idnits 2.17.1 draft-katz-ward-bfd-multipoint-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 16. -- Found old boilerplate from RFC 3978, Section 5.5, updated by RFC 4748 on line 1235. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1206. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1213. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1219. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** The document is more than 15 pages and seems to lack a Table of Contents. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There is 1 instance of too long lines in the document, the longest one being 22 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust Copyright Line does not match the current year == Line 226 has weird spacing: '... If a Multi...' -- 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.) -- Couldn't find a document date in the document -- date freshness check skipped. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'KEYWORDS' is mentioned on line 45, but not defined == Outdated reference: A later version (-11) exists of draft-ietf-bfd-base-07 Summary: 3 errors (**), 0 flaws (~~), 5 warnings (==), 7 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group D. Katz 3 Internet Draft Juniper Networks 4 D. Ward 5 Cisco Systems 6 Expires: July, 2008 January, 2008 8 BFD for Multipoint Networks 9 draft-katz-ward-bfd-multipoint-01.txt 11 Status of this Memo 13 By submitting this Internet-Draft, each author represents that any 14 applicable patent or other IPR claims of which he or she is aware 15 have been or will be disclosed, and any of which he or she becomes 16 aware will be disclosed, in accordance with Section 6 of BCP 79. 18 Internet-Drafts are working documents of the Internet Engineering 19 Task Force (IETF), its areas, and its working groups. Note that other 20 groups may also distribute working documents as Internet-Drafts. 22 Internet-Drafts are draft documents valid for a maximum of six months 23 and may be updated, replaced, or obsoleted by other documents at any 24 time. It is inappropriate to use Internet-Drafts as reference 25 material or to cite them other than as "work in progress." 27 The list of current Internet-Drafts can be accessed at 28 http://www.ietf.org/1id-abstracts.html 30 The list of Internet-Draft Shadow Directories can be accessed at 31 http://www.ietf.org/shadow.html 33 Abstract 35 This document describes extensions to the Bidirectional Forwarding 36 Detection (BFD) protocol for its use in multipoint and multicast 37 networks. Comments on this draft should be directed to rtg- 38 bfd@ietf.org. 40 Conventions used in this document 42 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 43 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 44 document are to be interpreted as described in RFC-2119 [KEYWORDS]. 46 1. Introduction 48 The Bidirectional Forwarding Detection protocol [BFD] specifies a 49 method for verifying unicast connectivity between a pair of systems. 50 This document defines a method for using BFD to provide verification 51 of multipoint or multicast connectivity between a multipoint sender 52 (the "head") and a set of multipoint receivers (the "tails"). 54 As multipoint transmissions are inherently unidirectional, this 55 mechanism purports only to verify this unidirectional connectivity. 56 Although this seems in conflict with the "Bidirectional" in BFD, it 57 is a natural fit for that protocol. 59 This application of BFD allows for the tails to detect a lack of 60 connectivity from the head. As an option, the tail may unreliably 61 notify the head of the lack of multipoint connectivity. As a further 62 option, this notification can be made reliable. Notification to the 63 head can be enabled for all tails, or for only a subset of the tails. 65 Multipoint BFD verifies only the head-to-tail connectivity over the 66 multipoint tree. Although it may use unicast paths in both 67 directions, Multipoint BFD does not verify those paths (and in fact 68 it is preferable if unicast paths share as little fate with the 69 multipoint tree as is feasible.) 71 Virtually all options and timing parameters are controlled by the 72 head. This is particularly important if head notifications are 73 enabled, since there are obvious scaling concerns in that case. 75 Throughout this document, the term "multipoint" is defined as a 76 mechanism by which more than one system receives packets sent by a 77 single sender. This specifically includes such things as IP 78 multicast and point-to-multipoint MPLS. The term "multipoint tree" 79 includes degenerate cases such as LAN multicast. 81 This document effectively modifies and adds to the base BFD 82 specification. It is the intention of the authors to fold these 83 extensions into the base specification at the appropriate time. 85 2. Goals 87 The primary goal of this mechanism is to allow tails to rapidly 88 detect the fact that multipoint connectivity from the head has 89 failed. An optional goal is for the head to reasonably rapidly have 90 knowledge of tails that have lost connectivity from the head. 92 Since scaling is a primary concern (particularly state implosion 93 toward the head), it is a goal that the head be in control of all 94 timing aspects of the mechanism, and that BFD packets from the tails 95 to the head not be synchronized. 97 Another goal is for the mechanism to work on any multicast or 98 multipoint medium. 100 A further goal is to support multiple, overlapping multipoint trees, 101 as well as multipoint trees with multiple heads, and to allow point- 102 to-point BFD sessions to operate simultaneously among the systems 103 participating in Multipoint BFD. 105 A final goal is to integrate multipoint operation into the base 106 specification in such a way as to make it relatively easy to support 107 both multipoint and point-to-point operation in a single 108 implementation. 110 It is a non-goal for this protocol to verify point-to-point 111 connectivity between the head and any tails. This can be done 112 independently (and with no penalty in protocol overhead) by using 113 point-to-point BFD. 115 3. Overview 117 The heart of this protocol is the periodic transmission of BFD 118 Control packets along a multipoint tree, from the head to all tails 119 on the tree. The contents of the BFD packets provide the means for 120 the tails to calculate the detection time for path failure. If no 121 BFD Control packets are received by a tail for a detection time, the 122 tail declares the path to have failed. For some applications this is 123 the only mechanism necessary; the head can remain ignorant of the 124 tails. In this mode, the tails never send any BFD traffic to the 125 head. 127 If the head wishes to be alerted to the tails' connectivity (or lack 128 thereof), there are a number of options. First, if all that is 129 needed is an unreliable failure notification, the head can direct the 130 tails to transmit unicast BFD Control packets back to the head when 131 the path fails. 