15:58:03 <shelikhoo> #startmeeting tor anti-censorship meeting
15:58:03 <shelikhoo> here is our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep
15:58:03 <shelikhoo> feel free to add what you've been working on and put items on the agenda
15:58:03 <MeetBot> Meeting started Thu Nov  3 15:58:03 2022 UTC.  The chair is shelikhoo. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:03 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:58:07 <shelikhoo> hi~ hi~
15:58:35 <nelson> hi all! i'm nelson (aka noah levenson) from lantern
15:58:40 <meskio> hello
15:59:31 <meskio> wellcome nelson, nice to have you here
16:00:13 <nelson> thanks, i'm happy to meet you all
16:01:46 <Adetoye[m]> Hello everyone
16:01:47 <Adetoye[m]> I'm the outreachy applicant from anti-censorship :)
16:02:10 <meskio> hello Adetoye[m]
16:02:31 <cohosh> hi!
16:03:42 <meskio> shelikhoo: should we start with the first point in the agenda?
16:04:01 <shelikhoo> yes
16:04:09 <shelikhoo> Lantern and snowflake
16:04:23 <nelson> cool
16:04:39 <meskio> nelson: do you want to introduce what you are working on?
16:05:14 <nelson> sure -- we're working on a p2p volunteer proxy system that's essentially the same concept as Snowflake
16:05:48 <nelson> it's early days for us, but we're hoping to identify ways that we can be helpful to the community and collaborate, so we'd love to be in touch
16:06:33 <meskio> this is pretty cool
16:07:00 <shelikhoo> nelson: are you expecting it to share the same proxy pool with snowflake?
16:07:13 <meskio> is there any reasons why you could not use snowflake directly? is it easier to write one for your needs than addapting snowflake?
16:08:03 <nelson> we were not planning to share the same proxy pool, no -- but i think we're all very interested in ways to pool resources in that way
16:09:43 <nelson> at first, we were hoping to either use or fork Snowflake, but our understanding is that we'd be instancing a new network rather than pooling resources, ostensibly because we wouldn't be able to use Snowflake's broker?
16:10:35 <meskio> yes, the easiest right now for you would be to set up your own broker, we've being thinking on ways to share the pool with other projects but haven't work on it much
16:11:03 <meskio> (things like let the proxies decide to which projects they want to contribute bandwith to)
16:11:26 <nelson> that's really cool -- it would be awesome to collaborate and get that working
16:11:47 <meskio> we'll be happy to discuss and review changes to make snowflake work for you
16:12:04 <shelikhoo> but otherwise we already support more than one snowflake servers, and it is not necessary for the server to run tor
16:13:31 <meskio> so, if you set up your own broker you don't need much from us and snowflake is designed to don't be tor specific, but we can fix it if that is not really the case
16:13:40 <nelson> well, once we discovered that we wouldn't be able to pool resources out of the box, we took a look at our Lantern-specific requirements and figured it made the most sense to roll our own system
16:13:45 <meskio> sharing the broker might be an option that will require some work together
16:14:27 <nelson> but we remain very interested in adapting what we're working on to interoperate with Snowflake and others
16:14:40 <meskio> I understand that
16:14:56 <meskio> I would be curious to hear what you do in your own system different than snowflake
16:15:09 <meskio> will be nice to learn from your fresh start design :)
16:15:15 <nelson> the main thing is the clients -- we have a single client codebase, written in Go, that runs on both desktop and the web
16:16:57 <shelikhoo> yes, right now snowflake have separate proxy based on go and browser
16:17:07 <nelson> we wanted to try to write a highly concurrent client (capable of managing many simultaneous peer connections) that could do N:M multiplexing between downstream and upstream connections
16:17:41 <nelson> and also make it very observable so that it could send lots of status information to the UI
16:18:31 <meskio> we have explored a bit using more than one proxy at once, but didn't get faster speeds, is in the planning to add support for this
16:19:18 <nelson> it would be very interesting to compare notes on this!
16:19:30 <meskio> let me find the issue
16:20:34 <meskio> do you have anything public of this project? will you publish the code?
