TSC 2018-01-25

Date

25 Jan 2018

Recording

IRC minutes

Full IRC Log  

Zoom Chat Log

Full IRC Log 

13:53:55 <kennypaul> #startmeeting tsc 2018-01-25
13:53:55 <collabot> Meeting started Thu Jan 25 13:53:55 2018 UTC.  The chair is kennypaul. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:53:55 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:53:55 <collabot> The meeting name has been set to 'tsc_2018_01_25'
13:54:13 <kennypaul> #chair phrobb
13:54:13 <collabot> Current chairs: kennypaul phrobb
13:54:31 <kennypaul> #topic rollcall
13:59:25 <kennypaul> #info Alla Goldner, AMDOCS
13:59:30 <JasonHunt> #info Jason Hunt, IBM
14:01:18 <gilbert> #info mazin gilbert
14:01:19 <alexvul> #info proxy for Rajesh Gadiyar (Intel)
14:01:26 <amir_levy> #info Amir Levy, Cloudify
14:01:47 <cdonley> #info Chris Donley Huawei
14:02:18 <xinhuili> #info Xinhui Li, VMware
14:02:21 <RannyHaiby> #info Ranny Haiby, Nokia
14:03:39 <DhananjayPavgi> #Info Dhananjay Pavgi, Tech Mahindra
14:03:46 <gilesheron> #info proxy Giles Heron, Cisco
14:04:40 <davidsauvageau> #info David Sauvageau, Bell Canada
14:05:15 <kennypaul> #info Chengli proxy for Lingli, CMCC
14:05:55 <SteveT> #info Stephen Terrill, Ericsson
14:07:03 <jamil> #infor jamil for Orange
14:07:05 <susana> #info Susana Sabater Vodafone
14:07:15 <kennypaul> #topic infrastructure coordinator proposal
14:07:20 <Zhaoxing> #info Zhaoxing Meng, ZTE
14:24:47 <kennypaul> #startvote does the TSC approve the Infrastructure Coordinator proposal as presented on slide 6 of the TSC meeting slide deck? +1, 0, -1
14:24:47 <collabot> Begin voting on: does the TSC approve the Infrastructure Coordinator proposal as presented on slide 6 of the TSC meeting slide deck? Valid vote options are +1, 0, -1.
14:24:47 <collabot> Vote using '#vote OPTION'. Only your last vote counts.
14:24:53 <gilbert> #vote +1
14:24:54 <cdonley> #vote +1
14:24:54 <amir_levy> #vote +1
14:24:55 <JasonHunt> #vote +1
14:24:56 <SteveT> +1
14:24:58 <RannyHaiby> #vote +1
14:24:59 <alexvul> #vote +1
14:25:01 <gilesheron> #vote +1
14:25:02 <SteveT> #vote +1
14:25:04 <Xiaojun> #vote +1
14:25:06 <davidsauvageau> #vote +1
14:25:15 <jamil> #vote +1
14:25:21 <ningso> #vote +1
14:25:22 <xinhuili> #vote +1
14:25:34 <Zhaoxing> #vote +1
14:25:46 <kennypaul> #endvote
14:25:46 <collabot> Voted on "does the TSC approve the Infrastructure Coordinator proposal as presented on slide 6 of the TSC meeting slide deck?" Results are
14:25:46 <collabot> +1 (14): davidsauvageau, jamil, Xiaojun, ningso, gilesheron, amir_levy, JasonHunt, xinhuili, SteveT, alexvul, cdonley, gilbert, RannyHaiby, Zhaoxing
14:28:55 <kennypaul> #agreed the TSC approves the Infrastructure Coordinator proposal as presented on slide 6 of the TSC meeting slide. deck
14:29:27 <kennypaul> #topic Project Definition Proposal
14:30:33 <kennypaul> #info @SteveT reviewed the slides
14:33:17 <kennypaul> #agreed that project proposal pages should be locked at the version approved
14:34:15 <kennypaul> #info project proposals that have been changed since approved need to be reviewed.
14:35:45 <kennypaul> #agreed the Project Definition Proposal as presented is approved
14:38:58 <kennypaul> #topic VF2F
14:40:08 <kennypaul> #info concerns regarding project meetings not being able occur
14:51:45 <kennypaul> #topic ONS
14:52:44 <kennypaul> #info concerns expressed over cost for attendance and impact to participation
14:54:13 <kennypaul> #info mazin is having conversations with the LFN Board regarding the impact
14:57:03 <kennypaul> #info will not be able to impact March event. Will focus on ONS Europe.
15:01:13 <kennypaul> #topic Runtime Catalog Proposal
15:01:56 <kennypaul> #info Maopeng reviewed the proposal deck.
15:14:32 <kennypaul> #info review by the Arch. Subcommittee
15:15:25 <kennypaul> #info concerns expressed about syncing 2 catalogs
15:19:09 <kennypaul> #info discussion around implementation of functionality SDC
15:20:50 <kennypaul> #info SDC PTL (not on call) is said to have agreed this should be a separate project
15:23:50 <kennypaul> #info need is recognized - concerns are on software side
15:27:27 <kennypaul> #info SO and VID are looking at this for Beijing - depends on the usecase
15:30:06 <kennypaul> #info software review at Arch meeting (Tuesday) , project approval vote next week's TSC meeting.
15:30:32 <kennypaul> #action kennypaul send Arch Sub invite to TSC list
15:33:30 <kennypaul> #endmeeting

