15:00:59 <asn> #startmeeting SponsorR
15:00:59 <MeetBot> Meeting started Tue Mar 31 15:00:59 2015 UTC.  The chair is asn. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:59 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:01:02 <asn> hello all
15:01:05 <dgoulet> ohmygodel: o/
15:01:10 <karsten> hi
15:01:35 <asn> ok let's get started
15:01:45 <isabela> hi
15:01:48 <asn> who wants to start with a briefing, while I organize my stuff.
15:02:07 <asn> oh well
15:02:08 <asn> i will go
15:02:10 <asn> hello syverson
15:02:11 <Yawning> "eeeeeeeeee HS DoS eeeeeeeeeeee"?
15:02:14 <asn> so during past week
15:02:31 <asn> i voted on the top-down spreadsheet
15:02:36 <asn> discussed it a bit more with people
15:02:53 <asn> did a tor-dev post about *bridge stats* (not realted to HS stats but whatever): https://lists.torproject.org/pipermail/tor-dev/2015-March/008541.html
15:03:02 <asn> we published the crowdfunding post
15:03:09 <asn> and now i'm working on #15463
15:03:17 <asn> not entirely sponsorr material for some of them, but well...
15:03:29 <asn> #15463 is our new HS annoyance.
15:03:36 <ohmygodel> very excited about the crowfunding !
15:03:40 <dgoulet> we have the first HS bi-weekly meeting tomorrow, that's a GREAT subject for it :)
15:03:41 <asn> stupid but effective DDoS without easy solution.
15:03:44 <asn> dgoulet: yes.
15:03:49 <asn> dgoulet: well timed indeed.
15:03:56 <asn> anyway, that's that from me.
15:04:01 <asn> who wants to go next?
15:04:11 * dgoulet can go
15:04:24 <asn> go
15:04:41 <Yawning> (the hs meeting is at the same time the pt meeting is)
15:05:15 <dgoulet> did a full timeline of an HS descriptor over 24 hours (validity time) with client/directory/service interactions, I'm almost done with the text part that had already comment on it for which we found stuff to fix
15:05:19 <asn> Yawning: fuck.
15:05:42 <asn> dgoulet: can you talk a bit more abou tthis timeline thing.
15:05:56 <asn> i still haven't had time to look at it :/
15:05:58 <dgoulet> worked on #15463 also lately, too much tor less R for me unfortunately, at least some progress on the HS improvement part, patch will come out of this
15:05:59 * dgoulet done
15:06:02 <dgoulet> asn: np
15:07:21 <dgoulet> next?
15:07:22 <asn> ok so no more about this timeling thing.
15:07:22 <Yawning> (is +HSPOST R?)
15:07:23 <asn> next?
15:07:30 * ohmygodel emailed an update already
15:07:34 <asn> Yawning: feel free to say any updates you have about it
15:07:39 <dgoulet> Yawning: not really
15:07:42 <asn> ohmygodel: it was about your SMC exploration
15:07:47 <ohmygodel> summary: i found some SMC code
15:07:49 <ohmygodel> yeah
15:08:10 <asn> ok
15:08:20 <Yawning> asn: I reviewed it?  And need to again sometime, but everything is on fire?
15:08:44 <asn> Yawning: ack
15:08:47 <asn> ok who next?
15:08:52 <Yawning> ohmygodel: is that the thing you mentiond "requires an MPI implementation"?
15:09:34 <ohmygodel> yawning: yes
15:10:01 <ohmygodel> (only for the StatAuths, relays have completely trivial stats reporting)
15:10:10 <Yawning> can we rip MPI out?
15:10:26 <ohmygodel> yawning: probably yes. i assume its because they ran the StatAuth over a cluster
15:10:31 <dgoulet> maybe discussion after status report ? :)
15:10:44 <Yawning> (sorry)
15:11:03 <asn> ok. who next?
15:11:24 * karsten doesn't have much
15:11:37 <karsten> voted on the top-down thing, done. (sorry.)
15:11:48 <asn> ack no worries
15:11:55 <asn> next? syverson ?
15:11:57 * syverson was mostly away and/or working on other things. I did enter my vote on the top-down thing, but just an hour ago.
15:12:02 <asn> ok
15:12:03 <asn> great
15:12:06 <asn> that makes sense i guess
15:12:09 <asn> let's move to discussion
15:12:40 <Yawning> so, yeah, MPI, I've written a ton of MPI code in the past
15:12:49 <asn> topics that you would like to see discussed?
