[07:25:36] --- anewton has become available
[07:36:12] <anewton> ecrit@ietf.xmpp.org
[07:37:08] --- spencerdawkins has become available
[07:38:27] --- rogermarshall has become available
[07:38:36] --- anabolism has become available
[07:42:04] <anewton> discussing agenda for the day?
[07:42:31] --- anabolism has left
[07:42:53] --- anabolism has become available
[07:43:01] <anewton> discussing milestones
[07:46:22] --- hardie has become available
[07:47:21] --- jon-ietf has become available
[07:49:23] --- jon-ietf has left: Replaced by new connection
[07:50:50] <anewton> current discussion is about revisions of the requirements and security requirements document. Current thoughts are that there will be revisions RSN and another in a couple of weeks. The intent is to have the drafts for wglc published before cut-off of the next IETF plenary meeting.
[07:53:02] <anewton> Planning a wglc for Feb. 20 on requirements document. Henning also promises to revise the service SOS draft.
[08:00:44] <anewton> Talk about wglc of the service SOS draft. Andy brought up the fact that accepting it as wg item and then wglc on it so quickly might look "funny".
[08:01:16] <hardie> I think the two items above are actually the the urn service draft, not the sos draft.
[08:01:33] <anewton> umm...yeah. sorry.
[08:05:32] <anewton> march 2006 for requirements document
[08:06:01] <anewton> march 2006 for urn sos document
[08:06:11] <anewton> may 2006 for security document
[08:11:51] <anewton> discussion on the bcp re strategy for associating session originators with physical locations, describing how to route an emergency call, and how to discover media stream types
[08:17:30] <anewton> Andy proposed that we should only have milestones that have an I-D that speaks to the item, either a wg or individual I-D.
[08:20:06] <anewton> With regard to putting a milestone for a grand-view architecture draft, Jon says that we should complete 2 of the early milestones before adding this one.
[08:25:09] <anewton> nov 2006 for PS for mapping protocol
[08:26:09] --- jon-ietf has become available
[08:27:39] --- HannesTschofenig has become available
[08:27:58] <HannesTschofenig> we have just discussed the charter milestones. here is the proposal: March 2006 Informational RFC containing terminology definitions and the requirements May 2006 An Informational document describing the threats and security considerations March 2006 A PS describing how to identify a session set-up request is to an emergency response center Nov 2006 A Standards Track RFC describing how to route an emergency call based on location information
[08:28:24] <hardie> I agreed to produce a document for the previous 2005/Nov document milestone, and we'll talk in Dallas about whether it needs to be a wg item.
[08:28:59] <HannesTschofenig> ted talks about the Nov 2005 A BCP describing how to discover the media stream types an ERC supports milestone
[08:31:46] <anewton> moving on to mapping protocols
[08:31:57] <anewton> hannes is giving a presentation on design criteria
[08:49:20] <HannesTschofenig> Agreement about - support for existing PSAP boundaries - end host and some servers (e.g., SIP proxies) to use the mapping protocol
[08:49:34] <hardie> Brian and henning agree that the shapes are what the shapes are
[08:54:36] <HannesTschofenig> Ted: Any entity along the signaling path might do the mapping.
[08:57:51] --- linsner has become available
[08:58:57] <linsner> Nadine is discussing differences in caching geo polygons and caching civic locations
[09:18:53] <linsner> discussion centers on using fresh data or cached data for call setup
[09:33:25] <anewton> discussion of trip times on when to use fresh data and cached data
[09:37:58] <anewton> All of the mapping protocols discussed can support both models of checking fresh first or using cache first.
[09:41:03] <HannesTschofenig> agreement that a protocol must support hints about the information it returns. this information returned by the server may contain - lifetime about the returned data - psap boundary how to use this information is a different story (i.e., when you place an emergency call whether you want to use it or use the query protocol to obtain fresh data)
[09:42:20] <HannesTschofenig> agreement that implementations will try to run the mapping protocol before an emergency call.
[09:42:33] <anewton> The agreement is after a long discussion is that at call time a mapping is attempted, but implementations will vary as to when to fall back to cache mapping data.
[09:43:03] <hardie> some will do it immediately, some will do after a trip time (and that trip time may vary)
[09:50:08] <HannesTschofenig> data at the mapping client expires when the TTL along provided by the mapping server expires. additionally, information about the psap boundary is provided with the response.
