idnits 2.17.1 draft-ietf-lisp-multicast-08.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 (September 9, 2011) is 4612 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- == Outdated reference: A later version (-06) exists of draft-ietf-lisp-interworking-02 == Outdated reference: A later version (-24) exists of draft-ietf-lisp-15 ** Obsolete normative reference: RFC 4601 (Obsoleted by RFC 7761) == Outdated reference: A later version (-10) exists of draft-ietf-lisp-alt-08 == Outdated reference: A later version (-26) exists of draft-ietf-mboned-mtrace-v2-08 Summary: 1 error (**), 0 flaws (~~), 5 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group D. Farinacci 3 Internet-Draft D. Meyer 4 Intended status: Experimental J. Zwiebel 5 Expires: March 12, 2012 S. Venaas 6 cisco Systems 7 September 9, 2011 9 LISP for Multicast Environments 10 draft-ietf-lisp-multicast-08 12 Abstract 14 This draft describes how inter-domain multicast routing will function 15 in an environment where Locator/ID Separation is deployed using the 16 LISP architecture. 18 Status of this Memo 20 This Internet-Draft is submitted in full conformance with the 21 provisions of BCP 78 and BCP 79. 23 Internet-Drafts are working documents of the Internet Engineering 24 Task Force (IETF). Note that other groups may also distribute 25 working documents as Internet-Drafts. The list of current Internet- 26 Drafts is at http://datatracker.ietf.org/drafts/current/. 28 Internet-Drafts are draft documents valid for a maximum of six months 29 and may be updated, replaced, or obsoleted by other documents at any 30 time. It is inappropriate to use Internet-Drafts as reference 31 material or to cite them other than as "work in progress." 33 This Internet-Draft will expire on March 12, 2012. 35 Copyright Notice 37 Copyright (c) 2011 IETF Trust and the persons identified as the 38 document authors. All rights reserved. 40 This document is subject to BCP 78 and the IETF Trust's Legal 41 Provisions Relating to IETF Documents 42 (http://trustee.ietf.org/license-info) in effect on the date of 43 publication of this document. Please review these documents 44 carefully, as they describe your rights and restrictions with respect 45 to this document. Code Components extracted from this document must 46 include Simplified BSD License text as described in Section 4.e of 47 the Trust Legal Provisions and are provided without warranty as 48 described in the Simplified BSD License. 50 Table of Contents 52 1. Requirements Notation . . . . . . . . . . . . . . . . . . . . 3 53 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 54 3. Definition of Terms . . . . . . . . . . . . . . . . . . . . . 6 55 4. Basic Overview . . . . . . . . . . . . . . . . . . . . . . . . 9 56 5. Source Addresses versus Group Addresses . . . . . . . . . . . 12 57 6. Locator Reachability Implications on LISP-Multicast . . . . . 13 58 7. Multicast Protocol Changes . . . . . . . . . . . . . . . . . . 14 59 8. LISP-Multicast Data-Plane Architecture . . . . . . . . . . . . 17 60 8.1. ITR Forwarding Procedure . . . . . . . . . . . . . . . . . 17 61 8.1.1. Multiple RLOCs for an ITR . . . . . . . . . . . . . . 17 62 8.1.2. Multiple ITRs for a LISP Source Site . . . . . . . . . 18 63 8.2. ETR Forwarding Procedure . . . . . . . . . . . . . . . . . 18 64 8.3. Replication Locations . . . . . . . . . . . . . . . . . . 19 65 9. LISP-Multicast Interworking . . . . . . . . . . . . . . . . . 20 66 9.1. LISP and non-LISP Mixed Sites . . . . . . . . . . . . . . 20 67 9.1.1. LISP Source Site to non-LISP Receiver Sites . . . . . 21 68 9.1.2. Non-LISP Source Site to non-LISP Receiver Sites . . . 22 69 9.1.3. Non-LISP Source Site to Any Receiver Site . . . . . . 23 70 9.1.4. Unicast LISP Source Site to Any Receiver Sites . . . . 24 71 9.1.5. LISP Source Site to Any Receiver Sites . . . . . . . . 24 72 9.2. LISP Sites with Mixed Address Families . . . . . . . . . . 25 73 9.3. Making a Multicast Interworking Decision . . . . . . . . . 27 74 10. Considerations when RP Addresses are Embedded in Group 75 Addresses . . . . . . . . . . . . . . . . . . . . . . . . . . 28 76 11. Taking Advantage of Upgrades in the Core . . . . . . . . . . . 29 77 12. Mtrace Considerations . . . . . . . . . . . . . . . . . . . . 30 78 13. Security Considerations . . . . . . . . . . . . . . . . . . . 31 79 14. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 32 80 15. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 33 81 16. References . . . . . . . . . . . . . . . . . . . . . . . . . . 34 82 16.1. Normative References . . . . . . . . . . . . . . . . . . . 34 83 16.2. Informative References . . . . . . . . . . . . . . . . . . 35 84 Appendix A. Document Change Log . . . . . . . . . . . . . . . . . 36 85 A.1. Changes to draft-ietf-lisp-multicast-08.txt . . . . . . . 36 86 A.2. Changes to draft-ietf-lisp-multicast-07.txt . . . . . . . 36 87 A.3. Changes to draft-ietf-lisp-multicast-06.txt . . . . . . . 36 88 A.4. Changes to draft-ietf-lisp-multicast-05.txt . . . . . . . 36 89 A.5. Changes to draft-ietf-lisp-multicast-04.txt . . . . . . . 36 90 A.6. Changes to draft-ietf-lisp-multicast-03.txt . . . . . . . 36 91 A.7. Changes to draft-ietf-lisp-multicast-02.txt . . . . . . . 37 92 A.8. Changes to draft-ietf-lisp-multicast-01.txt . . . . . . . 37 93 A.9. Changes to draft-ietf-lisp-multicast-00.txt . . . . . . . 37 94 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 38 96 1. Requirements Notation 98 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 99 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 100 document are to be interpreted as described in [RFC2119]. 102 2. Introduction 104 The Locator/ID Separation Architecture [LISP] provides a mechanism to 105 separate out Identification and Location semantics from the current 106 definition of an IP address. By creating two namespaces, an Endpoint 107 ID (EID) namespace used by sites and a Routing Locator (RLOC) 108 namespace used by core routing, the core routing infrastructure can 109 scale by doing topological aggregation of routing information. 111 Since LISP creates a new namespace, a mapping function must exist to 112 map a site's EID prefixes to its associated locators. For unicast 113 packets, both the source address and destination address must be 114 mapped. For multicast packets, only the source address needs to be 115 mapped. The destination group address doesn't need to be mapped 116 because the semantics of an IPv4 or IPv6 group address are logical in 117 nature and not topology-dependent. Therefore, this specification 118 focuses on to map a source EID address of a multicast flow during 119 distribution tree setup and packet delivery. 121 This specification will address the following scenarios: 123 1. How a multicast source host in a LISP site sends multicast 124 packets to receivers inside of its site as well as to receivers 125 in other sites that are LISP enabled. 127 2. How inter-domain (or between LISP sites) multicast distribution 128 trees are built and how forwarding of multicast packets leaving a 129 source site toward receivers sites is performed. 131 3. What protocols are affected and what changes are required to such 132 multicast protocols. 134 4. How ASM-mode (Any Source Multicast), SSM-mode (Single Source 135 Multicast), and Bidir-mode (Bidirectional Shared Trees) service 136 models will operate. 138 5. How multicast packet flow will occur for multiple combinations of 139 LISP and non-LISP capable source and receiver sites, for example: 141 A. How multicast packets from a source host in a LISP site are 142 sent to receivers in other sites when they are all non-LISP 143 sites. 145 B. How multicast packets from a source host in a LISP site are 146 sent to receivers in both LISP-enabled sites and non-LISP 147 sites. 149 C. How multicast packets from a source host in a non-LISP site 150 are sent to receivers in other sites when they are all LISP- 151 enabled sites. 153 D. How multicast packets from a source host in a non-LISP site 154 are sent to receivers in both LISP-enabled sites and non-LISP 155 sites. 157 This specification focuses on what changes are needed to the 158 multicast routing protocols to support LISP-Multicast as well as 159 other protocols used for inter-domain multicast, such as Multi- 160 protocol BGP (MBGP) [RFC4760]. The approach proposed in this 161 specification requires no packet format changes to the protocol and 162 no operational procedural changes to the multicast infrastructure 163 inside of a site when all sources and receivers reside in that site, 164 even when the site is LISP enabled. That is, internal operation of 165 multicast is unchanged regardless of whether or not the site is LISP 166 enabled or whether or not receivers exist in other sites which are 167 LISP-enabled. 