idnits 2.17.1 draft-ietf-tcpm-ecnsyn-03.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 1017. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1028. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1035. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1041. 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 (18 November 2007) is 5997 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Obsolete informational reference (is this intentional?): RFC 2309 (Obsoleted by RFC 7567) -- Obsolete informational reference (is this intentional?): RFC 2581 (Obsoleted by RFC 5681) -- Obsolete informational reference (is this intentional?): RFC 2988 (Obsoleted by RFC 6298) == Outdated reference: A later version (-05) exists of draft-irtf-tmrg-tools-04 Summary: 1 error (**), 0 flaws (~~), 2 warnings (==), 10 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Internet Engineering Task Force A. Kuzmanovic 2 INTERNET-DRAFT A. Mondal 3 Intended status: Proposed Standard Northwestern University 4 Expires: 18 May 2008 S. Floyd 5 ICIR 6 K.K. Ramakrishnan 7 AT&T 8 18 November 2007 10 Adding Explicit Congestion Notification (ECN) Capability 11 to TCP's SYN/ACK Packets 12 draft-ietf-tcpm-ecnsyn-03.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 December 2007. 39 Copyright Notice 41 Copyright (C) The IETF Trust (2007). 43 Abstract 45 This draft specifies a modification to RFC 3168 to allow TCP SYN/ACK 46 packets to be ECN-Capable. For TCP, RFC 3168 only specifies setting 47 an ECN-Capable codepoint on data packets, and not on SYN and SYN/ACK 48 packets. However, because of the high cost to the TCP transfer of 49 having a SYN/ACK packet dropped, with the resulting retransmit 50 timeout, this document specifies the use of ECN for the SYN/ACK 51 packet itself, when sent in response to a SYN packet with the two ECN 52 flags set in the TCP header, indicating a willingness to use ECN. 53 Setting TCP SYN/ACK packets as ECN-Capable can be of great benefit to 54 the TCP connection, avoiding the severe penalty of a retransmit 55 timeout for a connection that has not yet started placing a load on 56 the network. The sender of the SYN/ACK packet must respond to a 57 report of an ECN-marked SYN/ACK packet by reducing its initial 58 congestion window from two, three, or four segments to one segment, 59 thereby reducing the subsequent load from that connection on the 60 network. 62 Table of Contents 64 1. Introduction ....................................................4 65 2. Conventions .....................................................5 66 3. Proposal ........................................................6 67 4. Discussion ......................................................9 68 5. Related Work ...................................................12 69 6. Performance Evaluation .........................................13 70 6.1. The Costs and Benefit of Adding ECN-Capability ............13 71 6.2. An Evaluation of Different Responses to ECN-Marked SYN/ACK 72 Packets ........................................................14 73 7. Security Considerations ........................................15 74 8. Conclusions ....................................................16 75 9. Acknowledgements ...............................................17 76 A. Report on Simulations ..........................................17 77 A.1. Simulations with RED in Packet Mode .......................18 78 A.2. Simulations with RED in Byte Mode .........................19 79 Normative References ..............................................20 80 Informative References ............................................20 81 IANA Considerations ...............................................22 82 Full Copyright Statement ..........................................22 83 Intellectual Property .............................................23 85 NOTE TO RFC EDITOR: PLEASE DELETE THIS NOTE UPON PUBLICATION. 87 Changes from draft-ietf-tcpm-ecnsyn-02: 89 * Added to the discussion in the Security section of whether 90 ECN-Capable TCP SYN packets have problems with firewalls, 91 over and above the known problems of TCP data packets 92 (e.g., as in the Microsoft report). From a question raised 93 at the TCPM meeting at the July 2007 IETF. 95 * Added a sentence to the discussion of routers or middleboxes that 96 *might* drop TCP SYN packets on the basis of IP header fields. 97 Feedback from Remi Denis-Courmont. 99 * General editing. Feedback from Alfred Henes. 101 Changes from draft-ietf-tcpm-ecnsyn-01: 103 * Changes in response to feedback from Anil Agarwal. 105 * Added a look at the costs of adding ECN-Capability to 106 SYN/ACKs in a highly-congested scenario. 107 From feedback from Mark Allman and Janardhan Iyengar. 109 * Added a comparative evaluation of two possible responses 110 to an ECN-marked SYN/ACK packet. From Mark Allman. 112 Changes from draft-ietf-tcpm-ecnsyn-00: 114 * Only updating the revision number. 116 Changes from draft-ietf-twvsg-ecnsyn-00: 118 * Changed name of draft to draft-ietf-tcpm-ecnsyn. 120 * Added a discussion in Section 3 of "Response to 121 ECN-marking of SYN/ACK packets". Based on 122 suggestions from Mark Allman. 124 * Added a discussion to the Conclusions about adding 125 ECN-capability to relevant set-up packets in other 126 protocols. From a suggestion from Wesley Eddy. 128 * Added a description of SYN exchanges with SYN cookies. 129 From a suggestion from Wesley Eddy. 131 * Added a discussion of one-way data transfers, where the 132 host sending the SYN/ACK packet sends no data packets. 134 * Minor editing, from feedback from Mark Allman and Janardhan 135 Iyengar. 137 * Future work: a look at the costs of adding 138 ECN-Capability in a worst-case scenario. 139 From feedback from Mark Allman and Janardhan Iyengar. 141 * Future work: a comparative evaluation of two 142 possible responses to an ECN-marked SYN/ACK packet. 144 Changes from draft-kuzmanovic-ecn-syn-00.txt: 146 * Changed name of draft to draft-ietf-twvsg-ecnsyn. 148 END OF NOTE TO RFC EDITOR. 150 1. Introduction 152 TCP's congestion control mechanism has primarily used packet loss as 153 the congestion indication, with packets dropped when buffers 154 overflow. With such tail-drop mechanisms, the packet delay can be 155 high, as the queue at bottleneck routers can be fairly large. 