133 If the head wishes to know the identity of the tails on the 134 multipoint tree, it may solicit membership by sending a multipoint 135 BFD Control packet with the Poll (P) bit set, which will induce the 136 tails to return a unicast BFD Control packet with the Final (F) bit 137 set. The head can then create BFD session state for each of the 138 tails that have multipoint connectivity. If the head sends such a 139 packet on occasion, it can keep track of which tails answer, thus 140 providing a somewhat reliable mechanism for detecting which tails 141 fail to respond (implying a loss of multipoint connectivity.) 143 If the head wishes a reliable indication of the tails' connectivity, 144 it may do all of the above, but if it detects that a tail did not 145 answer the previous multipoint poll, it may initiate a Demand mode 146 Poll Sequence as a unicast to the tail. This covers the case where 147 either the multipoint poll or the single reply thereto is lost in 148 transit. If desired, the head may Poll one or more tails proactively 149 to track the tails' connectivity. 151 If some tails are more equal than others, in the sense that the head 152 needs to detect the lack of multipoint connectivity to a subset of 153 tails at a different rate, the head may transmit unicast BFD Polls to 154 that subset of tails. In this case, the timing may be independent on 155 a tail-by-tail basis. 157 Individual tails may be configured so that they never send BFD 158 control packets to the head, even when the head wishes notification 159 of path failure from the tail. Such tails will never be known to the 160 head, but will still be able to detect multipoint path failures from 161 the head. 163 Although this document describes a single head and a set of tails 164 spanned by a single multipoint tree, the protocol is capable of 165 supporting (and discriminating between) more than one multipoint tree 166 at both heads and tails. Furthermore, the same head and tail may 167 share multiple multipoint trees, and a multipoint tree may have 168 multiple heads. 170 4. Protocol Details 172 This section describes the operation of Multipoint BFD in detail. 174 4.1. Multipoint BFD Control Packets 176 Multipoint BFD Control packets (packets sent by the head over a 177 multipoint tree) are explicitly marked as such, via the setting of 178 the M bit (added to the latest revision of the BFD base 179 specification. This means that Multipoint BFD does not depend on the 180 recipient of a packet to know whether the packet was received over a 181 multipoint path. This can be useful in scenarios where this 182 information may not be available to the recipient. 184 4.2. Session Model 186 Multipoint BFD is modeled as a set of sessions of different types. 187 The elements of procedure differ slightly for each type. 189 Point-to-point sessions, as described in [BFD], are of type 190 PointToPoint. 192 The head has a session of type MultipointHead that is bound to a 193 multipoint tree. Multipoint BFD Control packets are sent by this 194 session over the multipoint tree, and no BFD Control packets are 195 received by it. 197 If the head is keeping track of some or all of the tails, it has a 198 session of type MultipointClient per tail that it cares about. All 199 of the MultipointClient sessions for tails on a particular particular 200 multipoint tree are grouped with the MultipointHead session to which 201 the clients are listening. A BFD Poll Sequence may be sent over such 202 a session to a tail if the head wishes to verify connectivity. These 203 sessions receive any BFD Control packets sent by the tails, and never 204 transmit periodic BFD Control packets other than Poll Sequences 205 (since periodic transmission is always done by the MultipointHead 206 session.) 208 Each tail has a session of type MultipointTail associated with a 209 multipoint tree. These sessions receive BFD Control packets from the 210 head, both as multipoint packets (the MultipointHead session) and as 211 unicast packets (the MultipointClient session, if it exists.) Any 212 BFD Control packets sent to the head are sent over this session. 214 4.3. Session Failure Semantics 216 The semantics of session failure are subtle enough to warrant further 217 explanation. 219 MultipointHead sessions cannot fail (since they are controlled 220 administratively.) 222 If a MultipointTail session fails, it means that the tail definitely 223 has lost contact with the head (or the head has been administratively 224 disabled) and the tail should take appropriate action. 226 If a MultipointClient session receives a BFD Control packet from the 227 tail with state Down or AdminDown, the head reliably knows that the 228 tail has lost multipoint connectivity. If the Detection Time expires 229 on a MultipointClient session, it is ambiguous as to whether the 230 multipoint connectivity failed or whether there was a unicast path 231 problem in one direction or the other, so the head does not reliably 232 know the tail state. 234 4.4. State Variables 236 Multipoint BFD introduces some new state variables, and modifies the 237 usage of a few existing ones. 239 4.4.1. New State Variables 241 A number of state variables are added to the base spec in support of 242 Multipoint BFD. 244 bfd.SessionType 246 The type of this session. Allowable values are: 248 PointToPoint: Classic point-to-point BFD. 250 MultipointHead: A session on the head responsible for the 251 periodic transmission of multipoint BFD Control packets 252 along the multipoint tree. 254 MultipointClient: A session on the head that tracks the 255 state of an individual tail, when desirable. 257 MultipointTail: A multipoint session on a tail. 259 This variable MUST be initialized to the appropriate type when 260 the session is created, according to the rules in section 4.16. 262 bfd.SilentTail 264 If 1, a tail will never transmit any BFD Control packets to the 265 head under any circumstances. If 0, a tail may send packets to 266 the head according to other parts of this specification. This 267 allows tails to be provisioned to always be silent, even when 268 the head is soliciting traffic from the tails. This can be 269 useful, for example, in deployments of a large number of tails 270 when the head wishes to track the state of a subset of them. 271 This variable MUST be initialized based on configuration. 273 This variable is only pertinent when bfd.SessionType is 274 MultipointTail. 