16:22:05 <meskio> nelson: that is our ticket about using multiple proxies: https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/snowflake/-/issues/25723
16:22:07 <nelson> it is not public yet, but i think we want to open source it if it's something that people find useful
16:22:50 <meskio> nice, I think it will be useful to look at it yes
16:23:16 <nelson> thanks for that ticket link, i'm excited to read about this
16:24:05 <shelikhoo> okay, anything more we wants to discuss on this topic?
16:24:29 <meskio> not from my side
16:24:30 <nelson> well i'm looking forward to continuing the conversation... we will send you all updates and things to look at!
16:24:40 <meskio> keep us posted
16:24:51 <meskio> exciting to see more work around webrtc
16:24:55 <meskio> ahh, just one last question
16:25:01 <nelson> sure
16:25:01 <shelikhoo> the next topic for today's meeting is
16:25:03 <meskio> are you using the pion webrtc implementation?
16:25:07 <nelson> yes we are
16:25:11 <meskio> :)
16:25:28 <shelikhoo> for browser I think it will be using the browser api
16:25:29 <meskio> that is it from my side then
16:25:54 <nelson> yeah exactly, compiled to wasm, pion just wraps the browser api
16:26:27 <nelson> thank you all for making some time in your agenda to chat today!
16:26:46 <shelikhoo> no problem
16:26:53 <shelikhoo> thanks for chatting with us
16:26:55 <shelikhoo> Distributed Snowflake Server Support will be included in Tor Browser Testing Channel(Alpha) Soon: https://gitlab.torproject.org/tpo/applications/tor-browser-build/-/merge_requests/571
16:27:30 <shelikhoo> we have already included the secondary snowflake's bridgeline into tor browser alpha
16:27:41 <shelikhoo> maybe we can find a way to get users to test it
16:28:15 <dcf1> I have been testing it myself in the past week, using a manually entered bridge line.
16:28:33 <dcf1> Do you want me to post instructions to the BBS thread for Iran?
16:28:47 <dcf1> I guess it's not even specific to Iran.
16:28:56 <shelikhoo> oh! yes! that would be super helpful, but it is not specific to iran
16:29:26 <shelikhoo> but the one in tor browser is different in the way that
16:29:38 <shelikhoo> it will keep connection to both 2 bridges
16:29:46 <dcf1> right
16:30:44 <shelikhoo> so the testing for snowflake in tor browser alpha is still necessary
16:31:03 <shelikhoo> I will write a post about that as soon as it is available
16:31:29 <meskio> cool
16:31:37 <shelikhoo> anything other ideas about this topic?
16:32:21 <itchyonion> nice work
16:32:27 <shelikhoo> anything else we wants to discuss in this meeting?
16:33:20 <dcf1> Will just mention that bandwidth at snowflake-01 is increased about 20% in the past 24 hours, after being very stable for a few weeks
16:33:40 <dcf1> I tried to indentify whether an Orbot release or something had happened, but did not find anything obvious
16:34:08 <dcf1> There is an Orbot release with the utls change in GitHub (https://github.com/guardianproject/orbot/releases/tag/16.6.3-RC-1-tor.0.4.7.10), but as far as I can tell that version is not on Google Play yet
16:34:28 <meskio> ohh, maybe people install it manually in Iran
16:34:44 <shelikhoo> I don't think people in iran have access to google play
16:34:57 <shelikhoo> without using a proxy first
16:35:05 <shelikhoo> because of us policy
16:35:08 <dcf1> I know, I'm using Google Play as a proxy for F-Droid or other distribution channels
16:35:30 <meskio> 20% is still far from what we had some weeks ago, isn't it? we are not hitting the bridge limits anymore
16:35:49 <dcf1> It's still well below the highest levels we've seen, yes.
16:35:57 <shelikhoo> I find it is quite common for user to download things from github and use it
16:36:08 <shelikhoo> f driod is not really commonly used by user
16:36:14 <dcf1> But the daily cycle has been super stable since then.
16:38:07 <shelikhoo> yes, let keep monitoring the situation
16:38:13 <meskio> I have another short notice, I created an issue for last weeks discussion on PT args: https://gitlab.torproject.org/tpo/anti-censorship/team/-/issues/104
16:38:30 <meskio> I'm not sure I do fully understand all the implications of it, feel free to chim in
16:39:15 <meskio> and that is all I have from my side for today
16:39:44 <shelikhoo> #endmeeting