[04:26:55] Qin W leaves the room
[04:41:25] Qin W joins the room
[04:56:24] gorryf joins the room
[05:16:40] Qin W leaves the room
[07:12:47] Meetecho joins the room
[07:16:53] spencerdawkins joins the room
[07:20:03] Julien Maisonneuve joins the room
[07:20:03] Juhamatti Kuusisaari joins the room
[07:20:03] Philip Eardley joins the room
[07:20:03] Daniel Havey joins the room
[07:20:03] Chris Lemmons joins the room
[07:20:03] Pete Resnick joins the room
[07:20:03] Peter Heist joins the room
[07:20:03] Paolo Saviano joins the room
[07:20:03] Ken Takayama joins the room
[07:20:03] Spencer Dawkins joins the room
[07:20:03] Jonathan Morton joins the room
[07:20:03] Christian Huitema joins the room
[07:20:03] Tobia Castaldi joins the room
[07:20:03] Carsten Bormann joins the room
[07:20:25] Ken Takayama leaves the room
[07:20:34] Spencer Dawkins leaves the room
[07:20:38] Spencer Dawkins joins the room
[07:20:43] Ken Takayama joins the room
[07:20:46] Ken Takayama leaves the room
[07:21:07] Bob Hinden joins the room
[07:21:28] Nicolas Kuhn joins the room
[07:21:47] Timothy Panton joins the room
[07:21:56] Frode Kileng joins the room
[07:22:09] Alessandro Amirante joins the room
[07:22:24] <Bob Hinden> Hello
[07:22:42] frodek joins the room
[07:23:40] Ingemar Johansson joins the room
[07:24:23] 菊池正史 joins the room
[07:24:26] Gorry Fairhurst joins the room
[07:25:02] <Christian Huitema> Hello!
[07:25:18] 菊池正史 leaves the room
[07:25:22] Wesley Eddy joins the room
[07:25:31] Mihail Yanev joins the room
[07:25:31] dahavey joins the room
[07:25:50] Alissa Cooper joins the room
[07:26:05] Mike Kosek joins the room
[07:26:35] Iain Learmonth joins the room
[07:26:59] Qin W joins the room
[07:27:04] <Ingemar Johansson> Good morning
[07:27:07] Michael Scharf joins the room
[07:27:08] Renan Krishna joins the room
[07:27:10] Renan Krishna leaves the room
[07:27:18] Mike Coleman joins the room
[07:27:29] irl joins the room
[07:27:33] <Spencer Dawkins> Looking at the chair video - "in my defense, I was left unsupervised".
[07:27:37] alex-meetecho joins the room
[07:27:42] Renan Krishna joins the room
[07:28:11] Mirja Kühlewind joins the room
[07:28:17] Ruediger Geib joins the room
[07:28:18] Ted Hardie joins the room
[07:28:25] Philip Eardley leaves the room
[07:28:25] Kuniyasu Suzaki joins the room
[07:28:25] Kuniyasu Suzaki leaves the room
[07:28:28] Philip Eardley joins the room
[07:28:29] David Black joins the room
[07:28:34] David Lawrence joins the room
[07:28:40] Stephen McQuistin joins the room
[07:28:48] <Chris Lemmons> Yay, your mic works. I was in the middle of trying to figure out why I couldn't hear anything. :D
[07:28:48] Mike Bishop joins the room
[07:28:59] Yoshifumi Atarashi joins the room
[07:29:13] David Schinazi joins the room
[07:29:13] Chi-Jiun Su joins the room
[07:29:16] Andrew McGregor joins the room
[07:29:21] Markus Amend joins the room
[07:29:21] Joakim Misund joins the room
[07:29:23] Jake Holland joins the room
[07:29:30] Igor Gashinsky joins the room
[07:29:33] Qin W leaves the room
[07:29:42] Kazuaki Ueda joins the room
[07:29:49] Ruediger Geib leaves the room
[07:29:50] Ram Ranganathan joins the room
[07:29:53] Richard Scheffenegger joins the room
[07:29:54] Greg Wood joins the room
[07:29:58] Donald Eastlake joins the room
[07:30:01] Ana Custura joins the room
[07:30:01] Ram Ranganathan leaves the room
[07:30:03] Greg White joins the room
[07:30:04] Sebastian Moeller joins the room
[07:30:05] Bob Briscoe joins the room
[07:30:06] Ram Ranganathan joins the room
[07:30:07] Mike Kosek leaves the room
[07:30:08] Li Yizhou joins the room
[07:30:10] Mike Kosek joins the room
[07:30:16] Esko Dijk joins the room
[07:30:20] Martin Duke joins the room
[07:30:21] <Mirja Kühlewind> day light here now! yeah!