276 bfd.ReportTailDown 278 Set to 1 if the head wishes tails to notify the head, via 279 periodic BFD Control packets, when they see the BFD session 280 fail. If 0, the tail will never send periodic BFD Control 281 packets, and the head will not be notified of session failures 282 by the tails. This variable MUST be initialized based on 283 configuration. 285 This variable is only pertinent when bfd.SessionType is 286 MultipointHead or MultipointClient. 288 bfd.UnicastRcvd 290 Set to 1 if a tail receives a unicast BFD Control packet from 291 the head. This variable MUST be set to zero if the session 292 transitions from Up state to some other state. 294 This variable MUST be initialized to zero. 296 This variable is only pertinent when Bfd.SessionType is 297 MultipointTail. 299 4.4.2. State Variable Initialization and Maintenance 301 Some state variables defined in section 6.8.1 of the BFD base spec 302 need to be initialized or manipulated differently depending on the 303 session type. 305 bfd.LocalDiscr 307 For session type MultipointClient, this variable MUST always 308 match the value of bfd.LocalDiscr in the associated 309 MultipointHead session. 311 bfd.DesiredMinTxInterval 313 For session type MultipointClient, this variable MUST always 314 match the value of bfd.DesiredMinTxInterval in the associated 315 MultipointHead session. 317 bfd.RequiredMinRxInterval 319 This variable MUST always be 0 for session type MultipointHead 320 if bfd.ReportTailDown is 0. 322 It should be noted that for sessions of type MultipointTail, 323 this variable only affects the rate of unicast Polls sent by 324 the head; the rate of multipoint packets is necessarily 325 unaffected by it. 327 bfd.DemandMode 329 This variable MUST be initialized to 1 for session types 330 MultipointHead and MultipointClient, and MUST be initialized to 331 0 for session type MultipointTail. 333 bfd.DetectMult 335 For session type MultipointClient, this variable MUST always 336 match the value of bfd.DetectMult in the associated 337 MultipointHead session. 339 4.5. Controlling Multipoint BFD Options 341 The state variables defined above are used to choose which 342 operational options are active. 344 The most basic form of operation, in which BFD Control packets flow 345 only from the head and no tracking is desired of tail state at the 346 head, is accomplished by setting bfd.ReportTailDown to 0 in the 347 MultipointHead session. 349 If the head wishes to know the identity of the tails, it sends 350 multipoint Polls as needed. Previously known tails that don't 351 respond to the Polls will be detected. 353 If the head wishes to be notified by the tails when they lose 354 connectivity, it sets bfd.ReportTailDown to 1 in either the 355 MultipointHead session (if such notification is desired from all 356 tails) or in the MultipointClient session (if notification is desired 357 from a particular tail.) Note that the setting of this variable in a 358 MultipointClient session for a particular tail overrides the setting 359 in the MultipointHead session. 361 If the head wishes to verify the state of a tail on an ongoing basis, 362 it sends a Poll Sequence from the MultipointClient session associated 363 with that tail as needed. 365 If the head wants to more quickly be alerted to a session failure 366 from a particular tail, it sends a BFD Control packet from the 367 MultipointClient session associated with that tail. This has the 368 effect of eliminating the initial delay that the tail would otherwise 369 insert prior to transmission of the packet. 371 If a tail wishes to operate silently (sending no BFD Control packets 372 to the head) it sets bfd.SilentTail to 1 in the MultipointTail 373 session. This allows a tail to be silent independent of the settings 374 on the head. 376 4.6. State Machine 378 The BFD state machine works slightly differently in the multipoint 379 application. In particular, since there is a many-to-one mapping, 380 three-way handshakes for session establishment and teardown are 381 neither possible nor appropriate. As such there is no Init state. 383 The following diagram provides an overview of the state machine for 384 session types MultipointClient and MultipointTail. The notation on 385 each arc represents the state of the remote system (as received in 386 the State field in the BFD Control packet) or indicates the 387 expiration of the Detection Timer. 389 DOWN, ADMIN DOWN, 390 +------+ TIMER +------+ 391 +----| |<---------------------| |----+ 392 DOWN,| | DOWN | | UP | |UP 393 ADMIN DOWN,+--->| |--------------------->| |<---+ 394 TIMER +------+ UP +------+ 396 Sessions of type MultipointHead never receive packets and have no 397 Detection Timer, and as such all state transitions are 398 administratively driven. 400 4.7. Session Establishment 402 Unlike Point-to-point BFD, Multipoint BFD provides a form of 403 discovery mechanism for tails to discover the head, and vice versa. 404 The minimum amount of a priori information required both on the head 405 and tails is the binding to the multipoint tree over which BFD is 406 running. The head transmits Multipoint BFD packets on that tree, and 407 the tails listen for BFD packets on that tree. All other information 408 MAY be determined dynamically. 410 A session of type MultipointHead is created for each multipoint tree 411 over which the head wishes to run BFD. This session runs in the 412 Active role. Except when terminating BFD service, this session is 413 always in state Up and always operates in Demand mode. No received 414 packets are ever demultiplexed to the MultipointHead session. In 415 this sense it is a degenerate form of a session. 417 Sessions on the tail MAY be established dynamically, based on the 418 receipt of a Multipoint BFD Control packet from the head, and are of 419 type MultipointTail. Tail sessions always take the Passive role. 421 If BFD Control packets are received at the head, they are 422 demultiplexed to sessions of type MultipointClient, which represent 423 the set of tails that the head is interested in tracking. These 424 sessions will typically also be established dynamically based on the 425 receipt of BFD Control packets. The head has broad latitude in 426 choosing which tails to track, if any, without affecting the basic 427 operation of the protocol. The head directly controls whether or not 428 tails are allowed to send BFD Control packets back to the head. 430 4.8. Discriminators and Packet Demultiplexing 432 The use of Discriminators is somewhat different in Multipoint BFD 433 than in Point-to-point BFD. 435 The head sends Multipoint BFD Control packets over the MultipointHead 436 session with My Discr set to a value bound to the multipoint tree, 437 and with Your Discr set to zero. The tails MUST demultiplex these 438 packets based on a combination of the source address and My Discr, 439 which together uniquely identify the head and the multipoint tree. 441 When the tails send BFD Control packets to the head from the 442 MultipointTail session, the contents of Your Discr (the discriminator 443 received from the head) will not be sufficient for the head to 444 demultiplex the packet, since the same value will be received from 445 all tails on the multicast tree. In this case, the head MUST 446 demultiplex packets based on the source address and the value of Your 447 Discr, which together uniquely identify the tail and the multipoint 448 tree. 450 When the head sends unicast BFD Control packets to a tail from a 451 MultipointClient session, the value of Your Discr will be valid, and 452 the tail MUST demultiplex the packet based solely on Your Discr. 