14:59:54 <asn_> #startmeeting SponsorR
14:59:54 <MeetBot> Meeting started Tue Jun 16 14:59:54 2015 UTC.  The chair is asn_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:54 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:00 <asn_> hello all
15:00:05 * isabela lurking
15:00:18 <asn_> so we have ohmygodel karsten dgoulet
15:00:25 <asn_> syverson is not around, but it's ok for now.
15:00:44 <asn_> let's start with status reports as always?
15:01:06 * ohmygodel can go
15:01:11 <asn_> ohmygodel: ah go for it!
15:01:31 <ohmygodel> I primarily came to let you all know that the publication release for the stats tech report finally came through today
15:01:37 <asn_> ah
15:01:40 <asn_> so we can publish!
15:01:45 <ohmygodel> this is "Hidden-Service Statistics Reported by Relays”
15:01:48 <asn_> yep
15:01:53 <asn_> the one about how obfuscation works etc.
15:01:55 <ohmygodel> so yes we can publish that now
15:01:59 <ohmygodel> right
15:02:02 <asn_> karsten: !
15:02:05 <ohmygodel> ill send an email as well
15:02:08 <asn_> ack
15:02:13 <karsten> sounds great!
15:02:35 <ohmygodel> other than that
15:02:36 <karsten> let me know where the updated (?) sources are, and I'll publish tonight.
15:02:51 <ohmygodel> ok karsten will do
15:02:55 <karsten> thanks
15:02:56 <ohmygodel> i dont have anything else to report
15:03:44 <asn_> great
15:03:45 <asn_> thx
15:03:46 <asn_> i can go next
15:03:53 <asn_> during early last week, i finished my msc thesis.
15:04:04 <karsten> yay!
15:04:09 <asn_> i posted it on the net here: https://moderncrypto.org/mail-archive/messaging/2015/001743.html
15:04:16 <asn_> then i started getting back into tor development.
15:04:23 <ohmygodel> congrats !
15:04:26 <asn_> i reviewed a few branches
15:04:42 <asn_> discussed the HS health tool results
15:04:55 <asn_> and organized the HS meeting on July  a bit
15:05:00 <asn_> i still need to do more of all the above.
15:05:07 <asn_> and that's that for now :)
15:05:08 <asn_> next?
15:05:17 * karsten can go next
15:05:23 <asn_> karsten: go!
15:05:24 <karsten> 1) updated graphs on HS descs archived by asn
15:05:25 <karsten> 2) discussed HS health tool scope, dependencies, etc. with dgoulet
15:05:26 <karsten> done
15:05:51 <asn_> ah yes. we have more graphs on #15513!!!!
15:05:56 <asn_> haven't had time to look at them in detail.
15:06:11 <asn_> i think the new graphs tell us that the agorahoo HS was under attack during the old graphs
15:06:12 <karsten> I didn't look in detail either. they didn't look broken.
15:06:29 <karsten> but I didn't look further.
15:06:30 <asn_> and now it's not anymore. and it almost always has 10 IPs.
15:06:43 <asn_> karsten: ack thx
15:06:52 * dgoulet can go
15:06:54 <asn_> dgoulet: please
15:07:09 <dgoulet> HS health, quite a bit of coding/review and bug hunting lately. Analyzed the HS Usenix paper and helped organized the HS meeting in DC.
15:07:21 <dgoulet> about done
15:07:22 <ohmygodel> dgoulet did you read the usenix paper ?
15:07:30 <dgoulet> ohmygodel: yup
15:07:36 <ohmygodel> ok so did i
15:07:41 <ohmygodel> would be happy to discuss
15:07:42 <dgoulet> ohmygodel: did you follow the email thread about it ?
15:07:48 <dgoulet> ohmygodel:  you should reply to it :)
15:07:49 <ohmygodel> skimmed it, yes
15:07:58 <ohmygodel> ok ill post my thoughts there
15:08:13 <dgoulet> ohmygodel: cool! also worth thinking how we can improve 224 with it
15:08:27 <asn_> yes
15:08:36 <ohmygodel> i also met with fabian lanze, a phd student of andriy pachenko last week
15:08:43 <ohmygodel> he was also working on an HS fingerprinting paper
15:09:07 <asn_> aha
15:09:17 <ohmygodel> so too bad for him :-(
15:09:18 <ohmygodel> but now we are considering working together on defenses
15:09:29 <ohmygodel> so good for us :-)
15:09:45 <nex`> is anyone else experiencing Xorg segmentation faults with the new torbrowser?
