18:02:53 <SumitNaiksatam> #startmeeting networking_policy
18:02:54 <openstack> Meeting started Thu Sep 25 18:02:53 2014 UTC and is due to finish in 60 minutes.  The chair is SumitNaiksatam. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:02:55 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:02:56 <banix> looks like a long weekend around here
18:02:57 <s3wong> SumitNaiksatam: what holiday?
18:02:58 <openstack> The meeting name has been set to 'networking_policy'
18:03:21 <rms_13> Hi
18:03:27 <SumitNaiksatam> #info agenda: https://wiki.openstack.org/wiki/Meetings/Neutron_Group_Policy
18:03:32 <SumitNaiksatam> rms_13: hi there
18:03:43 <banix> s3wong: jewish new year
18:03:56 <s3wong> banix: OK, didn't know :-)
18:03:58 <SumitNaiksatam> #chairs banix s3wong songole rms_13
18:05:01 <SumitNaiksatam> #topic StackForge repos and process
18:05:06 <banix> saw the email by SumitNaiksatam on all the repos
18:05:18 <SumitNaiksatam> #link https://wiki.openstack.org/wiki/GroupBasedPolicy/StackForge/repos
18:05:34 <SumitNaiksatam> banix: yeah, documented it in the wiki page above
18:05:39 <ivar-lazzaro> nice!
18:05:42 <banix> great thanks
18:05:58 <SumitNaiksatam> also registered in launchpad so that we can track bugs
18:06:17 <s3wong> SumitNaiksatam: are we still responsible for our own piece? Meaning that I will at some point need to post the Contract=>SG patch on gerrit?
18:06:33 <s3wong> (once all the dependencies are posted)
18:06:50 <SumitNaiksatam> s3wong: yes
18:07:06 <s3wong> SumitNaiksatam: OK, I will be waiting then :-)
18:07:10 <SumitNaiksatam> s3wong: though you ivar-lazzaro might volunteer to do that, if you are busy ;-P
18:07:22 <ivar-lazzaro> SumitNaiksatam s3wong: yup
18:07:23 <SumitNaiksatam> or rather i just volunteered Ivar
18:07:28 <s3wong> SumitNaiksatam: so I can +2 it right away :-)
18:07:31 <SumitNaiksatam> ivar-lazzaro: sorry!
18:07:55 <ivar-lazzaro> SumitNaiksatam s3wong: It's ok for me, I'm familiar with the code and I'll have to include some changes anyway (contract update)
18:08:17 <s3wong> SumitNaiksatam: what kind of timeline are we looking at --- I am guessing we want to have an official/tested built before K-Summit
18:08:21 <SumitNaiksatam> so in general, as a step zero, we just need to get the patches which are already in review in neutron, trasfered to this review queue
18:08:30 <SumitNaiksatam> s3wong: ok good question
18:08:41 <SumitNaiksatam> so in launchpad, i have created two milestones
18:08:53 <SumitNaiksatam> Sept 30th and Oct 24th
18:09:04 <SumitNaiksatam> i know very agressive, but thats all the time we have
18:09:24 <s3wong> SumitNaiksatam: actually it sounds reasonable
18:09:29 <ivar-lazzaro> *starts brewing as much coffee as possible*
18:09:36 <SumitNaiksatam> so for Sept 30th, we have to have moved all the code into these repos (including the WIP patches for client, horizon and heat)
18:09:39 <SumitNaiksatam> ivar-lazzaro: :-)
18:09:48 <SumitNaiksatam> s3wong: whew, at least one supporter, thanks! :-)
18:10:02 <s3wong> couple nights ago a Sachi Gupta IRC chat with me at 10pm --- and it seems like he or she is interested to try
18:10:26 <SumitNaiksatam> s3wong: okay, cook, yeah he/she pinged me a few times as well
18:10:36 <SumitNaiksatam> *cool
18:10:45 <SumitNaiksatam> s3wong: thanks for taking the time to helping out
18:10:46 <s3wong> so we do have early adopters, and I told him/her we should definitely be ready by K-Summit timeframe
18:10:55 <s3wong> so don't let him/her down :-)
18:11:02 <SumitNaiksatam> s3wong: thats the plan
18:11:29 <SumitNaiksatam> okay, so firstly, any questions on the things mentioned in #link https://wiki.openstack.org/wiki/GroupBasedPolicy/StackForge/repos
18:11:30 <SumitNaiksatam> ?
