idnits 2.17.1 draft-ietf-ice-trickle-00.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 : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) == There are 2 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 10 instances of lines with private range IPv4 addresses in the document. If these are generic example addresses, they should be changed to use any of the ranges defined in RFC 6890 (or successor): 192.0.2.x, 198.51.100.x or 203.0.113.x. -- The document has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (October 19, 2015) is 3110 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 normative reference: RFC 4566 (Obsoleted by RFC 8866) ** Obsolete normative reference: RFC 5245 (Obsoleted by RFC 8445, RFC 8839) == Outdated reference: A later version (-18) exists of draft-ietf-mmusic-trickle-ice-sip-03 -- Obsolete informational reference (is this intentional?): RFC 2543 (Obsoleted by RFC 3261, RFC 3262, RFC 3263, RFC 3264, RFC 3265) -- Obsolete informational reference (is this intentional?): RFC 3388 (Obsoleted by RFC 5888) -- Obsolete informational reference (is this intentional?): RFC 5389 (Obsoleted by RFC 8489) -- Obsolete informational reference (is this intentional?): RFC 5766 (Obsoleted by RFC 8656) Summary: 3 errors (**), 0 flaws (~~), 4 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group E. Ivov 3 Internet-Draft Jitsi 4 Intended status: Standards Track E. Rescorla 5 Expires: April 21, 2016 RTFM, Inc. 6 J. Uberti 7 Google 8 P. Saint-Andre 9 &yet 10 October 19, 2015 12 Trickle ICE: Incremental Provisioning of Candidates for the Interactive 13 Connectivity Establishment (ICE) Protocol 14 draft-ietf-ice-trickle-00 16 Abstract 18 This document describes an extension to the Interactive Connectivity 19 Establishment (ICE) protocol that allows ICE agents to send and 20 receive candidates incrementally rather than exchanging complete 21 lists. With such incremental provisioning, ICE agents can begin 22 connectivity checks while they are still gathering candidates and 23 considerably shorten the time necessary for ICE processing to 24 complete. This mechanism is called "trickle ICE". 26 Status of This Memo 28 This Internet-Draft is submitted in full conformance with the 29 provisions of BCP 78 and BCP 79. 31 Internet-Drafts are working documents of the Internet Engineering 32 Task Force (IETF). Note that other groups may also distribute 33 working documents as Internet-Drafts. The list of current Internet- 34 Drafts is at http://datatracker.ietf.org/drafts/current/. 36 Internet-Drafts are draft documents valid for a maximum of six months 37 and may be updated, replaced, or obsoleted by other documents at any 38 time. It is inappropriate to use Internet-Drafts as reference 39 material or to cite them other than as "work in progress." 41 This Internet-Draft will expire on April 21, 2016. 43 Copyright Notice 45 Copyright (c) 2015 IETF Trust and the persons identified as the 46 document authors. All rights reserved. 48 This document is subject to BCP 78 and the IETF Trust's Legal 49 Provisions Relating to IETF Documents 50 (http://trustee.ietf.org/license-info) in effect on the date of 51 publication of this document. Please review these documents 52 carefully, as they describe your rights and restrictions with respect 53 to this document. Code Components extracted from this document must 54 include Simplified BSD License text as described in Section 4.e of 55 the Trust Legal Provisions and are provided without warranty as 56 described in the Simplified BSD License. 58 Table of Contents 60 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 61 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 62 3. Determining Support for Trickle ICE . . . . . . . . . . . . . 5 63 4. Sending the Initial Offer . . . . . . . . . . . . . . . . . . 6 64 4.1. Encoding the SDP . . . . . . . . . . . . . . . . . . . . 7 65 5. Receiving the Initial Offer . . . . . . . . . . . . . . . . . 7 66 5.1. Sending the Initial Answer . . . . . . . . . . . . . . . 8 67 5.2. Forming Check Lists and Beginning Connectivity 68 Checks . . . . . . . . . . . . . . . . . . . . . . . . . 8 69 5.3. Encoding the SDP . . . . . . . . . . . . . . . . . . . . 9 70 6. Receiving the Initial Answer . . . . . . . . . . . . . . . . 9 71 7. Performing Connectivity Checks . . . . . . . . . . . . . . . 9 72 7.1. Scheduling Checks . . . . . . . . . . . . . . . . . . . . 9 73 7.2. Check List and Timer State Updates . . . . . . . . . . . 9 74 8. Discovering and Sending Additional Local Candidates . . . . . 10 75 8.1. Pairing Newly Learned Candidates and Updating 76 Check Lists . . . . . . . . . . . . . . . . . . . . . . . 12 77 8.2. Encoding the SDP for Additional Candidates . . . . . . . 13 78 8.3. Announcing End of Candidates . . . . . . . . . . . . . . 14 79 9. Receiving Additional Remote Candidates . . . . . . . . . . . 15 80 10. Receiving an End-Of-Candidates Notification . . . . . . . . . 16 81 11. Trickle ICE and Peer Reflexive Candidates . . . . . . . . . . 16 82 12. Concluding ICE Processing . . . . . . . . . . . . . . . . . . 16 83 13. Subsequent Offer/Answer Exchanges . . . . . . . . . . . . . . 16 84 14. Interaction with ICE Lite . . . . . . . . . . . . . . . . . . 17 85 15. Unilateral Use of Trickle ICE (Half Trickle) . . . . . . . . 18 86 16. Example Flow . . . . . . . . . . . . . . . . . . . . . . . . 19 87 17. Security Considerations . . . . . . . . . . . . . . . . . . . 20 88 18. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 20 89 19. References . . . . . . . . . . . . . . . . . . . . . . . . . 20 90 19.1. Normative References . . . . . . . . . . . . . . . . . . 20 91 19.2. Informative References . . . . . . . . . . . . . . . . . 21 92 Appendix A. Open Issues . . . . . . . . . . . . . . . . . . . . 22 93 A.1. MID/Stream Indices in SDP . . . . . . . . . . . . . . . . 22 94 A.2. Starting Checks . . . . . . . . . . . . . . . . . . . . . 23 95 A.3. Checklist States . . . . . . . . . . . . . . . . . . . . 23 96 A.4. Relationship to Continuous Nomination and 97 Passive Nomination . . . . . . . . . . . . . . . . . . . 23 98 A.5. ICE Restarts . . . . . . . . . . . . . . . . . . . . . . 23 99 A.6. Candidate Redundancy and Priority . . . . . . . . . . . . 23 100 A.7. Make Trickle ICE SDP-Agnostic . . . . . . . . . . . . . . 23 101 Appendix B. Interaction with ICE . . . . . . . . . . . . . . . . 23 102 Appendix C. Changes from Earlier Versions . . . . . . . . . . . 25 103 C.1. Changes from draft-mmusic-trickle-ice-02 . . . . . . . . 25 104 C.2. Changes from draft-ivov-01 and draft-mmusic-00 . . . . . 25 105 C.3. Changes from draft-ivov-00 . . . . . . . . . . . . . . . 26 106 C.4. Changes from draft-rescorla-01 . . . . . . . . . . . . . 26 107 C.5. Changes from draft-rescorla-00 . . . . . . . . . . . . . 27 108 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 27 110 1. Introduction 112 The Interactive Connectivity Establishment (ICE) protocol [RFC5245] 113 describes mechanisms for gathering candidates, prioritizing them, 114 choosing default ones, exchanging them with the remote party, pairing 115 them and ordering them into check lists. Once all of the above have 116 been completed, and only then, the participating agents can begin a 117 phase of connectivity checks and eventually select the pair of 118 candidates that will be used in the following session. 