[00:40:06] Varun Singh joins the room
[00:49:23] JLCjohn joins the room
[00:58:54] Pete McCann joins the room
[00:59:44] <JLCjohn> is someone in Beijing?
[01:00:02] <Pete McCann> Not me.
[01:00:11] <Pete McCann> Audio stream is quiet for me.
[01:00:33] <JLCjohn> likewise...
[01:01:08] bob.briscoe joins the room
[01:01:56] <JLCjohn> Hi, Bob
[01:02:37] <JLCjohn> Audio is _very_ quiet!
[01:03:32] Atarashi Yoshifumi joins the room
[01:04:31] Alissa Cooper joins the room
[01:05:25] <JLCjohn> Alissa, are you in Beijing?
[01:05:56] <JLCjohn> (or anyone else, for that matter... BTW Bob Briscoe isn't)
[01:07:55] <bob.briscoe> Yup, I'm in the UK. I was in Beijing until Tue, but had to fly home.
[01:09:12] <bob.briscoe> Do we have a jabber scribe?
[01:11:05] <Alissa Cooper> I'm in Beijing
[01:11:50] <Alissa Cooper> downloading slides so I can give slide numbers, one sec
[01:11:52] <JLCjohn> evidently not... :^( !!!
[01:12:16] <Pete McCann> We cannot hear anything on the audio - has meeting started yet?
[01:13:01] <Alissa Cooper> yes, I will go find the AV people
[01:13:02] shep joins the room
[01:13:18] <JLCjohn> Thanks, but without audio it's not terribly useful...
[01:13:40] <bob.briscoe> With volume full up I can just make out that Andrea Soppera is talking. Only because I know the rhythm of his voice and I know he should be talking in my place according to the agenda.
[01:13:56] <JLCjohn> (not sure why my responses were delayed)
[01:14:08] <bob.briscoe> But I cannot hear anything he is saying.
[01:14:47] <Pete McCann> I opened a trouble ticket with the NOC.
[01:14:51] <bob.briscoe> Thanks
[01:15:47] <Alissa Cooper> supposedly the AV people are on their way
[01:16:59] Pete McCann leaves the room
[01:17:04] <Alissa Cooper> andrea is wrapping up abstract-meth
[01:17:30] Pete McCann joins the room
[01:17:53] <Alissa Cooper> marcelo asking for comments on whether doc is ready to be adopted as WG item
[01:18:36] <JLCjohn> Can you summarize who responds?
[01:19:08] <Alissa Cooper> dave mcdysan asking where new use cases go since on ml ppl have said they aren't covered in either use cases or abstract-meth
[01:20:17] <Alissa Cooper> dave's use cases would need more/different mechanism than what is in abstract-meth
[01:21:02] <Alissa Cooper> leslie daigle: supports adoption as WG item
[01:21:45] <Alissa Cooper> dave: fine with adopting this doc if we can still talk about further use cases
[01:22:23] <Alissa Cooper> rich woundy: want to be able to move some appendix material from use-cases into abstract-mech
[01:22:53] <Alissa Cooper> marcelo: that will be easier once it is a WG item
[01:23:48] <Alissa Cooper> took a vote: many people in favor, none opposed
[01:23:57] <Alissa Cooper> adoption as WG item complete
[01:24:03] <Alissa Cooper> rich up next
[01:24:25] <Alissa Cooper> how is the audio?
[01:24:35] <JLCjohn> same :^(
[01:24:37] <Pete McCann> Still unusable.
[01:31:37] <JLCjohn> (more noise on audio now, but still useless)
[01:32:49] <JLCjohn> Leslie Daigle was audible, what was she asking?
[01:34:05] <Alissa Cooper> it was me (or Mirja)
[01:34:23] <Alissa Cooper> the accountability use case was not on the slide (slide 4?). so I was asking where it was.
[01:34:24] <JLCjohn> sorry (audio _is_ that bad!
[01:42:25] <Alissa Cooper> rich: asking for comments on aligning terminology with mechanism draft and moving appendix material to mechanism
[01:42:27] <shep> looks like they are fiddling with the audio gear now.