18:11:36 <KrishnaK> Question: Are there any steps to setup the environment and test ? ( did not see it here https://wiki.openstack.org/wiki/GroupBasedPolicy/StackForge/repos)
18:12:05 <SumitNaiksatam> KrishnaK: process is exactly the same as you would if you were working with Neutron
18:12:37 <KrishnaK> ok. thx Sumit.
18:12:40 <SumitNaiksatam> KrishnaK: but i am happy to document in more detais if required
18:13:01 <SumitNaiksatam> KrishnaK: clone the repo, and then run: tox -e py27
18:13:04 <SumitNaiksatam> ;-P
18:13:22 <SumitNaiksatam> oh wait, sorry the patches are still in review
18:13:40 <SumitNaiksatam> so if you just clone the repo now, you wont get much in terms of code artifacts
18:13:55 <SumitNaiksatam> the last patch in the current series is: #link https://review.openstack.org/#/c/123689/
18:14:24 <SumitNaiksatam> i believe there are about 5 patches now, and we will be adding the others as well shortly
18:14:53 <SumitNaiksatam> perhaps we need a wiki page to document this - #action SumitNaiksatam to add wiki page to document patches in review
18:15:28 <SumitNaiksatam> the hope is that we will be able to review quickly (since we already had several rounds of reviews on the same code in neutron), and land the patches in
18:15:51 <KrishnaK> SumitNaiksatam: Thx. Will look at the wiki and follow the instructions.
18:16:40 <SumitNaiksatam> any other questions on this?
18:17:04 <SumitNaiksatam> or thoughts/suggestions/concerns?
18:18:28 <SumitNaiksatam> so my request is to please review the patches which are in flight, so that we can land them in at the earliest, and actually start testing a working setup
18:18:53 <SumitNaiksatam> #topic IRC channel
18:19:22 <SumitNaiksatam> #info a few days back IRC channel was created for GBP: #openstack-gbp
18:19:34 <SumitNaiksatam> its mostly me and banix in it
18:19:50 <SumitNaiksatam> it will be nice if others join, so that we can quickly discuss and make progress
18:20:23 <SumitNaiksatam> #topic Resource renaming
18:20:24 <ivar-lazzaro> SumitNaiksatam: +1
18:20:53 <SumitNaiksatam> so we earlier decided to rename EP to Policy Targe, and EPG to Policy Target Group
18:21:04 <SumitNaiksatam> *Target
18:21:13 <SumitNaiksatam> those changes have not made it to the patches currently in review
18:21:28 <SumitNaiksatam> but we will add those as a set of follow up patches
18:21:40 <SumitNaiksatam> so everyone still on board with those name changes?
18:22:06 <s3wong> SumitNaiksatam: sure --- but you will have to forgive me if I still refer to EP and EPG...
18:22:14 * ivar-lazzaro still likes endpoint better
18:22:22 <SumitNaiksatam> s3wong: :-)
18:22:32 <banix> wonder if we need to make the change
18:23:01 <SumitNaiksatam> banix: i agree, i like the earlier one better
18:23:04 <banix> now taht we run our own business
18:23:26 <SumitNaiksatam> banix: but we still listen to the community :-)
18:23:35 <ivar-lazzaro> banix SumitNaiksatam: right, but endpoint is still a Keystone terminology
18:23:37 <SumitNaiksatam> that said actually endpoint is being used in places
18:23:41 <banix> but in the long run, i think its better to make the change
18:23:42 <SumitNaiksatam> ivar-lazzaro: yeah
18:23:46 <SumitNaiksatam> banix: yeah
18:23:50 <banix> yeah agree
18:23:57 <SumitNaiksatam> because even if you run the python-client
18:24:01 <SumitNaiksatam> and you do a help
18:24:09 <SumitNaiksatam> it has all these things about the endpoint
18:24:12 <ivar-lazzaro> banix SumitNaiksatam: I would do it in the short term if possible
18:24:20 <SumitNaiksatam> ivar-lazzaro: okay
18:24:23 <banix> ivar-lazzaro: sounds good
18:24:26 <ivar-lazzaro> banix SumitNaiksatam: The first API we show is the one which sticks in people's mind
18:24:29 <SumitNaiksatam> so if you do: neutron —help
18:24:40 <SumitNaiksatam> with the client
18:24:57 <SumitNaiksatam> it talks about the keystone endpoints right at the beginning
18:25:02 <SumitNaiksatam> so that tends to be confusing
18:25:08 <banix> yup
18:25:18 <SumitNaiksatam> i realized this when i was recently working on the demo
18:25:21 <SumitNaiksatam> okay
18:25:40 <SumitNaiksatam> so on “contracts”
18:25:49 <SumitNaiksatam> perhaps better to avoid that as well?
