20:00:31 #startmeeting anti-censorship checkin 2019/03/28 20:00:31 Meeting started Thu Mar 28 20:00:31 2019 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic. 20:00:35 hello everyone 20:00:46 hi 20:00:51 our pad is at https://pad.riseup.net/p/tor-censorship-2019-keep - it was down earlier so let's do updates 20:00:57 hi 20:01:20 o/ uniqx 20:01:23 hi 20:01:54 * arma1 is nearby. trying to stay afloat in the logistics items that i'm doing now that i'm back in nyc 20:02:37 let's start with snowflake roadmap 20:02:45 https://storm.torproject.org/shared/OdNtwrtRrqklh76l4PfcngBbQFDbjv_jRroj0WeSY0B 20:03:00 * sysrqb lurks 20:03:05 i think it looks right? 20:03:20 yes i think so 20:03:33 21315 is sort of in progress 20:03:42 there we go 20:03:59 and #21304 seems to still be in review 20:04:13 yes 20:04:16 cool! 20:04:19 yup we've done one round of reviewing and now it's in review again 20:04:26 I plan to re-reivew #21304 this week 20:04:34 thanks! 20:04:38 awesome 20:04:48 gettor, i think we skip 20:05:08 https://storm.torproject.org/grain/wTcdAfPCjooEoAnRQYcHBr/ any changes to the bridgedb roadmap? 20:06:20 ahf: not that i know of. though maybe i should move the release task to backlog 20:06:26 gah, storm is slow for me today 20:06:50 me too 20:07:18 when i add a filter nothing happens except the "in progress" one becomes empty 20:07:52 is the overall roadmap synced automatically to the snowflake one? it looks like no 20:08:01 cohosh: nope :-/ 20:08:10 catalyst: hm, which ticket is this? 20:08:13 the release task 20:08:38 ah, #29276 20:08:58 i see it in the backlog i think 20:09:15 ahf: i just moved it :) 20:09:34 (after storm finally woke up for me) 20:10:01 ah 20:10:03 thanks 20:10:25 * ahf tries to apply the PT filter 20:10:45 ok, that one looks like we haven't updated it in a while :-) 20:11:08 #29272 should go into 'done', right cohosh? 20:11:09 <_hc> hi all, I actually managed to remember this time slot today :-) kids are sleeping... 20:11:16 _hc: o/ 20:11:33 cohosh: or maybe in progress, i guess? 20:11:58 yeah i'd say done for now 20:12:21 moved it over to done now 20:12:24 cool! 20:12:26 we can open new tickets if we want to do something with it 20:12:43 yeah, i think that is a good idea 20:13:15 hm, kat5 have a needs help with entry with gettor 20:13:49 but i also see an item with syncing with hiro next week about it, i think that is the best option right now. i don't know the roadmap for gettor 20:13:56 * catalyst opened a few tickets about gettor but that was a long time ago 20:14:17 but i could imagine hiro being very busy with getting the shiny new website into shape right now and is usually not around for these meetings 20:15:03 i see some items from dcf1. i think we should get #29863 into the roadmap if it isn't already 20:15:46 and ugh, #29875 looks like another PT regression 20:15:57 i wonder if they used one of the alpha tor browser releases :-/ 20:16:00 was #29863 dependent on moving the snowflake broker to a TPA machine? 20:16:19 i know anarcat mentioned they already have some tools available for this 20:16:24 cohosh: no, I think if we install the right prometheus package it can happen before being moved to TPA. 20:16:26 hm 29875 sounds like geko's earlier bug where he tried to switch PTs early in tor browser startup 20:16:33 yeah 20:16:37 dcf1: ah ok, i'm happy to pick that one up then 20:16:40 it looks like the wrong component here though 20:16:46 ahf: yes, that's with an alpha 20:16:52 i reproduced it 20:16:55 isn't this a tor problem rather than a snowflake/torbrowser launcher issue? 20:16:56 Although I don't know how the prometheus-node works or how it authenticates whatever is interrogating it. 20:17:08 GeKo: ah, awesome (and hi!) 20:17:11 ahf: my first assumption is that it would be a tor bug 20:17:15 arma1: it's not during start-up 20:17:23 GeKo: shouldn't we move this one out of the snowflake component on trac? 20:17:35 arma1: yeah, ditto 20:17:40 i guess to core tor, yes 20:17:55 (hi!) 20:17:58 and it's a 040 issue, right? 20:18:03 yes 20:18:04 it didn't happen with 35x ? 20:18:09 dcf1: me neither, i will look into it and talk to anarcat. do you want me to check with you before installing packages on the snowflake machines? 