454 Note that, unlike PointToPoint sessions, the discriminator values on 455 all multipoint session types MUST NOT be changed during the life of a 456 session. This is a side effect of the more complex demultiplexing 457 scheme. 459 4.9. Controlling Tail Packet Transmission 461 As the fan-in from the tails to the head may be very large, it is 462 critical that the flow of BFD Control packets from the tails is 463 controlled. 465 The head always operates in Demand mode. This means that no tail 466 will send an asynchronous BFD Control packet as long as the session 467 is Up. 469 The value of Required Min Rx Interval received by a tail in a unicast 470 BFD Control packet, if any, always takes precedence over the value 471 received in Multipoint BFD Control packets. This allows the packet 472 rate from individual tails to be controlled separately as desired by 473 sending a BFD Control packet from the corresponding MultipointClient 474 session. This also eliminates the random delay prior to transmission 475 from the tail that would otherwise be inserted, reducing the latency 476 of reporting a failure to the head. 478 If the head wishes to suppress traffic from the tails when they 479 detect a session failure, it MAY set bfd.RequiredMinRxInterval to 480 zero, which is a reserved value that indicates that the sender wishes 481 to receive no periodic traffic. This can be set in the 482 MultipointHead session (suppressing traffic from all tails) or it can 483 be set in a MultipointClient session (suppressing traffic from only a 484 single tail.) 486 Any tail may be provisioned to never send *any* BFD Control packets 487 to the head by setting bfd.SilentTail to 1. This provides a 488 mechanism by which only a subset of tails report their session status 489 to the head. 491 4.10. Bringing Up and Shutting Down Multipoint BFD Service 493 Because there is no three-way handshake in Multipoint BFD, a newly 494 started head (that does not have any previous state information 495 available) SHOULD start with bfd.SessionState set to Down and with 496 bfd.RequiredMinRxInterval set to zero in the MultipointHead session. 497 The session SHOULD remain in this state for a time equal to 498 (bfd.DesiredMinTxInterval * bfd.DetectMult). This will ensure that 499 all MultipointTail sessions are reset (so long as the restarted head 500 is using the same or larger value of bfd.DesiredMinTxInterval than it 501 did previously.) 503 Multipoint BFD service is brought up by administratively setting 504 bfd.SessionState to Up in the MultipointHead session. 506 A head may wish to shut down its BFD service in a controlled fashion. 507 This is desirable because the tails need not wait a detection time 508 prior to declaring the multipoint session to be down (and taking 509 whatever action is necessary in that case.) 511 To shut down a multipoint session a head MUST administratively set 512 bfd.SessionState in the MultipointHead session to either Down or 513 AdminDown and SHOULD set bfd.RequiredMinRxInterval to zero (to keep 514 the tails from sending any BFD Control packets back.) The session 515 SHOULD send BFD Control packets in this state for a period equal to 516 (bfd.DesiredMinTxInterval * bfd.DetectMult). The tail SHOULD destroy 517 all MultipointClient sessions associated with the MultipointHead 518 session. 520 The semantic difference between Down and AdminDown state is for 521 further discussion. 523 4.11. Soliciting the Tails 525 If the head wishes to know the identities of the tails, the 526 MultipointHead session MAY send a BFD Control packet as specified in 527 section 4.16.3, with the Poll (P) bit set to 1. This will cause all 528 of the tails to reply with a unicast BFD Control Packet, randomized 529 across one packet interval. 531 The decision as to when to send a multipoint Poll is outside the 532 scope of this specification. However, it must never be sent more 533 often than the regular multipoint BFD Control packet. Since the tail 534 will treat a multipoint Poll like any other multipoint BFD Control 535 packet, Polls may be sent in lieu of non-Poll packets. 537 Soliciting the tails also starts the Detection Timer for each 538 associated MultipointClient session, which will cause those sessions 539 to time out if the associated tails do not respond. 541 Note that for this mechanism to work properly, the Detection Time 542 (which is equal to bfd.DesiredMinTxInterval) MUST be greater than the 543 round trip time of BFD Control packets from the head to the tail (via 544 the multipoint tree) and back (via a unicast path.) See section 4.14 545 for more details. 547 4.12. Verifying Connectivity to Specific Tails 549 If the head wishes to verify connectivity to a specific tail, the 550 corresponding MultipointClient session MAY send a BFD Poll Sequence 551 to said tail. This might be done in reaction to the expiration of 552 the Detection Timer (the tail didn't respond to a multipoint Poll), 553 or it might be done on a proactive basis. 555 The interval between transmitted packets in the Poll Sequence MUST be 556 calculated as specified in the base specification (the greater of 557 bfd.DesiredMinTxInterval and bfd.RemoteMinRxInterval.) 559 The poll sequence is terminated the same way as for unicast BFD; if 560 a reply from the tail with Final (F) set is received, the poll 561 sequence is terminated, and if the poll sequence times out, the 562 session with the tail is declared to be Down. 564 The value transmitted in Required Min RX Interval will be used by the 565 tail (rather than the value received in any multipoint packet) when 566 it transmits BFD Control packets to the head notifying it of a 567 session failure, and the transmitted packets will not be delayed. 