15:12:52 <Yawning> so if we need to rip it out I probably could
15:12:52 <asn> ok let's start with SMC.
15:13:01 <Yawning> but I haven't looked at said code yet
15:13:21 <kernelcorn> asn: different topic, but I think I'll take your advice and sent out an early version of my paper to tor-assistants very soon
15:13:26 <isabela> asn: topic -> my summary about the voting at top-down list
15:13:31 <asn> here is the code: https://github.com/benkreuter/secure-computation-uva
15:13:45 <asn> kernelcorn: hah tor-assistants. that's a private list. but ok.
15:13:52 <asn> isabela: yes
15:13:59 <asn> isabela: top-down discussion is queued up after that.
15:14:03 <isabela> ok
15:14:22 <kernelcorn> asn: well I've posted to it before, I'll take another look
15:14:25 <asn> this code has a compiler/ dir.
15:14:35 <asn> kernelcorn: you can post there, but only tor devs will be able to read it.
15:14:46 <asn> kernelcorn: i will be able to read it if you post there.
15:14:54 <kernelcorn> that's fine, I just need some feedback like you suggested
15:14:56 <ohmygodel> ok all sorry i have to go now
15:15:03 <asn> ohmygodel: alright
15:15:06 <dgoulet> ohmygodel: o/
15:15:07 <ohmygodel> i think the SMC stuff is still early evaluation
15:15:07 <asn> ohmygodel: anything you want to talk about?
15:15:08 <asn> before leaving?
15:15:13 <asn> ohmygodel: ack
15:15:21 <ohmygodel> it might be a good idea
15:15:35 <ohmygodel> to get some thoughts from you all
15:15:51 <ohmygodel> about if it seems like a reasonable approach
15:15:54 <asn> ack
15:16:02 <Yawning> (after the top down thing, we should talk about stats and our current fire)
15:16:10 <ohmygodel> ok adios
15:16:18 <asn> maybe we can talk about the current fire before the top-down?
15:16:22 <asn> in case we can add stuff to the top-down?
15:16:24 <kernelcorn> what's on fire?
15:16:29 <asn> kernelcorn: #15463
15:16:42 <asn> kernelcorn: HS DoS through circuit creation.
15:16:56 <kernelcorn> oh wow
15:17:03 <asn> 14:07 < armadev> it would be nice if we had network wide statistics about how many intro cells we saw
15:17:06 <asn> 14:07 < armadev> and how many rend points were established
15:17:09 <asn> 14:07 < armadev> all those things i wanted a few months ago and everybody was like "what good would that be" ;)
15:17:42 <asn> that would indeed show the scale of the attack.
15:18:05 <asn> ok let's talk about this attack for the next 15 minutes (till 15:30) in case we can find of any nice statistics to collect
15:18:10 <asn> after that, let's move to top-down.
15:18:23 <karsten> so, intro cells is hard to learn safely, right?
15:18:33 <karsten> and established rend points is mostly harmless.
15:18:44 <asn> that's also my impression yes
15:19:10 <karsten> we could totally work on est rend points and have stats in 2 months from now.
15:19:16 <karsten> is that useful? ..
15:19:27 <qwerty1> seems like overkill for this
15:19:28 <asn> maybe.
15:19:39 <asn> for this specific attack, it would show us how much the attackers are trying.
15:19:48 <asn> theoretically, they could not even estalbish rend points though
15:20:07 <Yawning> but they are I think
15:20:11 <asn> yes
15:20:12 <Yawning> because we see successes?
15:20:29 <asn> just saying that if we do this stat just to watch this attacker, then maybe they could adapt in the future.
15:20:38 <qwerty1> indeed
15:20:47 <Yawning> it would be useful to have the sort of stats that highlight that someone is doing the smarter variant
15:20:48 <asn> another useful stat would be something that tells us what we should set MAX_REND_FAILURES to
15:20:53 <Yawning> if/when that happens
15:21:01 <asn> Yawning: that would be number of INTRODUCE1 cells.
15:21:11 <Yawning> (or we can try to kill the smarter variants with fire)
15:21:28 <asn> but number of INTRODUCE1 cells will leak the popularity of HSes.
15:21:36 <asn> so we've been avoiding it so far.
15:21:36 <dgoulet> hrm stats of failures could be very useful to us, no idea on the privacy side :S
15:21:54 <karsten> what's MAX_REND_FAILURES?
