idnits 2.17.1 draft-ietf-tsvwg-sctp-failover-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 2, 2014) is 3708 days in the past. Is this intentional? Checking references for intended status: Experimental ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 4960 (Obsoleted by RFC 9260) Summary: 1 error (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group Y. Nishida 3 Internet-Draft GE Global Research 4 Intended status: Experimental P. Natarajan 5 Expires: September 3, 2014 Cisco Systems 6 A. Caro 7 BBN Technologies 8 P. Amer 9 University of Delaware 10 K. Nielsen 11 Ericsson 12 March 2, 2014 14 Quick Failover Algorithm in SCTP 15 draft-ietf-tsvwg-sctp-failover-03.txt 17 Abstract 19 One of the major advantages of SCTP is supporting multi-homed 20 communication. If a multi-homed end-point has a redundant network 21 connections, the SCTP associations have a good chance to survive 22 network failures by migrating traffic from inactive networks to 23 active ones. However, if the SCTP standard is followed, there can be 24 a significant delay during the migration. During this period, SCTP 25 might not be able to transmit much data to the peer. This issue 26 drastically impairs the usability of SCTP in some situations. This 27 memo describes the issue of the SCTP failover mechanism and specifies 28 an alternative failover procedure for SCTP that improves its 29 performance during and after failover. The procedures require only 30 minimal modifications to the current specification. 32 Status of this Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at http://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on September 3, 2014. 49 Copyright Notice 51 Copyright (c) 2014 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents 56 (http://trustee.ietf.org/license-info) in effect on the date of 57 publication of this document. Please review these documents 58 carefully, as they describe your rights and restrictions with respect 59 to this document. Code Components extracted from this document must 60 include Simplified BSD License text as described in Section 4.e of 61 the Trust Legal Provisions and are provided without warranty as 62 described in the Simplified BSD License. 64 Table of Contents 66 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 67 2. Conventions and Terminology . . . . . . . . . . . . . . . . . 4 68 3. Issues with the SCTP Path Management . . . . . . . . . . . . . 5 69 4. Existing Solutions for Smooth Failover . . . . . . . . . . . . 6 70 4.1. Reduce Path.Max.Retrans (PMR) . . . . . . . . . . . . . . 6 71 4.2. Adjust RTO related parameters . . . . . . . . . . . . . . 6 72 5. SCTP with Potentially-Failed Destination State (SCTP-PF) . . . 8 73 5.1. SCTP-PF Description . . . . . . . . . . . . . . . . . . . 8 74 5.2. Effect of Path Bouncing . . . . . . . . . . . . . . . . . 10 75 5.3. Permanent Failover . . . . . . . . . . . . . . . . . . . . 10 76 6. Socket API Considerations . . . . . . . . . . . . . . . . . . 12 77 6.1. Support for the Potentially Failed Path State . . . . . . 12 78 6.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket 79 Option . . . . . . . . . . . . . . . . . . . . . . . . . . 13 80 6.3. Exposing the Potentially Failed Path State 81 (SCTP_EXPOSE_POTENTIALLY_FAILED_STATE) Socket Option . . . 14 82 7. Security Considerations . . . . . . . . . . . . . . . . . . . 15 83 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 16 84 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 17 85 9.1. Normative References . . . . . . . . . . . . . . . . . . . 17 86 9.2. Informative References . . . . . . . . . . . . . . . . . . 17 87 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 19 89 1. Introduction 91 The Stream Control Transmission Protocol (SCTP) as specified in 92 [RFC4960] supports multihoming at the transport layer -- an SCTP 93 association can bind to multiple IP addresses at each endpoint. 94 SCTP's multihoming features include failure detection and failover 95 procedures to provide network interface redundancy and improved end- 96 to-end fault tolerance. 98 In SCTP's current failure detection procedure, the sender must 99 experience Path.Max.Retrans (PMR) number of consecutive failed 100 retransmissions on a destination before detecting a path failure. 