idnits 2.17.1 draft-farinacci-pim-port-01.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 19. -- Found old boilerplate from RFC 3978, Section 5.5, updated by RFC 4748 on line 884. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 895. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 902. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 908. 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 Copyright Line does not match the current year -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (May 20, 2008) is 5819 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 761 (Obsoleted by RFC 793, RFC 7805) ** Obsolete normative reference: RFC 4601 (Obsoleted by RFC 7761) ** Obsolete normative reference: RFC 4960 (Obsoleted by RFC 9260) -- Duplicate reference: RFC4601, mentioned in 'HELLO-OPT', was also mentioned in 'RFC4601'. -- Obsolete informational reference (is this intentional?): RFC 4601 (ref. 'HELLO-OPT') (Obsoleted by RFC 7761) == Outdated reference: A later version (-10) exists of draft-ietf-l3vpn-2547bis-mcast-05 Summary: 4 errors (**), 0 flaws (~~), 2 warnings (==), 9 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Dino Farinacci 3 Internet-Draft IJsbrand Wijnands 4 Intended status: Experimental Apoorva Karan 5 Expires: November 21, 2008 Arjen Boers 6 cisco Systems 7 Maria Napierala 8 AT&T Labs 9 May 20, 2008 11 A Reliable Transport Mechanism for PIM 12 draft-farinacci-pim-port-01.txt 14 Status of this Memo 16 By submitting this Internet-Draft, each author represents that any 17 applicable patent or other IPR claims of which he or she is aware 18 have been or will be disclosed, and any of which he or she becomes 19 aware will be disclosed, in accordance with Section 6 of BCP 79. 21 Internet-Drafts are working documents of the Internet Engineering 22 Task Force (IETF), its areas, and its working groups. Note that 23 other groups may also distribute working documents as Internet- 24 Drafts. 26 Internet-Drafts are draft documents valid for a maximum of six months 27 and may be updated, replaced, or obsoleted by other documents at any 28 time. It is inappropriate to use Internet-Drafts as reference 29 material or to cite them other than as "work in progress." 31 The list of current Internet-Drafts can be accessed at 32 http://www.ietf.org/ietf/1id-abstracts.txt. 34 The list of Internet-Draft Shadow Directories can be accessed at 35 http://www.ietf.org/shadow.html. 37 This Internet-Draft will expire on November 21, 2008. 39 Copyright Notice 41 Copyright (C) The IETF Trust (2008). 43 Abstract 45 This draft describes how a reliable transport mechanism can be used 46 by the PIM protocol to optimize CPU and bandwidth resource 47 utilization by eliminating periodic Join/Prune message transmission. 48 This draft proposes a modular extension to PIM to use either the TCP 49 or SCTP transport protocol. 51 Table of Contents 53 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 54 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 55 1.2. Definitions . . . . . . . . . . . . . . . . . . . . . . . 5 56 2. Protocol Overview . . . . . . . . . . . . . . . . . . . . . . 7 57 3. New PIM Hello Options . . . . . . . . . . . . . . . . . . . . 8 58 3.1. PIM over the TCP Transport Protocol . . . . . . . . . . . 8 59 3.2. PIM over the SCTP Transport Protocol . . . . . . . . . . . 9 60 4. Establishing Transport Connections . . . . . . . . . . . . . . 10 61 4.1. TCP Connection Maintenance . . . . . . . . . . . . . . . . 11 62 4.2. Transitional Periods . . . . . . . . . . . . . . . . . . . 11 63 4.3. On-demand versus Pre-configured Connections . . . . . . . 12 64 4.4. Possible Hello Suppression Considerations . . . . . . . . 12 65 4.5. Avoiding a Pair of Connections between Neighbors . . . . . 13 66 5. Common Header Definition . . . . . . . . . . . . . . . . . . . 14 67 6. Join/Prune Processing . . . . . . . . . . . . . . . . . . . . 18 68 7. Outgoing Interface List Explicit Tracking . . . . . . . . . . 19 69 8. Multiple Instances and Address-Family Support . . . . . . . . 20 70 9. Miscellany . . . . . . . . . . . . . . . . . . . . . . . . . . 21 71 10. Security Considerations . . . . . . . . . . . . . . . . . . . 22 72 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 23 73 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 24 74 13. References . . . . . . . . . . . . . . . . . . . . . . . . . . 25 75 13.1. Normative References . . . . . . . . . . . . . . . . . . . 25 76 13.2. Informative References . . . . . . . . . . . . . . . . . . 25 77 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 26 78 Intellectual Property and Copyright Statements . . . . . . . . . . 27 80 1. Introduction 82 The goals of this specification are: 84 o To create a simple incremental mechanism to provide reliable PIM 85 message delivery in PIM version 2. 87 o The reliable transport mechanism will be used for Join-Prune 88 message transmission only. 90 o Can be used for link-local transmission of Join-Prune messages or 91 multi-hop for use in a multicast VPN environments. 93 o When a router supports this specification, it need not use the 94 reliable transport mechanism on every interface. That is, 95 negotiation on per interface basis (or MDT basis) will occur. 