169 Therefore, we see changes only to PIM-ASM [RFC4601], MSDP [RFC3618], 170 and PIM-SSM [RFC4607]. Bidir-PIM [RFC5015], which typically does not 171 run in an inter-domain environment is not addressed in depth in this 172 version of the specification. 174 Also, the current version of this specification does not describe 175 multicast-based Traffic Engineering relative to the TE-ITR (Traffic 176 Engineering based Ingress Tunnel Router) and TE-ETR (Traffic 177 Engineering based Egress Tunnel Router) descriptions in [LISP]. 178 Futher work is also needed to determine the detailed behavior for 179 multicast proxy ITRs (mPITRs) (Section 9.1.3), mtrace (Section 12), 180 and locator reachability (Section 6). Finally, further deployment 181 and experimentation would be useful to understand the real-life 182 performance of the LISP-Multicast solution. For instance, the design 183 optimizes for minimal state and control traffic in the core, but can 184 in some cases cause extra multicast traffic to be sent Section 8.1.2. 186 3. Definition of Terms 188 The terminology in this section is consistent with the definitions in 189 [LISP] but is extended specifically to deal with the application of 190 the terminology to multicast routing. 192 LISP-Multicast: a reference to the design in this specification. 193 That is, when any site that is participating in multicast 194 communication has been upgraded to be a LISP site, the operation 195 of control-plane and data-plane protocols is considered part of 196 the LISP-Multicast architecture. 198 Endpoint ID (EID): a 32-bit (for IPv4) or 128-bit (for IPv6) value 199 used in the source address field of the first (most inner) LISP 200 header of a multicast packet. The host obtains a destination 201 group address the same way it obtains one today, as it would when 202 it is a non-LISP site. The source EID is obtained via existing 203 mechanisms used to set a host's "local" IP address. An EID is 204 allocated to a host from an EID prefix block associated with the 205 site the host is located in. An EID can be used by a host to 206 refer to another host, as when it joins an SSM (S-EID,G) route 207 using IGMP version 3 [RFC4604]. LISP uses Provider Independent 208 (PI) blocks for EIDs; such EIDs MUST NOT be used as LISP RLOCs. 209 Note that EID blocks may be assigned in a hierarchical manner, 210 independent of the network topology, to facilitate scaling of the 211 mapping database. In addition, an EID block assigned to a site 212 may have site-local structure (subnetting) for routing within the 213 site; this structure is not visible to the global routing system. 215 Routing Locator (RLOC): the IPv4 or IPv6 address of an ingress 216 tunnel router (ITR), the router in the multicast source host's 217 site that encapsulates multicast packets. It is the output of a 218 EID-to-RLOC mapping lookup. An EID maps to one or more RLOCs. 219 Typically, RLOCs are numbered from topologically-aggregatable 220 blocks that are assigned to a site at each point to which it 221 attaches to the global Internet; where the topology is defined by 222 the connectivity of provider networks, RLOCs can be thought of as 223 Provider Assigned (PA) addresses. Multiple RLOCs can be assigned 224 to the same ITR device or to multiple ITR devices at a site. 226 Ingress Tunnel Router (ITR): a router which accepts an IP multicast 227 packet with a single IP header (more precisely, an IP packet that 228 does not contain a LISP header). The router treats this "inner" 229 IP destination multicast address opaquely so it doesn't need to 230 perform a map lookup on the group address because it is 231 topologically insignificant. The router then prepends an "outer" 232 IP header with one of its globally-routable RLOCs as the source 233 address field. This RLOC is known to other multicast receiver 234 sites which have used the mapping database to join a multicast 235 tree for which the ITR is the root. In general, an ITR receives 236 IP packets from site end systems on one side and sends LISP- 237 encapsulated multicast IP packets out all external interfaces 238 which have been joined. 240 An ITR would receive a multicast packet from a source inside of 241 its site when 1) it is on the path from the multicast source to 242 internally joined receivers, or 2) when it is on the path from the 243 multicast source to externally joined receivers. 245 Egress Tunnel Router (ETR): a router that is on the path from a 246 multicast source host in another site to a multicast receiver in 247 its own site. An ETR accepts a PIM Join/Prune message from a site 248 internal PIM router destined for the source's EID in the multicast 249 source site. The ETR maps the source EID in the Join/Prune 250 message to an RLOC address based on the EID-to-RLOC mapping. This 251 sets up the ETR to accept multicast encapsulated packets from the 252 ITR in the source multicast site. A multicast ETR decapsulates 253 multicast encapsulated packets and replicates them on interfaces 254 leading to internal receivers. 256 xTR: is a reference to an ITR or ETR when direction of data flow is 257 not part of the context description. xTR refers to the router that 258 is the tunnel endpoint. Used synonymously with the term "Tunnel 259 Router". For example, "An xTR can be located at the Customer Edge 260 (CE) router", meaning both ITR and ETR functionality can be at the 261 CE router. 263 LISP Header: a term used in this document to refer to the outer 264 IPv4 or IPv6 header, a UDP header, and a LISP header. An ITR 265 prepends headers and an ETR strips headers. A LISP encapsulated 266 multicast packet will have an "inner" header with the source EID 267 in the source field; an "outer" header with the source RLOC in the 268 source field: and the same globally unique group address in the 269 destination field of both the inner and outer header. 271 (S,G) State: the formal definition is in the PIM Sparse Mode 272 [RFC4601] specification. For this specification, the term is used 273 generally to refer to multicast state. Based on its topological 274 location, the (S,G) state resides in routers can be either 275 (S-EID,G) state (at a location where the (S,G) state resides) or 276 (S-RLOC,G) state (in the Internet core). 278 (S-EID,G) State: refers to multicast state in multicast source and 279 receiver sites where S-EID is the IP address of the multicast 280 source host (its EID). An S-EID can appear in an IGMPv3 report, 281 an MSDP SA message or a PIM Join/Prune message that travels inside 282 of a site. 284 (S-RLOC,G) State: refers to multicast state in the core where S is 285 a source locator (the IP address of a multicast ITR) of a site 286 with a multicast source. The (S-RLOC,G) is mapped from (S-EID,G) 287 entry by doing a mapping database lookup for the EID prefix that 288 S-EID maps to. An S-RLOC can appear in a PIM Join/Prune message 289 when it travels from an ETR to an ITR over the Internet core. 291 uLISP Site: a unicast only LISP site according to [LISP] which has 292 not deployed the procedures of this specification and therefore, 293 for multicast purposes, follows the procedures from Section 9. 295 mPETR: this is a multicast proxy-ETR that is responsible for 296 advertising a very coarse EID prefix which non-LISP and uLISP 297 sites can target their (S-EID,G) PIM Join/Prune message to. mPETRs 298 are used so LISP source multicast sites can send multicast packets 299 using source addresses from the EID namespace. mPETRs act as Proxy 300 ETRs for supporting multicast routing in a LISP infrastructure. 301 It is likely an uPITR [INTWORK] and a mPETR will be co-located 302 since the single device advertises a coarse EID-prefix in the 303 underlying unicast routing system. 305 Mixed Locator-Sets: this is a locator-set for a LISP database 306 mapping entry where the RLOC addresses in the locator-set are in 307 both IPv4 and IPv6 format. 