#opnfv-meeting: Weekly Technical Discussion

Meeting started by bh526r at 13:00:07 UTC (full logs).

Meeting summary

  1. Roll Call (bh526r, 13:00:25)
    1. Bin Hu (bh526r, 13:00:36)
    2. Manuel Buil (mbuil, 13:00:47)
    3. Mark Shostak (bh526r, 13:03:40)
    4. Mark Beierl (bh526r, 13:04:56)
    5. Al Morton (bh526r, 13:05:01)
    6. Sunku Ranganath (bh526r, 13:05:28)
    7. Mark Beierl (mbeierl, 13:07:57)

  2. Closed Loop Platform Automation & OPNFV by Sunku Ranganath (bh526r, 13:10:49)
    1. https://wiki.opnfv.org/download/attachments/2925877/ClosedLoopPlatformAutomationWiOPNFV.PDF?version=1&modificationDate=1555637511000&api=v2 (bh526r, 13:11:12)
    2. Sunko described the background, and the initial discussion in OPNFV Develoepr Gathering on April 2 (bh526r, 13:12:45)
    3. Sunko introduced use case and how current OPNFV projects are related (bh526r, 13:13:17)
    4. e.g. Doctor, NFV Bench, Barometer etc. (bh526r, 13:14:12)
    5. The question was whether or not we need a new project (bh526r, 13:14:28)
    6. The outcome from discussion in meetup was that we may need a new WG (bh526r, 13:15:19)
    7. Sunku is asking for the feedback from community (bh526r, 13:16:16)
    8. Mark B posted comments in chat "Shameless plug for StorPerf here: OPNFV also has that project for doing storage capacity and optimization. Can do things like measure the impact of Barbican for software encryption, etc." (bh526r, 13:17:30)
    9. Al M echoed "Yes, exactly - Storperf should be included here" (bh526r, 13:18:05)
    10. Mark B further indicated "We also should note performance tools like VSPERF and StorPerf should not be used when live traffic is also running as these tools can flood the system and cause degradation of live traffic. Or, used with caution or only for path verification" (bh526r, 13:18:53)
    11. Al M indicated "I see some synergy with O-RAN. They have the Near-RT and non-RT loops Identified in their API framework. Establishing the Control Loop is possible, but designing and applying a specific policy in the loop that is not just a simplistic example is HARD, IMO. " (bh526r, 13:19:52)
    12. Trevor Cooper (bh526r, 13:26:24)
    13. Al Morton indicated that the challenge is when policy is beyond the simplistic examples. (bh526r, 13:28:02)
    14. Al asked if there is reason why OpenStack is pushing back (bh526r, 13:34:42)
    15. Sunku explained that Nova assigns resources when it starts a VM. They think Nova should control resource assignment, and thus the local agent seems extension of Nova. (bh526r, 13:36:26)
    16. Community doesn't understand the need of this closed-loop use case (bh526r, 13:39:56)
    17. Bin asked what is the execution plan, i.e. implement in own repo or coordinate with testing projects and implement it in their repos (bh526r, 13:55:56)
    18. Sunku said that eventually it needs own repo for analytics, but start with working with Doctor, Barometer, Bottleneck etc. projects (bh526r, 13:56:50)
    19. Sunku will draft a formal proposal in wiki for this WG by working with Doctor, Bottleneck, Barometer etc. projects. (bh526r, 13:58:17)
    20. Once the WG proposal is ready, Sunku will bring it to weekly technical discussion again. (bh526r, 13:58:42)
    21. Then it will be brought to TSC for approval. (bh526r, 13:59:06)

  3. AOB (bh526r, 13:59:13)
    1. Al asked what is the plan for PlugFest in June (bh526r, 13:59:28)
    2. Bin shared that Intel has expressed their support of hosting a low-cost OPNFV PlugFest+ONAP DDF in the week of June 10. (bh526r, 14:01:00)
    3. Bin has asked LFN (Phil Robb), but never got response (bh526r, 14:01:19)
    4. Intel followed up with Phil Robb in ONS, and the response was that LFN is now trying to find a location which is non US or China because of geo-political conditions, hence the Intel SC location might not be a good fit for the ONAP DDF + OPNFV PlugFest. (bh526r, 14:02:30)
    5. There is no more information from LFN regarding OPNFV PlugFest+ONAP DDF, while it is only 7 weeks away from the planned date (the week of June 10) (bh526r, 14:03:43)
    6. No other questions, meeting adjourned (bh526r, 14:04:12)


Meeting ended at 14:04:18 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. bh526r (40)
  2. mbuil (3)
  3. collabot (3)
  4. mbeierl (1)


Generated by MeetBot 0.1.4.