15:00:48 <mtahhan> #startmeeting VSPERF weekly meeting 15:00:48 <collabot> Meeting started Wed Apr 29 15:00:48 2015 UTC. The chair is mtahhan. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:48 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic. 15:00:48 <collabot> The meeting name has been set to 'vsperf_weekly_meeting' 15:05:03 <mtahhan> let's start with documenting the attendees 15:05:09 <mtahhan> #info Maryam Tahhan 15:05:29 <billyom> hi 15:05:34 <acm_> #info Al Morton 15:05:43 <billyom> #info 15:06:12 <mtahhan> sorry about the last minute change... 15:06:12 <mtahhan> had a bit of an issue with the company goto meeting account 15:06:28 <mtahhan> but hopefully normal services will resume next week 15:06:28 <acm_> np 15:07:22 <mtahhan> #info let's start with an open call for agenda items. is there any opens that people would like to discuss today? 15:07:22 <edube> #info Eric Dubé 15:08:01 <mtahhan> I have a few for today 15:08:16 <mtahhan> #info IETF - draft for the LTD 15:08:24 <mtahhan> #info TOIT update 15:09:01 <mtahhan> #info Setting up a backlog of items for discussion on this weekly call 15:09:18 <mtahhan> does that sound/read ok for eveyone? 15:09:29 <acm_> ok go 15:10:21 <billyom> sounds good. Some LTD reviews going on at the moment so maybe people could keep an eye on gerrit and add some comments on them 15:10:39 <mtahhan> ok we can add that as an agenda item also 15:10:50 <mtahhan> #info LTD open testcases reviews 15:11:07 <mtahhan> #topic IETF - draft submission of LTD 15:11:35 <mtahhan> I think this is a great idea and I had a few open questions about this from last week 15:11:56 <mtahhan> #info what's the best way to start this and how do we work together on this? 15:12:09 <acm_> I have an idea 15:12:25 <mtahhan> sure fire ahead 15:13:07 <acm_> we should first agree on the goal of the draft, I suggest to first summarize... 15:13:46 <acm_> how we have incorporated and expanded the existing RFCs to cover vSwitch 15:14:49 <acm_> what do folks think about that? 15:16:05 <mtahhan> That sounds like a good start. 15:16:05 <mtahhan> I suppose now's a good time to pose the question, do we think the LTD is covering enough of the existing RFCs to allow us to do that, or is the idea to keep this as a working copy that we upgrade in parallel to the LTD? 15:17:32 <acm_> yes, I think we upgrade/expand in parallel 15:18:04 <billyom> I'd suggest a summary once we get a reasonably stable & complete LTD. 15:18:15 <acm_> but eventually we could follow with very specific method drafts 15:19:08 <billyom> We can bring in further items from the evolving LTD later based on feedback from IETF or if a the LTD has evolved sufficiently far to warrant an upgrade from the LTD 15:19:35 <billyom> Basicallly you branch from the LTD when it feels right to do so. 15:20:14 <billyom> It would cause a lot of churn to be constantly adding small new pieces to the IETF forum 15:20:36 <acm_> I underastand. Since we've been moving sections to a "stable" state, I would suggest to summarize those at first. 15:20:49 <acm_> Churn is ok at IETF 15:20:59 <billyom> Sounds good when you type it! 15:21:20 <acm_> This is about communication as much as stability in direction. 15:22:08 <acm_> We want others to know about our progress in vSwitch - step 1 perhaps we avoid some overlap that way... 15:23:31 <mtahhan> I think it's something we're going to have to do anyway, and doing it now when our base is small is a lot less difficult then leaving it till later when the LTD is massive. 15:23:31 <mtahhan> plus we are agreed on general methodologies and the overall RFCs... and like you say it's a great way of communicating out on what we are doing 15:23:31 <mtahhan> I'd be happy to get the ball rolling on this. 15:24:11 <acm_> our work in OPNFV is among the first attempts to apply IETF BMWG RFCs to the NFVI 15:25:35 <acm_> agree 15:25:57 <acm_> as for "how" we do this... 15:26:34 <acm_> There are lost of tools to prepare Internet Drafts s/lost/lots/ 15:27:39 <acm_> If we keep the draft simple, there is a wayto do it using MarkDown. new area for me. 15:28:51 <acm_> I've been using XML, specifically RFC2XML and a slightly cranky but almost WYSIWYG editior. 15:29:12 <mtahhan> I don't mind the format... as long as we store in git 15:29:13 <acm_> but MD would work well with our OPNFV tool set 15:30:54 <acm_> sorry, that's XML2RFC above - I have a new laptop/keyboard affecting thinking... 15:31:49 <mtahhan> XML is fine (I think unless anyone has another opinion about this), my only concern is that we have a history of the doc and it's shareable... 