101 The sender fails over to an alternate active destination only after 102 failure detection. Until detecting the failover, the sender 103 continues to transmit data on the failed path, which degrades the 104 SCTP performance. Concurrent Multipath Transfer (CMT) [IYENGAR06] is 105 an extension to SCTP and allows the sender to transmit data on 106 multiple paths simultaneously. Research [NATARAJAN09] shows that the 107 current failure detection procedure worsens CMT performance during 108 failover and can be significantly improved by employing a better 109 failover algorithm. 111 This document specifies an alternative failure detection procedure 112 for SCTP (and CMT) that improves the SCTP (and CMT) performance 113 during a failover. 115 Also the operation after a failover impacts the performance of the 116 protocol. With [RFC4960] procedures, SCTP will, after a failover 117 from the primary path, switch back to use the primary path for data 118 transfer as soon as this path becomes available. From a performance 119 perspective, as confirmed in research [CARO02], such a switchback of 120 the data transmission path is not optimal in general. As an 121 alternative option to the switchback operation of [RFC4960], this 122 document specifies the support the Permanent Failover switchover 123 procedures proposed by [CARO02]. 125 2. Conventions and Terminology 127 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 128 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 129 document are to be interpreted as described in [RFC2119]. 131 3. Issues with the SCTP Path Management 133 SCTP can utilize multiple IP addresses for a single SCTP association. 134 Each SCTP endpoint exchanges the list of its usable addresses during 135 initial negotiation with its peer. Then the endpoints select one 136 address from the peer's list and define this as the primary 137 destination. During normal transmission, SCTP sends all user data to 138 the primary destination. Also, it sends heartbeat packets to all 139 idle destinations at a certain interval to check the reachability of 140 the path. Idle destinations normally include all non-primary 141 destinations. 143 If a sender has multiple active destination addresses, it can 144 retransmit data to secondary destination address, when the 145 transmission to the primary times out. 147 When a sender receives an acknowledgment for DATA or HEARTBEAT chunks 148 sent to one of the destination addresses, it considers that 149 destination to be active. If it fails to receive acknowledgments, 150 the error count for the address is increased. If the error counter 151 exceeds the protocol parameter 'Path.Max.Retrans', SCTP endpoint 152 considers the address to be inactive. 154 The failover process of SCTP is initiated when the primary path 155 becomes inactive (error counter for the primary path exceeds 156 Path.Max.Retrans). If the primary path is marked inactive, SCTP 157 chooses a new destination address from one of the active destinations 158 and start using this address to send data to. If the primary path 159 becomes active again, SCTP uses the primary destination for 160 subsequent data transmissions and stop using non-primary one. 162 One issue with this failover process is that it usually takes 163 significant amount of time before SCTP switches to the new 164 destination. Let's say the primary path on a multi-homed host 165 becomes unavailable and the RTO value for the primary path at that 166 time is around 1 second, it usually takes over 60 seconds before SCTP 167 starts to use the secondary path. This is because the recommended 168 value for Path.Max.Retrans in the standard is 5, which requires 6 169 consecutive timeouts before failover takes place. Before SCTP 170 switches to the secondary address, SCTP keeps trying to send packets 171 to the primary and only retransmitted packets are sent to the 172 secondary can be reached at the receiver. This slow failover process 173 can cause significant performance degradation and will not be 174 acceptable in some situations. 176 Another issue is that once the primary path is active again, the 177 traffic is switched back. This is not optimal in general. 179 4. Existing Solutions for Smooth Failover 181 The following approaches are conceivable for the solutions of this 182 issue. 184 4.1. Reduce Path.Max.Retrans (PMR) 186 Smaller values for Path.Max.Retrans shorten the failover duration. 