...
- Action items
- Edward Ting will share gerrit patch list to the team today
- Darien Hirotsu will coordinate with Jill on the release announcement
- Randy Bias still coordinating with Juniper team about NCDC content, more news by next week
- Several still pending
- 5.1 release
- Still working to get folks together to figure out how to rebuild the code, removing Contrail stuff
- No ETA yet (because there's no scope yet)
- Will we be tagging the release?
- Don't have a TF repo, so could we even tag a Juniper repo?
- SK asked internally for it to be tagged, so it may already be → No, 5.1 is not yet tagged
- vRouter performance
- Document to help newcomers to build vRouter on their own and then pass traffic to it
- Isolated vRouter setup and testing.docx
- Perf analysis of vRouter not yet included in the published doc
- Have enough information to start doing some optimizations, removing the bottlenecks that have been spotted
- Can cover technical details in the TC call
- PS: Is this related to vsperf? (in OPNFV)
- Yes and no: vsperf will give us an apples to apples comparison with other data planes, but we (TF) will need to make optimizations
- Are we interested in getting into vsperf?
- RLB: Interested, but may be a resource issue
- PS: OPNFV can do most of it, if they can get docs on how to do setup (this doc may do the trick), PS offers to help them
- JG: Already spoke with Sridhar Rao from Spirent, tech lead of vsperf. Integration is ongoing.
- SK: How does this perf compare to ovs?
- vsperf will run identical tests on ovs
- Don't seem to have the data yet for comparisons
- Integration w/other LFN projects helps get TF a step closer to becoming an LFN TAC member project
- Folks are +1 on the idea
- Document to help newcomers to build vRouter on their own and then pass traffic to it
- NCDC update
- Tweet announcement
- Have a workshop and a speaking slot
- Just need the content now (gated on Randy)
- networkservicemesh.io collaboration
- It's the new hotness
- Reference arch uses VPP only
- SK recommended TF be a part of their ref arch
- Will's already been doing a lot of work on this
- SK will intro networkservicemesh.io core dev w/Will to sort out initial ideas
- Have the core dev come present for ~15 minutes in an upcoming TSC call
- DH: Why networkservicemesh instead of other tech (istio, linkerd, for instance)?
- Please share this to the mailing list?
- Istio & linkerd are at layers 4+
- nsm are at layers 2 & 3, which is a better fit for us
- Redapt could potentially lend a hand with this
- CNCF is building a common test framework for all their projects (cross-cloud CI)
- nsm is one of those projects, so being a part of that means we'd be a part of that testing
- CLA update
- Pono set up CLA testing repo in gerrit.tungsten.io, so please give it a try
- Also a docs repo ready for merge: https://gerrit.tungsten.io/r/c/docs/+/19
- Still no process for current opencontrail gerrit
- Will Stevens has had a very negative time doing this
- https://jira.tungsten.io/browse/TFB-1394
- ^^ old process, questions about same, some answers
- Any way to get the email sent somewhere after submitted so we have eyes on it?
- casey: private email list
- Pono set up CLA testing repo in gerrit.tungsten.io, so please give it a try
Action items
- Sukhdev Kapurwhat's different with opencontrail and tungsten 5.1 releases
- Sukhdev Kapur email marketing@ about 5.1 cool stuff
- Sukhdev Kapur Get someone w/in Juniper to tag the 5.1 release
- Daniel Pono Takamori Continue the email thread about 5.1 release (incl Edward Ting & Prabhjot Singh Sethi )
- Prabhjot Singh Sethi will get the vsperf ball rolling with the OPNFV team
- Casey Cain Schedule reboot of TC call
- Casey Cain Get vsperf+vrouter onto TC call agenda
- Joseph Gasparakis Get a call with TF folks on vrouter perf (this action item duplicates the one above for Casey to add it to the TC agenda)
- Sukhdev Kapur Start email conversation between himself, Will Stevens, nsm core dev to get that ball rolling
- Sukhdev Kapur Report back next week on nsm conversations
- Casey Cain Create a CLA email address so we can capture CLA submissions