Versions Compared

Old Version 5

changes.mady.by.user Karine Sevilla

Saved on

compared with

New Version Current

changes.mady.by.user kumar.suman

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Agenda and Minutes:

  • Antitrust notices
  • Walk-in items
    • RM 5.0 alignment - delta here
      • alignment with 3.8 - can be broken down into labeling (Riccardo Gasparetto to create issue and track) and controller/APIs/interfaces (likely a number of gaps - another issue)
        • Not all controllers for accelerators may be in scope - do we need to specify the interface if the accelerator is not part of the infrastructure?
        • Likely overlap with RA1 - problem space is overlapping 
      • add issue to align with the whole RM security chapter
      • align with chapter 8 - track with issue - may need to change or relax (or add exceptions) to ch2 requirements - also need to track as flavours with labels
      • chapter 9 - discuss whether Cluster LCM should be part of the scope, and if so, how (eg should we constrain a solution or not, to what level should we specify its properties...)
        • PROS: useful for operators, provides guidance to reference implementation, may help direct vendors to existing standards
        • CONS: no immediate impact on workloads, may constrain 
    • https://github.com/cntt-n/CNTT/issues/2068 - do we need more profiles? RAN CaaS?
      • need to differentiate between Profiles and Flavours,
      • need to differentiate between workload profiles vs infrastructure/node profiles
      • Pankaj.Goyal : Operators are generally against proliferation of profiles
      • Profiles segment/partition the infrastructure: nodes belonging to different profiles are managed separately
      • Gergely Csatari: too much optionality in profile specs, gives too little assurance to users
      •  Riccardo Gasparetto Hierarchical profiles: flavours as sub-profiles? 
      • Pankaj.Goyal: need to maximise capacity - segmenting infrastructure with mono-dimensional profiles leads to waste of capacity when workloads are forced to choose a set partition to be allocated onto at runtime
      • need to rename the network-intensive profile to something else as it's misleading - eg basic vs advanced?
  • AOB & Project review
  • Permanent FYI
  • Actions/Next steps

...