[07:32:57] chi.jiun.su leaves the room
[08:37:34] Meetecho joins the room
[08:50:03] Daniel Havey joins the room
[08:50:03] Chris Lemmons joins the room
[08:50:03] David Millman joins the room
[08:50:03] Pyung-Soo Kim joins the room
[08:50:03] Mike Kosek joins the room
[08:50:03] Bernard Aboba joins the room
[08:50:03] Shuai Zhao joins the room
[08:50:03] Sean Turner joins the room
[08:50:04] Yoshifumi Atarashi joins the room
[08:50:19] David Schinazi joins the room
[08:50:19] Alessandro Amirante joins the room
[08:50:23] David Schinazi leaves the room
[08:50:38] Daniel Havey leaves the room
[08:50:52] David Schinazi joins the room
[08:51:39] Alan Frindell joins the room
[08:52:01] <Chris Lemmons> Audio coming through loud and clear.
[08:52:17] Sun Kj joins the room
[08:53:04] Roberto Polli joins the room
[08:54:12] <Sean Turner> whoops  wrong button
[08:54:17] Tommy Pauly joins the room
[08:54:27] <Chris Lemmons> I see that 2 people raised their hands and 1 did not.
[08:54:31] <Shuai Zhao> you can remove me in the queue.
[08:54:45] Victor Vasiliev joins the room
[08:54:46] Barry Leiba joins the room
[08:54:57] Virat Tara joins the room
[08:54:58] <Shuai Zhao> oh, i just removed myself :)
[08:55:01] <Sean Turner> you can also take yourself out :)
[08:55:07] Bernie Innocenti joins the room
[08:55:38] Luke Curley joins the room
[08:56:14] Nick Banks joins the room
[08:56:18] Harald Alvestrand joins the room
[08:56:19] Chi-Jiun Su joins the room
[08:56:30] Virat Tara leaves the room
[08:56:36] Virat Tara joins the room
[08:56:39] <Alan Frindell> I can relay jabber messages
[08:57:06] Mike Kosek leaves the room
[08:57:15] Mike Kosek joins the room
[08:57:23] Eric Orth joins the room
[08:57:39] Mike Kosek leaves the room
[08:57:51] Andrew S joins the room
[08:57:52] Mike Kosek joins the room
[08:57:52] Virat Tara leaves the room
[08:58:06] Eric Kinnear joins the room
[08:58:14] Mohammed Khan joins the room
[08:58:22] Virat Tara joins the room
[08:58:26] Joey Salazar joins the room
[08:58:47] Lohith Bellad joins the room
[08:59:09] alex-meetecho joins the room
[08:59:12] Barry Leiba_ joins the room
[08:59:59] Jake Holland joins the room
[09:00:10] Kazuho Oku joins the room
[09:00:20] Luca Niccolini joins the room
[09:00:22] Mike English joins the room
[09:00:26] Ali Hussain joins the room
[09:00:29] Stephan Wenger joins the room
[09:00:55] Ali Hussain leaves the room
[09:00:58] Dan McArdle joins the room
[09:00:59] Ali Hussain joins the room
[09:00:59] Marten Seemann joins the room
[09:01:04] Dragana Damjanovic joins the room
[09:01:13] Younghan Kim joins the room
[09:01:39] Mirja Kühlewind joins the room
[09:01:40] Yanmei Liu joins the room
[09:01:45] Valery Smyslov joins the room
[09:01:57] frodek joins the room
[09:01:58] Jonathan Lennox joins the room
[09:02:03] Mark Nottingham joins the room
[09:02:03] Ian Swett joins the room
[09:02:15] Jonathan Lennox leaves the room
[09:02:16] Xavier de Foy joins the room
[09:02:19] Jonathan Lennox joins the room
[09:02:21] Kazuho Oku leaves the room
[09:02:27] Bron Gondwana joins the room
[09:02:28] Lucas Pardue joins the room
[09:02:28] Kazuho Oku joins the room
[09:02:49] Kyle Rose joins the room
[09:03:01] Vittorio Bertola joins the room
[09:03:06] Roberto Polli leaves the room
[09:03:20] Lauren Bennett joins the room
[09:03:26] Martin Thomson joins the room
[09:03:33] <Lucas Pardue> Is that David or keanu reeves?
[09:03:46] lnicco@jabb.im joins the room
[09:03:48] <Bron Gondwana> duuuude
[09:04:23] Mike Bishop joins the room
[09:04:24] Tommy Jensen joins the room
[09:04:30] Mihail Yanev joins the room
[09:04:51] Philipp Tiesel joins the room
[09:04:53] <Lucas Pardue> "I know YANG"
[09:04:55] mnot joins the room
[09:05:13] Daniel Havey joins the room
[09:05:19] Ali Hussain leaves the room
[09:05:33] Tommy Jensen (Jabber) joins the room
[09:05:47] <Alan Frindell> Please preface any comments you would like relayed with 'mic:'
[09:05:53] Hiroyuki Goto joins the room
[09:06:16] Xing Jiang joins the room
[09:06:18] Luke joins the room
[09:06:25] Kazuhisa Matsuzono joins the room
[09:06:40] <Eric Kinnear> https://codimd.ietf.org/notes-ietf-109-webtrans?both
[09:07:08] Ali Hussain joins the room
[09:07:10] frodek leaves the room
[09:07:21] <Ian Swett> Whoa...
[09:07:55] Luke leaves the room: Disconnected: BOSH client silent for over 60 seconds
[09:09:34] Jana Iyengar joins the room
[09:09:48] Mike Kosek leaves the room
[09:13:46] <Alan Frindell> Doesn't multiplexing and priority sounds like...HTTP/3?
[09:14:05] <Martin Thomson> alan no it doesn't, that's just quic
[09:14:14] <Ian Swett> Agreed MT
[09:14:18] <Marten Seemann> What's the name of the QUIC implementation?
[09:14:24] ihlar joins the room
[09:14:44] <Alan Frindell> QUIC has priority ? ;)
[09:14:45] <mnot> Ad: Chrome + Fastly are experimenting with 103 Early Hints as an alternative to Server Push…
[09:15:06] <Martin Thomson> alan: yes, it doesn't have priority signaling, but you have to prioritize for the multiplexing not to suck
[09:16:09] <Jana Iyengar> The app would prioritize. Whatever they do has got to be better than what we had with HTTP priorities.
