IETF
QUIC
quic@jabber.ietf.org
Thursday, March 22, 2018< ^ >
kaduk@jabber.org/barnowl has set the subject to: QUIC Interim
Room Configuration
Room Occupants

GMT+0
[00:26:39] nygren leaves the room
[01:14:29] hgn leaves the room: Disconnected: Replaced by new connection
[01:14:30] hgn joins the room
[05:45:00] hgn leaves the room
[05:59:08] hgn joins the room
[06:07:10] hgn leaves the room
[06:08:30] hgn joins the room
[06:27:55] hgn leaves the room
[06:29:10] hgn joins the room
[06:45:47] hgn leaves the room
[06:49:59] hgn joins the room
[08:51:13] nygren joins the room
[09:02:36] Adam Roach leaves the room
[09:03:39] nygren leaves the room
[09:11:04] Yoshiro Yoneya joins the room
[09:17:38] meetecho joins the room
[09:19:49] Kurt Andersen joins the room
[09:24:59] aaron joins the room
[09:25:09] TOSHIO HIRAGA joins the room
[09:25:10] Martin Stiemerling joins the room
[09:25:10] John Border joins the room
[09:25:12] Kaoru Maeda joins the room
[09:25:12] Daniel Havey joins the room
[09:25:20] Martin Thomson joins the room
[09:25:25] <aaron> I am jabber scribe for this meeting.  Please preface any comments for the mic with "MIC:"
[09:25:27] blassey joins the room
[09:25:29] Pierre Pfister joins the room
[09:25:33] Martin Thomson has set the subject to: Spinning
[09:25:35] Daniel Stenberg joins the room
[09:25:56] ted.h joins the room
[09:26:15] g.e.montenegro joins the room
[09:26:26] yuki Goto joins the room
[09:26:34] Martin Duke joins the room
[09:27:17] <Martin Thomson> https://www.youtube.com/watch?v=WEhS9Y9HYjU
[09:27:50] Mukesh Chouhan joins the room
[09:28:24] Lars Eggert joins the room
[09:28:35] Magnus Westerlund joins the room
[09:29:31] Mauro Cociglio joins the room
[09:29:59] <aaron> https://www.youtube.com/watch?v=Ywa6glFr6io
[09:30:19] Thom Peterson joins the room
[09:30:24] aaron has set the subject to: Start of Thurs meeting
[09:30:54] aaron leaves the room: Replaced by new connection
[09:30:57] aaron joins the room
[09:31:41] Piers O'Hanlon joins the room
[09:31:52] aaron has set the subject to: Discussion of invariants
[09:32:31] Piers O'Hanlon leaves the room
[09:32:32] Piers O'Hanlon joins the room
[09:32:47] Maxime Piraux joins the room
[09:32:54] Magnus Westerlund leaves the room: Replaced by new connection
[09:32:54] Magnus Westerlund joins the room
[09:33:16] Simon Pietro Romano joins the room
[09:33:23] Adam Roach joins the room
[09:33:39] <aaron> speaking: Ekr
[09:34:09] Pierre Pfister leaves the room
[09:34:15] Pierre Pfister joins the room
[09:34:22] <aaron> speaking: Martin Thompson
[09:34:37] <aaron> (Thomson)
[09:35:11] lucaspardue joins the room
[09:35:24] ted.h leaves the room
[09:35:29] Jason Polis joins the room
[09:36:02] Mauro Cociglio leaves the room
[09:36:11] Mauro Cociglio joins the room
[09:36:13] <aaron> speaking: Mirja
[09:36:48] <aaron> speaking: Er
[09:36:51] <aaron> Ekr
[09:37:23] Martin Stiemerling leaves the room
[09:37:41] <Piers O'Hanlon> There's no audio on the YouTube Live feed..?
[09:38:00] <aaron> meetecho: is this something you can help with?
[09:38:07] <aaron> speaking: Ted Hardie
[09:38:22] <aaron> speaking: Jana
[09:38:44] ted.h joins the room
[09:38:55] brian@trammell.ch joins the room
[09:39:06] <meetecho> We don't handle the YouTube feed, but we can relay the issue
[09:39:27] Jonathan Lennox joins the room
[09:39:35] aaron has set the subject to: ECN Support in QUIC, Ingemar Johansson
[09:39:40] Mauro Cociglio leaves the room
[09:39:48] Mauro Cociglio joins the room
[09:40:00] nygren joins the room
[09:40:03] c.j. joins the room
[09:40:17] brian@trammell.ch leaves the room
[09:40:19] <aaron> meetecho: thanks!