15:22:00 <Yawning> 8
15:22:06 <asn> karsten: it's how many relaunches
15:22:07 <karsten> I mean, what is it used for?
15:22:12 <kernelcorn> is there any indication that the DOS attack is being used in conjunction with traffic analysis and observations for that traffic flood?
15:22:15 <asn> karsten: the hidden service will do, if it's first rendezvous circuit fails
15:22:20 <Yawning> number of times you relaunch when you fail to connect to the RP
15:22:28 <karsten> 8 times? oh wow.
15:22:34 <asn> yeah it's loco. it used to be 30.
15:22:36 <asn> lol.
15:22:48 <dgoulet> 30 was crazy yolo :)
15:22:58 <asn> kernelcorn: no.
15:23:05 <kernelcorn> ty
15:23:06 <karsten> isn't that something we could measure on a HS?
15:23:09 <Yawning> kernelcorn: no, this is a terribad way to mount this sort of thing
15:23:14 <asn> i think 1 relaunch or 0 relaunch is fine. and we should make the client retry.
15:23:30 <Yawning> fairly sure the client does retry
15:23:31 <kernelcorn> Yawning: all right
15:24:13 <asn> if the client retry logic is sensible, I'm fine with 0 relaunches too... but need more thinking.
15:24:13 <Yawning> kernelcorn: could be adapted to me a traffic confirmation thing, but they're randomizing the RP in each intro cell, which makes that unlikely
15:24:14 <dgoulet> FYI, if the circuit times out, only one relaunch is done
15:24:17 <syverson> asn: INTRODUCE1 need not leak popularity (which I continue to not see as a problem anyway, but that's another matter) if these are system aggregated numbers only.
15:24:34 <asn> syverson: you mean if we had a stats aggregation thing? i agree.
15:24:43 <syverson> asn: yes
15:24:48 <asn> yes i agree
15:25:33 <dgoulet> asn:  I would also think 1 is enough fwiw
15:25:41 <asn> ack
15:25:46 <asn> i will try to look into this tomorrow
15:25:50 <asn> or later today
15:26:24 <dgoulet> so everything trends toward "We need a stats aggregation system" ...
15:26:47 <asn> well, i would prefer to just plug the attack entirely.
15:26:53 <asn> then i wouldn't care if someone is doing it ;)
15:27:05 <dgoulet> oh well I meant for the stats part of detecting these attacks
15:27:06 <karsten> it seems stats are not the answer for this attack.
15:27:13 <asn> karsten: heh
15:27:29 <Yawning> yeah
15:27:34 <asn> ok
15:27:38 <asn> topic exhausted.
15:27:40 <Yawning> just thought it was worth bringin gup
15:27:47 <asn> let's move to top-down list
15:27:55 <asn> the spreadsheet is here https://docs.google.com/spreadsheets/d/1mY8wax7FBUIAPAmDLpGcEtYCYCcsxW21KtLCntRbSk0/edit?pli=1#gid=0
15:28:00 <asn> and most actors have voted
15:28:15 <asn> interesting patterns merge from the votes
15:28:22 <Yawning> (was I supposed to?  no right?)
15:28:25 <isabela> yes
15:28:36 <asn> Yawning: you can if you want!
15:28:37 <isabela> (I mean yes to asn comment)
15:28:40 <Yawning> oh ok
15:28:51 <asn> in any case, it seems that most people are conservative here, and I like this.
15:29:19 <asn> the green rows are mainly things that don't require tor modifications at all.
15:29:41 <asn> and we can just collect from testnets, or from looking at historical consensus documents
15:29:42 <karsten> which also means that we can work on them more quickly.
15:29:47 <asn> yes
15:29:52 <asn> personally, I like this.
15:29:55 <karsten> yep
15:29:56 <dgoulet> row 4 and 5 are being worked on by me
15:30:09 <nickm> rehi
15:30:11 <asn> great
15:30:14 <dgoulet> row 3, DonnchaC_ has stuff on it
15:30:15 <kernelcorn> Yawning: is high memory usage associated with this attack?
15:30:16 <nickm> meeting still hppening?
15:30:24 <asn> dgoulet: exactly
15:30:35 <asn> kernelcorn: we don't know. CPU is the main issue.
15:30:48 <dgoulet> row 14, naif has a system for that
15:30:52 <dgoulet> nickm: yes
15:30:53 <asn> I can also take over row 3 from DonnchaC_, if we decide to work on that.