20:18:14 aye /o\ i was hoping for no more PT regressions haha 20:18:20 i could test with 035 20:18:25 if that helps 20:18:27 and then they are just hiding in the other components 20:18:33 GeKo: if you have the time to do that, that would be awesome 20:18:50 i'll do it tomorrow 20:19:01 i can even bisect ;) 20:19:32 cohosh: if it's fine with you it's fine with me, if it seems obvious go ahead and do it. 20:19:40 thanks, let's see if we should give it the 040-must mark tomorrow too then 20:19:45 okay thanks! 20:19:46 kk 20:19:58 i have moved it over to the tor component now 20:20:13 dcf1: thanks for bringing that one up here at the meeting, that one had gone over my head 20:20:47 np 20:21:01 cohosh: extra points if you keep track of which packages you installed, so once it moves to a TPA machine, you can say "and i need these eight packages" 20:21:04 i thiiink, that was all the "help with" on the pad? it sounds like dcf1 is going to look at #21304 20:21:25 arma1: cool, i will document the process and add it to the ticket 20:23:20 ok, no more help items i think 20:23:43 should i summarize the conversation with NewNode? 20:23:49 cohosh: you wanna run the newnode conversation? 20:23:50 yeep! 20:23:53 cool 20:24:08 so a group working on something called NewNode emailed us a while back 20:24:36 they are just wrapping up an OTF funded project to make what understand is a mobile-based P2P CDN 20:24:56 and the OTF funders suggested they adapt it to be a pluggable transport 20:25:33 i was a bit confused at first because the connection between CDNs and a PT for Tor isn't obvious, but i had a chat with Marina from the that project today 20:25:52 and she mentioned that they can also route any kind of encrypted internet traffic in a P2P way 20:25:59 Oh I see the name Stanislav Shalunov on http://www.newnode.com/, I know who that is. 20:26:08 ah yes! cool 20:26:40 they're hopeing that if the system is popular for CDN purposes it will a "too big to block" but P2P system that could potentially be used for censorship resistance 20:26:47 that's my current understanding 20:27:17 at this point they wanted to reach out and introduce the project, they mentioned wanting to apply for the OTF pluggable transports funding in May i think 20:27:18 arma1: do you know these guys ^^ 20:28:04 and wanted to know if we are interested and would possibly show some support for the idea for the funding application (i'm not quite sure how that works) 20:28:17 is stanislav the guy who worked on the...local wireless comms network that was big in hong kong a few years ago? 20:28:28 firechat? 20:28:30 yes 20:28:39 yup, marina mentioned that in the call 20:28:47 ok. then yes i have met stanislav 20:28:58 i met him at, of all places, that fbi conference i go to 20:29:07 he spent the whole time telling me that tor is stupid and insecure and worthless 20:29:15 kind of a weird conversation 20:29:17 o.O 20:29:27 #winner 20:29:46 like, in that way or more like weird awkward nerd kind of technical discussion about all the tor issues that sometimes comes up at conferences? 20:29:49 the generous interpretation is that he has different priorites and a different threat model 20:29:49 Also https://en.wikipedia.org/wiki/LEDBAT 20:30:04 i sometimes meet people like that, and i don't think they are bad, i think they are just maybe a bit too focused on whatever it is they are doing 20:30:29 right 20:31:15 it sounds like maybe for now they'd be interesting to stay in contact with and/or maybe help them with something related to OTF if we can and have that knowhow? i guess we have more knowhow with those than most people do? 20:31:32 cohosh: so eventually, if this works, we would ship a little newnode client shim in tor browser, and run a bridge on the other side of the newnode cdn? 20:32:07 arma1: i think so, to be honest i don't yet have a clear idea of how the NewNode system works 20:32:26 the white paper doesn't go into much technical detail and i haven't found a specification, just the source code 20:33:18 i emailed them the PT evaluation page yawning angel had made, a link to the traffic-obf mailing list, and a link to the tor pt spec 20:33:39 that's a good start, thanks for ambassadoring 20:33:40 and they emailed me a link to the code repository 20:35:18 np, it looks interesting. i'm not sure what support from our side would look like at this point 20:36:18 maybe they should reach out when they have something they think we can be helpful with in different contexts? 