18:26:05 <SumitNaiksatam> a lot of people suggested just using “Policy Rule Set"
18:26:22 <banix> yeah i agree
18:26:25 <ivar-lazzaro> SumitNaiksatam: Contract really is the right word though...
18:26:25 <SumitNaiksatam> so we have “Policy Rule” and “Policy Rule Set” is a collection of policy rules
18:26:39 <SumitNaiksatam> ivar-lazzaro: not everyone likes it though
18:26:46 <banix> we had those terms early on
18:26:58 <s3wong> banix: right we had policy before
18:27:03 <banix> i mean policy rule and policy rule set
18:27:03 <SumitNaiksatam> ivar-lazzaro: i know that rkukura is also not in favor of changing the contract terminology
18:27:15 <SumitNaiksatam> banix: ah, did not realize that
18:27:27 <SumitNaiksatam> banix: so we would just be reverting back, if we were to :-P
18:27:36 <banix> :)
18:27:50 <s3wong> we didn't have policy rule set --- as far as I remember...
18:28:02 <banix> s3wong: we did; let me dig it out
18:28:14 <SumitNaiksatam> so that is one, and the other one is “Policy Labels”
18:28:34 <SumitNaiksatam> so we dont have the implementation for this in the current set of patches
18:28:46 <SumitNaiksatam> but we had planned to
18:29:05 <SumitNaiksatam> and the thinking is that, the labels should actually have a name and a list of values
18:29:11 <s3wong> SumitNaiksatam: labels would be interesting: it can be label associated with policy-rule, label associated with EPG (both provider and consumer)
18:29:12 <SumitNaiksatam> so then its really not a label
18:29:20 <SumitNaiksatam> s3wong: yes
18:29:25 <ivar-lazzaro> SumitNaiksatam: That's really weird to have "policy" everywhere in the API... We are namespacing everything although the entry point already is /grouppolicy
18:29:28 <s3wong> and in ODL model, we further have functionality names under label :-)
18:29:34 <banix> got distracted; will find the refrence later
18:29:36 <SumitNaiksatam> ivar-lazzaro: i agree
18:29:52 <SumitNaiksatam> ivar-lazzaro: but, if this is part of neutron then it makes sense to differentiate
18:30:14 <SumitNaiksatam> ivar-lazzaro: hence the “policy-“ prefix to the names
18:30:32 <SumitNaiksatam> so back to  the earlier point of labels being actually name and values
18:30:44 <ivar-lazzaro> SumitNaiksatam: sounds reasonable
18:30:48 <s3wong> SumitNaiksatam: I think we own "group policy" --- we have done enough advertisement that people in community hearing this would immediately associate that to our project
18:30:57 <SumitNaiksatam> if we are to do that, perhaps we can this “Policy Tags”?
18:31:05 <SumitNaiksatam> s3wong: true
18:31:08 <s3wong> Policy, OTOH, is a bit too high level
18:31:17 <SumitNaiksatam> s3wong: okay
18:31:27 <s3wong> I mean, under neutron/neutron, there is a policy.py file :-)
18:31:37 <SumitNaiksatam> so “Policy Labels” -> “Policy Tags”
18:32:03 <SumitNaiksatam> s3wong: yes there is, and those policies are defined in policy.json
18:32:15 <s3wong> :-)
18:32:19 <SumitNaiksatam> s3wong: resource access control
18:32:58 <SumitNaiksatam> so i will let people think about this a little more: “Contracts” -> “Policy Rule Sets”, and “Policy Labels” -> “Policy Tags”
18:33:19 <SumitNaiksatam> let the team know if you have objections
18:34:01 <SumitNaiksatam> #topic Policy Summit
18:34:07 <s3wong> SumitNaiksatam: we have no concern about terminology discrepancy between Neutron GBP and ODL GBP, right?