[09:40:22] Samuel Jero joins the room
[09:40:30] Mirja Kuehlewind joins the room
[09:40:49] <Piers O'Hanlon> (Y)
[09:41:24] gryning joins the room
[09:41:55] brian@trammell.ch joins the room
[09:41:57] <Martin Thomson> WTF GitHub: I change the title of one issue and it changes the title of another one
[09:42:02] Narayanaswamy Baluswamy joins the room
[09:42:08] Mirja Kuehlewind leaves the room: Disconnected: closed
[09:42:17] <aaron> speaking: Mirja
[09:43:27] <Piers O'Hanlon> Audio working now on Youtube - thanks
[09:44:24] <aaron> speaking: Magnus Westerlund
[09:44:29] Narayanaswamy Baluswamy leaves the room
[09:45:15] Brook Schofield joins the room
[09:47:02] Adam Roach leaves the room
[09:47:05] <aaron> speaking: Ekr
[09:47:05] Adam Roach joins the room
[09:47:55] <aaron> speaking: Mirja
[09:48:40] <aaron> speaking: Martin
[09:49:01] brian@trammell.ch leaves the room
[09:49:41] Mirja Kuehlewind joins the room
[09:50:16] <aaron> speaking: Jana
[09:51:24] aaron has set the subject to: Spin Bit Overview, Brian Trammell
[09:53:16] Shwetha Bhandari joins the room
[09:53:35] ch joins the room
[09:54:19] Pierre Pfister leaves the room
[09:54:24] Pierre Pfister joins the room
[09:54:32] Martin Thomson leaves the room: Stream closed by us: Replaced by new connection (conflict)
[09:54:32] Martin Thomson joins the room
[09:55:10] Alissa Cooper joins the room
[09:56:11] TOSHIO HIRAGA leaves the room
[09:56:25] aaron leaves the room
[09:56:46] Tobias Bühler joins the room
[09:58:39] ted.h leaves the room
[09:59:17] Andra Lutu joins the room
[09:59:39] aaron (jabber scribe) joins the room
[09:59:46] <aaron (jabber scribe)> speaking: ekr
[09:59:50] <aaron (jabber scribe)> speaking: mirja
[10:00:34] Brook Schofield leaves the room
[10:00:52] ted.h joins the room
[10:02:23] <aaron (jabber scribe)> speaking: Dave Oran
[10:02:30] Spencer Dawkins joins the room
[10:03:11] <aaron (jabber scribe)> speaking: Ian Swett
[10:03:22] ted.h leaves the room
[10:03:49] Roman Shpount joins the room
[10:04:02] <Spencer Dawkins> I have the same question that Dave Oran just had about multipath, but I was thinking about asymmetric paths with one observation point.
[10:04:53] <Spencer Dawkins> I told the chairs that I wouldn't interrupt the discussion as an individual, but if anyone wants to explain things to me, I'm around for the next 48 hours
[10:05:20] <aaron (jabber scribe)> speaking: Javier
[10:05:23] <aaron (jabber scribe)> er Jose
[10:05:31] Daniel Havey leaves the room
[10:05:41] Daniel Havey joins the room
[10:05:51] lucaspardue leaves the room: Disconnected: Replaced by new connection
[10:05:51] lucaspardue joins the room
[10:06:01] <aaron (jabber scribe)> speaking: Colin Perkins
[10:06:13] Mauro Cociglio leaves the room
[10:06:20] Mauro Cociglio joins the room
[10:06:39] Ignas Bagdonas joins the room
[10:07:36] ch leaves the room: Disconnected: closed
[10:08:39] <aaron (jabber scribe)> speaking: Christian Huitema
[10:09:42] Piet De Vaere joins the room
[10:10:23] frodek joins the room
[10:13:11] gryning is now known as craigt
[10:13:24] <aaron (jabber scribe)> speaking: Igor Lubashev
[10:14:26] nygren leaves the room
[10:14:39] ted.h joins the room
[10:14:40] <aaron (jabber scribe)> speaking: Ekr
[10:15:04] <aaron (jabber scribe)> speaking: Markko
[10:15:15] Pierre Pfister leaves the room
[10:15:21] Pierre Pfister joins the room
[10:15:31] Mirja Kuehlewind leaves the room
[10:16:04] spencerdawkins joins the room
[10:16:07] <aaron (jabber scribe)> speaking: Mirja
[10:16:32] <aaron (jabber scribe)> speaking: Erik Nygren
[10:17:01] <aaron (jabber scribe)> speaking: Mirja
[10:17:09] Jonathan Thackray joins the room
[10:17:30] <aaron (jabber scribe)> speaking: Subodh Iyangar
[10:17:34] <aaron (jabber scribe)> speaking: Colin Perkins
[10:17:34] Spencer Dawkins leaves the room
[10:17:36] nygren joins the room
[10:18:00] aaron (jabber scribe) has set the subject to: Chair's comments on scope of discussion
[10:18:03] ted.h leaves the room
[10:19:13] aaron (jabber scribe) has set the subject to: Discussion on adopting Spin Bit proposal
[10:19:18] <aaron (jabber scribe)> speaking: Brian Trammell
[10:19:27] <aaron (jabber scribe)> speaking: Ted Hardie
[10:21:52] Pierre Pfister leaves the room
[10:22:21] <aaron (jabber scribe)> speaking: Daniel Kahn Gillmor
[10:23:14] Shwetha Bhandari leaves the room
[10:23:53] <aaron (jabber scribe)> speaking: Jana
[10:24:08] <lucaspardue> +1 dkg
[10:25:33] <aaron (jabber scribe)> speaking: kevin smith
[10:25:43] <kaduk@jabber.org/barnowl> Would it make sense to use a quic version to experiment with a spin bit,
or are there finer-grained ways to indicate that to the path?
[10:26:45] <aaron (jabber scribe)> speaking; Emile Stephan
[10:27:14] <aaron (jabber scribe)> speaking: Colin Perkins
[10:27:31] <kaduk@jabber.org/barnowl> (I am in another session and can't follow terribly closely, but if Aaron
thinks that's worth asking at the mic, feel free to do so.)
[10:27:58] <craigt> Jana's experience is basically my own. Only radio network operators have expressed any interest to me directly. On balance, I wonder if experimentation or solving this problem within an exterior frame is the right way forward
[10:28:27] <spencerdawkins> The responsible AD would appreciate it if Aaron asked Ben's question (above) ...
[10:28:28] <aaron (jabber scribe)> Ben: Jana suggested the same: move ahead as an experiment
[10:28:43] <kaduk@jabber.org/barnowl> ok, cool
[10:29:11] <kaduk@jabber.org/barnowl> I guess with 32 bits of version space it's a pretty easy thing to do
[10:29:34] Samuel Jero leaves the room
[10:29:43] Steve Olshansky joins the room
[10:29:53] <aaron (jabber scribe)> Ben: conveyed to mic.
[10:29:59] <kaduk@jabber.org/barnowl> thanks!