309 Unicast Encapsulated PIM Join/Prune Message: this is a standard PIM 310 Join/Prune message (encapsulated in a LISP Encapsulated Control 311 Message with destination UDP port 4342) which is sent by ETRs at 312 multicast receiver sites to an ITR at a multicast source site. 313 This message is sent periodically as long as there are interfaces 314 in the oif-list for the (S-EID,G) entry the ETR is joining for. 316 4. Basic Overview 318 LISP, when used for unicast routing, increases the site's ability to 319 control ingress traffic flows. Egress traffic flows are controlled 320 by the IGP in the source site. For multicast, the IGP coupled with 321 PIM can decide which path multicast packets ingress. By using the 322 traffic engineering features of LISP, a multicast source site can 323 control the egress of its multicast traffic. By controlling the 324 priorities of locators from a mapping database entry, a source 325 multicast site can control which way multicast receiver sites join to 326 the source site. 328 At this point in time, we don't see a requirement for different 329 locator-sets, priority, and weight policies for multicast than we 330 have for unicast. However, when traffic engineering policies are 331 different for unicast versus multicast flows, it will be desirable to 332 use multicast-based priority and weight values in Map-Reply messages. 334 The fundamental multicast forwarding model is to encapsulate a 335 multicast packet into another multicast packet. An ITR will 336 encapsulate multicast packets received from sources that it serves in 337 a LISP multicast header. The destination group address from the 338 inner header is copied to the destination address of the outer 339 header. The inner source address is the EID of the multicast source 340 host and the outer source address is the RLOC of the encapsulating 341 ITR. 343 The LISP-Multicast architecture will follow this high-level protocol 344 and operational sequence: 346 1. Receiver hosts in multicast sites will join multicast content the 347 way they do today, they use IGMP. When they use IGMPv3 where 348 they specify source addresses, they use source EIDs, that is they 349 join (S-EID,G). If the multicast source is external to this 350 receiver site, the PIM Join/Prune message flows toward the ETRs, 351 finding the shortest exit (that is the closest exit for the Join/ 352 Prune message and the closest entrance for the multicast packet 353 to the receiver). 355 2. The ETR does a mapping database lookup for S-EID. If the mapping 356 is cached from a previous lookup (from either a previous Join/ 357 Prune for the source multicast site or a unicast packet that went 358 to the site), it will use the RLOC information from the mapping. 359 The ETR will use the same priority and weighting mechanism as for 360 unicast. So the source site can decide which way multicast 361 packets egress. 363 3. The ETR will build two PIM Join/Prune messages, one that contains 364 a (S-EID,G) entry that is unicast to the ITR that matches the 365 RLOC the ETR selects, and the other which contains a (S-RLOC,G) 366 entry so the core network can create multicast state from this 367 ETR to the ITR. 369 4. When the ITR gets the unicast Join/Prune message (see Section 3 370 for formal definition), it will process (S-EID,G) entries in the 371 message and propagate them inside of the site where it has 372 explicit routing information for EIDs via the IGP. When the ITR 373 receives the (S-RLOC,G) PIM Join/Prune message it will process it 374 like any other join it would get in today's Internet. The S-RLOC 375 address is the IP address of this ITR. 377 5. At this point there is (S-EID,G) state from the joining host in 378 the receiver multicast site to the ETR of the receiver multicast 379 site. There is (S-RLOC,G) state across the core network from the 380 ETR of the multicast receiver site to the ITR in the multicast 381 source site and (S-EID,G) state in the source multicast site. 382 Note, the (S-EID,G) state is the same S-EID in each multicast 383 site. As other ETRs join the same multicast tree, they can join 384 through the same ITR (in which case the packet replication is 385 done in the core) or a different ITR (in which case the packet 386 replication is done at the source site). 388 6. When a packet is originated by the multicast host in the source 389 site, it will flow to one or more ITRs which will prepend a LISP 390 header by copying the group address to the outer destination 391 address field and insert its own locator address in the outer 392 source address field. The ITR will look at its (S-RLOC,G) state, 393 where S-RLOC is its own locator address, and replicate the packet 394 on each interface a (S-RLOC,G) joined was received on. The core 395 has (S-RLOC,G) so where fanout occurs to multiple sites, a core 396 router will do packet replication. 398 7. When either the source site or the core replicates the packet, 399 the ETR will receive a LISP packet with a destination group 400 address. It will decapsulate packets because it has receivers 401 for the group. Otherwise, it would have not received the packets 402 because it would not have joined. The ETR decapsulates and does 403 a (S-EID,G) lookup in its multicast FIB to forward packets out 404 one or more interfaces to forward the packet to internal 405 receivers. 407 This architecture is consistent and scalable with the architecture 408 presented in [LISP] where multicast state in the core operates on 409 locators and multicast state at the sites operates on EIDs. 411 Alternatively, [LISP] also has a mechanism where (S-EID,G) state can 412 reside in the core through the use of RPF-vectors [RFC5496] in PIM 413 Join/Prune messages. However, few PIM implementations support RPF 414 vectors and LISP should avoid S-EID state in the core. See Section 5 415 for details. 417 However, we have some observations on the algorithm above. We can 418 scale the control plane but at the expense of sending data to sites 419 which may have not joined the distribution tree where the 420 encapsulated data is being delivered. For example, one site joins 421 (S-EID1,G) and another site joins (S-EID2,G). Both EIDs are in the 422 same multicast source site. Both multicast receiver sites join to 423 the same ITR with state (S-RLOC,G) where S-RLOC is the RLOC for the 424 ITR. The ITR joins both (S-EID1,G) and (S-EID2,G) inside of the 425 site. The ITR receives (S-RLOC,G) joins and populates the oif-list 426 state for it. Since both (S-EID1,G) and (S-EID2, G) map to the one 427 (S-RLOC,G) packets will be delivered by the core to both multicast 428 receiver sites even though each have joined a single source-based 429 distribution tree. This behavior is a consequence of the many-to-one 430 mapping between S-EIDs and a S-RLOC. 432 There is a possible solution to this problem which reduces the number 433 of many-to-one occurrences of (S-EID,G) entries aggregating into a 434 single (S-RLOC,G) entry. If a physical ITR can be assigned multiple 435 RLOC addresses and these addresses are advertised in mapping database 436 entries, then ETRs at receiver sites have more RLOC address options 437 and therefore can join different (RLOC,G) entries for each (S-EID,G) 438 entry joined at the receiver site. It would not scale to have a one- 439 to-one relationship between the number of S-EID sources at a source 440 site and the number of RLOCs assigned to all ITRs at the site, but we 441 can reduce the "n" to a smaller number in the "n-to-1" relationship. 442 And in turn, reduce the opportunity for data packets to be delivered 443 to sites for groups not joined. 445 5. Source Addresses versus Group Addresses 447 Multicast group addresses don't have to be associated with either the 448 EID or RLOC namespace. They actually are a namespace of their own 449 that can be treated as logical with relatively opaque allocation. 450 So, by their nature, they don't detract from an incremental 451 deployment of LISP-Multicast. 453 As for source addresses, as in the unicast LISP scenario, there is a 454 decoupling of identification from location. In a LISP site, packets 455 are originated from hosts using their allocated EIDs, those addresses 456 are used to identify the host as well as where in the site's topology 457 the host resides but not how and where it is attached to the 458 Internet. 460 Therefore, when multicast distribution tree state is created anywhere 461 in the network on the path from any multicast receiver to a multicast 462 source, EID state is maintained at the source and receiver multicast 463 sites, and RLOC state is maintained in the core. That is, a 464 multicast distribution tree will be represented as a 3-tuple of 465 {(S-EID,G) (S-RLOC,G) (S-EID,G)} where the first element of the 466 3-tuple is the state stored in routers from the source to one or more 467 ITRs in the source multicast site, the second element of the 3-tuple 468 is the state stored in routers downstream of the ITR, in the core, to 469 all LISP receiver multicast sites, and the third element in the 470 3-tuple is the state stored in the routers downstream of each ETR, in 471 each receiver multicast site, reaching each receiver. Note that 472 (S-EID,G) is the same in both the source and receiver multicast 473 sites. 