16:58:45 <ahf> #startmeeting Network team meeting, 12th december 2022
16:58:45 <MeetBot> Meeting started Mon Dec 12 16:58:45 2022 UTC.  The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:58:45 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:58:49 <ahf> hello hello folks
16:58:51 <Diziet> o/
16:58:53 <dgoulet> o/
16:59:00 <jnewsome> o/
16:59:00 <GeKo> hi
16:59:03 <ahf> pad is at https://pad.riseup.net/p/tor-netteam-2022.1-keep
16:59:06 <nickm> hello!
16:59:52 <hiro> o/
16:59:56 <mikeperry> o/
17:00:04 <ahf> many people here
17:00:05 <ahf> awesome
17:00:11 <ahf> how are folks doing with their boards?
17:00:32 <Diziet> I've been picking things off the Arti 1.1.x undone board
17:00:40 <ahf> eta: i may need to pick your brain on some onionmasq stuff at the end of this week, but i think i can hack it myselkf, but i may still have a question or two
17:00:47 <ahf> very nice \o/
17:01:03 <ahf> arti folks, remember our sync tomorrow on 1.1.0 retrospective and a bit on what to do in january until we have done roadmap session there
17:02:05 <nickm> i'm generally trying to inch my brain forward on whatever tasks it sticks to :/
17:02:09 <nickm> winter is tough!
17:02:23 <Diziet> eta, nickm: I'd appreciate an early opinion about arti!936, because I think the wholesale adding of error causes called for by arti#680 is going to be a great help with arti#679 and PT cleanup in general
17:02:41 <eta> ahf: happy to answer any questions :)
17:03:05 <ahf> awesome
17:03:16 <ahf> dgoulet: anything on releases?
17:03:34 <dgoulet> nope except that next one is likely in January considering the new TROVE today
17:04:13 <ahf> sounds good
17:04:31 <ahf> don't see anything incoming
17:04:39 <ahf> i have one item i wanted to talk with folks about
17:05:05 <ahf> does it make sense for us to sync next monday? i'd think everybody should make their plans for the next week and a half around now and then we ditch meetings on next monday
17:05:20 <ahf> and anything we need we can do just ad-hoc by talking together on irc
17:05:26 <nickm> wfm
17:05:31 <Diziet> Fine by me
17:05:47 <ahf> is this OK with you too mikeperry ? i don't know if you are planning to do s61 meeting on monday?
17:05:54 <ahf> there is very few days in next week that is work days
17:06:01 <mikeperry> I will be out next week
17:06:52 <ahf> ah very easy then
17:07:00 <ahf> let's ditch netteam meeting on next monday that is then
17:07:02 <ahf> very good
17:07:15 <ahf> i see no other discussion topics for today
17:07:22 <ahf> i think people not doing s61 can go do something else now
17:07:26 <ahf> mikeperry: you wanna takeover?
17:07:40 <mikeperry> ok
17:08:26 <mikeperry> so I updated the grafana panels and organized them nicely for all the new metrics we added in 0.4.7.11/12. dgoulet had done some of those updates but missed a few. the panels are all nicely grouped now so it is easier to see what is going on
17:08:28 <Diziet> ttfn, have fun with s61
17:08:44 <mikeperry> we are ready for a batch of param updates: https://gitlab.torproject.org/tpo/network-health/analysis/-/issues/44#note_2862064
17:08:51 <mikeperry> GeKo/dgoulet: ^
17:09:24 <GeKo> right
17:09:32 <mikeperry> I think we should do that today, if possible, so we can monitor this week
17:09:38 <dgoulet> thumbs up
17:09:48 <GeKo> how do we monitor that?
17:09:58 <GeKo> with our grafana setup?
17:10:08 <mikeperry> mostly grafana, but we can keep an eye on the onionperfs, esp op-hk7a
17:10:16 <GeKo> okay
17:10:23 <GeKo> dgoulet: wanna push the params update?