[10:29:59] <aaron (jabber scribe)> speaking: Patrick McManus
[10:30:49] Piet De Vaere leaves the room
[10:30:58] Samuel Jero joins the room
[10:30:59] <kaduk@jabber.org/barnowl> (I was also thinking a little bit in terms of "get a v2 out really fast"
for exercising version negotiation, and "really fast" might be really
fast if it's just adding spin bit
[10:31:15] <kaduk@jabber.org/barnowl> But that's a fine topic for the list and doesn't need to go on the mic
[10:31:38] <aaron (jabber scribe)> speaking: Thomas Fossati
[10:32:07] <aaron (jabber scribe)> speaking: Brian Trammell
[10:34:23] <spencerdawkins> kaduk, Spencer thinks that we need to have versioning nailed down enough for Google to negotiate between gQUIC and IETF QUIC, and if that is true, I am hopeful that your answer is "yes".
[10:34:43] <spencerdawkins> If the answer is "no", I probably need to talk at the mike :D ...
[10:35:47] <John Border> Since the client starts things up, it is essentially opt in or opt out, depending on the way you set the default.
[10:35:55] <Lars Eggert> in terms of versions, i personally (chair hat off) see the wg doing a v1.1 which is more a bugfix endevour and v2 with new features (e.g., multipath) in parallel
[10:36:42] <John Border> I need to type faster.  Mark got to my question...
[10:36:50] lucaspardue leaves the room: Disconnected: Replaced by new connection
[10:36:51] lucaspardue joins the room
[10:37:04] Adam Roach leaves the room
[10:37:20] Miroslav Lichvar joins the room
[10:37:32] <aaron (jabber scribe)> speaking: Gabriel Montenegro
[10:37:41] Roman Shpount leaves the room
[10:38:01] Adam Roach joins the room
[10:38:21] Pierre Pfister joins the room
[10:38:27] <aaron (jabber scribe)> speaking: Zahed Sarker
[10:39:14] Jason Polis leaves the room
[10:39:57] ch joins the room
[10:40:12] <kaduk@jabber.org/barnowl> Tying it to version (instead of just relying on opt-in) allows for
additional experimentation with square wave vs. transcendental sequence
vs. other sequence
[10:40:26] <aaron (jabber scribe)> speaking: Ian Swett
[10:40:37] <craigt> "Opt-in" may well not mean, user perceptible opt-in...People should clarify why they mean
[10:41:43] Adam Roach leaves the room
[10:41:50] <spencerdawkins> Craigt - understood. That would be nice.
[10:42:50] Adam Roach joins the room
[10:42:54] ch leaves the room: Disconnected: Replaced by new connection
[10:42:55] ch joins the room
[10:43:53] Martin Thomson leaves the room
[10:44:14] <nygren > +1 to Ian  (I was on the fence, but I think Ian convinced me)
[10:44:32] <aaron (jabber scribe)> speaking: marcus Ihlar
[10:44:37] Piet De Vaere joins the room
[10:44:54] <Jonathan Thackray> Agreed, Ian's arguments were convincing.
[10:45:16] <aaron (jabber scribe)> speaking: Manasi Deval
[10:45:28] <kaduk@jabber.org/barnowl> I missed it: what did Ian argue for?
[10:45:34] <ch> I have to leave the meeting room due to a conflict, but my preference is to mark 3 bits as reserved in the main spec, and have an experiemental RFC explains how to use the reserved bits for SPIN + VET.
[10:45:38] <aaron (jabber scribe)> speaking: Ekr
[10:45:48] <craigt> kaduk@jabber.org/barnowl: Ian argued for experimental draft
[10:46:36] <aaron (jabber scribe)> an experimental draft and use of many bits to figure out what the best answer is
[10:46:46] Piet De Vaere leaves the room
[10:46:48] piet de vaere joins the room
[10:47:57] Adam Roach leaves the room
[10:50:11] <aaron (jabber scribe)> speaking: spencer AD
[10:50:11] <Thom Peterson> MIC: I am more in favour of having consensus in allocating n bits for experiments like SPIN, rather than baking SPIN in
[10:50:44] Adam Roach joins the room
[10:51:31] Martin Thomson joins the room
[10:52:20] lucaspardue leaves the room: Disconnected: Replaced by new connection
[10:52:20] lucaspardue joins the room
[10:52:53] <Samuel Jero> What was the result?
[10:53:11] <Martin Thomson> opera humming
[10:53:35] <Martin Thomson> research was slightly less apparently
[10:53:54] <Samuel Jero> Okay. That's what I was trying to figure out. It wasn't clear on the remote stream
[10:54:02] <aaron (jabber scribe)> Marten Seemann
[10:54:06] <aaron (jabber scribe)> speaking
[10:54:19] <Thom Peterson> Thanks, whomever spoke.
[10:54:37] <aaron (jabber scribe)> my opinion: two fairly loud hums, engineering distinctly louder.  Not a consensus IMO
[10:54:57] Martin Thomson leaves the room
[10:54:58] <aaron (jabber scribe)> speaking: Sanjay Mishra
[10:55:27] Andra Lutu leaves the room
[10:56:31] <aaron (jabber scribe)> speaking: Mirja
[10:57:44] Jonathan Lennox leaves the room
[10:58:06] <spencerdawkins> Not standing up again for a while, but I agere with Mirja and note that Mirja is a chair of MAPRG. You know, another research group in the IRTF.
[10:58:41] nygren leaves the room
[10:58:41] <spencerdawkins> MAPRG measures things ...
