrawn. 1169 Each ITR can thus observe the presence or lack of a default route 1170 originated by the others to determine the Locator-Status-Bits it sets 1171 for them. 1173 RLOCs listed in a Map-Reply are numbered with ordinals 0 to n-1. The 1174 Locator-Status-Bits in a LISP-encapsulated packet are numbered from 0 1175 to n-1 starting with the least significant bit. For example, if an 1176 RLOC listed in the 3rd position of the Map-Reply goes down (ordinal 1177 value 2), then all ITRs at the site will clear the 3rd least 1178 significant bit (xxxx x0xx) of the 'Locator-Status-Bits' field for 1179 the packets they encapsulate. 1181 When an ETR decapsulates a packet, it will check for any change in 1182 the 'Locator-Status-Bits' field. When a bit goes from 1 to 0, the 1183 ETR, if acting also as an ITR, will refrain from encapsulating 1184 packets to an RLOC that is indicated as down. It will only resume 1185 using that RLOC if the corresponding Locator-Status-Bit returns to a 1186 value of 1. Locator-Status-Bits are associated with a Locator-Set 1187 per EID-Prefix. Therefore, when a Locator becomes unreachable, the 1188 Locator-Status-Bit that corresponds to that Locator's position in the 1189 list returned by the last Map-Reply will be set to zero for that 1190 particular EID-Prefix. 1192 When ITRs at the site are not deployed in CE routers, the IGP can 1193 still be used to determine the reachability of Locators, provided 1194 they are injected into the IGP. This is typically done when a /32 1195 address is configured on a loopback interface. 1197 When ITRs receive ICMP Network Unreachable or Host Unreachable 1198 messages as a method to determine unreachability, they will refrain 1199 from using Locators that are described in Locator lists of Map- 1200 Replies. However, using this approach is unreliable because many 1201 network operators turn off generation of ICMP Destination Unreachable 1202 messages. 1204 If an ITR does receive an ICMP Network Unreachable or Host 1205 Unreachable message, it MAY originate its own ICMP Destination 1206 Unreachable message destined for the host that originated the data 1207 packet the ITR encapsulated. 1209 Also, BGP-enabled ITRs can unilaterally examine the RIB to see if a 1210 locator address from a Locator-Set in a mapping entry matches a 1211 prefix. If it does not find one and BGP is running in the Default- 1212 Free Zone (DFZ), it can decide to not use the Locator even though the 1213 Locator-Status-Bits indicate that the Locator is up. In this case, 1214 the path from the ITR to the ETR that is assigned the Locator is not 1215 available. More details are in [LOC-ID-ARCH]. 1217 Optionally, an ITR can send a Map-Request to a Locator, and if a Map- 1218 Reply is returned, reachability of the Locator has been determined. 1219 Obviously, sending such probes increases the number of control 1220 messages originated by Tunnel Routers for active flows, so Locators 1221 are assumed to be reachable when they are advertised. 1223 This assumption does create a dependency: Locator unreachability is 1224 detected by the receipt of ICMP Host Unreachable messages. When a 1225 Locator has been determined to be unreachable, it is not used for 1226 active traffic; this is the same as if it were listed in a Map-Reply 1227 with Priority 255. 1229 The ITR can test the reachability of the unreachable Locator by 1230 sending periodic Requests. Both Requests and Replies MUST be rate- 1231 limited. Locator reachability testing is never done with data 1232 packets, since that increases the risk of packet loss for end-to-end 1233 sessions. 1235 When an ETR decapsulates a packet, it knows that it is reachable from 1236 the encapsulating ITR because that is how the packet arrived. In 1237 most cases, the ETR can also reach the ITR but cannot assume this to 1238 be true, due to the possibility of path asymmetry. In the presence 1239 of unidirectional traffic flow from an ITR to an ETR, the ITR SHOULD 1240 NOT use the lack of return traffic as an indication that the ETR is 1241 unreachable. Instead, it MUST use an alternate mechanism to 1242 determine reachability. 1244 10.1. Echo Nonce Algorithm 1246 When data flows bidirectionally between Locators from different 1247 sites, a data-plane mechanism called "nonce echoing" can be used to 1248 determine reachability between an ITR and ETR. When an ITR wants to 1249 solicit a nonce echo, it sets the N- and E-bits and places a 24-bit 1250 nonce [RFC4086] in the LISP header of the next encapsulated data 1251 packet. 1253 When this packet is received by the ETR, the encapsulated packet is 1254 forwarded as normal. When the ETR next sends a data packet to the 1255 ITR, it includes the nonce received earlier with the N-bit set and 1256 E-bit cleared. The ITR sees this "echoed nonce" and knows that the 1257 path to and from the ETR is up. 1259 The ITR will set the E-bit and N-bit for every packet it sends while 1260 in the echo-nonce-request state. The time the ITR waits to process 1261 the echoed nonce before it determines the path is unreachable is 1262 variable and is a choice left for the implementation. 1264 If the ITR is receiving packets from the ETR but does not see the 1265 nonce echoed while being in the echo-nonce-request state, then the 1266 path to the ETR is unreachable. This decision may be overridden by 1267 other Locator reachability algorithms. Once the ITR determines that 1268 the path to the ETR is down, it can switch to another Locator for 1269 that EID-Prefix. 1271 Note that "ITR" and "ETR" are relative terms here. Both devices MUST 1272 be implementing both ITR and ETR functionality for the echo nonce 1273 mechanism to operate. 1275 The ITR and ETR may both go into the echo-nonce-request state at the 1276 same time. The number of packets sent or the time during which echo 1277 nonce requests are sent is an implementation-specific setting. 1278 However, when an ITR is in the echo-nonce-request state, it can echo 1279 the ETR's nonce in the next set of packets that it encapsulates and 1280 subsequently continue sending echo-nonce-request packets. 1282 This mechanism does not completely solve the forward path 1283 reachability problem, as traffic may be unidirectional. That is, the 1284 ETR receiving traffic at a site may not be the same device as an ITR 1285 that transmits traffic from that site, or the site-to-site traffic is 1286 unidirectional so there is no ITR returning traffic. 1288 The echo-nonce algorithm is bilateral. That is, if one side sets the 1289 E-bit and the other side is not enabled for echo-noncing, then the 1290 echoing of the nonce does not occur and the requesting side may 1291 erroneously consider the Locator unreachable. An ITR SHOULD only set 1292 the E-bit in an encapsulated data packet when it knows the ETR is 1293 enabled for echo-noncing. This is conveyed by the E-bit in the Map- 1294 Reply message. 1296 Note that other Locator reachability mechanisms are being researched 1297 and can be used to compliment or even override the echo nonce 1298 algorithm. See the next section for an example of control-plane 1299 probing. 1301 10.2. RLOC-Probing Algorithm 1303 RLOC-Probing is a method that an ITR or PITR can use to determine the 1304 reachability status of one or more Locators that it has cached in a 1305 Map-Cache entry. The probe-bit of the Map-Request and Map-Reply 1306 messages is used for RLOC-Probing. 