17:01:05 #startmeeting tacker 17:01:06 Meeting started Tue Jan 19 17:01:05 2016 UTC and is due to finish in 60 minutes. The chair is sridhar_ram. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:07 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:10 The meeting name has been set to 'tacker' 17:01:15 #topic Roll Call 17:01:20 who is here for tacker ? 17:01:24 o/ 17:01:31 o/ 17:01:54 o/ 17:01:58 o/ 17:02:08 o/ 17:02:26 hi everyone.. lets get started.. 17:02:34 #topic Agenda 17:02:44 #link https://wiki.openstack.org/wiki/Meetings/Tacker#Meeting_Jan_19.2C_2016 17:03:05 anything else beyond this ? 17:03:50 lets continue 17:03:58 #topic Announcements 17:04:24 Tacker midcycle planned for Jan 28 - 29 in San Jose, CA 17:04:46 etherpad is here #link https://etherpad.openstack.org/p/tacker-mitaka-midcycle 17:05:09 at this time I see very few folks signed up to show up.. virtual or physical 17:05:51 if anyone planning to attend but haven't signed up please do soon... 17:07:01 #topic Midcycle Meetup Agenda 17:07:17 I'd like to firm up the Agenda for the meetup... 17:08:01 One general theme we can go, use Day 1 to discuss the current Mitaka related arch, design & implementation 17:08:15 and use Day 2 for forward looking, evolution of Tacker 17:08:21 what do you folks think ? 17:08:32 sounds good to me 17:08:36 Fine by me 17:08:46 +1 17:09:05 okay... 17:09:09 +1 17:09:37 +1 17:09:41 So for Day 1 .. do we want the cycle through the major blueprints ? any other suggestion ? 17:10:32 Need a discussion on overall design for Tacker 17:11:05 Tacker general architecture as it evolves 17:11:08 brucet: sure, we can take that up on Day 2, how we want to evolve Tacker... 17:11:18 OK 17:11:46 For Day 1, lets start with TOSCA Parser work as it underpins many other activities... 17:12:15 bobh: I'd expect you to drive this..would that be fine ? 17:12:32 sridhar_ram: yes 17:13:24 perhaps we can take up Auto Resource Creation and Efficient VNF Placement 17:13:42 sure 17:14:01 we can cover these topics Thursday AM 17:14:18 and continue with SFC in the afternoon... 17:14:29 s3wong: would be able to drive the SFC portion of the meetup ? 17:14:43 sridhar_ram: sure -- though I may need to be remote 17:15:17 s3wong: okay... 17:15:47 Lets move to Day 2... 17:16:22 perhaps we can keep Day 2 agenda open ended.. 17:16:39 Tacker Architecture evolution... 17:17:33 any other midcycle topics you are interested ? 17:18:05 I've live edited the etherpad at https://etherpad.openstack.org/p/tacker-mitaka-midcycle 17:19:09 Thanks for the Newton correction! 17:19:27 lets move on.. 17:19:39 Oh --- N release has a name already, huh? 17:19:51 just announced today 17:19:51 haven't even been tracking openstack-dev mail... 17:20:05 (and I didn't vote for the names neither...) 17:20:09 s3wong: news to me as well.. nice pick though 17:20:33 sridhar_ram: I can foresee legal issue behind that one, though :-) 17:20:46 i voted for newton 17:20:53 s3wong: I was thinking the same... hope this one sticks 17:21:05 #topic Mitaka Multi-Site 17:21:20 brucet: as promised last week... :) 17:21:31 sripriya: take it away... 17:21:35 OK 17:21:40 sridhar_ram: sure 17:22:07 sridhar_ram: uploaded a new version for spec with some updates, the major thing being the vim_id attribute 17:22:40 sridhar_ram: vim_id selection was moved from VNFD template back to VNF creation operation 17:22:47 since vim 17:22:53 Is the log term plan to use Tricircle? 17:23:17 since vim_id is driven by NFVO to place the VNF and is given as an input dynamically to VNFM 17:23:46 brucet what is Tricircle? 17:24:03 and as the ETSI doc says the vim_id is part of VNFC 17:24:17 which is the actual VNF instance 17:24:17 Openstack project that is essentially doing the same thing as Multi-Site 17:25:00 thanks 17:25:20 brucet: yes, if there is enough interest from the operators / users .. we should consider Tacker -> Tricircle integration 17:25:45 brucet: .. but we need something basic, in-built. 17:26:22 So what about the standalone Heat multi cloud that I proposed before for short term? 17:26:48 If long term is Tricircle, then you want to invest as little as possible for short term. 17:27:09 brucet: No, I don't see long term is Tricircle.. 17:27:27 brucet: that requires a dedicated node running heat service only with specific configuration for stand alone 17:27:43 but tacker can talk to such a heat service too to orchestrate VNFs 17:28:12 You are saying that standalone Heat cannot run on same node as Tacker? 17:28:21 IMO, tacker will support multiple deployment scenarios...the next logical thing to support is multi-site with keystone federation 17:29:10 it can run, but cannot deploy on local infrastructure 17:30:05 brucet: there is an item in Day 2 midcycle, for what I'm putting as "Standalone Tacker"... 17:30:45 brucet: ... it is an evolution of Tacker MultiSite where Tacker service can just run by itself. 17:31:12 yeah 17:31:19 I think more discussion is needed on this topic. 17:31:41 brucet: true, lets use Day 2 of midcycle for some open ended discussion.. 