[10:58:43] <aaron (jabber scribe)> speaking: Kazuho Oku
[10:59:20] <aaron (jabber scribe)> speaking: Andrew McGregor
[10:59:44] Alissa Cooper leaves the room
[11:01:37] <aaron (jabber scribe)> speaking:  Emile Stephan
[11:02:02] <aaron (jabber scribe)> speaking: Ted Hardie
[11:03:40] Jonathan Lennox joins the room
[11:04:12] Alissa Cooper joins the room
[11:04:36] aaron (jabber scribe) has set the subject to: Ted asks chairs about hum on encrypting seq no; causes segfault
[11:05:25] Alissa Cooper_4946 joins the room
[11:06:14] aaron (jabber scribe) has set the subject to: discussion on adopting Spin Bit proposal
[11:06:19] <aaron (jabber scribe)> speaking: DKG
[11:07:00] lucaspardue leaves the room: Disconnected: Replaced by new connection
[11:07:00] lucaspardue joins the room
[11:08:00] Kyle Rose joins the room
[11:08:05] piet de vaere leaves the room
[11:08:08] ted.h joins the room
[11:08:36] <aaron (jabber scribe)> speaking: Brian Trammell
[11:08:50] <aaron (jabber scribe)> (well, interrupted by Spencer)
[11:11:39] Steve Olshansky leaves the room
[11:12:18] <aaron (jabber scribe)> speaking: Roni Evan
[11:13:58] ch leaves the room: Disconnected: Replaced by new connection
[11:13:58] ch joins the room
[11:14:19] <aaron (jabber scribe)> speaking: Al Morton
[11:16:31] <ted.h> MIC:  So it gets used by HTTP.
[11:16:53] <aaron (jabber scribe)> speaking: Martin Thomson
[11:17:02] Cullen Jennings joins the room
[11:17:40] Martin Thomson joins the room
[11:17:50] <aaron (jabber scribe)> speaking: Praveeen Balasubramanian
[11:18:18] Maxime Piraux leaves the room
[11:18:24] Maxime Piraux joins the room
[11:18:34] <aaron (jabber scribe)> speaking: Subodh Ihengar
[11:18:46] Kyle Rose leaves the room
[11:19:26] <Martin Thomson> connection ID as spin bit is interesting
[11:19:27] Kyle Rose joins the room
[11:19:54] <Martin Thomson> about as interesting as connection ID as packet number, that is, likely only academic
[11:20:22] <ted.h> @Martin Having people start putting semantics into the connection ID is greasing a slope that will get slippery fast.
[11:20:33] <kaduk@jabber.org/barnowl> Are the enumerated options in one of the slide decks?
[11:20:33] <Martin Thomson> FWIW, the necessary behaviour for connection ID as a spin bit already exist
[11:20:51] <aaron (jabber scribe)> kaduk: no, text app in real time
[11:21:01] <kaduk@jabber.org/barnowl> *nods* that's what it looked/sounded like
[11:21:05] <Martin Thomson> ted.h: I don't think that we're looking at semantics, it's an emergent property
[11:21:06] <ted.h> @Martin Except for the mechanism to signal that it should be used, eh?
[11:21:12] <aaron (jabber scribe)> 1. Spin bit(s) in Invariants
[11:21:21] <aaron (jabber scribe)> 2. Spin Bit(s) in QUICv1
[11:21:22] <aaron (jabber scribe)> 3
[11:21:23] <Martin Thomson> ted.h: see Kazuho's comments
[11:21:41] <aaron (jabber scribe)> 3. Spin Bit(s) in (negotiated?) extension (separate experimental RFC)
[11:21:47] <aaron (jabber scribe)> 4. No Spin Bit(s)
[11:21:52] <aaron (jabber scribe)> speaking: Erik Nygren
[11:21:59] <ted.h> @ted.h can you relay them again?  I must have missed this point.
[11:22:13] <aaron (jabber scribe)> speaking: Ekr
[11:22:45] nygren joins the room
[11:23:06] <aaron (jabber scribe)> Option 3a. reserve the bit(s)?
[11:23:19] <ted.h> @Martin Can you relay them?  I must have missed that element of his point.
[11:23:20] <craigt> I notice that the main interest is coming from radio networks, and that one of the use cases is building functionality to affect flows; see hackathon. What's the confidence of the room this isn't just enabling the ossification we're trrying to avoid.
[11:24:02] <aaron (jabber scribe)> speaking: Mike Bishop
[11:24:54] <Martin Thomson> ted.h: kazuho's comment was that you can detect compliance by measuring flows, which - to ekr's point just now - suggests that you don't need negotiation/signaling
[11:25:24] <aaron (jabber scribe)> speaking: Dan Druta
[11:25:42] <spencerdawkins> @craigt, speaking as the responsible AD - After the IAB MaRNEW workshop, I do not have confidence that we can have an open and informed decision about what radio network operators will/won't do. ("We tried that for more than a year after the workshop, and here we are")
[11:26:13] <kaduk@jabber.org/barnowl> here we are ... or here it isn't?
[11:26:37] <aaron (jabber scribe)> speaking: Jana
[11:27:34] <ted.h> So, based on this concern, we are also abandoning the entire DTLS conversation, right?
[11:28:03] <ted.h> I'm happy people have become schedule conscious, but let's be consistent if so.
[11:28:10] <spencerdawkins> @kaduk - we're having the same discussion thar we were having after we got a post "Pervasive Monitoring Considered An Attack" RFC liaison years ago. MaRNEW was in 2015 and we don't know very much more now than we knew then.
[11:28:43] <aaron (jabber scribe)> speaking: Brian
[11:29:20] <kaduk@jabber.org/barnowl> "it's on or it's off", but *what* is on, and do you want to be able to
change the nature of what is on?
[11:29:29] Jun Liu joins the room
[11:29:38] <craigt> spencerdawkins: I'm sat next to one, I can grab his ankles if that helps. I'll refresh myself on the outcome of that workshop, thank you.
