16:00:09 <meskio> #startmeeting tor anti-censorship meeting
16:00:09 <MeetBot> Meeting started Thu Jun  2 16:00:09 2022 UTC.  The chair is meskio. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:12 <meskio> hello everybody!
16:00:22 <meskio> here is our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep
16:00:23 <shelikhoo> hi~
16:00:33 <meskio> feel free to add what you've been working on and put items on the agenda
16:02:04 <meskio> I see for now we only have one point in the agenda: Endless loop on the latest snowflake
16:02:48 <meskio> is this from our random cypherpunk?
16:03:39 <meskio> I think people should open issues with this kind of problems instead of dumping them into our pad
16:03:53 <meskio> if the author is not around I think we can skip it
16:04:07 <meskio> or does anybody want to talk about it?
16:04:45 <meskio> ok, lets go to the next point in the agenda: Implement metrics to measure snowflake churn
16:05:06 <shelikhoo> yes
16:05:27 <shelikhoo> this the one that try to count distinct proxy ip
16:05:41 <shelikhoo> and allow set operation on these proxy ip
16:05:58 <shelikhoo> so that we can know how fast does these proxy ip change
16:06:16 <meskio> nice!!!
16:06:21 <shelikhoo> to see how resilient our network is
16:06:42 <shelikhoo> it used hyperloglog++ structure
16:06:49 <shelikhoo> to do the count distinct
16:07:12 <shelikhoo> instead of remembering individual IP
16:07:31 <shelikhoo> so user's ip info is not extractable from our data
16:07:38 <shelikhoo> making it possible to publish this data
16:08:26 <shelikhoo> with all probabilistic data structure, it is not fully accurate
16:08:57 <shelikhoo> which is a free bonus in our use case
16:09:04 <meskio> +1
16:09:22 <shelikhoo> because we would record user info
16:09:30 <shelikhoo> and may even publish this info
16:09:51 <shelikhoo> so comments are welcomed
16:09:55 <meskio> sounds pretty good
16:09:59 <meskio> thank you
16:10:03 <shelikhoo> no problem
16:10:51 <cohosh> thanks shelikhoo for working on that
16:11:05 <cohosh> i'm interested to see the results
16:11:16 <shelikhoo> (additionally, user ip is keyed-hashed by a secret key first, so attacker can not either proof a ip is in poor or not without that secret key)
16:11:35 <shelikhoo> (additionally, user ip is keyed-hashed by a secret key first, so attacker can not either proof a ip is in pool or not without that secret key)
16:12:11 <shelikhoo> (additionally, user ip is keyed-hashed by a secret key first, so attacker can not neither proof a ip is in pool nor proof it is not in the pool without that secret key)
16:12:49 <shelikhoo> no problem~ I am interested to see the result as well
16:13:13 <meskio> yeah, let's see what it shows
16:13:13 <shelikhoo> that's all. an invitation for comment
16:13:34 <meskio> cool, I guess we can move to the next topic
16:13:50 <meskio> last week we were talking about enforcing having 2FA in gitlab
16:14:21 <meskio> I have contacted everybody that had commit access to gitlab's anti-censorship group without 2FA and ask them to set it up
16:14:46 <meskio> there are few accounts with non-commit permitions in the group that doesn't have 2FA, but I don't consider that a problem
16:15:19 <meskio> I left the point just to inform that is done
16:15:54 <meskio> anything else for today?
16:16:29 <meskio> should we pick a paper to discuss?
16:17:21 <meskio> maybe "Even Censors Have a Backup: Examining China's Double HTTPS Censorship Middleboxes": https://dl.acm.org/doi/10.1145/3473604.3474559 ?
16:17:57 <meskio> any other ideas?
16:18:32 <shelikhoo> nothing from me
16:19:43 <meskio> I would propose to do it in 3 weeks (June 23), as I will be AFK the next 2 thursdays
16:20:40 <shelikhoo> sounds good to me....
16:20:41 <meskio> ok, so if no one has any better proposal let's do that one
16:20:48 <meskio> anything else for today?
16:21:04 <meskio> (nice to have sort meetings once in a while :)
16:22:05 <shelikhoo> EOF
16:22:11 <meskio> #endmeeting