120 While the above sequence has the advantage of being relatively 121 straightforward to implement and debug once deployed, it may also 122 prove to be rather lengthy. Gathering candidates or candidate 123 gathering often involves things like querying STUN [RFC5389] servers, 124 discovering UPnP devices, and allocating relayed candidates at TURN 125 [RFC5766] servers. All of these can be delayed for a noticeable 126 amount of time and while they can be run in parallel, they still need 127 to respect the pacing requirements from [RFC5245], which is likely to 128 delay them even further. Some or all of the above would also have to 129 be completed by the remote agent. Both agents would next perform 130 connectivity checks and only then would they be ready to begin 131 streaming media. 133 All of the above can lead to relatively lengthy session establishment 134 times and degraded user experience. 136 The purpose of this document is to define an alternative mode of 137 operation for ICE implementations, also known as "trickle ICE", where 138 candidates can be exchanged incrementally. This would allow ICE 139 agents to exchange candidates as soon as a session has been 140 initiated. Connectivity checks for a media stream would also start 141 as soon as the first candidates for that stream have become 142 available. 144 Trickle ICE allows reducing session establishment times in cases 145 where connectivity is confirmed for the first exchanged candidates 146 (e.g. where the host candidates for one of the agents are directly 147 reachable from the second agent). Even when this is not the case, 148 running candidate gathering for both agents and connectivity checks 149 all in parallel allows to considerably reduce ICE processing times. 151 It is worth pointing out that before being introduced to the IETF, 152 trickle ICE had already been included in specifications such as XMPP 153 Jingle [XEP-0176] and it has been in use in various implementations 154 and deployments. 156 In addition to the basics of trickle ICE, this document also 157 describes how to discover support for trickle ICE, how regular ICE 158 processing needs to be modified when building and updating check 159 lists, and how trickle ICE implementations interoperate with agents 160 that only implement [RFC5245] processing. 162 This specification does not define usage of trickle ICE with any 163 specific signalling protocol, different from [RFC5245] which contains 164 a usage for ICE with SIP [RFC3261]. Such usages would have to be 165 specified in separate documents such as for example 166 [I-D.ietf-mmusic-trickle-ice-sip]. However, trickle ICE does however 167 reuse and build upon the SDP syntax defined by [RFC5245]. 169 Although this document mostly describes trickle ICE in terms of the 170 offer/answer model [RFC3264], trickle ICE (and ICE itself) can be 171 used by application protocols that do not follow the offer/answer 172 model. 174 2. Terminology 176 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 177 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 178 document are to be interpreted as described in [RFC2119]. 180 This specification makes use of all terminology defined by the 181 protocol for Interactive Connectivity Establishment in [RFC5245]. 183 Vanilla ICE: The Interactive Connectivity Establishment protocol as 184 defined in [RFC5245]. Through the rest of the text, the terms 185 vanilla ICE and "RFC5245" are used interchangeably. 187 Candidate Harvester: A module used by an ICE agent to obtain local 188 candidates. Candidate gatherers use different mechanisms for 189 discovering local candidates. Some of them would typically make 190 use of protocols such as STUN or TURN. Others may also employ 191 techniques that are not referenced within [RFC5245]. UPnP based 192 port allocation and XMPP Jingle Relay Nodes [XEP-0278] are among 193 the possible examples. 195 Trickled Candidates: Candidates that a trickle ICE agent is sending 196 subsequently to but within the context defined by an offer or an 197 answer. Trickled candidates can be sent in parallel with 198 candidate gathering and connectivity checks. 200 Trickling/Trickle (v.): The act of sending trickled candidates. 202 Half Trickle: A trickle ICE mode of operation where the offerer 203 gathers its first generation of candidates strictly before 204 creating and sending the offer. Once sent, that offer can be 205 processed by vanilla ICE agents and does not require support for 206 this specification. It also allows trickle ICE capable answerers 207 to still gather candidates and perform connectivity checks in a 208 non-blocking way, thus roughly offering "half" the advantages of 209 trickle ICE. The mechanism is mostly meant for use in cases where 210 support for trickle ICE cannot be confirmed prior to sending a 211 initial offer. 213 Full Trickle: Regular mode of operation for trickle ICE agents, used 214 in opposition to the half trickle mode of operation. 216 3. Determining Support for Trickle ICE 218 According to [RFC5245], supported features are to be advertised in 219 the ice-options attribute. Therefore an agent supporting trickle ICE 220 MUST include a token of "trickle" in the ice-options attribute every 221 time it generates an offer or an answer. Syntax for this token is 222 defined in Section 4.1. 224 Agents that receive offers or answers can verify support by examining 225 them for the "trickle" ice-options token. However, agents that are 226 about to send an initial offer have no way of doing this. Thus 227 usages of trickle for specific protocols need to either: 229 o Provide a way for agents to verify support of trickle ICE prior to 230 initiating a session (XMPP's Service Discovery [XEP-0030] is an 231 example of one such mechanism); or 233 o Make support for trickle ICE mandatory so that support could be 234 assumed the agents. 236 Alternately, for cases where a protocol provides neither of the 237 above, agents may either rely on provisioning/configuration, or use 238 the half trickle procedure described in Section 15. 240 Prior to sending an initial offer, agents using signaling protocols 241 that support capabilities discovery MAY attempt to verify whether or 242 not the remote party supports trickle ICE. If an agent determines 243 that the remote party does not support trickle ICE, it MUST fall back 244 to using vanilla ICE or abandon the entire session. 246 All trickle ICE offers and answers MUST indicate support of this 247 specification, as explained in Section 4.1. 249 Note that out-of-band discovery semantics and half trickle are only 250 necessary prior to session initiation, or in other words, when 251 sending the initial offer. Once a session is established and trickle 252 ICE support is confirmed for both parties, either agent can use full 253 trickle for subsequent offers. 