IETF
QUIC
quic@jabber.ietf.org
Thursday, January 26, 2017< ^ >
Sean Turner has set the subject to: QUIC Interim - 20170124-20170126 | 9:30 am  |  Japan Time (Tokyo, GMT+09:00)  |  8 hrs
Room Configuration
Room Occupants

GMT+0
[00:04:29] jo.ietf joins the room
[00:07:48] jo.ietf leaves the room
[00:07:58] Martin Thomson joins the room
[00:08:07] lucasp joins the room
[00:08:24] ekr joins the room
[00:11:22] spencerdawkins joins the room
[00:15:14] Mirja Kuehlewind joins the room
[00:16:56] Kaoru Maeda joins the room
[00:19:55] gryning@jabber.org.uk joins the room
[00:20:09] <Daniel> sound!
[00:20:15] <Daniel> video!
[00:22:26] Sean Turner joins the room
[00:24:51] <gryning@jabber.org.uk> good morning, evening and afternoon
[00:24:56] gryning@jabber.org.uk is now known as craigt
[00:25:05] afrind joins the room
[00:25:19] <Daniel> morning craigt!
[00:25:55] <mnot> starting in 5
[00:26:09] Dirk Kutscher joins the room
[00:26:18] <craigt> sorry, and thanks…probably caught me talking to the dogs
[00:26:50] <Daniel> "the dogs" right ;-)
[00:27:20] jo.ietf joins the room
[00:29:04] <craigt> It's the first night I've been joined by the hounds…
[00:29:06] ranjeeth joins the room
[00:31:05] Lars Eggert joins the room
[00:31:38] <Sean Turner> yep
[00:31:43] <Daniel> loud and clear
[00:31:44] chi.jiun.su joins the room
[00:31:45] <Sean Turner> yep
[00:31:45] <craigt> yes thx to both
[00:32:40] mcmanus joins the room
[00:34:03] ted.h joins the room
[00:34:21] Brian Trammell joins the room
[00:34:50] <ted.h> New Minutes doc: https://docs.google.com/document/d/1RjkilnUUziZWqwC2uvQDwg2Vq2PW1wBQGaqynirryC4/edit?usp=sharing
[00:34:55] <Daniel> thanks
[00:35:55] Buck Krasic joins the room
[00:38:05] <Sean Turner> ack in Europe
[00:38:14] <ted.h> Luxembourg!
[00:38:35] g.e.montenegro joins the room
[00:39:06] lstewart joins the room
[00:40:34] <Sean Turner> June 2-4 I'll be Nairobi …
[00:40:47] <craigt> I heard "first full week" i.e. 5th onwards
[00:40:54] <craigt> …but nice!
[00:40:57] <mnot> Yes
[00:41:13] lstewart sighs at Webex
[00:41:36] <Sean Turner> yeah I'm hoping it's 5 onward :)
[00:42:00] <Sean Turner> I'm sure we could get somebody in Dublin :)
[00:52:37] <Sean Turner> who is doing the logo?
[00:52:54] <Sean Turner> ISOC?
[00:52:57] <ted.h> Sean, it's you!
[00:53:04] <Sean Turner> ahhh
[00:53:06] <ted.h> The BoT will be doing it over a glass of whisky
[00:53:13] <craigt> Mike's son?
[00:54:08] Dirk Kutscher leaves the room
[00:57:34] Dirk Kutscher joins the room
[00:59:06] <mnot> Remote folks: hear / see OK?
[00:59:39] <lstewart> Can now after fighting with Webex for 15 mins :/
[01:00:03] Roni Even joins the room
[01:00:06] ranjeeth leaves the room
[01:00:49] <Buck Krasic> i can see hear well
[01:01:35] <craigt> Fine here also
[01:02:05] ranjeeth joins the room
[01:14:39] <Daniel> a scribing challenge no doubt...
[01:24:44] <Buck Krasic> long lived idle connections are considered taboo in mobile networks no?
[01:24:48] <Buck Krasic> battery eaters
[01:25:22] <craigt> layer 3/4 yes, layer 7 no
[01:25:31] <ted.h> @Buck only if that idle connection requires traffic to maintain.
[01:26:35] <Buck Krasic> I thought it was believed by some that having the connection open, even without traffic, might cause the radio to stay up when it could otherwise shut down?  
[01:27:02] <ted.h> If the radio is staying up to accept traffic, you have the same problem.
