Versions Compared

Old Version 97

changes.mady.by.user Randa Maher

Saved on

compared with

New Version 98

changes.mady.by.user Randa Maher

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...


6/19/17 - APPC Project Kick-Off

...

  • Decision: VID will provide Controller Type Property to SO so that SO can post request to correct DMaaP topic; this should be configurable. VID & SO will work together to flesh out the details.
  • Assumption is that user creating the scaleout request will know which controller needs to handle the request to scaleout.
  • Seshu will follow-up to determine out SO uses the APPC Client library and get back to us if additional discussions are needed.
  • Action: Another meeting is needed to since SDC representative was not able to make meeting. Follow-up sessionwill be planned for Monday, 2/12.
    • Open issue: What identifier can APPC use that uniquiely identifies the VfModule and does not change with environment.
      • Update: vfModuleModelName is proposed. See notes below.

02/12/18 - Scale Out dependency & cross product coordination - Continued

Participants: Takamune Cho, Vidya Jog, Paul Miller, Scott Seabolt, Lior Nachmias, Steve Smokowski, Scott Blandford, Lauren Lewis, Michael Lando, Dan Timoney, Prabhuram Somasundaram, David Ho, Shailendra Borale, Michael Zinnikas

Purpose of Meeting: Determine what identifier can be used by APPC that uniquely identifies the VfModule and does not change with environment

Conclusion

  • APPC can use the vfModuleModelName. They will have to traverse the AAI data to get to it; VID must pass the vfModule ID in the payload of the ConfigScaleOut request. Using this vfModule ID, APPC can retrive the model name. APPC would have to traverse AAI data (VF Module ID → Model Version ID → Model  → Model Invariant ID → Model  Version → Model Name)
  • APPC will be flexible in it's design -
    • If it gets controller-template-id in the payload, it will use this value to retrieve the correct ConfigScaleOut template.
    • If it does not get controller-template-id in the payload, it will use the vfModule ID, also expected in the payload, to retrieve the vfModuleModelName from AAI and use that to map to the correct ConfigScaleOut template.

Action items:

  • Michael Lando- follow-up to determine if the content of the CSAR changes per environment
  • Michael Lando - follow-up to determine if SDC is setting workload context.
  • Randa Maher to check with Ofir Sonsino and Lauren Lewis to make sure VID has requirement to include one of the above cases, using vfModuleID would be prefereable.
  • Randa Maher to setup a follow-up call to have further discussion on VID to SO requirements to determine which controller a request needs to go to.

Additional Notes

  • Steve Smokowski had some concerns around doing throw-away work to support multiple controllers; ideally, the model should define which controller manages a given VNF and that should be information that is part of the onboarding process for any VNF. However, we are not at that level of maturity yet and the issue extends beyond scale out scenario. The question is what is doable in Beijing, which is why using VID was the interim solution proposed. Some more discussion will take place on this topic. 

02/13/18 - Scale Out dependency & cross product coordination - VID to SO Controller Type- round 2

Participants: Takamune Cho, Vidya Jog, Paul Miller, Scott Seabolt, Scott Blandford, Lauren Lewis, Michael Lando, Dan Timoney, Prabhuram Somasundaram, Ofir Sonsino

Purpose of Meeting: A follow-up session was requested to revisit the decision made on 2/7/18 regarding how we determine to which controller to route the request. 

Conclusion:

  • SO was not able to make meeting, so quoram for discussion was not achieved.
  • The decision made on this issue is transparent to APPC; i.e., does not affect APPC requirements.
  • Scott Blandford and Lauren Lewis will follow-up with VID and SO teams to get alignment on this particular issue.

02/14/18 - APPC Project Weekly Minutes

Agenda:

  • Sprint 2 status
  • Misc

Participants:

Actions from last meeting:

  • Need to follow-up with SDC again on their plans to move to onap from openecomp; we still have not gotten feedback on timeline for this dependency yet. Takamune Cho to follow-up with SDC,
  • Sai Gandham to schedule meeting with APPC team to review his design plans for AAF-91.
    • Update: Meeting moved to this Thurs, 2/8/18

New Actions Items:

Notes: