INSIPID

Date: Wednesday August 1st @ 1:00pm

Room: Regency A

 

Note takers: Miguel Garcia, Andrew Allen

Jabber Scribe: Adam Roach

 

Introduction & Status Update - Chairs: Keith Drage & Gonzalo Salgueiro

-       Chairs reviewed the agenda for the meeting.

 

-       Chairs reviewed the WG status slide and provided an update on both WG milestones.

 

-       Chairs noted that the cadence of virtual interims would need to continue in order to stick to the tight timeline of the WG milestone deadlines. Chairs encouraged attendance and participation at the virtual interims and on the mailing list in order to avoid having to continually rehash issues where consensus was already reached.

draft-ietf-insipid-session-id-reqts-00.txt - James Polk

-       Reviews of the retained and new use cases text from the requirements draft was presented.

 

-       There was support for keeping the text regarding the Session Recording (3.1) use case.

 

-       There was support for keeping the text dealing with the Protocol Interworking (3.2) use case.

 

-       Chairs encouraged the group to focus the discussion on the use case itself and not on wordsmithing the text.

 

-       During the discussion on use case for “End-to-end identification of a communication session” Hadriel Kaplan brought up the discussion of B2BUA versus SBC. It is suggested that the words “or SBCs” be removed from the use case text. The point is also raised that the last sentence of this use case refers to a Session-ID persisting a dialog. Andrew Hutton further clarifies what a dialog is, indicating that a B2BUA changing the From or the To tags will have two different dialogs, and this is not what the last sentence conveys. It is agreed that the last sentence needs to be cleaned up to state that the same identifier must exist across multiple dialogs within a communication session. Chairs suggest taking the actual wordsmithing of that use case to the list. James Polk agrees and takes the action item to do so.

 

-       In the course of discussing the use case for “Identification of devices taking part in the same multipoint Conference” Hadriel Kaplan stated that he believes this use case is out of scope for this WG. Mary Barnes supported the position and indicated that a separate identifier is not needed for this because you already have the conference ID. James Polk indicated that this is needed for associating with other protocols. Paul Kyzivat commented that it might be difficult to find a solution that satisfies all the requirements. Charles Eckel mentions that he thinks it is a reasonable use case but that it appears outside the scope of the current charter. Christer Holmberg says this is a use case that cannot be solved and the WG should not take it on. Adam Roach (as Jabber scribe) channeling Paul Jones states that when the caller calls into the conference server it is like a transfer to the actual conference. Mary insisted on the existence of the conference identifier for the purpose of identifying conferences. Different conferences have different conference ids. Hadriel indicated he is OK with keeping the requirement and discussing this later when the solution proposal is presented. Andrew Hutton thought this use case is not within scope of the current charter. Keith Drage (as individual) was not of the mind that this conferencing use case exceeded the constraints of the charter. Mary Barnes highlighted the trouble with accepting a requirement for which a solution will not be agreed to, but is OK to go with the requirement. Gonzalo Salgueiro (as individual) indicates that the current use case text only refers to a simple (single star) conference with single focus based multipoint conference.  Adam Roach is concerned with the requirement applying to only one type of conference. He feels that the requirement should cover all types of conferences.  Keith Drage (as chair): asks the floor whether the working group should address this problem. Gonzalo Camarillo (as AD) clarifies the point that the charter is narrow in focus and tries to refocus the discussion meet that objective. Gonzalo Camarillo (as AD) clarifies that the purpose of the charter was to restrict the scope of the WG. He suggests keeping this use case out, because it is outside the charter; perhaps it can be considered at a later time.  James expressed some concern regarding not tackling the conference use case now. Gonzalo Camarillo (as AD) clarifies that working groups typically start with a restricted charter and are re-chartered as needed. The chairs suggest taking the discussion offline and take the action item to have a discussion between chairs and AD to settle this point of whether the conferencing use case will be taken on by the WG or not. Chairs will come back to the list with the outcome of this meeting.