Zoom Chat Log 

06:11:30 From Catherine Lefevre (AT&T) : +1 on Infrastructure Coordinator Proposal
06:15:54 From Catherine Lefevre (AT&T) : +1 on committer -
06:22:14 From stephen terrill : This is fine. Do we capture this as a TSC policy or someething like that?
06:22:52 From Catherine Lefevre (AT&T) : As a first step, we can reduce the flow from PTL-> Gildas -> TSC to PTL -> Gildas
06:23:05 From Chris Donley : We certainly could.
06:23:29 From Catherine Lefevre (AT&T) : and if everything is going well with Beijing then we can explore Jamil's proposal
06:23:42 From Tony Hansen : two tasks, can both be assigned to one person
06:26:07 From Michael O'Brien (Logging) : Excellent - Gildas!
06:26:20 From Randa Maher (AT&T) : thanks, this is a great move! Thanks Gildas for taking it on.
06:26:46 From Catherine Lefevre (AT&T) : @Stephen, maybe we should enhance the TSC chart accordingly
06:27:14 From Gildas Lanilis Huawei : Thanks all for your confidence and support on this. It will help the community.
06:29:07 From Chris Donley : FYI, here was the committer promotion process from OPEN-O: https://wiki.open-o.org/display/TSC/Committer+Promotion+Process. We could write up something similar for ONAP, given today's discussion.
06:29:53 From Chris Donley : The repo creation process could be updated on the Resources and Repositories page.
06:31:18 From ONAP Meeting 1 - Alla : Can we cover virtual and f2f meetings now?
06:31:34 From ONAP Meeting 1 - Alla : I can stay only next 30 min, have conflicting meeting after
06:31:50 From Pamela Dragosh : +1 locking
06:33:01 From Catherine Lefevre (AT&T) : @Chris - +1 for repo on the resources qnd repositories - i think teams are doing this. I will check your link lqter
06:35:15 From Catherine Lefevre (AT&T) : Open for any process simplification - project definition might evolve based on architecture guidelines and use cases so we need to consider this as well
06:42:48 From Catherine Lefevre (AT&T) : can we also have a quick feedback about tools for Python and also additional test framework that will be considered for Beijing?
06:43:35 From Catherine Lefevre (AT&T) : Labs availability for developers to prepare their pair-wise testing including fail-over
06:43:49 From Michael O'Brien (Logging) : ONS talk approvals come in on the 6th Feb right?
06:44:06 From Michael O'Brien (Logging) : CNCF, OPNFV yes
06:45:28 From Michael O'Brien (Logging) : OPNFV auto project pulls HEAT and OOM deployments of OPNAP
06:45:32 From Michael O'Brien (Logging) : ONAP
06:45:52 From Michael O'Brien (Logging) : python - yes - for Multicloud
06:46:23 From shankar : The Python session would be useful for OOF as well. OOF is a pure python project.
06:48:52 From shankar : Would there be sessions in Virtual F2F happening in parallel ?
06:59:09 From Seshu : :)
07:05:24 From Seshu : Intel onap team would be much better
07:06:28 From Brian : what is the impact on the projects that get models from SDC distribution client today ?
07:07:35 From Alex Vul (Intel) : Why wouldn't we start with SDC catalog implemetnation instead of re-inventing the wheel?
07:07:52 From Pamela Dragosh : Projects would have to implement 2 api’s for the 2 catalogs. Big impact.
07:08:22 From Alex Vul (Intel) : projects would only work against RT catalog...
07:08:42 From Alex Vul (Intel) : i.e. projects would only deal with NS and VNF designs promoted to production...
07:08:51 From Pamela Dragosh : Not necessarily, some projects are BOTH design time and run time. So they would need both.
07:08:52 From Alex Vul (Intel) : Jason - yes...
07:08:59 From Ting Lu : Look at the initial step implenetation here.
07:09:25 From Ting Lu : There are not impact to the RT compoennts.
07:09:48 From Brian : how does ops role ops001 know that th emodels were succesffully distributed ?
07:10:08 From Jason Hunt : I’m worried about sync of two catalogs. Could these runtime functions be implemented into the SDC catalog (along line of Alex’s question)?
07:10:34 From Mazin : Was this verified by the architecture team.
07:10:59 From Alex Vul (Intel) : we will need multiple copies becasuse of the deployment scenarios, but these should be copies of the SAME implementation...
