idnits 2.17.1 draft-ietf-mmusic-trickle-ice-02.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 (January 15, 2015) is 3388 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) == Unused Reference: 'RFC3261' is defined on line 957, but no explicit reference was found in the text == Unused Reference: 'RFC3840' is defined on line 966, but no explicit reference was found in the text == Unused Reference: 'XEP-0115' is defined on line 987, but no explicit reference was found in the text ** Obsolete normative reference: RFC 4566 (Obsoleted by RFC 8866) ** Obsolete normative reference: RFC 5245 (Obsoleted by RFC 8445, RFC 8839) -- 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 (~~), 6 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: July 19, 2015 RTFM, Inc. 6 J. Uberti 7 Google 8 January 15, 2015 10 Trickle ICE: Incremental Provisioning of Candidates for the Interactive 11 Connectivity Establishment (ICE) Protocol 12 draft-ietf-mmusic-trickle-ice-02 14 Abstract 16 This document describes an extension to the Interactive Connectivity 17 Establishment (ICE) protocol that allows ICE agents to send and 18 receive candidates incrementally rather than exchanging complete 19 lists. With such incremental provisioning, ICE agents can begin 20 connectivity checks while they are still gathering candidates and 21 considerably shorten the time necessary for ICE processing to 22 complete. 24 The above mechanism is also referred to as "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 July 19, 2015. 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. Incompatibility with Standard ICE . . . . . . . . . . . . . . 5 63 4. Determining Support for Trickle ICE . . . . . . . . . . . . . 6 64 4.1. Unilateral Use of Trickle ICE (Half Trickle) . . . . . . 7 65 5. Sending the Initial Offer . . . . . . . . . . . . . . . . . . 8 66 5.1. Encoding the SDP . . . . . . . . . . . . . . . . . . . . 9 67 6. Receiving the Initial Offer . . . . . . . . . . . . . . . . . 9 68 6.1. Sending the Initial Answer . . . . . . . . . . . . . . . 10 69 6.2. Forming check lists and beginning connectivity 70 checks . . . . . . . . . . . . . . . . . . . . . . . . . 10 71 6.3. Encoding the SDP . . . . . . . . . . . . . . . . . . . . 11 72 7. Receiving the Initial Answer . . . . . . . . . . . . . . . . 11 73 8. Performing Connectivity Checks . . . . . . . . . . . . . . . 11 74 8.1. Check List and Timer State Updates . . . . . . . . . . . 11 75 9. Discovering and Sending Additional Local Candidates . . . . . 12 76 9.1. Pairing newly learned candidates and updating 77 check lists . . . . . . . . . . . . . . . . . . . . . . . 14 78 9.2. Encoding the SDP for Additional Candidates . . . . . . . 15 79 9.3. Announcing End of Candidates . . . . . . . . . . . . . . 15 80 10. Receiving Additional Remote Candidates . . . . . . . . . . . 17 81 11. Receiving an End Of Candidates Notification . . . . . . . . . 17 82 12. Trickle ICE and Peer Reflexive Candidates . . . . . . . . . . 17 83 13. Concluding ICE Processing . . . . . . . . . . . . . . . . . . 18 84 14. Subsequent Offer/Answer Exchanges . . . . . . . . . . . . . . 18 85 15. Interaction with ICE Lite . . . . . . . . . . . . . . . . . . 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 Appendix B. Changes From Earlier Versions . . . . . . . . . . . 23 96 B.1. Changes From draft-ivov-01 and draft-mmusic-00 . . . . . 23 97 B.2. Changes From draft-ivov-00 . . . . . . . . . . . . . . . 23 98 B.3. Changes From draft-rescorla-01 . . . . . . . . . . . . . 24 99 B.4. Changes From draft-rescorla-00 . . . . . . . . . . . . . 25 100 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 25 102 1. Introduction 104 The Interactive Connectivity Establishment (ICE) protocol [RFC5245] 105 describes mechanisms for gathering, candidates, prioritizing them, 106 choosing default ones, exchanging them with the remote party, pairing 107 them and ordering them into check lists. Once all of the above have 108 been completed, and only then, the participating agents can begin a 109 phase of connectivity checks and eventually select the pair of 110 candidates that will be used in the following session. 112 While the above sequence has the advantage of being relatively 113 straightforward to implement and debug once deployed, it may also 114 prove to be rather lengthy. Gathering candidates or candidate 115 harvesting would often involve things like querying STUN [RFC5389] 116 servers, discovering UPnP devices, and allocating relayed candidates 117 at TURN [RFC5766] servers. All of these can be delayed for a 118 noticeable amount of time and while they can be run in parallel, they 119 still need to respect the pacing requirements from [RFC5245], which 120 is likely to delay them even further. Some or all of the above would 121 also have to be completed by the remote agent. Both agents would 122 next perform connectivity checks and only then would they be ready to 123 begin streaming media. 125 All of the above could lead to relatively lengthy session 126 establishment times and degraded user experience. 128 The purpose of this document is to define an alternative mode of 129 operation for ICE implementations, also known as "trickle ICE", where 130 candidates can be exchanged incrementally. This would allow ICE 131 agents to exchange host candidates as soon as a session has been 132 initiated. Connectivity checks for a media stream would also start 133 as soon as the first candidates for that stream have become 134 available. 136 Trickle ICE allows reducing session establishment times in cases 137 where connectivity is confirmed for the first exchanged candidates 138 (e.g. where the host candidates for one of the agents are directly 139 reachable from the second agent). Even when this is not the case, 140 running candidate harvesting for both agents and connectivity checks 141 all in parallel allows to considerably reduce ICE processing times. 143 It is worth pointing out that before being introduced to the IETF, 144 trickle ICE had already been included in specifications such as XMPP 145 Jingle [XEP-0176] and it has been in use in various implementations 146 and deployments. 148 In addition to the basics of trickle ICE, this document also 149 describes how support for trickle ICE needs to be discovered, how 150 regular ICE processing needs to be modified when building and 151 updating check lists, and how trickle ICE implementations should 152 interoperate with agents that only implement [RFC5245] processing. 