156 Dropping packets only when the queue overflows, and having TCP react 157 only to such losses, results in: 158 1) significantly higher packet delay; 159 2) unnecessarily many packet losses; and 160 3) unfairness due to synchronization effects. 162 The adoption of Active Queue Management (AQM) mechanisms allows 163 better control of bottleneck queues [RFC2309]. This use of AQM has 164 the following potential benefits: 165 1) better control of the queue, with reduced queueing delay; 166 2) fewer packet drops; and 167 3) better fairness because of fewer synchronization effects. 169 With the adoption of ECN, performance may be further improved. When 170 the router detects congestion before buffer overflow, the router can 171 provide a congestion indication either by dropping a packet, or by 172 setting the Congestion Experienced (CE) codepoint in the Explicit 173 Congestion Notification (ECN) field in the IP header [RFC3168]. The 174 IETF has standardized the use of the Congestion Experienced (CE) 175 codepoint in the IP header for routers to indicate congestion. For 176 incremental deployment and backwards compatibility, the RFC on ECN 177 [RFC3168] specifies that routers may mark ECN-capable packets that 178 would otherwise have been dropped, using the Congestion Experienced 179 codepoint in the ECN field. The use of ECN allows TCP to react to 180 congestion while avoiding unnecessary retransmissions and, in some 181 cases, unnecessary retransmit timeouts. Thus, using ECN has several 182 benefits: 184 1) For short transfers, a TCP connection's congestion window may be 185 small. For example, if the current window contains only one packet, 186 and that packet is dropped, TCP will have to wait for a retransmit 187 timeout to recover, reducing its overall throughput. Similarly, if 188 the current window contains only a few packets and one of those 189 packets is dropped, there might not be enough duplicate 190 acknowledgements for a fast retransmission, and the sender might have 191 to wait for a delay of several round-trip times using Limited 192 Transmit [RFC3042]. With the use of ECN, short flows are less likely 193 to have packets dropped, sometimes avoiding unnecessary delays or 194 costly retransit timeouts. 196 2) While longer flows may not see substantially improved throughput 197 with the use of ECN, they experience lower loss. This may benefit TCP 198 applications that are latency- and loss-sensitive, because of the 199 avoidance of retransmissions. 201 RFC 3168 only specifies marking the Congestion Experienced codepoint 202 on TCP's data packets, and not on SYN and SYN/ACK packets. RFC 3168 203 specifies the negotiation of the use of ECN between the two TCP end- 204 points in the TCP SYN and SYN-ACK exchange, using flags in the TCP 205 header. Erring on the side of being conservative, RFC 3168 does not 206 specify the use of ECN for the SYN/ACK packet itself. However, 207 because of the high cost to the TCP transfer of having a SYN/ACK 208 packet dropped, with the resulting retransmit timeout, this document 209 specifies the use of ECN for the SYN/ACK packet itself. This can be 210 of great benefit to the TCP connection, avoiding the severe penalty 211 of a retransmit timeout for a connection that has not yet started 212 placing a load on the network. The sender of the SYN/ACK packet must 213 respond to a report of an ECN-marked SYN/ACK packet by reducing its 214 initial congestion window from two, three, or four segments to one 215 segment, reducing the subsequent load from that connection on the 216 network. 218 The use of ECN for SYN/ACK packets has the following potential 219 benefits: 220 1) Avoidance of a retransmit timeout; 221 2) Improvement in the throughput of short connections. 223 This draft specifies ECN+, a modification to RFC 3168 to allow TCP 224 SYN/ACK packets to be ECN-Capable. Section 3 contains the 225 specification of the change, while Section 4 discusses some of the 226 issues, and Section 5 discusses related work. Section 6 contains an 227 evaluation of the proposed change. 229 2. Conventions 231 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 232 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 233 document are to be interpreted as described in [RFC 2119]. 235 3. Proposal 237 This section specifies the modification to RFC 3168 to allow TCP 238 SYN/ACK packets to be ECN-Capable. We use the following terminology 239 from RFC 3168: 241 The ECN field in the IP header: 242 o CE: the Congestion Experienced codepoint; and 243 o ECT: either one of the two ECN-Capable Transport codepoints. 245 The ECN flags in the TCP header: 246 o CWR: the Congestion Window Reduced flag; and 247 o ECE: the ECN-Echo flag. 249 ECN-setup packets: 250 o ECN-setup SYN packet: a SYN packet with the ECE and CWR flags; 251 o ECN-setup SYN-ACK packet: a SYN-ACK packet with ECE but not CWR. 253 RFC 3168 in Section 6.1.1. states that "A host MUST NOT set ECT on 254 SYN or SYN-ACK packets." In this section, we specify that a TCP node 255 MAY respond to an ECN-setup SYN packet by setting ECT in the 256 responding ECN-setup SYN/ACK packet, indicating to routers that the 257 SYN/ACK packet is ECN-Capable. This allows a congested router along 258 the path to mark the packet instead of dropping the packet as an 259 indication of congestion. 261 Assume that TCP node A transmits to TCP node B an ECN-setup SYN 262 packet, indicating willingness to use ECN for this connection. As 263 specified by RFC 3168, if TCP node B is willing to use ECN, node B 264 responds with an ECN-setup SYN-ACK packet. 266 Figure 1 shows an interchange with the SYN/ACK packet dropped by a 267 congested router. Node B waits for a retransmit timeout, and then 268 retransmits the SYN/ACK packet. 270 --------------------------------------------------------------- 271 TCP Node A Router TCP Node B 272 ---------- ------ ---------- 274 ECN-setup SYN packet ---> 275 ECN-setup SYN packet ---> 277 <--- ECN-setup SYN/ACK, possibly ECT 278 3-second timer set 279 SYN/ACK dropped . 280 . 281 . 