187 In fact, this is recommended in some research results [JUNGMAIER02] 188 [GRINNEMO04] [FALLON08]. For example, if when Path.Max.Retrans=0, 189 SCTP switches to another destination on a single timeout. However, 190 smaller value for Path.Max.Retrans also results in spurious failover. 191 In addition, smaller Path.Max.Retrans values also affect 192 'Association.Max.Retrans' values. When the SCTP association's error 193 count (sum of error counts on all ACTIVE paths) exceeds 194 Association.Max.Retrans threshold, the SCTP sender considers the peer 195 endpoint unreachable and terminates the association. Therefore, 196 Section 8.2 in [RFC4960] recommends that Association.Max.Retrans 197 value should not be larger than the summation of the Path.Max.Retrans 198 of each of the destination addresses, else the SCTP sender considers 199 its peer reachable even when all destinations are INACTIVE. To avoid 200 such inconsistent behavior an SCTP implementation SHOULD reduce 201 Association.Max.Retrans accordingly whenever it reduces 202 Path.Max.Retrans. However, smaller Association.Max.Retrans value 203 increases chances of association termination during minor congestion 204 events. 206 Another issue is that the interval of heartbeat packet: 'HB.interval' 207 could be in the order of seconds (recommended value is 30 seconds). 208 When the primary path becomes inactive, the next HB can be 209 transmitted only seconds later. Meanwhile, the primary path may have 210 recovered. In such situations, post failover, an endpoint is forced 211 to wait on the order of seconds before the endpoint can resume 212 transmission on the primary path. 214 The advantage of tuning Path.Max.Retrans is that it requires no 215 modification to the current standard. However, as we discuss above 216 tuning Path.Max.Retrans ignores several recommendations in [RFC4960]. 217 In addition, some research results indicate path bouncing caused by 218 spurious failover does not cause serious problems. We discuss the 219 effect of path bouncing in Section 5.2. 221 4.2. Adjust RTO related parameters 223 As several research results indicate, we can also shorten the 224 duration of failover process by adjusting RTO related parameters 225 [JUNGMAIER02] [FALLON08]. During failover process, RTO keeps being 226 doubled. However, if we can choose smaller value for RTO.max, we can 227 stop the exponential growth of RTO at some point. Also, choosing 228 smaller values for RTO.initial or RTO.min can contribute to keep RTO 229 value small. 231 Similar to reducing Path.Max.Retrans, the advantage of this approach 232 is that it requires no modification to the current specification, 233 although it needs to ignore several recommendations described in the 234 Section 15 of [RFC4960]. However, this approach requires to have 235 enough knowledge about the network characteristics between end 236 points. Otherwise, it can introduce adverse side-effects such as 237 spurious timeouts. 239 5. SCTP with Potentially-Failed Destination State (SCTP-PF) 241 5.1. SCTP-PF Description 243 SCTP-PF stems from the following two observations about SCTP's 244 failure detection procedure: 246 o In order to minimize performance impact during failover, the 247 sender should avoid transmitting data to the failed destination as 248 early as possible. In the current SCTP path management scheme, 249 the sender stops transmitting data to a destination only after the 250 destination is marked Failed. Thus, a smaller PMR value is ideal 251 so that the sender transitions a destination to the Failed state 252 quicker. 254 o Smaller PMR values increase the chances of spurious failure 255 detection where the sender incorrectly marks a destination as 256 Failed during periods of temporary congestion. Larger PMR values 257 are preferable to avoid spurious failure detection. 259 From the above observations it is clear that tweaking the PMR value 260 involves the following tradeoff -- a lower value improves performance 261 but increases the chances of spurious failure detection, whereas a 262 higher value degrades performance and reduces spurious failure 263 detection in a wide range of path conditions. Thus, tweaking the 264 association's PMR value is an incomplete solution to address 265 performance impact during failure. 