13:00:22 #startmeeting OPNFV Day Planning at OpenStack Summit 13:00:22 Meeting started Wed Apr 29 13:00:22 2015 UTC. The chair is rpaik. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:22 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:22 The meeting name has been set to 'opnfv_day_planning_at_openstack_summit' 13:00:43 hello everyone… 13:00:53 Morning Ray 13:01:05 hello 13:01:10 #info Daniel Smith 13:01:15 hello 13:01:28 #info Howard Huang 13:01:31 is this the sec meeting? 13:01:35 dneary, are you with us? 13:01:48 rpaik, Yes 13:01:52 Here 13:01:59 #info Heather Kirksey 13:02:00 @mike - this is a meeting for OPNFV Event during OStack Summit 13:02:00 lmcdasm: Error: "mike" is not a valid command. 13:02:05 0ike - this is a meeting for OPNFV Event during OStack Summit 13:02:06 this is the OPNFV day planning meeting for OpenStack 13:02:06 #info Bryan Sullivan 13:02:09 mike* 13:02:15 k. thx. bi 13:02:25 # 13:02:28 #info Ray Paik 13:02:31 #info dneary 13:02:39 #info 13:02:53 we only have ~30 mins so let’s get started…. 13:03:26 rpaik, Are you chairing? 13:03:30 As discussed yesterday, we want to organize the “breakout sessions” in the afternoon on May 18th 13:04:12 can you outline quickly (for those of us not in yuesterday;s mneeting) what the goals of the breakout sessions are 13:04:19 I can, but we can have multiple chairs :-) 13:04:19 and what we want to highlight in them? 13:04:31 and add the etherpad link here 13:04:39 #link https://etherpad.opnfv.org/p/Vancouver_OpenStack 13:05:33 So this event was originally caterogized as a Hackfest — it’s evolved but in general, the goals of the breakout sessions would be what’s useful to the technical community keeping in mind that we’re at OpenStack, so it’s a good chance to engage with that community 13:05:42 lmcdasm, The goal of an individual session, or of the day as a whole? 13:05:47 lmcdasm, They're not quite the same 13:05:57 hey.,. well. im just reading throuygh all this material that i guess was put together yuesterday 13:06:02 from the etherpad 13:06:09 lmcdasm, It's been there for quite a while 13:06:12 so im trying to catch up i guess.. seems to me the format is already laid out 13:06:41 ok ok - well this is the first meeting about OStack summit that i have had and as with everything in our OPNFV< unless someone says this is the link we are working on ytou cant find anything really 13:06:44 (no indea / points). 13:06:48 so thanks for that info Dave 13:06:54 The morning will kickoff with a general/overview session in a reverse fishbowl layout 13:07:04 lmcdasm, The goal of the event is to (a) provide OPNFV members and non-members an opportunity to collaborate on promoting NFV use-cases to the OpenStack community, and (b) provide an opportunity for collaboration with OpenStack and OPNFV 13:07:35 So, promotion of telco stuff + refinement & communication 13:07:41 I think it would be good for OPNFV members and OpenStack members to discuss results from Functest and other use case testing (as the prerequisite customizations necessary to arrange them) of Juno under the OPNFV R1 13:07:49 Excessive use of collaborate there, but... 13:08:00 In the afternoon, we’re planning on rearranging the room into a breakout format. 13:08:14 Our room can accomodate upto 275 people 13:08:53 one thing to consider, openstack folks might not be up to speed on OPNFV and it's projects 13:09:01 bryan_att, interesting suggestion. Do you think we’ll have much insight by then, given Arno’s schedule? 13:09:01 lmcdasm, The goal of an individual session would be similar to a design summit session - work through any potential issues/disagreements, and end with concrete plan of action for next OpenStack release (I'd like to see sessions essentially set the big picture and have participants then driving their agenda later in design summit sessions) 13:09:13 if we are focusing on Telco, then im a bit fuzzy on talking functest results, since we havent done a test with a real VNF yet (no applications running anywhere that i have seen) 13:09:19 RayNugent, we’re planning in kickin off the morning with an overview 13:09:44 Do you think we should get into individual project level w/ that overview? 