282 3-second timer expires 283 <--- ECN-setup SYN/ACK, not ECT 284 <--- ECN-setup SYN/ACK 285 Data/ACK ---> 286 Data/ACK ---> 287 <--- Data (one to four segments) 288 --------------------------------------------------------------- 290 Figure 1: SYN exchange with the SYN/ACK packet dropped. 292 If the SYN/ACK packet is dropped in the network, the TCP host (node 293 B) responds by waiting three seconds for the retransmit timer to 294 expire [RFC2988]. If a SYN/ACK packet with the ECT codepoint is 295 dropped, the TCP node SHOULD resend the SYN/ACK packet without the 296 ECN-Capable codepoint. (Although we are not aware of any middleboxes 297 that drop SYN/ACK packets that contain an ECN-Capable codepoint in 298 the IP header, we have learned to design our protocols defensively in 299 this regard [RFC3360].) 301 We note that if syn-cookies were used by Node B in the exchange in 302 Figure 1, TCP Node B wouldn't set a timer upon transmission of the 303 SYN/ACK packet [SYN-COOK]. In this case, if the SYN/ACK packet was 304 lost, the initiator (Node A) would have to timeout and retransmit the 305 SYN packet in order to trigger another SYN-ACK. 307 Figure 2 shows an interchange with the SYN/ACK packet sent as ECN- 308 Capable, and ECN-marked instead of dropped at the congested router. 310 --------------------------------------------------------------- 311 TCP Node A Router TCP Node B 312 ---------- ------ ---------- 314 ECN-setup SYN packet ---> 315 ECN-setup SYN packet ---> 317 <--- ECN-setup SYN/ACK, ECT 318 <--- Sets CE on SYN/ACK 319 <--- ECN-setup SYN/ACK, CE 321 Data/ACK, ECN-Echo ---> 322 Data/ACK, ECN-Echo ---> 323 Window reduced to one segment. 324 <--- Data, CWR (one segment only) 325 --------------------------------------------------------------- 327 Figure 2: SYN exchange with the SYN/ACK packet marked. 329 If the receiving node (node A) receives a SYN/ACK packet that has 330 been marked by the congested router, with the CE codepoint set, the 331 receiving node MUST respond by setting the ECN-Echo flag in the TCP 332 header of the responding ACK packet. As specified in RFC 3168, the 333 receiving node continues to set the ECN-Echo flag in packets until it 334 receives a packet with the CWR flag set. 336 When the sending node (node B) receives the ECN-Echo packet reporting 337 the Congestion Experienced indication in the SYN/ACK packet, the node 338 MUST set the initial congestion window to one segment, instead of two 339 segments as allowed by [RFC2581], or three or four segments allowed 340 by [RFC3390]. If the sending node (node B) was going to use an 341 initial window of one segment, and receives an ECN-Echo packet 342 informing it of a Congestion Experienced indication on its SYN/ACK 343 packet, the sending node MAY continue to send with an initial window 344 of one segment, without waiting for a retransmit timeout. We note 345 that this updates RFC 3168, which specifies that "the sending TCP 346 MUST reset the retransmit timer on receiving the ECN-Echo packet when 347 the congestion window is one." As specified by RFC 3168, the sending 348 node (node B) also sets the CWR flag in the TCP header of the next 349 data packet sent, to acknowledge its receipt of and reaction to the 350 ECN-Echo flag. 352 If the data transfer in Figure 2 is entirely from Node A to Node B, 353 then data packets from Node A continue to set the ECN-Echo flag in 354 data packets, waiting for the CWR flag from Node B acknowledging a 355 response to the ECN-Echo flag. 357 4. Discussion 359 Motivation: 360 The rationale for the proposed change is the following. When node B 361 receives a TCP SYN packet with ECN-Echo bit set in the TCP header, 362 this indicates that node A is ECN-capable. If node B is also ECN- 363 capable, there are no obstacles to immediately setting one of the 364 ECN-Capable codepoints in the IP header in the responding TCP SYN/ACK 365 packet. 367 There can be a great benefit in setting an ECN-capable codepoint in 368 SYN/ACK packets, as is discussed further in [ECN+], and reported 369 briefly in Section 5 below. Congestion is most likely to occur in 370 the server-to-client direction. As a result, setting an ECN-capable 371 codepoint in SYN/ACK packets can reduce the occurrence of three- 372 second retransmit timeouts resulting from the drop of SYN/ACK 373 packets. 375 Flooding attacks: 376 Setting an ECN-Capable codepoint in the responding TCP SYN/ACK 377 packets does not raise any novel security vulnerabilities. For 378 example, provoking servers or hosts to send SYN/ACK packets to a 379 third party in order to perform a "SYN/ACK flood" attack would be 380 highly inefficient. Third parties would immediately drop such 381 packets, since they would know that they didn't generate the TCP SYN 382 packets in the first place. Moreover, such SYN/ACK attacks would 383 have the same signatures as the existing TCP SYN attacks. Provoking 384 servers or hosts to reply with SYN/ACK packets in order to congest a 385 certain link would also be highly inefficient because SYN/ACK packets 386 are small in size. 388 However, the addition of ECN-Capability to SYN/ACK packets could 389 allow SYN/ACK packets to persist for more hops along a network path 390 before being dropped, thus adding somewhat to the ability of a 391 SYN/ACK attack to flood a network link. 393 The TCP SYN packet: 394 There are several reasons why an ECN-Capable codepoint MUST NOT be 395 set in the IP header of the initiating TCP SYN packet. First, when 396 the TCP SYN packet is sent, there are no guarantees that the other 397 TCP endpoint (node B in Figure 2) is ECN-capable, or that it would be 398 able to understand and react if the ECN CE codepoint was set by a 399 congested router. 401 Second, the ECN-Capable codepoint in TCP SYN packets could be misused 402 by malicious clients to `improve' the well-known TCP SYN attack. By 403 setting an ECN-Capable codepoint in TCP SYN packets, a malicious host 404 might be able to inject a large number of TCP SYN packets through a 405 potentially congested ECN-enabled router, congesting it even further. 