18:03:11 <colindixon> #startmeeting tws 18:03:11 <odl_meetbot> Meeting started Mon Feb 23 18:03:11 2015 UTC. The chair is colindixon. Information about MeetBot at http://ci.openstack.org/meetbot.html. 18:03:11 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:03:11 <odl_meetbot> The meeting name has been set to 'tws' 18:03:28 <colindixon> #topic slides, agenda, waiting, etc. 18:03:35 <colindixon> #chair tbachman alagalah regXboi dfarrell07 18:03:35 <odl_meetbot> Current chairs: alagalah colindixon dfarrell07 regXboi tbachman 18:03:39 <colindixon> #link https://wiki.opendaylight.org/view/File:OVSDB_Neutron_Support_in_Li_and_Beyond.pdf the slides 18:03:39 <tbachman> :) 18:03:52 * regXboi mutters "no good deed goes unpunished" 18:04:17 * tbachman gives vapid stare 18:04:28 <colindixon> #info today, we’ll cover the Lithium plans for Neutron support in the OVSDB project 18:04:33 <alagalah> #topic OVSDB/Neutron support in Li - Colin Dixon, Kyle Mestery, Chris Wright 18:06:23 <tbachman> #info OpenSTack Neutron has been a target use case for ODL since the beginning (OpenDOVE, OVSDB, and VTN all provided Neutron interfaces in Hydrogen) 18:07:00 <tbachman> #info OpenStack Neutron continues to be important (raised at the board and TSC levels; considerable developer/community) 18:07:44 <tbachman> #info shague says a lot has changed with OpenStack neutron. Since the initial integration with Hydrogen, neutrons progressed a lot 18:08:21 <tbachman> #info One fo the questions is how can we foster the relationship betweeen ODL and OpenStack 18:09:24 <tbachman> #info colindixon says even though ODL community recognizes the importance of the use case with Neutron, the ODL community hasn’t engaged enough with OpenStack 18:11:37 <tbachman> #info alagalah asks what the groups thoughts are on the newly created neutron project — will that project act as a liason to the OpenStack project 18:11:54 <tbachman> #info shague says that will be discussed later on in the presentation 18:12:12 <tbachman> #info shague says we’ve got some great first steps, but we need to go beyond what we’ve done 18:12:54 <tbachman> #info relevant projects in ODL include OVSDB, OpenDOVE, VTN, Plugin to OpenContrail, LISP, VPN Services, and Group Based Policy 18:13:14 <tbachman> #info shague points out that this list will probably only grow in subsequent releases 18:13:23 <tbachman> regXboi sets the rules 18:14:05 <tbachman> #info regXboi says projects shouldn’t go to the neutron community with an ask for an interface feature in a blueprint that hasn’t been approved 18:15:39 <tbachman> #info Prem asks for the best way to engage with the OpenStack community to do this 18:16:09 <tbachman> #info colindixon asks if the current neutron interface covers everything in neutron 18:16:19 <tbachman> #info regXboi says there are things that need to be implemented 18:16:24 <tbachman> #undo 18:16:24 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x1a8be10> 18:16:35 <tbachman> #info regXboi says there are things that need to be approved and implemented 18:16:46 <tbachman> #info edwarnicke says there are calls in tempest tests that aren’t documented 18:17:08 <tbachman> #info edwarnicke says he doesn’t know how you can tell what’s been approved and implemented, b/c it’s not covered in the documentation 18:17:26 <tbachman> #info edwarnicke asks for guidance on how to determine what’s been approved in Neutron 18:18:02 <tbachman> #info regXboi says the two best answers are: 1) he’ll ask for an approved BP in OpenStack; the other is some acronym which I didn’t get 18:18:30 <tbachman> #info alagalah says these are good questions, but lets reserve those until the end of the presentation 18:18:40 <regXboi> #undo 18:18:40 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x1b71fd0> 18:18:44 <regXboi> #info 2) is RTC 18:18:46 <regXboi> #undo 18:18:46 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x1b71990> 18:18:46 <edwarnicke> alagalah: Thank you for pulling us back on track :) 18:18:54 * edwarnicke sometimes gets rat holed :( 18:18:54 <regXboi> #info 2) is RTSC 18:18:59 <regXboi> #info alagalah says these are good questions, but lets reserve those until the end of the presentation 18:19:09 <regXboi> tbachman: wanted to get things back in order 18:19:12 <tbachman> regXboi: I think you removed item #1 :) 18:19:27 <tbachman> ah 18:19:30 <tbachman> no you idn't 18:19:31 <tbachman> sorry 18:19:46 <regXboi> no, I reorder alagalah's point of order 18:20:17 <tbachman> #info The OVSDB project in ODL houses the ODL OpenStack CI for the default ODL Neutron plugin ( 18:20:52 <tbachman> #info ODL developers need to engage the OpenStack community 18:21:07 <tbachman> #info Engagement at the upcoming OpenStack Summit is important 18:21:35 <tbachman> #info shague says the overall goal is can we make neutron better, networking better 18:22:06 <alagalah> shague: colindixon flaviof I have a question on the previous slide. 