13:09:51 bryan_att, I wonder how much interest there will be in Juno functest results when Kilo is out... 13:09:55 who's doing that session? 13:10:13 agree Bryan 13:10:21 Before we get into the "who" details, can I get a clarification on the format? 13:10:43 so.. there are lots of buzz words, but im missing the actual content 13:10:51 is this is simply a session where we sit and chat tech 13:10:55 HKirksey, From yesterday, my understanding is that there will be 1 room in the morning (for a shortish time), split into 2 rooms after lunch (for 4 hours) 13:11:01 HKirksey, Is that right? 13:11:04 or are we showing something (since im building up a IMS core right now to demo on our ARno release0? 13:11:08 dneary that’s correct 13:11:16 rpaik, How long is the morning? 13:11:22 lmcdasm, That's what we're here to talk about 13:11:24 we have some flexibility in what we request 13:11:29 honestly 13:11:31 it’s the same room but can accommodate upto 275 13:11:36 we can keep the whole fishbowl setup al day? 13:11:44 HKirksey, OK, nice to know, but we should nail it down 13:11:53 we can subdivide it….we can keep it whole but have breakout style scattered around 13:12:02 ok.. sorry .. im confused, ive heard a panel session to talk, ive heard breakout session in the same room (which im not sure will work without dividers - some people (like me) have a voice that carriers) 13:12:21 and i have requests to have a live demo showing Arno with a IMS Core and Scaling on it 13:12:23 lmcdasm, That's why I wanted to start with the space & time availability, and agree on a format 13:12:28 ok. 13:12:33 Makes it easier to slot things into a grid afterwards 13:12:42 I think a pannel in either the AM or PM would be good 13:12:48 HKirksey, I think if we’re to subdivide we should do that during lunch? 13:13:07 Yes, any room changes would need to happen over lunch and then need to stay that way 13:13:10 rpaik: How much time do we have in the morning? It's 11am to 12:30 or something? 13:13:37 11:15 to 6:10PM 13:13:49 bryan_att, I mean, before lunch 13:14:05 is there a Agenda / daily outline anywhere for times(other than the etherpad which is hard to pick out) for the week thus far? or is the etherpad it? 13:14:31 The time we have available is from after the opening keynotes until 6 PM when the show follor opens for the evening reception 13:14:41 lmcdasm, Yes, the agenda for the full week is here: https://openstacksummitmay2015vancouver.sched.org/ 13:14:45 #link https://openstacksummitmay2015vancouver.sched.org/ 13:14:50 dneary, it looks like lunch is between 12:45-2? 13:15:01 Also we can do as many breaks as we want... 13:15:09 rpaik, So, concretely, we have 11:15-12:45 and 2pm-6pm? 13:15:23 Right now on the "OPNFV Day at OpenStack " section of the etherpad, there is a topic list. We should refine the list of topics to be addressed in the first session. It looks like we have 1 hr 45 min before lunch (~1pm). 13:16:03 bryan_att, If we can, I'd like to figure out what time we want to use before we figure out how to use it 13:16:07 maybe time for 3 30-min and 3 15-min talks/panels? 13:16:21 Before lunch I was imagining we present an overview of OPNFV, how we envision working upstream, and our projects for anyone attending from OpenStack to understand what we’re up to 13:16:33 HKirksey, That makes sense to me 13:16:39 The telco workg group also requested some time for us to talk about the upcoming week, various design summit sessions and how to prioritize them from a telco perpsective 13:16:40 dneary, I think that’s correct. HKirksey, let me know if you hear otherwise from OpenStack folks.... 13:16:54 We have 1h30 - that is either 3x30 min sessions or 2x45 min sessions 13:16:58 So right now that’s on the schedule — I thought that would be good as a collaborative activity that we would all care about 13:17:02 start some bonding and concrete discussion 13:17:22 im sorry - which point are we discussion now 13:17:28 between, OPNFV intro, gapa analysis and TWG we already have a full day 13:17:29 cause i think there is three threads here 13:17:36 for the breakoutsession are you doing 2 or 3 13:17:36 gap 13:17:37 ? 