16:00:02 <meskio> #startmeeting tor anti-censorship meeting
16:00:02 <MeetBot> Meeting started Thu Jan 15 16:00:02 2026 UTC.  The chair is meskio. Information about MeetBot at https://wiki.debian.org/MeetBot.
16:00:02 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:05 <meskio> hello everyone!!!
16:00:09 <meskio> here is our meeting pad: https://pad.riseup.net/p/r.9574e996bb9c0266213d38b91b56c469
16:00:11 <meskio> 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 <meskio> I'll wait few minutes for everybody to add you've been working on and put items on the agenda
16:00:14 <onyinyang> hihi!
16:00:17 <ln5> helo
16:00:21 <theodorsm> hiii^^
16:01:15 <cohosh> hi
16:01:23 <Shelikhoo[mds]> hi~hi~
16:02:31 <meskio> should we start with the first topic?
16:02:39 <meskio> Add rate limiting back to broker for just /proxy endpoint for now
16:02:47 <cohosh> that one is mine
16:03:12 <cohosh> the proposal is here: https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/snowflake/-/issues/40506
16:03:34 <cohosh> i wanted to bring it up here because there's no code review process for this
16:03:49 <cohosh> i removed rate limiting from the broker because i thought it was causing client timeout errors
16:04:11 <cohosh> we'd originally put it in place to protect against DoS attacks on broker endpoints
16:04:28 <cohosh> i want to bring it back just for the /proxy endpoing (and maybe robots.txt)
16:04:52 <cohosh> the /client endpoint is still tricky to implement
16:05:20 <cohosh> and the /answer endpoint is also tricky because we don't want to prevent an honest proxy from responding after a client match
16:05:41 <cohosh> if anyone is opposed please say so here or on the issue and i'll deploy it early next week
16:06:02 <Shelikhoo[mds]> I wonder if this limit would impact users in school, or in environment where they are sharing outbound address with other persons
16:06:21 <cohosh> i'm not overly concerned about it affecting proxy polls
16:06:46 <cohosh> arguably, we want more IP address diversity so preventing multiple proxies with the same address from polling too often is also good
16:07:07 <Shelikhoo[mds]> yes, I think this is true...
16:07:28 <Shelikhoo[mds]> we can set this limit now and see if there is anyone complains
16:07:38 <meskio> I've been running in some places multiple proxies on the same IP, when the network was more conjested
16:08:31 <meskio> but makes sense to me to add the rate limit
16:08:35 <cohosh> meskio: ok, that's good to know, there's probably some tradeoff with proxy pool availability and poll rate
16:08:48 <cohosh> which is why we might actually want https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/snowflake/-/issues/25598
16:08:56 <cohosh> but this would only apply to honest proxies
16:09:06 <cohosh> maybe i'll set the rate limit a bit higher
16:10:53 <Shelikhoo[mds]> no thing more from me on this topic
16:11:11 <meskio> during the iran-israel war the proxies were overloaded, I think is the time set this up, not sure if I ever tear it down, in most of the situations it looks like we don't have such overloaded network and as you say we want IP diversity
16:11:46 <cohosh> yeah we can also remove it if we need more proxy pool capacity
16:11:55 <meskio> +1
16:12:00 <meskio> let's add it we can always change
16:12:30 <cohosh> ok that's it from me on this
16:12:40 <meskio> cool, let's move to the next topic
16:12:44 <meskio> "Conjure test network"
16:12:47 <meskio> ln5: ??
16:13:08 <ln5> yes, thanks. i'd like to know how much effort to put in, and electricity
16:13:29 <ln5> i shut it down on dec 1 to save electricity
16:13:45 <meskio> is it easy to bring it back when we need it?
16:14:16 <ln5> easy enough to start it and verify that it's working, but if not... might take some time
16:14:32 <ln5> (if i figured out how to *not* do the userland polling the power consumption would be far less)
16:14:44 <meskio> I see
16:14:58 <ln5> but the other issue is to keep it updated
16:15:10 <ln5> which will take some hours per time unit
16:15:29 <ln5> and also set up monitoring, so that we know when it needs care
16:15:40 <cohosh> we don't have active conjure development planned on our team for now
16:15:41 <ln5> all can be done, but if there is no value i don't want to spend cycles
16:15:48 <meskio> we don't have a clear plan on when we'll get back to work on conjure, the next steps are to get a bridge up (I think is mostly there) and start giving it to users
16:16:08 <meskio> so it will be useful in the future, but I think is fine to keep it down until we get back into developing conjure
16:16:14 <meskio> and it might take some time
16:16:21 <ln5> like months?
16:16:38 <meskio> yes
16:16:49 <meskio> I don't expect us to do much with conjure in Q1
16:17:04 <ln5> ok, that's enough for me to know how to handle it for now, thanks
16:17:19 <meskio> we might not do any serious development until we try it a bit with users and apply for another grant with what we learn
16:17:25 <meskio> if so it will take many months
16:18:19 <cohosh> ln5: thank you for putting so much care into the setup and documentation process for the test network
16:18:40 <onyinyang> yes, thank you ln5!
16:19:06 <ln5> thanks, i will point someone i met earlier towards the docs and see if they're able to set it up themselves
16:19:24 <Shelikhoo[mds]> thanks ln5!
16:19:25 <cohosh> that's amazing
16:19:25 <meskio> nice :)
16:20:14 <meskio> any other topics to discuss today?
16:20:27 <meskio> there are two interesting links, but I think they are from last week, I'll remove them
16:21:24 <meskio> do we want to pick up a date for the reading paper we have in the queue? "Fingerprint-resistant DTLS for usage in Snowflake"
16:22:11 <Shelikhoo[mds]> yes, I think we can set a day for that!
16:22:20 <meskio> Jan 29th?
16:22:22 <meskio> in two weeks?
16:22:30 <meskio> or do we want more time for it?
16:22:49 <Shelikhoo[mds]> I thin Jan 29th would work for me
16:23:07 <onyinyang> I'd prefer the following week if possible
16:23:27 <meskio> sure, let's do Feb 5
16:23:28 <Shelikhoo[mds]> I am happy with the following week as well
16:23:39 <Shelikhoo[mds]> nice!
16:23:46 <meskio> set then
16:23:46 <theodorsm> I'll join to answer questions etc
16:23:52 <meskio> :)
16:24:07 <meskio> I'll give it 1min to see if someone has something else and I'll close the meeting if not
16:25:09 <meskio> #endmeeting