...
- Scot Steele (AT&T)
- Rick Tennant (AT&T)
- Jim Baker (LF)
- Walter Kozlowski (Telstra)
- Rabi Abdel (Vodafone)
- Beth Cohen(Verizon)
- Nick Chase (Mirantis)
- Phil Robb (Ericsson)
- Ulrich Kleber (Huawei)
- William Diego Maza (Orange)
- Bob Monkman (Intel)
- Qiao Fu (China Mobile)
- Zhiqiang Yu (China Mobile)
- Jonne Soininen (Nokia)
- Jonathan Beltran (AT&T)
- Toshiyasu Wakayama (KDDI)
Topics:
- Anti-Trust Policies:
- Walk On Items
- F2F LA Status
- Room Arrangements (Jim Baker LFN Rick Tennant, Scot Steele AT&T)
- Overall Status
- Trials Documentation - Review: Chapter 9 - Trials Additions Scot Steele (AT&T)
- Next Steps (Scot Steele Jonathan Beltran AT&T)
- Test Case Identification & Pass Criteria
- Complete Artifacts (Questionnaires, results report template, End of trail report template)
- Secure participants
- Suppliers: Red Hat, Nokia, Ericsson, Canonical, VMware
- Operators: Orange, Verizon, AT&T
- Next Steps (Scot Steele Jonathan Beltran AT&T)
- Baldy Priority Items - Unassigned
- Project Scope of what is needed in initial Field Trials from RC1/OVP #1100 (Scot & Jonathan and Close)
- Secure lab participation/capacity/diversity for initial Field Trials #1101 ( RC-1 AT&T has a lab, Intel provides community lab, looking to others to provide)
- Define MVP support model / structure for trials #1102 (in our doc)
- Confirm trial participants and SPOCs #1103 (In the adoption work stream)
- Trial Kickoff Meetings #1105 (Step in plan)
...
CNTT LA F2F needs to happen even if virtual.
Baldy Action Items
1101, 1103 – not progressing as hoped, interdependent
We have Intel already for RC1.
RedHat is reviewing
Need to identify company to do trials first
If we sign up RedHat, then do they want to do it in their lab, if not then who can we get?
Review of Lab Types
Intel (Community)
Self-Lab (Run RC against self-hosted lab)
LaaS (UNH) “paid” service
Priority is to find companies to do a trial.
Rabi Abdel trying to get VMWare to participate. Awaiting feedback.
Action Item: Scot Steele to reassign all items appropriately
Action Item - Field Trial Approach Doc Edits - Scot Steele
9.5.3.1 #2 Add “Not to badge NFVI vendor Implementations”
Use term ‘VI’ instead of NFVI
“When something fails a test, what is the arbitration process for system under test being wrong, or test being wrong?” – user-20f1f
Need to clearer about the process here.
Talk more about how trial participant would interact with the community.
Each trial participant gets a SPOC.
SPOC will work with participant to identify what is wrong with the spec. Then assign it to appropriate work stream.
E.g. Was the test suite actually runnable?
Focus on validating the test suite. (Not pass/fail of TCs).
Run on a variety of NFVIs.
If things don’t pass, is there a simple agreement because the NFVI was deficient in some way (local analysis). Not trying to address the veracity of the system under test?
Anonymize (i.e. ''reverse' blind testing/assessment) the results:
Here’s situation.
Here’s the result.
“Keep a level of anonymity.”
Repositories: CNTT SPOC needs to be identified to take the feedback and where it should land: CNTT repository or OPNFV bug (then OPNFV)
Make sure language throughout says: NFVI participant will work with the SPOC/ Community member, and vendor - not entire community.
E.g. 9.5.6.1
“Community shall review” becomes “CNTT SPOC shall review”
“Overall CNTT Decision making leads” need revise that issues will be take issues to TSC/GSC.
Overall Trial Project Manager / Lead responsible for Field Trials – is who ?? – looking over entire trials process (cat herder)
Add: one person assigned to vendor as SPOC “CNTT SPOC
Developing Test Criteria: RC-1 has all this specified. RI/RC work stream is weekly at seven central. Existing effort to review results. Remove sections detailing test criteria from Doc. Point to RC doc—get Michael Fix (AT&T) to help here.
Action Item: Scot Steele to schedule sub-group calls to develop approach.
...