 

-       The call transfer use case was presented and Partha indicates that it is possible to have two different IDs, say, call transfer and call join. James Polk claims he is willing to make this only about call transfer. Keith Drage (as chair) indicates that transfer might also be outside the scope of the charter, as a result of the previous discussion regarding the conferencing use case.  James Polk highlights that the no one has questioned requirements where the session ID was changing.  Hadriel Kaplan indicates that he wants the WG to consider this use case. Keith Drage (as chair) indicates consensus for accepting the transfer use case, though the current text will require a bit of wordsmithing.

 

-       Chairs indicate there is consensus to accept the Session Signal Logging use case.

 

-       Hadriel Kaplan prefers to eliminate the RSVP requirement and use case as it could lead to some bad things. Keith Drage (as chair) mentions that H.323 is specifically mentioned in the charter and hence requirement 9A should remain. James Polk clarifies that the 9A, 9B and 9C requirements are just for awareness and the procedure for any of these protocols will not be specified. Gonzalo Salgueiro (as individual) states that these can simply be formatting requirements without the need to specify RSVP, RTCP, etc.  Gonzalo Camarillo (as AD) that we should be careful not to boil the ocean and stick to the narrow problem space laid out by the charter. He further comments that if we begin to overload this we will find ourselves in the same place we were years ago where we couldn’t charter this work because we couldn’t come to an agreement. Andy Hutton adds that it is a bad idea to talk about media. We should limit the discussion to session establishment protocols. This use case will not be considered by the WG.

 

-       Similarly, it was argued that the RTCP use case and requirement (9C) should also not be considered by the WG. There didn’t seem to be string consensus either way on this one (vote of 4 in favor and 2 against). Gonzalo Camarillo asks if these requirements are purely about formatting. Answer is yes, they are. Also to be discussed by the chairs and AD.

 

-       There was clear support in favor of the WG retaining the Traffic Monitoring use case.

 

-       The final use case being considered is the 3PCC use case. James Polk states that the current Session ID solution proposal will work for the call flow in Figure 1 in Section 4.1 of RFC 3725. He also verified it would work for four others. Paul Kyzivat indicates that there is many more 3PCC call flows to consider and would need to be checked individually and in detail. Andy Hutton comments this is a common use case, and it is needed. Decision is to keep requirement 10 and its associated use case. The rest of the 3PCC call flows will require discussion on the list.

 

-       As James presented the summary of changes to requirements list it was decided that the absence of requirement 1A would be discussed on the list.

draft-jones-insipid-session-id-00.txt – James Polk

-       As time was running short James Polk quickly presented the proposed session-ID solution in draft-jones-insipid-session-id-00.txt. The ‘JOIN’ case was heavily debated and Hadriel Kaplan commented the intent was to trace on a leg-by-leg basis. Time ran short and there was no time to cover how the proposed solution would look in the conferencing case.

 

-       James Polk quickly covered the construction of the Session-ID using UUIDs “half keys” concatenated together. There is ongoing debate on the list regarding the ordering of the “half-keys”.

 

-       Chairs indicate that so far this is the only solution that has come forward. Question was asked if the group feels that this is the way forward or whether we need to be looking in a different direction. Gonzalo Salgueiro (as chair) indicated time is running short on submitting new solution proposals since the WG will need to adopt something soon.

 

-       Christer Holmberg is in favor of the proposed solution but requested that the WG make an effort to make it backwards compatible with the Kaplan draft since that is deployed by several for quite some time now. Gonzalo Salgueiro (as individual) indicated he wasn’t sure how this could be managed since the construction of the ID is entirely different. Christer will think on it and take it to the list.

Closing & Summary - Chairs: Keith Drage & Gonzalo Salgueiro

-       Keith Drage (as chair) indicates that the requirements drafts should go to WGLC fairly shortly.

 

-       Keith Drage (as chair) says the WG will need to adopt a solution soon and in order to keep to the timeline will need to do additional virtual interims.