ATOCA 83

Chair presented agenda
Chair presented working group status
General agreement that milestones were behind schedule, but no changes were need to the actual milestones other than their dates


Hannes Tschofenig presented on requirements document

Discussion of scaling issues, discussion focussed on scenarios where stateless delivery is possible, vs. stateful (i.e. pub-sub)

General agreement that there will be multiple mechanisms for delivery, potentially including a minimal SIP SUB-NOT, possibly a 
link layer mechanism, and maybe other - already standardized - mechanisms such as XMPP or ATOM. 
Martin Thomson : Requirements and solution might not require state maintenance in order to provide service.  Scaling cheaply depends on stateless operation.
Richard Barnes : State may be unavoidable in many deployments
General Agreement that a solution that depends on pub-sub or unicasting will probably require state, but requirements should not assume that state exists

Discussion naturally progressed into the topic of discovery


Brian Rosen presented on discovery

Use cases for discovery (local vs. non-local) were non-controversial

Information needed from discovery was non-controversial

Mechanisms were a little controversial, but there was little time to continue


Impromptu discussion on SIP-CAP

Discussion on whether this should be adopted
Questions about why it was dead, since it came up in discussion frequently
James Polk asked about status as a WG item - a: it is not one
Suggestion from the chair that the authors revise the draft and provide an updated individual submission, agreement recorded