475 The concatenation/mapping from the first element to the second 476 element of the 3-tuples is done by the ITR and from the second 477 element to the third element is done at the ETRs. 479 6. Locator Reachability Implications on LISP-Multicast 481 Multicast state as it is stored in the core is always (S,G) state as 482 it exists today or (S-RLOC,G) state as it will exist when LISP sites 483 are deployed. The core routers cannot distinguish one from the 484 other. They don't need to because it is state that RPFs against the 485 core routing tables in the RLOC namespace. The difference is where 486 the root of the distribution tree for a particular source is. In the 487 traditional multicast core, the source S is the source host's IP 488 address. For LISP-Multicast the source S is a single ITR of the 489 multicast source site. 491 An ITR is selected based on the LISP EID-to-RLOC mapping used when an 492 ETR propagates a PIM Join/Prune message out of a receiver multicast 493 site. The selection is based on the same algorithm an ITR would use 494 to select an ETR when sending a unicast packet to the site. In the 495 unicast case, the ITR can change on a per-packet basis depending on 496 the reachability of the ETR. So an ITR can change relatively easily 497 using local reachability state. However, in the multicast case, when 498 an ITR goes unreachable, new distribution tree state must be built 499 because the encapsulating root has changed. This is more significant 500 than an RPF-change event, where any router would typically locally 501 change its RPF-interface for its existing tree state. But when an 502 encapsulating LISP-Multicast ITR goes unreachable, new distribution 503 state must be rebuilt and reflect the new encapsulator. Therefore, 504 when an ITR goes unreachable, all ETRs that are currently joined to 505 that ITR will have to trigger a new Join/Prune message for (S-RLOC,G) 506 to the new ITR as well as send a unicast encapsulated Join/Prune 507 message telling the new ITR which (S-EID,G) is being joined. 509 This issue can be mitigated by using anycast addressing for the ITRs 510 so the problem does reduce to an RPF change in the core, but still 511 requires a unicast encapsulated Join/Prune message to tell the new 512 ITR about (S-EID,G). The problem with this approach is that the ETR 513 really doesn't know when the ITR has changed so the new anycast ITR 514 will get the (S-EID,G) state only when the ETR sends it the next time 515 during its periodic sending procedures. 517 7. Multicast Protocol Changes 519 A number of protocols are used today for inter-domain multicast 520 routing: 522 IGMPv1-v3, MLDv1-v2: These protocols do not require any changes for 523 LISP-Multicast for two reasons. One being that they are link- 524 local and not used over site boundaries and second they advertise 525 group addresses that don't need translation. Where source 526 addresses are supplied in IGMPv3 and MLDv2 messages, they are 527 semantically regarded as EIDs and don't need to be converted to 528 RLOCs until the multicast tree-building protocol, such as PIM, is 529 received by the ETR at the site boundary. Addresses used for IGMP 530 and MLD come out of the source site's allocated addresses which 531 are therefore from the EID namespace. 533 MBGP: Even though MBGP is not a multicast routing protocol, it is 534 used to find multicast sources when the unicast BGP peering 535 topology and the multicast MBGP peering topology are not 536 congruent. When MBGP is used in a LISP-Multicast environment, the 537 prefixes which are advertised are from the RLOC namespace. This 538 allows receiver multicast sites to find a path to the source 539 multicast site's ITRs. MBGP peering addresses will be from the 540 RLOC namespace. There are no MBGP protocol changes required to 541 support LISP-Multicast. 543 MSDP: MSDP is used to announce active multicast sources to other 544 routing domains (or LISP sites). The announcements come from the 545 PIM Rendezvous Points (RPs) from sites where there are active 546 multicast sources sending to various groups. In the context of 547 LISP-Multicast, the source addresses advertised in MSDP will 548 semantically be from the EID namespace since they describe the 549 identity of a source multicast host. It will be true that the 550 state stored in MSDP caches from core routers will be from the EID 551 namespace. An RP address inside of site will be from the EID 552 namespace so it can be advertised and reached by internal unicast 553 routing mechanism. However, for MSDP peer-RPF checking to work 554 properly across sites, the RP addresses must be converted or 555 mapped into a routable address that is advertised and maintained 556 in the BGP routing tables in the core. MSDP peering addresses can 557 come out of either the EID or a routable address namespace. And 558 the choice can be made unilaterally because the ITR at the site 559 will determine which namespace the destination peer address is out 560 of by looking in the mapping database service. There are no MSDP 561 protocol changes required to support LISP-Multicast. 563 PIM-SSM: In the simplest form of distribution tree building, when 564 PIM operates in SSM mode, a source distribution tree is built and 565 maintained across site boundaries. In this case, there is a small 566 modification to the operation of the PIM protocol (but not to any 567 message format) to support taking a Join/Prune message originated 568 inside of a LISP site with embedded addresses from the EID 569 namespace and converting them to addresses from the RLOC namespace 570 when the Join/Prune message crosses a site boundary. This is 571 similar to the requirements documented in [RFC5135]. 573 PIM-Bidir: Bidirectional PIM is typically run inside of a routing 574 domain, but if deployed in an inter-domain environment, one would 575 have to decide if the RP address of the shared-tree would be from 576 the EID namespace or the RLOC namespace. If the RP resides in a 577 site-based router, then the RP address is from the EID namespace. 578 If the RP resides in the core where RLOC addresses are routed, 579 then the RP address is from the RLOC namespace. This could be 580 easily distinguishable if the EID address were well-known address 581 allocation block from the RLOC namespace. Also, when using 582 Embedded-RP for RP determination [RFC3956], the format of the 583 group address could indicate the namespace the RP address is from. 584 However, refer to Section 10 for considerations core routers need 585 to make when using Embedded-RP IPv6 group addresses. When using 586 Bidir-PIM for inter-domain multicast routing, it is recommended to 587 use staticly configured RPs so core routers think the Bidir group 588 is associated with an ITR's RLOC as the RP address and site 589 routers think the Bidir group is associated with the site resident 590 RP with an EID address. With respect to DF-election in Bidir PIM, 591 no changes are required since all messaging and addressing is 592 link-local. 594 PIM-ASM: The ASM mode of PIM, the most popular form of PIM, is 595 deployed in the Internet today is by having shared-trees within a 596 site and using source-trees across sites. By the use of MSDP and 597 PIM-SSM techniques described above, we can get multicast 598 connectivity across LISP sites. Having said that, that means 599 there are no special actions required for processing (*,G) or 600 (S,G,R) Join/Prune messages since they all operate against the 601 shared-tree which is site resident. Just like with ASM, there is 602 no (*,G) in the core when LISP-Multicast is in use. This is also 603 true for the RP-mapping mechanisms Auto-RP and BSR. 605 Based on the protocol description above, the conclusion is that there 606 are no protocol message format changes, just a translation function 607 performed at the control-plane. This will make for an easier and 608 faster transition for LISP since fewer components in the network have 609 to change. 