18:02:46 <regXboi> #startmeeting
18:02:46 <odl_meetbot> Meeting started Fri Apr 18 18:02:46 2014 UTC.  The chair is regXboi. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:02:46 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:02:56 <regXboi> #topic architecture meeting
18:03:26 <dconde> rah
18:04:33 <regXboi> #info mickey_spiegel asks that since architecture touches the MD-SAL a lot, are people with comfortable with using the MD-SAL
18:04:59 <mickey_spiegel> #info For now, assume MD-SAL for purposes of discussion. Data store subgroup will figure out if we actually go with MD-SAL
18:08:20 <dconde> #info jmedved said we need to find appropriate sharding mechanisms
18:08:54 <dconde> #info regxboi says it makes sense and he spoke w/ Colin
18:09:13 <alagalah> #info jmedved wants to leverage a lot of the work the APIC team has done, seems like the idea is AKKA would create the shards and leverage APIC's model of consistency
18:09:13 <regXboi> #chair dconde
18:09:13 <odl_meetbot> Current chairs: dconde regXboi
18:09:20 <regXboi> #chair alagalah
18:09:20 <odl_meetbot> Current chairs: alagalah dconde regXboi
18:10:08 <dconde> #info dvorkinsta state we need to figure out how the requirements we have on MD-SAL and whether it satisfied our requirements.
18:10:16 <dconde> #info we need to do that quickly.
18:10:51 <dconde> #info WE need to understand if AD-SAL can provide that.
18:11:19 <regXboi> #info edit: above should be MD-SAL
18:11:23 <alagalah> #info jmedved says he has a interesting architectural blueprint for the MDSAL
18:12:57 <regXboi> #info concerning using akka for knitting together shards that are stored within existing in-memory structures
18:12:59 <dconde> #info sharding via Akka. Matthew will have preentation on Tuesday. need experimentation to see
18:13:13 <dconde> #info want to take a step back
18:13:27 <dconde> #info ask whether to go w/ MD-SAL or not.
18:13:45 <dconde> #info if not, what do we use ?
18:14:04 <dconde> #info the alternative, regxboi states, is that we better come up with alternative.
18:14:23 <dconde> #info SAL is the backbone of the controller.
18:14:40 <dconde> #info so if we do not use it (model driven or not), then we are building a different system.
18:14:52 <dconde> #info the msg bus lies within SAL.
18:15:11 <regXboi> #info note: above three statements are from jmedved
18:15:21 <dconde> #info the APIs are eventually provided in the SAL. IF not, we need to build all these mechanisms ourselves.
18:16:08 <dconde> #info regxboi says let's make explicit. DATASTORE group was asked with a way to store policies in a datastore
18:16:18 <dconde> #info that is different from whether MD-SAL is the routing backbone.
18:16:45 <dconde> #info if we need to the DATASTORE people to take that task, it's a separate issue.
18:17:24 <dconde> #info jmedved says SAL can be made independent of data store, and stll be within architectural footprint of controller and
18:17:35 <dconde> lost audio?
18:17:54 <readams> jmedved: you dropped off hangout
18:17:57 <alagalah> jmedved: you dropped
18:18:16 <dconde> #info we lost Jan on audio
18:19:15 <dconde> #info regxboi says rate and amount are important.
18:19:38 <dconde> #info dvokinista states JMedVed's Mac crapped out.
18:20:34 <dconde> #info mickey_spiegel states thhat datastore's requirements are more important than SAL.
18:20:56 <mickey_spiegel> #info not more important, I just thought that comes first, whereas Jan is looking at it the other way around
18:20:59 <dconde> #info dvorkinista says let's don't do it bottoms up.
18:21:32 <dconde> #info alagalah says that it is in Hydrogen
18:21:51 <dconde> #info regxboi say we hope that it gets better in Helium (perf wise)
18:21:58 <dconde> #info Jmedved is back.
18:22:20 <dconde> #info jmedved MD-SAL perf #s.
18:22:39 <dconde> #info one is pure RPC notifivation perf CBENCH.
18:24:07 <dconde> #info readams asks where we want to use these components.
18:24:37 <dconde> #info dvorkista say as long as it's fast enough it's good t
18:24:59 <dconde> need someone to scribe
18:25:22 <regXboi> #info quick 101 of MD-SAL
18:25:40 <regXboi> I'll help out... I'm done with getting food (I was crashing hard)
18:27:48 <regXboi> #info one piece jmedved says is missing in MD-SAL is ability to do queries
18:28:28 <regXboi> #info dvorkinista wants to use policy requirements to drive how MD-SAL does queries
18:28:33 <regXboi> #info jmedved agrees
18:29:13 <regXboi> #info jmedved claims on RPC that performance is similair to floodlight
18:29:24 <regXboi> #info edit: similar
18:30:10 <regXboi> #info mickey_spiegel asks about details
18:30:24 <mickey_spiegel> #info about subscribing to subtrees
18:30:24 <regXboi> #info jmedved says we can, but he's not prepared to do it right now
18:30:52 <regXboi> #info mickey_spiegel asks about classes versus objects
18:31:08 <regXboi> #info jmedved says the yang model defines the api that are generated
18:31:59 <regXboi> #info but he's not sure whether subscription is to class or object
18:33:20 <mickey_spiegel> #info sounds like object is currently supported, class is not yet supported, requires some development that can be done
18:36:57 <mickey_spiegel> #info Ryan is OK with the idea of having an implementation of the data store that is packaged with the controller and ships
18:36:57 <mickey_spiegel> Not OK with there being only one possible implementation
18:37:09 <mickey_spiegel> # info Not OK with there being only one possible implementation
18:37:15 <regXboi> thanks you got it
18:37:23 <mickey_spiegel> #info Jan notes already have two implementations of data store
18:39:35 <regXboi> #info mickey_spiegel asks about NB APIs
18:40:07 <regXboi> #info Jan answers that you define the NB API in yang and it gets translated into RESTCONF
18:41:18 <regXboi> #info dvorkinista asks for clarity that every object gets an API
18:41:31 <regXboi> #info jmedved says yes and that is a plus and a minus
18:41:46 <regXboi> #info because there aren't access policies today
18:42:50 <regXboi> #info discussion that access control will be applied to trees
18:42:57 <regXboi> #info today everything is open
18:43:32 <regXboi> #info no RBAC today... that needs to be done
18:47:53 <mickey_spiegel> #info Need discussion with OpenStack group-based policy at some point to make sure ODL RESTCONF is compatible
18:48:21 <regXboi> #info also need to nail down the NB API yang early in the game to launch that discussion
18:48:53 <s3wong> from the OpenStack GBP side, I can tell you that we don't care
18:49:04 <s3wong> this is more a problem with the ODL driver / policy-driver
18:50:02 <regXboi> s3wong: exactly... I want to make sure the folks who wrote the ODL driver (mestery) are aware of what's happening
18:50:46 * mestery listens in.
18:50:54 <regXboi> folks --  I need to go get ready for my next meeting ... somebody else gets to close the meeting
18:51:18 <regXboi> #chair mickey_spiegel
18:51:18 <odl_meetbot> Current chairs: alagalah dconde mickey_spiegel regXboi
18:51:31 <dconde> bye regxboi
18:52:00 <dconde> #info mickey_spiegel question is how to bind to neutron
18:52:18 <dconde> #info port bindings in a normal way vs. GBP?
18:52:43 <dconde> #info s3wong say absolutely we want ot use ODL GBP without neutron GBP.
18:53:13 <dconde> can't hear readams well
18:53:35 <dconde> #info dvorkistia requests agenda for Friday meetings.
18:55:20 <alagalah> #endmeeting