407 For both these reasons, we continue the restriction that the TCP SYN 408 packet MUST NOT have the ECN-Capable codepoint in the IP header set. 410 Backwards compatibility: 411 In order for TCP node B to send a SYN/ACK packet as ECN-Capable, node 412 B must have received an ECN-setup SYN packet from node A. However, 413 it is possible that node A supports ECN, but either ignores the CE 414 codepoint on received SYN/ACK packets, or ignores SYN/ACK packets 415 with the ECT or CE codepoint set. If the TCP sender ignores the CE 416 codepoint on received SYN/ACK packets, this would mean that the TCP 417 connection would not respond to this congestion indication. However, 418 this seems to us an acceptable cost to pay in the incremental 419 deployment of ECN-Capability for TCP's SYN/ACK packets. It would 420 mean that the sender of the SYN/ACK packet would not reduce the 421 initial congestion window from two, three, or four segments down to 422 one segment, as it should. However, the TCP sender would still 423 respond correctly to any subsequent CE indications on data packets 424 later on in the connection. Thus, to be explicit, when a TCP 425 connection includes a sender that supports ECN but *does not* support 426 ECN-Capability for SYN/ACK packets, in combination with a receiver 427 that *does* support ECN-Capabililty for SYN/ACK packets, it is quite 428 possible that the ECN-Capable SYN/ACK packets will be marked rather 429 than dropped in the network, and that the sender will not respond to 430 the ECN mark on the SYN/ACK packet. 432 It is also possible that in some older TCP implementation, the TCP 433 sender would ignore arriving SYN/ACK packets that had the ECT or CE 434 codepoint set. This would result in a delay in connection set-up for 435 that TCP connection, with the TCP sender re-sending the SYN packet 436 after a retransmit timeout. We are not aware of any TCP 437 implementations with this behavior. 439 SYN/ACK packets and packet size: 440 There are a number of router buffer architectures that have smaller 441 dropping rates for small (SYN) packets than for large (data) packets. 442 For example, for a Drop Tail queue in units of packets, where each 443 packet takes a single slot in the buffer regardless of packet size, 444 small and large packets are equally likely to be dropped. However, 445 for a Drop Tail queue in units of bytes, small packets are less 446 likely to be dropped than are large ones. Similarly, for RED in 447 packet mode, small and large packets are equally likely to be dropped 448 or marked, while for RED in byte mode, a packet's chance of being 449 dropped or marked is proportional to the packet size in bytes. 451 For a congested router with an AQM mechanism in byte mode, where a 452 packet's chance of being dropped or marked is proportional to the 453 packet size in bytes, the drop or marking rate for TCP SYN/ACK 454 packets should generally be low. In this case, the benefit of making 455 SYN/ACK packets ECN-Capable should be similarly moderate. However, 456 for a congested router with a Drop Tail queue in units of packets or 457 with an AQM mechanism in packet mode, and with no priority queueing 458 for smaller packets, small and large packets should have the same 459 probability of being dropped or marked. In such a case, making 460 SYN/ACK packets ECN-Capable should be of significant benefit. 462 We believe that there are a wide range of behaviors in the real world 463 in terms of the drop or mark behavior at routers as a function of 464 packet size [Tools] (Section 10). We note that all of these 465 alternatives listed above are available in the NS simulator (Drop 466 Tail queues are by default in units of packets, while the default for 467 RED queue management has been changed from packet mode to byte mode). 469 Response to ECN-marking of SYN/ACK packets: 470 One question is why TCP SYN/ACK packets should be treated differently 471 from other packets in terms of the packet sender's response to an 472 ECN-marked packet. Section 5 of RFC 3168 specifies the following: 474 "Upon the receipt by an ECN-Capable transport of a single CE packet, 475 the congestion control algorithms followed at the end-systems MUST be 476 essentially the same as the congestion control response to a *single* 477 dropped packet. For example, for ECN-Capable TCP the source TCP is 478 required to halve its congestion window for any window of data 479 containing either a packet drop or an ECN indication." 481 In particular, Section 6.1.2 of RFC 3168 specifies that when the TCP 482 congestion window consists of a single packet and that packet is ECN- 483 marked in the network, then the sender must reduce the sending rate 484 below one packet per round-trip time, by waiting for one RTO before 485 sending another packet. If the RTO was set to the average round-trip 486 time, this would result in halving the sending rate; because the RTO 487 is in fact larger than the average round-trip time, the sending rate 488 is reduced to less than half of its previous value. 490 TCP's congestion control response to the *dropping* of a SYN/ACK 491 packet is to wait a default time before sending another packet. This 492 document argues that ECN gives end-systems a wider range of possible 493 responses to the *marking* of a SYN/ACK packet, and that waiting a 494 default time before sending a data packet is not the desired 495 response. 497 On the conservative end, one could assume an effective congestion 498 window of one packet for the SYN/ACK packet, and respond to an ECN- 499 marked SYN/ACK packet by reducing the sending rate to one packet 500 every two round-trip times. As an approximation, the TCP end-node 501 could measure the round-trip time T between the sending of the 502 SYN/ACK packet and the receipt of the acknowledgement, and reply to 503 the acknowledgement of the ECN-marked SYN/ACK packet by waiting T 504 seconds before sending a data packet. 506 However, we note that for an ECN-marked SYN/ACK packet, halving the 507 *congestion window* is not the same as halving the *sending rate*; 508 there is no `sending rate' associated with an ECN-Capable SYN/ACK 509 packet, as such packets are only sent as the first packet in a 510 connection from that host. Further, a router's marking of a SYN/ACK 511 packet is not affected by any past history of that connection. 