Conversation with esw07@conference.ecotroph.net at Tue 30 Oct 2007 09:58:09 AM EDT

Conversation with esw07@conference.ecotroph.net at Tue 30 Oct 2007

(04:29:47 AM) richard.barnes entered the room.
(04:29:47 AM) SDO Emergency Services Workshop 2007
(10/30/2007 04:07:45 AM) jchiaramonte: end host asks for a reference, gets the reference, sends to a recipient and then can be dereferenced
(10/30/2007 04:07:54 AM) richard.barnes: right.
(10/30/2007 04:08:25 AM) richard.barnes: end host gets a reference through HELD, sends it to the recipient however he wants, then the recipient uses the reference to get the location object
(10/30/2007 04:08:45 AM) andres.kytt@gmail.com: As Hannes noted, this is mostly useful for helping roaming customers avoid constant binging
(10/30/2007 04:08:46 AM) richard.barnes: next slide - proxy fetching location on behalf of target
(10/30/2007 04:09:08 AM) richard.barnes: It also has nice security implications
(04:10:07 AM) richard.barnes: Winterbottom: link to HELD identity extensions should be -03 instead of -02
(04:10:20 AM) richard.barnes: next slide - ES use case
(04:10:28 AM) JamesWinterbottom: 03 is much shorter than 02
(04:10:46 AM) JamesWinterbottom: and defines a set of URIs for the most common identifier types
(04:11:43 AM) JamesWinterbottom: http://tools.ietf.org/html/draft-winterbottom-geopriv-held-identity-extensions-03
(04:11:55 AM) richard.barnes: next slide - Dereference protocols
(04:13:33 AM) richard.barnes: Linsner: Please explain difference between individual ID and WG item
(04:13:45 AM) richard.barnes: (ID -> I-D)
(04:14:45 AM) richard.barnes: next slide - location filters
(04:18:12 AM) richard.barnes: next slide - identifying an emergency call
(04:19:55 AM) richard.barnes: next slide - emergency numbers
(04:23:15 AM) richard.barnes: next slide - LoST
(04:25:19 AM) richard.barnes: next slide - LoST features
(04:28:54 AM) JamesWinterbottom: And there goes the scribe
(04:29:47 AM) HannesTschofenig has set the topic to: 3rd Emergency Services Workshop
(04:30:00 AM) richard.barnes: (back now)
(04:30:15 AM) richard.barnes: next slide - LoST example
(04:31:20 AM) richard.barnes: Linsner: Recalling what XMPP is?
(04:31:47 AM) richard.barnes: Hannes: XMPP is like SIP presence, but a different protocol.
(04:31:58 AM) richard.barnes: (I think all of us in this room implicitly know what XMPP is?)
(04:32:56 AM) richard.barnes: next slide - location validation
(04:33:37 AM) richard.barnes: Linsner / Hannes: LoST can provide URIs of many types (SIP, XMPP, etc), but the contact point will always be a URI
(04:33:50 AM) richard.barnes: next slide - listServices example
(04:34:21 AM) richard.barnes: Linsner: Note that this is not only for emergency services, also urn:service:pizza
(04:35:43 AM) richard.barnes: Next slide : location conveyance
(04:37:44 AM) richard.barnes: next slide : conveyance example
(04:38:26 AM) JamesWinterbottom: Example shows depricated geo form
(04:39:57 AM) jchiaramonte: james - can you add that comment here?
(04:40:26 AM) JamesWinterbottom: Sure.. the example was showing the use of N and W location differentiators for the point
(04:40:42 AM) JamesWinterbottom: GML 3.1.1 deprecated that form
(04:40:55 AM) richard.barnes: next example: proxy adding lbyr
(04:41:06 AM) JamesWinterbottom: +- indicators should be used now, and this is shown in PIDF-LO profile
(04:41:11 AM) richard.barnes: next slide: architectural models
(04:41:11 AM) JamesWinterbottom: Does that help?
(04:41:24 AM) jchiaramonte: yes - thanks
(04:44:20 AM) richard.barnes: next slide - endpoint-driven model
(04:45:19 AM) JamesWinterbottom: I hope Hannes points out that the ned host-based model works in any environment, where the proxy-based model doesn't
(04:48:37 AM) richard.barnes: next slide - notes on the host-based model
(04:50:54 AM) richard.barnes: next slide - proxy-based model
(04:53:35 AM) richard.barnes: next slide - legacy equipment
(04:54:16 AM) richard.barnes: next slide - legacy equipment disadvantages
(04:55:20 AM) richard.barnes: next slide - media
(04:58:00 AM) richard.barnes: Hannu: In the last case, who is expected to respond?
(04:58:14 AM) richard.barnes: Hannes: The PSAP -- and it's supposed to play back audio as well
(04:58:34 AM) richard.barnes: next slide: Mapping arch
(05:01:37 AM) richard.barnes: next slide: Mapping arch example
(05:05:21 AM) richard.barnes: next slide: Location hiding
(05:07:11 AM) richard.barnes: Winterbottom: Why would you want to do that?
(05:07:20 AM) richard.barnes: Linsner / Hannes: Talk about that later
(05:08:20 AM) JamesWinterbottom: Not quite what I said, but close enough
(05:08:40 AM) richard.barnes: next slide - unauthenticated / unauthorized calls
(05:10:23 AM) richard.barnes: next slide: conclusion
(05:11:52 AM) richard.barnes: end of Hannes' presentation
(05:11:58 AM) richard.barnes: 5 min break
(05:12:28 AM) andres.kytt@gmail.com left the room.
