-
Notifications
You must be signed in to change notification settings - Fork 0
Minutes 2017 09 27
ericvoit edited this page Sep 27, 2017
·
3 revisions
Meeting Materials | Attending |
---|---|
WebEx Recording password: YgtY8KMn | Tianran Zhou, Xufeng Liu, Eric Voit, Zhengguangying (Walker), Henk Birkholz, Alex Clemm, Igor |
Reviewed the scope proposals (hyperlinked below)
-
COAP Facts relevant to our subscription work
- COAP has sequence numbers
- After 1000 messages/sec, COAP is not guaranteed to keep messages in order
-
Existing COAP subscription efforts are in flight for SmartThings.
- Michael Costa, Ari (Ericsson), Carston Borman, etc. involved here what level of interaction should we have.
- Does dynamic subscriptions, perhaps not configured. Might this be interesting?
- There subscriptions will be simpler than what we are attempting to subscribe against.
- As COAP pub-sub doesn't have a yang datastore, could subscribed-notifications be reused? Coap could identify a resource via "Stream".
- Henk to investigate whether the other authors might be interested in joint work
- Eric & Henk to meet on developing the "what" statement for next week.
-
Alex' proposed scope mostly accepted
- Need to nail down the selected stateful filters
-
Intersections with the Automation team
- Can incrementally add in the expressions as defined in the automation team, but there will be others as well not coming from but these are not a gating factor to progress
- Reviewed were the discovery and configuration elements
- People ok with using capabilities exchange mechansim
- Floated a deferral of the HTTP2 draft as there is no capabilities exchange. Instead, we could mandate the new header for this draft. I.e., this makes HTTP2 draft dependent on this one.
- It was interesting to consider per-subscription signaled header requests.
- For example, I want to subscribe to "foo", but only if you can send me encrypted, signed, and/or attested push updates.
- This would be its own draft, as it would enhance the subscribed-notifications, rather than being incorporated into this document.
- Eric to send out a link to current version
- Should refine and propose a problem statement document for non-adopted elements
- Determine and review with potential authors, as shown from first minutes on the site
- People will commit (or not) to being authors on the drafts after a review of problem statements next week.
- Note: Likely it is good to also float a problem statement draft for multi-line card. This avoids any questions on whether this was adopted or not.