00:01:24 <teor> #startmeeting Sponsor 55
00:01:24 <MeetBot> Meeting started Thu Feb  6 00:01:24 2020 UTC.  The chair is teor. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:24 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
00:01:31 <ahf> i'm here, but also fading out pretty soon if i need to get up at the time that i usually get up
00:01:55 <teor> Should I start with a status update?
00:02:18 <ahf> WFM
00:02:29 <gaba> https://pad.riseup.net/p/tor-s55-ipv6-project-keep that is the pad
00:02:33 <gaba> for this meeting
00:02:41 <gaba> and yes, status updte please :)
00:02:48 * gaba is also fading
00:03:03 <teor> I spent the last week doing prop#311 and prop#312, which are now merged to torspec as drafts
00:03:14 <teor> zwiebelbot?
00:03:22 <teor> https://gitweb.torproject.org/torspec.git/tree/proposals/311-relay-ipv6-reachability.txt
00:03:28 <teor> https://gitweb.torproject.org/torspec.git/tree/proposals/312-relay-auto-ipv6-addr.txt
00:04:13 <teor> I did a round of revisions on both proposals, after getting feedback from nickm
00:04:25 <ahf> cool
00:04:27 <nickm> (do you need more feedback from me on them?)
00:04:27 <teor> (and some specific ideas from s7r on proposal 312)
00:05:07 <ahf> after my talk at fosdem there was a guy who was really interested in ipv6 and tor and i told him to write to you now that this grant is around, so you might get some random person who wants to help write to you one of the days
00:05:27 <teor> nickm: my next step is to create tickets for all the essential tasks in prop 311 and 312
00:05:48 <gaba> \o/
00:07:03 <teor> The essential tasks are in sections 3, 4 (except 4.4 and 4.5), 5, and 6 of prop 311
00:08:11 <gaba> teor: are you going to review the plans that we had in the pad https://pad.riseup.net/p/JSrYZgX3npU0v4XFp3Ih ?
00:08:23 <teor> And 3 (except 3.4 and 3.5), 4, and 5 of prop 312
00:09:18 <teor> nickm: So if there are issues with the designs in those sections, or you want to change essential/optional for some tasks, please let me know
00:09:27 <teor> ahf: great!
00:09:56 <teor> gaba: I'm going to work from the proposals to create tickets. I might use some of the notes from the pad.
00:10:14 <teor> What do you mean by "review the plans that we had in the pad" ?
00:11:09 <ahf> maybe it is to get the tickets encoded into a roadmap with estimated time on it ?
00:11:09 <gaba> ok
00:11:15 <ahf> i think that depends on ticket creation, no?
00:11:19 <gaba> yes, that is what we need
00:11:22 <nickm> I don't think I can do essential vs not-essential tonight, but I coudl look over it tomorrow?
00:11:36 <teor> nickm: it's not urgent, by early next week will be fine
00:11:40 <nickm> ok
00:12:01 <teor> gaba: I'm still confused. Is it ok if I create tickets from the proposals?
00:12:25 <gaba> yes, go ahead
00:12:39 <teor> Thanks
00:13:25 <teor> As well as creating tickets, I need to write proposal 313, which covers IPv6 statistics and monitoring (Sponsor 55 Objectives 1.4 and 1.5)
00:14:03 <ahf> cool
00:14:15 <teor> I'd like to write the proposal draft, and then create tickets for 311 and 312 while I'm waiting for a review
00:14:18 <ahf> getting the tickets written is very nice so they can be added to the roadmap
00:14:38 <ahf> teor: sounds like a smart order such that the proposal can be reviewed while you create tickets
00:14:54 <teor> Yes, I understand. We made some significant changes to the proposals during the first review cycle, so I think getting a review is important before I make tickets.
00:15:06 <teor> I expect 313 will be much smaller than 311 or 312.
00:15:46 <ahf> ye
00:15:50 <ahf> cool!
00:15:54 <teor> So I think that's about it from me?
00:16:29 <nickm> I don't have much to add.  I'm ready to review 313 when it's out.
00:17:03 <nickm> I think for 312, getting the abstractions and infrastructure right will have a lot to do for making the implementation simple
00:17:25 <teor> There's a few related things I should mention
00:17:57 <teor> ahf is going to set me up with an IPv6 relay, I also have my own personal VPS. 2 relays will be enough to start testing.