(05:24:15 AM) richard.barnes: ok, we're starting again
(05:24:47 AM) richard.barnes: Hannu talking about the 3GPP/IMS ES architecture
(05:27:34 AM) richard.barnes: Marc reviewed logistics, same as before
(05:28:08 AM) richard.barnes: Hannu Hietalahti taking the podium
(05:29:27 AM) richard.barnes: next slide: agenda
(05:30:10 AM) richard.barnes: next slide: IMS calls in Rel-7
(05:30:16 AM) richard.barnes: (emergency calls)
(05:34:01 AM) richard.barnes: next slide: 3GPP specifications
(05:37:08 AM) richard.barnes: next slide: agenda - service requirements
(05:37:12 AM) richard.barnes: next slide: service requirements
(05:39:07 AM) HannesTschofenig entered the room.
(05:39:20 AM) HannesTschofenig: I am going to put all slides to the following Wiki
(05:39:21 AM) HannesTschofenig: http://www.tschofenig.com/twiki/bin/view/EmergencyServices/EswAgenda2007
(05:44:51 AM) richard.barnes: next slides: Service requirements for UE
(05:46:46 AM) richard.barnes: Hannu: experience has shown that "red button" on phone-like devices is a bad idea, but usable for special-purpose devices (like airbag sensors)
(05:48:00 AM) richard.barnes: Rosen: This will be a problem because of disabled/TTY users
(05:48:54 AM) richard.barnes: Rosen: Might have a terminal that only supports text (i.e., not voice), but still wants to make emergency calls
(05:49:27 AM) richard.barnes: Kytt: What is the border that defines the terminal? VoIP over cellular modem?
(05:49:47 AM) richard.barnes: Hannu: That sort of configuration would not trigger the mandate; only the modem is tested for conformance
(05:50:11 AM) richard.barnes: Hannu: Regulators may think different
(05:50:19 AM) richard.barnes: Van Gaever: Regulators do think different
(05:50:31 AM) richard.barnes: vG: disabled users are a very important consideration
(05:50:53 AM) richard.barnes: Hannu: Will relay to supporting companies in 3GPP
(05:51:37 AM) richard.barnes: Rosen: Chairs - should we bring up questions during these tutorial sessions?
(05:51:49 AM) richard.barnes: Linsner: Yes, we can at least capture them in the notes
(05:53:08 AM) linsner: slides are at: http://www.tschofenig.com/twiki/bin/view/EmergencyServices/EswAgenda2007
(05:53:15 AM) richard.barnes: many different classes of "limited service", i.e., there's not a single "unauthorized access" use case
(05:53:45 AM) richard.barnes: current slides are <http://www.tschofenig.com/twiki/pub/EmergencyServices/EswAgenda2007/3GPP_IMS_emergency_call_overview.ppt>
(05:54:29 AM) richard.barnes: Kytt: Does the 4th case (Illegal mobile station) mean that there's no way to block people making emergency calls?
(05:54:52 AM) richard.barnes: Hannu: Some people say this is a problem, up to 20-30% of PSAP's work
(05:55:10 AM) richard.barnes: Hannu: Nonetheless, still a regulatory requirement
(05:56:50 AM) richard.barnes: Hannu: More directly, if local authorities consider it's a problem, they can authorize blocking of emergency calls (the mechanism is there)
(05:57:42 AM) richard.barnes: Dirk: Do your use cases include expired/empty prepaid cards?
(05:58:00 AM) richard.barnes: Hannu: Yes [RB: think it's the 3rd case]
(05:59:57 AM) richard.barnes: next slide - identification of the emergency call
(06:00:39 AM) jchiaramonte: If the regulators block limited service calls - are all limited service types blocked (treated identically)?
(06:06:14 AM) richard.barnes: Q: If you have lots of default numbers, you get lots of accidental calls
(06:06:48 AM) richard.barnes: Hannu: I would prefer to block these completely. Maybe we should keep this in mind for the panel discussion
(06:07:51 AM) richard.barnes: next slide - interaction between CS and PS
(06:08:21 AM) richard.barnes: Terminal has to decide whether to use CS or PS for emergency calls
(06:08:36 AM) richard.barnes: Default is to use CS when capable
(06:10:42 AM) richard.barnes: next slide: location
(06:10:57 AM) andres.kytt@gmail.com entered the room.
(06:11:37 AM) richard.barnes: next slide: IMS calls in 3GPP environment
(06:14:31 AM) richard.barnes: Q: Is there a preference among different IMS accesses (GPRS vs fixed vs WLAN)
(06:14:54 AM) richard.barnes: Hannu: No, there is not preferences; terminals will usually not be attached to more than one at a time
(06:16:34 AM) richard.barnes: next slide - architecture
(06:19:01 AM) richard.barnes: Barnes: In Hannes' model, E-CSCF is the proxy doing LCP/LoST with the LRF?
(06:19:06 AM) richard.barnes: Hannu: Yes
(06:19:59 AM) richard.barnes: Q: Is location carried in SIP or with other 3GPP mechanisms?
(06:20:38 AM) richard.barnes: Hannu: At a minimum, terminal will need to include serving network/cell, in WLAN you include something else
(06:20:58 AM) richard.barnes: If the UE supports another positioning mechanism, it will include that
(06:21:19 AM) richard.barnes: Q: SUPL will invoke 3GPP protocols to get to the terminal
(06:21:39 AM) richard.barnes: Milan: Protocols for getting location are not defined in R7
(06:21:55 AM) richard.barnes: (i.e., the interface between E-CSCF and LRF)
(06:22:58 AM) richard.barnes: Liess: This doesn't work without a P-CSCF [Hannu: correct]
(06:23:47 AM) richard.barnes: Liess: As companies start to implement IMS, they may not have a P-CSCF in the visited network
(06:24:21 AM) richard.barnes: Liess: No real reason to use a P-CSCF in the visited network
(06:24:59 AM) richard.barnes: Hannu: As long as you get IP connectivity back to home network, you're OK
(06:25:21 AM) richard.barnes: Hannu: Only a problem if there's no CS in the terminal
(06:26:57 AM) richard.barnes: Q: Is IPv6 still the only IP for SIP? How does IPv6 affect all of this?