97 The explicit non-goals of this specification are: 99 o Changes to the PIM protocol machinery as defined in [RFC4601]. 100 The reliable transport mechanism will be used as a plugin layer so 101 the PIM component does not know it is really there. 103 o Provide support for both Datagram mode and Transport mode (see 104 Section 1.2 for definitions) on the same physical interface or 105 MDT. 107 This document will specify how periodic JP message transmission can 108 be eliminated by using TCP [RFC0761] or SCTP [RFC4960] as the 109 reliable transport mechanism for JP messages. 111 This specification enables greater scalability in multicast 112 deployment since the processing required for protocol state 113 maintenance can be reduced. These enhancements to PIMv2 are 114 applicable to IP multicast over routed services and VPNs [MCAST-VPN]. 115 In addition to reduced processing on PIM enabled routers, another 116 important feature is the reduced join and leave latency provided 117 through a reliable transport. 119 In many existing and emerging networks, particularly wireless and 120 mobile satellite systems, link degradation due to weather, 121 interference, and other impairments can result in temporary spikes in 122 the packet loss. In these environments, periodic PIM joining can 123 cause join latency when messages are lost causing a retransmission 124 only 60 seconds later. By applying a reliable transport, a lost join 125 is retransmitted rapidly. Furthermore, when the last user leaves a 126 multicast group, any lost prune is similarly repaired and the 127 multicast stream is quickly removed from the wireless/satellite link. 129 Without a reliable transport, the multicast transmission could 130 otherwise continue until it timed out, roughly 3 minutes later. As 131 network resources are at a premium in many of these environments, 132 rapid termination of the multicast stream is critical to maintaining 133 efficient use of bandwidth. 135 1.1. Requirements Notation 137 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 138 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 139 document are to be interpreted as described in [RFC2119]. 141 1.2. Definitions 143 PORT: Stands for PIM Over Reliable Transport. Which is the short 144 form for describing the mechanism in this specification where PIM 145 can use the TCP or SCTP transport protocol. 147 JP Message: An abbreviation for a Join-Prune message. 149 Periodic JP: A JP message sent periodically to refresh state. 151 Incremental JP: A JP message sent as a result of state creation or 152 deletion events. Also known as a triggered message. 154 Native JP: A JP message which is carried with an IP protocol type 155 of PIM. 157 Reliable JP: A JP message using TCP or SCTP for transport. 159 Datagram Mode: The current procedures PIM uses by encapsulating JP 160 messages in IP packets sent either triggered or periodically. 162 Transport Mode: Procedures used by PIM defined in this 163 specification for sending JP messages over the TCP or SCTP 164 transport layer. 166 MDT/PMSI: Used interchangeably in this document. An MDT tunnel is 167 one used between PE router to provide support for a Multicast VPN. 168 The new standards term for an MDT tunnel is a Provider-Network 169 Multicast Service Interface or PMSI. 171 Segmented Multi-Access LAN: A segmented (or partitioned) LAN is 172 like a virtual overlay network using the physical LAN to realize 173 control and data packets. Multiple overlay networks may be 174 created using the physical LAN, much like how VLANs or PMSI 175 overlays are configured over a multi-access phsyical LAN. The 176 interface associated with the partitioned LAN is like an NBMA 177 interface type so explicit tracking can be accomplished. Each 178 partitioned or segmented LAN has it's own data-link encapsulation 179 and link-layer multicast is still used to avoid head-end 180 replication. This concept also applies to MDTs/PMSIs and is 181 called "Segmented MDTs/PMSIs". A Segmented MDT/PMSI is a MDT/PMSI 182 that has a single forwarder (i.e. a single ingress PE) for any 183 multicast stream. 185 2. Protocol Overview 187 PIM Over Reliable Transport (PORT) is a simple extension to PIMv2 for 188 refresh reduction of PIM JP messages. It involves sending 189 incremental rather than periodic JPs over a TCP/SCTP connection 190 between PIM neighbors. 192 PORT can be incrementally used on an interface between PORT capable 193 neighbors. Routers which are not PORT capable can continue to use 194 PIM in Datagram Mode. PORT capability is detected using a new PORT 195 Capable PIM Hello Option. 197 When PORT is used, only incremental JPs are sent from downstream 198 routers to upstream routers. As such, downstream routers do not 199 generate periodic JPs for routes which RPF to a PORT-capable 200 neighbor. 202 For Joins and Prunes, which are received over a TCP/SCTP connection, 203 the upstream router does not start or maintain timers on the outgoing 204 interface entry. Instead, it explicitly tracks downstream routers 205 which have expressed interest. An interface is deleted from the 206 outgoing interface list only when all downstream routers on the 207 interface, no longer wish to receive traffic. 