1308 RLOC-Probing is done in the control plane on a timer basis, where an 1309 ITR or PITR will originate a Map-Request destined to a locator 1310 address from one of its own locator addresses. A Map-Request used as 1311 an RLOC-probe is NOT encapsulated and NOT sent to a Map-Server or to 1312 the mapping database system as one would when soliciting mapping 1313 data. The EID record encoded in the Map-Request is the EID-Prefix of 1314 the Map-Cache entry cached by the ITR or PITR. The ITR may include a 1315 mapping data record for its own database mapping information that 1316 contains the local EID-Prefixes and RLOCs for its site. RLOC-probes 1317 are sent periodically using a jittered timer interval. 1319 When an ETR receives a Map-Request message with the probe-bit set, it 1320 returns a Map-Reply with the probe-bit set. The source address of 1321 the Map-Reply is set according to the procedure described in 1322 [REF_TO_RFC6833bis]. The Map-Reply SHOULD contain mapping data for 1323 the EID-Prefix contained in the Map-Request. This provides the 1324 opportunity for the ITR or PITR that sent the RLOC-probe to get 1325 mapping updates if there were changes to the ETR's database mapping 1326 entries. 1328 There are advantages and disadvantages of RLOC-Probing. The greatest 1329 benefit of RLOC-Probing is that it can handle many failure scenarios 1330 allowing the ITR to determine when the path to a specific Locator is 1331 reachable or has become unreachable, thus providing a robust 1332 mechanism for switching to using another Locator from the cached 1333 Locator. RLOC-Probing can also provide rough Round-Trip Time (RTT) 1334 estimates between a pair of Locators, which can be useful for network 1335 management purposes as well as for selecting low delay paths. The 1336 major disadvantage of RLOC-Probing is in the number of control 1337 messages required and the amount of bandwidth used to obtain those 1338 benefits, especially if the requirement for failure detection times 1339 is very small. 1341 Continued research and testing will attempt to characterize the 1342 tradeoffs of failure detection times versus message overhead. 1344 11. EID Reachability within a LISP Site 1346 A site may be multihomed using two or more ETRs. The hosts and 1347 infrastructure within a site will be addressed using one or more EID- 1348 Prefixes that are mapped to the RLOCs of the relevant ETRs in the 1349 mapping system. One possible failure mode is for an ETR to lose 1350 reachability to one or more of the EID-Prefixes within its own site. 1351 When this occurs when the ETR sends Map-Replies, it can clear the 1352 R-bit associated with its own Locator. And when the ETR is also an 1353 ITR, it can clear its Locator-Status-Bit in the encapsulation data 1354 header. 1356 It is recognized that there are no simple solutions to the site 1357 partitioning problem because it is hard to know which part of the 1358 EID-Prefix range is partitioned and which Locators can reach any sub- 1359 ranges of the EID-Prefixes. This problem is under investigation with 1360 the expectation that experiments will tell us more. Note that this 1361 is not a new problem introduced by the LISP architecture. The 1362 problem exists today when a multihomed site uses BGP to advertise its 1363 reachability upstream. 1365 12. Routing Locator Hashing 1367 When an ETR provides an EID-to-RLOC mapping in a Map-Reply message to 1368 a requesting ITR, the Locator-Set for the EID-Prefix may contain 1369 different Priority values for each locator address. When more than 1370 one best Priority Locator exists, the ITR can decide how to load- 1371 share traffic against the corresponding Locators. 1373 The following hash algorithm may be used by an ITR to select a 1374 Locator for a packet destined to an EID for the EID-to-RLOC mapping: 1376 1. Either a source and destination address hash or the traditional 1377 5-tuple hash can be used. The traditional 5-tuple hash includes 1378 the source and destination addresses; source and destination TCP, 1379 UDP, or Stream Control Transmission Protocol (SCTP) port numbers; 1380 and the IP protocol number field or IPv6 next-protocol fields of 1381 a packet that a host originates from within a LISP site. When a 1382 packet is not a TCP, UDP, or SCTP packet, the source and 1383 destination addresses only from the header are used to compute 1384 the hash. 1386 2. Take the hash value and divide it by the number of Locators 1387 stored in the Locator-Set for the EID-to-RLOC mapping. 1389 3. The remainder will yield a value of 0 to "number of Locators 1390 minus 1". Use the remainder to select the Locator in the 1391 Locator-Set. 1393 Note that when a packet is LISP encapsulated, the source port number 1394 in the outer UDP header needs to be set. Selecting a hashed value 1395 allows core routers that are attached to Link Aggregation Groups 1396 (LAGs) to load-split the encapsulated packets across member links of 1397 such LAGs. Otherwise, core routers would see a single flow, since 1398 packets have a source address of the ITR, for packets that are 1399 originated by different EIDs at the source site. A suggested setting 1400 for the source port number computed by an ITR is a 5-tuple hash 1401 function on the inner header, as described above. 1403 Many core router implementations use a 5-tuple hash to decide how to 1404 balance packet load across members of a LAG. The 5-tuple hash 1405 includes the source and destination addresses of the packet and the 1406 source and destination ports when the protocol number in the packet 1407 is TCP or UDP. For this reason, UDP encoding is used for LISP 1408 encapsulation. 1410 13. Changing the Contents of EID-to-RLOC Mappings 1412 Since the LISP architecture uses a caching scheme to retrieve and 1413 store EID-to-RLOC mappings, the only way an ITR can get a more up-to- 1414 date mapping is to re-request the mapping. However, the ITRs do not 1415 know when the mappings change, and the ETRs do not keep track of 1416 which ITRs requested its mappings. For scalability reasons, we want 1417 to maintain this approach but need to provide a way for ETRs to 1418 change their mappings and inform the sites that are currently 1419 communicating with the ETR site using such mappings. 1421 When adding a new Locator record in lexicographic order to the end of 1422 a Locator-Set, it is easy to update mappings. We assume that new 1423 mappings will maintain the same Locator ordering as the old mapping 1424 but will just have new Locators appended to the end of the list. So, 1425 some ITRs can have a new mapping while other ITRs have only an old 1426 mapping that is used until they time out. When an ITR has only an 1427 old mapping but detects bits set in the Locator-Status-Bits that 1428 correspond to Locators beyond the list it has cached, it simply 1429 ignores them. However, this can only happen for locator addresses 1430 that are lexicographically greater than the locator addresses in the 1431 existing Locator-Set. 1433 When a Locator record is inserted in the middle of a Locator-Set, to 1434 maintain lexicographic order, the SMR procedure in Section 13.2 is 1435 used to inform ITRs and PITRs of the new Locator-Status-Bit mappings. 