(06:27:21 AM) richard.barnes: Hannu: We have included IPv4 now.
(06:28:14 AM) richard.barnes: Rosen: NENA accepts both v4 and v6 (probably/hopefully based on v6)
(06:28:38 AM) richard.barnes: Hannu: Haven't made any effort to remove it from emergency services, intention is to support IPv6
(06:29:35 AM) richard.barnes: Rosen: We should be able to come up with a way to mark calls from PSAPs
(06:30:09 AM) richard.barnes: Hannu: We also need to understand the requirements better
(06:30:23 AM) richard.barnes: Also need to correlate call-backs to the original emergency calls
(06:30:53 AM) richard.barnes: Hannes: This is in phonebcp, right? [Rosen: Yes]
(06:31:04 AM) richard.barnes: next slide - Main system elements
(06:32:36 AM) linsner: next slide...(slide 16)
(06:32:45 AM) linsner: protocol requirements..
(06:33:37 AM) linsner: slide 17 - functionality
(06:36:01 AM) socratez entered the room.
(06:36:52 AM) richard.barnes: Bajko: The slide says that only the local network knows which PSAP serves the UE. In the general case, this information is public, so the home network could determine it as well.
(06:37:41 AM) richard.barnes: Barnes: That's the point of LoST, right?
(06:38:01 AM) richard.barnes: Hannu: Yes -- if operators deploy LoST. There's no 3GPP requirement for that yet.
(06:38:48 AM) richard.barnes: Q: Do you mean that only the local network has the location information? Just like with IETF, LIS has to be in the local network.
(06:39:53 AM) richard.barnes: [Schramm was previous speaker]
(06:40:55 AM) richard.barnes: Murhammer: Regulations only apply to networks within the jurisdiction.
(06:41:50 AM) richard.barnes: Rosen: That used to be the case. With wireline IP roaming, the difficulty is the separation between packet service and calling service
(06:42:22 AM) richard.barnes: Requirement is for the whole thing to work when the packet service is not aware that an emergency call is being made
(06:42:49 AM) richard.barnes: next slide - 18 - differences from normal calls
(06:43:39 AM) socratez left the room.
(06:44:27 AM) richard.barnes: next slide - 19 - UE
(06:46:36 AM) richard.barnes: Q: Which part of the UE has to conform to these requirements? The device, the application?
(06:46:55 AM) richard.barnes: Hannu: In your configuration with a laptop/modem, where is the SIP stack?
(06:47:09 AM) richard.barnes: Q: On the laptop - the card just provide IP connectivity
(06:47:23 AM) richard.barnes: Hannu: The card is a UE, but doesn't support IMS
(06:49:04 AM) richard.barnes: Emergency services requirements only apply if the UE is doing SIP with the IMS core
(06:49:31 AM) richard.barnes: A UE that's just a modem doesn't do SIP, so it's not tested; if you bring the SIP stack, it's your problem
(06:50:30 AM) richard.barnes: McCann: Does 3GPP intend to have a relationship with WiFi Alliance or WiMAX forum? This could help ensure that multi-mode terminals are fully compliant.
(06:51:08 AM) richard.barnes: Hannu: This is entering into the domain of the "Common IMS". This is an ongoing discussion.
(06:51:28 AM) klaas entered the room.
(06:52:02 AM) richard.barnes: But you have to make the division between user-provided software and the UE
(06:52:54 AM) richard.barnes: next slide - 20 - P-CSCF
(06:54:10 AM) richard.barnes: next slide - 21 - E-CSCF
(06:54:15 AM) richard.barnes: next slide - 22 - LRF
(06:54:50 AM) richard.barnes: next slide - 23 - S-CSCF
(06:56:16 AM) richard.barnes: Barnes: What's the status of LoST in the 3GPP? Not required now, but in the future?
(06:56:48 AM) richard.barnes: Hannu: Not forbidden in current release, and we've just entered the R8 process
(06:57:24 AM) richard.barnes: Marcus: Recent studies in the US showed that network-derived location is worse than GPS. Is there a move toward this?
(06:57:59 AM) richard.barnes: Rosen: Difference between "network knows location" and "network-based determination" -- AGPS means network knows location
(06:58:44 AM) richard.barnes: End of 3GPP presentation
(06:58:50 AM) andres.kytt@gmail.com left the room.
(06:59:21 AM) richard.barnes: Breaking for lunch
(06:59:41 AM) jchiaramonte left the room.
(07:52:29 AM) linsner: we're back at it..
(07:52:38 AM) richard.barnes: Brian now briefing on NENA i2
(07:52:48 AM) richard.barnes: First slide - Purpose and scope of i2
(07:52:55 AM) linsner: Brian Rosen presenting for James Winterbottom
(07:53:07 AM) richard.barnes: <http://www.tschofenig.com/twiki/pub/EmergencyServices/EswAgenda2007/NENA-i2.ppt>
(07:53:17 AM) HannesTschofenig left the room (Replaced by new connection).
(07:53:59 AM) jchiaramonte entered the room.
(07:54:23 AM) jchiaramonte left the room.
(07:54:44 AM) richard.barnes: next slide - 3 - existing interfaces
(07:54:55 AM) jchiaramonte entered the room.
(07:55:19 AM) linsner: brian said 'tdm analog'....hmm.