255 4. Sending the Initial Offer 257 An agent starts gathering candidates as soon as it has an indication 258 that communication is imminent (e.g. a user interface cue or an 259 explicit request to initiate a session). Contrary to vanilla ICE, 260 implementations of trickle ICE do not need to gather candidates in a 261 blocking manner. Therefore, unless half trickle is being used, 262 agents SHOULD generate and transmit their initial offer as early as 263 possible, in order to allow the remote party to start gathering and 264 trickling candidates. 266 Trickle ICE agents MAY include any set of candidates in an offer. 267 This includes the possibility of generating one with no candidates, 268 or one that contains all the candidates that the agent is planning on 269 using in the following session. 271 For optimal performance, it is RECOMMENDED that the candidates in an 272 initial offer (if any) be host candidates only. This would allow 273 both agents to start gathering server reflexive, relayed and other 274 non-host candidates simultaneously, and it would also enable them to 275 begin connectivity checks. 277 If the privacy implications of revealing host addresses are a 278 concern, agents MAY generate an offer that contains no candidates and 279 then only trickle candidates that do not reveal host addresses (e.g. 280 relayed candidates). 282 Methods for calculating priorities and foundations, as well as 283 determining redundancy of candidates, work just as with vanilla ICE. 285 4.1. Encoding the SDP 287 The process of encoding the SDP [RFC4566] is mostly the same as the 288 one used by vanilla ICE. Still, trickle ICE does require a few 289 differences described here. 291 Agents MUST indicate support for Trickle ICE by including the 292 "trickle" token for the "a=ice-options" attribute: 294 a=ice-options:trickle 296 As mentioned earlier in this section, offers and answers can contain 297 any set of candidates, which means that a trickle ICE session 298 description MAY contain no candidates at all. Doing so enables the 299 offerer to receive the answerer's initial candidate list sooner, and 300 also enables the answerer to begin candidate gathering more quickly. 301 In such cases the agent would still need to place an address in the 302 "c=" line(s). If the use of a host address there is undesirable 303 (e.g., for privacy reasons), the agent MAY set the connection address 304 to 0.0.0.0. In this case it MUST also set the port number to 9 305 (Discard). There is no need to include a fictitious candidate for 306 the 0.0.0.0 address when doing so. 308 It is worth noting that the use of IP6 :: has been selected over IP4 309 0.0.0.0, even though [RFC3264] already gives the latter semantics 310 appropriate for such use. The reason for this choice is the historic 311 use of 0.0.0.0 as a means of putting a stream on hold [RFC2543] and 312 the ambiguity that this may cause with legacy libraries and 313 applications. 315 It is also worth mentioning that use of IP6 :: here does not 316 constitute any kind of indication as to the actual use of IPv6 317 candidates in a session and it can very well appear in a negotiation 318 that only involves IPv4 candidates. 320 5. Receiving the Initial Offer 322 When an agent receives an initial offer, it will first check if it 323 indicates support for trickle ICE as explained in Section 3. If this 324 is not the case, the agent MUST process the offer according to the 325 [RFC5245] procedures or standard [RFC3264] processing in case no ICE 326 support is detected at all. 328 It is worth pointing out that in case support for trickle ICE is 329 confirmed, an agent will automatically assume support for vanilla ICE 330 as well even if the support verification procedure in [RFC5245] 331 indicates otherwise. Specifically, the rules from RFC 5245 would 332 imply that ICE itself is not supported if the initial offer includes 333 no candidates in the offer; however, such a conclusion is not 334 warranted if the answerer can confirm that the offerer supports 335 trickle ICE. In this case, the IP6 :: address present in the "c=" 336 line would not "appear in a candidate attribute". Fallback to 337 [RFC3264] is not necessary in this scenario. 339 If, the offer does indicate support for trickle ICE, the agent will 340 determine its role, start gathering and prioritizing candidates and, 341 while doing so it will also respond by sending its own answer, so 342 that both agents can start forming check lists and begin connectivity 343 checks. 345 5.1. Sending the Initial Answer 347 An agent can respond to an initial offer at any point while gathering 348 candidates. The answer can again contain any set of candidates, 349 including all candidates or no candidates. (The benefit of including 350 no candidates is to send the answer as quickly as possible, so that 351 both parties can consider the overall session to be under active 352 negotiation as soon as possible.) Unless the answering agent is 353 protecting host addresses for privacy reasons, it would typically 354 construct this initial answer including only them, thus allowing the 355 remote party to also start forming checklists and performing 356 connectivity checks. 358 The answer MUST indicate support for trickle ICE as described by 359 Section 3. 361 5.2. Forming Check Lists and Beginning Connectivity Checks 363 After exchanging offer and answer, and as soon as they have obtained 364 local and remote candidates, agents will begin forming candidate 365 pairs, computing their priorities and creating check lists according 366 to the vanilla ICE procedures described in [RFC5245]. Obviously in 367 order for candidate pairing to be possible, candidates would need to 368 be provided in both the offer and the answer. If not, then the 369 agents will still create the check lists (so that their Active/Frozen 370 state could be monitored and updated) but they will only populate the 371 check lists once they actually have the candidate pairs. 373 Initially, all check lists will have their Active/Frozen state set to 374 Frozen. 376 Trickle ICE agents will then inspect the first check list and attempt 377 to unfreeze all candidates belonging to the first component on the 378 first media stream (i.e. the first media stream that was reported to 379 the ICE implementation from the using application). If this 380 checklist is still empty however, agents will hold off further 381 processing until this is no longer the case. 383 Respecting the order in which lists have been reported to an ICE 384 implementation, or in other words, the order in which they appear in 385 SDP, is crucial to the frozen candidates algorithm and important when 386 making sure that connectivity checks are performed simultaneously by 387 both agents. 389 5.3. Encoding the SDP 391 The process for encoding the SDP at the answerer is identical to the 392 process followed by the offerer for both full and lite 393 implementations, as described in Section 4.1. 395 6. Receiving the Initial Answer 397 When receiving an answer, agents will follow vanilla ICE procedures 398 to determine their role and they would then form check lists (as 399 described in Section 5.2) and begin connectivity checks . 