07:11:28 From Alex Vul (Intel) : much like we talk about "service" and "resource" orchestration being implemented using the same orchestration funciton....
07:11:43 From nagu : does it mean RT catalog must be always available for other ONAP components e.g., SO to function correctly
07:12:03 From Alex Vul (Intel) : Mazin - i brought up ,many of these concerrns before
07:12:13 From Chris Donley : @Mazin - yes, RT Catalog got the green light from ARC.
07:12:13 From Alex Vul (Intel) : but apparently we can just forge ahead
07:12:36 From Chris Donley : they met with us a few times and incorporated the feedback.
07:13:53 From Dhananjay Pavgi : It defeats the Architecture philosophy of separating Design and Run time envrionment. If we have different catalogs in Design and Run time.
07:14:58 From Dhananjay Pavgi : Whole rational is to have model driven models driving run time envrionment. Catalog once driven at Design Time should be consumed for "service specifics" or instances
07:15:54 From Michael O'Brien (Logging) : Ting Lu - thanks for the design/runtime clarification - was going to ask which is the source of truth - as in keeping them in sync - I need to understand if SDC is also the RT source in that case and not just distribution
07:18:16 From Michael O'Brien (Logging) : Ideally I would like to see a model driven "expected state" - where the system maintains this - like in Kubernetes where it is both distribution and runtime - will educate myself more - good discussion
07:18:24 From Dhananjay Pavgi : Isn't it TOSCA model that gets distributed as a metadata model?
07:18:37 From Michael O'Brien (Logging) : Yes good TOSCA point
07:19:28 From Michael O'Brien (Logging) : Yes need MIchael Definitely
07:20:05 From Jason Hunt : If we think of models as source code, would we have design-time vs run-time source code repositories?
07:21:01 From Huabing Zhao : @Michale Runtime Catalog will not change the TOSCA model generated by SDC
07:21:27 From Brian : many of these components use the SDC client to interact with SDC and the DMaaP topics - will it be the same SDC client from SDC with th eRT catalog ?
07:21:30 From Helen Chen : I am second that the catalog should have the same implementation code base. But it could be deployment in two places.
07:22:00 From Pamela Dragosh : Good question Brian, that is not clear at all.
07:22:09 From Helen Chen : Or we refectory the catalog out into a separate project?
07:23:35 From Michael O'Brien (Logging) : +1 for Jason meet on software impact
07:23:44 From Brian : +1
07:23:45 From nagu : +1
07:23:47 From Pamela Dragosh : +1 also for Jason and software architects
07:24:04 From Alex Vul (Intel) : +1
07:25:32 From Michael O'Brien (Logging) : (if arch team approves) - I am +1 creation (defer to arch team) just need to understand API impact
07:25:59 From Dhananjay Pavgi : We could certainly.
07:26:16 From Dhananjay Pavgi : Can't think about a use case that would break for the lack of this project for Beijing, Mazin
07:26:36 From Helen Chen : Prefer to do it right instead of rush.
07:27:13 From Dhananjay Pavgi : Can we discuss and debate use case that breaks existing architecture and warrants this proposed project to be quintessential?
07:27:19 From Helen Chen : duplicated code and inconsistency is scaring.
07:28:53 From Dhananjay Pavgi : Spot on Helen.
07:29:01 From Alex Vul (Intel) : +1
07:30:04 From Alex Vul (Intel) : I've built distributed catalogs before, and this is typically a function of a single catalog implementation...
07:30:14 From Michael O'Brien (Logging) : will do - Tue 10EST (GMT-5) - see you there Chris
07:30:41 From Catherine Lefevre (AT&T) : it will be important to understand the prioirty of RT Catalog vs the other use cases and SP3.
07:31:14 From Catherine Lefevre (AT&T) : not sure that all PTLs will be able to commit for additional work beyond what they were committing at M1 milestone
07:32:06 From Catherine Lefevre (AT&T) : for Beijing - we should at least have an agreement on the project + low level design details so the development team will have all what they need to move to the implementation as early as they have the bandwidth