ecrit - Monday, July 20, 2015

ecrit

ecrit@jabber.ietf.org

Monday, July 20, 2015< ^ >

martin.thomson has set the subject to: ECRIT

Room Configuration

Room Occupants

GMT+0
[16:45:43] Meetecho joins the room
[16:47:49] Paul Kyzivat joins the room
[16:50:08] marc linsner joins the room
[16:50:42] Simon Romano joins the room
[16:56:02] Alissa Cooper joins the room
[16:56:09] <Alissa Cooper> Agenda bashing now
[16:56:19] <Alissa Cooper> Document status
[16:57:32] marc linsner leaves the room
[16:58:10] <Alissa Cooper> Randy talking about additional-data
[16:59:45] marc linsner joins the room
[17:00:20] <Alissa Cooper> Brian: NENA is waiting for this and implementors are waiting for it
[17:00:46] <Alissa Cooper> … discussion within NENA about transport of additional data outside of SIP calls
[17:02:46] <Alissa Cooper> Alissa: will read the latest diff after Randy submits it today and then will go to IETF LC
[17:02:59] <Alissa Cooper> Randy talking about eCall
[17:03:48] <Alissa Cooper> Randy talking about car-crash
[17:05:24] <Alissa Cooper> Roger: what were Keith's concerns at IETF 92?
[17:05:52] <Alissa Cooper> Randy: Concerns were that multiple call setup mechanisms could cause confusion. Fix was for both drafts to share same mechanism with different data sets.
[17:06:14] <Alissa Cooper> … also had a liaison statement to 3GPP SA-1, SA-2, and CT-1
[17:06:21] marc linsner leaves the room
[17:06:26] <Alissa Cooper> Andrew: SA-2 had a meeting last week
[17:06:40] <Alissa Cooper> Roger: want to hear back from 3GPP before proceeding with drafts
[17:06:50] <Alissa Cooper> Randy: want to hear concerns or changes needed
[17:06:59] marc linsner joins the room
[17:07:03] <Alissa Cooper> Brian talking about lost-planned-changes
[17:08:04] ben joins the room
[17:09:54] marc linsner leaves the room
[17:11:20] <Alissa Cooper> Roger: need some reviewers
[17:11:46] <Alissa Cooper> Randy talking about SLIM
[17:13:23] Keith Drage joins the room
[17:14:35] Keith Drage leaves the room
[17:19:32] <Alissa Cooper> Christer: why don't we use existing media tags in SIP? unaware of evolution in thinking
[17:20:24] <Alissa Cooper> Henning: don't appreciate declaring consensus when there isn't one
[17:20:40] <Alissa Cooper> … two problems: 1) negotiation problem, 2) call routing problem
[17:22:22] <Alissa Cooper> … not an either/or proposition. emergency calling and single call center cases are different. in first, may have dedicated interpreter in one state, separate logical entity from PSAPs. in second, have single SBC/PBX that does queueing mgmt, resource assignment, etc. doesn't matter which way you do this because entity looks at SIP and SDP. in first case, mucking around with SDP for call routing is against SIP architecture.
[17:22:54] <Alissa Cooper> … once it's routed to your only call center and it doesn't work there, you have a bigger problem
[17:22:56] marc linsner joins the room
[17:23:51] <Alissa Cooper> … not a contradiction problem beyond an implementation bug. simple cases can and should be done as a call routing function.
[17:24:00] <Alissa Cooper> Bernard: second what Henning said
[17:24:30] <Alissa Cooper> … distinction between emergency case and non-emergency case. someone speech impaired claling a barber shop needs to have call go through.
[17:25:33] <Alissa Cooper> … some disability call flows are complex — multiple media. speech/hearing impaired combination. RTT using English, ASL using sign language. call routing aspect is distinct. can be many different deployments. sending to PSAP without interpreters or any capability at all — you want that call to succeed.
[17:25:46] <Alissa Cooper> … need full problem statement and review form disability community
[17:26:06] <Alissa Cooper> … not saying language tag is wrong but needs to be done in a group with expertise like this one
[17:26:14] <Alissa Cooper> … not in same group doing email at same time
[17:26:30] <Alissa Cooper> Brian: no difference whatsoever between emergency case and non-emergency case.
[17:26:38] <Alissa Cooper> … no one wants to go through two-step dialing
[17:26:49] <Alissa Cooper> … but can't do that now because they can't label their calls now
[17:27:10] <Alissa Cooper> … need to know every piece of information to make routing decision
[17:27:25] <Alissa Cooper> … you could replicate this syntax of media-language relationship in the signaling, but why?