209 Because incremental JPs are sent over a TCP/SCTP connection, no Join 210 suppression or Prune-Override of incremental JPs is possible on 211 multi-access LANs. As a result, upstream routers, which receive an 212 incremental Join or Prune that creates state, explicitly track all 213 downstream nodes. Note, for point-to-point links there is no need 214 for explicitly tracking downstream nodes. 216 There is no change proposed for the PIM JP packet format. However, 217 for JPs sent over TCP/SCTP connections, no IP Header is included. 218 The message begins with the PIM common header, followed by the JP 219 message. See section Section 5 for details on the common header. 221 3. New PIM Hello Options 223 3.1. PIM over the TCP Transport Protocol 225 Option Type: PIM-over-TCP Capable 227 0 1 2 3 228 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 1 229 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 230 | Type = 65006 | Length = X + 4 | 231 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 232 | TCP Connection ID AFI | Reserved | 233 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 234 | TCP Connection ID | 235 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 237 Allocated Hello Type values can be found in [HELLO-OPT]. 239 When a router is configured to use PIM over TCP on a given interface, 240 it MUST include the PORT Capable hello option in its Hello messages 241 for that interface. If a router is explicitly disabled from using JP 242 over TCP it MUST NOT include the PORT Capable hello option in its 243 Hello messages. When the router cannot setup a TCP connection, it 244 will refrain from including this option. 246 This option is only used when an interface is point-to-point, 247 segmented multi-access LAN or a PMSI [MCAST-VPN]. In all other 248 cases, such as multi-access LANs, Datagram Mode is used. 250 Implementation may provide a configuration option to enable or 251 disable PORT functionality. We recommend that this capability be 252 disabled by default. 254 Length: In bytes for the value part of the Type/Length/Value 255 encoding. Where X is 4 bytes if IP AFI of value 1 is used and 16 256 bytes when IPv6 AFI of 2 is used [AFI]. 258 TCP Connection ID AFI: The AFI value to describe the address-family 259 of the address of the TCP Connection ID field. 261 Reserved: Set to zero on transmission and ignored on receipt. 263 TCP Connection ID: An IP or IPv6 address used to establish the TCP 264 connection. When this field is 0, a mechanism outside the scope 265 of this spec is used to obtain the addresses used to establish the 266 TCP connection. 268 3.2. PIM over the SCTP Transport Protocol 270 Option Type: PIM-over-SCTP Capable 272 0 1 2 3 273 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 1 274 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 275 | Type = 65007 | Length = X + 4 | 276 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 277 | SCTP Connection ID AFI | Reserved | 278 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 279 | SCTP Connection ID | 280 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 282 Allocated Hello Type values can be found in [HELLO-OPT]. 284 When a router is configured to use PIM over SCTP on a given 285 interface, it MUST include the PORT Capable hello option in its Hello 286 messages for that interface. If a router is explicitly disabled from 287 using JP over SCTP it MUST NOT include the PORT Capable hello option 288 in its Hello messages. When the router cannot setup a SCTP 289 connection, it will refrain from including this option. 291 This option is only used when an interface is point-to-point or when 292 a multi-access LAN or MDT is segmented (also known as "Partitioned 293 MDTs" in a non-broadcast multi-access (NBMA) mode. In all other 294 cases, such as general purpose multi-access LANs, Datagram Mode is 295 used. 297 Implementation may provide a configuration option to enable or 298 disable PORT functionality. We recommend that this capability be 299 disabled by default. 301 Length: In bytes for the value part of the Type/Length/Value 302 encoding. Where X is 4 bytes if IP AFI of value 1 is used and 128 303 bytes when IPv6 AFI of 2 is used [AFI]. 305 SCTP Connection ID AFI: The AFI value to describe the address- 306 family of the address of the SCTP Connection ID field. 308 Reserved: Set to zero on transmission and ignored on receipt. 310 SCTP Connection ID: An IP or IPv6 address used to establish the 311 SCTP connection. When this field is 0, a mechanism outside the 312 scope of this spec is used to obtain the addresses used to 313 establish the SCTP connection. 