1437 When a Locator record is removed from a Locator-Set, ITRs that have 1438 the mapping cached will not use the removed Locator because the xTRs 1439 will set the Locator-Status-Bit to 0. So, even if the Locator is in 1440 the list, it will not be used. For new mapping requests, the xTRs 1441 can set the Locator AFI to 0 (indicating an unspecified address), as 1442 well as setting the corresponding Locator-Status-Bit to 0. This 1443 forces ITRs with old or new mappings to avoid using the removed 1444 Locator. 1446 If many changes occur to a mapping over a long period of time, one 1447 will find empty record slots in the middle of the Locator-Set and new 1448 records appended to the Locator-Set. At some point, it would be 1449 useful to compact the Locator-Set so the Locator-Status-Bit settings 1450 can be efficiently packed. 1452 We propose here three approaches for Locator-Set compaction: one 1453 operational mechanism and two protocol mechanisms. The operational 1454 approach uses a clock sweep method. The protocol approaches use the 1455 concept of Solicit-Map-Requests and Map-Versioning. 1457 13.1. Clock Sweep 1459 The clock sweep approach uses planning in advance and the use of 1460 count-down TTLs to time out mappings that have already been cached. 1461 The default setting for an EID-to-RLOC mapping TTL is 24 hours. So, 1462 there is a 24-hour window to time out old mappings. The following 1463 clock sweep procedure is used: 1465 1. 24 hours before a mapping change is to take effect, a network 1466 administrator configures the ETRs at a site to start the clock 1467 sweep window. 1469 2. During the clock sweep window, ETRs continue to send Map-Reply 1470 messages with the current (unchanged) mapping records. The TTL 1471 for these mappings is set to 1 hour. 1473 3. 24 hours later, all previous cache entries will have timed out, 1474 and any active cache entries will time out within 1 hour. During 1475 this 1-hour window, the ETRs continue to send Map-Reply messages 1476 with the current (unchanged) mapping records with the TTL set to 1477 1 minute. 1479 4. At the end of the 1-hour window, the ETRs will send Map-Reply 1480 messages with the new (changed) mapping records. So, any active 1481 caches can get the new mapping contents right away if not cached, 1482 or in 1 minute if they had the mapping cached. The new mappings 1483 are cached with a TTL equal to the TTL in the Map-Reply. 1485 13.2. Solicit-Map-Request (SMR) 1487 Soliciting a Map-Request is a selective way for ETRs, at the site 1488 where mappings change, to control the rate they receive requests for 1489 Map-Reply messages. SMRs are also used to tell remote ITRs to update 1490 the mappings they have cached. 1492 Since the ETRs don't keep track of remote ITRs that have cached their 1493 mappings, they do not know which ITRs need to have their mappings 1494 updated. As a result, an ETR will solicit Map-Requests (called an 1495 SMR message) from those sites to which it has been sending 1496 encapsulated data for the last minute. In particular, an ETR will 1497 send an SMR to an ITR to which it has recently sent encapsulated 1498 data. 1500 An SMR message is simply a bit set in a Map-Request message. An ITR 1501 or PITR will send a Map-Request when they receive an SMR message. 1502 Both the SMR sender and the Map-Request responder MUST rate-limit 1503 these messages. Rate-limiting can be implemented as a global rate- 1504 limiter or one rate-limiter per SMR destination. 1506 The following procedure shows how an SMR exchange occurs when a site 1507 is doing Locator-Set compaction for an EID-to-RLOC mapping: 1509 1. When the database mappings in an ETR change, the ETRs at the site 1510 begin to send Map-Requests with the SMR bit set for each Locator 1511 in each Map-Cache entry the ETR caches. 1513 2. A remote ITR that receives the SMR message will schedule sending 1514 a Map-Request message to the source locator address of the SMR 1515 message or to the mapping database system. A newly allocated 1516 random nonce is selected, and the EID-Prefix used is the one 1517 copied from the SMR message. If the source Locator is the only 1518 Locator in the cached Locator-Set, the remote ITR SHOULD send a 1519 Map-Request to the database mapping system just in case the 1520 single Locator has changed and may no longer be reachable to 1521 accept the Map-Request. 1523 3. The remote ITR MUST rate-limit the Map-Request until it gets a 1524 Map-Reply while continuing to use the cached mapping. When 1525 Map-Versioning as described in Section 13.3 is used, an SMR 1526 sender can detect if an ITR is using the most up-to-date database 1527 mapping. 1529 4. The ETRs at the site with the changed mapping will reply to the 1530 Map-Request with a Map-Reply message that has a nonce from the 1531 SMR-invoked Map-Request. The Map-Reply messages SHOULD be rate- 1532 limited. This is important to avoid Map-Reply implosion. 1534 5. The ETRs at the site with the changed mapping record the fact 1535 that the site that sent the Map-Request has received the new 1536 mapping data in the Map-Cache entry for the remote site so the 1537 Locator-Status-Bits are reflective of the new mapping for packets 1538 going to the remote site. The ETR then stops sending SMR 1539 messages. 1541 Experimentation is in progress to determine the appropriate rate- 1542 limit parameters. 1544 For security reasons, an ITR MUST NOT process unsolicited Map- 1545 Replies. To avoid Map-Cache entry corruption by a third party, a 1546 sender of an SMR-based Map-Request MUST be verified. If an ITR 1547 receives an SMR-based Map-Request and the source is not in the 1548 Locator-Set for the stored Map-Cache entry, then the responding Map- 1549 Request MUST be sent with an EID destination to the mapping database 1550 system. Since the mapping database system is a more secure way to 1551 reach an authoritative ETR, it will deliver the Map-Request to the 1552 authoritative source of the mapping data. 1554 When an ITR receives an SMR-based Map-Request for which it does not 1555 have a cached mapping for the EID in the SMR message, it MAY not send 1556 an SMR-invoked Map-Request. This scenario can occur when an ETR 1557 sends SMR messages to all Locators in the Locator-Set it has stored 1558 in its map-cache but the remote ITRs that receive the SMR may not be 1559 sending packets to the site. There is no point in updating the ITRs 1560 until they need to send, in which case they will send Map-Requests to 1561 obtain a Map-Cache entry. 1563 13.3. Database Map-Versioning 1565 When there is unidirectional packet flow between an ITR and ETR, and 1566 the EID-to-RLOC mappings change on the ETR, it needs to inform the 1567 ITR so encapsulation to a removed Locator can stop and can instead be 1568 started to a new Locator in the Locator-Set. 1570 An ETR, when it sends Map-Reply messages, conveys its own Map-Version 1571 Number. This is known as the Destination Map-Version Number. ITRs 1572 include the Destination Map-Version Number in packets they 1573 encapsulate to the site. When an ETR decapsulates a packet and 1574 detects that the Destination Map-Version Number is less than the 1575 current version for its mapping, the SMR procedure described in 1576 Section 13.2 occurs. 1578 An ITR, when it encapsulates packets to ETRs, can convey its own Map- 1579 Version Number. This is known as the Source Map-Version Number. 