[09:16:09] <Ian Swett> Yesterday would be the right time on early hints, but we are where we are.
[09:16:15] <Kazuho Oku> primitive prioritization = pure server-driven prioritazation would work well if you have good knowledge of the application.
[09:16:34] <Ian Swett> (hi QUIC WG &lt;wave&gt;)
[09:16:42] Lauren Bennett leaves the room
[09:17:17] <Jana Iyengar> Ah, so no priorities yet
[09:17:54] <Martin Thomson> no ability to tell Chrome how to prioritize sending; we have the same limitations
[09:18:52] <Alan Frindell> mt: you mean for the API to tell Chrome, or for the server?
[09:20:01] <Jana Iyengar> @Alan: I believe this is the API to Chrome
[09:20:13] <Martin Thomson> the API to tell chrome what to do
[09:20:31] <Alan Frindell> Are they trying to set the server priority for its egress streams, or for streams chrome is sending, or both?
[09:20:43] <Martin Thomson> that is the only relevant API, priority signaling of the sort you are talking about for h3 is the client telling the server stack what to do, which isn't always necessary
[09:20:59] <Martin Thomson> presumably the server knows how to prioritize properly for their use case
[09:21:32] <Luke Curley> we're prioritizing when we control the sender implementation (ex. distribution)
[09:22:04] Ali Hussain leaves the room
[09:22:07] <Luke Curley> but if we wanted to use the browser to send media, there's no mechanism to do that
[09:22:12] <Alan Frindell> luke: i wanted to ask - what classes of stream types are you using primarily for your application?
[09:22:20] <Martin Thomson> you do that based on what you know about the content, Luke?
[09:23:36] <Jana Iyengar> You don't need app layer acks for datagrams if you have loss/ack notification as part of the API, right?
[09:23:50] Jana Iyengar leaves the room
[09:24:48] <Martin Thomson> wow, that was a serious hitch in the chat
[09:25:02] <Lucas Pardue> that seems to be a bone of contention jana
[09:25:07] <mnot> Regarding 'Will Http3Transport be supported by CDNs' — it depends on what abstractions you want to have exposed. If you just want them to forward streams 1:1, sure. Beyond that, there's not a lot to grab onto…
[09:25:23] <Martin Thomson> say more Lucas?
[09:25:34] <Lucas Pardue> sec
[09:25:46] <Luke Curley> worth mentioning that we went down the WebRTC route, both media and data channels
[09:25:49] <Martin Thomson> mnot: yep; because you need application protocols for a CDN to really do something
[09:25:58] <Luke Curley> QuicTransport is better for our use-case
[09:26:04] <Lucas Pardue> see https://github.com/w3c/webtransport/issues/168
[09:26:04] ihlar joins the room
[09:26:18] <Chris Lemmons> Yeah, it'd be interesting to know what people mean when they say, "Supported by CDNs."
[09:26:25] <mnot> WebTransport is all about exposing low, low level primitives. You can't have it both ways...
[09:26:39] <Mike Bishop> Luke, I was wanting to ask -- since QUIC is coming to the end of last call, did you encounter anything that would be feedback to the core protocol?
[09:27:03] <mnot> (why is the video window frame… pulsating?)
[09:27:14] <Lucas Pardue> there's a github issue, jabber won't let me paste it?
[09:27:23] ihlar leaves the room
[09:27:27] <Ian Swett> Anyone know what an "Event" is for the purpose of these slides?
[09:29:16] <Eric Kinnear> @Jana, I had the same thought -- if you've got loss vs. ack in the API then the ack-eliciting behavior of a datagram ought to eliminate that particular need for application layer acks
[09:30:02] <Daniel Havey> If UDP is blocked then you must fallback to TCP.
[09:31:13] <Chris Lemmons> Chat seems to be having issues. :/
It'd be useful to understand what people mean when they say that they want "CDNs to support WebTranport."
[09:31:29] 一穂 奥 joins the room
[09:31:55] Tommy Pauly leaves the room
[09:31:55] Barry Leiba leaves the room
[09:31:55] Victor Vasiliev leaves the room
[09:31:55] Kazuhisa Matsuzono leaves the room
[09:31:55] Stephan Wenger leaves the room
[09:31:55] Eric Orth leaves the room
[09:31:55] Sean Turner leaves the room
[09:31:55] Virat Tara leaves the room
[09:31:55] Mark Nottingham leaves the room
[09:31:55] Xavier de Foy leaves the room
[09:31:55] Bron Gondwana leaves the room
[09:31:55] Jonathan Lennox leaves the room
[09:31:55] Mihail Yanev leaves the room
[09:31:55] Martin Thomson leaves the room
[09:31:55] Kyle Rose leaves the room
[09:31:55] Tommy Jensen leaves the room
[09:31:55] Younghan Kim leaves the room
[09:31:55] Philipp Tiesel leaves the room
[09:31:55] Bernie Innocenti leaves the room
[09:31:55] David Millman leaves the room
[09:31:55] Pyung-Soo Kim leaves the room
[09:31:55] Andrew S leaves the room
[09:31:55] Daniel Havey leaves the room
[09:31:55] Sun Kj leaves the room
[09:31:55] Yoshifumi Atarashi leaves the room
[09:31:55] Mohammed Khan leaves the room
[09:31:55] Harald Alvestrand leaves the room
[09:31:55] Alan Frindell leaves the room
[09:31:55] Nick Banks leaves the room
[09:31:55] Joey Salazar leaves the room
[09:31:55] Lucas Pardue leaves the room
[09:31:55] Chi-Jiun Su leaves the room
[09:31:55] Valery Smyslov leaves the room
[09:31:55] Marten Seemann leaves the room
[09:31:55] Alessandro Amirante leaves the room
[09:31:55] Luca Niccolini leaves the room
[09:31:55] Luke Curley leaves the room
[09:31:55] Eric Kinnear leaves the room
[09:31:55] Mike English leaves the room
[09:31:55] David Schinazi leaves the room
[09:31:55] Vittorio Bertola leaves the room
[09:31:55] Hiroyuki Goto leaves the room
[09:31:55] Dan McArdle leaves the room
[09:31:55] Dragana Damjanovic leaves the room
[09:31:55] Ian Swett leaves the room
[09:31:55] Chris Lemmons leaves the room
[09:31:55] Xing Jiang leaves the room
[09:31:55] Shuai Zhao leaves the room
[09:31:55] Bernard Aboba leaves the room
[09:31:55] Mirja Kühlewind leaves the room
[09:31:55] Yanmei Liu leaves the room
[09:31:55] Lohith Bellad leaves the room
[09:31:55] Jake Holland leaves the room
[09:31:55] Kazuho Oku leaves the room
[09:31:55] Mike Bishop leaves the room
[09:32:28] lpardue joins the room
[09:32:59] <mnot> wait, what?