267 This proposal introduces a new "Potentially-failed" (PF) destination 268 state in SCTP's path management procedure. The PF state was 269 originally proposed to improve CMT performance [NATARAJAN09]. The PF 270 state is an intermediate state between Active and Failed states. 271 SCTP's failure detection procedure is modified to include the PF 272 state. The new failure detection algorithm assumes that loss 273 detected by a timeout implies either severe congestion or failure en- 274 route. After a number of consecutive timeouts on a path, the sender 275 is unsure, and marks the corresponding destination as PF. A PF 276 destination is not used for data transmission except in special cases 277 (discussed below). The new failure detection algorithm requires only 278 sender-side changes. Details are: 280 1. The sender maintains a new tunable parameter called Potentially- 281 failed.Max.Retrans (PFMR). The recommended value of PFMR = 0 282 when quick failover is used. When PFMR is larger or equal to 283 PMR, quick failover is turned off. 285 2. Each time the T3-rtx timer expires on an active destination, the 286 error counter of that destination address will be incremented. 288 When the value in the error counter exceeds PFMR, the endpoint 289 should mark the destination transport address as PF. 291 3. The sender SHOULD avoid data transmission to PF destinations. 292 When all destinations are in either PF or Inactive state, the 293 sender MAY either move the destination from PF to Active state 294 (and transmit data to the active destination) or the sender MAY 295 transmit data to a PF destination. In the former scenario, (i) 296 the sender MUST NOT notify the ULP about the state transition, 297 and (ii) MUST NOT clear the destination's error counter. It is 298 recommended that the sender picks the PF destination with least 299 error count (fewest consecutive timeouts) for data transmission. 300 In case of a tie (multiple PF destinations with same error 301 count), the sender MAY choose the last active destination. 303 4. Only heartbeats MUST be sent to PF destination(s) once per RTO. 304 This means the sender SHOULD ignore HB.interval for PF 305 destinations. If an heartbeat is unanswered, the sender 306 increments the error counter and exponentially backs off the RTO 307 value. If error counter is less than PMR, the sender SHOULD 308 transmit another heartbeat immediately after T3-timer expiration. 310 5. When the sender receives an heartbeat ACK from a PF destination, 311 the sender clears the destination's error counter and transitions 312 the PF destination back to Active state. The sender should 313 perform slow-start as specified in Section 7.2.1 of [RFC4960] 314 when it sends data on this destination. 316 6. Additional (PMR - PFMR) consecutive timeouts on a PF destination 317 confirm the path failure, upon which the destination transitions 318 to the Inactive state. As described in [RFC4960], the sender (i) 319 SHOULD notify ULP about this state transition, and (ii) transmit 320 heartbeats to the Inactive destination at a lower frequency as 321 described in Section 8.3 of [RFC4960]. 323 7. When all destinations are in the Inactive state, the sender picks 324 one of the Inactive destinations for data transmission. This 325 proposal recommends that the sender picks the Inactive 326 destination with least error count (fewest consecutive timeouts) 327 for data transmission. In case of a tie (multiple Inactive 328 destinations with same error count), the sender MAY choose the 329 last active destination. 331 8. ACKs for retransmissions do not transition a PF destination back 332 to Active state, since a sender cannot disambiguate whether the 333 ack was for the original transmission or the retransmission(s). 335 9. SCTP shall provide the means to expose the PF state of its 336 destinations as well as SCTP SHOULD notify the ULP of the state 337 transitions from Active to PF and from PF to Active state. SCTP 338 can provide the means to suppress exposure of PF state and 339 association state transitions and in this case the ULP MAY make 340 SCTP suppress exposure of PF state to ULP. In this case the ULP 341 will rely solely on the [RFC4960] state machine even if quick 342 failover function is activated in SCTP. 344 5.2. Effect of Path Bouncing 346 The methods described above can accelerate the failover process. 