1580 When an ETR decapsulates a packet and detects that the Source Map- 1581 Version Number is greater than the last Map-Version Number sent in a 1582 Map-Reply from the ITR's site, the ETR will send a Map-Request to one 1583 of the ETRs for the source site. 1585 A Map-Version Number is used as a sequence number per EID-Prefix, so 1586 values that are greater are considered to be more recent. A value of 1587 0 for the Source Map-Version Number or the Destination Map-Version 1588 Number conveys no versioning information, and an ITR does no 1589 comparison with previously received Map-Version Numbers. 1591 A Map-Version Number can be included in Map-Register messages as 1592 well. This is a good way for the Map-Server to assure that all ETRs 1593 for a site registering to it will be synchronized according to Map- 1594 Version Number. 1596 See [RFC6834] for a more detailed analysis and description of 1597 Database Map-Versioning. 1599 14. Multicast Considerations 1601 A multicast group address, as defined in the original Internet 1602 architecture, is an identifier of a grouping of topologically 1603 independent receiver host locations. The address encoding itself 1604 does not determine the location of the receiver(s). The multicast 1605 routing protocol, and the network-based state the protocol creates, 1606 determine where the receivers are located. 1608 In the context of LISP, a multicast group address is both an EID and 1609 a Routing Locator. Therefore, no specific semantic or action needs 1610 to be taken for a destination address, as it would appear in an IP 1611 header. Therefore, a group address that appears in an inner IP 1612 header built by a source host will be used as the destination EID. 1613 The outer IP header (the destination Routing Locator address), 1614 prepended by a LISP router, can use the same group address as the 1615 destination Routing Locator, use a multicast or unicast Routing 1616 Locator obtained from a Mapping System lookup, or use other means to 1617 determine the group address mapping. 1619 With respect to the source Routing Locator address, the ITR prepends 1620 its own IP address as the source address of the outer IP header. 1621 Just like it would if the destination EID was a unicast address. 1622 This source Routing Locator address, like any other Routing Locator 1623 address, MUST be globally routable. 1625 There are two approaches for LISP-Multicast, one that uses native 1626 multicast routing in the underlay with no support from the Mapping 1627 System and the other that uses only unicast routing in the underlay 1628 with support from the Mapping System. See [RFC6831] and 1629 [I.D-ietf-lisp-signal-free-multicast], respectively, for details. 1630 Details for LISP-Multicast and interworking with non-LISP sites are 1631 described in [RFC6831] and [RFC6832]. 1633 15. Router Performance Considerations 1635 LISP is designed to be very "hardware-based forwarding friendly". A 1636 few implementation techniques can be used to incrementally implement 1637 LISP: 1639 o When a tunnel-encapsulated packet is received by an ETR, the outer 1640 destination address may not be the address of the router. This 1641 makes it challenging for the control plane to get packets from the 1642 hardware. This may be mitigated by creating special Forwarding 1643 Information Base (FIB) entries for the EID-Prefixes of EIDs served 1644 by the ETR (those for which the router provides an RLOC 1645 translation). These FIB entries are marked with a flag indicating 1646 that control-plane processing should be performed. The forwarding 1647 logic of testing for particular IP protocol number values is not 1648 necessary. There are a few proven cases where no changes to 1649 existing deployed hardware were needed to support the LISP data- 1650 plane. 1652 o On an ITR, prepending a new IP header consists of adding more 1653 octets to a MAC rewrite string and prepending the string as part 1654 of the outgoing encapsulation procedure. Routers that support 1655 Generic Routing Encapsulation (GRE) tunneling [RFC2784] or 6to4 1656 tunneling [RFC3056] may already support this action. 1658 o A packet's source address or interface the packet was received on 1659 can be used to select VRF (Virtual Routing/Forwarding). The VRF's 1660 routing table can be used to find EID-to-RLOC mappings. 1662 For performance issues related to map-cache management, see 1663 Section 19. 1665 16. Mobility Considerations 1667 There are several kinds of mobility, of which only some might be of 1668 concern to LISP. Essentially, they are as follows. 1670 16.1. Slow Mobility 1672 A site wishes to change its attachment points to the Internet, and 1673 its LISP Tunnel Routers will have new RLOCs when it changes upstream 1674 providers. Changes in EID-to-RLOC mappings for sites are expected to 1675 be handled by configuration, outside of LISP. 1677 An individual endpoint wishes to move but is not concerned about 1678 maintaining session continuity. Renumbering is involved. LISP can 1679 help with the issues surrounding renumbering [RFC4192] [LISA96] by 1680 decoupling the address space used by a site from the address spaces 1681 used by its ISPs [RFC4984]. 1683 16.2. Fast Mobility 1685 Fast endpoint mobility occurs when an endpoint moves relatively 1686 rapidly, changing its IP-layer network attachment point. Maintenance 1687 of session continuity is a goal. This is where the Mobile IPv4 1688 [RFC5944] and Mobile IPv6 [RFC6275] [RFC4866] mechanisms are used and 1689 primarily where interactions with LISP need to be explored, such as 1690 the mechanisms in [I-D.portoles-lisp-eid-mobility] when the EID moves 1691 but the RLOC is in the network infrastructure. 1693 In LISP, one possibility is to "glean" information. When a packet 1694 arrives, the ETR could examine the EID-to-RLOC mapping and use that 1695 mapping for all outgoing traffic to that EID. It can do this after 1696 performing a route-returnability check, to ensure that the new 1697 network location does have an internal route to that endpoint. 1698 However, this does not cover the case where an ITR (the node assigned 1699 the RLOC) at the mobile-node location has been compromised. 1701 Mobile IP packet exchange is designed for an environment in which all 1702 routing information is disseminated before packets can be forwarded. 1703 In order to allow the Internet to grow to support expected future 1704 use, we are moving to an environment where some information may have 1705 to be obtained after packets are in flight. Modifications to IP 1706 mobility should be considered in order to optimize the behavior of 1707 the overall system. Anything that decreases the number of new EID- 1708 to-RLOC mappings needed when a node moves, or maintains the validity 1709 of an EID-to-RLOC mapping for a longer time, is useful. 1711 In addition to endpoints, a network can be mobile, possibly changing 1712 xTRs. A "network" can be as small as a single router and as large as 1713 a whole site. This is different from site mobility in that it is 1714 fast and possibly short-lived, but different from endpoint mobility 1715 in that a whole prefix is changing RLOCs. However, the mechanisms 1716 are the same, and there is no new overhead in LISP. A map request 1717 for any endpoint will return a binding for the entire mobile prefix. 