154 This specification does not define usage of trickle ICE with any 155 specific signalling protocol, contrary to [RFC5245] which contains a 156 usage for ICE with SIP. Such usages would have to be specified in 157 separate documents such as for example 158 [I-D.ivov-mmusic-trickle-ice-sip]. 160 Trickle ICE does however reuse and build upon the SDP syntax defined 161 by [RFC5245]. 163 2. Terminology 165 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 166 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 167 document are to be interpreted as described in [RFC2119]. 169 This specification makes use of all terminology defined by the 170 protocol for Interactive Connectivity Establishment in [RFC5245]. 172 Vanilla ICE: The Interactive Connectivity Establishment protocol as 173 defined in [RFC5245]. Through the rest of the text, the terms 174 vanilla ICE and "RFC5245" are used interchangeably. 176 Candidate Harvester: A module used by an ICE agent to obtain local 177 candidates. Candidate harvesters use different mechanisms for 178 discovering local candidates. Some of them would typically make 179 use of protocols such as STUN or TURN. Others may also employ 180 techniques that are not referenced within [RFC5245]. UPnP based 181 port allocation and XMPP Jingle Relay Nodes [XEP-0278] are among 182 the possible examples. 184 Trickled Candidates: Candidates that a trickle ICE agent is sending 185 subsequently to but within the context defined by an offer or an 186 answer. Trickled candidates can be sent in parallel with 187 candidate harvesting and connectivity checks. 189 Trickling/Trickle (v.): The act of sending trickled candidates. 191 Half Trickle: A trickle ICE mode of operation where the offerer 192 gathers its first generation of candidates strictly before 193 creating and sending the offer. Once sent, that offer can be 194 processed by vanilla ICE agents and does not require support for 195 this specification. It also allows trickle ICE capable answerers 196 to still gather candidates and perform connectivity checks in a 197 non-blocking way, thus roughly offering "half" the advantages of 198 trickle ICE. The mechanism is mostly meant for use in cases where 199 support for trickle ICE cannot be confirmed prior to sending a 200 first offer. 202 Full Trickle: Regular mode of operation for trickle ICE agents, used 203 in opposition to the half trickle mode of operation. 205 3. Incompatibility with Standard ICE 207 The ICE protocol was designed to be fairly flexible so that it would 208 work in and adapt to as many network environments as possible. It is 209 hence important to point out at least some of the reasons why, 210 despite its flexibility, the specification in [RFC5245] would not 211 support trickle ICE. 213 [RFC5245] describes the conditions required to update check lists and 214 timer states while an ICE agent is in the Running state. These 215 conditions are verified upon transaction completion and one of them 216 stipulates that: 218 If there is not a pair in the valid list for each component of the 219 media stream, the state of the check list is set to Failed. 221 This could be a problem and cause ICE processing to fail prematurely 222 in a number of scenarios. Consider the following case: 224 o Alice and Bob are both located in different networks with Network 225 Address Translation (NAT). Alice and Bob themselves have 226 different address but both networks use the same [RFC1918] block. 228 o Alice sends Bob the candidate 10.0.0.10 which also happens to 229 correspond to an existing host on Bob's network. 231 o Bob creates a check list consisting solely of 10.0.0.10 and starts 232 checks. 234 o These checks reach the host at 10.0.0.10 in Bob's network, which 235 responds with an ICMP "port unreachable" error and per [RFC5245] 236 Bob marks the transaction as Failed. 238 At this point the check list only contains Failed candidates and the 239 valid list is empty. This causes the media stream and potentially 240 all ICE processing to Fail. 242 A similar race condition would occur if the initial offer from Alice 243 only contains candidates that can be determined as unreachable (per 244 [I-D.keranen-mmusic-ice-address-selection]) from any of the 245 candidates that Bob has gathered. This would be the case if Bob's 246 candidates only contain IPv4 addresses and the first candidate that 247 he receives from Alice is an IPv6 one. 249 Another potential problem could arise when a non-trickle ICE 250 implementation sends an offer to a trickle one. Consider the 251 following case: 253 o Alice's client has a non-trickle ICE implementation 255 o Bob's client has support for trickle ICE. 257 o Alice and Bob are behind NATs with address-dependent filtering 258 [RFC4787]. 260 o Bob has two STUN servers but one of them is currently unreachable 262 After Bob's agent receives Alice's offer it would immediately start 263 connectivity checks. It would also start gathering candidates, which 264 would take long because of the unreachable STUN server. By the time 265 Bob's answer is ready and sent to Alice, Bob's connectivity checks 266 may well have failed: until Alice gets Bob's answer, she won't be 267 able to start connectivity checks and punch holes in her NAT. The 268 NAT would hence be filtering Bob's checks as originating from an 269 unknown endpoint. 271 4. Determining Support for Trickle ICE 273 According to [RFC5245] every time an agent supporting trickle ICE 274 generates an offer or an answer, it MUST include the "trickle" token 275 in the ice-options attribute. Syntax for this token is defined in 276 Section 5.1. 278 Additionally, in order to avoid interoperability problems such as 279 those described in Section 3, it is important that trickle ICE 280 negotiation is only attempted in cases where the remote party 281 actually supports this specification. Agents that receive offers or 282 answers can verify support by examining them for the "trickle" ice- 283 options token. However, agents that are about to send a first offer, 284 have no immediate way of doing this. This means that usages of 285 trickle for specific protocols would need to either: 287 o Provide a way for agents to verify support of trickle ICE prior to 288 initiating a session. XMPP's Service discovery [XEP-0030] is an 289 example for one such mechanism; 291 o Make support for trickle ICE mandatory so that support could be 292 assumed the agents. 