[07:30:21] <Jonathan Morton> I do have daylight, in Finland, which is downright remarkable at this time of year
[07:30:25] Mike Kosek leaves the room
[07:30:39] Peter van Dijk joins the room
[07:30:40] <Spencer Dawkins> Three chairs, one in each IETF region timezone, really does need to be a BCP. TCPM showed the way.
[07:30:42] Kazuhisa Matsuzono joins the room
[07:30:52] Jiao Kang joins the room
[07:31:10] Mike Kosek joins the room
[07:31:12] Brian Trammell joins the room
[07:31:16] Ruediger Geib joins the room
[07:31:17] Lv Yunping joins the room
[07:31:21] Michael Scharf leaves the room
[07:31:24] Michael Scharf joins the room
[07:31:25] Frode Sorensen joins the room
[07:31:27] <Andrew McGregor> Sure, I'll do that again
[07:31:30] Igor Lubashev joins the room
[07:31:43] Zaheduzzaman Sarker joins the room
[07:31:54] <Brian Trammell> i can backup as well
[07:32:10] Colin Perkins joins the room
[07:32:18] Magnus Westerlund joins the room
[07:32:24] Francois Ortolan joins the room
[07:32:28] Hirochika Asai joins the room
[07:32:31] Stuart Cheshire joins the room
[07:32:34] dschinazi joins the room
[07:32:45] Wei Pan joins the room
[07:32:48] Wei Pan leaves the room
[07:32:54] <Jake Holland> https://codimd.ietf.org/notes-ietf-109-tsvwg_2?both
[07:32:56] Nick Banks joins the room
[07:33:03] <Brian Trammell> i've got https://codimd.ietf.org/notes-ietf-109-tsvwg?both
[07:33:06] Mohit Tahiliani joins the room
[07:33:17] Jana Iyengar joins the room
[07:33:17] <dahavey> Brian has the right link.
[07:33:18] Dmitri Tikhonov joins the room
[07:33:26] Francois Ortolan leaves the room
[07:33:36] Lars Eggert joins the room
[07:33:51] Kazuho Oku joins the room
[07:34:04] Zhen Cao joins the room
[07:34:11] Hannu Flinck joins the room
[07:34:28] Josh Cohen joins the room
[07:34:38] Jörg Deutschmann joins the room
[07:34:43] <David Black> Go to bottom of that page - 2nd session agenda is pasted there.