1719 If mobile networks become a more common occurrence, it may be useful 1720 to revisit the design of the mapping service and allow for dynamic 1721 updates of the database. 1723 The issue of interactions between mobility and LISP needs to be 1724 explored further. Specific improvements to the entire system will 1725 depend on the details of mapping mechanisms. Mapping mechanisms 1726 should be evaluated on how well they support session continuity for 1727 mobile nodes. See [REF_TO_I-D.farinacci-lisp-predictive-rlocs] for 1728 more recent mechanisms which can provide near-zero packet loss during 1729 handoffs. 1731 16.3. LISP Mobile Node Mobility 1733 A mobile device can use the LISP infrastructure to achieve mobility 1734 by implementing the LISP encapsulation and decapsulation functions 1735 and acting as a simple ITR/ETR. By doing this, such a "LISP mobile 1736 node" can use topologically independent EID IP addresses that are not 1737 advertised into and do not impose a cost on the global routing 1738 system. These EIDs are maintained at the edges of the mapping system 1739 (in LISP Map-Servers and Map-Resolvers) and are provided on demand to 1740 only the correspondents of the LISP mobile node. 1742 Refer to [LISP-MN] for more details for when the EID and RLOC are co- 1743 located in the roaming node. 1745 17. LISP xTR Placement and Encapsulation Methods 1747 [FIXME: The Authors/Ed. agree on adding further scenarios.] 1749 This section will explore how and where ITRs and ETRs can be placed 1750 in the networkand will discuss the pros and cons of each scenario. 1751 For a more detailed networkd design deployment recommendation, refer 1752 to [RFC7215]. 1754 There are two basic deployment tradeoffs to consider: centralized 1755 versus distributed caches; and flat, Recursive, or Re-encapsulating 1756 Tunneling. When deciding on centralized versus distributed caching, 1757 the following issues should be considered: 1759 o Are the xTRs spread out so that the caches are spread across all 1760 the memories of each router? A centralized cache is when an ITR 1761 keeps a cache for all the EIDs it is encapsulating to. The packet 1762 takes a direct path to the destination Locator. A distributed 1763 cache is when an ITR needs help from other Re-Encapsulating Tunnel 1764 Routers (RTRs) because it does not store all the cache entries for 1765 the EIDs it is encapsulating to. So, the packet takes a path 1766 through RTRs that have a different set of cache entries. 1768 o Should management "touch points" be minimized by only choosing a 1769 few xTRs, just enough for redundancy? 1771 o In general, using more ITRs doesn't increase management load, 1772 since caches are built and stored dynamically. On the other hand, 1773 using more ETRs does require more management, since EID-Prefix-to- 1774 RLOC mappings need to be explicitly configured. 1776 When deciding on flat, Recursive, or Re-Encapsulating Tunneling, the 1777 following issues should be considered: 1779 o Flat tunneling implements a single encapsulation path between the 1780 source site and destination site. This generally offers better 1781 paths between sources and destinations with a single encapsulation 1782 path. 1784 o Recursive Tunneling is when encapsulated traffic is again further 1785 encapsulated in another tunnel, either to implement VPNs or to 1786 perform Traffic Engineering. When doing VPN-based tunneling, the 1787 site has some control, since the site is prepending a new 1788 encapsulation header. In the case of TE-based tunneling, the site 1789 may have control if it is prepending a new tunnel header, but if 1790 the site's ISP is doing the TE, then the site has no control. 1791 Recursive Tunneling generally will result in suboptimal paths but 1792 with the benefit of steering traffic to parts of the network that 1793 have more resources available. 1795 o The technique of Re-Encapsulation ensures that packets only 1796 require one encapsulation header. So, if a packet needs to be re- 1797 routed, it is first decapsulated by the RTR and then Re- 1798 Encapsulated with a new encapsulation header using a new RLOC. 1800 The next sub-sections will examine where xTRs and RTRs can reside in 1801 the network. 1803 17.1. First-Hop/Last-Hop xTRs 1805 By locating xTRs close to hosts, the EID-Prefix set is at the 1806 granularity of an IP subnet. So, at the expense of more EID-Prefix- 1807 to-RLOC sets for the site, the caches in each xTR can remain 1808 relatively small. But caches always depend on the number of non- 1809 aggregated EID destination flows active through these xTRs. 1811 With more xTRs doing encapsulation, the increase in control traffic 1812 grows as well: since the EID granularity is greater, more Map- 1813 Requests and Map-Replies are traveling between more routers. 1815 The advantage of placing the caches and databases at these stub 1816 routers is that the products deployed in this part of the network 1817 have better price-memory ratios than their core router counterparts. 1818 Memory is typically less expensive in these devices, and fewer routes 1819 are stored (only IGP routes). These devices tend to have excess 1820 capacity, both for forwarding and routing states. 1822 LISP functionality can also be deployed in edge switches. These 1823 devices generally have layer-2 ports facing hosts and layer-3 ports 1824 facing the Internet. Spare capacity is also often available in these 1825 devices. 1827 17.2. Border/Edge xTRs 1829 Using Customer Edge (CE) routers for xTR placement allows the EID 1830 space associated with a site to be reachable via a small set of RLOCs 1831 assigned to the CE-based xTRs for that site. 1833 This offers the opposite benefit of the first-hop/last-hop xTR 1834 scenario: the number of mapping entries and network management touch 1835 points is reduced, allowing better scaling. 1837 One disadvantage is that fewer network resources are used to reach 1838 host endpoints, thereby centralizing the point-of-failure domain and 1839 creating network choke points at the CE xTR. 1841 Note that more than one CE xTR at a site can be configured with the 1842 same IP address. In this case, an RLOC is an anycast address. This 1843 allows resilience between the CE xTRs. That is, if a CE xTR fails, 1844 traffic is automatically routed to the other xTRs using the same 1845 anycast address. However, this comes with the disadvantage where the 1846 site cannot control the entrance point when the anycast route is 1847 advertised out from all border routers. Another disadvantage of 1848 using anycast Locators is the limited advertisement scope of /32 (or 1849 /128 for IPv6) routes. 1851 17.3. ISP Provider Edge (PE) xTRs 1853 The use of ISP PE routers as xTRs is not the typical deployment 1854 scenario envisioned in this specification. This section attempts to 1855 capture some of the reasoning behind this preference for implementing 1856 LISP on CE routers. 1858 The use of ISP PE routers for xTR placement gives an ISP, rather than 1859 a site, control over the location of the ETRs. That is, the ISP can 1860 decide whether the xTRs are in the destination site (in either CE 1861 xTRs or last-hop xTRs within a site) or at other PE edges. The 1862 advantage of this case is that two encapsulation headers can be 1863 avoided. By having the PE be the first router on the path to 1864 encapsulate, it can choose a TE path first, and the ETR can 1865 decapsulate and Re-Encapsulate for a new encapsuluation path to the 1866 destination end site. 1868 An obvious disadvantage is that the end site has no control over 1869 where its packets flow or over the RLOCs used. Other disadvantages 1870 include difficulty in synchronizing path liveness updates between CE 1871 and PE routers. 