#opendaylight-group-policy: Group Policy Status

Meeting started by readams1 at 19:56:55 UTC (full logs).

Meeting summary

    1. https://plus.google.com/hangouts/_/calendar/ZHZvcmtpbkBub2lyb25ldHdvcmtzLmNvbQ.nf43n3np1b0kc48it41dsg4v2o (alagalah, 19:59:27)
    2. https://plus.google.com/hangouts/_/calendar/ZHZvcmtpbkBub2lyb25ldHdvcmtzLmNvbQ.nf43n3np1b0kc48it41dsg4v2o (readams1, 20:02:13)

  2. Actions from last week (readams1, 20:05:30)
    1. Datastore: they tentatively stated we'd used MD-SAL and figure out how to implement sharding (readams1, 20:06:55)
    2. Need to benchmark and verify working assumptions (readams1, 20:07:15)
    3. repository/endpoint: working on defining basic outline, need help with MD-SAL best practices (readams1, 20:07:44)
    4. Jan will help to build the model (readams1, 20:09:23)
    5. ACTION: alagalah readams1 jmedved to meet Fri Apr25 after ARCH meeting for 90min to build Yang for Endpoint Reg and a subset of the UML Model with potentially a follow up Tue Apr29 with jmedved and Tony (alagalah, 20:12:16)
    6. Jan may invite other people to help with this. (readams1, 20:13:00)
    7. ACTION: Prior commitment to have Endpoint Reg strawman for Mon Apr28 may not be feasible. (alagalah, 20:13:15)
    8. Renderer meetings on hold until readams1/alagalah get strawman Renderer. May be predicated on YANG/MD-SAL discussions so will have better idea Monday Apr28 (alagalah, 20:14:44)

  3. Operational state (readams1, 20:15:05)
    1. What does "operational state" mean? (readams1, 20:16:10)
    2. is it about instrumentation? Or is it about making sure that the end-user readily consume group-based policy? (readams1, 20:17:41)
    3. Or is it operational data such as link-state information, etc? (readams1, 20:18:40)
    4. Does it have to do with the "enforcement/exception database"? (readams1, 20:19:14)
    5. People will be want to analyze logs and also based on specific policies or objects (readams1, 20:20:34)
    6. compliance (readams1, 20:20:39)
    7. ACTION: alagalah will talk to dvorkinista to clarify (readams1, 20:22:21)
    8. divorkinsta says it's about: capture of the general state and exceptions associated with it (readams1, 20:26:08)
    9. counters, measurements, exceptions, overall health, debugging, etc (readams1, 20:26:49)
    10. ACTION: dvorkinista attends and clarifies OPSTATE. (alagalah, 20:27:35)

  4. Renderer (readams1, 20:28:06)
    1. We need to settle on what the downstream prototype plugin would work (readams1, 20:28:39)
    2. we'll assume for now we'll use OpenFlow implementation, but we'll investigate what's the easiest (readams1, 20:29:46)
    3. Affinity might also provide a reasonable renderer (readams1, 20:30:14)
    4. Openflow/ovsdb may be the most understandable, but we'll have to see (readams1, 20:31:43)
    5. ACTION: alagalah will look at the different projects to figure out what has the most reasonable northbound API (readams1, 20:32:54)
    6. Another possibility would be netconf/ACL/whatever (readams1, 20:33:42)
    7. options are: opendove, ovsdb, netconf, openflow (readams1, 20:34:17)
    8. openflow/ovsdb are known quantity, not tied to a vendor (readams1, 20:34:52)
    9. Note that this isn't the only rendered: just the first renderer (readams1, 20:35:13)

  5. Helium release plan (readams1, 20:35:21)
    1. https://wiki.opendaylight.org/view/Simultaneous_Release:Helium_Release_Plan (alagalah, 20:35:57)
    2. I don't see the desktop. (dconde, 20:36:25)
    3. now I do. (dconde, 20:36:27)
    4. Do declare intent to participate in simultaneous release, we need to publish our release plan for public comment (readams1, 20:37:59)
    5. project plan needs to align to the milestone dates (readams1, 20:38:13)
    6. release plan template is still for Hydrogen (dconde, 20:38:33)
    7. release plan template needs to be updated (readams1, 20:38:48)
    8. does it resemble a waterfall project plan? (dconde, 20:39:18)
    9. regxboi states it is. (dconde, 20:39:23)
    10. https://wiki.opendaylight.org/view/Open_DOVE:Release_Plan_2013 (regXboi, 20:39:57)
    11. can change the draft release plan until M2 (dlenrow_, 20:40:21)
    12. big milestones are M4 and M5 (dconde, 20:40:38)
    13. plans are not supposed to change after M2. (dconde, 20:41:19)
    14. most project didn't really hit milestones for Hydrogen (readams1, 20:41:48)
    15. there's some tolerance for minor misses (readams1, 20:43:08)
    16. two weeks left to create a plan. (dconde, 20:43:22)
    17. we cannot put features into patches. (dconde, 20:44:16)
    18. ACTION: alagalah will build a sophisticated and complete draft plan (readams1, 20:44:52)
    19. release plans are an honor system. (dconde, 20:44:55)
    20. agendas = good (readams1, 20:48:28)


Meeting ended at 20:49:42 UTC (full logs).

Action items

  1. alagalah readams1 jmedved to meet Fri Apr25 after ARCH meeting for 90min to build Yang for Endpoint Reg and a subset of the UML Model with potentially a follow up Tue Apr29 with jmedved and Tony
  2. Prior commitment to have Endpoint Reg strawman for Mon Apr28 may not be feasible.
  3. alagalah will talk to dvorkinista to clarify
  4. dvorkinista attends and clarifies OPSTATE.
  5. alagalah will look at the different projects to figure out what has the most reasonable northbound API
  6. alagalah will build a sophisticated and complete draft plan


Action items, by person

  1. alagalah
    1. alagalah readams1 jmedved to meet Fri Apr25 after ARCH meeting for 90min to build Yang for Endpoint Reg and a subset of the UML Model with potentially a follow up Tue Apr29 with jmedved and Tony
    2. alagalah will talk to dvorkinista to clarify
    3. alagalah will look at the different projects to figure out what has the most reasonable northbound API
    4. alagalah will build a sophisticated and complete draft plan
  2. jmedved
    1. alagalah readams1 jmedved to meet Fri Apr25 after ARCH meeting for 90min to build Yang for Endpoint Reg and a subset of the UML Model with potentially a follow up Tue Apr29 with jmedved and Tony
  3. readams1
    1. alagalah readams1 jmedved to meet Fri Apr25 after ARCH meeting for 90min to build Yang for Endpoint Reg and a subset of the UML Model with potentially a follow up Tue Apr29 with jmedved and Tony
  4. UNASSIGNED
    1. Prior commitment to have Endpoint Reg strawman for Mon Apr28 may not be feasible.
    2. dvorkinista attends and clarifies OPSTATE.


People present (lines said)

  1. readams1 (40)
  2. dconde (10)
  3. alagalah (8)
  4. regXboi (5)
  5. odl_meetbot (4)
  6. dlenrow_ (4)
  7. s3wong (2)
  8. jmedved (1)


Generated by MeetBot 0.1.4.