15:09:59 <asn_> ohmygodel: great :)
15:10:09 <asn_> i'll add this to the HS meeting agenda
15:10:47 <ohmygodel> asn: great good idea
15:10:50 <asn_> did we finish all reports
15:10:51 <asn_> ?
15:11:09 <asn_> i guess so
15:11:13 <asn_> let's jump to discussion phase
15:11:15 <asn_> some ideas:
15:11:41 <asn_> #idea how should we wrap up the stats that we have for july?
15:12:20 <asn_> i guess that's it.
15:12:27 <asn_> but i can break it into smaller discussions.
15:12:49 <asn_> like "which stats are these?" "which stats do we need but we don't have currently?" "where should we publish analysis?"
15:12:59 <dgoulet> I think we are at the point where we can get good conclusions with the stats we have
15:13:02 <asn_> "which stats should the SponsorR slidedeck have?"
15:13:05 <dgoulet> we just need to make them
15:13:18 <asn_> so the stats we have are:
15:13:28 <asn_> #15513, #15744 and #13209
15:13:39 <asn_> we have presentable graphs for the two first, not really for the third.
15:14:08 <asn_> we have done some analysis in all of them, but nothing presentable.
15:14:17 <karsten> we also have a few weeks/months of torperf data that we could visualize.
15:14:23 <asn_> karsten: yes!
15:14:52 <asn_> karsten: i think you already have some graphs on this right?
15:15:00 <asn_> but you have newer data already
15:15:05 <asn_> what's the ticket again?
15:15:06 <asn_> is there one?
15:15:18 <karsten> #1944 ?
15:15:24 <karsten> yes, that one.
15:15:28 <karsten> not sure if that's updated.
15:15:39 <karsten> should we start a doc where we collect results?
15:15:59 <asn_> we could
15:16:06 <asn_> are we aiming for a tor-dev post?
15:16:11 <asn_> or for a tor-talk post?
15:16:12 <asn_> or a blog post?
15:16:21 <karsten> I don't know. something for a presentation in july?
15:16:41 <karsten> maybe a 15-minute presentation where we have 5 minutes?
15:16:48 <asn_> well, i think for the memex meeting, we just need the graphs and a 2 minute analysis
15:17:01 <karsten> as in, we have 2 minutes?
15:17:07 <asn_> no idea. but not too much.
15:17:14 <asn_> what i'm saying is that if we write an article like https://blog.torproject.org/blog/some-statistics-about-onions
15:17:25 <asn_> coming up with the stuff that is needfed for the memex presentation should be easy
15:17:25 <dgoulet> FYI, we'll have an hour or so one on one meeting with the PM so the more we have the better
15:18:00 <karsten> so, we're looking for something to present in 2--60 minutes. :)
15:18:05 <asn_> :)
15:18:20 <dgoulet> yup, epic requirement
15:18:23 <karsten> collecting things in a single place could be helpful for both.
15:18:23 <dgoulet> :P
15:18:27 <asn_> karsten: i agree
15:18:44 <asn_> OK. We can do the following:
15:18:49 <asn_> I take #15513 and #15744
15:18:50 <asn_> Karsten takes torperf
15:18:55 <asn_> Dgoulet takes #13209
15:19:09 <asn_> we all write a paragraph or three, blog post style, about the results
15:19:17 <asn_> we put them in a pad
15:19:20 <asn_> we concatenate them together
15:19:23 <asn_> and see if it makes sense
15:19:32 <asn_> then we spend a day or two, curating it to make it fit together
15:19:39 <asn_> and we post it to tor-talk or blog.
15:20:03 <asn_> then from there we pick the graphs we like, and the results we can present in 2-60 minutes ,and get it ready for the July meeting
15:20:08 <asn_> o.oi
15:20:20 <asn_> we can do this last step, before we post the article wherever fwiw.