294 Alternately, for cases where a protocol provides neither of the 295 above, agents may either rely on provisioning/configuration, or use 296 the half trickle procedure described in Section 4.1. 298 Note that out-of-band discovery semantics and half trickle are only 299 necessary prior to session initiation, or in other words, when 300 sending the initial offer. Once a session is established and trickle 301 ICE support is confirmed for both parties, either agent can use full 302 trickle for subsequent offers. 304 4.1. Unilateral Use of Trickle ICE (Half Trickle) 306 The idea of using half trickle is about having the caller send a 307 regular, vanilla ICE offer, with a complete set of candidates. This 308 offer still indicates support for trickle ice, so the answerer is 309 able to respond with an incomplete set of candidates and continue 310 trickling the rest. Half trickle offers will typically contain an 311 end-of-candidates indication, although this is not mandatory as, in 312 case trickle support is confirmed, the offerer may choose to trickle 313 additional candidates (e.g., additional relay candidates) before it 314 declares end of trickling. 316 The half trickle mechanism can be used in cases where there is no way 317 for an agent to verify in advance whether a remote party supports 318 trickle ice. Because it contains a full set of candidates, its first 319 offer can thus be handled by a regular vanilla ICE agent, while still 320 allowing a trickle one to use the optimisation defined in this 321 specification. This prevents negotiation from failing in the former 322 case while still giving roughly half the trickle ICE benefits in the 323 latter (hence the name of the mechanism). 325 Use of half trickle is only necessary during an initial offer/answer 326 exchange. Once both parties have received a session description from 327 their peer, they can each reliably determine trickle ICE support and 328 use it for all subsequent offer/answer exchanges. 330 It is worth pointing out that using half trickle may actually bring 331 more than just half the improvement in terms of user experience. 332 This can happen in cases where an agent starts gathering candidates 333 upon user interface cues that a call is pending, such as activity on 334 a keypad or the phone going off hook. This would mean a part or all 335 candidate harvesting could have completed before the agent actually 336 needs to send the offer. Given that the answerer will be able to 337 trickle candidates, both agents will be able to start connectivity 338 checks and complete ICE processing earlier than with vanilla ICE and 339 potentially even as early as with full trickle. 341 However, such anticipation is not not always possible. For example, 342 a multipurpose user agent or a WebRTC web page where communication is 343 a non-central feature (e.g. calling a support line in case of a 344 problem with the main features) would not necessarily have a way of 345 distinguishing between call intentions and other user activity. 346 Still, even in these cases, using half trickle would be an 347 improvement over vanilla ICE as it would optimize performance for 348 answerers. 350 5. Sending the Initial Offer 352 An agent starts gathering candidates as soon as it has an indication 353 that communication is imminent (e.g. a user interface cue or an 354 explicit request to initiate a session). Contrary to vanilla ICE, 355 implementations of trickle ICE do not need to gather candidates in a 356 blocking manner. Therefore, unless half trickle is being used, 357 agents SHOULD generate and transmit their initial offer as early as 358 possible, in order to allow the remote party to start gathering and 359 trickling candidates. 361 Trickle ICE agents MAY include any set of candidates in an offer. 362 This includes the possibility of generating one with no candidates, 363 or one that contains all the candidates that the agent is planning on 364 using in the following session. 366 For optimal performance, it is RECOMMENDED that an initial offer 367 contains host candidates only. This would allow both agents to start 368 gathering server reflexive, relayed and other non-host candidates 369 simultaneously, and it would also enable them to begin connectivity 370 checks. 372 If the privacy implications of revealing host addresses are a 373 concern, agents MAY generate an offer that contains no candidates and 374 then only trickle candidates that do not reveal host addresses (e.g. 375 relayed candidates). 377 Prior to actually sending an initial offer, agents MAY verify if the 378 remote party supports trickle ICE, where such mechanisms actually 379 exist. If absence of such support is confirmed agents MUST fall back 380 to using vanilla ICE or abandon the entire session. 382 All trickle ICE offers and answers MUST indicate support of this 383 specification, as explained in Section 5.1. 385 Calculating priorities and foundations, as well as determining 386 redundancy of candidates work the same way they do with vanilla ICE. 388 5.1. Encoding the SDP 390 The process of encoding the SDP [RFC4566] is mostly the same as the 391 one used by vanilla ICE. Still, trickle ICE does require a few 392 differences described here. 394 Agents MUST indicate support for Trickle ICE by including the 395 "trickle" token for the "a=ice-options" attribute: 397 a=ice-options:trickle 399 As mentioned earlier in this section, Offers and Answers can contain 400 any set of candidates, which means that a trickle ICE session 401 description MAY contain no candidates at all. In such cases the 402 agent would still need to place an address in the "c=" line(s). If 403 the use of a host address there is undesirable (e.g. for privacy 404 reasons), the agent MAY set the connection address to IP6 ::. In this 405 case it MUST also set the port number to 9 (Discard). There is no 406 need to include a fictitious candidate for the IP6 :: address when 407 doing so. 409 It is worth noting that the use of IP6 :: has been selected over IP4 410 0.0.0.0, even though [RFC3264] already gives the latter semantics 411 appropriate for such use. The reason for this choice is the historic 412 use of 0.0.0.0 as a means of putting a stream on hold [RFC2543] and 413 the ambiguity that this may cause with legacy libraries and 414 applications. 