[10:02:26] <anewton> back from break.
[10:04:44] --- Screen has become available
[10:04:56] --- anewton has left
[10:07:06] --- anewton has become available
[10:07:56] <anewton> test
[10:10:17] <anewton> talk about caching.
[10:10:32] <anewton> Hannes is now talking about how to move forward with the mapping protocols.
[10:12:59] <anewton> Discussion about merging proposals.
[10:13:15] <HannesTschofenig> Ted suggests to merge ECON-IRIS and LUMP.
[10:13:24] <linsner> proposal is to merge econ and lump
[10:13:42] --- linsner has left
[10:16:49] --- Shida has become available
[10:17:19] --- jon-ietf has left: Lost connection
[10:17:19] --- spencerdawkins has left: Lost connection
[10:17:19] --- rogermarshall has left: Lost connection
[10:17:19] --- HannesTschofenig has left: Lost connection
[10:17:19] --- anabolism has left: Lost connection
[10:26:07] --- jon-ietf has become available
[10:26:32] --- hardie has left
[10:29:59] --- linsner has become available
[10:30:14] --- rogermarshall has become available
[10:30:36] <linsner> discussing the rigidity of the dns proposal
[10:31:41] <rogermarshall> that is, that DNS by nature is inflexible in its field definitions
[10:32:23] --- Screen has left: Logged out
[10:33:22] <anewton> test
[10:33:41] <linsner> Andy has to keep testing because he uses a mac..
[10:34:17] --- Screen has become available
[10:35:23] --- spencerdawkins has become available
[10:37:46] <linsner> discussing the reliability of dns - what makes it reliable?
[10:47:51] <anewton> lunch break
[11:33:25] --- spencerdawkins has left
[11:33:26] --- jon-ietf has left: Lost connection
[11:33:35] --- spencerdawkins has become available
[11:37:23] <anewton> back
[11:37:23] --- linsner has left: Lost connection
[11:37:23] --- rogermarshall has left: Lost connection
[11:37:59] --- hardie has become available
[11:40:36] --- HannesTschofenig has become available
[11:41:13] --- linsner has become available
[11:41:44] <linsner> Brian is giving an overview of his DNS proposal.
[11:43:17] <HannesTschofenig> Ted is asking questions about the DNS caching mechanism and its usage in this context.
[11:44:18] <anewton> Ted is asking about who is doing the querying. And if this application of DNS is expected to use the current DNS cache infrastructure.
[11:50:06] --- hardie has left
[11:50:32] <anewton> Talking about which caching resolver gets queried. Ted mentions that Mobile IP in 3GPP and 3GPP2 uses the home network.
[11:51:05] --- anabolism has become available
[12:05:56] <anewton> Discussion about the caching.
[12:06:56] <HannesTschofenig> Ted asked whether someone shares Brian's assumptions about DNS.
[12:07:32] <linsner> Henning is showing a graph showing 5 year old data associated with timing of dns latency
[12:15:56] --- anabolism has left
[12:16:12] --- anabolism has become available
[12:16:35] <HannesTschofenig> Marc asked the room and folks on the conference bridge whether there is interest in the work on the DNS-SOS proposal.
[12:17:01] <HannesTschofenig> Conclusion: Only Brian was in favior of his proposal.
[12:19:34] <linsner> Proposal made to capture positive attributes of the DNS mechanism and carry them forward in design of 'LumpE'
[12:20:53] <spencerdawkins> actually, I'm also concerned about the NEGATIVE attributes Ted was listing off - 5 second DNS timeouts aren't that much slower than 3-second initial RTO TCP timeouts, right?
[12:22:31] <linsner> Henning is showing a demo of a java based LUMP client talking to a LUMP server cluster
[12:22:33] <linsner> I think the attributes that Brian is/was hanging onto include 24/7/365 uptime, etc.
[12:27:10] <anewton> now looking for the attributes of what we want in a mapping protocol
[12:27:48] --- rogermarshall has become available
[12:28:00] <linsner> brian's positive attributes to DNS
[12:32:07] <HannesTschofenig> Brian wants to have a description about different failure cases and the impact on the system.