15:31:49 <mtahhan> ok so the real question now is: are there any volunteers for this activity? 15:31:51 <acm_> in any case, I think the MD converts to XML2RFC and to .txt eventualy 15:32:10 <acm_> yes, obviously I'm in. 15:32:25 <mtahhan> cool 15:32:25 <mtahhan> I am also in 15:33:12 <billyom> I'll help out. 15:34:45 <acm_> that's proabbly enough to make it interesting... 15:36:07 <mtahhan> sounds good, Ok so we can start drafting something (maybe even a base/outline) and push a base to gerrit and move from there. 15:36:16 <acm_> So, I'll investigate how *simple* the MD path is to follow 15:36:27 <mtahhan> that would be great 15:36:52 <acm_> we can always take what we've got and move up to XML instead, I think 15:37:04 <acm_> that's something I'll investigate too. 15:37:07 <dneary> Hi everyone. 15:38:13 <dneary> The DPACC project is looking for space in the OpenStack Summit OPNFV day - one question that has come up is how much impact there will be, short-term, on OpenStack. If there were a dpacc session, would it make sense to have vsperf participants there too? 15:38:39 <dneary> (sorry for changing the topic mid-meeting, I can't tell where it would fit in the agenda) 15:40:59 <mtahhan> I think it would make sense for vsperf participants to be at a dpacc session, esp for phase one 15:41:10 <mtahhan> there could be some interesting intersections 15:42:55 <acm_> it might come down to logistics - I'm booked elsewhere, sadly. 15:43:15 <mtahhan> I'm also away at the time 15:44:04 <mtahhan> But Billy might be covering for me there. 15:44:37 <billyom> Cool! 15:45:46 <mtahhan> I'll add something in the minutes that go out to see how many vsperf folks will be around 15:46:13 <acm_> good, best we can do 15:48:35 <mtahhan> ok so Al, if you have a look at the MD open and the ease of conversion to XML, we can create a simple/base doc (next week or so) and work from there. 15:49:01 <mtahhan> does anyone have any comments/concerns? 15:49:05 <acm_> I'm on it, 15:50:55 <acm_> so, TOIT update? 15:50:55 <mtahhan> Ok cool 15:50:55 <mtahhan> so I think the next topic we should hit is the agenda backlog for the weekly meeting. 15:50:55 <mtahhan> I think we should start putting together a list of items/topics relevant to vsperf we'd like to discuss at the weekly meeting 15:51:08 <mtahhan> :) I can do the TOIT update also 15:51:24 <acm_> go with back log 15:51:24 <mtahhan> OK let's do toit and go back to agenda backlog 15:51:42 <mtahhan> on TOIT 15:51:42 <mtahhan> at the moment we have 3 developement streams 15:51:43 <billyom> http://pandoc.org/index.html is well worth a look when it comes to converting between many of these markdown et al languages. 15:52:13 <mtahhan> 1st stream is porting the exisiting tests we had into the agreed on dir structure 15:52:21 <acm_> Yeah, that's what they are using https://github.com/miekg/pandoc2rfc/ 15:52:47 <billyom> cool 15:53:05 <mtahhan> this is going well, we have all the low level components (vswitches/trafficgens/vnfs) bar one moved and old toit still functioning 15:53:44 <acm_> nice, functioning is nice 15:54:50 <mtahhan> the 2nd stream we have is implementing a stubbed out skeleton framework into which the low level components can be plugged in. 15:54:50 <mtahhan> we;re making good progress on this front also. 15:54:50 <mtahhan> and started plugging in vswitches under the hood this week 15:55:25 <mtahhan> the 3rd stream which is more of a path finding stream, is taking the exisiting code and porting it from python 2.x to python 3 15:55:57 <mtahhan> the methodologies and lessons learned from here can be used to port the skeleton code to python 3 also... 15:56:40 <mtahhan> so that's where we are with the framework at the moment, we're on track to making our first code drop in 4 weeks time 15:57:10 <acm_> all good progress, thanks! 15:57:24 <mtahhan> np 15:58:29 <mtahhan> so I don't think we'll hit the agenda topic and finish it in the next 3 mins. 15:58:29 <mtahhan> but I would ask for people to have a think about what discussion topics they would like to cover in this forum over the next few weeks 15:58:51 <mtahhan> for example Gurpreet might do a readout on vswitch performance for us 15:59:59 <mtahhan> does this sound OK for people? 16:00:33 <acm_> yes, real stuff always worth hearing about 16:02:06 <mtahhan> cool, alright let's adjourn for today and I will send out the mins 16:02:12 <mtahhan> thanks everyone for attending 16:02:20 <mtahhan> and sorry about the goto issue. 16:02:23 <mtahhan> #endmeeting