IETF CLUE WG Virtual Interim meeting May 29, 2014 15:00-17:00 Central US Chaired by Mary Barnes and Paul Kyzivat Attendees: ------------------ Alissa Cooper Keith Drage Mark Duckworth Roni Even Christian Groves Rob Hansen Christer Holmberg Jonathan Lennox John Leslie Matt Ryan Minutes by Mary Barnes with consensus/actions summarized by Christer Holmberg. Recording: https://ietf.webex.com/ietf/ldr.php?RCID=40ad666b0d5da9dfdcd5527f54d1c03b --------- Summary: --------- The objective of the meeting was to review the current WG status, focusing discussion on issue resolution and details of the signaling solution. The discussions and actions are summarized below, followed by Christer's notes. Status Review: =============== - Use Cases: Published as RFC 7205. - Requirements: In AUTH48. - Framework: The audio issue(s) that has been discussed on the mailing list and at the May xx design team meeting is the primary issue whose resolution could impact the framework. The issue with regards to consistency with data model will remain open until the data model is deemed near completion. Mark noted there are a couple other issues marked as Editor's notes in the document. Actions: 1) Mark: send the details to chairs on editor's notes so ticket(s) can be opened. 2) Chairs: Since the document is very near completion, the chairs will start a 2 week pre-WGLC to ensure there are no other outstanding issues to be resolved prior to forwarding for publication. - Data Channel: There was no specific discussion around the data channel (which was previously discussed at the design team meeting). The primary issues are related to dependencies in other WGs. - CLUE Protocol: There was no specific discussion around the CLUE protocol. The document is under consideration as a WG document (review to close on June 10th). Once the data model is finalized, the CLUE protocol document can be completed, with final changes coordinated with the signaling document. - Data Model: There was no discussion of the data model at the meeting. Actions: 1) Mary: follow-up with Roberta on the current status of data model. 2) Chairs: schedule a discussion of the data model open issues (#48 and #35) at an upcoming design team meeting if deemed necessary. - Call Flows: Rob and Roni agreed to work on a first pass at this. Some material can be pulled from the signaling document as well as from Simon's demo slides from IETF-89. Roni will start with a simple use case to be reviewed at an upcoming design team call. Way Forward =========== - Update wiki with proposed topics for upcoming design team meetings. - Chairs to start pre-WGLC on framework. - Individuals with assigned action items to complete those within the next few weeks (no later than June 28th to allow time for documents to be updated prior to IETF-90) Detailed Discussions: ===================== Signaling (Rob Hansen) ----------------------- - Group restrictions. There is already some text. At this time, this isn't something that should be done. However, it's possible that this might be useful in the future, thus the wording needs to be along the lines of MUST NOT…unless OR SHOULD NOT…unless. - CLUE controlled media. There's a restriction on sending CLUE controlled m-lines until the same has been negotiated in CLUE. Conclusion: use a=inactive - CLUE & SDP state machine interaction. Some changes are non-atomic. The participant must be able to cope with a transitory state. Conclusion: General agreement that we need some recommendations (along the lines of slide 7). Action: WG to discuss proposed text on mailing list. With regards to specifics: - inconsistencies are transitory - is this a "SHOULD be" or "MUST be". Can't really defi - don’t request anything until details are available in both CLUE and SDP. In most situations this won't require and additional O/A. - send changes in CLUE before SDP. Review of open action items: ---------------------------- The following action items from the May 13th, 2014 design team meeting remain open: 1) John L to propose a follow-up action for the Issue Tracker on separate mixer inputs. 2) Paul K to issue a call on the mailing list for consensus on what to do about echo cancellation. 3) TP vendors to provide an example advertisement listing their capabilities. All the other action items are either OBE or work underway. Review of open tickets: ----------------------- - #21: close - #35: Keep open until FW is complete and consistency with data model is achieved. - #47: Remains open. Christer to follow-up with Bo and Magnus to see if they plan to complete Simulcast document. - #11: should be updated to indicate that AppID needs to be considered. Is it sufficient for CLUE or do we need something else? Action: Jonathan to delve into this. - #12: close. - #29: close. - #25: Christian is still working on this - #27: close. - #49: close (state machines are not coupled) - #24 & #30: These relate to the use of BFCP. No one on the call was interested in looking at any possible interactions or usage of such. Thus, chairs will post a note to the mailing list. If no one is willing to investigate issue (within next 2 weeks) ticket will be closed as it's not within the obvious scope of CLUE. In addition, we need to open a new ticket with regards to the use of BUNDLE with CLUE. Consensus and Action item summary by Christer Holmberg: ======================================================= Topic: Call Flows AP: Roni E is going to start working on call flows, which will be placed in a dedicated draft. AP: Christer H will follow the 3GPP work on call flows, to see whether some call flows can be re-used between the organizations. Topic: Clue Signalling AP: Rob H will, in the signaling draft, describe handling of media channels for which CLUE content has not yet been negotiated. Such media channels will be identical to “inactive” media channels, which means that, while media content will not be sent, protocols like RTCP, DTLS and ICE will be sent. Topic: Tickets Ticket #21 closed. Ticket #12 closed. Ticket #29 closed. Ticket #23 closed. Ticket #27 closed. AP: Mary B will send an e-mail to the CLUE list, and suggest ticket #30 to be closed.