15:31:18 <asn> I'm very interested in working more on *stresstesting* HSes.
15:31:19 <dgoulet> row 18 also I have an HS that I use for profiling, that should be something tpo could host though
15:31:31 <kernelcorn> all right, I have a Tor instance that's consuming 1.48 GB of RAM, I guess it's unrelated
15:31:52 <asn> isabela:
15:31:53 <asn> isabela: hello
15:31:56 <dgoulet> asn: tbh, a tor network with rate limiting links would be AMAZING for that :)
15:31:59 <asn> isabela: would you mind telling us a bit about your email?
15:32:07 <asn> isabela: i think you mainly told us which rows are green, right?
15:32:57 <asn> dgoulet: wow, that seems hard.
15:33:25 <dgoulet> asn: I actually have what you need for that! I just need time to set it up
15:33:27 <asn> dgoulet: i have no idea how you would do that the right way. that's the sort of stuff that ns/opnet etc. does, right?
15:33:28 <Yawning> dgoulet: http://info.iet.unipi.it/~luigi/dummynet/
15:33:33 <Yawning> ^_^
15:33:43 <dgoulet> asn: https://github.com/nsec/the-internet
15:33:44 <isabela> now we have more votes there - but when I looked at it there were 2 questions where all the votes were negative and which I was wondering if we should drop or think of other approaches
15:33:59 <asn> isabela: fine with dropping them.
15:34:08 <asn> isabela: the big question here, I think
15:34:16 <asn> isabela: is from the rows that are green, which ones would the sponsor appreciate?
15:34:49 <dgoulet> top 3 rows are definitely something the sponsor will like imo
15:34:51 <asn> and should we care about this? or should we go full greedy tactic (do the best for Tor) and hope that it works for the best.
15:34:58 <asn> dgoulet: ok
15:35:04 <karsten> how much time is left?
15:35:13 <asn> karsten: till july I guess :)
15:35:18 <dgoulet> yeah July
15:35:20 <karsten> ah, not april?
15:35:27 <asn> karsten: april long gone man
15:35:28 <dgoulet> karsten: well quaterly meeting in 2 weeks
15:35:40 <dgoulet> but the real juice of our work should be presented in July
15:35:53 <dgoulet> full month of meeting and decision will be taken on our funding ;)
15:36:17 <karsten> okay, great. plenty of time.
15:36:39 <isabela> 14 is something they would like, no?
15:36:43 <isabela> row 14
15:36:59 <dgoulet> isabela: not entirely sure because they do have crawlers :)
15:37:20 <isabela> hehe
15:37:31 * isabela will hide the rows that are all red (just for now)
15:37:38 <asn> isabela: ok
15:37:48 <dgoulet> isabela: that sponsor developed intelligent crawlers that do not abuse the Tor network and use them for fingerprinting bad websites
15:38:44 <syverson> I don't know if this is the time/place for this. But I have yet to have someone tell me the harm from leaking that the most popular HSes are roughly popular after a delay. I think the assumption that this is bad accounts for a lot of the "no safe way to do this" entries.
15:40:20 <karsten> isabela: are you hiding rows that are all red or rows that have red in them?
15:40:33 <asn> let me pick three rows that we could do for July and you tell me if you enjoy:
15:40:42 <asn> * HSDir health stuff that dgoulet is working on
15:40:53 <asn> * IP lifetime stuff that donncha started work and I could take over or karsten
15:40:59 <karsten> yep
15:41:10 <asn> * Preemptive circuit building and stresstesting hidden services
15:41:17 <asn> that I could do or dgoulet or karsten
15:41:28 <dgoulet> asn: for the "IP lifetime stuff", you could also take on 23 easily also
15:41:45 <isabela> karsten: I hided some that had majority red as well
15:41:59 <asn> dgoulet: 23 is hsdir hash ringstuff?
15:42:00 <karsten> if I choose, I'd prefer crunching numbers/descriptors over setting up things.
15:42:04 <karsten> if I may*
15:42:11 <isabela> karsten: just trying to have a perspective of how many are actually something to consider (seems 6 to 9 now)
15:42:19 <asn> karsten: yes that's also my preference :P
15:42:24 <dgoulet> asn: yeah sounds like could be easy to do and useful to detect maybe attacks?