[17:27:34] <Alissa Cooper> Henning: because that is the SIP architecture
[17:27:45] <Alissa Cooper> … making same mistake twice does not count as progress
[17:28:10] <Alissa Cooper> Brian: charter the particular work that just deals with labeling the media and does not deal with routing. then do a different effort that deals with routing.
[17:28:30] <Alissa Cooper> … getting media labeled is useful independent of routing
[17:28:37] <Alissa Cooper> Bernard: agree with Brian on that piece
[17:28:46] <Alissa Cooper> … draft tries to do the whole picture
[17:29:02] <Alissa Cooper> … there is a very big need to solve the bigger problem
[17:29:28] <Alissa Cooper> Brian: to the contrary, right now we know exactly how we route calls for emergencies: the same way as any other call
[17:29:40] <Alissa Cooper> … that is what we would do with or without this draft
[17:29:51] <Alissa Cooper> … there is a desire to change that by some folks, need mechanisms to do that
[17:30:04] <Alissa Cooper> … could still use this draft within the PSAP
[17:30:13] <Alissa Cooper> … would not handle relay
[17:31:00] <Alissa Cooper> Henning: general advantage is that because you don't have fine-grained state knowledge when doing coarse-grained routing, want to go to high probability choice of being right
[17:31:30] <Paul Kyzivat> mic: this draft could eliminate "press 1 for Spanish" in the world. That would make *many* people happy.
[17:32:01] <Alissa Cooper> Henning: this going to be a pref indicator of the caller
[17:32:08] <Alissa Cooper> … need back up so call goes through
[17:32:18] <Alissa Cooper> … preference indicator unlike media indicator
[17:32:48] <Alissa Cooper> … fundamentally different from media negotiation
[17:33:00] <Alissa Cooper> … this decision is "if possible, optimize call behavior"
[17:34:10] <Alissa Cooper> Ban: easier way to solve it is for psap to be able to choose relay service
[17:34:22] <Alissa Cooper> Brian: today we can't do that in the US
[17:35:11] <Alissa Cooper> … it is a negotiation. but you accept what you get even if it's sub-optimal.
[17:35:21] <Alissa Cooper> … arrange a translation service. language line.
[17:35:46] <Alissa Cooper> … question is how does the psap do that? has no effect on routing outside
[17:36:11] <Alissa Cooper> … getting language line on the call takes minutes
[17:36:19] <Alissa Cooper> … they accept the call as-is
[17:37:12] <Alissa Cooper> Randy: this work has been going on for many years, nobody cared SIP vs. SDP. we re-wrote the draft entirely. the draft has been fair.
[17:37:25] <Alissa Cooper> … nobody was surprised
[17:37:56] <Alissa Cooper> Brian: and more than anything, it's been going on for 2 years and there's only 1 draft
[17:38:15] <Alissa Cooper> Bernard: we have been discussing aspects of this, but what has not happened is discussion of the context of using this
[17:38:29] <Alissa Cooper> … separate the work here from the large context
[17:38:46] <Alissa Cooper> Brian: send text
[17:39:19] <Alissa Cooper> Christer: did not want to throw away SDP vs. something else. have asked why we can't use existing mechanisms. why can't we use existing mechanisms, why don't they work?
[17:40:44] <Alissa Cooper> Alissa: good warm up for dispatch
[17:40:57] <Alissa Cooper> Randy: and there will be a working lunch on Weds
[17:41:21] <Alissa Cooper> Ban talking about total conversation for emergency communications
[17:45:44] <Paul Kyzivat> If this is *total* conversation, why doesn't it have smellovision?:p:p:p
[17:46:05] <Alissa Cooper> that would be helpful for emergencies
[17:49:13] <Paul Kyzivat> Maybe, instead of "total communication" it should just be called "triple play" communication.
[17:54:14] <Paul Kyzivat> mic: which is more important: trust by PSAP, or trust by calling user?
[17:54:33] <Alissa Cooper> Brian: look at NENA document, it deals with many of these issues.
[17:54:48] <Alissa Cooper> … additional-data designed for any entity to be able to add data, incl. relays etc.
[17:55:12] <Alissa Cooper> … i3 0803
[17:55:26] ben leaves the room
[17:56:27] Meetecho leaves the room
[17:56:49] Paul Kyzivat leaves the room
[17:57:13] marc linsner leaves the room
[18:06:04] Simon Romano leaves the room
[18:09:58] Alissa Cooper leaves the room
[21:25:49] Alissa Cooper joins the room