(07:56:32 AM) richard.barnes: next slide - VoIP
(07:58:31 AM) richard.barnes: next slide - key functions of location
(07:58:33 AM) andres.kytt@gmail.com entered the room.
(08:00:41 AM) richard.barnes: next slide - Goal - adding a new voice nework
(08:01:59 AM) richard.barnes: next slide - Basic i2
(08:03:29 AM) richard.barnes: Linsner: Do we have experience to indicate what's worked and what hasn't?
(08:03:40 AM) richard.barnes: Rosen: Bring that up later
(08:04:24 AM) richard.barnes: next slide - i2 with a Routing proxy
(08:07:34 AM) richard.barnes: next slide - which ERDB to choose?
(08:07:49 AM) richard.barnes: Linsner: Is this protocol published?
(08:08:32 AM) richard.barnes: Rosen: To be clear, the ERDBs publish to the root discovery node, who hands out pointers to the ERDBs. NENA has documents for these interfaces
(08:09:04 AM) richard.barnes: Next slide - (no title)
(08:09:23 AM) richard.barnes: Next slide - Internal view [ of ERDB ]
(08:11:28 AM) richard.barnes: Next slide - Routing on Civic
(08:13:35 AM) richard.barnes: Next slide - Message Flow Diagram
(08:13:45 AM) richard.barnes: End of i2 discussion
(08:14:03 AM) richard.barnes: Back to slide - Basic i2
(08:14:13 AM) richard.barnes: This is not really implemented
(08:14:37 AM) richard.barnes: FCC required that all VoIP carriers had to implement emergency calling just before this, and gave them 90 days
(08:14:50 AM) richard.barnes: there are VPCs and ESGW operators
(08:14:58 AM) richard.barnes: ESGW more or less per standards
(08:15:24 AM) richard.barnes: VPCs work differently, by geocoding (to geodetic location, then use the polygons from the mobile routing)
(08:15:49 AM) richard.barnes: LIS operated by VPC or same as VPC
(08:16:09 AM) richard.barnes: There's no ERDB lookup
(08:16:24 AM) richard.barnes: This has been pretty successful; the major problem is that location is self-reported
(08:17:56 AM) richard.barnes: Some places (e.g., TX) are moving to require ERDBs and VDBs
(08:18:38 AM) richard.barnes: Palm: If LIS user entry is voluntary, what happens if you don't enter it?
(08:18:43 AM) richard.barnes: Rosen: You're denied service
(08:19:18 AM) richard.barnes: Medland: [what was the question?]
(08:19:20 AM) jchiaramonte: q: how does the user actually enter the information?
(08:19:41 AM) richard.barnes: Rosen: It's hard, a lot of the time the user's idea of location doesn't match the MSAG
(08:19:52 AM) klaas: usually via a webform
(08:19:53 AM) richard.barnes: So there's a process that you have to go through to make them match
(08:20:19 AM) richard.barnes: The thing that fails is that you fail to update it when you go places
(08:21:11 AM) richard.barnes: Stark: Do VPCs have nationwide coverage now?
(08:21:17 AM) jchiaramonte: N.B. not all Vonage products provide 9-1-1 service (i.e. their "softphone" application)
(08:21:29 AM) richard.barnes: Rosen: I haven't seen stats, but we're in the high 90% range
(08:21:44 AM) richard.barnes: Marshall: It is in the high 80s or in the low/mid 90s.
(08:22:03 AM) richard.barnes: The reason it's not near 100% is that some PSAPs aren't very proactive in wanting to deploy VoIP services
(08:22:25 AM) richard.barnes: One of the weaknesses of self-provisioned location is that people can (purposely) input WRONG location
(08:22:46 AM) richard.barnes: Also, the process has gotten to where validation is more or less real time
(08:23:41 AM) richard.barnes: Linsner: some jurisdictions have requirements for tighter locations, e.g., within 7k sq ft, on the exact floor
(08:24:04 AM) richard.barnes: Never seen the V3/V7 interfaces from the LIS to a VPC in an enterprise scenario
(08:24:21 AM) richard.barnes: Rosen: V3 is really location by reference
(08:24:55 AM) richard.barnes: Q: Has there been a conclusion following the recent FCC call for comment?
(08:25:01 AM) richard.barnes: Rosen: Not with respect to VoIP
(08:25:57 AM) jchiaramonte: Status (dated) of MLTS/PBX Legislation: http://www.nena.org/pages/Content.asp?CID=156&CTID=41
(08:26:20 AM) richard.barnes: [ Previous speaker: (couldn't understand what he said) ]
(08:26:49 AM) richard.barnes: Marshall: While V3/V7 aren't done by NENA standards, there are internal, proprietary protocols that work
(08:26:59 AM) richard.barnes: When there's a need for external interfaces, they'll come along
(08:27:22 AM) richard.barnes: Hellstrom: Is there real-time text in V1/V4?
(08:27:58 AM) richard.barnes: Rosen: There's nothing in the spec about it. The paths exist, but the ESGWs aren't specified to translate from any text over IP to TTY
(08:28:28 AM) richard.barnes: NB: i2.5 should have a provision so that TTY survives through the VoIP leg
(08:28:38 AM) richard.barnes: Tschofenig: Can you speak a little more about i2.5?
(08:28:51 AM) richard.barnes: Rosen: We have learned a few things since we released i2.
(08:29:26 AM) richard.barnes: E.g., you can't use one form of addressing in the gateways, i.e., one URI that's uniform across gateways
(08:29:59 AM) richard.barnes: We're hoping to get wider deployment of i2.5
(08:30:35 AM) richard.barnes: Palm: You mention nomadic usage of VoIP; is that covered?