15:42:27 <karsten> asn: ah :)
15:42:31 <asn> karsten: in that case, I think #15513 might be nice for you?
15:42:43 <asn> karsten: it has potential for nice visuals
15:42:53 <karsten> asn: sounds great
15:42:55 <asn> karsten: and also the fact that donncha is checking out actual hidden services might excite the sponsor.
15:42:58 <dgoulet> ah yeah that intro point algorithm is still a mistery to me ...
15:43:05 <asn> dgoulet: oh my...
15:43:27 <dgoulet> would be useful to me to have a comprehensive analysis of it for the HS Descriptor timeline and reachability also
15:43:34 <isabela> (people can unhide things if they want)
15:43:40 <karsten> isabela: sounds good. I wasn't sure about one row, and then it went away before I could check.
15:44:05 <asn> syverson: i'm also not sure if this is the time/place for this :)
15:45:34 <syverson> asn: OK sure.
15:46:20 <asn> k so what should we do with the top-down thing now?
15:46:39 <asn> should we look at the green stuff and see what excites us over the next week?
15:46:40 <asn> isabela: ?
15:46:53 <isabela> yes
15:47:29 <isabela> another question is how many can be done giving the time
15:48:06 <asn> i will need to dig a bit into each problem to understand how deep it is.
15:48:16 <karsten> so, next week is pretty full here. (it's easter..) can we make plans for the next two weeks?
15:49:18 <asn> isabela: ^ ?
15:49:20 <isabela> karsten: related to the top-down list?
15:49:39 <karsten> 15:46:58 < asn> should we look at the green stuff and see what excites us over  the next week?
15:49:46 <syverson> karsten: following week is during Sponsor R PI meeting.
15:49:47 <karsten> related to that.
15:49:51 <karsten> oh, ok.
15:50:12 <karsten> how about next friday?
15:50:26 <asn> karsten: will you be away from IRC over the next two weeks?
15:50:35 <isabela> karsten: 10?
15:50:37 <karsten> somewhat, yes.
15:50:47 <karsten> yep, 10th.
15:50:55 <dgoulet> I,m ok with 10
15:50:56 <asn> would you like to check out #15513 and tell me what you think about it?  if you do it in a week that's good. if you can't do it that fast, that's ok too?
15:50:56 <karsten> sorry if I'm interrupting here, btw.
15:51:15 <karsten> #15513 doesn't look too hard.
15:51:20 <asn> yes it looks quite easy t bh
15:51:27 <karsten> I can work on that and try to get something done by tue, more by fri.
15:51:33 <asn> yes
15:51:34 <asn> like check it out
15:51:41 <asn> and see if it has lots of depth
15:51:46 <karsten> sure.
15:51:49 <asn> if not, we can do it fast and move to another one.
15:51:55 <karsten> right.
15:51:56 <asn> if it has depth, just report about it over the next two weeks
15:51:56 <dgoulet> is SMC stuff should be considered for July? (not finished but being worked on?)
15:52:05 <dgoulet> aka row 13 I just unhide
15:52:48 <asn> i have no idea about SMC. someone of us should look at the code...
15:53:06 <dgoulet> well I just mean the "concept" in Tor
15:53:25 <nickm> I was hoping that we'd have a list of SMC implementations to review for QOI.
15:53:31 <nickm> And for usability
15:53:32 <dgoulet> the concept of stats aggregation system, we've been talking about that since December but no decision on should we work towards one :S
15:53:33 <asn> nickm: we have
15:53:38 <nickm> great; where's the list? :)
15:53:39 <dgoulet> nickm: aaron email
15:53:42 <asn> nickm: [1] https://github.com/benkreuter/secure-computation-uva
15:53:42 <asn> [2] https://github.com/cryptouva/pcf
15:53:49 <nickm> ok.
15:53:51 <asn> nickm: aaron sent "Subject: Weekly update in lieu of attending Sponsor R meeting"
15:54:26 <asn> one of them is in C++ and has a compiler/ directory.
15:54:31 <asn> who knows.
15:54:41 <dgoulet> and my guess is that the Sponsor would like to know in april if we are working towards something like that
15:54:44 <dgoulet> syverson: am I wrong?
15:55:05 <asn> ok
15:55:09 <syverson> dgoulet: if we are working towards using SMC?
15:55:23 <dgoulet> syverson: well a "stats aggregation system" in Tor (tbd)
15:55:48 <nickm> Aggregation through some means or other is something IMO that we should do.  SMC or not is TBD IMO. I think whether SMC mostly depends on the QOI and usability of the SMC implementations.