[09:33:02] <mnot> many people in the queue
[09:33:21] <mnot> did we segment again?
[09:33:32] <Meetecho> Investigating
[09:33:57] alficles joins the room
[09:34:10] <lpardue> I was in the queue
[09:34:31] Jonathan Lennox (Jabber) joins the room
[09:34:37] <mnot> I was in the queue. The chat window and my jabber client say wildly different things.
[09:34:49] <mnot> I hit reload and now it's hanging.
[09:34:54] <alficles> Hrm, messages I have sent appear here, but aren't appearing in the chat in Meetecho.
[09:35:08] <Jonathan Lennox (Jabber)> Yeah, I also see the chat window in Meetecho stuck at 4:25
[09:35:11] <Jonathan Lennox (Jabber)> (EST)
[09:35:22] <alficles> Ditto.
[09:35:23] <mnot> well now I can't get audio or video
[09:36:30] <mnot> nope, it's dead
[09:36:51] <Jonathan Lennox (Jabber)> I still have audio and video from Meetecho, but not chat
[09:36:53] <Meetecho> We're trying to fix that ASAP, sorry
[09:37:10] <Meetecho> Yes, chat seems very lagged behind, but those still in can still see audio/video fine
[09:37:13] <mnot> if you still have audio/video, don't hit reload
[09:37:20] <Meetecho> There's problems getting back in again that we're working on
[09:37:26] <mnot> 'unable to retrieve room's information'
[09:37:38] Mirja Kuehlewind joins the room
[09:39:12] Jana Iyengar joins the room
[09:39:55] <Jana Iyengar> I'm in the same boat as mnot. I hit reload to get chat back, and I'm completely out now
[09:40:36] <Jana Iyengar> I have audio with the audio stream, but I can't participate
[09:41:00] <Jonathan Lennox (Jabber)> Well, we can relay with "mic:" if you need
[09:41:19] <mnot> We have no context for what's happening in the meeting.
[09:41:24] <mnot> I'm going to e-mail.
[09:41:29] Nasrul joins the room
[09:41:42] <Jana Iyengar> mnot — use the audio channel, that's something, but yeah, this is quite bad
[09:45:28] <mnot> What slides are we looking at, and what number are we on?
[09:45:45] <Jonathan Lennox (Jabber)> "PR#4: headers", slide 24
[09:45:47] <mnot> And could we ask speakers to say what slide number they're on for each one?
[09:45:49] <mnot> thx
[09:46:00] <Jonathan Lennox (Jabber)> Single slide deck it looks like
[09:47:07] <mnot> got it, thanks!
[09:47:29] <Jonathan Lennox (Jabber)> Slide 25, "Discussion"
[09:47:54] dschinazi@jab.im joins the room
[09:47:55] afrind joins the room
[09:48:58] <mnot> I sent my comments to the mailing list.
[09:49:24] alficles leaves the room
[09:49:25] alficles joins the room
[09:49:32] alficles leaves the room
[09:49:32] alficles joins the room
[09:49:34] alficles leaves the room
[09:49:43] alficles joins the room
[09:49:59] <Jonathan Lennox (Jabber)> Slide 26, "WebTransport using HTTP/2"
[09:51:00] ekinnear joins the room
[09:51:09] <Jonathan Lennox (Jabber)> 2 (apparently separate slide numbering?), "Since IETF 108"
[09:51:17] <mnot> _nice_ :)
[09:51:20] <ekinnear> Meetecho seems to have kicked me out
[09:51:28] <mnot> welcome
[09:51:29] <Jonathan Lennox (Jabber)> Alan is presenting your slides
[09:51:33] <Jana Iyengar> welcome eric
[09:51:35] <ekinnear> Excellence!
[09:51:39] <ekinnear> Thanks Alan :)
[09:51:42] <Jonathan Lennox (Jabber)> "#5 Unidirectional streams"
[09:51:49] englishm joins the room
[09:51:49] <Mirja Kuehlewind> you can join the audio eric
[09:51:58] <Jana Iyengar> the audio stream works, so you can listen to that
[09:51:59] <mnot> slide 3 (after slide 26)
[09:51:59] <ekinnear> TLDR: “Can we leave the zoo” :D
[09:52:14] <Jonathan Lennox (Jabber)> "#3 New streams without additional roundtrips"
[09:52:59] <ekinnear> Sneaky, thanks — audio stream is working
[09:53:27] <Jonathan Lennox (Jabber)> "#6 Datagrams"
[09:54:02] <Jonathan Lennox (Jabber)> "What's next?"
[09:54:16] <ekinnear> Read faaaaast :)
[09:55:11] <mnot> From Alissa: "There is a problem with an authentication VM that means it needs a reboot, which will kill all sessions for five minutes or so."