[01:30:35] <craigt> depends on the stack and a few socket params…keep alive L7-state, but if you keep alive tcp then you can have that issue…tradeoff is, you're more likely to be rebound NAT and need to restart…
[01:33:52] <craigt> sounds like the same cost as TLS session tickets today…
[01:34:28] <craigt> This sounds like a deployment consideration note…
[01:40:38] Mirja Kuehlewind leaves the room
[01:42:02] ekr leaves the room
[01:45:59] ekr joins the room
[01:47:35] Mirja Kuehlewind joins the room
[01:48:35] mnot leaves the room
[01:49:10] mnot joins the room
[01:49:24] <mnot> coming back
[01:55:19] <craigt> As I've heard ECMP balancing mentioned a lot today: I wonder if it's worth noting that ICMP will hash differently to the udp/tcp on most venodr implementations. i.e. You can't guarantee the sender is receiving ICMP, in a non-traditional 'black hole' sense
[01:56:34] <craigt> I like that idea! (IPv6 min)
[02:03:24] <spencerdawkins> Is it obvious why QUIC doesn't REQUIRE path MTU discovery (in 2017/2018)?
"Asking for a friend" ...
[02:04:05] <Buck Krasic> I think practice around PTMU discovery in 2017/2018 is non obvious
[02:04:29] <craigt> RFC4821?
[02:05:47] <spencerdawkins> craigt- exactly, and that was published 10 years ago.
[02:06:49] <craigt> I know it's still not on by default in many OS deployments…but it's certainly more robust than icmp in my experience
[02:09:34] <lstewart> ICMP is typically very expensive to deal with in the OS because you have to find the appropriate socket to deliver details to...
[02:10:00] <Sean Turner> https://datatracker.ietf.org/doc/rfc4821/
[02:18:37] <craigt> Mike Bishop / Mike Bishop: What's Windows PLPMTUD like?
[02:26:01] <ted.h> MUST be at least 1280 and MAY be the PMTU.
[02:26:11] <ted.h> or any size between 1280 and the PMTU.
[02:33:23] <Mike Bishop / Mike Bishop> Honestly, no idea.  Not my layer.  If we see Praveen surface again, he might be able to tell us.
[02:42:41] <mnot> (remote folks: please feel free to ask to have the camera moved)
[02:42:50] <craigt> thx
[02:43:56] fujisaki.tomohiro joins the room
[02:44:55] <craigt> mnot: will get you a steadicam for the next interim
[02:45:02] <mnot> heh
[02:45:08] Dirk Kutscher leaves the room
[02:45:23] <mnot> I’m actually looking at a camera that has a wider field of view (90 degrees vs. current 78 degrees)
[02:45:29] <ted.h> I want a foley operator instead, so that we get sound effects.
[02:45:54] <lucasp> mnot: during breaks can you take people to one side and get the "human side" of the story?
[02:46:04] <Sean Turner> ha1
[02:46:19] <mnot> hmm.
[02:46:59] <lucasp> or perhaps a big brother-style diary room
[02:51:05] Dirk Kutscher joins the room
[02:57:24] afrind leaves the room
[02:59:01] <Sean Turner> yeah mass sickness would be bad
[02:59:04] <mnot> back in one hour
[02:59:09] ekr leaves the room
[02:59:14] ted.h leaves the room
[02:59:22] jo.ietf leaves the room
[02:59:24] Brian Trammell leaves the room
[02:59:50] Mirja Kuehlewind leaves the room
[03:00:19] <Mike Bishop / Mike Bishop> I may or may not be back on for part of the afternoon.  If not, thanks for a great interim!
[03:00:29] Martin Thomson leaves the room
[03:00:51] <craigt> likewise, good to see you again…hope to see you irl in Chicago
[03:01:03] Dirk Kutscher leaves the room
[03:01:08] <Mike Bishop / Mike Bishop> Yes, I'm planning on being in Chicago.
[03:01:23] <lucasp> cya mike
[03:02:59] <craigt> lucasp: …so do we offer to do something in London 2018? …get the 'team' on-board and do an 8K, hlg hdr, 22.2 affair ;)
[03:03:17] Mike Bishop / Mike Bishop leaves the room: offline
[03:08:55] <lucasp> craigt: BBC BBQ (Big Bash with QUIC)
[03:10:28] <mnot> Remote folks: If you want to be not-remote for an interim, consider hosting us.
[03:10:31] <mnot> (Hi Craig!)
[03:10:47] <Daniel> =)
[03:12:08] <Daniel> my office is a bit on the smaller side for that ;-P
[03:12:43] <mnot> You have a yard, Daniel.
[03:12:54] <mnot> Mmm, Summer in Stockholm!