513 Adding ECN-Capability to SYN/ACK packets allows the simple response 514 of setting the initial congestion window to one packet, instead of 515 its allowed default value of two, three, or four packets, with the 516 host proceeding with a cautious sending rate of one packet per round- 517 trip time. If that packet is ECN-marked or dropped, then the sender 518 will wait an RTO before sending another packet. This document argues 519 that this approach is useful to users, with no dangers of congestion 520 collapse or of starvation of competing traffic. This is discussed in 521 more detail below in Section 6.2. 523 We note that if the data transfer is entirely from Node A to Node B, 524 then there is no effective difference between the two possible 525 responses to an ECN-marked SYN/ACK packet outlined above. In either 526 case, Node B sends no data packets, only sending acknowledgement 527 packets in response to received data packets. 529 5. Related Work 531 The addition of ECN-capability to TCP's SYN/ACK packets was proposed 532 in [ECN+]. The paper includes an extensive set of simulation and 533 testbed experiments to evaluate the effects of the proposal, using 534 several Active Queue Management (AQM) mechanisms, including Random 535 Early Detection (RED) [RED], Random Exponential Marking (REM) [REM], 536 and Proportional Integrator (PI) [PI]. The performance measures were 537 the end-to-end response times for each request/response pair, and the 538 aggregate throughput on the bottleneck link. The end-to-end response 539 time was computed as the time from the moment when the request for 540 the file is sent to the server, until that file is successfully 541 downloaded by the client. 543 The measurements from [ECN+] show that setting an ECN-Capable 544 codepoint in the IP packet header in TCP SYN/ACK packets 545 systematically improves performance with all evaluated AQM schemes. 546 When SYN/ACK packets at a congested router are ECN-marked instead of 547 dropped, this can avoid a long initial retransmit timeout, improving 548 the response time for the affected flow dramatically. 550 [ECN+] shows that the impact on aggregate throughput can also be 551 quite significant, because marking SYN ACK packets can prevent larger 552 flows from suffering long timeouts before being "admitted" into the 553 network. In addition, the testbed measurements from [ECN+] show that 554 web servers setting the ECN-Capable codepoint in TCP SYN/ACK packets 555 could serve more requests. 557 As a final step, [ECN+] explores the co-existence of flows that do 558 and don't set the ECN-capable codepoint in TCP SYN/ACK packets. The 559 results in [ECN+] show that both types of flows can coexist, with 560 some performance degradation for flows that don't use ECN+. Flows 561 that do use ECN+ improve their end-to-end performance. At the same 562 time, the performance degradation for flows that don't use ECN+, as a 563 result of the flows that do use ECN+, increases as a greater fraction 564 of flows use ECN+. 566 6. Performance Evaluation 568 6.1. The Costs and Benefit of Adding ECN-Capability 570 [ECN+] explores the costs and benefits of adding ECN-Capability to 571 SYN/ACK packets with both simulations and experiments. The addition 572 of ECN-capability to SYN/ACK packets could be of significant benefit 573 for those ECN connections that would have had the SYN/ACK packet 574 dropped in the network, and for which the ECN-Capability would allow 575 the SYN/ACK to be marked rather than dropped. 577 The percent of SYN/ACK packets on a link can be quite high. In 578 particular, measurements on links dominated by web traffic indicate 579 that 15-20% of the packets can be SYN/ACK packets [SCJO01]. 581 The benefit of adding ECN-capability to SYN/ACK packets depends in 582 part on the size of the data transfer. The drop of a SYN/ACK packet 583 can increase the download time of a short file by an order of 584 magnitude, by requiring a three-second retransmit timeout. For 585 longer-lived flows, the effect of a dropped SYN/ACK packet on file 586 download time is less dramatic. However, even for longer-lived 587 flows, the addition of ECN-capability to SYN/ACK packets can improve 588 the fairness among long-lived flows, as newly-arriving flows would be 589 less likely to have to wait for retransmit timeouts. 591 One question that arises is what fraction of connections would see 592 the benefit from making SYN/ACK packets ECN-capable, in a particular 593 scenario. Specifically: 595 (1) What fraction of arriving SYN/ACK packets are dropped at the 596 congested router when the SYN/ACK packets are not ECN-capable? 597 (2) Of those SYN/ACK packets that are dropped, what fraction would 598 have been ECN-marked instead of dropped if the SYN/ACK packets had 599 been ECN-capable? 601 To answer (1), it is necessary to consider not only the level of 602 congestion but also the queue architecture at the congested link. As 603 described in Section 4 above, for some queue architectures small 604 packets are less likely to be dropped than large ones. In such an 605 environment, SYN/ACK packets would have lower packet drop rates; 606 question (1) could not necessarily be inferred from the overall 607 packet drop rate, but could be answered by measuring the drop rate 608 for SYN/ACK packets directly. In such an environment, adding ECN- 609 capability to SYN/ACK packets would be of less dramatic benefit than 610 in environments where all packets are equally likely to be dropped 611 regardless of packet size. 613 As question (2) implies, even if all of the SYN/ACK packets were ECN- 614 capable, there could still be some SYN/ACK packets dropped instead of 615 marked at the congested link; the full answer to question (2) depends 616 on the details of the queue management mechanism at the router. If 617 congestion is sufficiently bad, and the queue management mechanism 618 cannot prevent the buffer from overflowing, then SYN/ACK packets will 619 be dropped rather than marked upon buffer overflow whether or not 620 they are ECN-capable. 622 For some AQM mechanisms, ECN-capable packets are marked instead of 623 dropped any time this is possible, that is, any time the buffer is 624 not yet full. For other AQM mechanisms however, such as the RED 625 mechanism as recommended in [RED], packets are dropped rather than 626 marked when the packet drop/mark rate exceeds a certain threshold, 627 e.g., 10%, even if the packets are ECN-capable. For a router with 628 such an AQM mechanism, when congestion is sufficiently severe to 629 cause a high drop/mark rate, some SYN/ACK packets would be dropped 630 instead of marked whether or not they were ECN-capable. 