06-21-2017 Policy Weekly Meeting

Owned by Pamela Dragosh
Last updated: Jun 21, 2017
2 min read

ONAP Policy Weekly - 06-21-2017.pptx

Link to Attendee List - TBD

Meeting notes - June 21, 2017

 

New members:

Aimee Ukasick - AT&T - focuses on OPNFV, closed loop control, VNF policies.

Bin - Windriver - committer for the VoLTE use case project. (PLD - I think MultiVIM??)

Lizi - ZTE, Open-O, interested in Policy, could be a contributor.

Ruan - from Orange, works on XACML Policies.

Xinyuan Wang - from ZTE, contributor in Policy.

Yuan Liu - China Mobil, would like to be a contributor.

Vijay VK - AT&T, DCAE team.

 

Current efforts:

  1. We are currently working on pushing the Yaml framework (high level policy specification) and the Policy Guard framework to ONAP.
  2. Jorge and Pam are working on initial plan for release planning. Watch the wiki page: https://lf-onap.atlassian.net/wiki/display/DW/Policy+R1+Release+Planning
  3. Pam will figure out the PLT selection process (there was an email describing it…).

 

We will have 3 educational presentations/demos on Policy:

  1. Jorge - how does the current platform work, focus on control loops. Vijay’s suggestion: maybe Jorge can focus on what others (e.g., Alex) need (e.g., specific modules for VNF orchestration).
  2. Chenfei and Michael - TOSCA for Policy.
  3. Ankit/Sastry/? - SNIRO and how Policy supports it.

 

On the wiki there are two presentations on PLT and Release Planning. It would be useful for everybody to review them and if you have any questions, we can discuss.

https://wiki.onap.org/display/DW/Presentations?preview=%2F3246513%2F6592765%2FONAP-Project+Roles+v1.pdf

https://wiki.onap.org/display/DW/Presentations?preview=%2F3246513%2F6593134%2FONAP-ReleasePlanningTraining.pdf

 

Jorge question - when integration team will be ready with the guidelines? - maybe in October. Pam and Jorge will need to prioritize how to deal with it.

 

Alex’s questions/comments: 

  • Currently, all the knowledge is with AT&T folks, thus, creating multiple projects may create more pain. So maybe this is good for future releases but for now, maybe it is better to keep everything together. (Vijay Gopalakrishnan - agrees with Alex.)
  • Are policies tenant-centric, system-centric, what are the requirements for VNFs - hardware centric… what recourses get consumed by what tenant, by what VNF,… service plans, service levels… Does Policy support this today? What are the generic policies we can define today to optimize VNFs? Then, after answering this, we can identify the gaps.
  • VFC vs APPC - they are pretty much different. We will need to figure the differences. Pam - we should write code that supports both, we should capture the information at the highest possible level. How do we easily capture the policy information.
  • We don’t have any notion of resource pools - to group things together (for VNFs, for network services…). David Chadmi and the folks from Israel - from SDC, we need to discuss with them the integration with Policy.
  • Tool for UML. Alex will ask for the UML plugin for Confluence.

 

Pam - Some folks are interested in Control Loop and some in VNF orchestration. So, we have two tracks: VNFs orchestration and Control Loops. Then we can decide who can contribute to what track.