#opendaylight-group-policy: gbp_status_arch
Meeting started by tbachman at 15:03:36 UTC
(full logs).
Meeting summary
- Agenda bashing/roll call (tbachman, 15:03:57)
- committers please #info in (tbachman,
15:04:02)
- tbachman (tbachman,
15:04:05)
- alagalah (alagalah,
15:04:16)
- https://lists.opendaylight.org/pipermail/groupbasedpolicy-dev/2015-August/001303.html
email from alagalah with today’s agenda (tbachman,
15:04:37)
- https://meetings.opendaylight.org/opendaylight-group-policy/2015/gbp_status_arch/opendaylight-group-policy-gbp_status_arch.2015-08-05-15.00.html
minutes from last week’s meeting (tbachman,
15:04:51)
- Stable Lithium Updates (tbachman, 15:06:14)
- alagalah says the focus of the current work is
on stability — CSIT standup for GBP/SFC, clustering support,
etc. (tbachman,
15:07:40)
- alagalah says that a lot of the team has done a
great job of adding unit tests (kinsey, kblagov, tomas_c|2, et. al.)
and CSIT (tomas_c|2, kblagov, et. al.) (tbachman,
15:09:36)
- alagalah says after CSIT, we’re creating
documentation, to help folks understand things like the robot
framework (tbachman,
15:10:27)
- alagalah says we’re looking at adding the GBP
neutron pieces into an OPNFV project (tbachman,
15:10:47)
- alagalah says an endgoal is to be able to test
upstream openstack against the GBP neutron integration (tbachman,
15:11:42)
- alagalah says multinode testing done by
tomas_c|2 will be a big help (tbachman,
15:12:49)
- Stable Lithium updates - Thomas Demo of how cool the demo env is without restarting controller (alagalah, 15:15:10)
- tbachman Is showing the standard 6 node GBPSFC
demo with ODL running on his laptop (not in the demo VMs)
(alagalah,
15:16:10)
- Doing a demo of asymmetric chain (alagalah,
15:17:50)
- showed the traffic working, then runs
./cleandemo.sh which removes the Datastore and Network element
data (alagalah,
15:21:48)
- then runs the standalone GBP demo (alagalah,
15:21:58)
- martin_sunal (martin_sunal,
15:22:33)
- alagalah says that the SFC project has
different scheduling algorithms; GBP isn’t leveraging this at the
moment, but could with some small modifications (tbachman,
15:30:37)
- alagalah says the RSP is referenced by name, so
we could come up with a naming convention to use to enable
this (tbachman,
15:31:10)
- alagalah is trying to modify the GBPSFC demo so
that SFs can be put in docker containers (tbachman,
15:32:22)
- Beryllium release planning (tbachman, 15:35:03)
- https://docs.google.com/document/d/1NZPhsEccY5lBOXI0v1uygE0Y_iIjCjRNcs27H4e4CMw/edit?usp=sharing
Brainstorming document for Beryllium release planning (tbachman,
15:35:22)
- https://wiki.opendaylight.org/view/Group_Based_Policy_(GBP)/Releases/Beryllium
GBP Beryllium release plan (tbachman,
15:35:38)
- https://trello.com/b/fDQss8ez/odlgbp-beryllium
Trell board for GBP Beryllium release (tbachman,
15:36:11)
- alagalah says we have until September 3rd to
get the final release plan completed (tbachman,
15:36:50)
- alagalah says code freeze is in December
(tbachman,
15:38:44)
- alagalah says that anything modeled in YANG can
be exposed to the outside world via REST (tbachman,
15:40:20)
- alagalah says that navigating the full model
can be daunting — led to discussion of how to map DSL from
northbound into the model (tbachman,
15:41:16)
- alagalah says this allows us to keep domain
specifics in a dictionary — allowing for DSL mappings (tbachman,
15:44:29)
- alagalah says this has implications into the UX
— how do we expose this to the user (tbachman,
15:44:51)
- alagalah says we could potentially use yang-ui
as a means of enabling this type of “expert-mode” (tbachman,
15:45:53)
- yapeng asks if the neutron northbound would
become a consumer of this expert-mode (tbachman,
15:48:07)
- alagalah says it’s actually the opposite — our
contract with NN is to map the interfaces it exposes (tbachman,
15:48:38)
- yapeng asks who would use the new API
(tbachman,
15:49:38)
- alagalah says anyone using ODL GBP for
Neutron (tbachman,
15:49:48)
- alagalah says the core engine should have no
domain specifics in it (tbachman,
15:52:04)
- alagalah says we need to separate the
forwarding and access models into different files (tbachman,
15:52:58)
- alagalah says the forwarding model is in the
main GBP project, and there is no alternate forwarding model; this
precludes others from easily adding new forwarding models
(tbachman,
15:53:30)
- alagalah says anyone wishing to contribute
should start with our wiki page (also see Trello board) (tbachman,
16:01:04)
- alagalah will be at LinuxCon next week - will
try to kick off discussion on the ML for the Beryllium
release (tbachman,
16:03:41)
Meeting ended at 16:03:43 UTC
(full logs).
Action items
- (none)
People present (lines said)
- tbachman (48)
- alagalah (6)
- odl_meetbot (5)
- yapeng (4)
- martin_sunal (2)
Generated by MeetBot 0.1.4.