632 Thus, the degree of benefit of adding ECN-Capability to SYN/ACK 633 packets depends not only on the overall packet drop rate in the 634 network, but also on the queue management architecture at the 635 congested link. 637 6.2. An Evaluation of Different Responses to ECN-Marked SYN/ACK Packets 639 This document specifies that the end-node responds to the report of 640 an ECN-marked SYN/ACK packet by setting the initial congestion window 641 to one segment, instead of its possible default value of two to four 642 segments. We call this ECN+ with NoWaiting. However, in Section 4 643 discussed another possible response to an ECN-marked SYN/ACK packet, 644 of the end-node waiting an RTT before sending a data packet. We call 645 this approach ECN+ with Waiting. 647 Simulations comparing the performance with Standard ECN (without ECN- 648 marked SYN/ACK packets), ECN+ with NoWaiting, and ECN+ with Waiting 649 show little difference, in terms of aggregate congestion, between 650 ECN+ with NoWaiting and ECN+ with Waiting. The details are given in 651 Appendix A below. Our conclusions are that ECN+ with NoWaiting is 652 perfectly safe, and there are no congestion-related reasons for 653 preferring ECN+ with Waiting over ECN+ with NoWaiting. That is, 654 there is no need for the TCP end-node to wait a round-trip time 655 before sending a data packet after receiving an acknowledgement of an 656 ECN-marked SYN/ACK packet. 658 7. Security Considerations 660 TCP packets carrying the ECT codepoint in IP headers can be marked 661 rather than dropped by ECN-capable routers. This raises several 662 security concerns that we discuss below. 664 "Bad" routers or middleboxes: 665 There are a number of known deployment problems from using ECN with 666 TCP traffic in the Internet. The first reported problem, dating back 667 to 2000, is of a small but decreasing number of routers or 668 middleboxes that reset a TCP connection in response to TCP SYN 669 packets using flags in the TCP header to negotiate ECN-capability 670 [Kelson00] [RFC3360] [MAF05]. Dave Thaler reported at the March 2007 671 IETF of new two problems encountered by TCP connections using ECN; 672 the first of the two problems concerns routers that crash when a TCP 673 data packet arrives with the ECN field in the IP header with the 674 codepoint ECT(0) or ECT(1), indicating that an ECN-Capable connection 675 has been established [SBT07]. 677 While there is no evidence that any routers or middleboxes drop 678 SYN/ACK packets that contain an ECN-Capable or CE codepoint in the IP 679 header, such behavior cannot be excluded. (There seems to be a 680 number of routers or middleboxes that drop TCP SYN packets that 681 contain known or unknown IP options [MAF05] (Figure 1).) Thus, as 682 specified in Section 3, if a SYN/ACK packet with the ECT or CE 683 codepoint is dropped, the TCP node SHOULD resend the SYN/ACK packet 684 without the ECN-Capable codepoint. There is also no evidence that 685 any routers or middleboxes crash when a SYN/ACK arrives with an ECN- 686 Capable or CE codepoint in the IP header (over and above the routers 687 already known to crash when a data packet arrives with either ECT(0) 688 or ECT(1)), but we have not conducted any measurement studies of this 689 [F07]. 691 Congestion collapse: 692 Because TCP SYN/ACK packets carrying an ECT codepoint could be ECN- 693 marked instead of dropped at an ECN-capable router, the concern is 694 whether this can either invoke congestion, or worsen performance in 695 highly congested scenarios. However, after learning that a SYN/ACK 696 packet was ECN-marked, the sender of that packet will only send one 697 data packet; if this data packet is ECN-marked, the sender will then 698 wait for a retransmission timeout. In addition, routers are free to 699 drop rather than mark arriving packets in times of high congestion, 700 regardless of whether the packets are ECN-capable. When congestion 701 is very high and a router's buffer is full, the router has no choice 702 but to drop rather than to mark an arriving packet. 704 The simulations reported in Appendix A show that even with demanding 705 traffic mixes dominated by short flows and high levels of congestion, 706 the aggregate packet dropping rates are not significantly different 707 with Standard ECN, ECN+ with NoWaiting, or ECN+ with Waiting. In 708 particular, the simulations show that in periods of very high 709 congestion the packet-marking rate is low with or without ECN+, and 710 the use of ECN+ does not significantly increase the number of dropped 711 or marked packets. 713 The simulations show that ECN+ is most effective in times of moderate 714 congestion. In these moderate-congested scenarios, the use of ECN+ 715 increases the number of ECN-marked packets, because ECN+ allows 716 SYN/ACK packets to be ECN-marked. At the same time, in these times 717 of moderate congestion, the use of ECN+ instead of Standard ECN does 718 not significantly affect the overall levels of congestion. 720 The simulations show that the use of ECN+ is less effective in times 721 of high congestion; the simulations show that in times of high 722 congestion more packets are dropped instead of marked, both with 723 Standard ECN and with ECN+. In times of high congestion, the buffer 724 can overflow, even with Active Queue Management and ECN; when the 725 buffer is full arriving packets are dropped rather than marked, 726 whether the packets are ECN-capable or not. Thus while ECN+ is less 727 effective in times of high congestion, it still doesn't result in a 728 significant increase in the level of congestion. More details are 729 given in the appendix. 