[07:34:44] Vidhi Goel joins the room
[07:34:58] Xavier de Foy joins the room
[07:34:58] Anna Brunstrom joins the room
[07:35:01] Koen Schepper joins the room
[07:35:10] Marcus Ihlar joins the room
[07:35:23] <Carsten Bormann> https://codimd.ietf.org/notes-ietf-109-tsvwg?both
[07:35:25] <Bob Briscoe> Thx, Wes
[07:35:40] Xavier de Foy leaves the room
[07:35:42] Corinne Cath joins the room
[07:35:43] Francois Ortolan joins the room
[07:35:47] Xavier de Foy joins the room
[07:37:29] Sergey Fomin joins the room
[07:37:58] Peter van Dijk leaves the room
[07:38:54] Greg Wood leaves the room
[07:38:58] Greg Wood joins the room
[07:39:30] Lucas Pardue joins the room
[07:39:30] Tommy Pauly joins the room
[07:39:37] Simon Hicks joins the room
[07:39:48] Mark McFadden joins the room
[07:39:52] Jingrong Xie joins the room
[07:39:57] Mike English joins the room
[07:41:57] Erik Kline joins the room
[07:42:10] John Border joins the room
[07:42:13] Simon Hicks leaves the room
[07:42:16] Timothy Panton leaves the room
[07:42:53] <Sebastian Moeller> I object to the first row, "across a range of conditions" is as correct as irrelevant. A number of relevant cases are not included in that assessment.
[07:42:56] Corinne Cath leaves the room
[07:44:43] <Gorry Fairhurst> Above comment to slide 3.
[07:45:08] Steffen Klassert joins the room
[07:45:48] Eric Orth joins the room
[07:46:53] Juan Cerezo joins the room
[07:47:32] Pyung-Soo Kim joins the room
[07:47:34] Mihail Zverev joins the room
[07:47:52] Pete Resnick leaves the room
[07:48:29] <Gorry Fairhurst> Slide 5
[07:48:36] <Sebastian Moeller> Sorry, confused by meetecho
[07:49:02] Eric Kinnear joins the room
[07:49:04] Dmitri Tikhonov leaves the room
[07:49:23] <Jake Holland> i didn't catch the question, that was from wes asking if there's something else to show with no response?  should that get into the notes as a question on slide 5?
[07:49:41] <Bob Briscoe> I think there was a hand remaining stuck up
[07:49:42] <Jonathan Morton> Sebastian had left his hand up, that's all
[07:49:47] <Gorry Fairhurst> no action in notes needed
[07:49:51] Yuexia Fu joins the room
[07:49:52] <Jake Holland> k thanks.
[07:50:59] Yuexia Fu leaves the room
[07:52:16] Pyung-Soo Kim leaves the room
[07:53:00] 藤﨑 智宏 joins the room
[07:54:50] <Colin Perkins> To make sure I’m understanding: the issue here is a flow sending ECT(1) that expects L4S marking, sharing a queue with a flow that sends ECT(0) and expects RFC3168 marking, where both receive RFC3168 marking?
[07:55:05] <Jake Holland> yes, that's issue #16
[07:55:24] Pete Resnick joins the room
[07:55:25] <Gorry Fairhurst> Slide 6
[07:55:27] <Bob Briscoe> Yes, both receive CE, 'cos of the CE ambiguity in L4S
[07:56:46] <Andrew McGregor> Routers exist that can do ECT1 as NotECT in hardware, but it's not all that common AFAIK.
[07:57:00] Frode Sorensen leaves the room
[07:57:11] Frode Sorensen joins the room
[07:57:17] <Jonathan Morton> it's technically contrary to RFC-3168
[07:57:34] Li Yizhou leaves the room
[07:57:39] <Sebastian Moeller> Fair enough, if such machines exist in the wild adding that section to the RFC makes some sense.
[07:57:39] Frode Sorensen leaves the room
[07:59:04] Corinne Cath joins the room
[07:59:56] Frode Sorensen joins the room
[08:00:01] <Sebastian Moeller> How is retroactively changing rfc3168 help with the deployed base?
[08:01:10] Anna Brunstrom leaves the room
[08:01:18] ihlar joins the room
[08:01:21] Frode Sorensen leaves the room
[08:01:41] Frode Sorensen joins the room
[08:01:51] <Gorry Fairhurst> They fail to comply with RFC3168, it doesn't remove/reconfigure them!
