GeoLocation and Privacy (GEOPRIV) Minutes Meeting : IETF68, Thursday 22 March 2007. Location: Prague, Czech Republic. Chairs : Henning Schulzrinne (acting), Jon Peterson (acting) Minutes : Robert Sparks Version 1 ============================================================= AGENDA A. Agenda Bashing B. Document Status C. Future Directions D. L7-LCP Problem Statement & Requirements E. L7-LCP Protocol Selection F. Geo URI ============================================================== MEETING REPORT A. Agenda Bashing Brian Rosen requested that the agenda be changed so as to focus more on L7-LCP and to not discuss location signing. The agenda was so modified. B. Document Status Jon Peterson points the group to the Last Call of the SIP Location Conveyance draft in the SIP working group. It is noted that draft-ietf-geopriv-pdif-lo-profile will be revised after the meeting. C. Future Directions Cullen Jennings asked the working group to make decisions around the analysis done so far. The group discussed the direction of having just one LCP protocol, more than one Layer 7 LCP protocol, and the basis for that LCP protocol, such as HTTP or SIP. Brian Rosen noted that the decision had already been made. James Winterbottom noted that the working group had accepted a requirements document on L7-LCPs. James Polk noted that there were already more than on LCP and that they were not work product of the IETF but of other SDOs. The chairs asked the room if there should be more than one IETF specified protocol. A strong hum in favar, and no hums opposed was recorded. Strong favor was recorded in the Jabber room. The group discussed subscription based approaches to LCP. Brian Rosen noted that each no option adds to the set of requirements needed by a phone. Rohan Mahy expressed his feelings that such a feature was not specific to LCP but was a general Location-by-Reference fetch mechanism. Keith Drage noted that he thought there was previous agreement that phones will have to implement all LCPs and that suggesting that each LCP was just an option went against that earlier agreement. The chairs asked the room if GEOPRIV should specify an event package that a target can use to subscribe to location, either directly or indirectly. A strong hum in favor and no hums opposed was recorded. No hums were seen in the Jabber room. The group next discussed explicit support for indirect mechanisms. Peter Blatherwick asked for clarification about these mechanisms being layer 2, to which the response from the room was no. The chairs asked the room if GEOPRIV should specifiy an indirect mechanism. A strong hum in favaor and no hums opposed was recoreded. The chairs asked the working group if GEOPRIV should specify a direct mechanism. A weak hum in favor and a strong hum opposed was recorded. Hums in favor from the Jabber room were recorded, but there was confusion about which question the Jabber participants were asking due to audio delay. The chairs asked the working group if GEOPRIV should specify layer 3 solutions with subscription, such as extending protocols like DHCP. A strong hum in favor and some hums opposed were recorded. The chairs noted consensus in favor. The chairs asked the working group if GEOPRIV should pursue solutions for indirect LCP though layer 7 mechanisms. A strong hum in favor and no hums opposed was recorded. The chairs asked the working group if GEOPRIV should pursue an HTTP based approach. A strong hum in favor and no hums opposed was recorded. The working group next discussed on reviewing requirements around dereferencing protocols. Kieth Drage urged the group to build towards the right things. Ted Hardy noted that some confusion may arise from which mechanisms are mandatory to implement. Henning Schulzrinne noted that the hums were intended to prioritize work. The chairs asked the working group if more than one dereferencing mechanism, at least HTTP, should be specified by GEOPRIV. A strong hum in favor and none opposed was recorded. D. L7-LCP Problem Statement & Requirements Henning Schulzrinne gave a presentation on the status of this document and some suggestions to reorganize it. Cullen Jennings stated that privacy and integrity need a viable story on these items before it could be brought in as a working group item. He stated it was not clear that GEOPRIV can build a protocol meeting On-Behalf-Of requirements, and that perhaps they should be pulled out and reintroduced when the working group is more convinced that they can be met. Henning noted that the document is now in working group last call and would need to be pulled back. James Winterbottom noted that he felt these requirements could be met in constrained environments. Cullen stated his belief that it is difficult to accept documents that only work in specialized networks. Henning stated his belief that delaying this document will make it irrelevant. The chairs asked the working group if On-Behalf-Of and LIS-to-LIS requirements should remain in this document. A strong hum in favor and a strong hum opposed were recorded, and the chairs deemed the hum inconclusive. Rohan Mahy asked if the requirements could be put in a separate document. The chairs then asked the room if putting On-Behalf-Of and LIS-to-LIS requirements in a separate document was acceptable. No opposition was recorded. Steve Norryes volunteered to edit this separate document. E. L7-LCP Protocol Selection James Winterbottom presented HELD. Henning Schulzrinne presented RELO. Jon Peterson asked if it is important to have exactly one answer to this question, and noted that other groups could chose one of many. Ted Hardy expressed his desire to go forward with one when the solutions are very similar. Marc Linser noted that if multiple solutions go forward, multiple implementation requirements go forward. Others noted their desire for only one solution. The chairs asked the working group if a single HTTP based L7 LCP solution should be picked. A strong hum in favor and no hum opposed was recorded. The chairs asked the working group if they felt informed enough to make this choice. A strong hum in favor and a weak hum opposed was recorded in the room; hums in favor were recorded from the Jabber room; and the chairs determined consensus to be in favor. Rohan Mahy noted that when we pick one, the current contents of that will be subject to change. The chairs asked if HELD should be a basis for the sole solution. A strong hum was recorded in both the room and the Jabber room. The chairs asked if RELO should be a basis for the sole solution. A strong hum was recorded in the room. The chairs ruled this hum inconclusive. The chairs then asked the working group if it was important to answer this question today. A strong hum in favor and a weak hum opposed was recorded. The chairs then discussed alternate consensus methods. Ted Hardy suggested a plurality vote. The chairs asked the working group if they would accept the decision of a plurality vote. A strong hum in favor and no hums opposed were recorded. The chairs then conducted a plurality vote. 22 votes in favor of HELD were counted in the room, 15 votes in favor of HELD were recorded in the Jabber room. 23 votes in favor of RELO were counted in the room, with Cullen Jennings casting the last vote in the room. The chairs conferred with Cullen Jennings and decided to accept the plurality vote. Ted Hardy noted that the task was now to bring back this decision to the list. He stressed that people who have not yet participated be given the opportunity to bring strong concerns with the process to the list. Marc Linser noted that this decision was not on the agenda. Henning expressed that he would feel better if there were a representative design team on the new basis document. Andy Newton asked that the chairs take a hum on deferring the non LCP parts of HELD. Otmar Lendl noted that the group had not included the identity extensions in the process just executed. F. Geo URI Alexander Mayrhofer presented his draft on a Geo: URI scheme. Several members pointed to issues with the proposal.