17:05:23 <regXboi> #startmeeting model subgroup 17:05:23 <odl_meetbot> Meeting started Fri May 16 17:05:23 2014 UTC. The chair is regXboi. Information about MeetBot at http://ci.openstack.org/meetbot.html. 17:05:23 <odl_meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:05:23 <odl_meetbot> The meeting name has been set to 'model_subgroup' 17:05:33 <regXboi> #chair tbachman 17:05:33 <odl_meetbot> Current chairs: regXboi tbachman 17:05:40 <tbachman> topic? 17:05:42 <regXboi> #chair readams alagalah 17:05:42 <odl_meetbot> Warning: Nick not in channel: alagalah 17:05:42 <odl_meetbot> Current chairs: alagalah readams regXboi tbachman 17:05:51 <tbachman> he's got a "_" 17:05:52 <tbachman> fancy 17:05:56 <regXboi> #nochair alagalah 17:05:56 <alagalah_> #link https://wiki.opendaylight.org/view/Group_Policy:Sub-Groups:MODEL#Team_Meeting 17:06:01 <regXboi> #chair alagalah_ 17:06:01 <odl_meetbot> Current chairs: alagalah alagalah_ readams regXboi tbachman 17:06:15 <regXboi> #unchair alagalah_ 17:06:15 <odl_meetbot> Current chairs: alagalah readams regXboi tbachman 17:06:22 <tbachman> #topic white board diagram of 3-tier to JSON 17:06:28 <tbachman> regXboi: lol 17:07:29 <tbachman> #info alagalah created JSON of 3-tier model 17:07:48 <alagalah> #info Regarding action of 3 tier app softcopy and JSON generation find information here: 17:07:53 <alagalah> #action https://drive.google.com/folderview?id=0BztNICcppsJKY0tMRl8tVlBQN2s&usp=sharing 17:08:24 <alagalah> #info work in progressw 17:08:38 <tbachman> #topic Create power point of 3-tier 17:08:50 <tbachman> #info work in progress -- task takes time 17:08:55 <alagalah> #info action from last meeting: alagalah to instead of writing whitepaper or annotating the UML to write a PPT with animations that describes the UML model with annotations. 17:09:01 <alagalah> #info WIP : blocked on time 17:09:24 <alagalah> #info Action: dvorkinista to restruct model in structure, definition use - has not had time. 17:09:31 <alagalah> #info no update, Mike at OS summit 17:09:44 <alagalah> #info Action: alagalah regXboi edwarnicke to discuss service chaining, specifically their datastore/subscription requirements and how they are planning on addressing this with MD-SAL etc 17:10:16 <tbachman> #topic Action Items (though look to other topics on resolution -- didn't get topics right at first) 17:10:35 <alagalah> #action alagalah to schedule meeting with edwarnicke and regxboi (possibly ChrisPrice) to discuss Action: alagalah regXboi edwarnicke to discuss service chaining, specifically their datastore/subscription requirements and how they are planning on addressing this with MD-SAL etc 17:11:29 <alagalah> #info Action: alagalah meeting with Uyen, Rex, Dave Lenrow, regXboi to discuss Use-Case -> UML -> JSON "conversion" ... 14:30 Pac Fri May9 17:12:10 <tbachman> #info meetings have been ongoing -- Unified communications use case with proactive flows. Team has been working on mapping use case to the model 17:13:21 <tbachman> #info Microsoft Lync uses many ports, so the question has been of how to provide information for classifier so that renderer can install flows between session peers. 17:13:42 <tbachman> (uchau: feel free to amend these if I get them wrong ;) 17:14:08 <uchau> not what alagalah is saying 17:14:15 <uchau> but more of what we decided on :) 17:16:27 <tbachman> #info plan is to manage call setup of 200+ per second 17:16:56 <tbachman> #info readams says that changing the config dynamically is not the way to do this 17:17:22 <tbachman> #info classifier represents dynamic concept, such as lookup in database (i.e. look up of metadata) 17:21:04 <tbachman> #info readams recommends move lookup of 5-tuple to different store, which better supports the applications needs 17:23:59 <tbachman> #info readams says operational data doesn't belong in the policy, which is a configuration store 17:28:32 <tbachman> #info mickey_spiegel notes that we're maybe talking about two APIs here. 17:29:10 <tbachman> #info readams asks if there is an orchestration system for the UC&C application to orchestrate this 17:29:26 <tbachman> #info uchau says there's an application on the controller (idea is based on ONF use case) 17:29:54 <tbachman> #info user app is integrated with UC&C system, network app that is part of the contoller, and is policy based system 17:32:09 <tbachman> #info readams says that GBP does not make sense as a generic midlleware layer 17:33:24 <tbachman> #info GBP model is different from the way normal network management is done. 17:35:00 <tbachman> #info mickey_spiegel notes asks for definition of what uchau means by policy 17:35:31 <tbachman> #info losing audio 17:36:49 <tbachman> #info regXboi is concerned about narrow definition of what GBP is trying to provide (e.g. SFC) 17:37:23 <tbachman> #info if there are orchestration systems NB with a different definition of policy model, he doesn't want us to say "oh we can't talk to you" 17:37:50 <tbachman> #info readams notes that if the model is based on fundamentally different principles, this might not make sense. 17:38:49 <tbachman> #info regXboi notes that this has to work with other orchestration systems (cough cough *openstack* cough cough) 17:38:53 <alagalah> ping we have network issues here... its seems new sessions cant be made 17:40:41 <tbachman> #info readams asserts that in cases where user sees endpoints and communication between these groups of endpoints, then GBP works well 17:41:45 <regXboi> uchau: not everybody is near the mike 17:41:49 <regXboi> er mic 17:43:03 <tbachman> #info uchau says that there are users who want to establish link calls 17:43:34 <tbachman> #info losing audio 17:53:03 <tbachman> #info there are two broad categories of data: operational state data and configuration data 17:53:17 <tbachman> #info for example, a new session could be considered operational state data 17:53:35 <tbachman> #info the policy needed by a session is considered configuration data 17:54:05 <tbachman> #info define a classifier that is "link communication", as "anything that is communication on an active link session" 17:55:47 <tbachman> #info semantics around operational and configuration state data are different 17:55:58 <tbachman> #info persistence, transactionality, referential integrity 17:56:16 <tbachman> #info This affects how we represent this data 17:57:14 <tbachman> #info question of can we change actions, classifiers, or tenants in a dynamic fashion? 18:00:17 <tbachman> #info it's 11am PST 18:00:26 <tbachman> we're officially over folks 18:00:33 <regXboi> we are going to bleed over 18:00:58 <tbachman> should we continue this in the next hangout? 18:01:57 <regXboi> let me break in and ask :) 18:02:01 <regXboi> and thanks for scribing 18:02:14 <tbachman> regXboi: no problem! 18:02:22 <tbachman> you carry the burden most of the time! 18:06:15 <tbachman> #endmeeting