416 It is also worth mentioning that use of IP6 :: here does not 417 constitute any kind of indication as to the actual use of IPv6 418 candidates in a session and it can very well appear in a negotiation 419 that only involves IPv4 candidates. 421 6. Receiving the Initial Offer 423 When an agent receives an initial offer, it will first check if it 424 indicates support for trickle ICE as explained in Section 4. If this 425 is not the case, the agent MUST process the offer according to the 426 [RFC5245] procedures or standard [RFC3264] processing in case no ICE 427 support is detected at all. 429 It is worth pointing out that in case support for trickle ICE is 430 confirmed, an agent will automatically assume support for vanilla ICE 431 as well even if the support verification procedure in [RFC5245] 432 indicates otherwise. Specifically, such verification would indicate 433 lack of support when the offer contains no candidates. The IP6 :: 434 address present in the c= line in that case would not "appear in a 435 candidate attribute". Obviously, a fallback to [RFC3264] is not 436 required when this happens. 438 If, the offer does indicate support for trickle ICE, the agent will 439 determine its role, start gathering and prioritizing candidates and, 440 while doing so it will also respond by sending its own answer, so 441 that both agents can start forming check lists and begin connectivity 442 checks. 444 6.1. Sending the Initial Answer 446 An agent can respond to an initial offer at any point while gathering 447 candidates. The answer can again contain any set of candidates 448 including none or all of them. Unless it is protecting host 449 addresses for privacy reasons, the agent would typically construct 450 this initial answer including only them, thus allowing the remote 451 party to also start forming checklists and performing connectivity 452 checks. 454 The answer MUST indicate support for trickle ICE as described by 455 Section 4. 457 6.2. Forming check lists and beginning connectivity checks 459 After exchanging offer and answer, and as soon as they have obtained 460 local and remote candidates, agents will begin forming candidate 461 pairs, computing their priorities and creating check lists according 462 to the vanilla ICE procedures described in [RFC5245]. Obviously in 463 order for candidate pairing to be possible, it would be necessary 464 that both the offer and the answer contained candidates. If this was 465 not the case agents will still create the check lists (so that their 466 Active/Frozen state could be monitored and updated) but they will 467 only populate them once they actually have the candidate pairs. 469 Initially, all check lists will have their Active/Frozen state set to 470 Frozen. 472 Trickle ICE agents will then inspect the first check list and attempt 473 to unfreeze all candidates belonging to the first component on the 474 first media stream (i.e. the first media stream that was reported to 475 the ICE implementation from the using application). If this 476 checklist is still empty however, agents will hold off further 477 processing until this is no longer the case. 479 Respecting the order in which lists have been reported to an ICE 480 implementation, or in other words, the order in which they appear in 481 SDP, is crucial to the frozen candidates algorithm and important when 482 making sure that connectivity checks are performed simultaneously by 483 both agents. 485 6.3. Encoding the SDP 487 The process for encoding the SDP at the answerer is identical to the 488 process followed by the offerer for both full and lite 489 implementations, as described in Section 5.1. 491 7. Receiving the Initial Answer 493 When receiving an answer, agents will follow vanilla ICE procedures 494 to determine their role and they would then form check lists (as 495 described in Section 6.2) and begin connectivity checks . 497 8. Performing Connectivity Checks 499 For the most part, trickle ICE agents perform connectivity checks 500 following vanilla ICE procedures. Of course, the asynchronous nature 501 of candidate harvesting in trickle ICE would impose a number of 502 changes described here. 504 8.1. Check List and Timer State Updates 506 The vanilla ICE specification requires that agents update check lists 507 and timer states upon completing a connectivity check transaction. 508 During such an update vanilla ICE agents would set the state of a 509 check list to Failed if the following two conditions are satisfied: 511 o all of the pairs in the check list are either in the Failed or 512 Succeeded state; 514 o if at least one of the components of the media stream has no pairs 515 in its valid list. 517 With trickle ICE, the above situation would often occur when 518 candidate harvesting and trickling are still in progress and it is 519 perfectly possible that future checks will succeed. For this reason 520 trickle ICE agents add the following conditions to the above list: 522 o all candidate harvesters have completed and the agent is not 523 expecting to discover any new local candidates; 525 o the remote agent has sent an end-of-candidates indication for that 526 check list as described in Section 9.3. 528 Vanilla ICE requires that agents then update all other check lists, 529 placing one pair in each of them into the Waiting state, effectively 530 unfreezing all remaining check lists. Given that with trickle ICE, 531 other check lists may still be empty at that point, a trickle ICE 532 agent SHOULD also maintain an explicit Active/Frozen state for every 533 check list, rather than deducing it from the state of the pairs it 534 contains. This state should be set to Active when unfreezing the 535 first pair in a list or when that couldn't happen because a list was 536 empty. 538 9. Discovering and Sending Additional Local Candidates 540 After an offer or an answer have been sent, agents will most likely 541 continue discovering new local candidates as STUN, TURN and other 542 non-host candidate harvesting mechanisms begin to yield results. 543 Whenever an agent discovers such a new candidate it will compute its 544 priority, type, foundation and component id according to normal 545 vanilla ICE procedures. 