[09:55:50] <mnot> five minutes, not two minutes
[09:56:20] <Jonathan Lennox (Jabber)> Slide 9, "Questions"
[09:56:22] <ekinnear> +1, where are we going
[09:56:36] <ekinnear> (And can we leave the zoo)
[09:56:38] <Jonathan Lennox (Jabber)> I can relay for anyone following the audio stream
[09:57:48] Yanmei Liu joins the room
[09:57:48] Yanmei Liu leaves the room
[09:57:58] Xavier de Foy joins the room
[09:57:58] Tommy Jensen joins the room
[09:57:58] Xavier de Foy leaves the room
[09:57:58] Tommy Jensen leaves the room
[09:57:58] Alan Frindell joins the room
[09:57:58] Alan Frindell leaves the room
[09:57:58] Virat Tara joins the room
[09:57:58] Virat Tara leaves the room
[09:57:59] Barry Leiba joins the room
[09:57:59] Barry Leiba leaves the room
[09:57:59] Harald Alvestrand joins the room
[09:57:59] Joey Salazar joins the room
[09:57:59] Harald Alvestrand leaves the room
[09:57:59] Joey Salazar leaves the room
[09:57:59] Jonathan Lennox joins the room
[09:57:59] Jonathan Lennox leaves the room
[09:57:59] Mirja Kühlewind joins the room
[09:57:59] Mirja Kühlewind leaves the room
[09:57:59] Alessandro Amirante joins the room
[09:57:59] Alessandro Amirante leaves the room
[09:57:59] Chris Lemmons joins the room
[09:57:59] Chris Lemmons leaves the room
[09:57:59] Yoshifumi Atarashi joins the room
[09:57:59] Yoshifumi Atarashi leaves the room
[09:58:00] Daniel Havey joins the room
[09:58:00] Daniel Havey leaves the room
[09:58:00] Shuai Zhao joins the room
[09:58:00] Shuai Zhao leaves the room
[09:58:00] Bron Gondwana joins the room
[09:58:00] David Schinazi joins the room
[09:58:00] Bron Gondwana leaves the room
[09:58:00] David Schinazi leaves the room
[09:58:01] Jake Holland joins the room
[09:58:01] Jake Holland leaves the room
[09:58:01] Chi-Jiun Su joins the room
[09:58:01] Mohammed Khan joins the room
[09:58:01] Chi-Jiun Su leaves the room
[09:58:01] Mohammed Khan leaves the room
[09:58:01] Hiroyuki Goto joins the room
[09:58:01] Hiroyuki Goto leaves the room
[09:58:01] Bernard Aboba joins the room
[09:58:01] Bernard Aboba leaves the room
[09:58:02] Lucas Pardue joins the room
[09:58:02] Lucas Pardue leaves the room
[09:58:02] Andrew S joins the room
[09:58:02] Andrew S leaves the room
[09:58:03] Nick Banks joins the room
[09:58:03] Nick Banks leaves the room
[09:58:03] Mike Bishop joins the room
[09:58:03] Mike Bishop leaves the room
[09:58:09] Luke Curley joins the room
[09:58:09] Luke Curley leaves the room
[09:58:10] Philipp Tiesel joins the room
[09:58:10] Philipp Tiesel leaves the room
[09:58:10] Dragana Damjanovic joins the room
[09:58:10] Dragana Damjanovic leaves the room
[09:58:11] Yanmei Liu joins the room
[09:58:11] Yanmei Liu leaves the room
[09:58:11] Kazuho Oku joins the room
[09:58:12] Kazuho Oku leaves the room
[09:58:13] Mihail Yanev joins the room
[09:58:13] Mihail Yanev leaves the room
[09:58:15] Mike English joins the room
[09:58:15] Mike English leaves the room
[09:58:16] Pyung-Soo Kim joins the room
[09:58:16] Pyung-Soo Kim leaves the room
[09:58:16] Marten Seemann joins the room
[09:58:16] Marten Seemann leaves the room
[09:58:19] Bernie Innocenti joins the room
[09:58:19] Bernie Innocenti leaves the room
[09:58:21] Xing Jiang joins the room
[09:58:21] Xing Jiang leaves the room
[09:58:22] Eric Orth joins the room
[09:58:22] Eric Orth leaves the room
[09:58:22] David Millman joins the room
[09:58:22] David Millman leaves the room
[09:58:23] Younghan Kim joins the room
[09:58:23] Younghan Kim leaves the room
[09:58:24] Tommy Pauly joins the room
[09:58:24] Tommy Pauly leaves the room
[09:58:25] Vittorio Bertola joins the room
[09:58:25] Vittorio Bertola leaves the room
[09:58:31] Stephan Wenger joins the room
[09:58:31] Stephan Wenger leaves the room
[09:58:33] <Jana Iyengar> The purge!
[09:58:35] Martin Thomson joins the room
[09:58:35] Martin Thomson leaves the room
[09:58:48] <Jonathan Lennox (Jabber)> As is likely evident from the joins/leaves here, the threatened meetecho reboot has happened
[09:58:51] <alficles> Sounds like we're trying "turning it off and back on again". :)
[09:59:07] <ekinnear> Thanks Jonathan :)
[09:59:11] Victor Vasiliev joins the room
[09:59:12] Bernard Aboba joins the room
[09:59:12] Ali Hussain joins the room
[09:59:12] Hiroyuki Goto joins the room
[09:59:23] Alan Frindell joins the room
[09:59:24] Eric Kinnear joins the room
[09:59:26] Alessandro Amirante joins the room
[09:59:34] englishm refills coffee
[09:59:35] Frédéric Fieau joins the room
[09:59:43] Kazuho Oku joins the room
[09:59:43] Jana Iyengar_820 joins the room
[09:59:44] Martin Thomson joins the room
[09:59:44] Dan McArdle joins the room
[09:59:49] Kazuaki Ueda joins the room
[09:59:51] David Schinazi joins the room
[09:59:58] Mike Bishop joins the room
[09:59:58] Mark Nottingham joins the room
[10:00:03] <Martin Thomson> I think we are back
[10:00:05] <Meetecho> It should be working again now, truly sorry for the issue...
[10:00:09] <dschinazi@jab.im> Yes, the VM is currently rebooting
[10:00:10] <Jana Iyengar_820> Yes, David
[10:00:19] Mike English joins the room
[10:00:24] Sean Turner joins the room
[10:00:24] Nicolas Kuhn joins the room
[10:00:37] Virat Tara joins the room
[10:00:37] Marcus Ihlar joins the room
[10:00:38] Tommy Pauly joins the room
[10:00:51] <ekinnear> What is even happening to this audio
[10:00:52] Stephan Wenger joins the room
[10:01:01] Mihail Yanev joins the room
[10:01:08] <Jana Iyengar_820> Thank you @meetecho -- I'm happy to be back in the room!