20:37:04 if they're only reaching out about a PT option because otf told them to, we're going to need to do most of the work 20:37:08 yeah perhaps, they probably know what they want from us more than we do at this point 20:37:14 arma1: agreed 20:37:18 and if we can't figure out what it is they're doing, then we're not going to have much reason to do work 20:37:30 my impression from this reach out was that they wanted to confirm we were interested and to say hello 20:37:46 i guess another option is that we should get that feedback loop back to the otf people, so they know what happened and what didn't happen 20:38:40 i agree that i don't think we should be doing the development there, i also don't think that is what they want? 20:39:34 yeah i think they are mostly trying to be sure that if they make something we'll be open to it being used 20:40:06 presumably before we integrate it into like Tor Browser we'll need to know more about how it works 20:40:56 but it might be too early for that, i guess it depends on what they want to propose to OTF 20:41:05 and what kinds of deliverables OTF will hold them to 20:41:21 i hope they don't make the "marionette mistake" with OTF 20:41:28 and get a deliverable that says "get X into tor browser" 20:41:50 oof, okay, that is something we should tell them about then 20:42:08 because that was a lot of stress i think for the marionette team(?) in those last 2 weeks or whatever it was when they started having a bit of panic 20:42:14 and we didn't have time to look at it 20:42:24 yeah 20:43:05 do we know who at otf pointed them toward us? 20:43:16 might be good, now that we have two data points, to have a periodic sync with that person too 20:43:19 (maybe it was adam?) 20:43:38 they didn't mention but if i had to guess i would guess adam 20:45:17 is this around some money they have already gotten or some money they are trying to get? 20:45:26 this is a new proposal they are writing 20:45:49 they are just finishing up a previous project with otf 20:46:00 which i believe was NewNode as a CDN 20:46:17 afaik 20:46:47 ah ok so if they are writing a proposal to otf to adapt newnode to be a pluggable transport, then ... it would sure be swell if we could understand their design 20:46:56 and should be setting off alarm bells for them and otf if we cannot 20:47:23 yes agreed, i would also prefer not to have to read C code to understand the main idea 20:47:35 +1 20:47:49 maybe when they have an early draft of their proposal, which surely will be good at explaining what's going on, they can show it to you too 20:48:02 ('surely') 20:48:35 cool, that is a good thing to mention to them 20:48:42 and also the marionette cautionary tale 20:49:03 maybe point them at ptlib too 20:49:11 and a couple of example PTs they can model their thing after 20:49:20 okay sounds good 20:50:10 awesome 20:50:21 thanks! 20:50:23 maybe tell them we have these meetings too if they want to come by and say hi 20:50:44 but maybe they should tell us before, because the first 20 min. or so of these meetings are pretty useless to them i think :-S 20:52:05 cool 20:52:16 cohosh: is the TPA machine thing next topic or was that solved already or? 20:52:39 <_hc> whenever there is time for a quick new topic, I'd like to introduce uniq and ask about the next Tor Meeting 20:53:24 ah i think there isn't any rush for moving snowflake infrastructure to TPA machines, and it will probably involve dcf1 20:53:56 oki :-) 20:53:59 to deal with sorting out domians correctly 20:54:17 cohosh: ln5 and anarcat can probably help there too 20:54:27 _hc: cool! let's hear? 20:54:36 yep. i'd suggest spending energy on moving to a TPA machine, rather than hacking together your own monitoring stuff, if you have energy to spend there 20:54:42 yes probably, i'm just thinking there are things i don't have access to that need to be modified 20:54:58 arma1: i think we can do the monitoring using anarcat's thing on the current setup 20:55:03 so these issues are a bit separate 20:55:26 cohosh: speaking of things only dcf has access to, one of our goals in general is to make sure that every thing has at least two people who can fix it 20:55:38 <_hc> ok, first the bad news, I can't make it to Stockholm, I have a conflict :-( but the good news is that uniq is available. 