(08:31:07 AM) richard.barnes: Rosen: The specs are designed with that in mind. The notion was that V0 would be some sort of IETF LCP and the endpoint would pass the location to the calling network
(08:31:25 AM) richard.barnes: Palm: But if I enter my info manually...
(08:31:52 AM) richard.barnes: Rosen: Location in the signaling would override the self-reported location
(08:32:09 AM) richard.barnes: Tschofenig: Do we have data on the end host side on what's to be implemented there?
(08:32:33 AM) richard.barnes: Rosen: NENA doesn't control that; AFAWK there's no implementation on that end [I guess of V0?]
(08:33:36 AM) linsner left the room.
(08:33:52 AM) linsner entered the room.
(08:34:03 AM) linsner: starting the location presentation...
(08:34:33 AM) linsner: slide 23 of the preso...
(08:37:41 AM) richard.barnes: Tschofenig: There was a requirements document sent around ECRIT and GEOPRIV -- was that this?
(08:37:48 AM) richard.barnes: Rosen: No, that was an ESIF document
(08:38:27 AM) richard.barnes: Linsner: Current incarnation of HELD doesn't support V3
(08:38:58 AM) richard.barnes: Barnes: Right, but that's because it's just a V0 protocol as specified, check the dereference usage doc for V3
(08:39:29 AM) richard.barnes: LEM: OMA has defined SUPL, which is usable for V3
(08:39:41 AM) richard.barnes: Rosen: SUPL doesn't meet NENA requirements right now (civic in particular)
(08:40:14 AM) richard.barnes: Marshall: My understanding is that HELD is being extended to reference/V3...
(08:40:30 AM) richard.barnes: Rosen/Barnes: Yes, that's the held-dereference and identity-extensions documents
(08:40:40 AM) richard.barnes: Moving to i3 briefing
(08:40:48 AM) klaas left the room (Replaced by new connection).
(08:40:51 AM) richard.barnes: http://www.tschofenig.com/twiki/pub/EmergencyServices/EswAgenda2007/ESW3-NENAi3Arch.ppt
(08:40:59 AM) richard.barnes: Slide 2
(08:42:04 AM) richard.barnes: Slide 3
(08:44:38 AM) richard.barnes: Slide 4
(08:45:17 AM) richard.barnes: Slide 5
(08:47:07 AM) richard.barnes: Slide 6
(08:49:07 AM) richard.barnes: Slide 7
(08:49:25 AM) richard.barnes: Slide 8
(08:50:49 AM) klaas entered the room.
(08:53:22 AM) richard.barnes: Slide 9
(08:55:29 AM) linsner: slide 10
(08:58:03 AM) linsner: up to slide 13
(08:59:02 AM) linsner: richardbarnes - so is the LIS just a proxy...?
(08:59:17 AM) linsner: it's just a protocol converter (Brian)
(08:59:51 AM) linsner: question - I don't understand the scope of NENA i3??
(09:01:23 AM) linsner: Brian - NENA i3 only defines the ecrf and esrp funtions, NENA is not defining the call control in the origination network
(09:02:11 AM) linsner:
Brian - for legacy wireline and wireless, NENA i3 is defining the gateway to the ESI net
(09:02:34 AM) linsner: question - the legacy gateway isn't special???
(09:02:37 AM) richard.barnes: Slide 14
(09:03:26 AM) linsner: Brian - well, it does need to perform the ecrf functions..
(09:04:55 AM) richard.barnes: Slide 15
(09:06:42 AM) richard.barnes: Slide 16
(09:10:21 AM) richard.barnes: Bajko: PSAP does not assume a carrier, so I could set up my own proxy
(09:10:34 AM) richard.barnes: Rosen: So in that case, you might not have the extra data
(09:10:48 AM) richard.barnes: Bajko: That means that an attacker could fake everything
(09:10:54 AM) richard.barnes: Rosen: That's the risk you take
(09:11:26 AM) richard.barnes: Hannu: That's why some networks don't take unauthenticated calls any more.
(09:12:24 AM) richard.barnes: I like the idea of using LoST as a central control point.
(09:12:27 AM) richard.barnes: Rosen: Was going to design the LoST server so that a carrier would have a local copy, controlled by the authoritative server
(09:12:49 AM) richard.barnes: Hannu: Can't help the work load at a PSAP. What's the point?
(09:13:36 AM) richard.barnes: Rosen: Maybe the carrier's DB points to a PSAP that's been blown up. PSAP should have backup, and the backup facility should be able to push the change
(09:14:20 AM) richard.barnes: Tschofenig: Multiple levels of load balancing. LoST can have multiple URIs, URI can resolve to multiple IPs through the DNS, ESRP can fan out further
(09:15:44 AM) richard.barnes: Palm: In Sweden, we have 4 duplicates of call data, and PSAPs can hand over through these
(09:15:50 AM) richard.barnes: Slide 17
(09:18:40 AM) richard.barnes: Hannu: Related to DoS attack: Do you have a recommendation on how to handle that? How do you distinguish from a disaster?
(09:18:43 AM) richard.barnes: Slide 18
(09:22:41 AM) richard.barnes: Wu: Before a responder can get a call, the DoS can take out the IP network as well
(09:23:15 AM) richard.barnes: Rosen: Building systems to try to be resilient to most known attacks.