568 This value can potentially be set much lower than in the multipoint 569 case, in order to speed up notification to the head, since the value 570 will be used only by the single tail. This value (and the lack of 571 delay) are "sticky", in that once the tail receives it, it will 572 continue to use it indefinitely. Therefore, if the head no longer 573 wishes to single out the tail, it SHOULD reset the timer to the 574 default by sending a Poll Sequence with the same value of Required 575 Min Rx Interval as is carried in the multipoint packets, or it MAY 576 reset the tail session by sending a Poll Sequence with state 577 AdminDown (after the completion of which the session will come back 578 up.) 580 Note that a failure of the head to receive a response to a Poll 581 Sequence does not necessarily mean that the tail has lost multipoint 582 connectivity, though a reply to a Poll Sequence does reliably 583 indicate connectivity or lack thereof (by virtue of the tail's state 584 not being Up in the BFD Control packet.) 586 4.13. Timer Manipulation 588 Because of the one-to-many mapping, a session of type MultipointHead 589 SHOULD NOT initiate a Poll Sequence in conjunction with timer value 590 changes. As such, such a session cannot wait for a Final before 591 increasing the transmit interval; such a session SHOULD send 592 bfd.DetectMult packets at the old transmit interval before using the 593 higher value in order to avoid false detection timeouts at the tails. 595 Since MultipointHead sessions do not calculate detection times, the 596 value of bfd.RequiredMinRxInterval may be changed at any time. 598 4.14. Detection Times 600 Multipoint BFD is inherently asymmetric. As such, each session type 601 has a different approach to detection times. 603 Since the MultipointHead session never receives packets, it does not 604 calculate a detection time. 606 MultipointClient sessions at the head are always in Demand mode, and 607 as such only care about detection time in two cases. First, if a 608 Poll Sequence is being sent on a MultipointClient session, the 609 detection time on this session is calculated according to the base 610 spec, that is, the transmission interval multiplied by 611 bfd.DetectMult. Second, when a multipoint Poll is sent to solicit 612 tail replies, the detection time on all associated MultipointClient 613 sessions that aren't currently sending Poll Sequences is set to a 614 value greater than or equal to bfd.RequiredMinRxInterval (one packet 615 time.) This value can be made arbitrarily large in order to ensure 616 that the detection time is greater than the BFD round trip time 617 between the head and the tail with no ill effects, other than 618 delaying the detection of unresponsive tails. Note that a detection 619 time expiration on a MultipointClient session at the head, while 620 indicating a BFD session failure, cannot be construed to mean that 621 the tail is not hearing multipoint packets from the head. 623 MultipointTail sessions cannot influence the transmission rate of the 624 MultipointHead session using the Required Min Rx Interval field 625 because of its one-to-many nature. As such, the Detection Time 626 calculation for a MultipointTail session does not use 627 bfd.RequiredMinRxInterval in the calculation. The detection time is 628 calculated as the product of the last received values of Desired Min 629 TX Interval and Detect Mult. 631 The value of bfd.DetectMult may be changed at any time on any session 632 type. 634 4.15. State Maintenance for Down/AdminDown Sessions 636 The length of time session state is kept after the session goes down 637 determines how long the session will continue to send BFD Control 638 packets (since no packets can be sent after the session is 639 destroyed.) 641 4.15.1. MultipointHead Sessions 643 When a MultipointHead session transitions to states Down or 644 AdminDown, the state SHOULD be maintained for a period equal to 645 (bfd.DesiredMinTxInterval * bfd.DetectMult) to ensure that the tails 646 more quickly detect the session going down (by continuing to transmit 647 BFD Control packets with the new state.) 649 4.15.2. MultipointTail Sessions 651 If bfd.SilentTail is 1, or bfd.RemoteMinRxInterval is zero, 652 MultipointTail sessions MAY be destroyed immediately upon leaving Up 653 state, since they will transmit no further packets. 655 Otherwise, MultipointTail sessions MUST be maintained as long as BFD 656 Control packets are being received by it (which by definition will 657 indicate that the head is not Up.) 659 MultipointTail sessions MUST be maintained after a Detection Time 660 expiration for at least the longer of an additional Detection Time 661 and the transmission of the first (delayed) BFD Control packet to the 662 head. The state MAY be maintained longer than this, but the session 663 MUST NOT transmit periodic BFD Control packets for a period longer 664 than the negotiated transmit interval multiplied by bfd.DetectMult; 665 after this time either the session MUST be destroyed or 666 bfd.RemoteMinRxInterval MUST be set to zero to suppress packet 667 transmission. 669 4.15.3. MultipointClient Sessions 671 If the MultipointHead session is going down (which only happens 672 administratively), all associated MultipointClient sessions SHOULD be 673 destroyed as they are superfluous. 675 If a MultipointClient session goes down due to the receipt of an 676 unsolicited BFD Control packet from the tail with state Down or 677 AdminDown (not in response to a Poll), and tail connectivity 678 verification is not being done, the session MAY be destroyed. If 679 verification is desired, the session SHOULD send a Poll Sequence and 680 the session SHOULD be maintained. 682 If the tail replies to a Poll Sequence with state Down or AdminDown, 683 it means that the tail session is definitely down. In this case, the 684 session MAY be destroyed. 686 If the Detection Time expires on a MultipointClient session (meaning 687 that the tail did not reply to a Poll Sequence) the session MAY be 688 destroyed. 690 4.16. Base Specification Text Replacement 692 The following sections are meant to replace the corresponding 693 sections in the base specification. 695 4.16.1. Reception of BFD Control Packets 697 The following procedure replaces section 6.8.6 of [BFD]. 699 When a BFD Control packet is received, the following procedure MUST 700 be followed, in the order specified. If the packet is discarded 701 according to these rules, processing of the packet MUST cease at that 702 point. 704 If the version number is not correct (1), the packet MUST be 705 discarded. 707 If the Length field is less than the minimum correct value (24 if 708 the A bit is clear, or 26 if the A bit is set), the packet MUST be 709 discarded. 711 If the Length field is greater than the payload of the 712 encapsulating protocol, the packet MUST be discarded. 714 If the Detect Mult field is zero, the packet MUST be discarded. 716 If the My Discriminator field is zero, the packet MUST be 717 discarded. 