[11:29:40] <kaduk@jabber.org/barnowl> (I assume the mic queue is closed and also long so no point trying to relay)
[11:29:52] <Martin Thomson> mic queue is long closed
[11:29:53] <ted.h> @kaduk it is closed
[11:30:55] Ignas Bagdonas leaves the room
[11:31:14] <spencerdawkins> @kaduk informstion is always helpful, but what we got with GSMA asking is that the operators wouldn't even tell each other what their plans were, much less tell US in public.
[11:31:40] <aaron (jabber scribe)> speaking: Mirja
[11:31:59] <spencerdawkins> @kaduk - so, grab away, but remember that you're taking a sample size of "1" ...
[11:32:00] <frodek> @spencerdawkings I couldn't agree mor and wish you would repeat this on the mike
[11:32:07] brian@trammell.ch joins the room
[11:32:15] Tobias Bühler leaves the room
[11:32:47] <craigt> Jonathan Thackray: ...are you in the room?
[11:32:51] <aaron (jabber scribe)> speaking: Chris Seal
[11:33:00] <Martin Duke> When we punted this issue  long ago, I thought the rationale was that we could easily drop it in at the last minute. Now the assertion is that it's too hard to drop in at the last minute.
[11:33:21] <Jonathan Thackray> @craigt - sorry, I'm remote
[11:33:22] <brian@trammell.ch> no the assertion is more or less "this is the last minute"
[11:33:34] <aaron (jabber scribe)> speaking: Joerg Ott
[11:33:50] Kyle Rose leaves the room
[11:33:56] <spencerdawkins> @frodek - @trammell or @ted.h can confirm or deny this, but I believe the MaRNEW IAB workshop report goes out for community comments (NOT IETF LAST CALL) in a week or two. I'll say it when I forward that call for comments to QUIC.
[11:34:33] <brian@trammell.ch> one thing i forgot to state at the mic: i'm actually convinced enough of the usefulness of this signal that i'm currently working on draft-trammell-tcpm-spin.
[11:34:39] <ted.h> @spencerdawkins  Yes.  Cindy has her hand on the button as we speak.
[11:34:45] <brian@trammell.ch> (admittedly, it started as a joke, but it isn't anymore.)
[11:34:53] <aaron (jabber scribe)> https://tools.ietf.org/html/draft-iab-marnew-report-01
[11:35:13] <spencerdawkins> Jonathan Thackery - thank you foro participating remotely. More knowledge is always helpful.
[11:35:40] <aaron (jabber scribe)> speaking: Ian Swett
[11:36:25] <kaduk@jabber.org/barnowl> Yay, Ian is reading my mind
[11:36:44] <kaduk@jabber.org/barnowl> Though I'm waiting for dkg to come up and say that middleboxes can drop
quicv1 and force quicvspin-bit
[11:37:07] Adam Roach leaves the room
[11:37:19] <frodek> @spencerdawkings. Great. Thx. Haven't seen the new version but looking forward to read the new version which is hopefully greatly improved
[11:37:27] <aaron (jabber scribe)> speaking: Ted Hardie
[11:37:38] <brian@trammell.ch> any middlebox that wants to force quicvspin can much more easily force tco
[11:37:42] <brian@trammell.ch> *tcp
[11:37:53] <Martin Thomson> defining vSPIN has the virtue of not including the bit definitions in the core transport doc
[11:37:54] <brian@trammell.ch> the cost to force tcp is lower. it's a single stateless rule.
[11:37:55] <kaduk@jabber.org/barnowl> For now, at least.
[11:38:01] <Martin Thomson> also, it removes HOLB
[11:38:01] <aaron (jabber scribe)> speaking: Mike Bishop
[11:38:22] <aaron (jabber scribe)> speaking: Ian
[11:38:34] <aaron (jabber scribe)> speaking: Brad Lassey
[11:39:05] brian@trammell.ch leaves the room
[11:39:20] <aaron (jabber scribe)> speaking: Brian
[11:39:55] geir joins the room
[11:40:05] yuki Goto leaves the room
[11:40:07] aaron (jabber scribe) has set the subject to: Spin bit discussion chairs' summary
[11:40:08] yuki Goto joins the room
[11:40:08] <Martin Thomson> management idea: detect reordering based on the bits and then hold back any packets that were out of order
[11:40:27] <Martin Thomson> the "let me straighten that out for you"
[11:40:43] <aaron (jabber scribe)> speaking: mirja
[11:40:47] <blassey> To be clear, wasn't thinking of mucking with the packets, was thinking of mucking with buffers and dropping
[11:41:11] <Martin Thomson> blassey: yes, dropping, reordering the queue, delaying, etc...
[11:41:12] yuki Goto leaves the room
[11:41:14] yuki Goto joins the room
[11:41:19] <blassey> exactly
[11:41:32] <aaron (jabber scribe)> speaking: brian
[11:42:38] Adam Roach joins the room
[11:42:39] <nygren > Are we getting close to a hum on 3b vs others?
[11:42:48] <Martin Thomson> nygren: it sounds like it
[11:43:07] <Martin Thomson> of course, this is all predicated on a confused understanding of version negotiation
[11:43:20] <aaron (jabber scribe)> speaking: jana
[11:43:26] Pierre Pfister leaves the room
[11:44:14] Kurt Andersen leaves the room
[11:44:19] <aaron (jabber scribe)> speaking: ekr
[11:44:26] Pierre Pfister joins the room
[11:45:10] ch leaves the room: Disconnected: Replaced by new connection
[11:45:10] ch joins the room
[11:45:12] geir leaves the room
[11:46:07] <aaron (jabber scribe)> speaking: roni even
[11:46:20] <Martin Thomson> quic v-1 uses little endian
[11:46:37] <kaduk@jabber.org/barnowl> Using a version lets you change around how you're using those bits at a
later time
[11:46:41] <kaduk@jabber.org/barnowl> Though I guess you only need to bump the version when you have the new
thing you want to do with the bits.