(09:23:41 AM) richard.barnes: e.g., recommending 4-10GB input bandwidth, which covers most cases
(09:25:22 AM) richard.barnes: Hannu: That's a problem at the bearer level. We've been trying to eliminate that in the 3GPP, even enabling the carrier to disable things in the terminal remotely
(09:25:56 AM) richard.barnes: Theoretically, you can't do anything about an endpoint that doesn't listen, just spews out noise
(09:26:26 AM) richard.barnes: Rosen: Might have border control functions be able to turn of IP addresses
(09:26:32 AM) richard.barnes: s/of/off/
(09:27:01 AM) jchiaramonte: or reprioritize calls within a queue
(09:28:25 AM) richard.barnes: Lange: Are you worried about QoS for emergency calls?
(09:28:40 AM) richard.barnes: Rosen: Spec is 2sec setup time
(09:29:05 AM) richard.barnes: There are other aspects to QoS, and we're trying to address those
(09:29:41 AM) richard.barnes: Currently, we're saying to mark emergency calls (signaling and media), but not taking a stance on how those markings are used
(09:29:52 AM) richard.barnes: within the ESInet, emergency calls will get some sort of priority
(09:30:43 AM) richard.barnes: To clarify, the ESInet is NOT the Internet, it's a private IP network
(09:31:27 AM) richard.barnes: Hixson: We don't intend PSAPs to be on the Internet directly
(09:32:00 AM) richard.barnes: Medland: How do we make the transition to multimedia?
(09:32:27 AM) richard.barnes: Rosen: The ES-side networks will accept multimedia, and they can also apply policy.
(09:33:00 AM) richard.barnes: That policy may have to do with who's calling, e.g., might accept video from known parties, not from just anybody
(09:33:46 AM) richard.barnes: Trying to give you the basis on which to discriminate
(09:35:48 AM) richard.barnes: Hannu: What about evolution to multimedia? The good thing here is that the process overall is agnostic to what type of media you're using
(09:36:03 AM) richard.barnes: [rb: Yes, that's the point of the separation between SIP and SDP and media!]
(09:36:26 AM) richard.barnes: Still need PSAPs that support multimedia
(09:36:37 AM) richard.barnes: Wu: What's the time-frame for deployment?
(09:36:54 AM) richard.barnes: Rosen: Demonstration next year, deployment after, overall 3-5yrs
(09:37:28 AM) richard.barnes: We have gone from "You guys are crazy" to "it'll never happen" to "it'll happen, we don't know when" to "ok, soon, but we don't know how to pay for it"
(09:37:36 AM) richard.barnes: Time scale will be clearer once funding is available
(09:38:10 AM) richard.barnes: Marcus: Along those lines: These are funded at the state/local level. Are there efforts at pursuing Federal $$ ?
(09:38:29 AM) richard.barnes: Rosen: There's no official NENA opinion on this, so this is my opinion
(09:38:52 AM) richard.barnes: 2 sources of revenue: taxes and surcharges. These vary a lot, but they're structurally the same
(09:39:14 AM) richard.barnes: That much might not change, except maybe the surcharge will move to the IP network
(09:39:49 AM) richard.barnes: ENHANCE Act authorizes $1.5 M, but it's not appropriated
(09:40:01 AM) richard.barnes: At best, they'll cover the cost of transition
(09:41:04 AM) richard.barnes: Cost of running parallel networks might speed up transition once we get into it(09:58:09 AM) spencerdawkins left the room.
(10:13:49 AM) richard.barnes: starting OMA presentation
(10:14:24 AM) richard.barnes: On slide 5 now
(10:14:27 AM) richard.barnes: Slide 6
(10:15:03 AM) richard.barnes: Slide 8
(10:15:20 AM) richard.barnes: Slide 9
(10:16:52 AM) richard.barnes: Slide 10
(10:17:49 AM) richard.barnes: Slide 11
(10:19:55 AM) richard.barnes: Slide 13
(10:20:27 AM) richard.barnes: Slide 14
(10:20:29 AM) richard.barnes: Slide 15
(10:20:35 AM) richard.barnes: Slide 16
(10:21:32 AM) richard.barnes: Slide 17
(10:22:32 AM) richard.barnes: Slide 18
(10:23:01 AM) richard.barnes: Slide 19
(10:23:26 AM) linsner left the room.
(10:24:35 AM) richard.barnes: Slide 20
(10:25:15 AM) richard.barnes: Slide 21
(10:25:18 AM) richard.barnes: Slide 22
(10:26:49 AM) richard.barnes: Slide 23
(10:27:48 AM) richard.barnes: Slide 24
(10:27:48 AM) jchiaramonte: difference between SUPL 1.0 - only GPS - SUPL 2.0 adds Galileo
(10:27:57 AM) jchiaramonte: (and something else)
(10:28:16 AM) richard.barnes: Slide 25
(10:28:30 AM) richard.barnes: SUPL 2.0 essentially adds more positioning mechanisms
(10:29:39 AM) spencerdawkins entered the room.
(10:29:45 AM) richard.barnes: Slide 26
(10:29:49 AM) richard.barnes: End of presentation
(10:30:37 AM) richard.barnes: Rosen: Does the system have a way to push location to the recipient (MLS client)?
(10:31:30 AM) richard.barnes: Lem: Yes. In SUPL 2.0, there is "triggered location". Client sends an area, gets location when the terminal enters the area
(10:32:04 AM) richard.barnes: Rosen: What you would like for a PSAP is a trigger that says "send an update when you move N meters, no more than M per second"
(10:32:10 AM) richard.barnes: Is there something like that?
(10:32:34 AM) richard.barnes: Lem: Both the area and for a certain amount of time can be constrained.
(10:34:26 AM) richard.barnes: Tschofenig: What identifier is being used by the application for the terminal (in MLP)?
(10:34:32 AM) richard.barnes: Lem: The ID is a phone number.
(10:35:08 AM) richard.barnes: Tschofenig: So translating this to the IP world, you would need a different identifier, like a SIP URI.