347 Hence, they might introduce the path bouncing effect where the sender 348 keeps changing the data transmission path frequently. This sounds 349 harmful to the data transfer, however several research results 350 indicate that there is no serious problem with SCTP in terms of path 351 bouncing effect [CARO04] [CARO05]. 353 There are two main reasons for this. First, SCTP is basically 354 designed for multipath communication, which means SCTP maintains all 355 path related parameters (CWND, ssthresh, RTT, error count, etc) per 356 each destination address. These parameters cannot be affected by 357 path bouncing. In addition, when SCTP migrates the data transfer to 358 another path, it starts with the minimal or the initial CWND. Hence, 359 there is little chance for packet reordering or duplicating. 361 Second, even if all communication paths between the end-nodes share 362 the same bottleneck, the quick failover results in a behavior already 363 allowed by [RFC4960]. 365 5.3. Permanent Failover 367 Post failover then, by [RFC4960] behavior, an SCTP sender migrates 368 the traffic back to the original primary destination once this 369 destination becomes active anew. As the CWND towards the original 370 primary destination has to be rebuilt once data transfer resumes, the 371 switch back to use the original primary path is not always optimal. 372 Indeed [CARO02] shows that the switch over to the original primary 373 may degrade SCTP performance compared to continuing data transmission 374 on the same path, especially, but not only, in scenarios where this 375 path's characteristics are better. In order to mitigate this 376 performance degradation, Permanent Failover operation was proposed in 377 [CARO02]. When SCTP changes the destination due to failover, 378 Permanent Failover marks it as new primary. This means Permanent 379 Failover allows SCTP sender to continue data transmission to the path 380 even after the old primary destination becomes active again. This is 381 achieved by having SCTP perform a switchover of the primary path to 382 an alternative working path rather than having SCTP switch back data 383 transfer to the (previous) primary path. 385 The manner of switchover operation that is most optimal in a given 386 scenario depends on the relative quality of a set primary path versus 387 the quality of alternative paths available as well as it depends on 388 the extent to which it is desired for the mode of operation to 389 enforce traffic distribution over a number of network paths. I.e., 390 load distribution of traffic from multiple SCTP associations may be 391 sought to be enforced by distribution of the set primary paths with 392 [RFC4960] switchback operation. However as [RFC4960] switchback 393 behavior is suboptimal in certain situations, especially in scenarios 394 where a number of equally good paths are available, it is recommended 395 for SCTP to support also, as alternative behavior, the Permanent 396 Failover modes of operation where forced switch back to a previously 397 failed primary path is not always performed. The Permanent Failover 398 operation requires only sender side changes. Details, as originally 399 outlined in [CARO02], are: 401 1. The sender maintains a new tunable parameter, called 402 Primary.Switchover.Max.Retrans (PSMR). When the path error 403 counter on a set primary path exceeds PSMR, the SCTP 404 implementation autonomously selects and sets a new primary path. 406 2. The primary path selected by the SCTP implementation shall be the 407 path which at the given time would be chosen for data transfer. 408 A previously failed primary path may come in use as data transfer 409 path as per normal path selection when the present data transfer 410 path fails. 412 3. The recommended value of PSMR is PFMR when Permanent failover is 413 used. This means that no forced switchback to a previously 414 failed primary path is performed. 416 4. It must be possible to disable the Permanent Failover and obtain 417 the standard switchback operation of [RFC4960]. 419 We recommend that SCTP-PF should stick to the standard RFC4960 420 behavior as default, i.e., switch back to the old primary destination 421 once the destination becomes active again. However, implementors MAY 422 implement Permanent Failover and MAY enable it based on network 423 configurations or users' requests. 