18:22:54 <tbachman> #info OVSDB tasks for Lithim: abstract net virt from OVSDB library and plugin; remove AD-SAL dependencies; Support existing OVSDB plugin interface for VTN; add MD-SAL southbound plugin; migrate net virt to MD-SAL southbound plugin; test and importve scale, stability and performance; close feature gap with Neutron and OpenSTack OVS plugin (e.g. L3, LBaaS, FWaaS, VPNaaS) 18:23:17 <flaviof> alagalah: np. want to as it on irc or webex? 18:23:47 <alagalah> flaviof: That's ok, lets keep trucking.... 18:23:57 <flaviof> alagalah: k 18:24:23 <tbachman> #info shague says a lot of code came in during Helium; project is trying to focus on the scale, stability and perfomance and closing the feature gaps is high priority 18:24:56 <colindixon> alagalah: what’s the questions? 18:24:58 <tbachman> #info shague says flaviof has done a lot of work on L3; srini084 has done a lot of work on LBaaS; FWaaS and VPNaaS still need more work 18:25:15 <tbachman> #info Uri asks if L3 means DVR 18:25:38 <tbachman> #info flaviof says it’s not DVR — it’s OVS-based, using OpenFlow rules to do L3 forwarding (TTL decrement, etc.) 18:25:51 <tbachman> #info flaviof says they’re still looking into how scalable that solution is 18:26:15 <tbachman> #info Uri asks if DVR can be added to the list 18:26:48 <tbachman> #info Uri asks for the projects thoughts on the “xaaS” features, and how they relate to SFC 18:27:02 <tbachman> #info shague says those are all good questions, but are still open items 18:27:23 <regXboi> alagalah: can I inject a point of order? 18:27:38 <regXboi> that last question needs an answer from neutron project as well 18:27:48 <alagalah> regXboi: Yes.. I am not the thought police dude, just keeping us on agenda 18:28:40 <tbachman> #info shague says that this all presumes that Neutron passes this information to ODL 18:29:11 <tbachman> #info Uri asks if the list on the presentation is prioritized 18:30:00 <tbachman> #info colindixon says that is a somewhat prioritized list — where the first 5 items are being done as part of the “paying down technical debt”, and the last two items are the things are the end-goal 18:31:16 <tbachman> #inf mohnish asks which pieces are considered plugin — seems like whole OVSDB is a plugin once its migrated to the MD-SAL 18:31:26 <tbachman> #info mohnish asks which pieces are considered plugin — seems like whole OVSDB is a plugin once its migrated to the MD-SAL 18:31:46 <tbachman> #info shague says the current architecture uses the AD-SAL Node; migrating to the MD-SAL includes migrating to the MD-SAL node 18:32:15 <tbachman> #info shague says the netvirt layer which sits on top of the lib, and the layers above it, aren’t considered plugins 18:33:18 <tbachman> #info shague says there are two parallel paths — existing plugin interface and the new MD-SAL interface; as time goes on, this will converge to the MD-SAL plugin 18:34:00 <tbachman> #info Neutron tasks for Lithium include: migrate Nuetron to its own project; rework bundles and features to better represent the Neutron implementation; Triage existing bugs; add YANG model and MD-SAL provider 18:34:15 <tbachman> #info edwarnicke, regXboi, and flaviof have been doing a fantastic job tackling these tasks 18:34:23 * tbachman gives applause 18:35:31 <tbachman> #info see regXboi for meeting minutes 18:35:41 <regXboi> nice :) 18:35:41 <tbachman> #undo 18:35:41 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x18b7d90> 18:35:43 <tbachman> lol 18:36:24 <tbachman> #info regXboi encourages the community to contribute code to the Neutron project 18:36:46 <tbachman> #info regXboi says if there’s something a project wants in Juno or Kilo, then please contribute 18:38:05 <tbachman> #info ODL stackforge: migrates