(10:37:32 AM) richard.barnes: (missed a couple of comments)
(10:38:02 AM) richard.barnes: Palm: If someone from DE comes to SE, how do the SE PSAPs get location?
(10:38:57 AM) richard.barnes: Lem: You access the mobile by its phone number, the mobile provider knows whether the phone is roaming
(10:39:09 AM) richard.barnes: Palm: We know who to ask right now based on ISUP
(10:39:34 AM) richard.barnes: Lem: That's another approach -- you can either ask where to go and go there, or you can go someplace and get redirected
(10:40:13 AM) richard.barnes: Nistal: SUPL 2.0 is pretty open with respect to positioning mechanisms. Working on WLAN, WiMAX
(10:40:38 AM) richard.barnes: OK as long as (1) you can get IP to the terminal and (2) the terminal can position itself
(10:41:37 AM) richard.barnes: [RB: In response to Hannes' comment -- From the IETF perspective, both MLP and RLP are location fetch mechanisms, hence effectively dereference protocols]
(10:42:11 AM) richard.barnes: Rosen: If you just get a call, and all you have is an MSISDN, you have to figure out who the home operator or visited operator is.
(10:42:30 AM) richard.barnes: That requires the SS7 network to figure out
(10:42:43 AM) richard.barnes: Is there a 3GPP equivalent? How do you figure it out?
(10:44:01 AM) richard.barnes: Rosen: To be clear, the concept is that you use the MSISDN to find the operator, then ask the operator?
(10:44:09 AM) richard.barnes: [Somebody]: Yes, that's the idea
(10:47:01 AM) richard.barnes: Tschofenig: How do you get to the terminal? A lot of times, NAT/firewall/etc will screw this up.
(10:47:15 AM) richard.barnes: Nistal: No matter what, the client has to listen somehow
(10:48:40 AM) richard.barnes: [???]: Have to consider that privacy is a critical issue. That's taken into account in this architecture through a set of internal protocols.
(10:49:13 AM) richard.barnes: Lem: Main difference between MLS and IP-based positioning is that the only interface to the mobile network is MLP
(10:49:48 AM) richard.barnes: Tschofenig: Good point re: privacy. Obviously, privacy policies are in the home network; how does the user modify these policies?
(10:50:59 AM) richard.barnes: Lem: Out of band. Part of establishing a subscription to a location-based service. Also, periodic notifications
(10:51:41 AM) richard.barnes: [???]: AFAIK, these things are usually all done within a single nation
(10:53:28 AM) richard.barnes: Barnes: Clarification -- privacy settings are not automated
(10:53:32 AM) richard.barnes: Lem: That's correct
(10:53:50 AM) richard.barnes: Nistal: You can also contact your provider to change the policy
(10:54:24 AM) richard.barnes: Lem: Other parts of OMA/OGC are working privacy
(10:55:16 AM) richard.barnes: Medland: In practice, we know at the PSAP what the visited network is, and we request location from them. If it is a roaming user, then we don't get location.
(10:55:41 AM) richard.barnes: This indicates that they're not using RLP. How long has RLP been out?
(10:56:03 AM) richard.barnes: Lem: There are operators that don't even support MLP; fewer support RLP.
(10:56:51 AM) richard.barnes: Marcus: Is RLP not covered by existing roaming agreements? If not, then you would expect delays.
(10:57:02 AM) richard.barnes: Lem: Sure. RLP doesn't work without roaming agreements.
(10:57:13 AM) richard.barnes: Marcus: So it wouldn't be incorporated in an existing roaming agreement.
(10:57:34 AM) richard.barnes: Lem: Right. And location is not high on operators' list of priorities
(10:57:43 AM) richard.barnes: (beyond cell-id)
(10:59:52 AM) richard.barnes: Moving to U-2010
(11:00:03 AM) richard.barnes: Bryan McLaughlin briefing
(11:00:24 AM) richard.barnes: http://www.tschofenig.com/twiki/pub/EmergencyServices/EswAgenda2007/U-2010.ppt
(11:01:36 AM) richard.barnes: Slide [X] - Key Data
(11:01:45 AM) richard.barnes: Slide 4
(11:03:08 AM) richard.barnes: Slide 5
(11:04:37 AM) linsner entered the room.
(11:06:05 AM) richard.barnes: Slide 6
(11:09:11 AM) richard.barnes: Slide 9
(11:10:11 AM) richard.barnes: Slide 10
(11:14:46 AM) richard.barnes: Slide 11
(11:17:45 AM) richard.barnes: Slide 14
(11:19:53 AM) richard.barnes: Slide 17
(11:20:00 AM) richard.barnes: Slide 18
(11:21:19 AM) richard.barnes: Slide 19
(11:21:48 AM) richard.barnes: Slide 21
(11:22:58 AM) richard.barnes: Slide 23
(11:24:42 AM) richard.barnes: Rosen: They just did this in the US (they = Comcare, OASIS, et al.)
(11:24:56 AM) richard.barnes: McLaughlin: Thanks for the info. This is broaded scope.
(11:25:03 AM) jchiaramonte: http://www.comcare.org/HAVE.html
(11:25:47 AM) jchiaramonte: and: http://www.comcare.org/Patient_Tracking.html
(11:26:02 AM) richard.barnes: Thanks John, duly noted. Will distribute.
(11:27:11 AM) richard.barnes: Slide 24
(11:27:41 AM) richard.barnes: Slide 26
(11:29:42 AM) richard.barnes: Rosen: The PSAP needs to be part of this network, i.e., the ESInet. Everyone is on this network, PSAPs, responders, response managers, etc
(11:30:20 AM) richard.barnes: E.g., requirements are affected by the inclusion of PSAPs, but they need to be on there
(11:30:49 AM) richard.barnes: McLaughlin: Your ESInet is a great vision. Right now, everything is a silo.