[10:01:15] Yoshifumi Atarashi joins the room
[10:01:21] Bron Gondwana joins the room
[10:01:30] Luke Curley joins the room
[10:01:37] Andrew S joins the room
[10:01:39] Nick Banks joins the room
[10:01:43] Lucas Pardue joins the room
[10:01:53] Luca Niccolini joins the room
[10:02:16] Barry Leiba joins the room
[10:02:56] Valery Smyslov joins the room
[10:02:57] Jake Holland joins the room
[10:03:00] Kentaro Teramoto joins the room
[10:03:02] <ekinnear> All better now, thank you @meetecho!
[10:03:20] Mirja Kühlewind joins the room
[10:03:42] Marten Seemann joins the room
[10:04:19] Jonathan Lennox joins the room
[10:05:18] Josh Cohen joins the room
[10:07:44] <Lucas Pardue> wonder who has the faster fan, Bernard or Daniel?
[10:08:20] Robin Marx joins the room
[10:08:36] <dschinazi@jab.im> HAHA
[10:08:37] Yanmei Liu joins the room
[10:09:53] <dschinazi@jab.im> I felt left out so I got one too
[10:10:22] <Lucas Pardue> fan-tastic
[10:10:36] <Mike Bishop> Fancy!
[10:10:39] <dschinazi@jab.im> slow clap
[10:11:41] Chi-Jiun Su joins the room
[10:13:42] <Jana Iyengar_820> Has anyone used any of the other WebTransport variants besides QuicTransport?
[10:14:07] <Alan Frindell> We have experience with something very like the H2 Transport
[10:14:32] <Jana Iyengar_820> priorities!
[10:14:34] <ekinnear> Probably unsurprisingly, I’ve used H2Transport in a totally different context but with pretty much the same set of needs
[10:16:11] Philipp Tiesel joins the room
[10:16:14] <Alan Frindell> Why is this a separate transport and not an option: let transport = new WebTransport(&lt;uri&gt;, &lt;pooled&gt;)
[10:16:15] <Mirja Kühlewind> I thought that's what http2transport kind of already gives you...
[10:16:18] <Martin Thomson> why would Bernard even need a fan?
[10:17:07] <Philipp Tiesel> Has anyone thought about re-adding pooling for QUIC Transport by using QUIC?
[10:17:32] <mnot> Maybe VIctor also has a fan, going at *exactly* the right speed
[10:17:34] <Martin Thomson> Philipp: you can't get shared use of connections between WebTransport and HTTP at that point
[10:17:50] <Philipp Tiesel> sorry… s/QUIC\?$/MASQ?/
[10:18:28] <Martin Thomson> Oh, so you put the QUIC over the same connection as HTTP.  That might work, yeah.
[10:18:49] <dschinazi@jab.im> That would work but it has the same downside as h3transport, you'd still require the server to implement h3
[10:18:51] <Philipp Tiesel> yep - that's the idea
[10:20:35] <Philipp Tiesel> So for QUIC Transport over MASQ, the proxy/lb needs to, but the backend wouldn't
[10:21:10] <Mike English> I'm a fan of Victor's presentation
[10:21:17] <dschinazi@jab.im> Yes, but most of the server scenarios I've talked to didn't want to load balance WebTransport
[10:21:47] <Philipp Tiesel> Do these need pooling?
[10:21:57] <Lucas Pardue> I think you could adapt Victor's "abstract headers" model in order to have a QUIC-level MASQUE initiation/management message that is used in a style of the CONNECT-UDP QUIC-aware mode. In other words, a low level message to "pool" mutliple QUIC connections and avoid the overhead of double encryption
[10:22:35] Mo Zanaty joins the room
[10:22:49] <dschinazi@jab.im> That wouldn't pool in congestion control terms though
[10:22:54] <Lucas Pardue> it pools fate sharing but not congestion control
[10:22:55] Mike English leaves the room
[10:23:00] Mike English joins the room
[10:23:02] <Lucas Pardue> yeah :)
[10:23:19] Eric Rescorla joins the room
[10:24:35] Paolo Saviano joins the room
[10:24:58] <Lucas Pardue> called it MUXQUE
[10:25:08] Alissa Cooper joins the room
[10:25:21] alficles leaves the room
[10:25:21] <dschinazi@jab.im> MUXIQUE?
[10:27:04] <Mike Bishop> You can't use 0-RTT across ALPNs, so that's not totally accurate.
[10:27:14] <ekinnear> Luke: Why do they need to transport a lot of data for that to be a benefit? Isn’t the reuse also a RTT count benefit?
[10:27:25] <ekinnear> And right, I think we need an API that can express that
[10:27:29] Jonathan Lennox (Jabber) leaves the room
[10:28:01] <Jonathan Lennox> This is sounding notably TAPS-ish. :-)
[10:28:24] <Lucas Pardue> webtaps
[10:28:24] <Mike Bishop> It is, a bit.  Also, the presenter doesn't need to join the queue to respond.  ;-)
[10:28:28] <Mike English> ^ My thoughts as well
[10:29:17] <Luke Curley> ekinnear: 0-RTT would remove the latency cost of dialing a new QUIC connection to the same host
[10:29:27] <Jana Iyengar_820> Yes - that is exactly what I've been thinking.
[10:29:37] <Jana Iyengar_820> ^ that = webtaps
[10:30:10] <ekinnear> Luke: I’m not sure that holds all the time here, especially see Mike’s comment about 0-RTT across ALPNs and mux-ing the traffic
[10:30:36] <ekinnear> May not be the most important part here, just trying to think through what’s happening underneath
[10:30:41] <Luke Curley> yeah, the benefit of connection pooling that I didn't mention is making sure both connections are routed to the same host
[10:31:36] <Lucas Pardue> the first use case I had for webtaps arose from the multipath QUIC discussion
[10:31:55] Harald Alvestrand joins the room
[10:32:38] <Martin Thomson> So I see ~zero value from intermediation and pooling right now.
