Client workflow support with API interaction with multiple service instances

[xi-yang]

• [ ] 1. Design a slim workflow definition schema similar to FabFed but more compact
• [ ] 2. Develop client methods for managing "instance group" (similar to FabFed workflow session)
• [ ] 3. Experiment and develop new API interaction with SENSE-O to better track and trouble shoot instances in a group.
  • SENSE-O server side change if any will be addressed as an StackV issue in the StackV repo

Target ETC will be in Dec 2024 to early 2025. The release will be aligned with SENSE-O r25.0.

[xi-yang]

Some initial thoughts on the design:

• We consider this workflow as a SENSE service "Client Session" which is like a "Workflow Session" in fabfed. Similarly sessions information will be persisted under a local storage folder. Directory of the folder can be configured in ~/.sense-o-auth.yaml.
• A workflow definition template will firstly include Service Instances each with reference to a service profile. Primitives such as "count" also apply here.
• Simple stitching policy can be embedded into the workflow definition. That will be explicit. I don't think we need the policy being inferred from dynamic information or by addition a policy definition configuration file.
• Additional statements for interaction with Manifest, Discovery, AddressManager and MetadataService API will be required. Those will retrieve information from the API as parameters to bind to variables in the Instance profiles.

[xi-yang]

@abessiari For initial testing scenarios:

1. Simplest test for N x L2VPN instances.

   • Use the Any-to-Any-L2VPN service profile to create multiple (say count = 3) instances in a group.
   • You can change the terminal URI, vlan tag and bandwidth of the L2VPN.
   • Exercise create, cancel and repvovision in a group.
   • Exercise modify of count (optional)

2. A L2VPN only stitching scenario:

   • Use the Any-to-Any-L2VPN-IPv4 profile to create 2x instances. With one stitching to the other. For example,
     • Instance A: urn:ogf:network:maxgigapop.net:2013:ptxn-sense-v1.maxgigapop.net VLAN any to urn:ogf:network:es.net:2013::star-cr6:2_1_c5_1:+ VLAN any.
     • Instance B: urn:ogf:network:icair.org:2013:mren8700:esnet VLAN depending on the produced at urn:ogf:network:es.net:2013::star-cr6:2_1_c5_1:+ by Instance A, to urn:ogf:network:starlight.org:2022:r740xd4.it.northwestern.edu VLAN any
   • Before the above two instances, a AddressManager API call to allocate pair of /30 IPv4 address from the AutoGOLE-IPv4-Test-Pool pool and assign that to Instance A terminal urn:ogf:network:maxgigapop.net:2013:ptxn-sense-v1.maxgigapop.net and Instance B terminal urn:ogf:network:starlight.org:2022:r740xd4.it.northwestern.edu respectively.
   • Exercise create, cancel and repvovision

     in a group.