1873 As mentioned in earlier sections, a combination of these scenarios is 1874 possible at the expense of extra packet header overhead; if both site 1875 and provider want control, then Recursive or Re-Encapsulating Tunnels 1876 are used. 1878 17.4. LISP Functionality with Conventional NATs 1880 LISP routers can be deployed behind Network Address Translator (NAT) 1881 devices to provide the same set of packet services hosts have today 1882 when they are addressed out of private address space. 1884 It is important to note that a locator address in any LISP control 1885 message MUST be a globally routable address and therefore SHOULD NOT 1886 contain [RFC1918] addresses. If a LISP xTR is configured with 1887 private RLOC addresses, they MUST be used only in the outer IP header 1888 so the NAT device can translate properly. Otherwise, EID addresses 1889 MUST be translated before encapsulation is performed when LISP VPNs 1890 are not in use. Both NAT translation and LISP encapsulation 1891 functions could be co-located in the same device. 1893 17.5. Packets Egressing a LISP Site 1895 When a LISP site is using two ITRs for redundancy, the failure of one 1896 ITR will likely shift outbound traffic to the second. This second 1897 ITR's cache may not be populated with the same EID-to-RLOC mapping 1898 entries as the first. If this second ITR does not have these 1899 mappings, traffic will be dropped while the mappings are retrieved 1900 from the mapping system. The retrieval of these messages may 1901 increase the load of requests being sent into the mapping system. 1902 Deployment and experimentation will determine whether this issue 1903 requires more attention. 1905 18. Traceroute Considerations 1907 When a source host in a LISP site initiates a traceroute to a 1908 destination host in another LISP site, it is highly desirable for it 1909 to see the entire path. Since packets are encapsulated from the ITR 1910 to the ETR, the hop across the tunnel could be viewed as a single 1911 hop. However, LISP traceroute will provide the entire path so the 1912 user can see 3 distinct segments of the path from a source LISP host 1913 to a destination LISP host: 1915 Segment 1 (in source LISP site based on EIDs): 1917 source host ---> first hop ... next hop ---> ITR 1919 Segment 2 (in the core network based on RLOCs): 1921 ITR ---> next hop ... next hop ---> ETR 1923 Segment 3 (in the destination LISP site based on EIDs): 1925 ETR ---> next hop ... last hop ---> destination host 1927 For segment 1 of the path, ICMP Time Exceeded messages are returned 1928 in the normal manner as they are today. The ITR performs a TTL 1929 decrement and tests for 0 before encapsulating. Therefore, the ITR's 1930 hop is seen by the traceroute source as having an EID address (the 1931 address of the site-facing interface). 1933 For segment 2 of the path, ICMP Time Exceeded messages are returned 1934 to the ITR because the TTL decrement to 0 is done on the outer 1935 header, so the destinations of the ICMP messages are the ITR RLOC 1936 address and the source RLOC address of the encapsulated traceroute 1937 packet. The ITR looks inside of the ICMP payload to inspect the 1938 traceroute source so it can return the ICMP message to the address of 1939 the traceroute client and also retain the core router IP address in 1940 the ICMP message. This is so the traceroute client can display the 1941 core router address (the RLOC address) in the traceroute output. The 1942 ETR returns its RLOC address and responds to the TTL decrement to 0, 1943 as the previous core routers did. 1945 For segment 3, the next-hop router downstream from the ETR will be 1946 decrementing the TTL for the packet that was encapsulated, sent into 1947 the core, decapsulated by the ETR, and forwarded because it isn't the 1948 final destination. If the TTL is decremented to 0, any router on the 1949 path to the destination of the traceroute, including the next-hop 1950 router or destination, will send an ICMP Time Exceeded message to the 1951 source EID of the traceroute client. The ICMP message will be 1952 encapsulated by the local ITR and sent back to the ETR in the 1953 originated traceroute source site, where the packet will be delivered 1954 to the host. 1956 18.1. IPv6 Traceroute 1958 IPv6 traceroute follows the procedure described above, since the 1959 entire traceroute data packet is included in the ICMP Time Exceeded 1960 message payload. Therefore, only the ITR needs to pay special 1961 attention to forwarding ICMP messages back to the traceroute source. 1963 18.2. IPv4 Traceroute 1965 For IPv4 traceroute, we cannot follow the above procedure, since IPv4 1966 ICMP Time Exceeded messages only include the invoking IP header and 1967 8 octets that follow the IP header. Therefore, when a core router 1968 sends an IPv4 Time Exceeded message to an ITR, all the ITR has in the 1969 ICMP payload is the encapsulated header it prepended, followed by a 1970 UDP header. The original invoking IP header, and therefore the 1971 identity of the traceroute source, is lost. 1973 The solution we propose to solve this problem is to cache traceroute 1974 IPv4 headers in the ITR and to match them up with corresponding IPv4 1975 Time Exceeded messages received from core routers and the ETR. The 1976 ITR will use a circular buffer for caching the IPv4 and UDP headers 1977 of traceroute packets. It will select a 16-bit number as a key to 1978 find them later when the IPv4 Time Exceeded messages are received. 1979 When an ITR encapsulates an IPv4 traceroute packet, it will use the 1980 16-bit number as the UDP source port in the encapsulating header. 1981 When the ICMP Time Exceeded message is returned to the ITR, the UDP 1982 header of the encapsulating header is present in the ICMP payload, 1983 thereby allowing the ITR to find the cached headers for the 1984 traceroute source. The ITR puts the cached headers in the payload 1985 and sends the ICMP Time Exceeded message to the traceroute source 1986 retaining the source address of the original ICMP Time Exceeded 1987 message (a core router or the ETR of the site of the traceroute 1988 destination). 1990 The signature of a traceroute packet comes in two forms. The first 1991 form is encoded as a UDP message where the destination port is 1992 inspected for a range of values. The second form is encoded as an 1993 ICMP message where the IP identification field is inspected for a 1994 well-known value. 1996 18.3. Traceroute Using Mixed Locators 1998 When either an IPv4 traceroute or IPv6 traceroute is originated and 1999 the ITR encapsulates it in the other address family header, one 2000 cannot get all 3 segments of the traceroute. Segment 2 of the 2001 traceroute cannot be conveyed to the traceroute source, since it is 2002 expecting addresses from intermediate hops in the same address format 2003 for the type of traceroute it originated. Therefore, in this case, 2004 segment 2 will make the tunnel look like one hop. All the ITR has to 2005 do to make this work is to not copy the inner TTL to the outer, 2006 encapsulating header's TTL when a traceroute packet is encapsulated 2007 using an RLOC from a different address family. This will cause no 2008 TTL decrement to 0 to occur in core routers between the ITR and ETR. 2010 19. Security Considerations 2012 FIXME: ToDo 2014 Security considerations for LISP are discussed in [LISP-THREATS], in 2015 addition [LISP-SEC] provides authentication and integrity to LISP 2016 mappings. 