[11:47:01] <aaron (jabber scribe)> speaking: subodh iyengar
[11:47:35] Samuel Jero leaves the room
[11:47:38] <nygren > (oh, I'd been assuming that 3b was to be thought of as experimental at this point?)
[11:47:38] Samuel Jero joins the room
[11:47:51] <ted.h> I am not in line, but the argument that this is an invariant is just wrong.
[11:48:01] <ted.h> It's a version-specific use of the bits.
[11:48:07] <aaron (jabber scribe)> speaking: ian swett
[11:49:06] <aaron (jabber scribe)> speaking: emile
[11:49:14] <kaduk@jabber.org/barnowl> extension is okay if only the peers need to know how the bits are used,
but if you need the path involved, right -- what Ian says
[11:49:38] <Martin Thomson> Ian is wrong of course: not every path sees, or will remember, the handshake
[11:49:48] <aaron (jabber scribe)> speaking: sanjay
[11:50:00] <Martin Thomson> you might accept the loss of those connections that migrate, of course
[11:51:17] aaron (jabber scribe) has set the subject to: hums & hands
[11:51:20] <Martin Duke> F5 would do it
[11:53:00] <aaron (jabber scribe)> speaking: brian
[11:53:00] ted.h leaves the room
[11:53:28] <aaron (jabber scribe)> speaking: brad
[11:54:03] <aaron (jabber scribe)> speaking: roni
[11:54:28] <aaron (jabber scribe)> speaking: jana
[11:55:16] <aaron (jabber scribe)> speaking: martin
[11:55:32] <nygren > blocking port >> blocking version >> blocking after seeing stream of all-zero spin bits
[11:55:36] <aaron (jabber scribe)> speaking: spencer
[11:55:55] <aaron (jabber scribe)> speaking: ted
[11:56:15] ch leaves the room: Disconnected: closed
[11:56:23] Piet De Vaere joins the room
[11:56:26] <aaron (jabber scribe)> speaking: lars
[11:56:36] ted.h joins the room
[11:57:18] <aaron (jabber scribe)> speaking: marcus
[11:57:37] <aaron (jabber scribe)> speaking: jana
[11:58:50] brian@trammell.ch joins the room
[11:59:00] Pierre Pfister leaves the room
[11:59:10] <brian@trammell.ch> clarification: GQUIC demux bits are coming back out
[11:59:14] <brian@trammell.ch> correct?
[12:00:02] Pierre Pfister joins the room
[12:00:16] <brian@trammell.ch> (PR is no longer clean, but this change is so small that proposing a new PR is essentially zero effort)
[12:00:17] aaron (jabber scribe) leaves the room
[12:00:19] <craigt> I'm still feeling this as an optimisation, rather some thing that is needed to work.
[12:00:34] aaron (jabber scribe) joins the room
[12:00:36] brian@trammell.ch leaves the room
[12:01:27] <aaron (jabber scribe)> speaking: brian
[12:01:38] <aaron (jabber scribe)> hum:
[12:01:43] <aaron (jabber scribe)> a: no spin bits in v1
[12:01:58] <aaron (jabber scribe)> b. reserve bits, separate doc, later consensus to incorporate in v1
[12:02:03] Kurt Andersen joins the room
[12:02:03] <aaron (jabber scribe)> c don't know yet / something else
[12:02:28] Cullen Jennings leaves the room
[12:02:31] <aaron (jabber scribe)> speaking: martin
[12:02:48] lucaspardue leaves the room
[12:02:52] Magnus Westerlund leaves the room
[12:03:08] <meetecho> For remote attendees: fixing the audio issue as a type
[12:03:19] <John Border> b
[12:03:20] <aaron (jabber scribe)> speaking: mirja
[12:03:27] <Jonathan Thackray> thanks, @meetecho.
[12:03:29] yuki Goto leaves the room
[12:03:31] ted.h leaves the room
[12:03:31] yuki Goto joins the room
[12:03:40] Mauro Cociglio leaves the room
[12:03:45] Mauro Cociglio joins the room
[12:03:54] Martin Duke leaves the room
[12:03:56] Martin Duke joins the room
[12:03:58] <aaron (jabber scribe)> any other remote participants want to weigh in?
[12:03:59] Piet De Vaere leaves the room
[12:04:00] Piet De Vaere joins the room
[12:04:10] Daniel Stenberg leaves the room
[12:04:11] <nygren > should we have an option for "we should incorporate Brian's PR into the current doc"?  