401 7. Performing Connectivity Checks 403 For the most part, trickle ICE agents perform connectivity checks 404 following vanilla ICE procedures. Of course, the asynchronous nature 405 of gathering and communicating candidates in trickle ICE would impose 406 a number of changes described here. 408 7.1. Scheduling Checks 410 The ICE specification [RFC5245], Section 5.8, requires that agents 411 terminate the timer for a triggered check in relation to an active 412 check list once the agent has exhausted all frozen pairs in check 413 list. This will not work with trickle ICE, because more pairs will 414 be added to the check list incrementally. 416 Therefore, a trickle ICE agent SHOULD NOT terminate the timer until 417 the state of the check list is completed or failed as specified 418 herein (see Section 8.3). 420 7.2. Check List and Timer State Updates 422 The ICE specification [RFC5245], Section 7.1.3.3, requires that 423 agents update check lists and timer states upon completing a 424 connectivity check transaction. During such an update vanilla ICE 425 agents would set the state of a check list to Failed if both of the 426 following two conditions are satisfied: 428 o all of the pairs in the check list are either in the Failed or 429 Succeeded state; and 431 o there is not a pair in the valid list for each component of the 432 media stream. 434 With trickle ICE, the above situation would often occur when 435 candidate gathering and trickling are still in progress, even though 436 it is quite possible that future checks will succeed. For this 437 reason trickle ICE agents add the following conditions to the above 438 list: 440 o all candidate gatherers have completed and the agent is not 441 expecting to discover any new local candidates; 443 o the remote agent has sent an end-of-candidates indication for that 444 check list as described in Section 8.3. 446 Vanilla ICE requires that agents then update all other check lists, 447 placing one pair in each of them into the Waiting state, effectively 448 unfreezing all remaining check lists. Given that with trickle ICE, 449 other check lists may still be empty at that point, a trickle ICE 450 agent SHOULD also maintain an explicit Active/Frozen state for every 451 check list, rather than deducing it from the state of the pairs it 452 contains. This state should be set to Active when unfreezing the 453 first pair in a list or when that couldn't happen because a list was 454 empty. 456 8. Discovering and Sending Additional Local Candidates 458 After an offer or an answer have been sent, agents will most likely 459 continue discovering new local candidates as STUN, TURN and other 460 non-host candidate gathering mechanisms begin to yield results. 461 Whenever an agent discovers such a new candidate it will compute its 462 priority, type, foundation and component id according to normal 463 vanilla ICE procedures. 465 The new candidate is then checked for redundancy against the existing 466 list of local candidates. If its transport address and base match 467 those of an existing candidate, it will be considered redundant and 468 will be ignored. This would often happen for server reflexive 469 candidates that match the host addresses they were obtained from 470 (e.g. when the latter are public IPv4 addresses). Contrary to 471 vanilla ICE, trickle ICE agents will consider the new candidate 472 redundant regardless of its priority. 474 Next the client sends (i.e. trickles) the newly learnt candidate(s) 475 to the remote agent. The actual delivery of the new candidates will 476 be specified by using protocols such as SIP. Trickle ICE imposes no 477 restrictions on the way this is done or whether it is done at all. 478 For example, some applications may choose not to send trickle updates 479 for server reflexive candidates and rely on the discovery of peer 480 reflexive ones instead. 482 When trickle updates are sent however, each candidate MUST be 483 delivered to the receiving Trickle ICE implementation not more than 484 once and in the same order that they were sent. In other words, if 485 there are any candidate retransmissions, they must be hidden from the 486 ICE implementation. 488 Also, candidate trickling needs to be correlated to a specific ICE 489 negotiation session, so that if there is an ICE restart, any delayed 490 updates for a previous session can be recognized as such and ignored 491 by the receiving party. 493 One important aspect of Vanilla ICE is that connectivity checks for a 494 specific foundation and component be attempted simultaneously by both 495 agents, so that any firewalls or NATs fronting the agents would 496 whitelist both endpoints and allow all except for the first (suicide) 497 packets to go through. This is also crucial to unfreezing candidates 498 in the right time. 500 In order to preserve this feature here, when trickling candidates 501 agents MUST respect the order of the components as they appear 502 (implicitly or explicitly) in the Offer/Answer descriptions. 503 Therefore a candidate for a specific component MUST NOT be sent prior 504 to candidates for other components within the same foundation. 506 For example, the following session description contains two 507 components (RTP and RTCP), and two foundations (host and the server 508 reflexive): 510 v=0 511 o=jdoe 2890844526 2890842807 IN IP4 10.0.1.1 512 s= 513 c=IN IP4 10.0.1.1 514 t=0 0 515 a=ice-pwd:asd88fgpdd777uzjYhagZg 516 a=ice-ufrag:8hhY 517 m=audio 5000 RTP/AVP 0 518 a=rtpmap:0 PCMU/8000 519 a=candidate:1 1 UDP 2130706431 10.0.1.1 5000 typ host 520 a=candidate:1 2 UDP 2130706431 10.0.1.1 5001 typ host 521 a=candidate:2 1 UDP 1694498815 192.0.2.3 5000 typ srflx 522 raddr 10.0.1.1 rport 8998 523 a=candidate:2 2 UDP 1694498815 192.0.2.3 5001 typ srflx 524 raddr 10.0.1.1 rport 8998 526 For this description the RTCP host candidate MUST NOT be sent prior 527 to the RTP host candidate. Similarly the RTP server reflexive 528 candidate MUST be sent together with or prior to the RTCP server 529 reflexive candidate. 531 Note that the order restriction only applies among candidates that 532 belong to the same foundation. 534 It is also equally important to preserve this order across media 535 streams and this is covered by the requirement to always start 536 unfreezing candidates starting from the first media stream 537 Section 5.2. 539 Once the candidate has been sent to the remote party, the agent 540 checks if any remote candidates are currently known for this same 541 stream. If this is not the case the new candidate will simply be 542 added to the list of local candidates. 544 Otherwise, if the agent has already learned of one or more remote 545 candidates for this stream and component, it will begin pairing the 546 new local candidates with them and adding the pairs to the existing 547 check lists according to their priority. 549 8.1. Pairing Newly Learned Candidates and Updating Check Lists 551 Forming candidate pairs will work the way it is described by the ICE 552 specification [RFC5245]. Actually adding the new pair to a check 553 list however, will happen according to the rules described below. 555 If the check list where the pair is to be added already contains the 556 maximum number of candidate pairs (100 by default as per [RFC5245]), 557 the new pair is discarded. 