425 6. Socket API Considerations 427 This section describes how the socket API defined in [RFC6458] is 428 extended to provide a way for the application to control and observe 429 the quick failover behavior. 431 Please note that this section is informational only. 433 A socket API implementation based on [RFC6458] is, by means of the 434 existing SCTP_PEER_ADDR_CHANGE event, extended to provide the event 435 notification when a peer address enters or leaves the potentially 436 failed state as well as the socket API implementation is extended to 437 expose the potentially failed state of a peer address in the existing 438 SCTP_GET_PEER_ADDR_INFO structure. 440 Furthermore, two new read/write socket options for the level 441 IPPROTO_SCTP and the name SCTP_PEER_ADDR_THLDS and 442 SCTP_EXPOSE_POTENTIALLY_FAILED_STATE are defined as described below. 443 The first socket option is used to control the values of the PFMR and 444 PSMR parameters described in Section 5. The second one controls the 445 exposition of the potentially failed path state. 447 Support for the SCTP_PEER_ADDR_THLDS and 448 SCTP_EXPOSE_POTENTIALLY_FAILED_STATE socket options need also to be 449 added to the function sctp_opt_info(). 451 6.1. Support for the Potentially Failed Path State 453 As defined in [RFC6458], the SCTP_PEER_ADDR_CHANGE event is provided 454 if the status of a peer address changes. In addition to the state 455 changes described in [RFC6458], this event is also provided, if a 456 peer address enters or leaves the potentially failed state. The 457 notification as defined in [RFC6458] uses the following structure: 459 struct sctp_paddr_change { 460 uint16_t spc_type; 461 uint16_t spc_flags; 462 uint32_t spc_length; 463 struct sockaddr_storage spc_aaddr; 464 uint32_t spc_state; 465 uint32_t spc_error; 466 sctp_assoc_t spc_assoc_id; 467 } 469 [RFC6458] defines the constants SCTP_ADDR_AVAILABLE, 470 SCTP_ADDR_UNREACHABLE, SCTP_ADDR_REMOVED, SCTP_ADDR_ADDED, and 471 SCTP_ADDR_MADE_PRIM to be provided in the spc_state field. This 472 document defines in addition to that the new constant 473 SCTP_ADDR_POTENTIALLY_FAILED, which is reported if the affected 474 address becomes potentially failed. 476 The SCTP_GET_PEER_ADDR_INFO socket option defined in [RFC6458] can be 477 used to query the state of a peer address. It uses the following 478 structure: 480 struct sctp_paddrinfo { 481 sctp_assoc_t spinfo_assoc_id; 482 struct sockaddr_storage spinfo_address; 483 int32_t spinfo_state; 484 uint32_t spinfo_cwnd; 485 uint32_t spinfo_srtt; 486 uint32_t spinfo_rto; 487 uint32_t spinfo_mtu; 488 }; 490 [RFC6458] defines the constants SCTP_UNCONFIRMED, SCTP_ACTIVE, and 491 SCTP_INACTIVE to be provided in the spinfo_state field. This 492 document defines in addition to that the new constant 493 SCTP_POTENTIALLY_FAILED, which is reported if the peer address is 494 potentially failed. 496 6.2. Peer Address Thresholds (SCTP_PEER_ADDR_THLDS) Socket Option 498 Applications can control the quick failover behavior by getting or 499 setting the number of consecutive timeouts before a peer address is 500 considered potentially failed or unreachable and before the primary 501 path is changed automatically. This socket option uses the level 502 IPPROTO_SCTP and the name SCTP_PEER_ADDR_THLDS. 504 The following structure is used to access and modify the thresholds: 506 struct sctp_paddrthlds { 507 sctp_assoc_t spt_assoc_id; 508 struct sockaddr_storage spt_address; 509 uint16_t spt_pathmaxrxt; 510 uint16_t spt_pathpfthld; 511 uint16_t spt_pathcpthld; 512 }; 514 spt_assoc_id: This parameter is ignored for one-to-one style 515 sockets. For one-to-many style sockets the application may fill 516 in an association identifier or SCTP_FUTURE_ASSOC. It is an error 517 to use SCTP_{CURRENT|ALL}_ASSOC in spt_assoc_id. 519 spt_address: This specifies which peer address is of interest. If a 520 wildcard address is provided, this socket option applies to all 521 current and future peer addresses. 523 spt_pathmaxrxt: Each peer address of interest is considered 524 unreachable, if its path error counter exceeds spt_pathmaxrxt. 526 spt_pathpfthld: Each peer address of interest is considered 527 potentially failed, if its path error counter exceeds 528 spt_pathpfthld. 