Attendees:
Bob Monkman (Deactivated) (Intel)
Timo Perala (Nokia)
kumarcv (Verizon)
cathy zhang(Intel)
Victor Morales (Samsung)
Sharad Mishra (Intel)
Ritu Sood (Deactivated)(Intel)
Anshul Sharma(Calsoft)
LJ Illuzzi(LFN Staff)
Review Open ARs:
TAC Induction- Open Items
- Issue with Technical Charter - TSC Composition. Update Charter language around TSC voting section 2b. Bob to propose draft language- Bob and Louis did meet agree the necessary small change required. Need to overcome an issue with the source file and then Louis will submit the edited file to LFN Legal for approval
- Address TSC Growth criteria- This is best handled on the Governance Page- Bob made a proposed edit
- Address PTL - Suggested at least Core and UI to start with- This is also best handled on the Governance Page- Placeholder language has been proposed- Need PTL to be determined. For discussion.
New Agenda Items:
- It has been asked to revisit the meeting time as it is very difficult for Asia participants, especially with DST ending soon.
- Also, it has been suggested that we revisit whether this TSC needs to meet every week or can we switch to meeting every two weeks at some point
- Update on Gitlab upgrade? This has been escalated both internally with LF IT and with GitLab
- Any update on CI scan tool integration? We will need those for 21.12 release?
- Logistics of 21.12 release with Aarna Networks.
- Nominations for TSC, TAC Rep and MAC Rep: Requirements for nomination
- High level requirements from Verizon rep (Ravi)
Open Discussion & Next Steps
GitLab "Free" CI blocking developers from normal day-to-day code submission with CI checks. Need upgrade to GitLab Premium/Ultimate.
Future Agenda Items:
Bob will send edits of Technical Charter to Louis to update the language for legal review
Cathy/Sharad to work with designated PTLs for Core and GUI on the Governance page
Carve out a dialog with OPS-5G Project who looks to be a constituent potentially.
Louis say the upgrade on GitLab has come through and Igor will coordinate with Louis to start the process of enabling other users
Louis has been made Owner of the repo now and can start to work on this.
On the CI Scan tool, the original plan remains to use the the Gitlab Scan tool and want to ensure that is available and running for the 21.12 release.
- Louis indicates a scan tool is available once the Ultimate privilege level is enabled for the project
- Louis will work with Sharad and team to explore this and getting it running offline.
Logistics with Aarna Networks in release 21.12- No Aarna rep today and so this can be deferred. Sharad will tee up the conversation with Amar
- Amar will check with his team, but thinks it should be coordinated as close as possible
- Integration, API changes etc need to be considered
Sharad, Ritu to figure out of we still have a particular dedicated pod available
Nominations
Hi Level Requirements from Verizon- 5g Edge Portal- Networking intelligence through APIs
- These APIs are developed based on 5GFF specifications
- Ravi will send a link t these APIs based on these standards
- Would like to see EMCO adopt these standards APIs for MEC deployments
- First round of APIs are related to latency SLAs to find the best MEC placement locations for application
- Get list of MEC locations
- Create Service Profiles- dictate CPU, memory storage and other requirements for each instance
- Track where apps are deployed
- Identify best locations, with fallback if best location not available
- 2nd item under MEC orchestration
- Push v Pull
- eg, AWS Wavelength instances
- Public locations have no inbound interface, outbound only
- Anyone who wants to deploy MEC apps will need a way to reach MEC applications /k8s clusters
- Need public exposure access, with an Operator exposed - to enable to Push Pod specs (Call home)
- Next the ability to be able to "Pull" pod specs
- Srini has noted that there is a 21.12 early engagement with Azure and Google Anthos teams to begin to build some of this support in
- need to ensure that there is a way to Pull the Pod information necessary
- Ravi notes that the multi cluster capability of EMCO is a key role it can play
- Another point- monitoring for observability
- Using Prometheus today but there are limitations when there is no outbound or inbound connection, only outbound, etc.
- Cluster Management
- How to setup the k8s clusters themselves , bring up additional workers capacity nodes
- Srini notes that today we use ClusterAPI as the way to bring up k8s Clusters themselves
- Suspects the requirements being discussed is mainly around bringing up additional worker nodes
- Onboarding MEC Services
Minutes: