Friday, August 2, 2013< ^ >
mark.jones has set the subject to: "dime wg meeting in Vancouver"
Room Configuration
Room Occupants

[00:49:05] Andrew Booth joins the room
[08:50:33] Andrew Booth joins the room
[09:19:45] joins the room
[09:21:32] benoit.claise joins the room
[09:21:52] benoit.claise has set the subject to: "dime wg meeting in Berlin"
[09:22:35] stefan.winter joins the room
[09:22:48] <stefan.winter> Hi, I'm the jabber scribe for the first dime slot.
[09:22:54] <stefan.winter> Agenda bashing started
[09:23:17] jgunn joins the room
[09:26:25] <stefan.winter> Diameter Group Signalling presentation started
[09:26:39] <stefan.winter> Slide: "Motivations"
[09:26:57] <stefan.winter> Slide: Two Problem Aspects
[09:27:23] <stefan.winter> Slide: Doc history and WI status
[09:28:33] <stefan.winter> Slide: Summary of 1st rev
[09:29:06] <stefan.winter> Slide: intended status and scope
[09:29:27] <stefan.winter> Slide: Introduction section
[09:30:53] <stefan.winter> Slide: section Grouping User Sessions
[09:33:40] <stefan.winter> Slide: Protocol Description
[09:36:18] <stefan.winter> Slide: ... Protocol Description #3
[09:37:35] <stefan.winter> Slide: Attribute-Value Pairs
[09:38:58] <stefan.winter> People in the room note that the audio bridge seems not to work?
[09:38:59] <jgunn> The audio is working for me
[09:39:03] <Andrew Booth> I can hear you...
[09:39:07] <stefan.winter> Anyone on Jabber can confirm?
[09:39:12] Hannes Tschofenig joins the room
[09:39:25] <jgunn> AUDIO IS WORKING FINE
[09:39:36] <stefan.winter> Thanks
[09:39:44] <Hannes Tschofenig> Now it is working and it did not work with the other microphone?
[09:40:09] <jgunn> It was working for me with the other mic
[09:40:12] <stefan.winter> Slide: Commands Formatting
[09:40:24] <stefan.winter> Slide: open items
[09:43:37] <stefan.winter> Slide: Open items (2)
[09:44:27] <stefan.winter> Slide: Next steps
[09:45:57] <stefan.winter> Mic: Ben Campbell: What about the use cases for mid-session changes? Seems expensive operation.
[09:46:17] <stefan.winter> Leaving out mid-session would remove a lot of complexity
[09:46:40] <stefan.winter> Hannes: be clearer about the use case...
[09:47:44] <stefan.winter> Ban: we should only include things which are useful - not sure about that at this point.
[09:48:18] <stefan.winter> Commenter N.N: again use case question: why client initiated?
[09:48:52] <stefan.winter> (i.e. why should a client, not only the server, be able to assign sessions to groups)
[09:49:28] <stefan.winter> Lionel: use case exists, e.g. in asynchronous operation where client needs to operate on its own
[09:50:18] <stefan.winter> add text about the use case to make clear
[09:51:11] <stefan.winter> Jouni: capability negotiation "supported features": there is a 3GPP equivalent mechanism, are you replicating that?
[09:51:14] <stefan.winter> Lionel: no.
[09:51:39] <stefan.winter> Also is only a possibility, not sure if it is going to be done.
[09:52:17] <stefan.winter> Hannes: other diameter apps do similar things; this would not be revolutionary
[09:52:36] <stefan.winter> So can do - maybe a different name for the AVP would be appropriate, it clashes.
[09:54:33] <stefan.winter> ### next presentation: Ben Campbell: Overload Control
[09:55:18] <stefan.winter> Slide: Issues Discussion
[09:56:30] <stefan.winter> Slide: Non-Adjacent Overload Control
[09:57:09] <stefan.winter> Slide: Non-Adj Use Cases
[09:58:05] <stefan.winter> Marten from AT&T: no sharing of overload control between carriers. Trust model between carries doesn't make this fly.
[09:58:13] <stefan.winter> Other mic comment: seconded.
[09:59:49] <stefan.winter> Hannes: load information - agreed; but for *overload* something else - you want others to know about it, because they should send less to you.
[10:02:16] <stefan.winter> So send "dropping packet" error report to client; but for simple re-routing for load-balancing, no need to tell client.
[10:03:02] <stefan.winter> Ben clarifies: this issue is only when there is a third-party between client and server which does not support overload control.
[10:06:05] <jgunn> what slide?
[10:06:45] <stefan.winter> Slide: Non-supporting agent
[10:07:26] <stefan.winter> Slide: Non-Supporting Agent Chacteristics
[10:10:50] <stefan.winter> Slide: Non-Adj Topology Issues
[10:13:58] <stefan.winter> Hannes: there is a way to specify source-to-dest all along the path
[10:14:05] <stefan.winter> Ben: in common use?