[12:04:15] <meetecho> aaron (jabber scribe): the audio died for remotes, we're fixing it
[12:04:15] Adam Roach leaves the room
[12:04:17] <Kaoru Maeda> hum for b
[12:04:37] <Martin Duke> hum for b
[12:04:52] <Mauro Cociglio> hum for b
[12:04:53] <Piers O'Hanlon> hum for b
[12:05:00] craigt leaves the room: Disconnected: closed
[12:05:00] gryning joins the room
[12:05:20] <Jonathan Thackray> hum for b
[12:05:28] c.j. leaves the room: offline
[12:05:36] Martin Thomson leaves the room
[12:05:38] Yoshiro Yoneya leaves the room
[12:05:40] yuki Goto leaves the room
[12:05:41] aaron (jabber scribe) has set the subject to: adjourn
[12:05:42] yuki Goto joins the room
[12:05:45] blassey leaves the room
[12:05:50] nygren leaves the room
[12:05:55] <aaron (jabber scribe)> for the remote folks: strong consensus for B
[12:06:00] aaron (jabber scribe) leaves the room
[12:06:09] Lars Eggert leaves the room
[12:06:14] Piers O'Hanlon leaves the room
[12:06:14] Miroslav Lichvar leaves the room
[12:06:14] Martin Duke leaves the room
[12:06:14] Piet De Vaere leaves the room
[12:06:14] Kaoru Maeda leaves the room
[12:06:14] Jun Liu leaves the room
[12:06:14] yuki Goto leaves the room
[12:06:14] Mauro Cociglio leaves the room
[12:06:14] Samuel Jero leaves the room
[12:06:14] Daniel Havey leaves the room
[12:06:14] Maxime Piraux leaves the room
[12:06:14] Jonathan Thackray leaves the room
[12:06:14] John Border leaves the room
[12:06:14] Alissa Cooper_4946 leaves the room
[12:06:14] Simon Pietro Romano leaves the room
[12:06:14] Mukesh Chouhan leaves the room
[12:06:38] g.e.montenegro leaves the room
[12:06:44] Pierre Pfister leaves the room
[12:06:44] Jonathan Lennox leaves the room
[12:07:09] meetecho leaves the room
[12:08:30] spencerdawkins leaves the room: Disconnected: closed
[12:11:07] ch joins the room
[12:13:35] Jonathan Lennox joins the room
[12:14:12] Jonathan Lennox leaves the room
[12:15:14] frodek leaves the room
[12:17:22] brian@trammell.ch joins the room
[12:17:40] brian@trammell.ch leaves the room
[12:20:14] Alissa Cooper leaves the room
[12:21:31] gryning leaves the room: Disconnected: closed
[12:29:09] Steve Olshansky joins the room
[12:30:26] Steve Olshansky leaves the room: Replaced by new connection
[12:30:32] Martin Thomson joins the room
[12:30:43] Steve Olshansky joins the room
[12:36:06] Adam Roach joins the room
[12:41:55] Kyle Rose joins the room
[12:42:46] ch leaves the room: Disconnected: Replaced by new connection
[12:42:46] ch joins the room
[12:45:45] Adam Roach leaves the room
[12:45:51] Steve Olshansky leaves the room
[12:45:52] Martin Thomson leaves the room
[12:47:45] Martin Thomson joins the room
[12:54:59] Steve Olshansky joins the room
[12:57:37] Kyle Rose leaves the room: Disconnected: closed
[12:59:22] Martin Thomson leaves the room
[13:00:54] Thom Peterson leaves the room
[13:01:35] Martin Thomson joins the room
[13:02:00] Martin Thomson leaves the room
[13:04:51] Mirja Kuehlewind joins the room
[13:06:28] Mirja Kuehlewind leaves the room
[13:09:16] ch leaves the room: Disconnected: Replaced by new connection
[13:09:16] ch joins the room
[13:10:30] ch leaves the room: Disconnected: closed
[13:13:14] Martin Thomson joins the room
[13:13:42] Adam Roach joins the room
[13:20:01] gryning joins the room
[13:20:44] Kurt Andersen leaves the room
[13:20:59] Martin Thomson leaves the room
[13:22:18] Kyle Rose joins the room
[13:22:36] Martin Thomson joins the room
[13:25:45] Kyle Rose leaves the room: Disconnected: closed
[13:27:03] Magnus Westerlund joins the room
[13:27:43] Kurt Andersen joins the room
[13:30:07] Magnus Westerlund leaves the room
[13:30:26] lucaspardue joins the room
[13:30:51] gryning joins the room
[13:30:57] gryning leaves the room: Disconnected: closed
[13:33:57] Pierre Pfister joins the room
[13:34:16] spencerdawkins joins the room
[13:34:42] Mirja Kuehlewind joins the room
[13:36:15] Pierre Pfister leaves the room
[13:37:37] blassey joins the room
[13:39:33] Pierre Pfister joins the room
[13:40:31] frodek joins the room
[13:42:36] Mirja Kuehlewind leaves the room
[13:44:32] ch joins the room
[13:46:26] Kyle Rose joins the room
[13:46:41] Kyle Rose leaves the room: Disconnected: Replaced by new connection
[13:46:41] Kyle Rose joins the room
[13:48:14] frodek leaves the room
[13:48:19] Steve Olshansky leaves the room: Replaced by new connection
[13:48:30] Steve Olshansky joins the room
[13:49:00] Kyle Rose leaves the room
[13:50:20] ch leaves the room
[13:58:05] Mirja Kuehlewind joins the room
[14:03:07] Adam Roach leaves the room
[14:05:27] spencerdawkins joins the room
[14:11:19] Adam Roach joins the room
[14:13:47] nygren joins the room
[14:21:50] spencerdawkins leaves the room: Disconnected: No route to host
[14:24:14] nygren leaves the room
[14:24:42] Adam Roach leaves the room
[14:24:52] Pierre Pfister leaves the room
[14:26:50] Adam Roach joins the room
[14:27:04] nygren joins the room
[14:28:09] spencerdawkins leaves the room: Disconnected: closed
[14:29:56] gryning leaves the room: Disconnected: No route to host
[14:31:57] lucaspardue leaves the room: Disconnected: Replaced by new connection
[14:31:57] lucaspardue joins the room
[14:33:44] Kurt Andersen leaves the room
[14:35:42] spencerdawkins joins the room
[14:35:47] Pierre Pfister joins the room