(11:31:41 AM) richard.barnes: McCann: You've got a problem with WLAN -- you need access.
(11:32:01 AM) richard.barnes: Authentications steps aren't included in your example
(11:32:23 AM) richard.barnes: Most of the time, public hotspots require either WWW interactions or cryptographic credentials
(11:32:28 AM) richard.barnes: That's a problem in an emergency
(11:32:41 AM) richard.barnes: Can't count on hotspots to give access.
(11:32:47 AM) richard.barnes: Maybe also a problem for WiMAX
(11:33:47 AM) richard.barnes: McLaughlin: We're aware of the issue. We're proving the ability of the tech to deliver the services. Access control is a policy requirement, that we can take in later
(11:34:24 AM) richard.barnes: McCann: You have to identify who the right people are to get the right policy enforced.
(11:35:11 AM) richard.barnes: Rosen: We've found that the definition of "responder" is much wider than people think. E.g., school bus drivers are responders for moving people.
(11:35:27 AM) richard.barnes: So the "BusDriverNet" needs to be tied into the ESInet
(11:36:29 AM) jchiaramonte: provided appropriate policies are in place to handle these "non-traditional" responder networks / systems
(11:36:34 AM) richard.barnes: McLaughlin: Trying to tie together IP with Tetra, etc
(11:38:05 AM) richard.barnes: Latif taking the mic
(11:38:09 AM) richard.barnes: <http://www.tschofenig.com/twiki/pub/EmergencyServices/EswAgenda2007/PSCE_Project.ppt>
(11:38:14 AM) richard.barnes: Slide 1
(11:40:04 AM) richard.barnes: Slide 2
(11:41:14 AM) richard.barnes: Slide 3
(11:41:28 AM) richard.barnes: Slide 4
(11:42:06 AM) HannesTschofenig entered the room.
(11:42:31 AM) richard.barnes: Slide 5
(11:42:33 AM) richard.barnes: Slide 6
(11:43:17 AM) HannesTschofenig left the room.
(11:44:09 AM) richard.barnes: Slide 7
(11:51:35 AM) richard.barnes: Slide 9
(11:54:32 AM) richard.barnes: Slide 10
(11:55:30 AM) richard.barnes: Slide 11
(11:58:00 AM) richard.barnes: Slide 12
(11:58:12 AM) richard.barnes: Slide 12
(11:58:15 AM) richard.barnes: Slide 13
(12:00:18 PM) richard.barnes: Marcus: Not familiar with regulations of 4.9GHz
(12:01:10 PM) richard.barnes: McCann: 802.11p is working on vehicle-to-vehicle in the 4.9GHz band; their work is maturing
(12:01:19 PM) richard.barnes: Since IEEE is international, their work should be applicable worldwide
(12:01:55 PM) richard.barnes: Worstell: 700MHz is now open for auction; UHF is being moved out, digital services moved in.
(12:02:07 PM) richard.barnes: FCC has authorized part of 4.9GHz band for emergency services
(12:02:37 PM) richard.barnes: Linsner: Believe 700MHz is split into multiple bands, one of which requires the buyer to provide ES access
(12:04:24 PM) richard.barnes: Slide 14
(12:06:12 PM) richard.barnes: Slide 15
(12:06:17 PM) jchiaramonte: Here's a reference link for 700 MHz auction: http://arstechnica.com/news.ars/post/20070815-700mhz-auction-whats-really-up-for-grabs-and-why-it-wont-be-monopolized.html
(12:08:01 PM) richard.barnes: Slide 16
(12:08:46 PM) klaas: what's the page that holds links to all presentations? i don't think i received the e-mail with the url...
(12:08:59 PM) jchiaramonte: http://www.tschofenig.priv.at/twiki/bin/view/EmergencyServices/EswAgenda2007
(12:09:05 PM) richard.barnes: http://www.tschofenig.com/twiki/bin/view/EmergencyServices/EswAgenda2007
(12:09:12 PM) richard.barnes: yeah, what John said
(12:09:28 PM) klaas: tks!
(12:09:33 PM) richard.barnes: Slide 17
(12:10:06 PM) richard.barnes: Slide 18
(12:11:15 PM) richard.barnes: Slide 19
(12:11:28 PM) richard.barnes: Slide 20
(12:15:43 PM) richard.barnes: Slide 21
(12:16:30 PM) richard.barnes: End of PSC briefing
(12:16:55 PM) richard.barnes: Marcus: You're probably aware the EU is in the final stage of its review of the regulatory framework
(12:17:20 PM) richard.barnes: Spectrum allocation was probably a part of that, but that's an area where the balance between EU and members needs to be managed
(12:17:50 PM) richard.barnes: One example that looms large is when a jet crashed on a bridge between DC and VA
(12:18:00 PM) richard.barnes: Police from both jurisdictions responded, but couldn't communicate
(12:19:04 PM) richard.barnes: Ladid: In the EU, you have the additional complication that there's no-one in charge
(12:19:24 PM) jchiaramonte: (Air Florida Flight 90 - Jan 1982)
(12:22:49 PM) klaas left the room.
(12:22:55 PM) andres.kytt@gmail.com left the room.
(12:23:53 PM) richard.barnes: oops, stopped scribing for a minute
(12:23:59 PM) richard.barnes: reviewing the agendas for the next few days
(12:24:14 PM) richard.barnes: adjourning for the day...
(12:26:36 PM) jchiaramonte left the room.
(12:31:17 PM) linsner left the room.
(12:31:17 PM) linsner entered the room.
(12:37:04 PM) linsner left the room.