731 8. Conclusions 733 This draft specifies a modification to RFC 3168 to allow TCP nodes to 734 send SYN/ACK packets as being ECN-Capable. Making the SYN/ACK packet 735 ECN-Capable avoids the high cost to a TCP transfer when a SYN/ACK 736 packet is dropped by a congested router, by avoiding the resulting 737 retransmit timeout. This improves the throughput of short 738 connections. The sender of the SYN/ACK packet responds to an ECN 739 mark by reducing its initial congestion window from two, three, or 740 four segments to one segment, reducing the subsequent load from that 741 connection on the network. The addition of ECN-capability to SYN/ACK 742 packets is particularly beneficial in the server-to-client direction, 743 where congestion is more likely to occur. In this case, the initial 744 information provided by the ECN marking in the SYN/ACK packet enables 745 the server to more appropriately adjust the initial load it places on 746 the network. 748 Future work will address the more general question of adding ECN- 749 Capability to relevant handshake packets in other protocols that use 750 retransmission-based reliability in their setup phase (e.g., SCTP, 751 DCCP, HIP, and the like). 753 9. Acknowledgements 755 We thank Anil Agarwal, Mark Allman, Wesley Eddy, Janardhan Iyengar, 756 and Pasi Sarolahti for feedback on earlier versions of this draft. 758 A. Report on Simulations 760 This section reports on simulations showing the costs of adding ECN+ 761 in highly-congested scenarios. This section also reports on 762 simulations for a comparative evaluation between ECN+ with NoWaiting 763 and ECN+ with Waiting. 765 The simulations are run with a range of file-size distributions. As 766 a baseline, they use the empirical heavy-tailed distribution reported 767 in [SCJO01], with a mean file size of around 7 KBytes. This flow- 768 size distribution is manipulated by skewing the flow sizes towards 769 lower and higher values to get distributions with mean file sizes of 770 3 KBytes, 5 KBytes, 14 KBytes and 17 KBytes. The congested link is 771 100 Mbps. RED is run in gentle mode, and arriving ECN-Capable 772 packets are only dropped instead of marked if the buffer is full (and 773 the router has no choice). 775 We explore two alternatives for a TCP node's response to a report of 776 an ECN-marked SYN/ACK packet. With ECN+ with NoWaiting, the TCP node 777 sends a data packet immediately (with an initial congestion window of 778 one segment). With the alternative ECN+ with Waiting, the TCP node 779 waits a round-trip time before sending a data packet; the sender 780 already has one measurement of the round-trip time when the 781 acknowledgement for the SYN/ACK packet is received. 783 In the tables below, ECN+ refers to ECN+ with NoWaiting, where the 784 sender starts transmitting immediately, and ECN+/wait refers to ECN+ 785 with Waiting, where the sender waits a round-trip time before sending 786 a data packet into the network. 788 The simulation scripts are available on [ECN-SYN], along with graphs 789 showing the distribution of response times for the TCP connections. 791 A.1. Simulations with RED in Packet Mode 793 The simulations with RED in packet mode and with the queue in packets 794 show that ECN+ is useful in times of moderate congestion, though it 795 adds little benefit in times of high congestion. The simulations 796 show a minimal increase in levels of congestion with either ECN+ with 797 Waiting or ECN+ with NoWaiting, either in terms of packet dropping or 798 marking rates or in terms of the distribution of responses times. 799 Thus, the simulations show no problems with ECN+ in times of high 800 congestion, and no reason to use ECN+ with Waiting instead of ECN+ 801 with NoWaiting. 803 Table 1 shows the congestion levels for simulations with RED in 804 packet mode, with a queue in packets. To explore a worst-case 805 scenario, these simulations use a traffic mix with an unrealistically 806 small flow size distribution, with a mean flow size of 3 Kbytes. For 807 each table showing a particular traffic load, the three rows show the 808 number of packets dropped, the number of packets ECN-marked, and the 809 aggregate packet drop rate, and the three columns show the 810 simulations with Standard ECN, ECN+ (NoWaiting) and ECN+/wait. 812 The usefulness of ECN+: The first thing to observe is that for the 813 simulations with the somewhat moderate load of 95%, with packet drop 814 rates of 5-6%, the use of ECN+ or ECN+/wait more than doubled the 815 number of packets marked. This indicates that with ECN+ or 816 ECN+/wait, many SYN/ACK packets are marked instead of dropped. 818 No increase in congestion: The second thing to observe is that in all 819 of the simulations, the use of ECN+ or ECN+/wait does not 820 significantly increase the aggregate packet drop rate. 822 Comparing ECN+ and ECN+/wait: The third thing to observe is that 823 there is little difference between ECN+ and ECN+/wait in terms of the 824 aggregate packet drop rate. Thus, there is no congestion-related 825 reason to prefer ECN+/wait over ECN+. 827 Traffic Load = 95%: 828 ECN ECN+ ECN+/wait 829 ------- ------- ------- 830 Dropped 74,645 64,034 64,983 831 Marked 7,639 17,681 16,914 832 Loss rate 6.05% 5.26% 5.33% 834 Traffic Load = 110%: 835 ECN ECN+ ECN+/wait 836 ------- ------- ------- 837 Dropped 161,644 163,620 165,196 838 Marked 4,375 6,653 6,144 839 Loss rate 10.38% 10.45% 10.53% 841 Traffic Load = 125%: 842 ECN ECN+ ECN+/wait 843 ------- ------- ------- 844 Dropped 257,671 268,161 264,437 845 Marked 2,885 3,712 3,359 846 Loss rate 14.52% 15.00% 14.83% 848 Traffic Load = 150%: 849 ECN ECN+ ECN+/wait 850 ------- ------- ------- 851 Loss rate 24.36% 24.61% 24.46% 853 Traffic Load = 200%: 854 ECN ECN+ ECN+/wait 855 ------- ------- ------- 856 Loss rate 29.99% 30.22% 30.23% 858 Table 1: Simulations with an average flow size of 3 Kbytes, RED in 859 packet mode, queue in packets. 861 A.2. Simulations with RED in Byte Mode 863 Table 3 below shows simulations with RED in byte mode and the queue 864 in bytes. Like the simulations with RED in packet mode, there is no 865 significant increase in aggregate congestion with the use of ECN+ or 866 ECN+/wait, and no congestion-related reason to prefer ECN+/wait over 867 ECN+. 