547 The new candidate is then checked for redundancy against the existing 548 list of local candidates. If its transport address and base match 549 those of an existing candidate, it will be considered redundant and 550 will be ignored. This would often happen for server reflexive 551 candidates that match the host addresses they were obtained from 552 (e.g. when the latter are public IPv4 addresses). Contrary to 553 vanilla ICE, trickle ICE agents will consider the new candidate 554 redundant regardless of its priority. 556 Next the client sends (i.e. trickles) the newly learnt candidate(s) 557 to the remote agent. The actual delivery of the new candidates will 558 be specified by using protocols such as SIP. Trickle ICE imposes no 559 restrictions on the way this is done or whether it is done at all. 560 For example, some applications may choose not to send trickle updates 561 for server reflexive candidates and rely on the discovery of peer 562 reflexive ones instead. 564 When trickle updates are sent however, each candidate MUST be 565 delivered to the receiving Trickle ICE implementation not more than 566 once and in the same order that they were sent. In other words, if 567 there are any candidate retransmissions, they must be hidden from the 568 ICE implementation. 570 Also, candidate trickling needs to be correlated to a specific ICE 571 negotiation session, so that if there is an ICE restart, any delayed 572 updates for a previous session can be recognized as such and ignored 573 by the receiving party. 575 One important aspect of Vanilla ICE is that connectivity checks for a 576 specific foundation and component be attempted simultaneously by both 577 agents, so that any firewalls or NATs fronting the agents would 578 whitelist both endpoints and allow all except for the first (suicide) 579 packets to go through. This is also crucial to unfreezing candidates 580 in the right time. 582 In order to preserve this feature here, when trickling candidates 583 agents MUST respect the order of the components as they appear 584 (implicitly or explicitly) in the Offer/Answer descriptions. 585 Therefore a candidate for a specific component MUST NOT be sent prior 586 to candidates for other components within the same foundation. 588 For example, the following session description contains two 589 components (RTP and RTCP), and two foundations (host and the server 590 reflexive): 592 v=0 593 o=jdoe 2890844526 2890842807 IN IP4 10.0.1.1 594 s= 595 c=IN IP4 10.0.1.1 596 t=0 0 597 a=ice-pwd:asd88fgpdd777uzjYhagZg 598 a=ice-ufrag:8hhY 599 m=audio 5000 RTP/AVP 0 600 a=rtpmap:0 PCMU/8000 601 a=candidate:1 1 UDP 2130706431 10.0.1.1 5000 typ host 602 a=candidate:1 2 UDP 2130706431 10.0.1.1 5001 typ host 603 a=candidate:2 1 UDP 1694498815 192.0.2.3 5000 typ srflx 604 raddr 10.0.1.1 rport 8998 605 a=candidate:2 2 UDP 1694498815 192.0.2.3 5001 typ srflx 606 raddr 10.0.1.1 rport 8998 608 For this description the RTCP host candidate MUST NOT be sent prior 609 to the RTP host candidate. Similarly the RTP server reflexive 610 candidate MUST be sent together with or prior to the RTCP server 611 reflexive candidate. 613 Note that the order restriction only applies among candidates that 614 belong to the same foundation. 616 It is also equally important to preserve this order across media 617 streams and this is covered by the requirement to always start 618 unfreezing candidates starting from the first media stream 619 Section 6.2. 621 Once the candidate has been sent to the remote party, the agent 622 checks if any remote candidates are currently known for this same 623 stream. If this is not the case the new candidate will simply be 624 added to the list of local candidates. 626 Otherwise, if the agent has already learned of one or more remote 627 candidates for this stream and component, it will begin pairing the 628 new local candidates with them and adding the pairs to the existing 629 check lists according to their priority. 631 9.1. Pairing newly learned candidates and updating check lists 633 Forming candidate pairs will work the way it is described by the 634 vanilla ICE specification. Actually adding the new pair to a check 635 list however, will happen according to the rules described below. 637 If the check list where the pair is to be added already contains the 638 maximum number of candidate pairs (100 by default as per [RFC5245]), 639 the new pair is discarded. 641 If the new pair's local candidate is server reflexive, the server 642 reflexive candidate MUST be replaced by its base before adding the 643 pair to the list. Once this is done, the agent examines the check 644 list looking for another pair that would be redundant with the new 645 one. If such a pair exists, the newly formed pair is ignored. 647 For all other pairs, including those with a server reflexive local 648 candidate that were not found to be redundant: 650 o if this check list is Frozen then the new pair will also be 651 assigned a Frozen state. 653 o else if the check list is Active and it is either empty or 654 contains only candidates in the Succeeded and Failed states, then 655 the new pair's state is set to Waiting. 657 o else if the check list is non-empty and Active, then the new pair 658 state will be set to 660 Frozen: if there is at least one pair in the list whose 661 foundation matches the one in the new pair and whose state is 662 neither Succeeded nor Failed (eventually the new pair will get 663 unfrozen after the the on-going check for the existing pair 664 concludes); 666 Waiting: if the list contains no pairs with the same foundation 667 as the new one, or, in case such pairs exist but they are all 668 in either the Succeeded or Failed states. 670 9.2. Encoding the SDP for Additional Candidates 672 To facilitate interoperability an ICE agent will encode additional 673 candidates using the vanilla ICE SDP syntax. For example: 675 a=candidate:2 1 UDP 1658497328 198.51.100.33 5000 typ host 677 Given that such lines do not provide a relationship between the 678 candidate and the m line that it relates to, signalling protocols 679 using trickle ICE MUST establish that relation themselves using an 680 MID [RFC3388]. Such MIDs use "media stream identification", as 681 defined in [RFC3388], to identify a corresponding m-line. When 682 creating candidate lines usages of trickle ICE MUST use the MID if 683 possible, or the m-line index if not. Obviously, agents MUST NOT 684 send individual candidates prior to generating the corresponding SDP 685 session description. 687 The exact means of transporting additional candidates to a remote 688 agent is left to the protocols using trickle ICE. It is important to 689 note, however, that these candidate exchanges are not part of the 690 offer/answer model. 