719 Demultiplex the packet to a session according to section 4.16.2 720 below. The result is either a session of the proper type, or the 721 packet is discarded (and packet processing MUST cease.) 723 If the A bit is set and no authentication is in use (bfd.AuthType 724 is zero), the packet MUST be discarded. 726 If the A bit is clear and authentication is in use (bfd.AuthType 727 is nonzero), the packet MUST be discarded. 729 If the A bit is set, the packet MUST be authenticated under the 730 rules of section 6.7, based on the authentication type in use 731 (bfd.AuthType.) This may cause the packet to be discarded. 733 Set bfd.RemoteDiscr to the value of My Discriminator. 735 Set bfd.RemoteState to the value of the State (Sta) field. 737 Set bfd.RemoteDemandMode to the value of the Demand (D) bit. 739 If bfd.SessionType is MultipointTail 741 If bfd.UnicastRcvd is 0 or the M bit is clear, set 742 bfd.RemoteMinRxInterval to the value of Required Min RX 743 Interval. 745 If the M bit is clear, set bfd.UnicastRcvd to 1. 747 Else (not MultipointTail) 749 Set bfd.RemoteMinRxInterval to the value of Required Min RX 750 Interval. 752 If the Required Min Echo RX Interval field is zero, the 753 transmission of Echo packets, if any, MUST cease. 755 If a Poll Sequence is being transmitted by the local system and 756 the Final (F) bit in the received packet is set, the Poll Sequence 757 MUST be terminated. 759 If bfd.SessionType is PointToPoint or MultipointClient, update the 760 transmit interval as described in [BFD] section 6.8.2. 762 If bfd.SessionType is PointToPoint, update the Detection Time as 763 described in [BFD] section 6.8.4. Otherwise, update the Detection 764 Time as described in section 4.14 above. 766 If bfd.SessionState is AdminDown 767 Discard the packet 769 If received state is AdminDown 770 If bfd.SessionState is not Down 771 Set bfd.LocalDiag to 3 (Neighbor signaled session down) 772 Set bfd.SessionState to Down 774 Else 776 If bfd.SessionState is Down 777 If bfd.SessionType is PointToPoint 778 If received State is Down 779 Set bfd.SessionState to Init 780 Else if received State is Init 781 Set bfd.SessionState to Up 783 Else (bfd.SessionType is not PointToPoint) 784 If received State is Up 785 Set bfd.SessionState to Up 787 Else if bfd.SessionState is Init 788 If received State is Init or Up 789 Set bfd.SessionState to Up 791 Else (bfd.SessionState is Up) 792 If received State is Down 793 Set bfd.LocalDiag to 3 (Neighbor signaled session down) 794 Set bfd.SessionState to Down 796 Check to see if Demand mode should become active or not 797 (see [BFD] section 6.6). 799 If bfd.RemoteDemandMode is 1, bfd.SessionState is Up, and 800 bfd.RemoteSessionState is Up, Demand mode is active on the 801 remote system and the local system MUST cease the periodic 802 transmission of BFD Control packets (see section 4.16.3.) 803 If bfd.RemoteDemandMode is 0, or bfd.SessionState is not Up, or 804 bfd.RemoteSessionState is not Up, Demand mode is not active on the 805 remote system and the local system MUST send periodic BFD Control 806 packets (see section 4.16.3.) 808 If the Poll (P) bit is set, and bfd.SilentTail is zero, send a BFD Control packet to the 809 remote system with the Poll (P) bit clear, and the Final (F) bit 810 set (see section 4.16.3.) 812 If the packet was not discarded, it has been received for purposes 813 of the Detection Time expiration rules in [BFD] section 6.8.4. 815 4.16.2. Demultiplexing BFD Control Packets 817 This section is part of the replacement for [BFD] section 6.8.6, 818 separated for clarity. 820 If the Multipoint (M) bit is set 822 If the Your Discriminator field is nonzero, the packet MUST be 823 discarded. 825 Select a session based on the source address and the My 826 Discriminator field. If a session is found, and 827 bfd.SessionType is not MultipointTail, the packet MUST be 828 discarded. If a session is not found, a new session of type 829 MultipointTail MAY be created, or the packet MAY be discarded. 830 This choice is outside the scope of this specification. 832 Else (Multipoint bit is clear) 834 If the Your Discriminator field is nonzero 836 Select a session based on the value of Your Discriminator. 837 If no session is found, the packet MUST be discarded. 839 If bfd.SessionType is MulticastHead 841 Select a session based on the source address and the 842 value of Your Discriminator. If no session is found, a 843 new session of type MultipointClient MAY be created, or 844 the packet MAY be discarded. This choice is outside the 845 scope of this specification. 847 If bfd.SessionType is not MulticastClient, the packet 848 MUST be discarded. 850 Else (Your Discriminator is zero) 852 If the State field is not Down or AdminDown, the packet MUST 853 be discarded. 855 Otherwise, the session MUST be selected based on some 856 combination of other fields, possibly including source 857 addressing information, the My Discriminator field, and the 858 interface over which the packet was received. The exact 859 method of selection is application-specific and is thus 860 outside the scope of this specification. 862 If a matching session is found, and bfd.SessionType is not 863 PointToPoint, the packet MUST be discarded. 865 If a matching session is not found, a new session of type 866 PointToPoint may be created, or the packet may be discarded. 867 This choice is outside the scope of this specification. 869 If the State field is Init and bfd.SessionType is not 870 PointToPoint, the packet MUST be discarded. 872 4.16.3. Transmitting BFD Control Packets 874 The following procedure replaces section 6.8.7 of [BFD]. 876 BFD Control packets MUST be transmitted periodically at the rate 877 determined according to [BFD] section 6.8.2, except as specified in 878 this section. 880 A system MUST NOT transmit any BFD Control packets if bfd.RemoteDiscr 881 is zero and the system is taking the Passive role. 883 A system MUST NOT transmit any BFD Control packets if bfd.SilentTail 884 is 1. 886 A system MUST NOT periodically transmit BFD Control packets if Demand 887 mode is active on the remote system (bfd.RemoteDemandMode is 1, 888 bfd.SessionState is Up, and bfd.RemoteSessionState is Up) and a Poll 889 Sequence is not being transmitted. 891 A system MUST NOT periodically transmit BFD Control packets if 892 bfd.RemoteMinRxInterval is zero. 894 A system MUST NOT periodically transmit BFD Control packets if 895 bfd.SessionType is MulticastClient and a Poll Sequence is not being 896 transmitted. 898 If bfd.SessionType is MultipointHead, the transmit interval MUST be 899 set to bfd.DesiredMinTxInterval (this should happen automatically, as 900 bfd.RemoteMinRxInterval will be zero.) 902 If bfd.SessionType is not MultipointHead, the transmit interval MUST 903 be recalculated whenever bfd.DesiredMinTxInterval changes, or 904 whenever bfd.RemoteMinRxInterval changes, and is equal to the greater 905 of those two values. See [BFD] sections 6.8.2 and 6.8.3 for details 906 on transmit timers. 