13:17:44 And then in the afternoon we can break out however we wish — we should probably reserve 10-15 m inutes to come back together for closing, even if the room setup won’t accomodate it perfect, we can squeeze in 13:17:53 Also, we’re doing a reception that evening. :) So beer for all 13:18:00 lmcdasm, What we do in the morning 13:18:12 travel 13:18:23 sorry I meant we should have time for 2 30-min and 2 15-min talks/panels in the AM 13:18:38 ? 13:18:40 #info We have 1h30 in the morning (11:15-12:45), and 4h in the afternoon (2pm-6pm) to plan 13:18:48 can the overview Heather mentioned be done in 30 min? 13:19:09 not reallt 4h in PM cuz there are breaks 13:19:20 #info HKirksey proposes that we use the morning to set the scene, present an overview of OPNFV, perhaps have a panel discussion and something with the Telco WG 13:19:32 #info The afternoon can then be break-out sessions 13:19:44 I think it’d be better to budget more than 30 mins for overview to accommodate questions 13:19:48 Do those #infos represent agreement on the format? 13:19:48 agreee with dneary 13:20:04 +1 13:20:14 sorry - its not clear what was agreed 13:20:17 rpaik, The freedom is whether to do 30' or 45' sessions 13:20:19 a 30 minute overview session? 13:20:23 or the afternoon format - 13:20:25 lmcdasm, Not yet 13:20:44 ok.. sorry.. seems im a blocker here.. ill drop out and let you guys set it since its well on the way 13:20:46 lmcdasm, If I may, here's how I propose we converge to a schedule: 13:20:53 45 min overview, 30 min breakouts 13:20:54 I think we need more than 30 mins. to accommodate questions 13:20:56 and you can send me the agenda. since im causing more question sthan not 13:21:11 1. Agree how much time we have in the morning and afternoon, and general goals of morning and afternoon (done) 13:21:24 2. Agree on format of morning and afternoon sessions (we are here now) 13:21:40 3. Figure out topics and people for morning and afternoon sessions (not here yet) 13:21:56 1.5 agree on how input to the AM overview will be collected 13:22:13 bryan_att, I'm not sure I understand 13:22:32 we need more agreement than the format. where is the content coming from 13:22:47 For the overview I can volunteer for that 13:22:54 And also rope in our illustrious TSC chair 13:22:58 rpaik, I don't know how many questions you are going to get in a 250 person room 13:23:07 It's not exactly amenable to discussion 13:23:38 if we have a project overview, what's included and where does the content come from - that was my question 13:23:58 bryan_att, Ah... I say pick a qualified presenter and trust them 13:24:15 * dneary is against micromanaging content people will present 13:24:29 so that would be you Dave :-) 13:24:48 So - 45' overview and 45' panel discussion with some ETSI & Telco WG people? 13:25:02 RayNugent, Heather volunteered (and also volunteered ChrisPriceAB) 13:25:25 Is the panel the second thing you’re suggesting we do w/ our morning? 13:25:42 you said "qualified" 13:25:43 my point was not to control the presenter, but make sure they have the necessary support so that the overview is comprehensive, at the necessary level for the time, and conveys the most effective info for the audience. that will require input IMO from the project leads, even if presented by one person 13:25:58 since you've done openstack before... 13:25:59 dneary, that seems like a good way to organize the morning 13:26:11 Wow, RayNugent 13:26:24 JK! :-) 13:26:29 I figured. ;-p 13:26:41 Keep throwing the darts. I can take em; 13:26:44 do we get a lot of attendees from ETSI @the summit? 13:26:44 RayNugent, Not a good one... 13:27:03 rpaik, Well, the ETSI NFV meeting is happening the same week in China 13:27:14 So I don't expect a lot of people prominent in ETSI 13:27:26 does ETSI ever attend openstack? 