559 If the new pair's local candidate is server reflexive, the server 560 reflexive candidate MUST be replaced by its base before adding the 561 pair to the list. Once this is done, the agent examines the check 562 list looking for another pair that would be redundant with the new 563 one. If such a pair exists, the newly formed pair is ignored. 565 For all other pairs, including those with a server reflexive local 566 candidate that were not found to be redundant: 568 o if this check list is Frozen then the new pair will also be 569 assigned a Frozen state. 571 o else if the check list is Active and it is either empty or 572 contains only candidates in the Succeeded and Failed states, then 573 the new pair's state is set to Waiting. 575 o else if the check list is non-empty and Active, then the new pair 576 state will be set to 578 Frozen: if there is at least one pair in the list whose 579 foundation matches the one in the new pair and whose state is 580 neither Succeeded nor Failed (eventually the new pair will get 581 unfrozen after the the on-going check for the existing pair 582 concludes); 584 Waiting: if the list contains no pairs with the same foundation 585 as the new one, or, in case such pairs exist but they are all 586 in either the Succeeded or Failed states. 588 8.2. Encoding the SDP for Additional Candidates 590 To facilitate interoperability an ICE agent will encode additional 591 candidates using the vanilla ICE SDP syntax. For example: 593 a=candidate:2 1 UDP 1658497328 198.51.100.33 5000 typ host 595 Given that such lines do not provide a relationship between the 596 candidate and the m line that it relates to, signalling protocols 597 using trickle ICE MUST establish that relation themselves using an 598 MID [RFC3388]. Such MIDs use "media stream identification", as 599 defined in [RFC3388], to identify a corresponding m-line. When 600 creating candidate lines usages of trickle ICE MUST use the MID if 601 possible, or the m-line index if not. Obviously, agents MUST NOT 602 send individual candidates prior to generating the corresponding SDP 603 session description. 605 The exact means of transporting additional candidates to a remote 606 agent is left to the protocols using trickle ICE. It is important to 607 note, however, that these candidate exchanges are not part of the 608 offer/answer model. 610 8.3. Announcing End of Candidates 612 Once all candidate gatherers for a specific media stream complete, or 613 expire, the agents will generate an "end-of-candidates" indication 614 for that stream and send it to the remote agent via the signalling 615 channel. Such indications are sent in the form of a media-level 616 attribute that has the following form: end-of-candidates. 618 a=end-of-candidates 620 The end-of-candidates indications can be sent in the following ways: 622 o As part of an offer (which would typically be the case with half 623 trickle initial offers) 625 o Along with the last candidate an agent can send for a stream 627 o As a standalone notification (e.g., after STUN Binding requests or 628 TURN Allocate requests to a server timeout and the agent has no 629 other active gatherers) 631 Controlled trickle ICE agents SHOULD always send end-of-candidates 632 indications once gathering for a media stream has completed unless 633 ICE processing terminates before they've had a chance to do so. 634 Sending the indication is necessary in order to avoid ambiguities and 635 speed up ICE conclusion. Controlling agents on the other hand MAY 636 sometimes conclude ICE processing prior to sending end-of-candidates 637 notifications for all streams. This would typically be the case with 638 aggressive nomination. Yet it is RECOMMENDED that controlling agents 639 do send such indications whenever possible for the sake of 640 consistency and keeping middle boxes and controlled agents up-to-date 641 on the state of ICE processing. 643 When sending end-of-candidates during trickling, rather than as a 644 part of an offer or an answer, it is the responsibility of the using 645 protocol to define means that can be used to relate the indication to 646 one or more specific m-lines. 648 Receiving an end-of-candidates notification allows an agent to update 649 check list states and, in case valid pairs do not exist for every 650 component in every media stream, determine that ICE processing has 651 failed. It also allows agents to speed ICE conclusion in cases where 652 a candidate pair has been validates but it involves the use of lower- 653 preference transports such as TURN. In such situations some 654 implementations may choose to wait in case higher-priority candidates 655 are received and end-of-candidates provides an indication that this 656 is not going to happen. 658 An agent MAY also choose to generate an end-of-candidates event 659 before candidate gathering has actually completed, if the agent 660 determines that gathering has continued for more than an acceptable 661 period of time. However, an agent MUST NOT send any more candidates 662 after it has send an end-of-candidates notification. 664 When performing half trickle agents SHOULD send end-of-candidates 665 together with their initial offer unless they are planning on 666 potentially sending additional candidates in case the remote party 667 turns out to actually support trickle ICE. 669 When end-of-candidates is sent as part of an offer or an answer it 670 can appear as a session-level attribute, which would be equivalent to 671 having it appear in all m-lines. 673 Once an agent sends the end-of-candidates event, it will update the 674 state of the corresponding check list as explained in Section 7.2. 675 Past that point agents MUST NOT send any new candidates within this 676 ICE session. Once an agent has received an end-of-candidates 677 indication, it MUST also ignore any newly received candidates for 678 that media stream, and adding new candidates to the negotiation is 679 only possible through an ICE restart. 681 This specification does not override vanilla ICE semantics for 682 concluding ICE processing. Therefore even if end-of-candidates 683 indications are sent agents will still have to go through pair 684 nomination. Also, if pairs have been nominated for components and 685 media streams, ICE processing will still conclude even if end-of- 686 candidate indications have not been received for all streams. 688 9. Receiving Additional Remote Candidates 690 At any point of ICE processing, a trickle ICE agent may receive new 691 candidates from the remote agent. When this happens and no local 692 candidates are currently known for this same stream, the new remote 693 candidates are simply added to the list of remote candidates. 695 Otherwise, the new candidates are used for forming candidate pairs 696 with the pool of local candidates and they are added to the local 697 check lists as described in Section 8.1. 699 Once the remote agent has completed candidate gathering, it will send 700 an end-of-candidates event. Upon receiving such an event, the local 701 agent MUST update check list states as per Section 7.2. This may 702 lead to some check lists being marked as Failed. 704 10. Receiving an End-Of-Candidates Notification 706 When an agent receives an end-of-candidates notification for a 707 specific check list, they will update its state as per Section 7.2. 