[14:39:44] Adam Roach leaves the room
[14:40:25] gryning joins the room
[14:42:52] Kurt Andersen joins the room
[14:45:29] Pierre Pfister leaves the room
[14:51:36] Steve Olshansky leaves the room: Replaced by new connection
[14:51:51] Steve Olshansky joins the room
[14:53:16] nygren leaves the room
[14:54:57] Adam Roach joins the room
[14:55:04] nygren joins the room
[14:56:55] nygren leaves the room
[14:59:32] Pierre Pfister joins the room
[15:00:32] Pierre Pfister leaves the room
[15:00:41] Pierre Pfister joins the room
[15:02:34] Steve Olshansky leaves the room
[15:02:39] Mirja Kuehlewind leaves the room
[15:03:42] nygren joins the room
[15:04:09] nygren leaves the room
[15:04:47] nygren joins the room
[15:06:27] spencerdawkins leaves the room: Disconnected: closed
[15:08:00] Steve Olshansky joins the room
[15:09:22] Steve Olshansky leaves the room
[15:09:37] Steve Olshansky joins the room
[15:09:53] Steve Olshansky leaves the room
[15:10:21] blassey leaves the room: Disconnected: closed
[15:10:41] Martin Thomson leaves the room
[15:10:42] nygren leaves the room
[15:10:49] Adam Roach leaves the room
[15:10:54] lucaspardue leaves the room
[15:11:27] Steve Olshansky joins the room
[15:13:27] spencerdawkins joins the room
[15:16:38] spencerdawkins joins the room
[15:16:52] Pierre Pfister leaves the room
[15:17:15] Kurt Andersen leaves the room
[15:19:34] Mirja Kuehlewind joins the room
[15:19:59] Mirja Kuehlewind leaves the room: Disconnected: closed
[15:23:12] spencerdawkins leaves the room: Disconnected: closed
[15:26:24] Steve Olshansky leaves the room: Replaced by new connection
[15:26:31] Steve Olshansky joins the room
[15:27:33] blassey joins the room
[15:27:43] gryning leaves the room: Disconnected: No route to host
[15:32:56] spencerdawkins leaves the room: Disconnected: closed
[15:35:15] Martin Thomson joins the room
[15:42:24] Steve Olshansky leaves the room: Replaced by new connection
[15:42:32] Steve Olshansky joins the room
[15:43:18] Martin Thomson leaves the room
[15:44:16] Martin Thomson joins the room
[15:44:26] Kurt Andersen joins the room
[15:46:14] Adam Roach joins the room
[15:46:32] Mirja Kuehlewind joins the room
[15:47:27] Mirja Kuehlewind leaves the room
[15:53:24] spencerdawkins joins the room
[15:54:03] ch joins the room
[15:54:38] gryning joins the room
[15:57:50] Martin Thomson leaves the room
[16:01:31] nygren joins the room
[16:01:42] Alissa Cooper joins the room
[16:04:45] Kurt Andersen leaves the room
[16:05:00] Pierre Pfister joins the room
[16:05:14] Alissa Cooper joins the room
[16:05:15] Alissa Cooper leaves the room
[16:08:38] ch leaves the room
[16:09:06] Mirja Kuehlewind joins the room
[16:12:00] Mirja Kuehlewind leaves the room: Disconnected: closed
[16:14:16] Alissa Cooper leaves the room
[16:14:35] Alissa Cooper joins the room
[16:16:29] Mirja Kuehlewind joins the room
[16:16:45] Kurt Andersen joins the room
[16:27:37] Steve Olshansky leaves the room
[16:38:15] Alissa Cooper leaves the room
[16:42:46] Pierre Pfister leaves the room
[16:50:23] Pierre Pfister joins the room
[16:50:45] Alissa Cooper joins the room
[16:51:20] nygren leaves the room
[16:54:30] spencerdawkins joins the room
[16:56:36] nygren joins the room
[17:06:02] nygren leaves the room
[17:10:08] nygren joins the room
[17:11:02] spencerdawkins leaves the room: Disconnected: closed
[17:23:18] Kurt Andersen leaves the room
[17:37:14] Mirja Kuehlewind leaves the room
[17:40:22] Adam Roach leaves the room
[17:42:00] blassey leaves the room
[17:52:06] Kurt Andersen joins the room
[17:54:45] Kurt Andersen leaves the room
[17:54:54] Kurt Andersen joins the room
[17:55:45] Alissa Cooper leaves the room
[17:55:46] gryning leaves the room: Disconnected: closed
[18:01:52] nygren leaves the room
[18:11:17] hgn leaves the room
[18:13:20] gryning joins the room
[18:13:38] hgn joins the room
[18:14:16] spencerdawkins leaves the room: Disconnected: closed
[18:16:04] ch joins the room
[18:17:18] nygren joins the room
[18:47:36] ch leaves the room: Disconnected: Replaced by new connection
[18:47:36] ch joins the room
[18:49:28] Alissa Cooper joins the room
[18:52:39] ch leaves the room
[19:00:33] Pierre Pfister leaves the room
[19:08:54] Pierre Pfister joins the room
[19:10:20] Pierre Pfister leaves the room
[19:10:56] Pierre Pfister joins the room
[19:13:29] gryning joins the room
[19:13:29] gryning leaves the room: Disconnected: closed
[19:14:06] nygren leaves the room
[19:25:25] hgn leaves the room
[19:27:04] hgn joins the room
[19:29:18] Pierre Pfister leaves the room
[19:29:55] gryning leaves the room: Disconnected: closed
[19:38:34] spencerdawkins joins the room
[20:09:05] hgn leaves the room
[20:11:28] spencerdawkins leaves the room: Disconnected: closed
[20:12:03] hgn joins the room
[20:32:12] spencerdawkins joins the room
[20:42:17] hgn leaves the room
[20:43:40] spencerdawkins joins the room
[20:57:15] Alissa Cooper leaves the room
[21:00:07] spencerdawkins leaves the room: Disconnected: No route to host
[21:16:52] spencerdawkins leaves the room: Disconnected: closed
[21:24:20] hgn joins the room
[23:20:46] Kurt Andersen leaves the room
[23:47:11] spencerdawkins joins the room
[23:55:36] spencerdawkins joins the room
[23:57:44] spencerdawkins joins the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!