#opnfv-meeting: Weekly Technical Discussion
Meeting started by bh526r at 13:02:40 UTC
(full logs).
Meeting summary
- Roll Call (bh526r, 13:02:48)
- Bin Hu (bh526r,
13:02:54)
- Gary Wang (bh526r,
13:03:01)
- Brian Skerry (bjskerry,
13:03:23)
- Trevor Cooper (bh526r,
13:05:34)
- KVM4NFV Requirement Gathering (bh526r, 13:07:30)
- https://wiki.opnfv.org/display/kvm
(bh526r,
13:07:46)
- Gary mentioned that this KVM project has been
in maintenance mode for 2 years (bh526r,
13:08:22)
- Gary and his colleagues, such as Isaku,
discussed to make the project more active (bh526r,
13:08:49)
- and convince his management to invest more
resources in this project and in OPNFV (bh526r,
13:09:17)
- Gary and Isaku have thought about some of the
areas to contribute to the project such as performance
improvement (bh526r,
13:09:52)
- and some optimization (bh526r,
13:10:02)
- and consider to contribute patches to this
project (bh526r,
13:10:20)
- Thus Gary wants to discuss some requirement for
this project (bh526r,
13:11:44)
- Brian indicated that one of the issue is the
adoption rather than spinning a fork (bh526r,
13:12:13)
- Gary indicated that some patches may be
contributed to OPNFV, and some may be submitted to upstream KVM
project (bh526r,
13:14:07)
- Trevor asked if those are existing patches or
new patches (bh526r,
13:14:55)
- In another word, if old patches are not
accepted by upstream Linux kernel, why new patches? (bh526r,
13:16:40)
- Gary indicated that we are looking for ways for
kernel to accept our patches (bh526r,
13:17:07)
- Brian asked if the use case is the real time
performance (bh526r,
13:17:44)
- Gary said yes, all existing patches and new
patches will be for real time feature (bh526r,
13:18:06)
- Trevor said that maybe some specifics are
needed, such as test framework, to show the benefits (bh526r,
13:20:00)
- Gary said that we were focusing on real time,
and we want to hear from community if there are more
requirement (bh526r,
13:23:57)
- In current project, we were focusing on
analyzing the bottleneck and improving the performance. We are
looking for some new requirement at overall NFV level (bh526r,
13:26:14)
- i.e. suggestions from industry and
ecosystem (bh526r,
13:26:57)
- Brian suggested the direction of
container (bh526r,
13:28:59)
- Gary said that the project focused on kernel
space, and if there are container requirement, they would be happy
to address it and support it (bh526r,
13:29:50)
- Brian also indicated that nevertheless, there
is the issue of being accepted by upstream kernel community
(bh526r,
13:30:20)
- Trevor suggested to write a proposal as if a
new project, and give people the insight of the real scenario that
shows the benefit (bh526r,
13:35:49)
- Bin suggested that maybe another way is to
analyze how KVM technology can support container (bh526r,
13:44:36)
- Gary said that because of resource constraint,
it may be challenging to do such analysis (bh526r,
13:45:10)
- Gary/team will communicate with existing
stakeholders/committers of project to get more information
(bh526r,
13:52:40)
- AOB (bh526r, 13:59:34)
- None (bh526r,
13:59:38)
Meeting ended at 13:59:41 UTC
(full logs).
Action items
- (none)
People present (lines said)
- bh526r (36)
- collabot (3)
- bjskerry (2)
Generated by MeetBot 0.1.4.