315 4. Establishing Transport Connections 317 Since this specification describes using Transport on point-to- point 318 links or NBMA configured MDTs, a router knows when a Transport is 319 established with the neighbor. When the Transport connection is not 320 established, Datagram Mode is used. When the Transport connection 321 becomes established Transport Mode is in effect where the router can 322 suppress sending periodic JPs. 324 When a router receives a Hello from a neighbor it has not previously 325 heard from, or the PORT-Capable Option is included in a Hello that 326 was not previously included by an existing neighbor, the router will 327 attempt to establish a Transport connection with the neighbor. When 328 the router is using TCP it will compare the IP address it uses to 329 send Hellos on the interface with the IP address the neighbor is 330 using to send Hellos. The router with the lower IP address will do 331 an active Transport open to the neighbor address. The higher IP 332 addressed neighbor will do a passive Transport open. When the router 333 is using SCTP, the IP address comparison not be done since the SCTP 334 protocol can handle call collision. 336 When a Transport connection goes down, Join or Prune state that was 337 sent over the Transport connection is still retained. The neighbor 338 should not be considered down until the neighbor timer has expired. 339 This allows routers to do a control-plane switchover without 340 disrupting the network. If a Transport connection is reestablished 341 before the neighbor timer expires, the previous state is intact and 342 any new JP messages sent cause state to be created or removed 343 (depending on if it was a Join or Prune). If the neighbor timer does 344 expire, only the upstream router, that has oif-list state, to the 345 expired downstream neighbor will need to clear state. A downstream 346 router, when an upstream neighboring router has expired, will simply 347 RPF to a new neighbor where it would trigger JP messages like it 348 would in [RFC4601]. It is required of a PIM router to clear it's 349 neighbor table for a neighbor who has timed out due to neighbor 350 holdtime expiration. 352 When a router is in Datagram Mode with a neighbor and has been 353 sending periodic JP messages to it and then the Transport connection 354 has been established to the neighbor, there is no requirement for the 355 downstream router to send JP messages to the upstream neighbor. The 356 upstream router can keep the state maintained from the Datagram Mode 357 creation. However when a router is in Transport Mode with a neighbor 358 and moves to Datagram Mode because the transport connection went down 359 (and several attempts to reestablish the transport connection fail), 360 the router cannot be sure that all the JP data was received by the 361 neighbor. Therefore, it is required to send a full set of JP 362 messages to refresh or re-create state in the upstream neighbor. 364 An upstream neighbor does have the responsibility of removing the 365 timer-activated timeout of an oif-list entry. When a Transport 366 connection is established, the timer-activated timeout is disabled. 367 When a Transport connection goes down, the timer-activated timeout 368 for an oif-list is enabled. Both the upstream and downstream routers 369 stay in sync based on the state of the Transport connection. If the 370 upstream router has timer-activated timeout on oif-lists, the 371 downstream router will be sending periodic JPs. Otherwise, the 372 downstream router suppresses sending periodic JPs because it assumes 373 the upstream router has disabled the timer-activated timeout of oif- 374 list entries the downstream router has previously joined. 376 4.1. TCP Connection Maintenance 378 TCP is designed to keep connections up indefinitely during a period 379 of network disconnection. If a PIM-over-TCP router fails, the TCP 380 connection may stay up until the neighbor actually reboots, and even 381 then it may continue to stay up until you actually try to send the 382 neighbor some information. This is particularly relevant to PIM, 383 since the flow of JPs might be in only one direction, and the 384 downstream neighbor might never get any indication via TCP that the 385 other end of the connection isn't really there. 387 Most applications using TCP want to detect when a neighbor is no 388 longer there, so that the associated application state can be 389 released. Also, one wants to clean up the TCP state, and not keep 390 half-open connections around indefinitely. This is accomplished by 391 using PIM Hellos and by not introducing an application-specific or 392 new PIM keep-alive message. Therefore, when a GENID changes from a 393 received PIM Hello message, and a TCP connection is established or 394 attempting to be established, the local side will tear down the 395 connection and attempt to reopen a new one for the new instance of 396 the neighbor coming up. 398 When PORT capable routers come up and try to establish transport 399 connections with their neighbors, but cannot for some reason, after 3 400 attempts to do so, the router should go into datagram mode and not 401 advertise the PORT Hello option anymore. Operator intervention is 402 required to restart the process after the problem is found. 