17:10:30 <dgoulet> GeKo: yes I can
17:10:37 <mikeperry> some of thsoe are updates from old values so watch for dups
17:10:39 <GeKo> i'll be afk starting with the end of this meeting
17:10:45 <GeKo> thanks
17:12:57 <mikeperry> other than that, dgoulet and I are making decent progress on conflux. we made an update to the proposal to move the sequence number to a relay command, which simplifies the implementation a lot, at the cost of some additional overhead until prop#340
17:13:36 <mikeperry> I also filed a ticket about prop#340, since most of the stuff for conflux and congestion control (and many other commands) does not need a stream_id: https://gitlab.torproject.org/tpo/core/torspec/-/issues/174
17:14:05 <mikeperry> we won't be using prop#340 in the first conflux impl tho; it will be able to run without it
17:14:19 <mikeperry> just with more overhead from these commands
17:15:09 <mikeperry> we're also going to restrict this first implementation to only using conflux through pre-emptively built circuits, as I mentioned before
17:15:15 <mikeperry> this also simplifies things
17:16:23 <nickm> I can make that change in 340; are we planning to build 340 right away though?
17:16:51 <mikeperry> no; probably around when we add UDP support
17:16:53 <nickm> ok
17:17:15 <nickm> assigning torspec#174 to me assuming it can wait a couple of weeks :)
17:17:20 <mikeperry> we also have funding in S112 for the crypto upgrade for tagging-resistent ciphers
17:17:33 <mikeperry> so we can do all of that at the same time
17:17:36 <mikeperry> if we want
17:17:48 <mikeperry> (as the proposal suggests)
17:17:50 <ahf> nice
17:17:54 <nickm> mikeperry: sounds excellent
17:18:12 <nickm> We should reach out to the counter galois onion team, see where they're at.
17:18:14 <nickm> What's our timeframe?
17:18:25 <nickm> I can write them an email and cc you (and whoever else is appropriate)
17:18:50 <mikeperry> yeah I assume you are in contact with them? yeah, that would be good. probably dgoulet too so he can have an idea wrt UDP timeline
17:19:06 <mikeperry> the UDP support is tied to the VPN funding.. not sure that timeline
17:19:39 <nickm> ok. I should know that before I reach out so I can tell them what our timeframes are
17:19:42 <nickm> who would know here?
17:19:52 <mikeperry> micah and dgoulet, I would guess?
17:19:58 <nickm> ok
17:20:26 <mikeperry> dgoulet has some UDP code and some plumbing for prop#340 laying around, but nowhere near merge-read IIUC
17:20:34 <mikeperry> err merge-ready
17:21:53 <dgoulet> I unfortunately don't know the UDP timeline :S ... gaba or ahf would know
17:22:27 <gaba> We were thinking you are going to work on UDP in q1 2023
17:22:32 <mikeperry> perhaps we can try to get some rough idea and then figure it out in the jan roadmap meeting?
17:22:39 <ahf> any time in s101 i believe, q1 2023 should be alright
17:22:46 <ahf> mikeperry: totally
17:23:19 <nickm> sounds good
17:25:06 <ahf> anything else for now?
17:25:06 <mikeperry> ok, I think that covers congestion control and conflux things.. geko+juga: any updates/issues with sbws/onbasca/PT load balancing?
17:25:10 <ahf> bueno
17:26:00 <juga> mikeperry: working to easy deployment on the bridge scanner so that someone can test it
17:26:19 <juga> hopefully we test it this week
17:26:24 <mikeperry> ok awesome
17:26:29 <juga> nothing else from my side
17:26:30 <mikeperry> very nice work
17:26:47 <GeKo> +1
17:27:30 <mikeperry> I think we're good then for this meeting? anything else?
17:27:39 * juga is good
17:27:44 * jnewsome is good
17:27:51 <ahf> \o/
17:27:53 * ahf good
17:28:49 <mikeperry> ok great
17:28:57 <ahf> thanks all <3 last network team meeting for 2022
17:29:04 <ahf> #endmeeting