Casey joined briefly to let us know he is unwell and is seeking a LF stand-in to work with the PNDA team in his absence. Charter and governance agenda items are currently on hold.
The ONAP DDF took place in Paris last week. Two topics of interest to the PNDA community:
PNDA - DCAE Discussion
We held a well attended round-table discussion about the status and future of PNDA in DCAE for Dublin and beyond.
- Interest from DCAE team in seeing further development of PNDA integration in Dublin – catalog integration and data integration.
- A lot of discussion around what data to collect and store in PNDA. Existing TCA application is policy driven and only collects minimum data for TCAs defined in policy.
- Wider data collection for offline analytics was not universally endorsed.
- Lots of interest in seeing a containerised PNDA that can be deployed on any Kubernetes infrastructure – no commitment to contribute from anyone.
- Much discussion about analytics apps on PNDA – no commitment to contribute.
- Interest in analytics app that can join layers of data collected from infra, compute, VNFs, etc. to calculate fault or performance impact.
- No commitment from anyone in the room to deliver any specific work into PNDA or the PNDA / DCAE integration effort.
Distributed Analytics as a Service
This work is ranked priority #3 for the Dublin release, with a firm resourcing commitment from Intel and VMware: https://wiki.onap.org/display/DW/Dublin+Release+Requirements
The common thread to both of these discussions is that there is considerable interest from the ONAP/DCAE community in having a fully containerised PNDA that can support flexible deployment scenarios as part of an ONAP system.
Cloud Native PNDA
Unknown User (trsmith2) provided these Notes on Cloud Native PNDA from a PNDA design meeting last year. This provides a very useful analysis of what is needed to produce a fully containerised cloud-native PNDA platform. We need to collate the concrete design work into PDP-12: Cloud-native PNDA and add items to JIRA for the pieces of work as we identify them.
Carlos Giraldo Rodriguez provided an update about the Kubernetes orchestrated hadoop experimentation that they are working on in their lab. They hope to contribute back their findings to the PNDA wiki.