404 4.2. Transitional Periods 406 There may be transitional periods when a router receives, from a 407 given neighbor, both datagram JP messages and JP messages sent over a 408 transport connection. When this happens, a transport connection to a 409 particular neighbor is established, and as long as it remains 410 established, the router MUST ignore PIM messages sent in Datagram 411 Mode from that neighbor. Otherwise, the datagram messages could get 412 out of order with respect to the transport messages, and the router 413 could end up in an erroneous state of pruning joined state or joining 414 pruned state which it is unable to recover from as long as the 415 transport connection stays up. 417 4.3. On-demand versus Pre-configured Connections 419 Transport connections could be established when they are needed or 420 when a router interface to other PIM neighbors has come up. The 421 advantages of on-demand Transport connection establishment are the 422 reduction of router resources. Especially in the case where there is 423 no need for n^2 connections on a network interface or MDT tunnel. 424 The disadvantages are deciding what to do when a JP message needs to 425 be sent and a Transport connection is not established yet. An 426 implementation can either send a Datagram Mode JP or queue the JP to 427 be sent as a Transport Mode JP after the Transport connection is 428 established. 430 If a router interface has become operational and PIM neighbors are 431 learned from Hello messages, at that time, Transport connections may 432 be established. The advantage is that a connection is ready to 433 transport data by the time a JP messages needs to be sent. The 434 disadvantage is there can be more connections established than 435 needed. This can occur when there is a small set of RPF neighbors 436 for the active distribution trees compared to the total number of 437 neighbors. Even when Transport connections are pre-established 438 before they are needed, a connection can go down and an 439 implementation will have to deal with an on-demand situation. 441 Therefore, this specification recommends but does not mandate the use 442 of on-demand Transport connection establishment. 444 4.4. Possible Hello Suppression Considerations 446 This specification indicates that a Transport connection cannot be 447 established until a Hello message is received. One reason for this 448 is to determine if the PIM neighbor supports this specification and 449 the other is to determine the remote address to use to establish the 450 Transport connection. 452 There are cases where it is desirable to suppress entirely the 453 transmission of Hello messages. In this case, it is outside the 454 scope of this document on how to determine if the PIM neighbor 455 supports this specification as well as an out-of-band (outside of the 456 PIM protocol) method to determine the remote address to establish the 457 Transport connection. 459 4.5. Avoiding a Pair of Connections between Neighbors 461 To ensure there are not two connections between a pair of PIM 462 neighbors, the following set of rules must be followed. Let A and B 463 be two PIM neighbors where A's IP address is numerically smaller than 464 B's IP address, and each is known to the other as having a potential 465 PIM adjacency relationship. 467 At node A: 469 o If there is already an established TCP connection to B, on the 470 PIM-over-TCP port, then A MUST NOT attempt to establish a new 471 connection to B. Rather it uses the established connection to send 472 JPs to B. (This is independent of which node initiated the 473 connection.) 475 o If A has initiated a connection to B, but the connection is still 476 in the process of being established, then A MUST refuse any 477 connection on the PIM-over-TCP port from B. 479 o At any time when A does not have a connection to B which is either 480 established or in the process of being established, A MUST accept 481 connections from B. 483 At node B: 485 o If there is already an established TCP connection to A, on the 486 PIM-over-TCP port, then B MUST NOT attempt to establish a new 487 connection to A. Rather it uses the established connection to send 488 JPs to A. (This is independent of which node initiated the 489 connection.) 491 o If B has initiated a connection to A, but the connection is still 492 in the process of being established, then if A initiates a 493 connection to, B MUST accept the connection initiated by A and 494 must release the connection which it (B) initiated. 496 5. Common Header Definition 498 It may be desirable for scaling purposes to include JP messages from 499 different PIM protocol instances to be sent over the same Transport 500 connection. Also, it may be desirable to have a set of JP messages 501 for one address-family sent over a Transport connection that is 502 established over a different address-family network layer. 504 To be able to do this we need a common header that is inserted and 505 parsed for each PIM JP message that is sent on a Transport 506 connection. This common header will provide both record boundary and 507 demux points when sending over a stream protocol like Transport. 509 Each JP message will have in front of it the following common header 510 in Type/Length/Value format. And multiple different TLV types can be 511 sent over the same Transport connection. 513 To make sure PIM JP messages are delivered as soon as the TCP 514 transport layer receives the JP buffer, the TCP Push flag will be set 515 in all outgoing JP messages sent over a TCP transport connection. 