708 In case the list is still in the Active state after the update, the 709 agent will persist the the fact that an end-of-candidates 710 notification has been received for and take it into account in future 711 list updates. 713 11. Trickle ICE and Peer Reflexive Candidates 715 Even though Trickle ICE does not explicitly modify the procedures for 716 handling peer reflexive candidates, their processing could be 717 impacted in implementations. With Trickle ICE, it is possible that 718 server reflexive candidates be discovered as peer reflexive in cases 719 where incoming connectivity checks are received from these candidates 720 before the trickle updates that carry them. 722 While this would certainly increase the number of cases where ICE 723 processing nominates and selects candidates discovered as peer- 724 reflexive it does not require any change in processing. 726 It is also likely that, some applications would prefer not to trickle 727 server reflexive candidates to entities that are known to be publicly 728 accessible and where sending a direct STUN binding request is likely 729 to reach the destination faster than the trickle update that travels 730 through the signalling path. 732 12. Concluding ICE Processing 734 This specification does not directly modify the procedures ending ICE 735 processing described in Section 8 of [RFC5245], and trickle ICE 736 implementations will follow the same rules. 738 13. Subsequent Offer/Answer Exchanges 740 Either agent MAY generate a subsequent offer at any time allowed by 741 [RFC3264]. When this happens agents will use [RFC5245] semantics to 742 determine whether or not the new offer requires an ICE restart. If 743 this is the case then agents would perform trickle ICE as they would 744 in an initial offer/answer exchange. 746 The only differences between an ICE restart and a brand new media 747 session are that: 749 o during the restart, media can continue to be sent to the 750 previously validated pair. 752 o both agents are already aware whether or not their peer supports 753 trickle ICE, and there is no longer need for performing half 754 trickle or confirming support with other mechanisms. 756 14. Interaction with ICE Lite 758 Behaviour of Trickle ICE capable ICE lite agents does not require any 759 particular rules other than those already defined in this 760 specification and [RFC5245]. This section is hence added with an 761 informational purpose only. 763 A Trickle ICE capable ICE Lite agent would generate offers or answers 764 as per [RFC5245]. Both will indicate support for trickle ICE 765 (Section 4.1) and given that they will contain a complete set of 766 candidates (the agent's host candidates) these offers and answers 767 would also be accompanied with an end-of-candidates notification. 769 When performing full trickle, a full ICE implementation could send an 770 offer or an answer with no candidates and an IP6 :: connection line 771 address. After receiving an answer that identifies the remote agent 772 as an ICE lite implementation, the offerer may very well choose to 773 not send any additional candidates. The same is also true in the 774 case when the ICE lite agent is making the offer and the full ICE one 775 is answering. In these cases the connectivity checks would be enough 776 for the ICE lite implementation to discover all potentially useful 777 candidates as peer reflexive. The following example illustrates one 778 such ICE session: 780 ICE Lite Bob 781 Agent 782 | Offer (a=ice-lite a=ice-options:trickle) | 783 |---------------------------------------------->| 784 | |no cand 785 | Answer (a=ice-options:trickle) |trickling 786 |<----------------------------------------------| 787 | Connectivity Checks | 788 |<--------------------------------------------->| 789 peer rflx| | 790 cand disco| | 791 | | 792 |<=============== MEDIA FLOWS =================>| 794 Figure 1: Example 796 In addition to reducing signaling traffic this approach also removes 797 the need to discover STUN bindings, or to make TURN or UPnP 798 allocations which may considerably lighten ICE processing. 800 15. Unilateral Use of Trickle ICE (Half Trickle) 802 In half trickle mode, the offerer sends a regular, vanilla ICE offer, 803 with a complete set of candidates. This ensures that the offer can 804 be processed by a vanilla ICE answerer and is mostly meant for use in 805 cases where support for trickle ICE cannot be confirmed prior to 806 sending a initial offer. The initial offer indicates support for 807 trickle ICE, so that the answerer can respond with an incomplete set 808 of candidates and continue trickling the rest. Half trickle offers 809 typically contain an end-of-candidates indication, although this is 810 not mandatory because if trickle support is confirmed then the 811 offerer can choose to trickle additional candidates before it 812 declares end of trickling. 814 The half trickle mechanism can be used in cases where there is no way 815 for an agent to verify in advance whether a remote party supports 816 trickle ICE. Because it contains a full set of candidates, its 817 initial offer can thus be handled by a regular vanilla ICE agent, 818 while still allowing a trickle one to use the optimisation defined in 819 this specification. This prevents negotiation from failing in the 820 former case while still giving roughly half the trickle ICE benefits 821 in the latter (hence the name of the mechanism). 823 Use of half trickle is only necessary during an initial offer/answer 824 exchange. Once both parties have received a session description from 825 their peer, they can each reliably determine trickle ICE support and 826 use it for all subsequent offer/answer exchanges. 828 In some instances, using half trickle might bring more than just half 829 the improvement in terms of user experience. This can happen when an 830 agent starts gathering candidates upon user interface cues that the 831 user will soon be initiating an offer, such as activity on a keypad 832 or the phone going off hook. This would mean that some or all of the 833 candidate gathering could be completed before the agent actually 834 needs to send the offer. Because that the answerer will be able to 835 trickle candidates, both agents will be able to start connectivity 836 checks and complete ICE processing earlier than with vanilla ICE and 837 potentially even as early as with full trickle. 839 However, such anticipation is not not always possible. For example, 840 a multipurpose user agent or a WebRTC web page where communication is 841 a non-central feature (e.g., calling a support line in case of a 842 problem with the main features) would not necessarily have a way of 843 distinguishing between call intentions and other user activity. In 844 such cases, using full trickle is most likely to result in an ideal 845 user experience. Even so, using half trickle would be an improvement 846 over vanilla ICE because it would improve the experience for 847 answerers. 