16:00:01 #startmeeting tor anti-censorship meeting 16:00:01 Meeting started Thu Jun 5 16:00:01 2025 UTC. The chair is meskio. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:01 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:00:05 hello everyone!!! 16:00:09 here is our meeting pad: https://pad.riseup.net/p/r.9574e996bb9c0266213d38b91b56c469 16:00:11 ask me in private to give you the link of the pad to be able to edit it if you don't have it 16:00:13 I'll wait few minutes for everybody to add you've been working on and put items on the agenda 16:00:16 hi 16:03:40 we have one discussion topic: 16:03:43 Adding more snowflake builtin bridge options 16:03:47 cohosh: ? 16:04:11 i opened tor-browser#43842 to discuss this 16:04:33 orbot has different snowflake builtin bridge options for each of the three rendezvous methods 16:04:59 i think this is, in general, a good idea 16:05:20 we have the circumvention settings api but if the front domain is unusable or blocked, users have trouble reaching it 16:05:46 we're seeing that our default settings are less and less generalizable lately as different places try to block snowflake in different ways 16:06:19 however, for sqs i think it's better to wait and continue to ramp up slowly before making it a builtin option for cost reasons 16:06:51 sounds good, I wonder how to explain the differencies to users, but something like "use this by default and it fails use the others" might work for most users 16:07:02 my costs jumped up to 8 USD last month, which is still low, but considering the fraction of SQS users compared to other rendezvous methods, i'm still not sure it will scale well 16:07:40 meskio: yeah i was thinking having something like "snowflake-google" (ampcache), "snowflake-cdn77", and "snowflake-amazon" (SQS, when it's ready) 16:07:54 we did soemthing similar with meek in the past 16:08:13 +1 16:08:18 the nice thing about this is it also acts as a disclosure of which third party is seeing your IP address and the fact you're connecting to Tor 16:08:44 dcf1: thanks for the link to https://gitlab.torproject.org/tpo/anti-censorship/team/-/wikis/Snowflake-costs 16:08:55 i spent a few minutes trying to find that page yesterday 16:09:44 that's it from me, i think the actionable thing is to add a snowflake-ampcache bridge as a builtin then 16:09:53 I had this dream once of a per-transport configuration menu in Tor Browser connection settings, where you pick a transport and it shows a custom control panel for that transport (i.e. for snowflake a dropdown with rendezvous methods). Rather than encode every combinatorial option into a pseudo–transport name in a single dropdown. 16:10:11 hm, interesting 16:11:17 that might make sense, we should ask the UX team about it 16:13:27 yeah, definitely 16:13:44 cohosh: should we look into moving SQS costs to TPI? 16:14:08 meskio: that would be preferred but even so we need to be careful about scaling 16:14:25 i'll send an email to start that conversation 16:14:25 sure, I'll ask around on how to do that 16:14:35 the nice thing is, AWS budget actions seem to work well 16:14:53 meskio: oh that would be great, thanks 16:16:10 I guess we'll wait for apps team to voice their opinion on the issue 16:16:17 anything more on this topic? 16:16:23 not from me 16:16:54 no more discussion topics, but few interesting links 16:17:02 A wall behind a wall 16:17:07 a gfw.report publication 16:18:20 nice! maybe we could do a reading group discussion on this 16:18:40 sure, that will be nice 16:19:19 should we try in two weeks (June 19)? 16:19:46 that can work for me 16:20:09 I'll check with onyinyang and shelikhoo_ once they are online, but let's try that date 16:20:57 https://github.com/ban6cat6/aparecium 16:21:38 * onyinyang agrees from absentia 16:21:45 XD 16:21:46 :) 16:22:06 on a fast look it doesn't look like this would apply to something like webtunnel were we use a real nginx/apache server 16:22:28 but maybe I'm not sure I understand all the details 16:22:46 anyway, yes nice graphs of ShadowTLS and REALITY 16:24:17 and the last link is about snowflake users in Turkmenistan, they remain elevated 16:24:38 BTW, since today we do recommend snowflake in turkmenistan as default option in circumvention settings 16:24:59 i'm glad it's continuing to work there 16:25:12 meskio: oh, is there a commit reference for that? I'll add it to the metrics timeline. 16:26:07 * meskio looks for the commit 16:26:25 https://gitlab.torproject.org/tpo/anti-censorship/rdsys-admin/-/commit/ccd914ba9c3828c77e7ffaa7458db79cb0ef93eb 16:28:03 we should somehow map the changes in the circumvention map to the metrics timeline, maybe I should get used to do changes in the timeline when we change the map 16:29:13 anything more for today? 16:29:32 not from me 16:30:00 meskio: my usual thinking is, document it if it might help explain a graph feature, like if someone is looking at the graph and might want to know possible causes or related events 16:30:44 I agree, but I haven't being giving much love to the metrics timeline 16:30:55 dcf1: there are few other changes that we might want to document together 16:31:16 https://gitlab.torproject.org/tpo/anti-censorship/rdsys-admin/-/merge_requests/39#note_3206264 16:31:38 you should be able to commit, and feel free to commit directly, no need for a MR or discussion. The "tidy" script checks the syntax of the file. 16:31:47 and then maybe the right commit is https://gitlab.torproject.org/tpo/anti-censorship/rdsys-admin/-/commit/85fc4032d15fc6776f6cb31d5f9fdb1ff983b2a8 16:31:53 as mine was just fixing a typo of this one 16:32:08 got it 16:32:27 thanks, I'll do 16:33:19 I guess we can finish this meeting here 16:33:31 #endmeeting