517 PIM messages will be sent using TCP port number TBD. When using SCTP 518 as the reliable transport, port number TBD will be used. See 519 Section 11 for IANA considerations. 521 If the buffer length of the received TLV message is less than what is 522 encoded in the TLV Length field, the entire TLV encoded message is 523 ignored and a error message is logged. Likewise, if the received 524 buffer length left to process at each record parsing level, is less 525 than the JP Message Length, the rest of the message is malformed and 526 not processed. 528 Each JP message that has passed the length checks above, contained in 529 the TLV encoding, will be error checked individually. This includes 530 a bad PIM checksum, illegal type fields, or illegal addresses. If 531 any parsing errors occur in a single JP message, it is skipped over 532 and not processed but other JP message records in the TLV are still 533 parsed and processed. 535 The current list of defined TLVs are: 537 IPv4 JP Message 539 0 1 2 3 540 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 1 541 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 542 | Type = 1 | Length = (12 * X) + Y | 543 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 544 | JP Message Length | Reserved |I-Type| 545 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 546 | Instance ID . . . | 547 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 548 | . . . Instance ID | 549 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 550 | PIMv2 JP Message | 551 | . | 552 | . | 553 | . | 554 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 555 | JP Message Length | Reserved |I-Type| 556 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 557 | Instance ID . . . | 558 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 559 | . . . Instance ID | 560 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 561 | PIMv2 JP Message | 562 | . | 563 | . | 564 | . | 565 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 567 The IPv4 JP common header is used when a JP message is sent that has 568 all IPv4 encoded addresses in the PIM payload. 570 Length: In bytes for the value part of the Type/Length/Value 571 encoding. Where there are 12 bytes per JP message (where X above 572 is the number of JP messages contained) enclosed in one 573 transmission plus Y which is the sum of each "JP Message Length" 574 field that appears in the transmission. 576 I-Type: Defines the encoding and semantics of the Instance ID 577 field. This is not specified in this specification. 579 Instance ID: This can be a VPN-ID. This field could also be a BGP 580 Route Target (RT) or BGP Route Distinguisher (RD) as defined in 581 [RFC4364]. Not specified in this specification. 583 Reserved: Set to zero on transmission and ignored on receipt. 585 JP Message Length: The number of bytes that follow which make up 586 the PIMv2 JP message. 588 PIMv2 JP Message: PIMv2 Join/Prune message and payload with no IP 589 header in front of it. As you can see from the packet format 590 diagram, multiple JP messages can go into one TCP/SCTP stream from 591 the same or different Instance IDs. 593 IPv6 JP Message 595 0 1 2 3 596 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 1 597 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 598 | Type = 2 | Length = (12 * X) + Y | 599 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 600 | JP Message Length | Reserved |I-Type| 601 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 602 | Instance ID . . . | 603 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 604 | . . . Instance ID | 605 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 606 | PIMv2 JP Message | 607 | . | 608 | . | 609 | . | 610 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 611 | JP Message Length | Reserved |I-Type| 612 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 613 | Instance ID . . . | 614 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 615 | . . . Instance ID | 616 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 617 | PIMv2 JP Message | 618 | . | 619 | . | 620 | . | 621 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 623 The IPv6 JP common header is used when a JP message is sent that has 624 all IPv6 encoded addresses in the PIM payload. 626 Length: In bytes for the value part of the Type/Length/Value 627 encoding. Where there are 12 bytes per JP message (where X above 628 is the number of JP messages contained) enclosed in one 629 transmission plus Y which is the sum of each "JP Message Length" 630 field that appears in the transmission. 632 I-Type: Defines the encoding and semantics of the Instance ID 633 field. This is not specified in this specification. 635 Instance ID: This can be a VPN-ID, BGP Route Target (RT) or BGP 636 Route Distinguisher (RD). Not specified in this specification. 638 Reserved: Set to zero on transmission and ignored on receipt. 640 JP Message Length: The number of bytes that follow which make up 641 the PIMv2 JP message. 643 PIMv2 JP Message: PIMv2 Join/Prune message and payload with no IP 644 header in front of it. As you can see from the packet format 645 diagram, multiple JP messages can go into one TCP/SCTP stream from 646 the same or different Instance IDs. 648 6. Join/Prune Processing 650 When a PORT neighbor transitions to using Transport Mode, the 651 downstream router sends JP messages for existing routes that RPF to 652 the neighbor over the Transport connection. In addition, periodic JP 653 messages are stopped and only incremental JPs are sent thereafter. 