[08:01:52] <Sebastian Moeller> Here is rfc3168:The
   ECN-Capable Transport (ECT) codepoints '10' and '01' are set by the
   data sender to indicate that the end-points of the transport protocol
   are ECN-capable; we call them ECT(0) and ECT(1) respectively.  The
   phrase "the ECT codepoint" in this documents refers to either of the
   two ECT codepoints.  Routers treat the ECT(0) and ECT(1) codepoints
   as equivalent.
[08:02:28] Frode Sorensen leaves the room
[08:02:42] <Jonathan Morton> yes, that last sentence is what I was referring to
[08:02:47] <Carsten Bormann> That is a statement of fact.
[08:02:55] <Carsten Bormann> Does anything rely on that?
[08:03:03] <Bob Briscoe> But that's not MUST treat them as equivalent
[08:03:04] <Carsten Bormann> If not, this is not an interoperability requirement.
[08:03:13] <Jonathan Morton> Nonce Sum did
[08:03:14] <Carsten Bormann> Bob: exactly.
[08:03:26] <Gorry Fairhurst> On PCN - I meant: RFC6660
[08:03:44] Corinne Cath leaves the room
[08:03:49] Corinne Cath joins the room
[08:04:36] Corinne Cath leaves the room
[08:04:37] <Gorry Fairhurst> although the arch is RFC5559
[08:05:05] Corinne Cath joins the room
[08:05:11] Juhamatti Kuusisaari leaves the room
[08:05:14] Juhamatti Kuusisaari joins the room
[08:05:39] <Christian Huitema> Do we have documentation on end-to-end assessment of AQM at bottleneck?
[08:05:59] Josh Cohen leaves the room
[08:06:21] <David Black> RFC 8311 allows ECT(0) and ECT(1) to be treated differently - see 3rd paragraph in Section 4.2.
[08:06:21] Frode Sorensen joins the room
[08:06:24] <David Black> https://tools.ietf.org/html/rfc8311#section-4.2
[08:06:38] Corinne Cath leaves the room
[08:06:57] <Bob Briscoe> See https://arxiv.org/abs/1911.00710 for instance.
[08:07:09] Corinne Cath joins the room
[08:07:15] <Bob Briscoe> That was @Christian
[08:07:16] Corinne Cath leaves the room
[08:07:19] Corinne Cath joins the room
[08:07:30] <Sebastian Moeller> Fair enough, but that does not help with gear in the field that complies to rfc3168 and was put in service before rfc8311 was published.
[08:07:47] Martin Thomson joins the room
[08:07:56] <Sebastian Moeller> This is IMHO more about the existing internet and not so much a fully RFC compliant and up to date internet, no?
[08:08:06] Corinne Cath leaves the room
[08:08:16] <Carsten Bormann> Both are of interest.
[08:08:41] Corinne Cath joins the room
[08:08:43] Corinne Cath leaves the room
[08:08:46] Corinne Cath joins the room
[08:09:05] <Bob Briscoe> Certainly - that's why we have to have multiple strategies - cos some kit will have more flexible config than others
[08:09:11] <David Black> Fair enough - the question would be whether such gear exists and what the nature of the dependency is.  RFC 8311 reclassified the ECN Nonce (RFC 3540) as Historic.
[08:09:33] Corinne Cath leaves the room
[08:09:43] <Sebastian Moeller> The leanest recommendation would be, no ECN on links that are likely to carry tunneled traffic.Unfortunatley that would be almost all links...
[08:09:59] <Bob Briscoe> Why?
[08:10:34] <Carsten Bormann> The answer to the question "is a router that distinguishes ECT0 and ECT1 non-compliant" is probably "no", even just from RFC 3168.
[08:10:59] <Colin Perkins> This seems to not be a severe problem, and not a likely problem. Not sure it’s a practical concern.
[08:11:12] <Sebastian Moeller> @Stuart, does this includes tests over wifi?
