15:58:01 <shelikhoo> #startmeeting tor anti-censorship meeting
15:58:01 <shelikhoo> here is our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep
15:58:01 <shelikhoo> feel free to add what you've been working on and put items on the agenda
15:58:01 <MeetBot> Meeting started Thu May 11 15:58:01 2023 UTC.  The chair is shelikhoo. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:01 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:58:19 <shelikhoo> Hi~ Hi~
15:58:34 <shelikhoo> I think the meetbot's clock is out of sync?
15:59:01 <itchyonion> hello
15:59:03 <gaba> yes, is it logging?
15:59:18 <shelikhoo> yes, but the timestamp seems off
15:59:21 <shelikhoo> but it is fine
16:02:22 <onyinyang[m]> hihi o/
16:02:47 <shelikhoo> okay, I didn't see a lot of new edits, let's begin the meeting
16:02:57 <shelikhoo> There is no new discussion topics
16:03:10 <shelikhoo> so let's start from interesting links
16:03:21 <shelikhoo> Research about designing an armored bridge line sharing URL format(https://gitlab.torproject.org/tpo/anti-censorship/team/-/issues/126)
16:03:28 <shelikhoo> The first link is from me
16:04:02 <shelikhoo> there is a new armored bridge line format being designed to help with sharing bridge lines
16:04:15 <shelikhoo> all comments welcome
16:04:31 <shelikhoo> we will discuss it within our team next week
16:04:44 <shelikhoo> (everyone is welcome)
16:05:38 <shelikhoo> and formally request opinion from other stake holders once we finished discussion on anti-censorship team side
16:06:00 <shelikhoo> okay, any comments?
16:06:32 <shelikhoo> okay, I think the next link is from dcf1
16:06:34 <shelikhoo> https://opencollective.com/censorship-circumvention/projects/snowflake-daily-operations/updates/2023-april-update
16:06:59 <dcf1> this is just the latest snowflake-01 update
16:07:09 <dcf1> its doesn't include snowflake-02
16:07:25 <dcf1> besides the fact that snowflake-01 is still at capacity, the third graph is interesting
16:07:56 <dcf1> it shows the effect of the release of snowflake-webext 0.7.2, which fixed a bug that caused clients to be attributed to the "??" country
16:08:45 <dcf1> across the 2 bridges, snowflake has transfered about 5 PB since the beginning, almost 1 PB of that just in the past month. currently about 30 TB per day.
16:09:17 <dcf1> I don't have a graph handy of the snowflake-02 contribution, but you can see here, snowflake-01 is about 70k and snowflake-02 about 10k
16:09:21 <dcf1> https://gitlab.torproject.org/dcf/snowflake-graphs/-/blob/main/userstats-bridge-transport-multi.csv?plain=1#L3352-3355
16:09:33 <dcf1> https://gitlab.torproject.org/dcf/snowflake-graphs/-/blob/93b1cdb35be5f71a5ae12f68c8813ef21e019eae/userstats-bridge-transport-multi.csv#L3352-3355
16:09:45 <dcf1> that's all
16:09:54 <shelikhoo> yes, I think there is a lot of connection from US, is this an issue related to GeoIP database?
16:10:12 <shelikhoo> or we actually have this amount of US user?
16:10:16 <dcf1> yes, I believe it is an error in geolocation, see the March 2023 update
16:10:41 <dcf1> https://opencollective.com/censorship-circumvention/projects/snowflake-daily-operations/updates/2023-march-update
16:11:00 <dcf1> "The second largest share geolocates to the United States; however we suspect this represents geolocation errors because of how strongly it correlates with the Iran count."
16:11:17 <shelikhoo> Yes, will it helps if we switch to another geoip database?
16:11:27 <dcf1> idk
16:11:59 <arma2> i also (just from looking at the graphs a lot) think many of those US users are actually IR users.
16:12:15 <arma2> in the past, the metrics team did a bunch of geoip database evaluations, and the conclusion was that all of them were flawed, in different ways
16:12:39 <arma2> if we had infinite engineering resources, we might start tracking with two different db's, to compare them. but, with limited resources, maybe not worth it. :)
16:13:19 <shelikhoo> anyway, that's a lot of traffics, and the counties based user count have limited accuracy...
16:14:22 <shelikhoo> I think we have might receive sponsored work about better user counting, or something like that, so just remember that might be something we could work on next
16:14:40 <shelikhoo> anyway, anything more on this topic?
16:14:54 <dcf1> no
16:15:24 <shelikhoo> I didn't see any other topics on the pad, anything else we would like to discuss in this meeting?
16:15:44 <dcf1> regarding user counting, I plan to pitch this idea at a PETS rump session or something: https://gitlab.torproject.org/tpo/anti-censorship/team/-/wikis/Research-ideas#calibrate-bridge-users-estimation-with-on-bridge-socket-counts
16:16:43 <dcf1> onyinyang[m] has a "needs help with"
16:17:04 <dcf1> "figuring out whether or not the metrics I added to rdsys actually collect what we want them to..."
16:17:18 <onyinyang[m]> I don't think I can get help til @meskio is back
16:17:30 <dcf1> ok
16:17:38 <onyinyang[m]> thanks tho :)
16:17:38 <dcf1> I appreciate your detailed updates btw
16:18:00 <shelikhoo> oh, yeah, thanks for the detailed status update!
16:18:38 <shelikhoo> no need to keep anyone any longer, let's call it a meeting
16:18:38 <shelikhoo> #endmeeting