655 A router which has a Transport connection established MUST send and 656 receive JP messages over the Transport session to that given peer as 657 well as accept and process native JP messages as described in 658 [RFC4601]. 660 When a Transport connection is established for a newly discovered 661 neighbor, the downstream router triggers JP messages for its existing 662 state. This is to allow the upstream router to build state it may 663 previously not had. If state had existed due to a Native JP, the 664 expiration timer would have been started. Now it can be stopped 665 because the state is being sent incrementally over the Transport 666 connection. 668 When a Transport connection goes down to a given neighbor, the 669 downstream router does not have to trigger native JP messages. It 670 can wait for its next periodic interval to send a native JP messages. 671 When the upstream router receives the native JP message, it will 672 start the expiration timer for the oif associated with the state from 673 the JP message. 675 Note, since JP messages are sent over a Transport connection, no 676 Prune Override or Join Suppression are possible for these messages. 678 7. Outgoing Interface List Explicit Tracking 680 Since this specification indicates the use of TCP/SCTP for PIM JP 681 messages over point-to-point or NBMA type links, explicit tracking 682 can be achieved by tracking only oif-list state and not per-neighbor 683 per oif-list state. This is true for segmented LANs and in segmented 684 MDT/PMSI environments. 686 By using explicit tracking of oifs, the router tracks all downstream 687 neighbors which have expressed interest in a route on a given 688 interface. The list of tracked routers is one of the checks used to 689 determine whether traffic needs to be forwarded on a given interface 690 or not. 692 For (*,G) and (S,G) routes, the router starts forwarding traffic on 693 an interface when a Join is received from a neighbor on such an 694 interface. This is tracking the oif to the neighbor. When the 695 neighbor sends a Prune, the interface is removed and forwarding of 696 traffic stops on the interface. 698 When all interfaces are removed from the oif-list, the route entry 699 can be removed. 701 For (S,G,R) routes, typically is tracking Prune state on the shared 702 tree. One at least one downstream neighbor sends a Prune over a 703 Transport connection, the (S,G,R) state is create with a empty 704 outgoing interface list. If a subsequent JP is received over a 705 Transport connection which has (*,G) in the join-list and does not 706 have (S,G,R) in the prune-list, the upstream router will add the 707 interface the JP message was received on to the oif-list. And oif- 708 list based explicit tracking will occur just like in the (*,G) and 709 (S,G) route case above. 711 The only difference in the (S,G,R) route case, is that when the 712 outgoing interface is pruned, the entry must stay in the route table 713 or else forwarding will occur on the interfaces for the (*,G) entry. 714 Therefore, explicit tracking for Prunes must be provided. Only when 715 the (S,G,R) oif-list interfaces match the interfaces in the (*,G) can 716 the (S,G,R) route be removed. 718 8. Multiple Instances and Address-Family Support 720 Multiple instances of the PIM protocol may be used to support 721 multiple VPNs or within a VPN to support multiple address families. 722 Multiple instances can cause a multiplier effect on the number of 723 router resources consumed. To be able to have an option to use 724 router resources more efficiently, muxing JP messages over fewer 725 Transport connections can be performed. 727 There are two ways this can be accomplished, one using a common 728 header format over a TCP connection and the other using multiple 729 streams over a single SCTP connection. 731 Using the Common Header format described previously in this 732 specification, using different TLVs, both IPv4 and IPv6 based JP 733 messages can be encoded within a Transport connection. Likewise, 734 within a TLV, multiple occurrences of JP messages can occur and are 735 tagged with an instance-ID so multiple JP messages for different VPNs 736 can use a single Transport connection. 738 When using SCTP multi-streaming, the common header is still used to 739 convey instance information but an SCTP association is used, on a 740 per-VPN basis, to send data concurrently for multiple instances. 741 When data is sent concurrently, head of line blocking, which can 742 occur when using TCP, is avoided. 744 9. Miscellany 746 No changes expected in processing of other PIM messages like PIM 747 Asserts, Grafts, Graft-Acks, Registers, and Register-Stops. This 748 goes for BSR and Auto-RP type messages as well. 750 This extension is applicable only to PIM-SM, PIM-SSM and Bidir-PIM. 751 It does not take requirements for PIM-DM into consideration. 