17:31:52 IMHO These issues cannot be resolved via IM. Too complex 17:31:55 OK 17:32:42 Another *major* I would like to reiterate is Tacker is not just a layer on top of Heat.. we tend to use Heat as it cuts our development time... 17:33:46 If something else make spins VMs easier we will cut over to that :) 17:34:04 Agree that there are some features in ETSI NFV that are out of scope of Heat and Tricircle 17:34:17 But there is major overlap. 17:35:31 I'd argue there is a overlap between Nova and libvirt .. but we still see the benefits of an Cloud API in Nova 17:35:45 same argument here... 17:36:07 NFV community need to string a bunch of services to get their NFV orchestration going ... 17:36:12 *need NOT 17:36:42 Tacker is an one stop service to gets things going for NFV 17:37:11 back to the spec, I think we should wrap this one soon.. 17:37:13 Is there a goal to avoid overlap in other projects where it makes sense? 17:38:07 brucet: true, again w.r.t to delivering faster features using overlap projects underneath Tacker... 17:38:39 OK. So, maybe we can make some progress in f2f on this topic. 17:38:55 brucet: sounds good.. 17:39:24 sripriya: back to the spec, having vim_id in the API for now sounds okay to me... 17:39:30 joined late when is f2f? 17:39:54 sridhar_ram: yeah 17:39:56 sripriya: we can always add to the template if there is an aske 17:40:00 *ask 17:40:20 sridhar_ram: ok 17:40:27 LouisF: Jan 28 - 29th, please sign up at the etherpad here.. https://etherpad.openstack.org/p/tacker-mitaka-midcycle 17:40:46 LouisF: https://etherpad.openstack.org/p/tacker-mitaka-midcycle 17:41:06 anything else on Multisite...? 17:41:19 thx 17:41:28 sridhar_ram: EOD, it would still be NFVO resource orchestration module which would guide vim_id decision as it evolves 17:41:59 sridhar_ram: tackerclient and horizon basic UI is done, probably this week to focus on VIM health monitoring 17:42:35 sripriya: cool.. good progress 17:42:41 team, specifically tacker-cores, please review this multi-site https://review.openstack.org/#/c/249085/ ... 17:42:53 sridhar_ram: NFVO alignment is very similar to VNFM with plugin/driver framework implementation 17:43:40 sripriya: cool, we are official foraying into NFVO territory with multi-site and SFC :) 17:44:06 we could discuss in the midycle meetup for further evolving the NFVO 17:44:14 we need get to network services area in the next cycle to get more things going in that area 17:44:21 yes 17:44:23 sounds good.. 17:44:39 lets moves on.. 17:44:57 #topic Service Function Chaining 17:45:12 since s3wong and LouisF are here .. 17:45:37 s3wong: LouisF: can you guys give a quick update on neutron networking-sfc ? 17:45:45 where does the project stand now ? 17:46:08 s3wong: I'm particularly interested in the Tacker --> neutron sfc integration path 17:46:25 * sridhar_ram apologies for putting you both in the spot 17:46:49 we will complete merging all parches this week and release in M2 17:47:04 patches 17:47:22 LouisF: cool, that would be with ovs driver ? 17:47:30 we have a ovs driver and an onos driver is soon to be added 17:48:11 LouisF: any plan to add ODL driver ? 17:48:33 we would welcome help in that area 17:48:55 the onos driver would serve as an example 17:49:28 LouisF: okay... I'll poke few people around for an ODL driver. 17:50:05 btw, my motivation for an ODL driver is to avoid direct odl api call from tacker.. instead go through neutron for all the scenarios 17:50:20 we have done extensive testing in several networking scenarios: multi-node, cross-subnet, scale-out/in... 17:50:22 LouisF: thanks for the update! 17:50:52 we would work with anyone on the odl driver 17:51:50 LouisF: sounds good.. I'm not sure who owns neutron-odl plugin .. they would be the good candidates to take this up. 17:52:02 sridhar_ram: yes 17:52:22 will ask armax and mestery 17:52:46 quick FYI, tim rozet is bit busy with OPNFV release these weeks... he plans to push the patchset for Tacker SFC in couple of weeks. 17:52:49 we are also monitoring closely work on adding nsh support in an official ovs release 17:53:30 an will incorporate nsh into the ovs driver as soon as that is complete 17:53:40 LouisF: true, that is a common denominator which ever way we go... 17:53:51 LouisF: current driver uses mpls ? 17:54:40 sridhar_ram: yes, but that is intended to be a temporary solution until nsh support in ovs is available 17:54:56 LouisF: sounds good... 17:55:19 is allows to test and validate the sfc overlay steering 17:55:54 LouisF: okay 17:56:07 LouisF: appreciate the update! 17:56:11 lets move on.. 17:56:17 #topic Open Discussion 17:56:25 sridhar_ram: you're welcome 17:56:39 a quick note, from the ETSI side.. 17:57:30 last week multiple SDOs met in Denver to evolve a common Information Model (eventually a Data Model) for the whole NFV community 17:58:03 I represented both TOSCA and Tacker .. and it was quite well received 17:58:16 See #link http://www.etsi.org/news-events/news/1053-2016-01-news-etsi-seeks-harmonization-of-nfv-information-modelling-across-standards-bodies-and-open-source 17:58:56 there is lot of respect for Tacker in the NFV community for *realizing* the vision of MANO ! 17:59:43 and a quick note on big tent, I plan to push the governance patchset sometime this week... 17:59:49 the write is here #link https://etherpad.openstack.org/p/tacker-big-tent-writeup 17:59:55 anything else team ? 18:00:07 lets wrap! 18:00:11 thanks everyone 18:00:17 #endmeeting