13:27:32 But I think we should make an effort to have some people who think ETSI is important on a panel, if we do one 13:27:46 RayNugent, Many ETSI participants do 13:28:01 Anthony Soong is ETSI, who else? 13:28:06 dneary, I’m a littl hesitant on the panel, just because in a panel format I’m not sure we’d accomplish a ton 13:28:06 RayNugent, ETSI has not exhibited there, as far as I know 13:28:32 HKirksey, I agree, panels suck. However, I thought you had proposed one 13:28:47 Oh. No. Probably being unclear this early in the morning. 13:28:56 HKirksey, It's not like it's 6am... 13:29:24 Can we move on to something we can agree on, temporarily? 13:29:24 The telco working group had said that they wanted to look at the week schedule and design summit sessions and talk about ones we might care about as a larger group and prioritize participations at them 13:29:26 yeah, west coast ;-) 13:29:40 The problem is that sounds interactive which might be difficult in the setup, but could be useful for everyone to be part of 13:29:46 * dneary thinks we can quickly agree on the afternoon format 13:29:46 and would be a point of concrete collaboration 13:30:03 Good. Let’s talk afternoon 13:30:10 could we invite telco to speak? 13:30:11 I propose that we do 2 parallel sessions (no more, to avoid dissipation of energy) 13:30:33 dneary, yes that’s where we left off yesterday 13:30:35 And that we have 1h sessions (30' is short if you want discussion and collaboration) 13:31:00 So that's 4x1h sessions in 2 tracks, 8 spots available for topics 13:31:12 RayNugent, Yes, we could. 13:31:31 RayNugent, They're working on use-cases that are very similar to "requirements projects" in OPNFV 13:31:36 They don't line up exactly 13:31:39 and proposed topics were IPv6, Doctor, Multisite, SFC, VNFFG 13:31:58 Yes - potentially putting SFC and VNFFG together 13:32:13 And Promise 13:32:25 That's 6 topics I see a need to have sessions on 13:32:41 good morning early birds! 13:32:51 Perhaps we can take the time to have a big cage match about installers & alternate stacks too 13:33:03 I think first we need to find out if project participants will be at OpenStack to lead discussions 13:33:25 Yeah!!! dneary 13:33:28 See…now we’re talking…. OPNFV Day = Mad Max at Vancouver Dome 13:33:34 rpaik, If we have a list of potential topics, it's easier to figure out who might lead them, and who needs to be there 13:34:03 dneary agreed. guess I’m stating the obvious 13:34:35 I propose that we have one topic that addresses policy-related use cases and projects in OPNFV, and OpenStack components - as suggested on the etherpad 13:34:43 rpaik: just ammend the comment - first get a list of topics then second see who might be there from those projects 13:34:45 I know that the Multisite team will be there, there is lots of IPv6 interest, Doctor and VNFFG will both be well represented... don't know about Promise 13:34:46 I can get folks for multisite and VNFFG 13:34:52 So, Copper, bryan_att? 13:34:53 ha also need 13:35:12 bryan_att, Yes, I forgot Copper. Sorry, Bryan 13:35:35 hui, are you saying we shouls also have a session on HA? 13:35:42 should 13:35:58 OK - so let me #info some of what we have agreed on: 13:36:00 I think so 13:36:02 no, overall. Copper, Doctor, Promise, etc... this will be the first time that OPNFV projects have been able/encouraged to take an umbrella view of the needs of OPNFV at a high-level in the context of OpenStack capabilities 13:36:13 #info Afternoon will be 2 parallel sessions, 1h in length 13:36:17 Ok, so Copper, HA, MultiSite, Doctor, VNFFG.SFC, IPv6, Doctor, Promise 13:36:21 #info We have 8 session spots available 13:36:23 That gets us to 8 total 13:36:26 Ha is in teleconf now 13:36:48 They are talking meet in Vancouver 13:36:50 All the projects that depend upon configuration or cloud state awareness for policy-related management 13:36:55 Oh, are you saying we’re now camping out on HA’s IRC meetings? 