[10:32:47] <Philipp Tiesel> I am not sure whether you really need do have "webtaps" – I think just a JS interface for regular TAPS with different defaults should work
[10:33:08] <Tommy Pauly> Right webtaps is just one language binding of TAPS
[10:33:18] <Martin Thomson> And MUXQUE might be the way we get things like shared congestion state out of this... maybe
[10:33:28] <ekinnear> +1 let’s just have one, the rest of this is a bit messy but let’s for sure not try to do them all at the same time
[10:33:34] <Lucas Pardue> web platform != JavaScript
[10:34:53] <mnot> @Lucas: are you sure?
[10:35:20] <Lucas Pardue> oh is it !=== ?
[10:35:31] <Philipp Tiesel> Lucas: right – but that does not invalidiate the one TAPS argument
[10:35:41] <Mike English> :P
[10:35:50] <dschinazi@jab.im> What's the TAPS argument?
[10:36:46] <Martin Thomson> I think that it would be really interesting to build a pubsub protocol for QUIC.  It seems like that is always step 2 here.
[10:36:46] <Jonathan Lennox> I was thinking in terms of the semantics to say "prefer this feature / require this feature" and find out what happened
[10:36:50] Nasrul leaves the room: Disconnected: BOSH client silent for over 60 seconds
[10:37:07] <Martin Thomson> Jonathan, use getUserMedia as an example of what NOT to do
[10:37:15] <Mike English> negotiation?
[10:37:19] <Jonathan Lennox> I'm sure we have no shortage of bad examples
[10:37:31] <Philipp Tiesel> dschinazi: That there should be no need to define "webtaps" –  TAPS without modifications should work fine in a web context
[10:37:35] Tommy Jensen (Jabber) leaves the room
[10:37:55] <Mike English> also, re webtaps... web platform != browser maybe?
[10:38:18] <Mike English> Have non-browser wasm use cases been considered?
[10:38:34] <Martin Thomson> thanks alan.  I can see how that pushes for something you can push into HTTP
[10:38:41] Ali Hussain leaves the room
[10:39:00] Ali Hussain joins the room
[10:39:05] <Martin Thomson> I think that is something of an exceptional circumstance, unfortunately
[10:39:12] alficles joins the room
[10:39:26] <dschinazi@jab.im> @Philipp what is TAPS in this context? My limited understanding of TAPS is that it's a way to think about protocols. I don't know what TAPS means in a Web context in practice
[10:39:44] <Lucas Pardue> well I'm not going to be able to use the TAPS API to start a server inside a browser am I? Or create a client connection to any old server
[10:39:59] <Alan Frindell> Note that websockets *can* be multiplexed with HTTP!
[10:40:28] <Martin Thomson> RFC 8441 ?
[10:40:38] Ali Hussain leaves the room
[10:40:45] <Alan Frindell> yes
[10:40:48] Ali Hussain joins the room
[10:41:14] <Martin Thomson> So would QuicTransport + MASQUE address your concerns?
[10:41:35] <Victor Vasiliev> That... sounds like an interesting construct
[10:41:37] Ali Hussain leaves the room
[10:41:38] Tommy Pauly leaves the room
[10:41:44] Ali Hussain joins the room
[10:41:50] <Alan Frindell> I need to think about it more
[10:41:51] <Jana Iyengar_820> @MT: QuicTransport + HTTP/3 within MASQUE?
[10:41:56] <Martin Thomson> That is, CONNECT-QUIC specifically.
[10:42:12] <Martin Thomson> CONNECT-UDP is workable, but probably suboptimal.
[10:42:23] <Martin Thomson> This is not my idea, credit to Philipp
[10:42:33] Ali Hussain leaves the room
[10:42:40] Ali Hussain joins the room
[10:42:51] <Jana Iyengar_820> Yeah - you want to terminate the QUIC connection; but I think CONNECT-QUIC ought to work
[10:42:58] <Lucas Pardue> I think you'd want a way to describe that the QUIC server terminates
[10:43:00] <Philipp Tiesel> dschinazi: The TAPS idea – for me – is you specify your endpoints and requirements/preferences and get the best transport available – that set of available transports will change in web conext, but the principle API should still work
[10:43:05] Jiao Kang joins the room
[10:43:05] Mihail Yanev leaves the room
[10:43:10] <Victor Vasiliev> CONNECT-QUIC does not dispatch based on URL, though, it dispatches based on network endpoints
[10:43:11] Mihail Yanev joins the room
[10:43:12] <Martin Thomson> The idea there is basically that most people want something simple: QuicTransport.  Those who also need multiplexing can MASQUE it.
[10:43:14] Stephan Wenger leaves the room
[10:43:14] Tommy Pauly joins the room
[10:43:17] <dschinazi@jab.im> @Philipp except you can't let the Web do that
[10:43:43] <Jonathan Lennox> Well, "available" is a more restricted subset in a web context
[10:43:48] <Martin Thomson> Victor: yes, you lose multiplexing of WebTransport sessions at the connection-level, but the cost of a new connection inside of MASQUE is pretty low
[10:44:04] <dschinazi@jab.im> Because the Web has a clear and specific security model that doesn't let you use any protocol in any way
[10:44:05] <Alan Frindell> Why is that better than H3 transport?
[10:44:06] Ali Hussain leaves the room
[10:44:17] <Mike English> To clarify - Philipp, you mean specifying properties like this? https://tools.ietf.org/html/draft-ietf-taps-interface-10#section-4.2
[10:44:19] <Mike Bishop> Can I suggest that MASQUE idea come up to the main discussion?
[10:44:22] Ali Hussain joins the room
[10:44:22] <Martin Thomson> Because it is a progressive complexity story.
[10:44:36] <Jana Iyengar_820> +1 to Mike Bishop
[10:44:47] <Jana Iyengar_820> I think that is a very interesting take on pooling
[10:44:49] <Martin Thomson> If I have requirements like Luke (which are, in my experience, overwhelming in the majority), then I can just do QuicTransport.
[10:44:50] alficles leaves the room
[10:45:13] <Alan Frindell> Is the complexity concern the pooling capability in particular (eg: 1 stream to define the session)
[10:45:15] <Martin Thomson> If I also need this to flow through the same path/flow/intermediary/whatever then I do the next layer.