692 9.3. Announcing End of Candidates 694 Once all candidate harvesters for a specific media stream complete, 695 or expire, the agents will generate an "end-of-candidates" indication 696 for that stream and send it to the remote agent via the signalling 697 channel. Such indications are sent in the form of a media-level 698 attribute that has the following form: end-of-candidates. 700 a=end-of-candidates 702 The end-of-candidates indications can be sent as part of an offer, 703 which would typically be the case with half trickle initial offers, 704 they can accompany the last candidate an agent can send for a stream, 705 and they can also be sent alone (e.g. after STUN Binding requests or 706 TURN Allocate requests to a server timeout and the agent has no other 707 active harvesters). 709 Controlled trickle ICE agents SHOULD always send end-of-candidates 710 indications once harvesting for a media stream has completed unless 711 ICE processing terminates before they've had a chance to do so. 712 Sending the indication is necessary in order to avoid ambiguities and 713 speed up ICE conclusion. This is necessary in order to avoid 714 ambiguities and speed up ICE conclusion. Controlling agents on the 715 other hand MAY sometimes conclude ICE processing prior to sending 716 end-of-candidates notifications for all streams. This would 717 typically be the case with aggressive nomination. Yet it is 718 RECOMMENDED that controlling agents do send such indications whenever 719 possible for the sake of consistency and keeping middle boxes and 720 controlled agents up-to-date on the state of ICE processing. 722 When sending end-of-candidates during trickling, rather than as a 723 part of an offer or an answer, it is the responsibility of the using 724 protocol to define means that can be used to relate the indication to 725 one or more specific m-lines. 727 Receiving an end-of-candidates notification allows an agent to update 728 check list states and, in case valid pairs do not exist for every 729 component in every media stream, determine that ICE processing has 730 failed. It also allows agents to speed ICE conclusion in cases where 731 a candidate pair has been validates but it involves the use of lower- 732 preference transports such as TURN. In such situations some 733 implementations may choose to wait in case higher-priority candidates 734 are received and end-of-candidates provides an indication that this 735 is not going to happen. 737 An agent MAY also choose to generate an end-of-candidates event 738 before candidate harvesting has actually completed, if the agent 739 determines that harvesting has continued for more than an acceptable 740 period of time. However, an agent MUST NOT send any more candidates 741 after it has send an end-of-candidates notification. 743 When performing half trickle agents SHOULD send end-of-candidates 744 together with their initial offer unless they are planning on 745 potentially sending additional candidates in case the remote party 746 turns out to actually support trickle ICE. 748 When end-of-candidates is sent as part of an offer or an answer it 749 can appear as a session-level attribute, which would be equivalent to 750 having it appear in all m-lines. 752 Once an agent sends the end-of-candidates event, it will update the 753 state of the corresponding check list as explained in section 754 Section 8.1. Past that point agents MUST NOT send any new 755 candidates. Once an agent has received an end-of-candidates 756 indication, it MUST also ignore any newly received candidates for 757 that media stream. Adding new candidates to the negotiation is hence 758 only possible through an ICE restart. 760 It is important to note that This specification does not override 761 vanilla ICE semantics for concluding ICE processing. This means that 762 even if end-of-candidates indications are sent agents will still have 763 to go through pair nomination. Also, if pairs have been nominated 764 for components and media streams, ICE processing will still conclude 765 even if end-of-candidate indications have not been received for all 766 streams. 768 10. Receiving Additional Remote Candidates 770 At any point of ICE processing, a trickle ICE agent may receive new 771 candidates from the remote agent. When this happens and no local 772 candidates are currently known for this same stream, the new remote 773 candidates are simply added to the list of remote candidates. 775 Otherwise, the new candidates are used for forming candidate pairs 776 with the pool of local candidates and they are added to the local 777 check lists as described in Section 9.1. 779 Once the remote agent has completed candidate harvesting, it will 780 send an end-of-candidates event. Upon receiving such an event, the 781 local agent MUST update check list states as per Section 8.1. This 782 may lead to some check lists being marked as Failed. 784 11. Receiving an End Of Candidates Notification 786 When an agent receives an end-of-candidates notification for a 787 specific check list, they will update its state as per Section 8.1. 788 In case the list is still in the Active state after the update, the 789 agent will persist the the fact that an end-of-candidates 790 notification has been received for and take it into account in future 791 list updates. 793 12. Trickle ICE and Peer Reflexive Candidates 795 Even though Trickle ICE does not explicitly modify the procedures for 796 handling peer reflexive candidates, their processing could be 797 impacted in implementations. With Trickle ICE, it is possible that 798 server reflexive candidates be discovered as peer reflexive in cases 799 where incoming connectivity checks are received from these candidates 800 before the trickle updates that carry them. 802 While this would certainly increase the number of cases where ICE 803 processing nominates and selects candidates discovered as peer- 804 reflexive it does not require any change in processing. 806 It is also likely that, some applications would prefer not to trickle 807 server reflexive candidates to entities that are known to be publicly 808 accessible and where sending a direct STUN binding request is likely 809 to reach the destination faster than the trickle update that travels 810 through the signalling path. 812 13. Concluding ICE Processing 814 This specification does not directly modify the procedures ending ICE 815 processing described in Section 8 of [RFC5245], and trickle ICE 816 implementations will follow the same rules. 