13:37:04 #info Ideas for session topics are: Copper, HA, Multisite, Doctor, VNFFG/SFC, IPv6, Doctor, Promise, installer/stack cage match 13:37:25 bryan_att, are there addtional projects you conceive in that? 13:37:38 please do not do installer 13:37:43 #info We need to identify topic leaders for each of these, ensure they are in Vancouver and prepared to lead a session, before we schedule 13:37:49 What I mean is that I don't think it will be helpful to have a session per project 13:37:54 that could take a week all on its own 13:38:09 we need sessions that address the project types and harmonize what they are trying to get from OpenStack 13:38:26 bryan_att, So your suggestion is to have "Policy-driven projects" as a topic? 13:38:30 all the policy-related projects are very similar in this. we should address them all in one session 13:38:32 dneary, I can help identify/chase topic leaders 13:38:49 I guess fu qiao will apply for ha slot later after teleconf 13:38:54 agree with bryan_att 13:38:55 rpaik, I can help with that too - I've been chasing down several project leaders for blueprints 13:39:16 dneary, great let me know if I can help 13:39:22 hui, Would HA and multisite be close enough to share a session? 13:39:29 yes, so we don't have to deep-dive but can pull out the common thread of OpenStack dependency and especially highlight the cross-VIM dependencies we will be addressing in OLD, Contrail, ONOS, etc 13:39:35 rpaik, Perhaps I can suggest leaders, and you can chase them down? 13:39:41 instead of calling the installer/distro topic a debate we should just present an overview of the status of each "experiment" 13:39:43 * dneary is choc-a-block this week 13:39:59 They need to talk, I can't answer for them 13:40:12 this should be of first value to the OPNFV community to help us harmonize on broad themes of NFVI functionality 13:40:22 dneary, sounds like a plan 13:40:25 HA/ Multisite not close enough to share space 13:40:36 in the context of OpenStack as a key VIM 13:40:52 iben_, There is a fundamental issue there of how we scale the project for multiple deployment platforms, but I agree we're not going to converge or even get something useful in an hour 13:41:25 i can collect the infro and present it 13:41:35 I would suggest we stay as close to openstack related topics as possible. When in Rome... 13:41:37 #action Dave to suggest topic leaders for the topic suggestions to Ray, Ray to chase them down and request commitment to lead sessions 13:41:50 let’s get confirmation from project leads and start slotting sessions into timeslots 13:42:11 Once we have topics and leaders nailed down, we will want to promote these topics to the right OpenStack people 13:42:15 RayNugent: +1 13:42:35 So, policy related projects will need to reach out to Congress, Heat, Ceilometer project participants 13:42:40 that's why the installer topic should be mentioned too 13:42:45 we probably need at least one break in the afternoon and a wrap-up session (5 mins) may also be a good idea 13:42:47 Otherwise we miss the boat on the collaboration opportunity 13:42:50 I'll start an OPNFV thread with the policy-related projects on a common session theme 13:43:14 rpaik, afternoon break is 30 mins 13:43:17 rpaik, I don't think you need breaks 13:43:19 rpaik, yes 13:43:21 dneary, yes and other projects likely too - there are several ways/components via which policy can be implemented 13:43:33 We need one afternoon break 13:43:45 cuz cookies 13:43:57 HKirksey, So, 45' session before & after break, 1h session after lunch and from 5'6 13:44:02 5-6 13:44:38 Hmm... makes scheduling a bit trickier 13:45:08 we’re ~15 minutes over…. I think we have some consensus for the afternoon 13:45:14 lets keep all sessions at 1 hour 13:45:15 We should make sure session starts & finishes align with summit talks if possible - makes it easier for people to attend talks 13:45:15 Lunch is 12:45-2. 13:45:37 dneary +1 13:45:40 rpaik, That is a detail, can be taken care of in email, I'm sure 13:45:56 So 2-3, 3:-3:30 break, 3:30-4:30, and then wrap up or anything else that we determine we want to do coming out of the afternoon sessions? 