[10:14:16] <stefan.winter> Hannes: topology is usually not so complicated to need this
[10:14:26] <stefan.winter> Ben: disagree, hidden topologies are common
[10:15:24] <stefan.winter> Hannes: Is there an expectation to control paths inside third-party domain?
[10:15:43] <stefan.winter> So: non-use-case
[10:17:40] <stefan.winter> Topologies exist; client doesn't see them; and doesn't know what his choice of next-hop actually triggers elsewhere.
[10:19:59] <stefan.winter> Hannes: previous agent will know if its own next hop has problems, and can fix those - no need for signalling to client
[10:20:54] <jgunn> MIC - if you try to control the overlaod solely form the server, you get congestion collapse.  The resources needed to "reject" the messages overwhelm the  server.
[10:21:34] <stefan.winter> Hannes: carriers' probelms seems to be overloaded servers, not intermediate agents. Load-balancing can fix this.
[10:25:18] <stefan.winter> waiting in Q...
[10:25:37] <jgunn> thanks
[10:26:22] <stefan.winter> who are you?
[10:26:25] <jgunn> Janet
[10:26:31] <jgunn> Janet Gunn
[10:27:02] <jgunn> That is based on the analysis and modeling done for the SIP overload work, presented at an eiarier IETF meeting
[10:27:29] <Hannes Tschofenig> SIP overload is different than Diameter overload because SIP is just entirely different
[10:28:09] <stefan.winter> not sure I can make another comment... people are going way over time
[10:28:26] <stefan.winter> Slide: Non-Adj OC Scopes
[10:28:30] <Hannes Tschofenig> That's the problem in just applying the SIP stuff to this work. So, I encourage to be careful here.
[10:28:42] <Hannes Tschofenig> (And: The SIP overload work is not particularly successful.)
[10:29:05] <stefan.winter> Slide: OVerload scopes
[10:29:12] <jgunn> That is OK,  I earleir sent an email on Hannes draft that expresses my concerns in more detail
[10:30:43] <Hannes Tschofenig> Thanks. That's helpful
[10:31:13] stefan.winter leaves the room
[10:31:41] <> starting point slide
[10:32:26] <> new proposal slide
[10:32:52] <> slide: Complexity
[10:33:11] <> slide: Principles
[10:34:07] <> Ben at mic
[10:36:42] <> Eric at mic
[10:38:54] <jgunn> Name of verizon speaker?
[10:39:03] <> I don't know
[10:39:13] <> I'll get it later.
[10:39:32] <jgunn> thanks
[10:40:01] <> Martin at mic
[10:41:09] <> Ben
[10:42:09] <> back on principles slide
[10:42:31] <> Eric at mic
[10:44:23] <> Karilani at mic
[10:44:36] <> sp?
[10:44:58] <> Lionel from the floor
[10:45:55] <jgunn> Her name is Kalyani  Bogineni - looked it up in the attendees list
[10:46:31] <> Eric
[10:47:37] <> still on Principles slide
[10:48:29] <> slide  - Overload Signal
[10:48:46] <> slide Load balancing
[10:49:00] <> Information Model
[10:49:15] <> Overload Communication Basic Design Options
[10:49:42] <> Getting Implementation - skipped
[10:50:35] <> Way forward slide
[10:53:18] <> Eric
[10:56:38] <> Glen at mic
[10:57:38] <> Hannes behind Glen
[10:58:59] <> Hannes at mic
[11:00:01] <> Keith
[11:02:54] <> Ben at mic
[11:04:24] <> Martin at mic
[11:11:05] <> Kalyani at mic
[11:11:48] <> lionel
[11:12:04] <> hannes at mic
[11:13:20] <> Ben at mic
[11:15:24] <> Martin at mic
[11:18:14] <> Hannes at mic
[11:18:32] <> Ben at mic
[11:19:19] <> Ben at mic
[11:20:04] <> Keith at mic
[11:21:35] <> slide - Motivation
[11:21:49] benoit.claise leaves the room
[11:21:53] <> Cathy at the mic
[11:22:57] <> Parameters Required by each transition method
[11:23:11] <> attribute-value pair defined
[11:24:15] <> attribute-value pair defined
[11:24:37] <> Next step
[11:24:48] <> Lionel from the floor
[11:26:10] <> attribute-value pair defined
[11:28:12] Hannes Tschofenig leaves the room
[11:29:35] <> martin
[11:30:45] <> and that's all folks. Sorry I wasn't the best jabber scribe, I was also the note taker
[11:30:55] <Andrew Booth> Thanks
[11:32:21] benoit.claise joins the room
[11:48:57] Andrew Booth leaves the room
[11:59:52] benoit.claise leaves the room
[12:04:30] benoit.claise joins the room
[12:07:19] leaves the room
[12:09:19] joins the room
[12:12:36] leaves the room
[12:13:55] jgunn leaves the room
[12:40:02] benoit.claise leaves the room
[13:05:18] jgunn joins the room
[13:06:15] jgunn leaves the room