753 10. Security Considerations 755 Transport connections can be authenticated using HMACs MD5 and SHA-1 756 similar to use in BGP [RFC4271] and MSDP [RFC3618]. 758 When using SCTP as the transport protocol, [RFC4895] can be used, on 759 a per SCTP association basis to authenticate PIM data. 761 11. IANA Considerations 763 This specification requests IANA to allocate a TCP port number and a 764 SCTP port number for the use of PIM-Over-Reliable-Transport. 766 12. Acknowledgments 768 The authors would like to give a special thank you and appreciation 769 to Nidhi Bhaskar for her initial design and early prototype of this 770 idea. 772 Appreciation goes to Randall Stewart for his authoritative review and 773 recommendation for using SCTP. 775 Thanks also goes to the following for their ideas and commentary 776 review of this specification, Mike McBride, Toerless Eckert, Yiqun 777 Cai, Albert Tian, Suresh Boddapati, Nataraj Batchu, Daniel Voce, John 778 Zwiebel, Yakov Rekhter, and Lenny Giuliano. 780 A special thank you goes to Eric Rosen for his very detailed review 781 and commentary. Many of his comments are reflected as text in this 782 specification. 784 13. References 786 13.1. Normative References 788 [RFC0761] Postel, J., "DoD standard Transmission Control Protocol", 789 RFC 761, January 1980. 791 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 792 Requirement Levels", BCP 14, RFC 2119, March 1997. 794 [RFC3618] Fenner, B. and D. Meyer, "Multicast Source Discovery 795 Protocol (MSDP)", RFC 3618, October 2003. 797 [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway 798 Protocol 4 (BGP-4)", RFC 4271, January 2006. 800 [RFC4364] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private 801 Networks (VPNs)", RFC 4364, February 2006. 803 [RFC4601] Fenner, B., Handley, M., Holbrook, H., and I. Kouvelas, 804 "Protocol Independent Multicast - Sparse Mode (PIM-SM): 805 Protocol Specification (Revised)", RFC 4601, August 2006. 807 [RFC4895] Tuexen, M., Stewart, R., Lei, P., and E. Rescorla, 808 "Authenticated Chunks for the Stream Control Transmission 809 Protocol (SCTP)", RFC 4895, August 2007. 811 [RFC4960] Stewart, R., "Stream Control Transmission Protocol", 812 RFC 4960, September 2007. 814 13.2. Informative References 816 [AFI] IANA, "Address Family Indicators (AFIs)", ADDRESS FAMILY 817 NUMBERS http://www.iana.org/numbers.html, February 2007. 819 [HELLO-OPT] 820 IANA, "PIM Hello Options", PIM-HELLO-OPTIONS per 821 RFC4601 http://www.iana.org/assignments/pim-hello-options, 822 March 2007. 824 [MCAST-VPN] 825 Rosen and Aggarwal, "Multicast in MPLS/BGP VPNs", Internet 826 Draft draft-ietf-l3vpn-2547bis-mcast-05.txt, July 2007. 828 Authors' Addresses 830 Dino Farinacci 831 cisco Systems 832 Tasman Drive 833 San Jose, CA 95134 834 USA 836 Email: dino@cisco.com 838 IJsbrand Wijnands 839 cisco Systems 840 Tasman Drive 841 San Jose, CA 95134 842 USA 844 Email: ice@cisco.com 846 Apoorva Karan 847 cisco Systems 848 170 Tasman Drive 849 San Jose, CA 850 USA 852 Email: apoorva@cisco.com 854 Arjen Boers 855 cisco Systems 856 Tasman Drive 857 San Jose, CA 95134 858 USA 860 Email: aboers@cisco.com 862 Maria Napierala 863 AT&T Labs 864 200 Laurel Drive 865 Middletown, New Jersey 07748> 866 USA 868 Email: mnapierala@att.com 870 Full Copyright Statement 872 Copyright (C) The IETF Trust (2008). 874 This document is subject to the rights, licenses and restrictions 875 contained in BCP 78, and except as set forth therein, the authors 876 retain all their rights. 878 This document and the information contained herein are provided on an 879 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 880 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 881 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 882 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 883 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 884 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 886 Intellectual Property 888 The IETF takes no position regarding the validity or scope of any 889 Intellectual Property Rights or other rights that might be claimed to 890 pertain to the implementation or use of the technology described in 891 this document or the extent to which any license under such rights 892 might or might not be available; nor does it represent that it has 893 made any independent effort to identify any such rights. Information 894 on the procedures with respect to rights in RFC documents can be 895 found in BCP 78 and BCP 79. 897 Copies of IPR disclosures made to the IETF Secretariat and any 898 assurances of licenses to be made available, or the result of an 899 attempt made to obtain a general license or permission for the use of 900 such proprietary rights by implementers or users of this 901 specification can be obtained from the IETF on-line IPR repository at 902 http://www.ietf.org/ipr. 904 The IETF invites any interested party to bring to its attention any 905 copyrights, patents or patent applications, or other proprietary 906 rights that may cover technology that may be required to implement 907 this standard. Please address the information to the IETF at 908 ietf-ipr@ietf.org. 910 Acknowledgment 912 Funding for the RFC Editor function is provided by the IETF 913 Administrative Support Activity (IASA).