NETCONF WG 67 th IETF San Diego, CA, USA November 6, 2006
NETCONF WG Details Mailing List Discussion: Subscribe: ‘subscribe’ in the message body Archive: WG Chairs Simon Leinen Andy Bierman WG Charter Page WG Home Page
Administrativia Volunteers Note takers for the minutes Jabber scribe (please announce slide numbers) Jabber proxy Please use the microphones when asking questions and don't forget to state your name
Four drafts are in “Authors 48 hours” RFC numbers have been assigned 4741 – NETCONF base protocol 4742 – NETCONF over SSH 4743 – NETCONF over SOAP 4744 – NETCONF over BEEP Expect publication this month Upcoming RFCs
Active Drafts WG Draft: NETCONF Event Notifications draft-ietf-netconf-notification-04.txt Individual Submissions: Netconf System Architecture draft-atarashi-netconfmodel-architecture-03.txt Experience of implementing NETCONF over SOAP draft-iijima-netconf-soap-implementation-01.txt Generic Action RPC for Netconf draft-lengyel-action-rpc-00.txt Netconf Master-agent Sub-agent Communication Protoc'l draft-kulkarni-netconf-subagent-prot-00.txt
Agenda Agenda bashing and other administrativia (5') NETCONF notifications (as long as it takes) Experiences with NETCONF over SOAP (5', if time permits)
notification-04 issues Meta-issue: apparent lack of thrust 4 th meeting on this doc (including interim) little review, slow convergence would like to WGLC before next meeting Suggested approach Reduce to include only essential mechanisms
notification-04 issue: datamodel I-D includes extensive (read-mostly) data model Application of “eat or own dog food” Introduces (as done here) a normative reference on datamodel draft Do we need to query subscriptions? Do we need to query event streams?
notification-04 issue: filtering Several ways of selecting notifications: event streams one or 1–n per subscription? (StreamsList) filters XPath and subtree specified ad-hoc or by reference to named profile
notification-04 issue: replay Retrieval of saved notifications possible could use normal ? No combined replay/real-time (“tail -f”) replay followed by real-time notifications “temporal ordering” problem Balasz suggests to address this by deferring live notifications on a session during replay Client can specify start- but not end-time