...
- CNTT ... reference implementations and compliance frameworks, etc.
- Serve and grow the community
- Grow awareness of OPNFV work and making sure we are delivering what carriers need.
- Collaborate with vendors and all stakeholders to ensure we are focused on areas of importance to them.
- Track use of OPNFV artifacts (infrastructure, test cases, etc.) both with upstream communities and commercial entities.
- Convince member companies to invest and prioritize developer resources
- Refocus waht what we deliver and refresh tools
- Modularity, loosely couple components, re-usable components: We’ve had two types of deliverables in the past: Scenarios (as “fully tested integrated solution stacks”) and Tools (to conduct that testing). While scenarios have been the focus, it is tools that have survived. Functest, NFVbench, Barometer, Dovetail, … If those would have been even more generic and reusable, we’d have greater reach.
- To achieve Modularity, encourage the various tools to evolve toward more common architecture. Defining an OPNFV target architecture would help, in areas where agreement is possible.
- Evolve to modern tool chains – cloud based solutions which are easy to leverage and replicate.
...
- Which projects are up and/or downstream for your project?
- OpenStack
- ODL
- TF
- Kubernetes
- OVS
- VPP
- DPDK
- etc.
- What collaboration / interaction with other projects is working well?
- OpenStack
- Openstack Edge Computing Group
- Airship (re. CNTT)
- ODL
- OSM collaboration with SFC (currently not active)
- other?
- What collaboration / interaction (existing or desired) with other projects is not working well or is a gap?
- Akraino
- ONAP
- TF
- other?
- In general, collaboration with all other LFN projects is poor in practice, i.e., anything beyond high-level discussions
- What collaboration / interaction would your project benefit from that is not already happening?
- CNTT/GSMA (spinning up)
- Akraino
- ONAP
- TF (switch performance tools)
- Kubernetes networking - Multus, DanM, Calico, etc.
- other?
- How do you suggest improving collaboration / interaction with other projects?
- Collaboration must be a natural fit and mutually beneficial to work
- Education as to benefits of collaboration (specifics)
- Cross-TSC presentations
- other?
- How/What should we drive for collaboration with projects outside LFN that would be beneficial?
- Open source software (OSS) (examples: CNCF, LF Edge, Hyperledger, Openstack, ONF, etc.)
- CNCF, OpenStack, LFEdge (Akraino, ...)
- other?
- Standards Development Organizations (SDOs) (examples: O-RAN, ETSI NFV, GSMA, 3GPP, ETSI ZSM, etc.)
- GSMA (CNTT)
- ETSI NFV - TST WG (testing methodologies for CNTT)
- ETSI-OSM
- other?
- Analytics, ML, AI: Collaborating with LF-Deep would be beneficial to OPNFV
- Other?
...
- Question needs clarification, there are 2 different types of responses ...
- more integration of LFN technologies is good and so more integration is the way to go
- Close integration is not the way to get broader interest and reach. Small re-usable and easily composable pieces is what quite often gives you reach. That way, people can grow their own solutions based on OSS components. Tightly integrated systems lead to a situation where it either “fits your needs” or “does not fit your needs” – and the latter is more common. Look at Unix. Unix is a tool box of loosely coupled components that can easily be combined. OPNFV could be more a tool box for composition and testing than an effort to build the one and only stack.
- OPNFV has been an integration project and is uniquely positioned for LFN technologies to be integrated and tested together
- OPNFV could become the "shop window" for LFN technologies
- Initiatives with important industry goals like CNTT should naturally result in more integration between LFN projects. This is a preferable way to increase industry adoption.
6. The LFN TAC made recommendations on increased efficiencies for LFN project infrastructure. (Link: Infrastructure Working Group Summary Report) Are there other areas of project operations that you would like to see the TAC look into?
...