[08:11:22] <Gorry Fairhurst> Tunnels are not a new problem
[08:12:02] Juhamatti Kuusisaari leaves the room
[08:12:06] Juhamatti Kuusisaari joins the room
[08:12:20] Martin Thomson leaves the room
[08:12:25] <dschinazi> @Sebastien last time I looked at the APple data it applied on all interface types
[08:12:39] Martin Thomson joins the room
[08:12:59] Martin Thomson leaves the room
[08:13:00] <Sebastian Moeller> @dschinazi, thanks.
[08:13:00] Maxime Piraux joins the room
[08:13:02] Martin Thomson joins the room
[08:13:11] <Brian Trammell> RIPE MAT
[08:13:18] <Brian Trammell> (says the RIPE MAT co-chair :)
[08:13:27] <Jana Iyengar> What is non-zero percent, Jake?
[08:13:57] Maxime Piraux leaves the room
[08:14:47] <Spencer Dawkins> OMG, we're talking about how many ECN marking routers can forward packets on the head of a pin ...
[08:14:50] <Jonathan Morton> IIRC, a small number of AS's had quite high penetration
[08:15:34] <Jana Iyengar> Chairs -- we are 45 minutes into the agenda time, with 30 minutes planned for the L4S discussion. Are we planning to get to the rest of this?
[08:15:54] <David Black> We're planning to run for 90 minutes, not just 60 ...
[08:16:06] <Jana Iyengar> Ah, ok, thanks.
[08:16:12] <dschinazi> Isn't meetecho going to cut you off after 5 min past time?
[08:16:19] <Sebastian Moeller> Just a data point, OpenWrt comes with rfc3168 smart queueing, that needs to be enabled by the user. So there er no reliable numbers on how many devices we are talking about here.
[08:16:29] <Meetecho> dschinazi: no we don't cut rooms off automatically
[08:16:31] <David Black> @MartinDuke extended this, IIUC
[08:16:36] <dschinazi> Ah, thanks meetecho!
[08:16:37] <Martin Duke> i did not
[08:16:57] <Spencer Dawkins> Can we go 25 minutes over, for a pre-plenary bio break? :o
[08:17:04] <David Black> @Meetecho - thanks - we will need to run to nearly 90 minutes ...
[08:17:29] <Jonathan Morton> we know a few EU ISPs have nearly universal ECN deployment in middleboxes
[08:17:31] <dschinazi> +1 to Spencer, it would be nice to have a break before the plenary
[08:17:33] <Carsten Bormann> The plenary has a built-in bio break at the start, no?
[08:17:36] <Meetecho> David Black: won't that overlap with the plenary?
[08:17:56] <Sebastian Moeller> Also relevant data point, those OpenWrt user enabling SQM most likely are interested in low latency.
[08:18:12] irl leaves the room
[08:18:29] <David Black> @Meetecho - going to 15:55 instead of 15:30, plenary starts at 16:00.
[08:18:47] <Brian Trammell> I'll need to drop out of backup note taking at 09:25
[08:18:56] Erik Kline leaves the room
[08:19:06] <Brian Trammell> have a $dayjob meeting in the break.
[08:19:14] <Brian Trammell> (would have been nice to stay on time :) )
[08:19:16] <Peter Heist> anecdotal note: my ISP in Czech uses fq_codel in the backhaul
[08:19:28] <Meetecho> David Black: ack
[08:19:31] <Colin Perkins> I'd encourage stating the poll results in the audio, so the recording captures them
[08:20:11] <Gorry Fairhurst> Ok will do.
[08:20:31] <Jake Holland> it's in the notes.  15 read it, 19 did not.  88 participants in room.