15:20:45 <ohmygodel> can i ask what is included in the torperf data?
15:20:48 <nickm> abraham: btw, if you're still there, I think that writing a pluggable transport might be a good idea
15:20:57 <asn_> ohmygodel: time to download pages of various sizes
15:21:17 <karsten> ohmygodel: let me find the details on the ticket..
15:21:34 <ohmygodel> are those times to download from a hidden service ?
15:21:40 <karsten> yes
15:22:00 <karsten> https://trac.torproject.org/projects/tor/attachment/ticket/1944/torperf-hs.5.md#L325
15:22:01 <ohmygodel> are those stats that rob has been collecting
15:22:01 <ohmygodel> or has Tor been collecting such stats as well ?
15:22:08 <dgoulet> asn_: so that would be a big blog post on all stats we've worked on, I'm good with the above!
15:22:32 <karsten> that's data we're collecting. I still didn't manage to look into what rob has.
15:22:33 <asn_> dgoulet: yeah. it doesn't need to be super big. it will be like a collage of 4-5 statistics.
15:23:17 <karsten> http://ec2-54-92-231-52.compute-1.amazonaws.com/
15:23:56 <karsten> what you see there is from last fall. the new data aren't graphed yet.
15:23:57 <ohmygodel> karsten: ok thanks
15:24:28 <karsten> other stuff I'm hoping to graph is "average number of hs descs fetched per request"
15:24:39 <asn_> the substep graph would be nice if it's done with tor master. there might be optimizations merged there.
15:25:07 <karsten> asn_: hmm, I'd have to look which tor version we're using.
15:26:12 <abraham> nickm: i am, just doing some paper reading :). Ok, how difficult/doable is it, and do you think it might get accepted as a thesis?
15:26:39 <asn_> if we have our texts/graphs by next week, then we have a week to put them together.
15:26:55 <asn_> i can ask for more feedback in PETS, and aim to have the article ready by the end of PETS.
15:27:08 <asn_> or too tight?
15:27:09 <asn_> discuss.
15:27:14 <nickm> abraham: it's not a huge coding problem; maybe look at soem of the existing code.
15:27:23 <nickm> abraham: I think that if you come up with a novel idea, it's thesis material.
15:27:30 <nickm> abraham: but ask your advisor. :)
15:28:01 <dgoulet> asn_: totally doable for next week (on my side, hs health)
15:28:16 <asn_> dgoulet: ack.
15:28:19 <karsten> asn_: I'll have something by next tuesday.
15:28:30 <asn_> dgoulet: you think you will be able to have answers for the #13208 question as well?
15:28:33 <asn_> karsten: ok great.
15:28:39 <asn_> karsten: will you be at PETS btw/
15:28:43 <karsten> asn_: no.
15:28:48 <asn_> ok np
15:29:20 <asn_> karsten: dgoulet: ohmygodel: https://etherpad.mozilla.org/F8P18C7wYp-hs-stats
15:29:21 <dgoulet> asn_: as we discussed, this one is tricky/weird but I can come up with similar answer like we talked about
15:30:00 <asn_> ok
15:31:11 <asn_> #idea stats pad: https://etherpad.mozilla.org/F8P18C7wYp-hs-stats
15:31:27 <ohmygodel> hey all, i must run
15:31:29 <asn_> ok this seems like a decent plan.
15:31:30 <ohmygodel> keep up the excellent work!
15:31:31 <asn_> ohmygodel: thx for the meeting
15:31:38 <dgoulet> ohmygodel: cheers!
15:31:41 <asn_> see you in a few days!
15:31:45 <ohmygodel> indeed !!
15:32:08 <asn_> anything else we should talk about?
15:32:40 <dgoulet> can't think of something else
15:32:45 <asn_> ok that's fine.
15:32:50 <asn_> we are done for now.
15:32:53 <asn_> i will be in IRC all week anyway
15:33:04 <karsten> nothing else from me.
15:33:09 <asn_> ok thx
15:33:10 <asn_> #endmeeting