15:58:09 <cohosh> #startmeeting anti-censorship meeting
15:58:09 <MeetBot> Meeting started Thu Feb  4 15:58:09 2021 UTC.  The chair is cohosh. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:09 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:58:13 <cohosh> hey everyone!
15:58:16 <agix> hi
15:58:18 <dcf1> hello
15:58:20 <cohosh> here is our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep
15:58:40 <hanneloresx> hi
15:58:58 <cohosh> we have a pretty light agenda today
15:59:38 <cohosh> first is the weekly reminder that the anti-censorship job opening is still up so be sure to apply soon if you're interested :)
15:59:55 <atagar> good morning
16:00:56 <cohosh> the next is that the snowflake and snowflake-webext repos are set up on anonticket.onionize.space
16:01:27 <cohosh> so you can now anonymously report and track tickets by visiting https://anonticket.onionize.space
16:01:49 <cohosh> i created snowflake-webext#21 this morning to try out the interface and it's pretty cool
16:02:04 <cohosh> we do need to approve the tickets before they are added to the repo
16:02:44 <cohosh> dcf1: you're also a maintainer of snowflake so let me know if you want to be added as a moderator to this system
16:03:57 <cohosh> alright, anyone have other discussion items for this week?
16:03:58 <dcf1> hmm probably not but I'll let you know if I change my mind
16:04:06 <cohosh> dcf1: sounds good
16:04:20 <dcf1> would we benefit from any interactive discussion on https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/snowflake/-/issues/30498
16:04:29 <cohosh> possibly
16:04:31 <dcf1> I've been following and it is a weird one
16:05:04 <cohosh> yeah, i'm not sure how much time we want to spend on it since standalone proxies are a very small percentage of snowflakes these days
16:05:24 <dcf1> It looks like, proxy connects to client, proxy connects to bridge, 20 seconds pass, then "OnClose channel" with 0 bytes transferred
16:05:24 <cohosh> but i'd still like to know what's happening here
16:05:36 <cohosh> since it might be we have faulty assumptions on how NATs work still
16:05:54 <cohosh> yup
16:05:55 <dcf1> could it be happening with web-based proxies too?
16:06:06 <cohosh> it's possible
16:06:25 <cohosh> it is also possible that this is due to a client acting strangely and not a problem with proxies at all
16:06:31 <dcf1> yeah I was just about to say
16:06:41 <cohosh> since I frequently do throughput tests from different types of NATs and haven't noticed a lot of issues
16:07:25 <cohosh> if there is a client scraping snowflake for proxy IPs this is the behaviour I'd expect, heh
16:07:47 <dcf1> mm, true
16:08:17 <dcf1> ok thanks, the situation is clearer to me now
16:09:07 <cohosh> yeah thanks for bringing it up
16:09:37 <cohosh> i have been curious about how the number of requests for snowflakes is much higher than our daily number of clients
16:11:32 <dcf1> the Tor Metrics graph is not exactly the daily number of clients
16:12:05 <dcf1> it's -- I forget how it's phrased exactly, let me look it up quick
16:12:37 <dcf1> https://gitweb.torproject.org/metrics-web.git/tree/src/main/resources/doc/users-q-and-a.txt
16:12:52 <dcf1> "Average number of concurrent users, estimated from data collected over a day."
16:13:19 <cohosh> ahh interesting
16:13:20 <dcf1> So the graph estimates, if you were to spot-check at a random time during the day, how many users would be currently connected.
16:14:05 <dcf1> So if the average user stays connected less than 24 hours, the daily unique users will be higher than the metrics graph; if they stay connected longer than 24 hours, lower.
16:14:12 <dcf1> that may explain it in part
16:14:31 <atagar> Estimates, and of course intentially imprecise for anonymity purposes. Many metrics stats are like that, but if you're concerned about a significant delta it wouldn't account for that.
16:15:58 <cohosh> we are getting about 10,000 snowflake matches per day
16:16:12 <cohosh> but i also haven't seen how many snowflakes i go through if i'm browsing on snowflake all day
16:16:26 <cohosh> that might be a worthwhile thing to track
16:16:32 <dcf1> we have a "research ideas" wiki page, right?
16:16:52 <cohosh> if we don't we should :)
16:17:02 <dcf1> I thought phw made one, let's see
16:17:46 <dcf1> hmm maybe not
16:18:05 <cohosh> we have https://research.torproject.org/
16:18:19 <cohosh> but it might be worth making one specifically on the anti-censorship team wiki
16:18:40 <dcf1> cuz I'm also interested to know, if an adversary were to try to do proxy enumeration, how many proxies would they see and how fast does the pool change
16:18:49 <cohosh> yup
16:19:21 <dcf1> now that we've created this thing worthy of study
16:19:29 <cohosh> :)
16:19:36 <cohosh> i can create this page after the meeting
16:19:55 <cohosh> i also have an email i keep re-sending to students that has a list of open snowflake-adjacent questions
16:20:17 <cohosh> or we can open a ticket for it and label it "research"
16:20:33 <cohosh> or both
16:21:25 <dcf1> as long as the idea doesn't get lost
16:22:15 <cohosh> yup, i'll definitely do the wiki, and check with phw first to see if this already exists and i'm just not finding it
16:23:16 <cohosh> any needs help with items for today?
16:23:51 <cohosh> hanneloresx: are you waiting for a review on anything?
16:23:59 <atagar> I am
16:24:11 <atagar> oops, sorry
16:24:20 <atagar> thought that said 'willing to review anything'
16:24:37 <hanneloresx> haha, i have an MR that phw is reviewing
16:24:40 <hanneloresx> so should be fine
16:25:00 <cohosh> atagar: heh awesome, i might have some bridgedb code for you to review soon :)
16:25:06 <cohosh> hanneloresx: cool!
16:25:14 <atagar> \o/
16:26:04 <cohosh> anything else before closing the meeting?
16:28:05 <cohosh> okay see you all next week!
16:28:10 <cohosh> #endmeeting