[10:45:38] <Martin Thomson> The complexity is that you need h3.  Not so much the multiplexing piece.
[10:46:07] <Lucas Pardue> If we want a low-level CONNECT-UDP mechanism, we need to remember why MASQUE decided to use a HTTP/3 mechanism in the first place.
[10:46:27] <Martin Thomson> which of those do you think is relevant Lucas?
[10:46:47] <ekinnear> Time for a meetecho hum-vote?
[10:46:55] <Lucas Pardue> I can't remember all the reasons, I'm sure that is a matter of record though
[10:47:03] <Martin Thomson> HIGHLANDER!
[10:47:04] <Victor Vasiliev> I have a question before the hum
[10:47:22] Kazuaki Ueda leaves the room
[10:47:25] Kazuaki Ueda joins the room
[10:47:26] <Lucas Pardue> HTTP-level authentication / access control is a decent reason
[10:47:29] Ali Hussain leaves the room
[10:48:01] Ali Hussain joins the room
[10:48:11] <Mirja Kühlewind> i think the main reason was to actually masque any traffic within web traffic...
[10:49:15] <Lucas Pardue> with ECH, is there a way for a passive observer to distibguish H3 from anything-over-QUIC?
[10:49:19] <Kazuho Oku> I kind of liked Victor's proposal to support both (by letting the server signal if the connection can be pooled)
[10:49:55] xdf joins the room
[10:50:26] <Alan Frindell> So the MASQUE option is: H3 connection, use a CONNECT stream to establish each "session"?
[10:50:41] Jiao Kang leaves the room
[10:50:50] <ekinnear> That’s my understanding, Alan
[10:50:53] <Martin Thomson> David: not necessarily double-encryption, depending on how you tunnel
[10:51:00] <Mirja Kühlewind> Lucas I think the point was to avoid a dependency on ECH
[10:51:06] Ali Hussain leaves the room
[10:51:06] <Eric Rescorla> I'm not sure how enthusiastic I am about this MASQUE layering
[10:51:07] <Victor Vasiliev> For what it's worth, I believe the hum is premature
[10:51:16] Ali Hussain joins the room
[10:51:20] <Jana Iyengar_820> If you want the datagrams and congestion control sharing, you need double enc... I think
[10:51:27] <Alan Frindell> Do the sub-streams of that CONNECT stream get first-order stream support in the H3 connection?  eg: no head of line blocking?
[10:51:44] <mnot> uh, what?
[10:51:45] <Kazuho Oku> Think of CONNECT-UDP over QUIC, there's no reason for double encryption
[10:51:47] <mnot> I see no interface for this
[10:52:05] Ali Hussain leaves the room
[10:52:10] Ali Hussain joins the room
[10:52:11] <Jonathan Lennox> "Show of hands tool", to the left of the chat and participants
[10:52:21] <Robin Marx> looks like a barchart icon
[10:53:11] <Philipp Tiesel> Mike: jepp - but as Jonathan said – the set of possible outcomes of fulfilling these will be very limited by the security model
[10:53:33] Kentaro Teramoto leaves the room
[10:53:34] <Martin Thomson> We should have a voting tool that has multiple options that can be edited.
[10:54:18] Bron Gondwana leaves the room
[10:54:18] <mnot> That's crazy talk, MT
[10:54:29] <Jonathan Lennox> I'm also perplexed that the hands tool says "57 participants" whereas the participants tab says 73 participants
[10:54:29] Ali Hussain leaves the room
[10:54:36] Ali Hussain joins the room
[10:54:37] <Lucas Pardue> we can call it the Diebold voting machine
[10:54:46] <Mike Bishop> Participants tab shows Jabber users.
[10:54:49] <mnot> I suspect it counts people who are logged into jabber separately too
[10:54:51] <Mirja Kühlewind> there are jabber only people in the list
[10:54:56] <Jonathan Lennox> Aha
[10:55:04] <mnot> Does that mean I get two votes?
[10:55:08] <Mirja Kühlewind> no
[10:55:14] <Mirja Kühlewind> there are no votes
[10:55:22] <Mike English> hmmmm
[10:55:23] <Jana Iyengar_820> @jonathan: also, presumably not everyone voted.
[10:55:32] <mnot> Not fun!
[10:55:38] <Jonathan Lennox> Yes, that's why the 19/10/57
[10:55:47] <Jana Iyengar_820> Wait, I just voted, why are there no votes?
[10:55:56] <mnot> It's a result, not a vote
[10:56:06] <Jana Iyengar_820> But I voted!
[10:56:14] <mnot> It's an appendage altitude test.
[10:56:16] <Lucas Pardue> so is the number that didn't raise their hand reall (57 - 19)?
[10:56:28] <Mirja Kühlewind> no you didn't because there is no decision
[10:56:41] <Eric Rescorla> @Lukas: I think, yes
[10:56:41] <Mike English> Jabber-only people have no hands to raise
[10:56:43] <Jonathan Lennox> 10 people clicked "don't raise your hand", whereas (57-19) clicked nothing
[10:56:47] <mnot> +1 lucas
[10:56:55] <mnot> but they didn't raise their hands
[10:56:56] <Mirja Kühlewind> yes lucas the people sleeping are 38
[10:57:05] <Mike Bishop> @Lucas:  57 minus 29, you mean.
[10:57:05] <Jonathan Lennox> Or who haven't made up their mind
[10:57:11] <Lucas Pardue> we need haptics!
[10:57:12] dragana joins the room
[10:57:23] <Kazuho Oku> zzzzz
[10:57:29] <mnot> Lucas wins jabber tonight; thank you everyone for playing.
[10:57:36] <Martin Thomson> we have 3 options: QUIC, H3, BOTH.  How do we pick A or B?
[10:57:42] <Mirja Kühlewind> no 29 are awake, otherwise do mails or went to bed or listen to another session or something
[10:57:57] <Jana Iyengar_820> +1 to Alan
[10:58:20] <Victor Vasiliev> My main objection to Ekr's idea of "we have to pick one", is that we tried to pick one, and we've not picked a decisive winner, because there isn't really one
[10:58:52] <Mirja Kühlewind> flip a coin or close the group...?
