Tentative SO Impact

(need verification of SO PTL and stakeholders)

Assumptions & Decisions

Possible SO Changes Needed

Opt. 1 Modify the existing flows (used by VCPE in amsterdam) to do service decomposition and call the OOF APIs to perform the optimize service/VNF homing and placement.

1. Change to Generic Service Level Workflows - accommodate decompose, homing/placement
   1. Call "Decomposition Building Block" and parse TOSCA Service Template Content
      1. Determine Resource Node Types
      2. Determine Sequence of Node Types using template content
   2. Call "Homing Building Block" - which calls SNIRO
      1. Flow sends "Customer_Location" in homing input for vCpeResCust Service Model
      2. SNIRO would determine vGMUX instance for TunnelXConn, based on latency policies for "Customer_Location"
      3. Returns to Service Level Flow - "homing solution":
         1. TunnelXConn Home = vGMuxInfra UUID
         2. vG VNF Home = vGMUX Cloud Zone
         3. BRG PNF Home = N/A
2. Change to Generic Resource (VNF) Level Workflows?
   1. TBD
3. Addition of HPA Sub-Workflow to handle OOF HAS API interaction?
   1. How does "Get Optimized Service/VNF Homing/Placement" happen? See: OOF-SO Interaction in R2
   2. What are changes needed to make HAS interaction happen, can they all be handled in a workflow (either, sub-flow or generic flow)?
4. Modify Sub-Workflows for HPA to handle interaction with SDNC, MultiCloud, APPC/VFC?
5. What are all of the specific SO workflows used in vCPE Use Case?

Opt. 2 Introduce new flows that would invoke the TOSCA DO inside SO to perform the decomposition and OOF interactions

Note: Will TOSCA based DO be delivered by R2 or earlier enough to add functionality?  See SO-22

1. Create new workflow invoking TOSCA DO
   1. Decompose Services/VNFs
   2. Get Homing/Placement from OOF

SO JIRA Epics & Stories