611 It should also be stated just like it is in [LISP] that no host 612 changes, whatsoever, are required to have a multicast source host 613 send multicast packets and for a multicast receiver host to receive 614 multicast packets. 616 8. LISP-Multicast Data-Plane Architecture 618 The LISP-Multicast data-plane operation conforms to the operation and 619 packet formats specified in [LISP]. However, encapsulating a 620 multicast packet from an ITR is a much simpler process. The process 621 is simply to copy the inner group address to the outer destination 622 address. And to have the ITR use its own IP address (its RLOC) as 623 the source address. The process is simpler for multicast because 624 there is no EID-to-RLOC mapping lookup performed during packet 625 forwarding. 627 In the decapsulation case, the ETR simply removes the outer header 628 and performs a multicast routing table lookup on the inner header 629 (S-EID,G) addresses. Then the oif-list for the (S-EID,G) entry is 630 used to replicate the packet on site-facing interfaces leading to 631 multicast receiver hosts. 633 There is no Data-Probe logic for ETRs as there can be in the unicast 634 forwarding case. 636 8.1. ITR Forwarding Procedure 638 The following procedure is used by an ITR, when it receives a 639 multicast packet from a source inside of its site: 641 1. A multicast data packet sent by a host in a LISP site will have 642 the source address equal to the host's EID and the destination 643 address equal to the group address of the multicast group. It is 644 assumed the group information is obtained by current methods. 645 The same is true for a multicast receiver to obtain the source 646 and group address of a multicast flow. 648 2. When the ITR receives a multicast packet, it will have both S-EID 649 state and S-RLOC state stored. Since the packet was received on 650 a site-facing interface, the RPF lookup is based on the S-EID 651 state. If the RPF check succeeds, then the oif-list contains 652 interfaces that are site-facing and external-facing. For the 653 site-facing interfaces, no LISP header is prepended. For the 654 external-facing interfaces a LISP header is prepended. When the 655 ITR prepends a LISP header, it uses its own RLOC address as the 656 source address and copies the group address supplied by the IP 657 header the host built as the outer destination address. 659 8.1.1. Multiple RLOCs for an ITR 661 Typically, an ITR will have a single RLOC address but in some cases 662 there could be multiple RLOC addresses assigned from either the same 663 or different service providers. In this case when (S-RLOC,G) Join/ 664 Prune messages are received for each RLOC, there is a oif-list 665 merging action that must take place. Therefore, when a packet is 666 received from a site-facing interface that matches on a (S-EID,G) 667 entry, the interfaces of the oif-list from all (RLOC,G) entries 668 joined to the ITR as well as the site-facing oif-list joined for 669 (S-EID,G) must be part be included in packet replication. In 670 addition to replicating for all types of oif-lists, each oif entry 671 must be tagged with the RLOC address, so encapsulation uses the outer 672 source address for the RLOC joined. 674 8.1.2. Multiple ITRs for a LISP Source Site 676 Note when ETRs from different multicast receiver sites receive 677 (S-EID,G) joins, they may select a different S-RLOC for a multicast 678 source site due to policy (the multicast ITR can return different 679 multicast priority and weight values per ETR Map-Request). In this 680 case, the same (S-EID,G) is being realized by different (S-RLOC,G) 681 state in the core. This will not result in duplicate packets because 682 each ITR in the multicast source site will choose their own RLOC for 683 the source address for encapsulated multicast traffic. The RLOC 684 addresses are the ones joined by remote multicast ETRs. 686 When different (S-EID,G) traffic is combined into a single (RLOC,G) 687 core distribution tree, this may cause traffic to go to a receiver 688 multicast site when it does not need to. This happens when one 689 receiver multicast site joins (S1-EID,Gi) through a core distribution 690 tree of (RLOC1,Gi) and another multicast receiver site joins (S2- 691 EID,Gi) through the same core distribution tree of (RLOC1,Gi). When 692 ETRs decapsulate such traffic, they should know from their local 693 (S-EID,G) state if the packet should be forwarded. If there is no 694 (S-EID,G) state that matches the inner packet header, the packet is 695 discarded. 697 8.2. ETR Forwarding Procedure 699 The following procedure is used by an ETR, when it receives a 700 multicast packet from a source outside of its site: 702 1. When a multicast data packet is received by an ETR on an 703 external-facing interface, it will do an RPF lookup on the S-RLOC 704 state it has stored. If the RPF check succeeds, the interfaces 705 from the oif-list are used for replication to interfaces that are 706 site-facing as well as interfaces that are external-facing (this 707 ETR can also be a transit multicast router for receivers outside 708 of its site). When the packet is to be replicated for an 709 external-facing interface, the LISP encapsulation header are not 710 stripped. When the packet is replicated for a site-facing 711 interface, the encapsulation header is stripped. 713 2. The packet without a LISP header is now forwarded down the 714 (S-EID,G) distribution tree in the receiver multicast site. 716 8.3. Replication Locations 718 Multicast packet replication can happen in the following topological 719 locations: 721 o In an IGP multicast router inside a site which operates on S-EIDs. 723 o In a transit multicast router inside of the core which operates on 724 S-RLOCs. 726 o At one or more ETR routers depending on the path a Join/Prune 727 message exits a receiver multicast site. 729 o At one or more ITR routers in a source multicast site depending on 730 what priorities are returned in a Map-Reply to receiver multicast 731 sites. 733 In the last case the source multicast site can do replication rather 734 than having a single exit from the site. But this only can occur 735 when the priorities in the Map-Reply are modified for different 736 receiver multicast site so that the PIM Join/Prune messages arrive at 737 different ITRs. 739 This policy technique, also used in [ALT] for unicast, is useful for 740 multicast to mitigate the problems of changing distribution tree 741 state as discussed in Section 6. 743 9. LISP-Multicast Interworking 745 This section will describe the multicast corollary to [INTWORK] which 746 describes the interworking of multicast routing among LISP and non- 747 LISP sites. 749 9.1. LISP and non-LISP Mixed Sites 751 Since multicast communication can involve more than two entities to 752 communicate together, the combinations of interworking scenarios are 753 more involved. However, the state maintained for distribution trees 754 at the sites is the same regardless of whether or not the site is 755 LISP enabled or not. So most of the implications are in the core 756 with respect to storing routable EID prefixes from either PA or PI 757 blocks. 759 Before we enumerate the multicast interworking scenarios, we must 760 define 3 deployment states of a site: 762 o A non-LISP site which will run PIM-SSM or PIM-ASM with MSDP as it 763 does today. The addresses for the site are globally routable. 765 o A site that deploys LISP for unicast routing. The addresses for 766 the site are not globally routable. Let's define the name for 767 this type of site as a uLISP site. 769 o A site that deploys LISP for both unicast and multicast routing. 770 The addresses for the site are not globally routable. Let's 771 define the name for this type of site as a LISP-Multicast site. 773 We will not consider a LISP site enabled for multicast purposes only 774 but do consider a uLISP site as documented in [INTWORK]. In this 775 section we don't discuss how a LISP site sends multicast packets when 776 all receiver sites are LISP-Multicast enabled; that has been 777 discussed in previous sections. 779 The following scenarios exist to make LISP-Multicast sites interwork 780 with non-LISP-Multicast sites: 782 1. A LISP site must be able to send multicast packets to receiver 783 sites which are a mix of non-LISP sites and uLISP sites. 785 2. A non-LISP site must be able to send multicast packets to 786 receiver sites which are a mix of non-LISP sites and uLISP sites. 