530 spt_pathcpthld: Each peer address of interest is not considered the 531 primary remote address anymore, if its path error counter exceeds 532 spt_pathcpthld. Using a value of 0xffff disables the selection of 533 a new primary peer address. If an implementation does not support 534 the automatically selection of a new primary address, it should 535 indicate an error with errno set to EINVAL if a value different 536 from 0xffff is used in spt_pathcpthld. 538 6.3. Exposing the Potentially Failed Path State 539 (SCTP_EXPOSE_POTENTIALLY_FAILED_STATE) Socket Option 541 Applications can control the exposure of the potentially failed path 542 state in the SCTP_PEER_ADDR_CHANGE event and the 543 SCTP_GET_PEER_ADDR_INFO as described in Section 6.1. The default 544 value is implementation specific. 546 This socket option uses the level IPPROTO_SCTP and the name 547 SCTP_EXPOSE_POTENTIALLY_FAILED_STATE. 549 The following structure is used to control the exposition of the 550 potentially failed path state: 552 struct sctp_assoc_value { 553 sctp_assoc_t assoc_id; 554 uint32_t assoc_value; 555 }; 557 assoc_id: This parameter is ignored for one-to-one style sockets. 558 For one-to-many style sockets the application may fill in an 559 association identifier or SCTP_FUTURE_ASSOC. It is an error to 560 use SCTP_{CURRENT|ALL}_ASSOC in assoc_id. 562 assoc_value: The potentially failed path state is exposed if and 563 only if this parameter is non-zero. 565 7. Security Considerations 567 There are no new security considerations introduced in this document. 569 8. IANA Considerations 571 This document does not create any new registries or modify the rules 572 for any existing registries managed by IANA. 574 9. References 576 9.1. Normative References 578 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 579 Requirement Levels", BCP 14, RFC 2119, March 1997. 581 [RFC4960] Stewart, R., "Stream Control Transmission Protocol", 582 RFC 4960, September 2007. 584 9.2. Informative References 586 [CARO02] Caro Jr., A., Iyengar, J., Amer, P., Heinz, G., and R. 587 Stewart, "A Two-level Threshold Recovery Mechanism for 588 SCTP", Tech report, CIS Dept, University of Delaware , 589 7 2002. 591 [CARO04] Caro Jr., A., Amer, P., and R. Stewart, "End-to-End 592 Failover Thresholds for Transport Layer Multihoming", 593 MILCOM 2004 , 11 2004. 595 [CARO05] Caro Jr., A., "End-to-End Fault Tolerance using Transport 596 Layer Multihoming", Ph.D Thesis, University of Delaware , 597 1 2005. 599 [FALLON08] 600 Fallon, S., Jacob, P., Qiao, Y., Murphy, L., Fallon, E., 601 and A. Hanley, "SCTP Switchover Performance Issues in WLAN 602 Environments", IEEE CCNC 2008, 1 2008. 604 [GRINNEMO04] 605 Grinnemo, K-J. and A. Brunstrom, "Performance of SCTP- 606 controlled failovers in M3UA-based SIGTRAN networks", 607 Advanced Simulation Technologies Conference , 4 2004. 609 [IYENGAR06] 610 Iyengar, J., Amer, P., and R. Stewart, "Concurrent 611 Multipath Transfer using SCTP Multihoming over Independent 612 End-to-end Paths.", IEEE/ACM Trans on Networking 14(5), 613 10 2006. 615 [JUNGMAIER02] 616 Jungmaier, A., Rathgeb, E., and M. Tuexen, "On the use of 617 SCTP in failover scenarios", World Multiconference on 618 Systemics, Cybernetics and Informatics , 7 2002. 620 [NATARAJAN09] 621 Natarajan, P., Ekiz, N., Amer, P., and R. Stewart, 622 "Concurrent Multipath Transfer during Path Failure", 623 Computer Communications , 5 2009. 625 [RFC6458] Stewart, R., Tuexen, M., Poon, K., Lei, P., and V. 626 Yasevich, "Sockets API Extensions for the Stream Control 627 Transmission Protocol (SCTP)", RFC 6458, December 2011. 629 Authors' Addresses 631 Yoshifumi Nishida 632 GE Global Research 633 2623 Camino Ramon 634 San Ramon, CA 94583 635 USA 637 Email: nishida@wide.ad.jp 639 Preethi Natarajan 640 Cisco Systems 641 510 McCarthy Blvd 642 Milpitas, CA 95035 643 USA 645 Email: prenatar@cisco.com 647 Armando Caro 648 BBN Technologies 649 10 Moulton St. 650 Cambridge, MA 02138 651 USA 653 Email: acaro@bbn.com 655 Paul D. Amer 656 University of Delaware 657 Computer Science Department - 434 Smith Hall 658 Newark, DE 19716-2586 659 USA 661 Email: amer@udel.edu 663 Karen E. E. Nielsen 664 Ericsson 665 Kistavaegen 25 666 Stockholm, 164 80 667 Sweden 669 Email: karen.nielsen@tieto.com