[03:12:57] <Daniel> good thinking
[03:15:44] <craigt> …did I miss any tentative locations for June?
[03:15:51] Martin Thomson joins the room
[03:16:02] <Daniel> no, just random ideas from people
[03:16:22] <mnot> You can supply a location for June, Craig :)
[03:16:31] <Daniel> +1
[03:18:03] <craigt> I can certainly ask; …but London would be very tricky given the amount of site works atm…
[03:18:16] <craigt> …something around 2018 would have been easier…
[03:18:20] <mnot> ah
[03:18:27] <craigt> …though there is always Salford
[03:18:45] <craigt> 'The North' does have more beautiful countryside
[03:19:02] <mnot> MediaCity?
[03:19:42] <mnot> ... and white city is *still* having works done?
[03:20:32] <craigt> yes, loads
[03:20:37] <craigt> it's shaping up nicely
[03:21:14] <craigt> I'll have a chat over the next few days…
[03:21:53] afrind joins the room
[03:22:18] <lucasp> they opened an Oliver Bonas, store, get that!
[03:28:09] Dirk Kutscher joins the room
[03:31:27] Lars Eggert leaves the room
[03:31:41] <mnot> thx, craig
[03:39:59] Dirk Kutscher leaves the room
[03:40:13] Dirk Kutscher joins the room
[03:40:23] Dirk Kutscher leaves the room
[03:43:13] Buck Krasic leaves the room
[03:43:20] Mirja Kuehlewind joins the room
[03:47:13] Brian Trammell joins the room
[03:54:15] mnot leaves the room
[03:56:36] mnot joins the room
[03:56:52] <mnot> restarting soon-ish
[03:59:20] Daniel is awake!
[04:00:06] afrind leaves the room
[04:00:13] afrind joins the room
[04:00:16] ted.h joins the room
[04:00:54] <craigt> …it's that horrible time when you don't know whether coffee is a good idea
[04:01:17] ekr joins the room
[04:02:00] <Daniel> we hear and see
[04:02:01] <ted.h> @craigt  it's always a good idea.
[04:03:48] <craigt> ted.h: it's 0400, and i'll be trying to sleep ~0800…
[04:04:05] <ted.h> @craigt, see that assumes sleep is a good idea.
[04:07:12] jo.ietf joins the room
[04:11:07] Dirk Kutscher joins the room
[04:16:16] spencerdawkins leaves the room
[04:16:50] <Brian Trammell> patrick: https://docs.google.com/document/d/1RjkilnUUziZWqwC2uvQDwg2Vq2PW1wBQGaqynirryC4/edit
[04:18:25] spencerdawkins joins the room
[04:26:48] Lars Eggert joins the room
[04:45:48] Dirk Kutscher leaves the room
[04:46:04] Dirk Kutscher joins the room
[04:49:47] ekr leaves the room
[04:50:40] Sean Turner leaves the room: Replaced by new connection
[04:52:23] ekr joins the room
[05:07:56] Sean Turner joins the room
[05:09:01] fujisaki.tomohiro leaves the room
[05:17:33] <ekr> http://memedad.com/meme/1089910
[05:17:50] <Sean Turner> haha
[05:17:53] <mnot> The first WG meme. So proud.
[05:18:42] <ted.h> That's not the first one, if you check ietfmemes.tumblr.com you will see more.
[05:18:52] <mnot> Right, but this is the first one INSIDE the WG.
[05:18:56] craigt waiting for an April 1st dkg bit draft
[05:19:01] <mnot> Thanks for the link tho - wasn’t aware of that.
[05:22:39] ted.h leaves the room
[05:22:45] Mirja Kuehlewind leaves the room
[05:22:53] Brian Trammell leaves the room
[05:22:54] ekr leaves the room
[05:23:56] Mirja Kuehlewind joins the room
[05:24:20] Martin Thomson leaves the room
[05:30:22] ekr joins the room
[05:32:30] Martin Thomson joins the room
[05:33:38] Brian Trammell joins the room
[05:34:00] ted.h joins the room
[05:34:31] Martin Thomson leaves the room
[05:35:00] Martin Thomson joins the room
[05:35:47] <Daniel> we're still on mute
[05:36:38] <spencerdawkins> Thanks to the remote folks for pointing that out ;-)
[05:38:24] <Daniel> obviously that was a test if we are still awake
[05:38:48] <spencerdawkins> the chairs needed that test :D
[05:39:59] Mirja Kuehlewind leaves the room
[05:45:52] <Martin Thomson> I think that we can reduce this to two problems: how packets are constructed before we have a connection and how packets are constructed once we have a connection.