13:45:57 rpaik, Agree with you, we're close to aligned on the afternoon, back to the morning? 13:46:07 we just need to agree on the morning 13:46:27 HKirksey, We're potentially going from 8 to 6 afternoon sessions with that plan 13:46:48 HKirksey, Which I'm OK with, if we agree it's the way to go, but it does change things for session placement & choice 13:47:29 2-3 Sesisoon 1, 3-3:30 break, 3:30-4:30 Session 2, 4:30-5:30 Session 3, 5:30-6 wrap up 13:47:29 do we need an hour slot for IPv6? 13:47:32 We have the run till 6 13:47:39 rpaik, I have a preference in the morning for 3x30' sessions, I don't think that you'll get a lot of questions, and I see it more as "setting the tone/framing the afternoon discussions" 13:47:39 We have the room til 6 13:47:41 So it still works 13:48:23 So morning could be, 30 minutes high level overview, 30 minutes project overview, 30 minutes set the tone for the rest of the week (the session the telco WG asked for) 13:48:29 I would do 30' OPNFV Arno overview, 30' Telco WG use-cases & opportunities for collaboration during the week, 30' lightning talks (5' overview of each OPNFV project) 13:48:40 30 min project oview goes into a bit more detail about the projects that touch openstack to give them insight 13:48:48 HKirksey, Great minds think alike :-) 13:48:51 :) 13:49:27 RayNugent, Yes, I think we need an hour for IPv6 13:49:34 Lightning talks are not non-native speaker friendly 13:49:35 I think that flows well for the morning 13:50:00 RayNugent, Especially if we get Neutron and Open vSwitch people in the room. 13:50:09 RayNugent, Good talk 13:50:31 RayNugent, Might be better to do what HKirksey proposed and have one person present all the requirements projects 13:50:47 s/"Good talk"/"Good point"/ 13:50:52 yes, agree 13:51:06 Sounds like we have general agreement 13:51:10 +1 on one speaker. reduced time between speakers 13:51:18 I am not sure dpacc will meet as well 13:51:53 dpacc probably not a great Openstack topic 13:51:53 Dpacc generate lots email discussion in the list 13:51:56 I don’t think DPACC has much synergy with OpenStack 13:52:15 #info morning: 30 minutes high level overview, 30 minutes project overview, 30 minutes set the tone for the rest of the week (the session the telco WG asked for) 13:52:18 #agreed Morning format will be 3x30' sessions, an overview of OPNFV (positioning wrt OpenStack and industry drivers), an overview of OPNFV requirements projects, and an invited speaker from the Telco Wg to present use-cases and opportunities for collaboration during the week 13:52:24 :-) 13:52:35 Well done, well done 13:52:41 They are discussing openstack gap and bp, they have 13:52:47 I can update the Etherpad with what we discussed 13:53:05 HKirksey, RayNugent: Agreed, although there will be some Nova & Scheduler/Congress impact later on. 13:53:54 #agreed Afternoon will be 2 tracks, with 3x1h sessions, 30' breaks between them, and a 30' "wrap up the day" where we get everyone in 1 room to summarise achievements 13:54:27 yes but hardware acceleration is going to be a new concept to stackers, maybe save it for another venue after we've introduced ourselves? 13:54:33 #agreed 6 slots for working group topics, topics to be determined and scheduled based on availaility of project leaders 13:54:34 Ah ok. Should we see if DPACC wants to join the party too? 13:54:57 HKirksey, There will certainly be some DPDK and ODP people present 13:55:04 I will let dpacc know 13:55:12 HKirksey, I still think it shouldn't be the focus of a session 13:55:33 I agree with RayNugent on that - most of the dpacc work is not OpenStack related right now 13:55:34 They will have teleconf in 8hours 13:56:02 Seems like we have marching orders 13:56:05 #action rpaik to update the Etherpad 13:56:12 are we good? 13:56:28 HKirksey and I have another call in 4 minutes :-( 13:57:08 I think we're good 13:57:11 I would prefer dpacc to answer it 13:57:18 seems like we covered everything. Now the details... 13:57:28 thanks everyone 13:57:39 #endmeeting