00:18:15 <teor> I'll also be testing using chutney (and maybe stem)
00:18:35 <ahf> yep! i got the gnupg moved over today so i could read your email 8)
00:19:46 <ahf> gnupg key*
00:19:51 <teor> :-)
00:20:14 <teor> I've also added a GSoC IPv6 project, in case we get a volunteer who wants to help out
00:20:22 <ahf> great
00:20:25 <teor> My plan is to get them working on the optional changes
00:20:47 <teor> nixkm, ahf: Did pili ask you if you want to be secondary GSoC mentors?
00:20:53 <teor> * nickm
00:21:32 <ahf> i don't remember such question, no
00:21:58 <nickm> i would be okay with being a secondary mentor though
00:22:02 <nickm> especially if ahf would too
00:22:40 <ahf> i am OK with being in the list but will be away some of may or june and 2 weeks(ish) of august
00:23:02 <teor> Ok, I will let pili know
00:23:11 <ahf> thanks
00:23:55 <teor> I don't think there's much more to add
00:24:15 <teor> I really appreciate nickm's feedback, the design is much better now
00:24:34 <gaba> teor: let me know when you start creating tickets then. I want to tag them per objective and track progress there.
00:24:45 <ahf> neat
00:25:09 <teor> gaba: it will probably be next week, I have already done about 4 days work this week
00:25:22 <teor> and I need to draft prop 313 first
00:25:28 <gaba> there are some KPI that we need to track before and after
00:25:52 <gaba> ok
00:26:17 <teor> It looks like most of this work will be merged into Tor 0.4.4, which goes stable in August/September, after Sponsor 55 ends
00:26:35 <teor> So be careful that your KPIs don't depend on deployment of a large number of relays
00:26:36 <gaba> that is ok. Last report is in November
00:27:22 <teor> nickm: Is 1-3 months usually enough for relay operators to upgrade to the next stable release?
00:27:46 <ahf> i think maybe metrics can help there?
00:28:01 <gaba> teor: this is something you reviewed and agreed when we were writing this proposal
00:29:01 <nickm> teor: https://metrics.torproject.org/versions.html
00:29:12 <teor> gaba: I'm just trying to check what's feasible
00:29:42 <teor> When we wrote the proposal, we didn't know when it would be scheduled, in relation to our releases
00:30:35 <nickm> from that it looks like 0.4.2.x got serious adoption within 1.5 months of stable...
00:30:53 <teor> yes, about 50% of relay operators upgraded, that's good
00:31:18 <nickm> if we need a bunch of relays to upgrade, we're fine.  If we need most or all relays to upgrade, that's harder
00:31:33 <teor> gaba: yes, the final report in November should be ok, as long as we get 0.4.4 stable by September (and we get Sponsor 55 in 0.4.4)
00:32:02 <teor> nickm: functionally, we need about 1%, see https://gitweb.torproject.org/torspec.git/tree/proposals/311-relay-ipv6-reachability.txt#n362
00:32:15 <teor> reachability checks are cheap
00:32:47 <nickm> we'll have 1% within a couple of weeks of the first alpha.  If not sooner.
00:32:51 <teor> But for good KPIs, we'd like as many as possible. 50% should be plenty.
00:32:51 <nickm> (I bet)
00:33:02 <teor> nickm: great, that's what we want to find bugs
00:33:34 <teor> I can also do a call out on tor-relays, for testers when we merge significant features to nightly, and when we have our first alpha
00:33:54 <teor> And we can do a reminder when 0.4.4 goes stable
00:34:14 <teor> We should do that anyway, because auto IPv6 may break a small number of relay configs
00:34:26 <teor> Or surprise some operators
00:34:58 <ahf> yeah
00:35:07 <teor> gaba: if we really want to drive adoption, we should change Relay Search, consensus-health, or the metrics graphs to show the new IPv6 relays
00:35:56 <teor> If we have time left over after the essential work is done, we should consider giving some of that time to metrics, to drive adoption
00:36:14 <teor> I'll think about these things as I write proposal 313
00:36:20 <gaba> ok
00:37:24 <gaba> Anything else for today?
00:38:03 <teor> I think that's pretty much it, we'll have more to discuss once we have proposal 313, and some tickets for proposals 311 and 312
00:38:14 <teor> Anyone have anything else?
00:38:20 * ahf is good
00:38:21 <nickm> i'm set
00:39:26 <teor> #endmeeting