788 3. A non-LISP site must be able to send multicast packets to 789 receiver sites which are a mix of LISP sites, uLISP sites, and 790 non-LISP sites. 792 4. A uLISP site must be able to send multicast packets to receiver 793 sites which are a mix of LISP sites, uLISP sites, and non-LISP 794 sites. 796 5. A LISP site must be able to send multicast packets to receiver 797 sites which are a mix of LISP sites, uLISP sites, and non-LISP 798 sites. 800 9.1.1. LISP Source Site to non-LISP Receiver Sites 802 In the first scenario, a site is LISP capable for both unicast and 803 multicast traffic and as such operates on EIDs. Therefore there is a 804 possibility that the EID prefix block is not routable in the core. 805 For LISP receiver multicast sites this isn't a problem but for non- 806 LISP or uLISP receiver multicast sites, when a PIM Join/Prune message 807 is received by the edge router, it has no route to propagate the 808 Join/Prune message out of the site. This is no different than the 809 unicast case that LISP-NAT in [INTWORK] solves. 811 LISP-NAT allows a unicast packet that exits a LISP site to get its 812 source address mapped to a globally routable address before the ITR 813 realizes that it should not encapsulate the packet destined to a non- 814 LISP site. For a multicast packet to leave a LISP site, distribution 815 tree state needs to be built so the ITR can know where to send the 816 packet. So the receiver multicast sites need to know about the 817 multicast source host by its routable address and not its EID 818 address. When this is the case, the routable address is the 819 (S-RLOC,G) state that is stored and maintained in the core routers. 820 It is important to note that the routable address for the host cannot 821 be the same as an RLOC for the site because we want the ITRs to 822 process a received PIM Join/Prune message from an external-facing 823 interface to be propagated inside of the site so the site-part of the 824 distribution tree is built. 826 Using a globally routable source address allows non-LISP and uLISP 827 multicast receiver to join, create, and maintain a multicast 828 distribution tree. However, the LISP multicast receiver site will 829 want to perform an EID-to-RLOC mapping table lookup when a PIM Join/ 830 Prune message is received on a site-facing interface. It does this 831 because it wants to find a (S-RLOC,G) entry to Join in the core. So 832 we have a conflict of behavior between the two types of sites. 834 The solution to this problem is the same as when an ITR wants to send 835 a unicast packet to a destination site but needs determine if the 836 site is LISP capable or not. When it is not LISP capable, the ITR 837 does not encapsulate the packet. So for the multicast case, when ETR 838 receives a PIM Join/Prune message for (S-EID,G) state, it will do a 839 mapping table lookup on S-EID. In this case, S-EID is not in the 840 mapping database because the source multicast site is using a 841 routable address and not an EID prefix address. So the ETR knows to 842 simply propagate the PIM Join/Prune message to a external-facing 843 interface without converting the (S-EID,G) because it is an (S,G) 844 where S is routable and reachable via core routing tables. 846 Now that the multicast distribution tree is built and maintained from 847 any non-LISP or uLISP receiver multicast site, the way packet 848 forwarding model is performed can be explained. 850 Since the ITR in the source multicast site has never received a 851 unicast encapsulated PIM Join/Prune message from any ETR in a 852 receiver multicast site, it knows there are no LISP-Multicast 853 receiver sites. Therefore, there is no need for the ITR to 854 encapsulate data. Since it will know a priori (via configuration) 855 that its site's EIDs are not routable (and not registered to the 856 mapping database system), it assumes that the multicast packets from 857 the source host are sent by a routable address. That is, it is the 858 responsibility of the multicast source host's system administrator to 859 ensure that the source host sends multicast traffic using a routable 860 source address. When this happens, the ITR acts simply as a router 861 and forwards the multicast packet like an ordinary multicast router. 863 There is an alternative to using a LISP-NAT scheme just like there is 864 for unicast [INTWORK] forwarding by using Proxy Tunnel Routers 865 (PxTRs). This can work the same way for multicast routing as well, 866 but the difference is that non-LISP and uLISP sites will send PIM 867 Join/Prune messages for (S-EID,G) which make their way in the core to 868 multicast PxTRs. Let's call this use of a PxTR as a "Multicast 869 Proxy-ETR" (or mPETR). Since the mPETRs advertise very coarse EID 870 prefixes, they draw the PIM Join/Prune control traffic making them 871 the target of the distribution tree. To get multicast packets from 872 the LISP source multicast sites, the tree needs to be built on the 873 path from the mPETR to the LISP source multicast site. To make this 874 happen the mPETR acts as a "Proxy ETR" (where in unicast it acts as a 875 "Proxy ITR", or an uPITR [INTWORK]). 877 The existence of mPETRs in the core allows source multicast site ITRs 878 to encapsulate multicast packets according to (S-RLOC,G) state. The 879 (S-RLOC,G) state is built from the mPETRs to the multicast ITRs. The 880 encapsulated multicast packets are decapsulated by mPETRs and then 881 forwarded according to (S-EID,G) state. The (S-EID,G) state is built 882 from the non-LISP and uLISP receiver multicast sites to the mPETRs. 884 9.1.2. Non-LISP Source Site to non-LISP Receiver Sites 886 Clearly non-LISP multicast sites can send multicast packets to non- 887 LISP receiver multicast sites. That is what they do today. However, 888 discussion is required to show how non-LISP multicast sites send 889 multicast packets to uLISP receiver multicast sites. 891 Since uLISP receiver multicast sites are not targets of any (S,G) 892 state, they simply send (S,G) PIM Join/Prune messages toward the non- 893 LISP source multicast site. Since the source multicast site, in this 894 case has not been upgraded to LISP, all multicast source host 895 addresses are routable. So this case is simplified to where a uLISP 896 receiver multicast site looks to the source multicast site as a non- 897 LISP receiver multicast site. 899 9.1.3. Non-LISP Source Site to Any Receiver Site 901 When a non-LISP source multicast site has receivers in either a non- 902 LISP/uLISP site or a LISP site, one needs to decide how the LISP 903 receiver multicast site will attach to the distribution tree. We 904 know from Section 9.1.2 that non-LISP and uLISP receiver multicast 905 sites can join the distribution tree, but a LISP receiver multicast 906 site ETR will need to know if the source address of the multicast 907 source host is routable or not. We showed in Section 9.1.1 that an 908 ETR, before it sends a PIM Join/Prune message on an external-facing 909 interface, does a EID-to-RLOC mapping lookup to determine if it 910 should convert the (S,G) state from a PIM Join/Prune message received 911 on a site-facing interface to a (S-RLOC,G). If the lookup fails, the 912 ETR can conclude the source multicast site is a non-LISP site so it 913 simply forwards the Join/Prune message (it also doesn't need to send 914 a unicast encapsulated Join/Prune message because there is no ITR in 915 a non-LISP site and there is namespace continuity between the ETR and 916 source). 918 For a non-LISP source multicast site, (S-EID,G) state could be 919 limited to the edges of the network with the use of multicast proxy- 920 ITRs (mPITRs). The mPITRs can take native, unencapsulated multicast 921 packets from non-LISP source multicast and uLISP sites and 922 encapsulate them to ETRs in receiver multicast sites or to mPETRs 923 that can decapsulate for non-LISP receiver multicast or uLISP sites. 924 The mPITRs are responsible for sending (S-EID,G) joins to the non- 925 LISP source multicast site. To connect the distribution trees 926 together, multicast ETRs will need to be configured with the mPITR's 927 RLOC addresses so they can send both (S-RLOC,G) joins to build a 928 distribution tree to the mPITR as well as for sending unicast joins 929 to mPITRs so they can propogate (S-EID,G) joins into source multicast 930 sites. The use of mPITRs is undergoing more study and is work in 931 progress. 