[05:46:13] Mirja Kuehlewind joins the room
[06:07:04] fujisaki.tomohiro joins the room
[06:30:42] Dirk Kutscher leaves the room
[06:34:06] <ted.h> I think for Lucas's case you will need a version uplift anyway, but I may be wrong.
[06:34:43] Dirk Kutscher joins the room
[06:36:35] <Sean Turner> 42
[06:37:22] ranjeeth joins the room
[06:41:56] Dirk Kutscher leaves the room
[06:50:40] ranjeeth leaves the room
[06:50:50] Roni Even leaves the room
[06:50:59] <ted.h> @Sean Do you have a library for 42bit integer math handy?
[06:51:56] <craigt> I'd be interested to see some of the planning for those 23day numbers…bitrates, video format, segment duration…
[06:52:24] <Sean Turner> I was thinking hitchhiker's guide ;)
[06:52:24] <ted.h> @craigt https://github.com/quicwg/base-drafts/issues/108 has the back of the envelope.
[06:52:34] <ted.h> There is no detail beyond what's in there.
[06:52:39] Dirk Kutscher joins the room
[06:57:15] spencerdawkins leaves the room
[06:57:41] <craigt> The 500Hz premise lacks a bit of clairty. Assumes you known all about webtrc segmentisation/stream strategies (which unfortunately i don't)
[06:57:45] spencerdawkins joins the room
[07:10:42] <Martin Thomson> 1 << (x>>11)*(0x800 + (x&0xfff)) maybe?
[07:10:57] fujisaki.tomohiro leaves the room
[07:12:02] <Martin Thomson> the base2 logarithm requires more care
[07:19:45] <Martin Thomson> float t1=(float)x; uint32_t t2; memcpy(&t2, &t2, 4); t3=htonl(t2); assert(t3>>31 == 0); int8_t e = (t3 >> 23) - 127; assert(e >= 0); uint16_t m = (t3 >> 12) & 0x7ff; y = e << 11 | m;
[07:19:50] <ekr> ugh
[07:20:18] <Martin Thomson> confidence of correctness ~= 0
[07:21:53] Dirk Kutscher leaves the room
[07:22:17] <Martin Thomson> well, the answer is derived from uninitialized memory
[07:22:47] Dirk Kutscher joins the room
[07:24:56] Brian Trammell leaves the room
[07:26:36] <Daniel> 13, 17, 23 pages of minutes. All prime numbers!
[07:27:06] Brian Trammell joins the room
[07:28:40] <Sean Turner> clap clap!
[07:28:43] g.e.montenegro leaves the room
[07:28:45] <craigt> Thank you Akamai
[07:28:50] <Sean Turner> night all
[07:28:53] <Daniel> thanks for a good remote experience
[07:29:17] <craigt> Thank you non-remote attendees for all your participation.
[07:29:29] <lucasp> yes it worked well, I've added some feedback on mnot's poll
[07:29:30] <craigt> It was a good remote experience.
[07:29:34] ted.h leaves the room
[07:29:36] Mirja Kuehlewind leaves the room
[07:29:45] Sean Turner leaves the room
[07:29:50] <jo.ietf> It was also a nice live experience
[07:30:18] Lars Eggert leaves the room: Disconnected: closed
[07:30:26] Dirk Kutscher leaves the room
[07:30:28] spencerdawkins leaves the room
[07:30:28] spencerdawkins joins the room
[07:30:28] spencerdawkins leaves the room
[07:30:29] Brian Trammell leaves the room
[07:30:30] afrind leaves the room
[07:30:31] chi.jiun.su leaves the room: offline
[07:30:35] Daniel waves
[07:30:42] Kaoru Maeda leaves the room
[07:30:51] mnot leaves the room
[07:31:31] mcmanus leaves the room
[07:31:50] Martin Thomson leaves the room
[07:32:05] patrick leaves the room
[07:33:29] <craigt> Bye all, see you soon.