15:56:06 <nickm> If not, then we need to return to some of the older proposals, which will limit what we can aggregate.
15:56:19 <nickm> I think that's where we stand; am I right?
15:56:23 <asn> yes
15:56:39 <syverson> I think giving them an idea of what we are planning to work on and why in general is good, but we don't have to give them everything.
15:57:12 <dgoulet> syverson: right not everything just what we have/planned to do
15:57:19 <syverson> Sorry, I didn't mean to say "SMC". I was using that as a stand-in for somehow doing stats aggregation.
15:57:28 <asn> ok i can take a look at the SMC implementations over next week, or nickm you can do it if you prefer?
15:57:51 <nickm> let's do it together
15:57:56 <nickm> and compare notes
15:57:59 <asn> ok
15:58:07 <syverson> dgoulet: yes.
15:58:23 <asn> ok
15:58:29 <asn> so let's sum up what we will be doing for next week?
15:58:54 <dgoulet> just to be clear, we don't *HAVE* to present a stats aggreation plan, just that if we want that, presenting some "agenda" in April sounds like a clever idea
15:58:55 <asn> - I will be looking at the top-down list, and specifically the stresstesting stuff.
15:59:14 <asn> I will also take a look at the SMC stuff with nickm.
15:59:21 <asn> And also at the timeline mail by dgoulet.
15:59:24 <syverson> Someone will need to coordinate with Phil about what will be presented at the PI meeting kickoff. I think that will be Roger and me.
16:00:03 <dgoulet> syverson: right armadev wants to be on the stage I think this time :)
16:00:06 <asn> you dgoulet, what will you be doing? i guess you will be working towards rows 4 and 5?
16:00:17 <asn> PI meeting is related to Memex?
16:00:26 <syverson> But may imply answers to e.g. what nickm thinks when he looks into what the stuff ohmygodel pointed at will look like.
16:00:37 <syverson> PI meeting is the Memex PI meeting.
16:00:42 <asn> ack
16:00:46 <dgoulet> asn: yes mostly and I would also like to help you out for stess testing if you need anything, I've played with that quite a bit lately
16:00:53 <asn> dgoulet: yes i will need your help.
16:00:56 <asn> ok
16:00:59 <asn> and karsten?
16:01:06 <karsten> that ticket
16:01:10 <asn> you will be lookin at #15513 and talking to me over IRC.
16:01:13 <asn> ok
16:01:15 <dgoulet> awesome
16:01:16 <karsten> yep
16:01:16 <asn> that seems reasonable.
16:01:22 <dgoulet> asn: wait
16:01:35 <dgoulet> asn: is the DoS performacne issue also something you are intersted in?
16:01:42 <asn> btw, with this and that and the DoS stuff, next week seems quite intense and I also have a plane to catch.
16:01:43 <dgoulet> asn: that fits in SponsorR
16:01:45 <asn> dgoulet: yes
16:01:51 <asn> i will try to look into that too.
16:01:58 <asn> i will do stuff best-effort and try to be useful.
16:02:02 <dgoulet> asn: cool, no rush, just fun stuff I think :)
16:02:14 <karsten> so, meet next tue or fri?
16:02:18 <Yawning> (as usual, lemmie know if y'all need more help)
16:02:25 <asn> next tuesday is not very convenient for me.
16:02:27 <dgoulet> Yawning: you are the ace in the hole for R! :P
16:02:40 <asn> because i will just have arrived to a different location.
16:02:47 <asn> do other people also prefer friday?
16:02:50 <Yawning> heh
16:02:53 <dgoulet> no problem for me on Friday
16:03:00 <isabela> at 16:00utc?
16:03:03 <dgoulet> sure
16:03:04 <karsten> works for me
16:03:11 <asn> we talking about next friday or this friday? next, right?
16:03:16 <dgoulet> 10
16:03:17 <isabela> april 10th
16:03:18 <karsten> neyt
16:03:19 <asn> yes
16:03:20 <karsten> next
16:03:21 <asn> ok great
16:03:26 <asn> i'm fine with it.
16:03:31 <isabela> ok
16:03:40 <asn> isabela: does our plan make sense?
16:03:44 <asn> isabela: or?
16:04:13 <asn> or would you like to resolve the top-down situation in a different manner?
16:05:46 <asn> #endmeeting