908 If bfd.SessionType is MulticastTail and periodic transmission of BFD 909 Control packets is just starting (due to Demand mode not being active 910 on the remote system), the first packet to be transmitted MUST be 911 delayed by a random amount of time between zero and (0.9 * 912 bfd.RemoteMinRxInterval). 914 If a BFD Control packet is received with the Poll (P) bit set to 1, 915 the receiving system MUST transmit a BFD Control packet with the Poll 916 (P) bit clear and the Final (F) bit, without respect to the 917 transmission timer or any other transmission limitations, without 918 respect to the session state, and without respect to whether Demand 919 mode is active on either system. A system MAY limit the rate at 920 which such packets are transmitted. If rate limiting is in effect, 921 the advertised value of Desired Min TX Interval MUST be greater than 922 or equal to the interval between transmitted packets imposed by the 923 rate limiting function. If the Multipoint (M) bit is set in the 924 received packet, the packet transmission MUST be delayed by a random 925 amount of time between zero and (0.9 * bfd.RemoteMinRxInterval). 926 Otherwise, the packet MUST be transmitted as soon as practicable. 928 A system MUST NOT set the Demand (D) bit if bfd.SessionType is 929 MultipointTail. 931 A system MUST NOT set the Demand (D) bit if bfd.SessionType is 932 MultipointClient or PointToPoint unless bfd.DemandMode is 1, 933 bfd.SessionState is Up, and bfd.RemoteSessionState is Up. 935 If bfd.SessionType is PointToPoint or MultipointHead, a BFD Control 936 packet SHOULD be transmitted during the interval between periodic 937 Control packet transmissions when the contents of that packet would 938 differ from that in the previously transmitted packet (other than the 939 Poll and Final bits) in order to more rapidly communicate a change in 940 state. 942 The contents of transmitted BFD Control packets MUST be set as 943 follows: 945 Version 947 Set to the current version number (1). 949 Diagnostic (Diag) 951 Set to bfd.LocalDiag. 953 State (Sta) 955 Set to the value indicated by bfd.SessionState. 957 Poll (P) 959 Set to 1 if the local system is sending a Poll Sequence or is a 960 session of type MultipointHead soliciting the identities of the 961 tails, or 0 if not. 963 Final (F) 965 Set to 1 if the local system is responding to a Control packet 966 received with the Poll (P) bit set, or 0 if not. 968 Control Plane Independent (C) 970 Set to 1 if the local system's BFD implementation is independent 971 of the control plane (it can continue to function through a 972 disruption of the control plane.) 974 Authentication Present (A) 976 Set to 1 if authentication is in use on this session (bfd.AuthType 977 is nonzero), or 0 if not. 979 Demand (D) 981 Set to bfd.DemandMode if bfd.SessionState is Up and 982 bfd.RemoteSessionState is Up. Set to 1 if bfd.SessionType is 983 MultipointHead or MultipointClient. Otherwise it is set to 0. 985 Multipoint (M) 987 Set to 1 if bfd.SessionType is MultipointHead. Otherwise it is 988 set to 0. 990 Detect Mult 992 Set to bfd.DetectMult. 994 Length 996 Set to the appropriate length, based on the fixed header length 997 (24) plus any Authentication Section. 999 My Discriminator 1001 Set to bfd.LocalDiscr. 1003 Your Discriminator 1005 Set to bfd.RemoteDiscr. 1007 Desired Min TX Interval 1009 Set to bfd.DesiredMinTxInterval. 1011 Required Min RX Interval 1013 Set to bfd.RequiredMinRxInterval. 1015 Required Min Echo RX Interval 1017 Set to the minimum required Echo packet receive interval for this 1018 session. If this field is set to zero, the local system is 1019 unwilling or unable to loop back BFD Echo packets to the remote 1020 system, and the remote system will not send Echo packets. 1022 Authentication Section 1024 Included and set according to the rules in section 6.7 if 1025 authentication is in use (bfd.AuthType is nonzero.) Otherwise 1026 this section is not present. 1028 5. Assumptions 1030 If head notification is to be used, it is assumed that a multipoint 1031 BFD packet encapsulation contains enough information so that a tail 1032 can address a unicast BFD packet to the head. 1034 If head notification is to be used, it is assumed that is that there 1035 is bidirectional unicast communication available (at the same 1036 protocol layer within which BFD is being run) between the tail and 1037 head. 1039 For the head to know reliably that a tail has lost multipoint 1040 connectivity, the unicast paths in both directions between that tail 1041 and the head must remain operational when the multipoint path fails. 1042 It is thus desirable that unicast paths not share fate with the 1043 multipoint path to the extent possible if the head wants reliable 1044 knowledge of tail state. 1046 Since the normal BFD three-way handshake is not used in this 1047 application, a tail transitioning from state Up to Down and back to 1048 Up again may not be reliably detected at the head. 1050 If authentication is in use, all tails must be configured to have a 1051 common authentication key in order to receive the multipoint BFD 1052 Control packets. 1054 6. Operational Scenarios 1056 It is worth analyzing how this protocol reacts to various scenarios. 1057 There are three path components present, namely, the multipoint tree, 1058 the forward unicast path (from head to a particular tail), and the 1059 reverse unicast path (from a tail to the head.) There are also four 1060 options as to how the head is notified about failures from the tail. 1062 6.1. No Head Notification 1064 Since the only path used in this scenario is the multipoint tree, 1065 none of the others matter. A failure in the multipoint tree will 1066 result in the tail noticing the failure within a detection time, and 1067 the head will remain ignorant of the tail state. 1069 6.2. Unreliable Head Notification 1071 In this scenario, the tail sends back asynchronous BFD packets in 1072 response to the detection of a multipoint path failure. It uses the 1073 reverse unicast path, but not the forward unicast path. 1075 If the multipoint path fails but the reverse unicast path stays up, 1076 the tail will detect the failure within a detection time, and the 1077 head will know about it within one reverse packet time (since the 1078 notification is delayed.) 