[12:32:10] <anewton> uniform mechanism
[12:32:19] <anewton> uniform mechanism for discovery
[12:32:35] <anewton> described ways of what servers to use
[12:32:47] <anewton> described ways for resolvers to know authoritative servers
[12:34:30] <anewton> timeout predicatibility
[12:34:41] <anewton> caching predicatibility
[12:36:22] <anewton> separation from application infrastructure
[12:36:33] <anewton> little fate sharing with other applications
[12:36:56] --- hardie has become available
[12:38:40] --- Dan Mongrain has become available
[12:39:03] <HannesTschofenig> http://www.ietf-ecrit.org/Interim2006/ECRIT-IRIS-20060130.ppt
[13:01:38] <HannesTschofenig> Discussion about the functionality we want.
[13:02:05] <HannesTschofenig> More than one transport binding will be needed. (Brian, Henning, Andy, Ted)
[13:04:25] <HannesTschofenig> Discussion about re-transmissions.
[13:04:49] <HannesTschofenig> Andy+Brian: If your UDP request gets lost then you might want to use TCP immediately afterwards.
[13:05:56] <HannesTschofenig> Henning points to the problems with NATs and Firewalls when UDP is used (based on his experience with SIP).
[13:12:30] <HannesTschofenig> Discussion about: If you query using UDP then you get only the URI back. If you use TCP then you get the URI+PSAP boundary+other info back.
[13:13:59] <HannesTschofenig> Discussion about privacy: Is there a privacy problem with the query/response of a mapping protocol?
[13:15:10] <HannesTschofenig> Ted: If we only use UDP in the emergency case then everyone knows that an emergency is happening.
[13:16:26] <anabolism> Also, there are privacy concerns even without an explicit identity (for example, IP addresses)
[13:16:55] <HannesTschofenig> you are certainly right. the whole work in mobile ip privacy just shows us this particular issue
[13:20:23] <HannesTschofenig> Henning suggests to provide a description of TLS/UDP/TCP usage. It might depend on a number of factors.
[13:21:56] <HannesTschofenig> Ted wants to split the seeker part from the backend part.
[13:22:18] <linsner> we're having noise on the bridge....
[13:23:25] <HannesTschofenig> henning suggests to put - seeker protocol into one document - protocol between trees (database replication) into another doc.
[13:25:57] <anewton> file serialization
[13:39:49] <anabolism> Hello? Why is there a computer-generated voice reciting a weather forecast on the audio bridge?
[13:41:17] <HannesTschofenig> we have turned on the weather forecast.
[13:41:25] <anabolism> ah
[13:41:42] <HannesTschofenig> but we are about to start in a few minutes.
[13:41:58] <anabolism> let me know when we are back, as I am not listening on the bridge at the moment
[13:41:59] <anabolism> thanks
[13:42:28] <HannesTschofenig> i wil put the slides to http://www.ietf-ecrit.org/Interim2006/
[13:42:44] <Screen> was the voice computer generated?
[13:43:19] <anabolism> It sounded so to me, but of course I have no way of knowing
[13:45:21] <HannesTschofenig> http://www.ietf-ecrit.org/Interim2006/ECRIT_Interim_Location_Conveyance.ppt
[13:45:36] <anewton> James blames the AD for moving this slowly!
[13:45:59] <anewton> Blames Jon as Jon and not Jon as the AD.
[13:46:31] <anewton> James Polk is talking about the Emergency SIP Location Conveyance
[13:46:43] <anewton> Will be cutting more cruft from the draft
[13:47:01] <anewton> moving examples to examples document
[13:50:54] <anewton> chat log for yesterday: 2006-02-01.html
[13:51:23] <anewton> chat log for yesterday: http://www.xmpp.org/ietf-logs/ecrit@ietf.xmpp.org/2006-02-01.html
[13:51:38] <anewton> chat log for today: http://www.xmpp.org/ietf-logs/ecrit@ietf.xmpp.org/2006-02-02.html
[14:08:25] --- rogermarshall has left
[14:23:06] --- Dan Mongrain has left
[14:36:02] <linsner> still going through location conveyance draft...
[15:03:22] --- linsner has left
[15:06:59] --- Shida has left
[15:07:43] --- Screen has left: Logged out
[15:07:50] --- hardie has left
[15:07:59] <anewton> IT IS OVER!
[15:08:01] --- anewton has left
[15:09:27] --- anabolism has left
[15:40:20] --- HannesTschofenig has left
[16:09:03] --- spencerdawkins has left: Disconnected
[22:59:05] --- spencerdawkins has become available