18:34:16 <SumitNaiksatam> #undo
18:34:17 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x263d9d0>
18:34:33 <SumitNaiksatam> s3wong: i have not kept pace with the changes in ODL :-(
18:34:47 <SumitNaiksatam> so i dont know, its for us to collectively decide if thats an issue
18:35:04 <s3wong> SumitNaiksatam: AFAIK there hasn't been any changes on ODL GBP terminology since May or June -ish timeframe
18:35:08 <SumitNaiksatam> personally i think we should decide what works best for us here
18:35:19 <s3wong> SumitNaiksatam: certainly
18:35:36 <SumitNaiksatam> s3wong: okay, you guys attend those meetings, so you will definitely have a better feel than me
18:35:59 <SumitNaiksatam> so yeah, i mean if anyone has objections, we are still discussing this, so open to changes
18:36:01 <s3wong> SumitNaiksatam: used to be a big deal in the beginning, honestly
18:36:03 <s3wong> :-)
18:36:12 <SumitNaiksatam> s3wong: :-)
18:36:23 <SumitNaiksatam> just want to put this renaming thing behind us for good
18:36:36 <SumitNaiksatam> thought it was done when we decided to rename EP/EPG
18:36:40 <Guest40815> ONF NBI arch group is using the GBP terminoloty in it models for service chaining (contracts, endpoints, endpointgroup, classifier)
18:36:53 <SumitNaiksatam> but people keep coming back on other things
18:37:16 <SumitNaiksatam> Guest40815: thanks, thats helps, i guess! :-(
18:37:28 <s3wong> Guest40815: Oh, really. That is good to know
18:37:36 <SumitNaiksatam> so now we are going to go in circles
18:38:12 <banix> yeah lets not be tied to odl terminology
18:38:36 <banix> odl will be one driver at the end of the day
18:38:48 <SumitNaiksatam> banix: okay
18:38:51 <banix> even though there are close ties in concepts and all
18:39:18 <SumitNaiksatam> Guest40815: but thanks for chiming in and providing the reference
18:39:24 <banix> we wont be able to change those and need to adapt to requirements by openstack
18:39:41 <SumitNaiksatam> Guest40815: its just that we have to make some progress here
18:39:55 <banix> s3wong: ivar-lazzaro make sense?
18:39:56 <SumitNaiksatam> Guest40815: it seems that endpoint/endpointgroup is not going work in the openstack context
18:40:08 <SumitNaiksatam> banix: i agree
18:40:12 <ivar-lazzaro> banix: +1
18:40:26 <s3wong> banix: sure - I should focus on the best terminology to gain support with the OpenStack community
18:40:28 <Guest40815> ONF NBI can adapt,
18:40:30 <SumitNaiksatam> Guest40815: any strong objections to what we are proposing here?
18:40:32 <s3wong> s/I/we
18:40:46 <SumitNaiksatam> s3wong: +1
18:41:07 <s3wong> Now I know who Guest40815 is :-)
18:41:17 <SumitNaiksatam> bouthors: good to know you are Guest40815 :-)
18:41:27 <SumitNaiksatam> s3wong: ^^^
18:41:29 <s3wong> bouthors: how are you doing, Nicolas
18:42:06 <bouthors> great thx, I was please to see that ONF NBI was following GBP tracks
18:42:22 <SumitNaiksatam> bouthors: ah, okay, thats a good data point
18:42:27 <SumitNaiksatam> i share that excitement as well
18:42:37 <SumitNaiksatam> #topic Policy Summit
18:43:02 <SumitNaiksatam> so the GBP team participated in the Policy Summit last week
18:43:22 <SumitNaiksatam> we did the presentation during the networking part of the agenda
18:43:45 <SumitNaiksatam> a big THANK YOU to all those who really slogged it out for getting the demo ready in time
18:44:15 <ivar-lazzaro> SumitNaiksatam: yeah, thank you all!
18:44:31 <SumitNaiksatam> i have posted the slides
18:44:39 <SumitNaiksatam> i cant seem to find the link though
18:44:44 <s3wong> ivar-lazzaro: special thanks to you! :-)
18:45:05 <ivar-lazzaro> s3wong: :-)
18:45:06 <banix> i just saw it googling; looking for earlier docs
18:45:26 <s3wong> SumitNaiksatam: http://www.slideshare.net/sumit_naik/groupbased-policyjunopolicysummit
18:45:29 <s3wong> this one?