818 14. Subsequent Offer/Answer Exchanges 820 Either agent MAY generate a subsequent offer at any time allowed by 821 [RFC3264]. When this happens agents will use [RFC5245] semantics to 822 determine whether or not the new offer requires an ICE restart. If 823 this is the case then agents would perform trickle ICE as they would 824 in an initial offer/answer exchange. 826 The only differences between an ICE restart and a brand new media 827 session are that: 829 o during the restart, media can continue to be sent to the 830 previously validated pair. 832 o both agents are already aware whether or not their peer supports 833 trickle ICE, and there is no longer need for performing half 834 trickle or confirming support with other mechanisms. 836 15. Interaction with ICE Lite 838 Behaviour of Trickle ICE capable ICE lite agents does not require any 839 particular rules other than those already defined in this 840 specification and [RFC5245]. This section is hence added with an 841 informational purpose only. 843 A Trickle ICE capable ICE Lite agent would generate offers or answers 844 as per [RFC5245]. Both will indicate support for trickle ICE 845 (Section 5.1) and given that they will contain a complete set of 846 candidates (the agent's host candidates) these offers and answers 847 would also be accompanied with an end-of-candidates notification. 849 When performing full trickle, a full ICE implementation could send an 850 offer or an answer with no candidates and an IP6 :: connection line 851 address. After receiving an answer that identifies the remote agent 852 as an ICE lite implementation, the offerer may very well choose to 853 not send any additional candidates. The same is also true in the 854 case when the ICE lite agent is making the offer and the full ICE one 855 is answering. In these cases the connectivity checks would be enough 856 for the ICE lite implementation to discover all potentially useful 857 candidates as peer reflexive. The following example illustrates one 858 such ICE session: 860 ICE Lite Bob 861 Agent 862 | Offer (a=ice-lite a=ice-options:trickle) | 863 |---------------------------------------------->| 864 | |no cand 865 | Answer (a=ice-options:trickle) |trickling 866 |<----------------------------------------------| 867 | Connectivity Checks | 868 |<--------------------------------------------->| 869 peer rflx| | 870 cand disco| | 871 | | 872 |<=============== MEDIA FLOWS =================>| 874 Figure 1: Example 876 In addition to reducing signaling traffic this approach also removes 877 the need to discover STUN bindings, or to make TURN or UPnP 878 allocations which may considerably lighten ICE processing. 880 16. Example Flow 882 A typical successful trickle ICE exchange with an Offer/Answer 883 protocol would look this way: 885 Alice Bob 886 | Offer | 887 |---------------------------------------------->| 888 | Additional Candidates | 889 |---------------------------------------------->| 890 | | 891 | Answer | 892 |<----------------------------------------------| 893 | Additional Candidates | 894 |<----------------------------------------------| 895 | | 896 | Additional Candidates and Connectivity Checks | 897 |<--------------------------------------------->| 898 | | 899 |<=============== MEDIA FLOWS =================>| 901 Figure 2: Example 903 17. Security Considerations 905 This specification inherits most of its semantics from [RFC5245] and 906 as a result all security considerations described there remain the 907 same. 909 18. Acknowledgements 911 The authors would like to thank Bernard Aboba, Christer Holmberg, 912 Dale R. Worley, Enrico Marocco, Flemming Andreasen, Jonathan Lennox 913 and Martin Thomson for their reviews and suggestions on improving 914 this document. 916 19. References 918 19.1. Normative References 920 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 921 Requirement Levels", BCP 14, RFC 2119, March 1997. 923 [RFC3264] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model 924 with Session Description Protocol (SDP)", RFC 3264, June 925 2002. 927 [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session 928 Description Protocol", RFC 4566, July 2006. 930 [RFC5245] Rosenberg, J., "Interactive Connectivity Establishment 931 (ICE): A Protocol for Network Address Translator (NAT) 932 Traversal for Offer/Answer Protocols", RFC 5245, April 933 2010. 935 19.2. Informative References 937 [I-D.ivov-mmusic-trickle-ice-sip] 938 Ivov, E., Marocco, E., and C. Holmberg, "A Session 939 Initiation Protocol (SIP) usage for Trickle ICE", draft- 940 ivov-mmusic-trickle-ice-sip-02 (work in progress), June 941 2014. 943 [I-D.keranen-mmusic-ice-address-selection] 944 Keraenen, A. and J. Arkko, "Update on Candidate Address 945 Selection for Interactive Connectivity Establishment 946 (ICE)", draft-keranen-mmusic-ice-address-selection-01 947 (work in progress), July 2012. 949 [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and 950 E. Lear, "Address Allocation for Private Internets", BCP 951 5, RFC 1918, February 1996. 953 [RFC2543] Handley, M., Schulzrinne, H., Schooler, E., and J. 954 Rosenberg, "SIP: Session Initiation Protocol", RFC 2543, 955 March 1999. 957 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 958 A., Peterson, J., Sparks, R., Handley, M., and E. 959 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 960 June 2002. 962 [RFC3388] Camarillo, G., Eriksson, G., Holler, J., and H. 963 Schulzrinne, "Grouping of Media Lines in the Session 964 Description Protocol (SDP)", RFC 3388, December 2002. 966 [RFC3840] Rosenberg, J., Schulzrinne, H., and P. Kyzivat, 967 "Indicating User Agent Capabilities in the Session 968 Initiation Protocol (SIP)", RFC 3840, August 2004. 970 [RFC4787] Audet, F. and C. Jennings, "Network Address Translation 971 (NAT) Behavioral Requirements for Unicast UDP", BCP 127, 972 RFC 4787, January 2007. 974 [RFC5389] Rosenberg, J., Mahy, R., Matthews, P., and D. Wing, 975 "Session Traversal Utilities for NAT (STUN)", RFC 5389, 976 October 2008. 978 [RFC5766] Mahy, R., Matthews, P., and J. Rosenberg, "Traversal Using 979 Relays around NAT (TURN): Relay Extensions to Session 980 Traversal Utilities for NAT (STUN)", RFC 5766, April 2010. 