Agenda: 

  • Architecture
    • Flush out low level detailed architecture, including what to focus now, vs and what to postpone
    • Detail work flow and API design
  • Release Planning for R1
    • Release scope and planning template (Release Planning). We need to deliver “Project Deliverables Defined” by 6/29
    • Project milestones
    • JIRA logistics
  • Integration Lab dependency update

Attendees:

Anbing Zhang, Bin Hu, Bin Yang, Brian, Danny Lin, Eric Troup, Jinhua, Matti Hiltunen, tapanm, Xinhui Li, Maopeng Zhang

Minutes:

Danny:

  • There should be consistent representation in Arch. Subcommittee and encourage anyone who is interested in that.
  • Now it is time to focus on what we can deliver for R1, we need to flush out API, detail sequence and workflow diagram.
  • For the arch. discussion for R1, In addition to what we want to support, we also want to highlight what we do not plan to support.

Eric Troup:

  • Suggest to have on-boarding process for each cloud platform , there should be some mechanisms to document what each cloud platform does and how it does it, where those APIs are. Then we can work through a set of use cases.
  • Danny talked about documentation projects help on that goal, and VoLTE and vEPC (do you mean vCPE actually?) use cases would help drive the APIs definition.

Brian:

  • Mentioned that vFW should be the supported use cases for R1 as well. Brian volunteered to help on sequence diagram for the existing vFW in OPEN-ECOMP. Brian raised the question on how other cloud (VMware, Azure) support mechanism similar to OpenStack's cloud-init

Action:

  1. Brian will help on sequence diagram for the existing vFW in OPEN-ECOMP.
    Danny: To brian's question, both Vmware and Azure support that. And the question will be how to abstract that in a way can be understood by different cloud providers. Suggest to do that in specific use cases.
    Brain raised the question about clients of Multi-VIM/Cloud, Danny and Bin Hu shared the perceptions from DCAE and SDN-C:  There will be a adapter in SDN-C which will interact with Multi-VIM. SO will also interact with Multi-VIM and the network will be embedded in the heat template.
    Brian raised the question whether and how Multi-VIM/Cloud plays roles in CI/CD.
    Called for need volunteers to reach out to the CI/CD to get details on robot

  2. Brain take the action to do research of Robot with integration team Helen.
    Danny had reach out to SO, no response yet. The intention is to clarify the confusion between Infra-C and Multi-VIM: Infra-C is a component inside Multi-VIM/Cloud. SO should go with Multi-VIM/Cloud.

  3. Danny will continue to reach out to SO team to clarify the concepts of Infra-C and Multi-VIM
    Talked about provider registration and think it is essential piece of this project in a long run, but might be not in scope of R1.
    Danny proposed to take the vFW and vEPC use cases, deep dive them to identify the points of interaction with Multi-VIM/Cloud and document them from perspective from Multi-VIM/Cloud.

  4. Danny, Bin Hu, Bin Yang and someone from Microsoft will work on these diagrams to identify interactions from perspective of Multi-VIM/Cloud.
    Release logistics: solicit volunteers to propose or define milestone for this project from point of view of  this project.

  5. Eric Troup will give some inputs on the framework approach which should be part of the plan.
    Eric Troup suggested that all participates put their name and company.
    Bin Hu updated the jira : Eric has pick up the action item.

  6. Bin Hu will make sure this project is created on the jira.
    Danny introduced the mesh network (the virtual lab on the wiki:https://wiki.onap.org/display/DW/Virtual+lab) which deploy clouds and network elements in a distributed environment and connect them via VPN.
    Eric and Brian figured out what need to be done from perspective of Azure to support this virtual lab. And it was mentioned that AT&T has been doing this work and can be leveraged.

  7. Eric will have guest account setup for this purpose. Eric will also make sure someone from microsoft to participate this discussion. In addtion to regular meetings occurs every Friday, there will be frequent meetings which at least all meetings should join.

  8. All committers should get consensus on time slot for frequent meetings (at least Monday and Wednesday next week) to work on architecture design and sequence diagram for MultiCloud project
  • No labels