1080 If both the multipoint path and the reverse unicast paths fail, the 1081 tail will detect the failure but the head will remain unaware of it. 1083 6.3. Semi-reliable Head Notification and Tail Solicitation 1085 In this scenario, the head sends occasional multipoint Polls in 1086 addition to (or in lieu of) non-Poll multipoint BFD Control packets, 1087 expecting the tails to reply with Final. This also uses the reverse 1088 unicast path, but not the forward unicast path. 1090 If the multipoint path fails but the reverse unicast path stays up, 1091 the tail will detect the failure within a detection time, and the 1092 head will know about it within one reverse packet time (the 1093 notification is delayed to avoid synchronization of the tails.) 1095 If both the multipoint path and the reverse unicast paths fail, the 1096 tail will detect the failure but the head will remain unaware of this 1097 fact. 1099 If the reverse unicast path fails but the multipoint path stays up, 1100 the head will see the BFD session fail, but the state of the 1101 multipoint path will be unknown to the head. The tail will continue 1102 to receive multipoint data traffic. 1104 If either the multipoint Poll or the unicast reply is lost in 1105 transit, the head will see the BFD session fail, but the state of the 1106 multipoint path will be unknown to the head. The tail will continue 1107 to receive multipoint data traffic. 1109 6.4. Reliable Head Notification 1111 In this scenario, the head sends occasional multipoint Polls in 1112 addition to (or in lieu of) non-Poll multipoint BFD control packets, 1113 expecting the tails to reply with Final. If a tail that had 1114 previously replied to a multipoint Poll fails to reply (or if the 1115 head simply wishes to verify tail connectivity,) the head issues a 1116 unicast Poll Sequence to the tail. This scenario makes use of all 1117 three paths. 1119 If the multipoint path fails but the two unicast paths stay up, the 1120 tail will detect the failure within a detection time, and the head 1121 will know about it within one reverse packet time (since the 1122 notification is delayed.) Note that the reverse packet time may be 1123 smaller in this case if the head has previously issued a unicast Poll 1124 (since the tail will not delay transmission of the notification in 1125 this case.) 1127 If both the multipoint path and the reverse unicast paths fail 1128 (regardless of the state of the forward unicast path), the tail will 1129 detect the failure but the head will remain unaware of this fact. 1130 The head will detect a BFD session failure to the tail but cannot 1131 make a determination about the state of the tail's multipoint 1132 connectivity. 1134 If the forward unicast path fails but the reverse unicast path stays 1135 up, the head will detect a BFD session failure to the tail if it 1136 happens to send a unicast Poll sequence, but cannot make a 1137 determination about the state of the tail's multipoint connectivity. 1138 If the multipoint path to the tail fails prior to any unicast Poll 1139 being sent, the tail will detect the failure within a detection time, 1140 and the head will know about it within one reverse packet time (since 1141 the notification is delayed.) 1143 If the multipoint path stays up but the reverse unicast path fails, 1144 the head will see the BFD session fail if it happens to send a Poll 1145 Sequence, but the state of the multipoint path will be unknown to the 1146 head. The tail will continue to receive multipoint data traffic. 1148 If the multipoint path and the reverse unicast path both stay up but 1149 the forward unicast path fails, neither side will notice so long as a 1150 unicast Poll Sequence is never sent by the head. If the head sends a 1151 unicast Poll Sequence, the head will see the BFD session fail, but 1152 the state of the multipoint path will be unknown to the head. The 1153 tail will continue to receive multipoint data traffic. 1155 Contributors 1157 Rahul Aggarwal of Juniper Networks and George Swallow of Cisco 1158 Systems provided the initial idea for this specification and 1159 contributed to its development. 1161 Normative References 1163 [BFD] Katz, D., and Ward, D., "Bidirectional Forwarding Detection", 1164 draft-ietf-bfd-base-07.txt, January, 2008. 1166 Security Considerations 1168 This specification does not raise any additional security issues 1169 beyond those of the specifications referred to in the list of 1170 normative references. 1172 IANA Considerations 1174 This document has no actions for IANA. 1176 Authors' Addresses 1178 Dave Katz 1179 Juniper Networks 1180 1194 N. Mathilda Ave. 1181 Sunnyvale, California 94089-1206 USA 1182 Phone: +1-408-745-2000 1183 Email: dkatz@juniper.net 1185 Dave Ward 1186 Cisco Systems 1187 170 W. Tasman Dr. 1188 San Jose, CA 95134 USA 1189 Phone: +1-408-526-4000 1190 Email: dward@cisco.com 1192 Changes from the previous draft 1194 This is a reissue of the previous version. There are only minor 1195 editorial changes. 1197 IPR Disclaimer 1199 The IETF takes no position regarding the validity or scope of any 1200 Intellectual Property Rights or other rights that might be claimed to 1201 pertain to the implementation or use of the technology described in 1202 this document or the extent to which any license under such rights 1203 might or might not be available; nor does it represent that it has 1204 made any independent effort to identify any such rights. Information 1205 on the procedures with respect to rights in RFC documents can be 1206 found in BCP 78 and BCP 79. 1208 Copies of IPR disclosures made to the IETF Secretariat and any 1209 assurances of licenses to be made available, or the result of an 1210 attempt made to obtain a general license or permission for the use of 1211 such proprietary rights by implementers or users of this 1212 specification can be obtained from the IETF on-line IPR repository at 1213 http://www.ietf.org/ipr. 1215 The IETF invites any interested party to bring to its attention any 1216 copyrights, patents or patent applications, or other proprietary 1217 rights that may cover technology that may be required to implement 1218 this standard. Please address the information to the IETF at ietf- 1219 ipr@ietf.org. 1221 Full Copyright Notice 1223 Copyright (C) The IETF Trust (2008). 1225 This document is subject to the rights, licenses and restrictions 1226 contained in BCP 78, and except as set forth therein, the authors 1227 retain all their rights. 1229 This document and the information contained herein are provided on an 1230 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1231 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 1232 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 1233 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 1234 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1235 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1237 Acknowledgement 1239 Funding for the RFC Editor function is currently provided by the 1240 Internet Society. 1242 This document expires in July, 2008.