#opendaylight-group-policy: policy status

Meeting summary

1. requirements (regXboi, 20:08:29)
   1. https://wiki.opendaylight.org/view/Group_Policy:Sub-Groups:REQUIREMENTS#Requirements_Sub-Group (regXboi, 20:08:40)
   2. requirements meeting to be set up (regXboi, 20:09:18)
   3. ACTION: dvokinista to set up requirements meetings (regXboi, 20:09:29)
   4. for those joining, the hangout is coming (regXboi, 20:10:43)
   5. https://plus.google.com/hangouts/_/noironetworks.com/odl-gbp-status (regXboi, 20:10:50)
   
   
2. functional spec (regXboi, 20:12:08)
   1. https://wiki.opendaylight.org/view/Group_Policy:Architecture/Policy_Model (regXboi, 20:13:41)
   2. folks need to read the functional spec and comment, as implementation is under way (tbachman, 20:13:57)
   3. The Architecture page has a link to the Policy Model page (tbachman, 20:14:36)
   4. the main wiki page now has a subpages section (regXboi, 20:17:15)
   5. readams is working on renderer-commons framework (tbachman, 20:19:13)
   6. Trying to get to a point where a renderer can map a parir of endpoints into endpoint groups (tbachman, 20:19:47)
   7. demo-renderer will demonstrate that the basic framework “works" (tbachman, 20:20:55)
   8. likely a simple overlay (regXboi, 20:21:17)
   9. readams says anybody who wants to take a swing at it can feel free (regXboi, 20:22:20)
   10. readams thinks that will be more easily filled in as we understand what the renderers expect (regXboi, 20:22:47)
   11. https://lists.opendaylight.org/pipermail/groupbasedpolicy-dev/2014-May/000193.html <— dlenrow mentioning the renderer group (tbachman, 20:28:29)
   12. regXboi recalls that dvorkinista made a statement that portions of the model are optional (tbachman, 20:30:32)
   13. If that is true, regXboi looked at the current YANG in the repo, but nothing is tagged as a yang feature, which would indicate that it was optiona (tbachman, 20:31:03)
   14. readams notes that entire pieces of the model aren’t necesarrily optional, but that the renderer may not be able to render parts of the model (tbachman, 20:31:32)
   15. which would result in an exception, then having retries (tbachman, 20:31:45)
   16. resulting in parts of the model not being required to fulfil the policy requirement (tbachman, 20:32:14)
   17. dvorkinista says that there are things that require change of placement, and stuff that does not. (tbachman, 20:32:40)
   18. If there are endpoints that can’t be satisified, then the endpoint must be moved somewhere else (tbachman, 20:32:57)
   19. because we can’t move workloads (tbachman, 20:33:06)
   20. We may want the option to have a way of expressing whether the policy has to be “all-or-nothing” (tbachman, 20:35:09)
   21. dvorkinista feels that this can fall on the renderer. The renderer is responsible for achieving the desired state. (tbachman, 20:36:25)
   22. sanjay asks if the renderer architecture been decided on? (regXboi, 20:38:07)
   23. readams says that we’re creating a demo-renderer (tbachman, 20:38:41)
   24. readams says that demo will be based on *something* (regXboi, 20:38:43)
   25. regXboi points out that *something* is a bit nebulous (regXboi, 20:40:32)
   26. and that it should be nailed down by the to-be reconstituted renderer meetings (regXboi, 20:40:52)
   27. AGREED: renderer meetings will be restarted (regXboi, 20:41:03)

Action items

1. dvokinista to set up requirements meetings

People present (lines said)

1. regXboi (25)
2. tbachman (18)
3. odl_meetbot (5)
4. Youcef (5)
5. Bhushan_ (2)
6. dconde (2)
7. alagalah (0)