ODL ML2 plugin to its own repo; leverage the plugin functionality in devsatck, so new plugin (aka networking-odl) is seamlessly integrated with local.conf; enable tempest tests; integrate with OVSDB Jenkins jobs to enable robust integration testing 18:44:18 <tbachman> #info Future Work Items: close feature gap with Neutron and OpenStack OVS plugin; migrate netvirt northbound to use MD-SAL neutron; migrate net virt from OVS specific to tunnel management and traffic direction (into tunnels) 18:45:41 * tbachman can’t #info diagrams 18:46:21 <regXboi> wow - that diagram is actually wrong 18:46:35 <tbachman> regXboi: speak :) 18:47:21 <tbachman> #info regXboi says that the OVSDB Evolution diagram doesn’t show the controller in the Current World 18:47:35 <tbachman> #info colindixon says the functionality provided is from the controller 18:48:32 <tbachman> #info LuisGomez says in “near future”, there is a NetCONF piece. Asks about this feature 18:48:58 <tbachman> #info colindixon says that the hope is to provide a set of YANG APIs where you could plugin whatever you want into tunnel management and traffic direction 18:50:46 <tbachman> #info mohnish ass what the policy adapter is in the diagram 18:51:47 <tbachman> #info colindixon says that there’s a collaborative push between GBP and NIC to define a broader policy framework to do things that get realized in neutron, but enable richer behaviors than what neutron currently provides 18:51:49 <Prem> tbachman: You may want to undo the last info message due to a typo 18:51:56 <tbachman> lol! 18:52:00 <tbachman> Prem: thx ;) 18:52:01 <tbachman> #undo 18:52:01 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x19aaf10> 18:52:02 <tbachman> #undo 18:52:02 <odl_meetbot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x19aa890> 18:52:15 <tbachman> #info mohnish asks what the policy adapter is in the diagram 18:52:22 <tbachman> #info colindixon says that there’s a collaborative push between GBP and NIC to define a broader policy framework to do things that get realized in neutron, but enable richer behaviors than what neutron currently provides 18:53:00 * dfarrell07 just got down giving an interview, *super sad* at having missed most of the meeting, will listen to recording ASAP /cc cdub 18:53:40 <tbachman> dfarrell07: yeah — alagalah recorded it, so you didn’t miss anything :) 18:54:04 <dfarrell07> tbachman: perfect, thanks alagalah :) 18:55:25 <tbachman> #info abhijitkumbhare asks about the opendaylight neutron plan and GBP — will they be mapping this architecture into the neutron diagram 18:55:43 <tbachman> #info alagalah says this is a bit out of scope for the TWS, but is happy to address in the GBP meetings 18:56:01 <abhijitkumbhare> https://docs.google.com/presentation/d/1vsYddlHFRnVG9cDwWxyldT2BNSfYUTPcR1lYtUrFA8U/edit#slide=id.g612a6349b_00 18:56:26 <tbachman> #link : https://docs.google.com/presentation/d/1vsYddlHFRnVG9cDwWxyldT2BNSfYUTPcR1lYtUrFA8U/edit#slide=id.g612a6349b_00 abhijitkumbhare just requests if GBP could map this to the slide deck presented here 18:58:06 <tbachman> #info Opportunities to help: MD-SAL migration; MD-SAL southbound plugin; Future work items: advanced services, MD-SAL neutron, tunnel-like southbound; OpenStack engagement 18:58:24 <tbachman> #link https://trello.com/b/ctunVAEr/ovsdb-mdsal-southbound MD-SAL southbound trello board 18:58:40 <tbachman> #link https://trello.com/b/lKnNc7PB/ovsdb-mdsal-migration OVSDB MD-SAL migration trello board 18:58:53 <tbachman> #link https://trello.com/b/ddIvDQE0/ovs-openstack OVSDB trello board 19:00:24 <tbachman> #info alagalah asks what the appropriate mailing list to be used for discussions on this topic 19:00:58 <tbachman> #info regXboi says the community interface should be on discuss 19:01:08 <tbachman> #info regXboi says code-related questions should be on neutron-dev 19:01:31 <tbachman> #info colindixon asks if we have topics set up on discuss 19:01:46 * tbachman suspects everyone is talking over phrobb 19:01:50 <shague> #info big thanks to Ed for getting the mdsal sb crud 19:02:21 <tbachman> #action alagalah to ask if we have topics set up on discuss 19:02:32 <zxiiro> +1 19:03:00 <mohnish> good presentation. thanks. 19:03:50 <Prem> +1 Good presentation and it cleared lots of doubts around OS and ODL integration 19:04:17 <tbachman> #endmeeting