[07:33:38] craigt leaves the room
[07:36:08] jo.ietf leaves the room
[07:36:22] lucasp leaves the room
[07:43:03] ekr leaves the room
[07:44:28] ranjeeth joins the room
[07:46:48] ranjeeth leaves the room
[08:07:21] ranjeeth leaves the room
[08:53:52] lstewart leaves the room
[09:02:07] patrick joins the room
[09:40:29] spencerdawkins joins the room
[11:08:03] spencerdawkins leaves the room
[11:33:30] Mirja Kuehlewind joins the room
[11:41:49] Mirja Kuehlewind leaves the room
[12:20:28] Mirja Kuehlewind joins the room
[12:38:42] Mirja Kuehlewind leaves the room
[13:27:01] ekr joins the room
[13:33:10] ekr leaves the room
[14:06:58] Daniel leaves the room: offline
[15:02:46] Dirk Kutscher joins the room
[15:13:14] Dirk Kutscher leaves the room
[15:16:43] jo.ietf joins the room
[15:16:50] jo.ietf leaves the room
[15:18:29] Sean Turner joins the room
[15:18:35] Sean Turner leaves the room
[15:43:44] Mirja Kuehlewind joins the room
[16:34:04] Mirja Kuehlewind leaves the room: Disconnected: Replaced by new connection
[16:34:06] Mirja Kuehlewind joins the room
[17:13:55] ranjeeth joins the room
[17:14:18] Martin Thomson joins the room
[17:31:20] Mirja Kuehlewind leaves the room
[17:52:05] ranjeeth leaves the room
[18:02:48] Martin Thomson leaves the room
[18:11:18] patrick leaves the room: Disconnected: Replaced by new connection
[18:11:18] patrick joins the room
[18:13:01] Martin Thomson joins the room
[18:14:50] Martin Thomson leaves the room
[18:26:35] Martin Thomson joins the room
[18:28:19] Martin Thomson leaves the room
[18:40:10] Martin Thomson joins the room
[18:42:02] Martin Thomson leaves the room
[18:53:46] Martin Thomson joins the room
[18:55:29] Martin Thomson leaves the room
[19:07:20] Martin Thomson joins the room
[19:09:04] Martin Thomson leaves the room
[19:20:56] Martin Thomson joins the room
[19:23:08] Martin Thomson leaves the room
[19:34:30] Martin Thomson joins the room
[19:36:29] Martin Thomson leaves the room
[19:48:06] Martin Thomson joins the room
[19:49:57] Martin Thomson leaves the room
[20:01:41] Martin Thomson joins the room
[20:03:14] Martin Thomson leaves the room
[20:15:15] Martin Thomson joins the room
[20:16:26] ted.h joins the room
[20:16:50] Martin Thomson leaves the room
[20:17:30] ted.h leaves the room
[20:28:51] Martin Thomson joins the room
[20:30:45] Martin Thomson leaves the room
[20:42:25] Martin Thomson joins the room
[20:43:59] Martin Thomson leaves the room
[20:52:17] ekr joins the room
[20:56:00] Martin Thomson joins the room
[20:57:42] Martin Thomson leaves the room
[21:09:34] Martin Thomson joins the room
[21:11:07] Martin Thomson leaves the room
[21:23:10] Martin Thomson joins the room
[21:24:42] Martin Thomson leaves the room
[21:36:46] Martin Thomson joins the room
[21:38:15] patrick joins the room
[21:39:01] Martin Thomson leaves the room
[21:41:58] patrick leaves the room
[21:50:19] Martin Thomson joins the room
[21:51:34] ekr leaves the room
[21:52:42] Martin Thomson leaves the room
[22:03:56] Martin Thomson joins the room
[22:04:20] Mike Bishop / Mike Bishop joins the room
[22:04:58] Mike Bishop / Mike Bishop leaves the room
[22:05:14] Mike Bishop / Mike Bishop joins the room
[22:05:26] Martin Thomson leaves the room
[22:05:52] ekr joins the room
[22:11:51] Martin Thomson joins the room
[22:13:42] Martin Thomson leaves the room
[22:16:03] ekr leaves the room
[22:20:38] ekr joins the room
[22:23:35] Martin Thomson joins the room
[22:25:20] Martin Thomson leaves the room
[22:27:20] ekr leaves the room
[22:37:10] Martin Thomson joins the room
[22:37:13] ekr joins the room
[22:39:54] Martin Thomson leaves the room
[22:43:27] Martin Thomson joins the room
[22:45:12] Martin Thomson leaves the room
[22:45:25] Martin Thomson joins the room
[22:46:28] ekr leaves the room
[22:57:50] ekr joins the room
[22:58:11] Martin Thomson leaves the room
[23:01:45] Mike Bishop / Mike Bishop leaves the room
[23:03:45] ekr leaves the room
[23:07:06] Martin Thomson joins the room
[23:08:39] Martin Thomson leaves the room
[23:20:41] Martin Thomson joins the room
[23:22:41] Martin Thomson leaves the room
[23:33:55] Martin Thomson joins the room
[23:36:08] Martin Thomson leaves the room
[23:47:57] Martin Thomson joins the room
[23:49:24] Martin Thomson leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!