[10:59:02] <Alan Frindell> +1 ekr
[10:59:23] <Mike Bishop> Flip a coin to decide whether to close the group?
[10:59:25] <ekinnear> I’m not sure I parsed the last bit of that Victor — is Alan correct in that H3Transport offers a superset of QuicTransport?
[10:59:29] <Lucas Pardue> Would anyone refuse to implement H3transport because it's too hard?
[10:59:29] <Jonathan Lennox> I feel like if there were guidance on "how to make implementing a h3 server less complicated if you just care about quictransport semantics" guidance, going with h3 would be clearer ... is that possible?
[10:59:44] <Sean Turner> INTERIM!
[10:59:50] <Alan Frindell> +1 interim
[10:59:52] <Jana Iyengar_820> Sean: Bingo!
[10:59:52] <Victor Vasiliev> It should be a superset in terms of features, but not in terms of complexity
[10:59:59] <Sean Turner> and distribute knives
[11:00:10] <Mirja Kühlewind> we could go over time a bit if needed but I don't think that will lead to a decisions
[11:00:18] <Mike English> +1 interim
[11:00:19] Eric Rescorla leaves the room
[11:00:22] Marcus Ihlar leaves the room
[11:00:39] <Alan Frindell> jonathan: I think we can do something like that
[11:00:49] Nick Banks leaves the room
[11:00:53] Eric Kinnear leaves the room
[11:00:53] Barry Leiba leaves the room
[11:00:54] Chi-Jiun Su leaves the room
[11:00:59] Ali Hussain leaves the room
[11:00:59] Mike Bishop leaves the room
[11:01:01] Tommy Pauly leaves the room
[11:01:01] Mihail Yanev leaves the room
[11:01:02] Andrew S leaves the room
[11:01:03] Martin Thomson leaves the room
[11:01:03] Dan McArdle leaves the room
[11:01:04] <Mirja Kühlewind> bye thanks!
[11:01:05] Alan Frindell leaves the room
[11:01:05] Marten Seemann leaves the room
[11:01:05] Bernard Aboba leaves the room
[11:01:06] Yoshifumi Atarashi leaves the room
[11:01:07] Alissa Cooper leaves the room
[11:01:07] Sean Turner leaves the room
[11:01:08] Victor Vasiliev leaves the room
[11:01:09] Lucas Pardue leaves the room
[11:01:09] Kazuho Oku leaves the room
[11:01:09] Jake Holland leaves the room
[11:01:09] Valery Smyslov leaves the room
[11:01:11] Luke Curley leaves the room
[11:01:11] Philipp Tiesel leaves the room
[11:01:12] Robin Marx leaves the room
[11:01:12] Luca Niccolini leaves the room
[11:01:13] Virat Tara leaves the room
[11:01:14] Mark Nottingham leaves the room
[11:01:17] Harald Alvestrand leaves the room
[11:01:18] Hiroyuki Goto leaves the room
[11:01:18] Kazuaki Ueda leaves the room
[11:01:19] Mike English leaves the room
[11:01:20] Jana Iyengar_820 leaves the room
[11:01:24] xdf leaves the room
[11:01:26] Jana Iyengar leaves the room
[11:01:38] ihlar joins the room
[11:01:39] 一穂 奥 leaves the room
[11:01:46] <Jonathan Lennox> https://www.youtube.com/watch?v=vhxttDEv_I8
[11:02:02] Alessandro Amirante leaves the room
[11:02:02] Mo Zanaty leaves the room
[11:02:02] Jonathan Lennox leaves the room
[11:02:02] Paolo Saviano leaves the room
[11:02:02] Frédéric Fieau leaves the room
[11:02:02] Mirja Kühlewind leaves the room
[11:02:02] Nicolas Kuhn leaves the room
[11:02:02] David Schinazi leaves the room
[11:02:02] Josh Cohen leaves the room
[11:02:02] Yanmei Liu leaves the room
[11:02:31] ekinnear leaves the room
[11:03:34] Meetecho leaves the room
[11:14:36] mnot leaves the room: Disconnected: closed
[11:17:30] ihlar leaves the room
[11:47:28] lnicco@jabb.im leaves the room: Connection failed: connection closed
[12:04:29] alex-meetecho leaves the room
[12:08:58] lpardue leaves the room: Disconnected: BOSH client silent for over 60 seconds
[12:40:51] Mirja Kuehlewind leaves the room: Disconnected: closed
[12:52:18] ihlar leaves the room
[13:16:11] dragana leaves the room
[13:48:27] dragana joins the room
[13:51:13] dragana leaves the room
[14:22:17] dragana joins the room
[14:24:50] dragana leaves the room
[14:28:14] ihlar joins the room
[16:23:07] dragana joins the room
[16:25:57] dragana leaves the room
[17:07:18] ihlar leaves the room
[17:15:37] dragana joins the room
[17:18:25] dragana leaves the room
[17:24:17] ihlar joins the room
[17:35:56] lnicco@jabb.im joins the room
[17:48:57] afrind leaves the room
[17:48:58] afrind joins the room
[17:59:17] lnicco@jabb.im leaves the room: Connection failed: connection closed
[17:59:22] lnicco@jabb.im joins the room
[18:06:24] afrind leaves the room
[18:27:47] Mirja Kuehlewind joins the room
[18:45:11] lnicco@jabb.im leaves the room: Connection failed: connection closed
[18:52:50] afrind joins the room
[18:52:50] afrind leaves the room
[18:53:14] afrind joins the room
[18:53:26] afrind leaves the room
[18:53:34] afrind joins the room
[18:53:39] afrind leaves the room
[18:54:49] dragana joins the room
[20:20:15] afrind joins the room
[20:20:15] afrind leaves the room
[20:21:21] ihlar leaves the room
[20:54:47] dschinazi@jab.im leaves the room
[22:18:19] afrind joins the room
[22:24:24] afrind leaves the room
[23:07:39] dragana leaves the room
[23:40:13] dragana joins the room
[23:42:52] dragana leaves the room