18:45:51 <SumitNaiksatam> s3wong: yeah, thanks
18:46:10 <SumitNaiksatam> unfortunately we dont have the recorded demo
18:46:28 <SumitNaiksatam> if we get the time, we will record one and upload it
18:46:47 <SumitNaiksatam> anyone else wants to fill in what happened at the Policy Summit?
18:48:01 <banix> how was the second day?
18:48:06 <banix> the workshop part
18:48:26 <SumitNaiksatam> banix: unfortunately i was not able to attend the second day
18:48:27 <banix> i see (and s3wong also noticed) some tasks/actions created related to gbp
18:48:35 <ivar-lazzaro> banix: mainly exercises on how to model scenarios using congress
18:48:56 <ivar-lazzaro> banix: was really good, I got the impression that people were very interested in both congress and GBP
18:49:11 <banix> ivar-lazzaro: ok thanks
18:49:31 <s3wong> SumitNaiksatam: echoing what banix said, there are a lot of action items on how GBP works with Congress
18:49:40 <hemanthravi> banix: second day folks broke up into group to come with bp and implement define policy with congres
18:49:49 <s3wong> SumitNaiksatam: and for most of them, LouisF and Cathy volunteered
18:49:52 <ivar-lazzaro> banix: the first gives a general way of describing things, while most of the people there agreed that GBP was a very simple way for describing policy scenarios
18:50:22 <hemanthravi> congress has come up with a language, for networking congress will use GBP APIs
18:50:43 <SumitNaiksatam> ivar-lazzaro: hemanthravi s3wong: good to know
18:51:09 <SumitNaiksatam> #topic Open Discussion
18:51:24 <SumitNaiksatam> anything else to discuss today?
18:51:41 <SumitNaiksatam> else we can get back 9 mins of our time :-)
18:51:48 <s3wong> SumitNaiksatam: for the K-Summit, hemanthravi, banix, you, and me would have to start looking at presentation stuff
18:52:02 <SumitNaiksatam> s3wong: yes working on that
18:52:12 <banix> s3wong: yes
18:52:13 <SumitNaiksatam> s3wong: mandeep had put together some slides
18:52:28 <SumitNaiksatam> s3wong: i will add the slides from the policy summit to that
18:52:43 <SumitNaiksatam> s3wong: just to create a pool of slides
18:53:02 <SumitNaiksatam> but, yeah fast approaching!
18:53:08 <s3wong> SumitNaiksatam: based on the reception we received on the demo during the policy summit, we may want to NOT show people SG
18:53:10 <SumitNaiksatam> people done with their travel arrangements?
18:53:12 <s3wong> just saying :-)
18:53:18 <SumitNaiksatam> s3wong: yes
18:53:19 <s3wong> seems to cause a lot of confusion
18:53:26 <SumitNaiksatam> s3wong: we dont have SG in the slides :-)
18:53:41 <SumitNaiksatam> s3wong: demo is a different beast
18:53:49 <s3wong> SumitNaiksatam: nice :-) jt was in the backup slide on mapping driver, though :-)
18:53:49 <SumitNaiksatam> s3wong:  we need to plan that carefully
18:53:57 <banix> yes regarding travel arrangements
18:54:04 <SumitNaiksatam> s3wong: ah okay, that was the cue for rkukura :-)
18:54:13 <SumitNaiksatam> banix: you booked?
18:54:16 <SumitNaiksatam> i haven’t
18:54:20 <banix> SumitNaiksatam: yes
18:54:38 <SumitNaiksatam> its more challenging from here
18:55:01 <banix> going 10 days before and staying 10 days after :)
18:55:04 <banix> just kidding
18:55:12 <SumitNaiksatam> banix: you got me fooled
18:55:18 <banix> plan to be there on Sat before the summit and return the next Sat
18:55:18 <SumitNaiksatam> banix: i think we all wish! :-)
18:55:28 <s3wong> banix is traveling around Europe during that time :-)
18:55:33 <SumitNaiksatam> banix: okay, i think rkukura has similar plans
18:55:53 <SumitNaiksatam> alirght then, if nothing else, lets call it a wrap
18:55:53 <s3wong> I will be there on Sunday and will leave on Saturday after
18:56:01 <banix> s3wong: no such plans unfortunately
18:56:03 <SumitNaiksatam> s3wong:  cool
18:56:21 <SumitNaiksatam> back to the junu-gbp-1 milestone ;-)
18:56:24 <SumitNaiksatam> thanks all!
18:56:27 <SumitNaiksatam> #endmeeting