2018 20. Network Management Considerations 2020 Considerations for network management tools exist so the LISP 2021 protocol suite can be operationally managed. These mechanisms can be 2022 found in [LISP-MIB] and [RFC6835]. 2024 21. IANA Considerations 2026 This section provides guidance to the Internet Assigned Numbers 2027 Authority (IANA) regarding registration of values related to the LISP 2028 specification, in accordance with BCP 26 [RFC5226]. 2030 There are four namespaces (listed in the sub-sections below) in LISP 2031 that have been registered. 2033 o LISP IANA registry allocations should not be made for purposes 2034 unrelated to LISP routing or transport protocols. 2036 o The following policies are used here with the meanings defined in 2037 BCP 26: "Specification Required", "IETF Review", "Experimental 2038 Use", and "First Come First Served". 2040 21.1. LISP ACT and Flag Fields 2042 New ACT values [REF_TO_RFC6833bis] can be allocated through IETF 2043 review or IESG approval. Four values have already been allocated by 2044 this specification ([REF_TO_RFC6833bis]. 2046 In addition, LISP has a number of flag fields and reserved fields, 2047 such as the LISP header flags field (Section 5.3). New bits for 2048 flags in these fields can be implemented after IETF review or IESG 2049 approval, but these need not be managed by IANA. 2051 21.2. LISP Address Type Codes 2053 LISP Address [LCAF] type codes have a range from 0 to 255. New type 2054 codes MUST be allocated consecutively, starting at 0. Type Codes 2055 0-127 are to be assigned by IETF review or IESG approval. 2057 Type Codes 128-255 are available according to the [RFC5226] First 2058 Come First Served policy. 2060 This registry, initially empty, is constructed for future use in 2061 experimental work related to LISP Canonical Address Format (LCAF) 2062 values. See [LCAF] for details of other possible unapproved address 2063 encodings. The unapproved LCAF encodings are an area for further 2064 study and experimentation. 2066 21.3. LISP UDP Port Numbers 2068 The IANA registry has allocated UDP port numbers 4341 and 4342 for 2069 lisp-data and lisp-control operation, respectively. IANA has updated 2070 the description for UDP ports 4341 and 4342 as follows: 2072 lisp-data 4341 udp LISP Data Packets 2073 lisp-control 4342 udp LISP Control Packets 2075 21.4. LISP Key ID Numbers 2077 The following Key ID values are defined by this specification as used 2078 in any packet type that references a 'Key ID' field: 2080 Name Number Defined in 2081 ----------------------------------------------- 2082 None 0 n/a 2083 HMAC-SHA-1-96 1 [RFC2404] 2084 HMAC-SHA-256-128 2 [RFC4868] 2086 Number values are in the range of 0 to 65535. The allocation of 2087 values is on a first come first served basis. 2089 22. References 2091 22.1. Normative References 2093 [I.D-ietf-lisp-crypto] 2094 Farinacci, D. and B. Weiss, "LISP Data-Plane 2095 Confidentiality", October 2016. 2097 [RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, 2098 DOI 10.17487/RFC0768, August 1980, 2099 . 2101 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 2102 DOI 10.17487/RFC0791, September 1981, 2103 . 2105 [RFC1918] Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G., 2106 and E. Lear, "Address Allocation for Private Internets", 2107 BCP 5, RFC 1918, DOI 10.17487/RFC1918, February 1996, 2108 . 2110 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2111 Requirement Levels", BCP 14, RFC 2119, 2112 DOI 10.17487/RFC2119, March 1997, 2113 . 2115 [RFC2404] Madson, C. and R. Glenn, "The Use of HMAC-SHA-1-96 within 2116 ESP and AH", RFC 2404, DOI 10.17487/RFC2404, November 2117 1998, . 2119 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 2120 (IPv6) Specification", RFC 2460, DOI 10.17487/RFC2460, 2121 December 1998, . 2123 [RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition 2124 of Explicit Congestion Notification (ECN) to IP", 2125 RFC 3168, DOI 10.17487/RFC3168, September 2001, 2126 . 2128 [RFC3232] Reynolds, J., Ed., "Assigned Numbers: RFC 1700 is Replaced 2129 by an On-line Database", RFC 3232, DOI 10.17487/RFC3232, 2130 January 2002, . 2132 [RFC4086] Eastlake 3rd, D., Schiller, J., and S. Crocker, 2133 "Randomness Requirements for Security", BCP 106, RFC 4086, 2134 DOI 10.17487/RFC4086, June 2005, 2135 . 2137 [RFC4632] Fuller, V. and T. Li, "Classless Inter-domain Routing 2138 (CIDR): The Internet Address Assignment and Aggregation 2139 Plan", BCP 122, RFC 4632, DOI 10.17487/RFC4632, August 2140 2006, . 2142 [RFC4868] Kelly, S. and S. Frankel, "Using HMAC-SHA-256, HMAC-SHA- 2143 384, and HMAC-SHA-512 with IPsec", RFC 4868, 2144 DOI 10.17487/RFC4868, May 2007, 2145 . 2147 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 2148 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 2149 DOI 10.17487/RFC5226, May 2008, 2150 . 2152 [RFC5496] Wijnands, IJ., Boers, A., and E. Rosen, "The Reverse Path 2153 Forwarding (RPF) Vector TLV", RFC 5496, 2154 DOI 10.17487/RFC5496, March 2009, 2155 . 2157 [RFC5944] Perkins, C., Ed., "IP Mobility Support for IPv4, Revised", 2158 RFC 5944, DOI 10.17487/RFC5944, November 2010, 2159 . 2161 [RFC6115] Li, T., Ed., "Recommendation for a Routing Architecture", 2162 RFC 6115, DOI 10.17487/RFC6115, February 2011, 2163 . 2165 [RFC6275] Perkins, C., Ed., Johnson, D., and J. Arkko, "Mobility 2166 Support in IPv6", RFC 6275, DOI 10.17487/RFC6275, July 2167 2011, . 2169 [RFC6833] Farinacci, D. and V. Fuller, "Locator/ID Separation 2170 Protocol (LISP) Map-Server Interface", RFC 6833, January 2171 2013. 2173 [RFC6834] Iannone, L., Saucez, D., and O. Bonaventure, "Locator/ID 2174 Separation Protocol (LISP) Map-Versioning", RFC 6834, 2175 January 2013. 2177 [RFC6836] Farinacci, D., Fuller, V., Meyer, D., and D. Lewis, 2178 "Locator/ID Separation Protocol Alternative Logical 2179 Topology (LISP+ALT)", RFC 6836, January 2013. 2181 [RFC7215] Jakab, L., Cabellos, A., Coras, F., Domingo-Pascual, J., 2182 and D. Lewis, "Locator/Identifier Separation Protocol 2183 (LISP) Network Element Deployment Considerations", 2184 RFC 6834, April 2014. 2186 22.2. Informative References 2188 [AFI] IANA, "Address Family Numbers", November 2007, 2189 . 2191 [BGP-SEC] Lepinski, M. and S. Turner, "An Overview of BGPSEC", Work 2192 in Progress, May 2012. 2194 [CHIAPPA] Chiappa, J., "Endpoints and Endpoint names: A Proposed 2195 Enhancement to the Internet Architecture", 1999, 2196 . 2198 [CONS] Brim, S., Chiappa, N., Farinacci, D., Fuller, V., Lewis, 2199 D., and D. Meyer, "LISP-CONS: A Content distribution 2200 Overlay Network Service for LISP", Work in Progress, April 2201 2008. 2203 [EMACS] Brim, S., Farinacci, D., Meyer, D., and J. Curran, "EID 2204 Mappings Multicast Across Cooperating Systems for LISP", 2205 Work in Progress, November 2007. 2207 [I-D.portoles-lisp-eid-mobility] 2208 Portoles, M., Ashtaputre, V., Moreno, V., Maino, F., and 2209 D. Farinacci, "LISP L2/L3 EID Mobility Using a Unified 2210 Control Plane", Work in Progress, April 2016. 