982 [XEP-0030] 983 Hildebrand, J., Millard, P., Eatmon, R., and P. Saint- 984 Andre, "XEP-0030: Service Discovery", XEP XEP-0030, June 985 2008. 987 [XEP-0115] 988 Hildebrand, J., Saint-Andre, P., Troncon, R., and J. 989 Konieczny, "XEP-0115: Entity Capabilities", XEP XEP-0115, 990 February 2008. 992 [XEP-0176] 993 Beda, J., Ludwig, S., Saint-Andre, P., Hildebrand, J., 994 Egan, S., and R. McQueen, "XEP-0176: Jingle ICE-UDP 995 Transport Method", XEP XEP-0176, June 2009. 997 [XEP-0278] 998 Camargo, T., "XEP-0278: Jingle Relay Nodes", XEP XEP-0278, 999 June 2011. 1001 Appendix A. Open issues 1003 At the time of writing of this document the authors have no clear 1004 view on how and if the following list of issues should be addressed. 1006 A.1. MID/Stream Indices in SDP 1008 This specification does not currently define syntax for candidate-to- 1009 stream bindings although it says that they should be implemented with 1010 MID or a stream index. Yet, it is reasonable to assume that most 1011 usages would need to do this within the SDP and it may make sense to 1012 agree on the format. Here's one possible way to do this: 1014 a=mid:1 1015 a=candidate:1 1 UDP 1658497328 192.168.100.33 5000 typ host 1016 a=candidate:2 1 UDP 1658497328 96.1.2.3 5000 typ srflx 1017 a=mid:2 1018 a=candidate:2 1 UDP 1658497328 96.1.2.3 5002 typ srflx 1019 a=end-of-candidates 1021 A.2. Starting checks 1023 Normally Vanilla ICE implementations would first activate a check 1024 list, validate at least one pair in every component and only then 1025 unfreeze all other checklists. With trickle ICE this would be 1026 suboptimal since, candidates can arrive randomly and we would be 1027 wasting time waiting for a checklist to fill (almost as if we were 1028 doing vanilla ICE). We need to decide if unfreezing everything 1029 solely based on foundation is good enough. 1031 Appendix B. Changes From Earlier Versions 1033 Note to the RFC-Editor: please remove this section prior to 1034 publication as an RFC. 1036 B.1. Changes From draft-ivov-01 and draft-mmusic-00 1038 o Added a requirement to trickle candidates by order of components 1039 to avoid deadlocks in the unfreezing algorithm. 1041 o Added an informative note on peer-reflexive candidates explaining 1042 that nothing changes for them semantically but they do become a 1043 more likely occurrence for Trickle ICE. 1045 o Limit the number of pairs to 100 to comply with 5245. 1047 o Added clarifications on the non-importance of how newly discovered 1048 candidates are trickled/sent to the remote party or if this is 1049 done at all. 1051 o Added transport expectations for trickled candidates as per Dale 1052 Worley's recommendation. 1054 B.2. Changes From draft-ivov-00 1056 o Specified that end-of-candidates is a media level attribute which 1057 can of course appear as session level, which is equivalent to 1058 having it appear in all m-lines. Also made end-of-candidates 1059 optional for cases such as aggressive nomination for controlled 1060 agents. 1062 o Added an example for ICE lite and trickle ICE to illustrate how, 1063 when talking to an ICE lite agent doesn't need to send or even 1064 discover any candidates. 1066 o Added an example for ICE lite and trickle ICE to illustrate how, 1067 when talking to an ICE lite agent doesn't need to send or even 1068 discover any candidates. 1070 o Added wording that explicitly states ICE lite agents have to be 1071 prepared to receive no candidates over signalling and that they 1072 should not freak out if this happens. (Closed the corresponding 1073 open issue). 1075 o It is now mandatory to use MID when trickling candidates and using 1076 m-line indexes is no longer allowed. 1078 o Replaced use of 0.0.0.0 to IP6 :: in order to avoid potential 1079 issues with RFC2543 SDP libraries that interpret 0.0.0.0 as an on- 1080 hold operation. Also changed the port number here from 1 to 9 1081 since it already has a more appropriate meaning. (Port change 1082 suggested by Jonathan Lennox). 1084 o Closed the Open Issue about use about what to do with cands 1085 received after end-of-cands. Solution: ignore, do an ice restart 1086 if you want to add something. 1088 o Added more terminology, including trickling, trickled candidates, 1089 half trickle, full trickle, 1091 o Added a reference to the SIP usage for trickle ICE as requested at 1092 the Boston interim. 1094 B.3. Changes From draft-rescorla-01 1096 o Brought back explicit use of Offer/Answer. There are no more 1097 attempts to try to do this in an O/A independent way. Also 1098 removed the use of ICE Descriptions. 1100 o Added SDP specification for trickled candidates, the trickle 1101 option and 0.0.0.0 addresses in m-lines, and end-of-candidates. 1103 o Support and Discovery. Changed that section to be less abstract. 1104 As discussed in IETF85, the draft now says implementations and 1105 usages need to either determine support in advance and directly 1106 use trickle, or do half trickle. Removed suggestion about use of 1107 discovery in SIP or about letting implementing protocols do what 1108 they want. 1110 o Defined Half Trickle. Added a section that says how it works. 1111 Mentioned that it only needs to happen in the first o/a (not 1112 necessary in updates), and added Jonathan's comment about how it 1113 could, in some cases, offer more than half the improvement if you 1114 can pre-gather part or all of your candidates before the user 1115 actually presses the call button. 1117 o Added a short section about subsequent offer/answer exchanges. 1119 o Added a short section about interactions with ICE Lite 1120 implementations. 1122 o Added two new entries to the open issues section. 1124 B.4. Changes From draft-rescorla-00 1126 o Relaxed requirements about verifying support following a 1127 discussion on MMUSIC. 1129 o Introduced ICE descriptions in order to remove ambiguous use of 1130 3264 language and inappropriate references to offers and answers. 1132 o Removed inappropriate assumption of adoption by RTCWEB pointed out 1133 by Martin Thomson. 1135 Authors' Addresses 1137 Emil Ivov 1138 Jitsi 1139 Strasbourg 67000 1140 France 1142 Phone: +33 6 72 81 15 55 1143 Email: emcho@jitsi.org 1145 Eric Rescorla 1146 RTFM, Inc. 1147 2064 Edgewood Drive 1148 Palo Alto, CA 94303 1149 USA 1151 Phone: +1 650 678 2350 1152 Email: ekr@rtfm.com 1154 Justin Uberti 1155 Google 1156 747 6th St S 1157 Kirkland, WA 98033 1158 USA 1160 Phone: +1 857 288 8888 1161 Email: justin@uberti.name