933 9.1.4. Unicast LISP Source Site to Any Receiver Sites 935 In the last section, it was explained how an ETR in a multicast 936 receiver site can determine if a source multicast site is LISP- 937 enabled by looking into the mapping database. When the source 938 multicast site is a uLISP site, it is LISP enabled but the ITR, by 939 definition is not capable of doing multicast encapsulation. So for 940 the purposes of multicast routing, the uLISP source multicast site is 941 treated as non-LISP source multicast site. 943 Non-LISP receiver multicast sites can join distribution trees to a 944 uLISP source multicast site since the source site behaves, from a 945 forwarding perspective, as a non-LISP source site. This is also the 946 case for a uLISP receiver multicast site since the ETR does not have 947 multicast functionality built-in or enabled. 949 Special considerations are required for LISP receiver multicast sites 950 since they think the source multicast site is LISP capable, the ETR 951 cannot know if ITR is LISP-Multicast capable. To solve this problem, 952 each mapping database entry will have a multicast 2-tuple (Mpriority, 953 Mweight) per RLOC. When the Mpriority is set to 255, the site is 954 considered not multicast capable. So an ETR in a LISP receiver 955 multicast site can distinguish whether a LISP source multicast site 956 is LISP-Multicast site from a uLISP site. 958 9.1.5. LISP Source Site to Any Receiver Sites 960 When a LISP source multicast site has receivers in LISP, non-LISP, 961 and uLISP receiver multicast sites, it has a conflict about how it 962 sends multicast packets. The ITR can either encapsulate or natively 963 forward multicast packets. Since the receiver multicast sites are 964 heterogeneous in their behavior, one packet forwarding mechanism 965 cannot satisfy both. However, if a LISP receiver multicast site acts 966 like a uLISP site then it could receive packets like a non-LISP 967 receiver multicast site making all receiver multicast sites have 968 homogeneous behavior. However, this poses the following issues: 970 o LISP-NAT techniques with routable addresses would be required in 971 all cases. 973 o Or alternatively, mPETR deployment would be required forcing 974 coarse EID prefix advertisement in the core. 976 o But what is most disturbing is that when all sites that 977 participate are LISP-Multicast sites but then a non-LISP or uLISP 978 site joins the distribution tree, then the existing joined LISP 979 receiver multicast sites would have to change their behavior. 980 This would create too much dynamic tree-building churn to be a 981 viable alternative. 983 So the solution space options are: 985 1. Make the LISP ITR in the source multicast site send two packets, 986 one that is encapsulated with (S-RLOC,G) to reach LISP receiver 987 multicast sites and another that is not encapsulated with 988 (S-EID,G) to reach non-LISP and uLISP receiver multicast sites. 990 2. Make the LISP ITR always encapsulate packets with (S-RLOC,G) to 991 reach LISP-Multicast sites and to reach mPETRs that can 992 decapsulate and forward (S-EID,G) packets to non-LISP and uLISP 993 receiver multicast sites. 995 9.2. LISP Sites with Mixed Address Families 997 A LISP database mapping entry that describes the locator-set, 998 Mpriority and Mweight per locator address (RLOC), for an EID prefix 999 associated with a site could have RLOC addresses in either IPv4 or 1000 IPv6 format. When a mapping entry has a mix of RLOC formatted 1001 addresses, it is an implicit advertisement by the site that it is a 1002 dual-stack site. That is, the site can receive IPv4 or IPv6 unicast 1003 packets. 1005 To distinguish if the site can receive dual-stack unicast packets as 1006 well as dual-stack multicast packets, the Mpriority value setting 1007 will be relative to an IPv4 or IPv6 RLOC See [LISP] for packet format 1008 details. 1010 If you consider the combinations of LISP, non-LISP, and uLISP sites 1011 sharing the same distribution tree and considering the capabilities 1012 of supporting IPv4, IPv6, or dual-stack, the number of total 1013 combinations grows beyond comprehension. 1015 Using some combinatorial math, we have the following profiles of a 1016 site and the combinations that can occur: 1018 1. LISP-Multicast IPv4 Site 1020 2. LISP-Multicast IPv6 Site 1022 3. LISP-Multicast Dual-Stack Site 1024 4. uLISP IPv4 Site 1026 5. uLISP IPv6 Site 1027 6. uLISP Dual-Stack Site 1029 7. non-LISP IPv4 Site 1031 8. non-LISP IPv6 Site 1033 9. non-LISP Dual-Stack Site 1035 Lets define (m n) = m!/(n!*(m-n)!), pronounced "m choose n" to 1036 illustrate some combinatorial math below. 1038 When 1 site talks to another site, the combinatorial is (9 2), when 1 1039 site talks to another 2 sites, the combinatorial is (9 3). If sum 1040 this up to (9 9), we have: 1042 (9 2) + (9 3) + (9 4) + (9 5) + (9 6) + (9 7) + (9 8) + (9 9) = 1044 36 + 84 + 126 + 126 + 84 + 36 + 9 + 1 1046 Which results in the total number of cases to be considered at 502. 1048 This combinatorial gets even worse when you consider a site using one 1049 address family inside of the site and the xTRs use the other address 1050 family (as in using IPv4 EIDs with IPv6 RLOCs or IPv6 EIDs with IPv4 1051 RLOCs). 1053 To rationalize this combinatorial nightmare, there are some 1054 guidelines which need to be put in place: 1056 o Each distribution tree shared between sites will either be an IPv4 1057 distribution tree or an IPv6 distribution tree. Therefore, we can 1058 avoid head-end replication by building and sending packets on each 1059 address family based distribution tree. Even though there might 1060 be an urge to do multicast packet translation from one address 1061 family format to the other, it is a non-viable over-complicated 1062 urge. Multicast ITRs will only encapsulate packets where the 1063 inner and outer headers are from the same address family. 1065 o All LISP sites on a multicast distribution tree must share a 1066 common address family which is determined by the source site's 1067 locator-set in its LISP database mapping entry. All receiver 1068 multicast sites will use the best RLOC priority controlled by the 1069 source multicast site. This is true when the source site is 1070 either LISP-Multicast or uLISP capable. This means that priority- 1071 based policy modification is prohibited. When a receiver 1072 multicast site ETR receives a (S-EID,G) join, it must select a 1073 S-RLOC for the same address family as S-EID. 1075 o When a multicast locator-set has more than one locator, only 1076 locators from the same address-family MUST be set to the same best 1077 priority value. A mixed locator-set can exist (for unicast use), 1078 but the multicast priorities MUST be the set for the same address 1079 family locators. 1081 o When the source site is not LISP capable, it is up to how 1082 receivers find the source and group information for a multicast 1083 flow. That mechanism decides the address family for the flow. 1085 9.3. Making a Multicast Interworking Decision 1087 This Multicast Interworking section has shown all combinations of 1088 multicast connectivity that could occur. As you might have already 1089 concluded, this can be quite complicated and if the design is too 1090 ambitious, the dynamics of the protocol could cause a lot of 1091 instability. 1093 The trade-off decisions are hard to make and we want the same single 1094 solution to work for both IPv4 and IPv6 multicast. It is imperative 1095 to have an incrementally deployable solution for all of IPv4 unicast 1096 and multicast and IPv6 unicast and multicast while minimizing (or 1097 eliminating) both unicast and multicast EID namespace state. 1099 Therefore the design decision to go with uPITRs [INTWORK] for unicast 1100 routing and mPETRs for multicast routing seems to be the sweet spot 1101 in the solution space so we can optimize state requirements and avoid 1102 head-end data replication at ITRs. 1104 10. Considerations when RP Addresses are Embedded in Group Addresses 1106 When ASM and PIM-Bidir is used in an IPv6 inter-domain environment, a 1107 technique exists to embed the unicast address of an RP in a IPv6 1108 group address [RFC3956]. When routers in end sites process a PIM 1109 Join/Prune message which contain an embedded-RP group address, they 1110 extract the RP address from the group address and treat it from the 1111 EID namespace. However, core routers do not have state for the EID 1112 namespace, need to extract an RP address from the RLOC namespace. 