869 However, unlike the simulations with RED in packet mode, the 870 simulations with RED in byte mode show little benefit from the use of 871 ECN+ or ECN+/wait, in that the packet marking rate with ECN+ or 872 ECN+/wait is not much different than the packet marking rate with 873 Standard ECN. This is because with RED in byte mode, small packets 874 like SYN/ACK packets are rarely dropped or marked - that is, there is 875 no drawback from the use of ECN+ in these scenarios, but not much 876 need for ECN+ either, in a scenario where small packets are unlikely 877 to be dropped or marked. 879 Traffic Load = 95%: 880 ECN ECN+ ECN+/wait 881 ------- ------- ------- 882 Dropped 13,044 13,323 14,855 883 Marked 18,880 19,175 19,049 884 Loss rate 1.13% 1.16% 1.29% 886 Traffic Load = 110%: 887 ECN ECN+ ECN+/wait 888 ------- ------- ------- 889 Dropped 84,809 83,013 83,564 890 Marked 4,086 4,644 4,826 891 Loss rate 5.90% 5.78% 5.81% 893 Traffic Load = 125%: 894 ECN ECN+ ECN+/wait 895 ------- ------- ------- 896 Dropped 157,305 157,435 158,368 897 Marked 2,183 2,363 2,663 898 Loss rate 9.89% 9.87% 9.93% 900 Table 3: Simulations with an average flow size of 3 Kbytes, RED in 901 byte mode, queue in bytes. 903 Normative References 905 [RFC 2119] S. Bradner, Key words for use in RFCs to Indicate 906 Requirement Levels, RFC 2119, March 1997. 908 [RFC3168] K.K. Ramakrishnan, S. Floyd, and D. Black, The Addition of 909 Explicit Congestion Notification (ECN) to IP, RFC 3168, Proposed 910 Standard, September 2001. 912 Informative References 914 [ECN+] A. Kuzmanovic, The Power of Explicit Congestion Notification, 915 SIGCOMM 2005. 917 [ECN-SYN] ECN-SYN web page with simulation scripts, URL to be added. 919 [F07] S. Floyd, "[BEHAVE] Response of firewalls and middleboxes to 920 TCP SYN packets that are ECN-Capable?", August 2, 2007, email sent to 921 the BEHAVE mailing list, URL "http://www1.ietf.org/mail- 922 archive/web/behave/current/msg02644.html".` 924 [Kelson00] Dax Kelson, note sent to the Linux kernel mailing list, 925 September 10, 2000. 927 [MAF05] A. Medina, M. Allman, and S. Floyd. Measuring the Evolution 928 of Transport Protocols in the Internet, ACM CCR, April 2005. 930 [PI] C. Hollot, V. Misra, W. Gong, and D. Towsley, On Designing 931 Improved Controllers for AQM Routers Supporting TCP Flows, April 932 1998. 934 [RED] Floyd, S., and Jacobson, V. Random Early Detection gateways 935 for Congestion Avoidance . IEEE/ACM Transactions on Networking, V.1 936 N.4, August 1993. 938 [REM] S. Athuraliya, V. H. Li, S. H. Low and Q. Yin, REM: Active 939 Queue Management, IEEE Network, May 2001. 941 [RFC2309] B. Braden et al., Recommendations on Queue Management and 942 Congestion Avoidance in the Internet, RFC 2309, April 1998. 944 [RFC2581] M. Allman, V. Paxson, and W. Stevens, TCP Congestion 945 Control, RFC 2581, April 1999. 947 [RFC2988] V. Paxson and M. Allman, Computing TCP's Retransmission 948 Timer, RFC 2988, November 2000. 950 [RFC3042] M. Allman, H. Balakrishnan, and S. Floyd, Enhancing TCP's 951 Loss Recovery Using Limited Transmit, RFC 3042, Proposed Standard, 952 January 2001. 954 [RFC3360] S. Floyd, Inappropriate TCP Resets Considered Harmful, RFC 955 3360, August 2002. 957 [RFC3390] M. Allman, S. Floyd, and C. Partridge, Increasing TCP's 958 Initial Window, RFC 3390, October 2002. 960 [SCJO01] F. Smith, F. Campos, K. Jeffay, D. Ott, What {TCP/IP} 961 Protocol Headers Can Tell us about the Web, SIGMETRICS, June 2001. 963 [SYN-COOK] Dan J. Bernstein, SYN cookies, 1997, see also 964 966 [SBT07] M. Sridharan, D. Bansal, and D. Thaler, Implementation Report 967 on Experiences with Various TCP RFCs, Presentation in the TSVAREA, 968 IETF 68, March 2007. URL 969 "http://www3.ietf.org/proceedings/07mar/slides/tsvarea-3/sld6.htm". 971 [Tools] S. Floyd and E. Kohler, Tools for the Evaluation of 972 Simulation and Testbed Scenarios, Internet-draft draft-irtf-tmrg- 973 tools-04, work in progress, July 2007. 975 IANA Considerations 977 There are no IANA considerations regarding this document. 979 Authors' Addresses 981 Aleksandar Kuzmanovic 982 Phone: +1 (847) 467-5519 983 Northwestern University 984 Email: akuzma at northwestern.edu 985 URL: http://cs.northwestern.edu/~a 987 Amit Mondal 988 Northwestern University 989 Email: a-mondal at northwestern.edu 991 Sally Floyd 992 Phone: +1 (510) 666-2989 993 ICIR (ICSI Center for Internet Research) 994 Email: floyd@icir.org 995 URL: http://www.icir.org/floyd/ 997 K. K. Ramakrishnan 998 Phone: +1 (973) 360-8764 999 AT&T Labs Research 1000 Email: kkrama at research.att.com 1001 URL: http://www.research.att.com/info/kkrama 1003 Full Copyright Statement 1005 Copyright (C) The IETF Trust (2007). 1007 This document is subject to the rights, licenses and restrictions 1008 contained in BCP 78, and except as set forth therein, the authors 1009 retain all their rights. 1011 This document and the information contained herein are provided on an 1012 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1013 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 1014 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 1015 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 1016 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1017 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1019 Intellectual Property 1021 The IETF takes no position regarding the validity or scope of any 1022 Intellectual Property Rights or other rights that might be claimed to 1023 pertain to the implementation or use of the technology described in 1024 this document or the extent to which any license under such rights 1025 might or might not be available; nor does it represent that it has 1026 made any independent effort to identify any such rights. Information 1027 on the procedures with respect to rights in RFC documents can be 1028 found in BCP 78 and BCP 79. 1030 Copies of IPR disclosures made to the IETF Secretariat and any 1031 assurances of licenses to be made available, or the result of an 1032 attempt made to obtain a general license or permission for the use of 1033 such proprietary rights by implementers or users of this 1034 specification can be obtained from the IETF on-line IPR repository at 1035 http://www.ietf.org/ipr. 1037 The IETF invites any interested party to bring to its attention any 1038 copyrights, patents or patent applications, or other proprietary 1039 rights that may cover technology that may be required to implement 1040 this standard. Please address the information to the IETF at ietf- 1041 ipr@ietf.org.