849 16. Example Flow 851 A typical successful trickle ICE exchange with an Offer/Answer 852 protocol would look this way: 854 Alice Bob 855 | Offer | 856 |---------------------------------------------->| 857 | Additional Candidates | 858 |---------------------------------------------->| 859 | | 860 | Answer | 861 |<----------------------------------------------| 862 | Additional Candidates | 863 |<----------------------------------------------| 864 | | 865 | Additional Candidates and Connectivity Checks | 866 |<--------------------------------------------->| 867 | | 868 |<=============== MEDIA FLOWS =================>| 870 Figure 2: Example 872 17. Security Considerations 874 This specification inherits most of its semantics from [RFC5245] and 875 as a result all security considerations described there remain the 876 same. 878 18. Acknowledgements 880 The authors would like to thank Bernard Aboba, Flemming Andreasen, 881 Rajmohan Banavi, Christer Holmberg, Jonathan Lennox, Enrico Marocco, 882 Pal Martinsen, Martin Thomson, Dale R. Worley, and Brandon Williams 883 for their reviews and suggestions on improving this document. 885 19. References 887 19.1. Normative References 889 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 890 Requirement Levels", BCP 14, RFC 2119, March 1997. 892 [RFC3264] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model 893 with Session Description Protocol (SDP)", RFC 3264, June 894 2002. 896 [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session 897 Description Protocol", RFC 4566, July 2006. 899 [RFC5245] Rosenberg, J., "Interactive Connectivity Establishment 900 (ICE): A Protocol for Network Address Translator (NAT) 901 Traversal for Offer/Answer Protocols", RFC 5245, April 902 2010. 904 19.2. Informative References 906 [I-D.ietf-mmusic-trickle-ice-sip] 907 Ivov, E., Thomas, T., Marocco, E., and C. Holmberg, "A 908 Session Initiation Protocol (SIP) usage for Trickle ICE", 909 draft-ietf-mmusic-trickle-ice-sip-03 (work in progress), 910 October 2015. 912 [I-D.keranen-mmusic-ice-address-selection] 913 Keraenen, A. and J. Arkko, "Update on Candidate Address 914 Selection for Interactive Connectivity Establishment 915 (ICE)", draft-keranen-mmusic-ice-address-selection-01 916 (work in progress), July 2012. 918 [RFC1918] Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G., 919 and E. Lear, "Address Allocation for Private Internets", 920 BCP 5, RFC 1918, DOI 10.17487/RFC1918, February 1996, 921 . 923 [RFC2543] Handley, M., Schulzrinne, H., Schooler, E., and J. 924 Rosenberg, "SIP: Session Initiation Protocol", RFC 2543, 925 DOI 10.17487/RFC2543, March 1999, 926 . 928 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 929 A., Peterson, J., Sparks, R., Handley, M., and E. 930 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 931 June 2002. 933 [RFC3388] Camarillo, G., Eriksson, G., Holler, J., and H. 934 Schulzrinne, "Grouping of Media Lines in the Session 935 Description Protocol (SDP)", RFC 3388, DOI 10.17487/ 936 RFC3388, December 2002, 937 . 939 [RFC4787] Audet, F., Ed. and C. Jennings, "Network Address 940 Translation (NAT) Behavioral Requirements for Unicast 941 UDP", BCP 127, RFC 4787, DOI 10.17487/RFC4787, January 942 2007, . 944 [RFC5389] Rosenberg, J., Mahy, R., Matthews, P., and D. Wing, 945 "Session Traversal Utilities for NAT (STUN)", RFC 5389, 946 DOI 10.17487/RFC5389, October 2008, 947 . 949 [RFC5766] Mahy, R., Matthews, P., and J. Rosenberg, "Traversal Using 950 Relays around NAT (TURN): Relay Extensions to Session 951 Traversal Utilities for NAT (STUN)", RFC 5766, April 2010. 953 [XEP-0030] 954 Hildebrand, J., Millard, P., Eatmon, R., and P. Saint- 955 Andre, "XEP-0030: Service Discovery", XEP XEP-0030, June 956 2008. 958 [XEP-0176] 959 Beda, J., Ludwig, S., Saint-Andre, P., Hildebrand, J., 960 Egan, S., and R. McQueen, "XEP-0176: Jingle ICE-UDP 961 Transport Method", XEP XEP-0176, June 2009. 963 [XEP-0278] 964 Camargo, T., "XEP-0278: Jingle Relay Nodes", XEP XEP-0278, 965 June 2011. 967 Appendix A. Open Issues 969 At the time of writing of this document the authors have no clear 970 view on how and if the following list of issues should be addressed. 972 A.1. MID/Stream Indices in SDP 974 This specification does not currently define syntax for candidate-to- 975 stream bindings although it says that they should be implemented with 976 MID or a stream index. Yet, it is reasonable to assume that most 977 usages would need to do this within the SDP and it may make sense to 978 agree on the format. Here's one possible way to do this: 980 a=mid:1 981 a=candidate:1 1 UDP 1658497328 192.168.100.33 5000 typ host 982 a=candidate:2 1 UDP 1658497328 96.1.2.3 5000 typ srflx 983 a=mid:2 984 a=candidate:2 1 UDP 1658497328 96.1.2.3 5002 typ srflx 985 a=end-of-candidates 987 A.2. Starting Checks 989 Normally vanilla ICE implementations would first activate a check 990 list, validate at least one pair in every component and only then 991 unfreeze all other checklists. With trickle ICE this would be 992 suboptimal since candidates can arrive randomly and we would be 993 wasting time waiting for a checklist to fill (almost as if we were 994 doing vanilla ICE). We need to decide if unfreezing everything 995 solely based on foundation is good enough. 997 A.3. Checklist States 999 It's been proposed that we add a waiting-for-candidates state (e.g., 1000 if the checklist is empty and no candidate pairs have been sent or 1001 received yet). 1003 A.4. Relationship to Continuous Nomination and Passive Nomination 1005 Does it make sense to tie trickle ICE more explicitly the continuous 1006 nomination and passive nomination specs? In particular, is address 1007 mobility a goal for the trickle ICE specification? 1009 A.5. ICE Restarts 1011 We need to describe how trickle ICE interacts with ICE restarts. 1012 Specifically, is it sufficient to modify the ufrag and pwd without 1013 starting a full offer/answer exchange, if the signaling protocol 1014 being used does not require it or if the restarting entity does not 1015 include a media description? 1017 A.6. Candidate Redundancy and Priority 1019 We need to clarify the relationship between RFC 5245 and trickle ICE 1020 with respect to candidate redundancy and priority. 1022 A.7. Make Trickle ICE SDP-Agnostic 1024 Would it make sense to remove the tie to SDP in the spec? This is 1025 similar to what's being done with the ICEbis spec, so consistency 1026 might be desirable. 1028 Appendix B. Interaction with ICE 1030 The ICE protocol was designed to be flexible enough to would work in 1031 and adapt to as many network environments as possible. Despite that 1032 flexibility, ICE as specified in [RFC5245] does not by itself support 1033 trickle ICE. This section describes how trickling of candidates 1034 interacts with ICE. 1036 [RFC5245] describes the conditions required to update check lists and 1037 timer states while an ICE agent is in the Running state. These 1038 conditions are verified upon transaction completion and one of them 1039 stipulates that: 1041 If there is not a pair in the valid list for each component of the 1042 media stream, the state of the check list is set to Failed. 1044 This could be a problem and cause ICE processing to fail prematurely 1045 in a number of scenarios. Consider the following case: 1047 1. Alice and Bob are both located in different networks with Network 1048 Address Translation (NAT). Alice and Bob themselves have 1049 different address but both networks use the same [RFC1918] block. 