1114 Therefore, it is the responsibility of ETRs in multicast receiver 1115 sites to map the group address into a group address where the 1116 embedded-RP address is from the RLOC namespace. The mapped RP- 1117 address is obtained from a EID-to-RLOC mapping database lookup. The 1118 ETR will also send a unicast (*,G) Join/Prune message to the ITR so 1119 the branch of the distribution tree from the source site resident RP 1120 to the ITR is created. 1122 This technique is no different than the techniques described in this 1123 specification for translating (S,G) state and propagating Join/Prune 1124 messages into the core. The only difference is that the (*,G) state 1125 in Join/Prune messages are mapped because they contain unicast 1126 addresses encoded in an Embedded-RP group address. 1128 11. Taking Advantage of Upgrades in the Core 1130 If the core routers are upgraded to support [RFC5496], then we can 1131 pass EID specific data through the core without, possibly, having to 1132 store the state in the core. 1134 By doing this we can eliminate the ETR from unicast encapsulating PIM 1135 Join/Prune messages to the source site's ITR. 1137 However, this solution is restricted to a small set of workable cases 1138 which would not be good for general use of LISP-Multicast. In 1139 addition due to slow convergence properties, it is not being 1140 recommended for LISP-Multicast. 1142 12. Mtrace Considerations 1144 Mtrace functionality must be consistent with unicast traceroute 1145 functionality where all hops from multicast receiver to multicast 1146 source are visible. 1148 The design for mtrace for use in LISP-Multicast environments is to be 1149 determined but should build upon the mtrace version 2 specified in 1150 [MTRACE]. 1152 13. Security Considerations 1154 Refer to the [LISP] specification. 1156 14. Acknowledgments 1158 The authors would like to gratefully acknowledge the people who have 1159 contributed discussion, ideas, and commentary to the making of this 1160 proposal and specification. People who provided expert review were 1161 Scott Brim, Greg Shepherd, and Dave Oran. Other commentary from 1162 discussions at Summer 2008 Dublin IETF were Toerless Eckert and 1163 Ijsbrand Wijnands. 1165 We would also like to thank the MBONED working group for constructive 1166 and civil verbal feedback when this draft was presented at the Fall 1167 2008 IETF in Minneapolis. In particular, good commentary came from 1168 Tom Pusateri, Steve Casner, Marshall Eubanks, Dimitri Papadimitriou, 1169 Ron Bonica, Lenny Guardino, Alia Atlas, Jesus Arango, and Jari Arkko. 1171 An expert review of this specification was done by Yiqun Cai and 1172 Liming Wei. We thank them for their detailed comments. 1174 This work originated in the Routing Research Group (RRG) of the IRTF. 1175 The individual submission [MLISP] was converted into this IETF LISP 1176 working group draft. 1178 15. IANA Considerations 1180 This document makes no request of the IANA. 1182 16. References 1184 16.1. Normative References 1186 [INTWORK] Lewis, D., Meyer, D., and D. Farinacci, "Interworking LISP 1187 with IPv4 and IPv6", draft-ietf-lisp-interworking-02.txt 1188 (work in progress). 1190 [LISP] Farinacci, D., Fuller, V., Meyer, D., and D. Lewis, 1191 "Locator/ID Separation Protocol (LISP)", 1192 draft-ietf-lisp-15.txt (work in progress). 1194 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1195 Requirement Levels", BCP 14, RFC 2119, March 1997. 1197 [RFC3618] Fenner, B. and D. Meyer, "Multicast Source Discovery 1198 Protocol (MSDP)", RFC 3618, October 2003. 1200 [RFC3956] Savola, P. and B. Haberman, "Embedding the Rendezvous 1201 Point (RP) Address in an IPv6 Multicast Address", 1202 RFC 3956, November 2004. 1204 [RFC4601] Fenner, B., Handley, M., Holbrook, H., and I. Kouvelas, 1205 "Protocol Independent Multicast - Sparse Mode (PIM-SM): 1206 Protocol Specification (Revised)", RFC 4601, August 2006. 1208 [RFC4604] Holbrook, H., Cain, B., and B. Haberman, "Using Internet 1209 Group Management Protocol Version 3 (IGMPv3) and Multicast 1210 Listener Discovery Protocol Version 2 (MLDv2) for Source- 1211 Specific Multicast", RFC 4604, August 2006. 1213 [RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for 1214 IP", RFC 4607, August 2006. 1216 [RFC4760] Bates, T., Chandra, R., Katz, D., and Y. Rekhter, 1217 "Multiprotocol Extensions for BGP-4", RFC 4760, 1218 January 2007. 1220 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 1221 "Bidirectional Protocol Independent Multicast (BIDIR- 1222 PIM)", RFC 5015, October 2007. 1224 [RFC5135] Wing, D. and T. Eckert, "IP Multicast Requirements for a 1225 Network Address Translator (NAT) and a Network Address 1226 Port Translator (NAPT)", BCP 135, RFC 5135, February 2008. 1228 [RFC5496] Wijnands, IJ., Boers, A., and E. Rosen, "The Reverse Path 1229 Forwarding (RPF) Vector TLV", RFC 5496, March 2009. 1231 16.2. Informative References 1233 [ALT] Farinacci, D., Fuller, V., and D. Meyer, "LISP Alternative 1234 Topology (LISP-ALT)", draft-ietf-lisp-alt-08.txt (work in 1235 progress). 1237 [MLISP] Farinacci, D., Meyer, D., Zwiebel, J., and S. Venaas, 1238 "LISP for Multicast Environments", 1239 draft-farinacci-lisp-multicast-01.txt (work in progress). 1241 [MTRACE] Asaeda, H., Jinmei, T., Fenner, W., and S. Casner, "Mtrace 1242 Version 2: Traceroute Facility for IP Multicast", 1243 draft-ietf-mboned-mtrace-v2-08.txt (work in progress). 1245 Appendix A. Document Change Log 1247 A.1. Changes to draft-ietf-lisp-multicast-08.txt 1249 o Posted September 2011. Minor editorial changes from Jari's 1250 commentary. 1252 A.2. Changes to draft-ietf-lisp-multicast-07.txt 1254 o Posted July 2011. Fixing IDnits errors. 1256 A.3. Changes to draft-ietf-lisp-multicast-06.txt 1258 o Posted June 2011 to complete working group last call. 1260 o Added paragraph to section 8.1.2 based on Jesus comment about 1261 making it more clear what happens when two (S-EID,G) trees use the 1262 same (RLOC,G) tree. 1264 o Make more references to [INTWORK] when mentioning uPITRs and 1265 uPETRs. 1267 o Made many changes based on editorial and wordsmithing comments 1268 from Alia. 1270 A.4. Changes to draft-ietf-lisp-multicast-05.txt 1272 o Posted April 2011 to reset expiration timer. 1274 o Updated references. 1276 A.5. Changes to draft-ietf-lisp-multicast-04.txt 1278 o Posted October 2010 to reset expiration timer. 1280 o Updated references. 1282 A.6. Changes to draft-ietf-lisp-multicast-03.txt 1284 o Posted April 2010. 1286 o Added section 8.1.2 to address Joel Halpern's comment about 1287 receiver sites joining the same source site via 2 different RLOCs, 1288 each being a separate ITR. 1290 o Change all occurences of "mPTR" to "mPETR" to become more 1291 consistent with uPITRs and uPETRs described in [INTWORK]. That 1292 is, an mPETR is a LISP multicast router that decapsulates 1293 multicast packets that are encapsulated to it by ITRs in multicast 1294 source sites. 1296 o Add clarifications in section 9 about how homogeneous multicast 1297 encapsulation should occur. As well as describing in this 1298 section, how to deal with mixed-locator sets to avoid 1299 heterogeneous encapsulation. 1301 o Introduce concept of mPITRs to help reduce (S-EID,G) to the edges 1302 of LISP global multicast network. 1304 A.7. Changes to draft-ietf-lisp-multicast-02.txt 1306 o Posted September 2009. 1308 o Added Document Change Log appendix. 1310 o Specify that the LISP Encapsulated Control Message be used for 1311 unicasting PIM Join/Prune messages from ETRs to ITRs. 1313 A.8. Changes to draft-ietf-lisp-multicast-01.txt 1315 o Posted November 2008. 1317 o Specified that PIM Join/Prune unicast messages that get sent from 1318 ETRs to ITRs of a source multicast site get LISP encapsulated in 1319 destination UDP port 4342. 1321 o Add multiple RLOCs per ITR per Yiqun's comments. 1323 o Indicate how static RPs can be used when LISP is run using Bidir- 1324 PIM in the core. 1326 o Editorial changes per Liming comments. 1328 o Add Mttrace Considersations section. 1330 A.9. Changes to draft-ietf-lisp-multicast-00.txt 1332 o Posted April 2008. 1334 o Renamed from draft-farinacci-lisp-multicast-01.txt. 1336 Authors' Addresses 1338 Dino Farinacci 1339 cisco Systems 1340 Tasman Drive 1341 San Jose, CA 1342 USA 1344 Email: dino@cisco.com 1346 Dave Meyer 1347 cisco Systems 1348 Tasman Drive 1349 San Jose, CA 1350 USA 1352 Email: dmm@cisco.com 1354 John Zwiebel 1355 cisco Systems 1356 Tasman Drive 1357 San Jose, CA 1358 USA 1360 Email: jzwiebel@cisco.com 1362 Stig Venaas 1363 cisco Systems 1364 Tasman Drive 1365 San Jose, CA 1366 USA 1368 Email: stig@cisco.com