20:55:41 so if there are things that only dcf can do, those are bugs :) 20:55:42 <_hc> He's been working with Guardian Proect for a while, and is a Tor user since 2008 20:55:59 <_hc> specifically on PTs 20:56:11 i have access to the snowflake machines 20:56:20 <_hc> incluuding the Android and devops sides of it. so I think it makes sense for him to go in my place 20:56:25 but i think not the domain forwarding 20:56:32 _hc: you have seen that this meeting is slightly different, right? in that i don't think we have any open days, but, i think there should still be some PT related sessions there for sure 20:56:57 i'm not sure how it works with invitations to these meetings now 20:57:02 i could try to figure it out 20:57:14 <_hc> I responded to Jon with that same info 20:57:45 ahf: i think the way it works is that we go to isa and explain about the extra person, and then she juggles all the information and makes a decision 20:57:54 i think right now we have more people-we-want-to-invite than spaces-for-humans 20:57:58 <_hc> that's my bit. I'm happy to answer quesrions about IETF, ESNI, Android PTs, etc. 20:58:25 arma1: oh, ok. i'll talk to isa 20:59:00 actually, _hc, next week gaba is back, so we should probably also involve her in this and figure out what PT related things we want to do at this meeting 21:00:25 _hc: i read your IETF mail, that was very useful 21:00:32 ahf, cohosh: oh, speaking of PT things, i have word from josh, the program manager for race, that we won't need to deal with pub review restrictions for race. this is good news. 21:00:38 <_hc> glad to hear it 21:02:05 arma1: cool 21:02:14 are "pub review restrictions" that think where the govt wants to read our work before we publish it? 21:02:15 arma1: but, so, do we have the money or what? 21:02:18 *that thing 21:02:30 cohosh: yeah :-/ 21:03:02 thanks friends, I gotta dip. 21:03:07 dcf1: o/ 21:03:17 dcf1: thanks! 21:03:18 dcf1: btw, are you going to stockholm? 21:03:28 it would be very nice to finally meet you 21:04:34 cohosh: yes, exactly. and the workaround is to never actually do any work for them. which is kind of awkward. 21:04:47 ahf: no, nothing is signed yet. but things are still moving forward. 21:04:50 yeah i was gonna ask if we'd even take money with those restrictions on it 21:05:23 cohosh: i almost turned down the sponsor19 money, because they wouldn't get rid of the pub review requirement 21:05:36 they eventually contacted me and suggested that perhaps we would just not have any deliverables 21:05:43 ahh ok 21:05:43 and we worked out a workaround 21:05:54 still awkward, and requires lots of trust on both sides, etc 21:06:15 such a weird system :-S 21:06:19 in general we want to never have to deal with that, 21:06:30 but also i general we want to not fire our developers because we ran out of money 21:06:36 but i guess as long as we can do some meaningful work and they think what we do is meaningful, then it's all good 21:06:39 we all want things i guess 21:06:48 yeah 21:07:07 hey all I didn't know I needed to be around this meeting... can I be part of the list where you people send your updates and organise this? 21:07:30 hiro: hey! i don't think we have a list :/ just a usual meeting time 21:07:31 hiro: no, i don't think you needed to be 8) we just had a gettor item on a list 21:07:36 ok 21:07:43 * cohosh wonders if it is even posted on the wiki 21:07:44 hiro: but it sounds like kat5 is going to write to you next week :-) 21:07:51 yes kat5 wrote me 21:07:53 cohosh: i don't think we have a team page yet there 21:08:08 well feel free to ping me a bit before hand if you want ame around 21:08:09 oh hey, creating a team page on the wiki seems smart 21:08:12 at this time I am finishing dinner 21:08:15 we have a pad: https://pad.riseup.net/p/tor-censorship-2019-keep 21:08:18 okay, we are 8min. passed the hour and i haven't had dinner yet and it's 22 here. should we close it off? 21:08:25 :-) 21:08:27 okay sounds good 21:08:33 thanks ahf! 21:08:40 cool. thanks all o/ next week is exciting because phw begins on monday! 21:08:45 \o/ 21:08:56 #endmeeting