1051 2. Alice sends Bob the candidate 10.0.0.10 which also happens to 1052 correspond to an existing host on Bob's network. 1054 3. Bob creates a check list consisting solely of 10.0.0.10 and 1055 starts checks. 1057 4. These checks reach the host at 10.0.0.10 in Bob's network, which 1058 responds with an ICMP "port unreachable" error and per [RFC5245] 1059 Bob marks the transaction as Failed. 1061 At this point the check list only contains Failed candidates and the 1062 valid list is empty. This causes the media stream and potentially 1063 all ICE processing to Fail. 1065 A similar race condition would occur if the initial offer from Alice 1066 only contains candidates that can be determined as unreachable (per 1067 [I-D.keranen-mmusic-ice-address-selection]) from any of the 1068 candidates that Bob has gathered. This would be the case if Bob's 1069 candidates only contain IPv4 addresses and the first candidate that 1070 he receives from Alice is an IPv6 one. 1072 Another potential problem could arise when a non-trickle ICE 1073 implementation sends an offer to a trickle one. Consider the 1074 following case: 1076 1. Alice's client has a non-trickle ICE implementation 1078 2. Bob's client has support for trickle ICE. 1080 3. Alice and Bob are behind NATs with address-dependent filtering 1081 [RFC4787]. 1083 4. Bob has two STUN servers but one of them is currently unreachable 1084 After Bob's agent receives Alice's offer it would immediately start 1085 connectivity checks. It would also start gathering candidates, which 1086 would take long because of the unreachable STUN server. By the time 1087 Bob's answer is ready and sent to Alice, Bob's connectivity checks 1088 may well have failed: until Alice gets Bob's answer, she won't be 1089 able to start connectivity checks and punch holes in her NAT. The 1090 NAT would hence be filtering Bob's checks as originating from an 1091 unknown endpoint. 1093 Appendix C. Changes from Earlier Versions 1095 Note to the RFC-Editor: please remove this section prior to 1096 publication as an RFC. 1098 C.1. Changes from draft-mmusic-trickle-ice-02 1100 o Addressed feedback from Rajmohan Banavi and Brandon Williams. 1102 o Clarified text about determining support and about how to proceed 1103 if it can be determined that the answering agent does not support 1104 trickle ICE. 1106 o Clarified text about check list and timer updates. 1108 o Clarified when it is appropriate to use half trickle or to send no 1109 candidates in an offer or answer. 1111 o Updated the list of open issues. 1113 C.2. Changes from draft-ivov-01 and draft-mmusic-00 1115 o Added a requirement to trickle candidates by order of components 1116 to avoid deadlocks in the unfreezing algorithm. 1118 o Added an informative note on peer-reflexive candidates explaining 1119 that nothing changes for them semantically but they do become a 1120 more likely occurrence for Trickle ICE. 1122 o Limit the number of pairs to 100 to comply with 5245. 1124 o Added clarifications on the non-importance of how newly discovered 1125 candidates are trickled/sent to the remote party or if this is 1126 done at all. 1128 o Added transport expectations for trickled candidates as per Dale 1129 Worley's recommendation. 1131 C.3. Changes from draft-ivov-00 1133 o Specified that end-of-candidates is a media level attribute which 1134 can of course appear as session level, which is equivalent to 1135 having it appear in all m-lines. Also made end-of-candidates 1136 optional for cases such as aggressive nomination for controlled 1137 agents. 1139 o Added an example for ICE lite and trickle ICE to illustrate how, 1140 when talking to an ICE lite agent doesn't need to send or even 1141 discover any candidates. 1143 o Added an example for ICE lite and trickle ICE to illustrate how, 1144 when talking to an ICE lite agent doesn't need to send or even 1145 discover any candidates. 1147 o Added wording that explicitly states ICE lite agents have to be 1148 prepared to receive no candidates over signalling and that they 1149 should not freak out if this happens. (Closed the corresponding 1150 open issue). 1152 o It is now mandatory to use MID when trickling candidates and using 1153 m-line indexes is no longer allowed. 1155 o Replaced use of 0.0.0.0 to IP6 :: in order to avoid potential 1156 issues with RFC2543 SDP libraries that interpret 0.0.0.0 as an on- 1157 hold operation. Also changed the port number here from 1 to 9 1158 since it already has a more appropriate meaning. (Port change 1159 suggested by Jonathan Lennox). 1161 o Closed the Open Issue about use about what to do with cands 1162 received after end-of-cands. Solution: ignore, do an ICE restart 1163 if you want to add something. 1165 o Added more terminology, including trickling, trickled candidates, 1166 half trickle, full trickle, 1168 o Added a reference to the SIP usage for trickle ICE as requested at 1169 the Boston interim. 1171 C.4. Changes from draft-rescorla-01 1173 o Brought back explicit use of Offer/Answer. There are no more 1174 attempts to try to do this in an O/A independent way. Also 1175 removed the use of ICE Descriptions. 1177 o Added SDP specification for trickled candidates, the trickle 1178 option and 0.0.0.0 addresses in m-lines, and end-of-candidates. 1180 o Support and Discovery. Changed that section to be less abstract. 1181 As discussed in IETF85, the draft now says implementations and 1182 usages need to either determine support in advance and directly 1183 use trickle, or do half trickle. Removed suggestion about use of 1184 discovery in SIP or about letting implementing protocols do what 1185 they want. 1187 o Defined Half Trickle. Added a section that says how it works. 1188 Mentioned that it only needs to happen in the first o/a (not 1189 necessary in updates), and added Jonathan's comment about how it 1190 could, in some cases, offer more than half the improvement if you 1191 can pre-gather part or all of your candidates before the user 1192 actually presses the call button. 1194 o Added a short section about subsequent offer/answer exchanges. 1196 o Added a short section about interactions with ICE Lite 1197 implementations. 1199 o Added two new entries to the open issues section. 1201 C.5. Changes from draft-rescorla-00 1203 o Relaxed requirements about verifying support following a 1204 discussion on MMUSIC. 1206 o Introduced ICE descriptions in order to remove ambiguous use of 1207 3264 language and inappropriate references to offers and answers. 1209 o Removed inappropriate assumption of adoption by RTCWEB pointed out 1210 by Martin Thomson. 1212 Authors' Addresses 1214 Emil Ivov 1215 Jitsi 1216 Strasbourg 67000 1217 France 1219 Phone: +33 6 72 81 15 55 1220 Email: emcho@jitsi.org 1221 Eric Rescorla 1222 RTFM, Inc. 1223 2064 Edgewood Drive 1224 Palo Alto, CA 94303 1225 USA 1227 Phone: +1 650 678 2350 1228 Email: ekr@rtfm.com 1230 Justin Uberti 1231 Google 1232 747 6th St S 1233 Kirkland, WA 98033 1234 USA 1236 Phone: +1 857 288 8888 1237 Email: justin@uberti.name 1239 Peter Saint-Andre 1240 &yet 1242 Email: peter@andyet.com 1243 URI: https://andyet.com/