2212 [I.D-ietf-lisp-signal-free-multicast] 2213 Moreno, V. and D. Farinacci, "Signal-Free LISP Multicast", 2214 Work in Progress, October 2016. 2216 [LCAF] Farinacci, D., Meyer, D., and J. Snijders, "LISP Canonical 2217 Address Format (LCAF)", Work in Progress, January 2013. 2219 [LISA96] Lear, E., Tharp, D., Katinsky, J., and J. Coffin, 2220 "Renumbering: Threat or Menace?", Usenix Tenth System 2221 Administration Conference (LISA 96), October 1996. 2223 [LISP-DDT] 2224 Fuller, V., Lewis, D., Ermagan, V., Jain, A., and A. 2225 Smirnov, "LISP Delegated Database Tree", Work in Progress, 2226 April 2015. 2228 [LISP-INTRO] 2229 Cabellos, A. and D. Saucez, "An Architectural Introduction 2230 to the Locator/ID Separation Protocol (LISP)", Work 2231 in Progress, April 2015. 2233 [LISP-MIB] 2234 Schudel, G., Jain, A., and V. Moreno, "LISP MIB", Work 2235 in Progress, January 2013. 2237 [LISP-MN] Farinacci, D., Lewis, D., Meyer, D., and C. White, "LISP 2238 Mobile Node", Work in Progress, October 2012. 2240 [LISP-SEC] 2241 Maino, F., Ermagan, V., Cabellos, A., Saucez, D., and O. 2242 Bonaventure, "LISP-Security (LISP-SEC)", Work in Progress, 2243 October 2012. 2245 [LISP-THREATS] 2246 Saucez, D., Iannone, L., and O. Bonaventure, "LISP Threats 2247 Analysis", Work in Progress, January 2016. 2249 [LOC-ID-ARCH] 2250 Meyer, D. and D. Lewis, "Architectural Implications of 2251 Locator/ID Separation", Work in Progress, January 2009. 2253 [OPENLISP] 2254 Iannone, L., Saucez, D., and O. Bonaventure, "OpenLISP 2255 Implementation Report", Work in Progress, July 2008. 2257 [RADIR] Narten, T., "On the Scalability of Internet Routing", Work 2258 in Progress, February 2010. 2260 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 2261 STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987, 2262 . 2264 [RFC2784] Farinacci, D., Li, T., Hanks, S., Meyer, D., and P. 2265 Traina, "Generic Routing Encapsulation (GRE)", RFC 2784, 2266 DOI 10.17487/RFC2784, March 2000, 2267 . 2269 [RFC3056] Carpenter, B. and K. Moore, "Connection of IPv6 Domains 2270 via IPv4 Clouds", RFC 3056, DOI 10.17487/RFC3056, February 2271 2001, . 2273 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 2274 A., Peterson, J., Sparks, R., Handley, M., and E. 2275 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 2276 DOI 10.17487/RFC3261, June 2002, 2277 . 2279 [RFC4107] Bellovin, S. and R. Housley, "Guidelines for Cryptographic 2280 Key Management", BCP 107, RFC 4107, DOI 10.17487/RFC4107, 2281 June 2005, . 2283 [RFC4192] Baker, F., Lear, E., and R. Droms, "Procedures for 2284 Renumbering an IPv6 Network without a Flag Day", RFC 4192, 2285 DOI 10.17487/RFC4192, September 2005, 2286 . 2288 [RFC4866] Arkko, J., Vogt, C., and W. Haddad, "Enhanced Route 2289 Optimization for Mobile IPv6", RFC 4866, 2290 DOI 10.17487/RFC4866, May 2007, 2291 . 2293 [RFC4984] Meyer, D., Ed., Zhang, L., Ed., and K. Fall, Ed., "Report 2294 from the IAB Workshop on Routing and Addressing", 2295 RFC 4984, DOI 10.17487/RFC4984, September 2007, 2296 . 2298 [RFC6480] Lepinski, M. and S. Kent, "An Infrastructure to Support 2299 Secure Internet Routing", RFC 6480, DOI 10.17487/RFC6480, 2300 February 2012, . 2302 [RFC6518] Lebovitz, G. and M. Bhatia, "Keying and Authentication for 2303 Routing Protocols (KARP) Design Guidelines", RFC 6518, 2304 DOI 10.17487/RFC6518, February 2012, 2305 . 2307 [RFC6831] Farinacci, D., Meyer, D., Zwiebel, J., and S. Venaas, "The 2308 Locator/ID Separation Protocol (LISP) for Multicast 2309 Environments", RFC 6831, January 2013. 2311 [RFC6832] Lewis, D., Meyer, D., Farinacci, D., and V. Fuller, 2312 "Interworking between Locator/ID Separation Protocol 2313 (LISP) and Non-LISP Sites", RFC 6832, January 2013. 2315 [RFC6835] Farinacci, D. and D. Meyer, "The Locator/ID Separation 2316 Protocol Internet Groper (LIG)", RFC 6835, January 2013. 2318 [RFC6837] Lear, E., "NERD: A Not-so-novel Endpoint ID (EID) to 2319 Routing Locator (RLOC) Database", RFC 6837, January 2013. 2321 [UDP-TUNNELS] 2322 Eubanks, M., Chimento, P., and M. Westerlund, "IPv6 and 2323 UDP Checksums for Tunneled Packets", Work in Progress, 2324 January 2013. 2326 [UDP-ZERO] 2327 Fairhurst, G. and M. Westerlund, "Applicability Statement 2328 for the use of IPv6 UDP Datagrams with Zero Checksums", 2329 Work in Progress, December 2012. 2331 Appendix A. Acknowledgments 2333 An initial thank you goes to Dave Oran for planting the seeds for the 2334 initial ideas for LISP. His consultation continues to provide value 2335 to the LISP authors. 2337 A special and appreciative thank you goes to Noel Chiappa for 2338 providing architectural impetus over the past decades on separation 2339 of location and identity, as well as detailed reviews of the LISP 2340 architecture and documents, coupled with enthusiasm for making LISP a 2341 practical and incremental transition for the Internet. 2343 The authors would like to gratefully acknowledge many people who have 2344 contributed discussions and ideas to the making of this proposal. 2345 They include Scott Brim, Andrew Partan, John Zwiebel, Jason Schiller, 2346 Lixia Zhang, Dorian Kim, Peter Schoenmaker, Vijay Gill, Geoff Huston, 2347 David Conrad, Mark Handley, Ron Bonica, Ted Seely, Mark Townsley, 2348 Chris Morrow, Brian Weis, Dave McGrew, Peter Lothberg, Dave Thaler, 2349 Eliot Lear, Shane Amante, Ved Kafle, Olivier Bonaventure, Luigi 2350 Iannone, Robin Whittle, Brian Carpenter, Joel Halpern, Terry 2351 Manderson, Roger Jorgensen, Ran Atkinson, Stig Venaas, Iljitsch van 2352 Beijnum, Roland Bless, Dana Blair, Bill Lynch, Marc Woolward, Damien 2353 Saucez, Damian Lezama, Attilla De Groot, Parantap Lahiri, David 2354 Black, Roque Gagliano, Isidor Kouvelas, Jesper Skriver, Fred Templin, 2355 Margaret Wasserman, Sam Hartman, Michael Hofling, Pedro Marques, Jari 2356 Arkko, Gregg Schudel, Srinivas Subramanian, Amit Jain, Xu Xiaohu, 2357 Dhirendra Trivedi, Yakov Rekhter, John Scudder, John Drake, Dimitri 2358 Papadimitriou, Ross Callon, Selina Heimlich, Job Snijders, Vina 2359 Ermagan, Fabio Maino, Victor Moreno, Chris White, Clarence Filsfils, 2360 Alia Atlas, Florin Coras and Alberto Rodriguez. 2362 This work originated in the Routing Research Group (RRG) of the IRTF. 2363 An individual submission was converted into the IETF LISP working 2364 group document that became this RFC. 2366 The LISP working group would like to give a special thanks to Jari 2367 Arkko, the Internet Area AD at the time that the set of LISP 2368 documents were being prepared for IESG last call, and for his 2369 meticulous reviews and detailed commentaries on the 7 working group 2370 last call documents progressing toward experimental RFCs. 2372 Authors' Addresses 2373 Dino Farinacci 2374 Cisco Systems 2375 Tasman Drive 2376 San Jose, CA 95134 2377 USA 2379 EMail: farinacci@gmail.com 2381 Vince Fuller 2382 Cisco Systems 2383 Tasman Drive 2384 San Jose, CA 95134 2385 USA 2387 EMail: vaf@vaf.net 2389 Dave Meyer 2390 Cisco Systems 2391 170 Tasman Drive 2392 San Jose, CA 2393 USA 2395 EMail: dmm@1-4-5.net 2397 Darrel Lewis 2398 Cisco Systems 2399 170 Tasman Drive 2400 San Jose, CA 2401 USA 2403 EMail: darlewis@cisco.com 2405 Albert Cabellos 2406 UPC/BarcelonaTech 2407 Campus Nord, C. Jordi Girona 1-3 2408 Barcelona, Catalunya 2409 Spain 2411 EMail: acabello@ac.upc.edu