[01:42:38] <Alissa Cooper> dave: not objecting to that
[01:42:43] <JLCjohn> sounds like it, too!
[01:43:08] <Alissa Cooper> is it fixed?
[01:43:14] <Alissa Cooper> joel is walking away
[01:43:16] <bob.briscoe> Audio is perfect now.
[01:43:19] <Alissa Cooper> cool
[01:43:26] <JLCjohn> audio mostly good...
[01:43:31] <Alissa Cooper> we are on rich's slide 14
[01:44:02] <JLCjohn> audio glitch, came back OK
[01:51:27] <bob.briscoe> I didn't think we were asked to remove it. Just to tone it down. Can someone say this for me?
[01:51:27] <Alissa Cooper> many in favor
[01:51:58] <bob.briscoe> OK, too late - don't bother.
[01:52:23] <Alissa Cooper> rich said he will have a conversation offline with you
[01:52:50] <Alissa Cooper> dave up next
[02:02:39] <bob.briscoe> If the heavy volume users are using a LEDBAT-like congestion control, and the light volume users are using TCP, the light volume users could be the heavy cost users
[02:02:41] <JLCjohn> The 20%/80% is pretty much a given for human activity.
[02:09:57] <bob.briscoe> ConEx information is designed to be the best way to measure when burstiness matters and when it doesn't. It works over both queue timescales and days/weeks months, by design. Would have to explain this on the list. Pls relay this.
[02:11:37] <Alissa Cooper> moved on to next slide
[02:21:01] <bob.briscoe> As Marcelo is saying, in ConEx we need to hold in mind that people will want to build these things on top of ConEx. We should encourage experience reports from people who try to build this on ConEx.
[02:40:51] <bob.briscoe> We ought to limit use-cases we choose to those that skeptics cannot disagree with (esp. skeptics on the IESG). No need to be exhaustive.
[02:52:36] <bob.briscoe> Q for clarification: If a new hop-by-hop says silently forward if don't understand and a router doesn't understand a new hop-by-hop, will this punt to the slow path?
[02:53:51] <Alissa Cooper> sorry, did this question get answered? was distracted
[02:54:03] <bob.briscoe> Nope. Pls ask it.
[02:55:09] <bob.briscoe> Rephrasing to make it understandable. Q for clarification: If a new hop-by-hop has the bit set that means "silently forward if don't understand" and a router indeed doesn't understand this new option, will this punt to the slow path?
[02:59:11] <bob.briscoe> Clarification Q: "flow label field ... has not been widely used". Has it been used /at all/ in production networks?
[03:02:25] <bob.briscoe> ConEx doesn't need mutable
[03:03:53] <bob.briscoe> Would this fly in 6man?: Constrain any network node using flow label to mask the last 4 bits (e.g. for ECMP).
[03:04:46] <Alissa Cooper> can you phrase a q if you still have a q -- I am not following this real well
[03:05:49] <bob.briscoe> Might this proposal be acceptable to 6man?: Any network node using flow label (e.g. for ECMP) must mask the last 4 bits.
[03:08:36] YJS joins the room
[03:26:43] YJS leaves the room
[03:29:59] Alissa Cooper leaves the room
[03:30:23] <bob.briscoe> Thanks Alissa for taking on scribe
[03:30:29] Atarashi Yoshifumi leaves the room
[03:30:30] <bob.briscoe> Too late!
[03:30:46] bob.briscoe leaves the room
[03:33:17] JLCjohn leaves the room
[03:42:24] shep leaves the room
[04:28:45] Pete McCann leaves the room
[05:13:51] Alissa Cooper joins the room
[05:17:44] Alissa Cooper leaves the room
[05:19:16] Atarashi Yoshifumi joins the room
[05:19:48] Atarashi Yoshifumi leaves the room
[06:56:03] Varun Singh leaves the room
[07:37:16] Varun Singh joins the room
[07:39:48] Varun Singh leaves the room