[08:22:08] <Martin Duke> Jake's maprg presentation: https://datatracker.ietf.org/meeting/interim-2020-maprg-01/materials/slides-interim-2020-maprg-01-sessa-latency-aqm-observations-on-the-internet-01
[08:23:15] <Gorry Fairhurst> /88/101
[08:24:47] <Jake Holland> also here: https://www.youtube.com/watch?v=1326B7YYwLM&t=1h12m19s
[08:25:06] cabo joins the room
[08:25:17] <Martin Duke> @David Black I didn't request more time because we completed the agenda. I am not going to be able to go for 25 minutes extra
[08:25:56] <Gorry Fairhurst> @Martin: You're AD - tell us what we need to do.
[08:26:31] <Martin Duke> Can Bob hit the highlights?
[08:27:50] Sebastian Moeller leaves the room
[08:27:53] Sebastian Moeller joins the room
[08:28:02] <David Black> @martin: When do you have to drop?
[08:28:51] <Jake Holland> @gorry: it was 88 at the time of the poll, not 101.  not sure why the fluctuation, but it's there in the poll report and matches what the participants count had at that time, i was looking.
[08:29:04] <Jake Holland> i think new people connected since then. unclear why :)
[08:29:17] <Martin Duke> :40 ?
[08:29:18] <Gorry Fairhurst> Anyway, record 88
[08:29:26] <Jake Holland> yep
[08:29:28] <Martin Thomson> I'm sure that there are many people who depend on the gap.
[08:29:40] <David Black> Ok, let's get that deadline inserted at end of current slide.
[08:29:45] <Jake Holland> maybe we should have a flame war about it.
[08:30:27] Ruediger Geib leaves the room
[08:30:49] Sebastian Moeller leaves the room
[08:30:51] <Jana Iyengar> Um, let's hit the highlights and then get to discussion?
[08:30:52] Sebastian Moeller joins the room
[08:31:05] Nicolas Kuhn leaves the room
[08:31:05] <Philip Eardley> Apparently the poll is the number of people, the participants double counts those on jabber
[08:31:16] Mike Bishop leaves the room
[08:31:27] <Martin Thomson> Chairs have a counter that they can set.
[08:31:41] Brian Trammell leaves the room
[08:31:43] <Martin Thomson> It's one of the widgets at the top.
[08:31:44] Nick Banks leaves the room
[08:31:47] <Martin Duke> I agree with Jana -- let's have a break after this slide for some commetns
[08:32:33] Mike Coleman leaves the room
[08:32:41] Ted Hardie leaves the room
[08:33:02] Sebastian Moeller leaves the room
[08:33:05] <Gorry Fairhurst> We'll do this anyway,... prepare questions if any!
[08:33:09] Sebastian Moeller joins the room
[08:33:29] Sebastian Moeller leaves the room
[08:33:32] Sebastian Moeller joins the room
[08:34:05] Martin Thomson leaves the room
[08:34:16] Paolo Saviano leaves the room
[08:34:42] alex-meetecho leaves the room
[08:34:45] <Sebastian Moeller> Nope the biggest issue is that dualq is currently broken.
[08:34:58] Iain Learmonth leaves the room
[08:35:08] Pete Resnick leaves the room
[08:35:15] Huanan Chen joins the room
[08:35:40] Frode Sorensen leaves the room
[08:35:45] Frode Sorensen joins the room
[08:36:05] Frode Sorensen leaves the room
[08:36:07] <Colin Perkins> +1 to Jana
[08:36:11] Frode Sorensen joins the room
[08:36:26] <Richard Scheffenegger> @jana +1
[08:36:31] <Mirja Kühlewind> yes let's get the RFCs out and start the experiment!
[08:36:38] <Ingemar Johansson> +1
[08:36:40] Ana Custura leaves the room: Disconnected: BOSH client silent for over 60 seconds
[08:37:06] <Philip Eardley> +1
[08:37:10] <Mirja Kühlewind> what's the issue about dualq?
[08:37:14] <Mark McFadden> +1
[08:37:23] <Sebastian Moeller> Sorry, an experiment does not require an RFC, and atm general safety of L4S has not been demonstrated.
[08:37:35] <Jonathan Morton> I do not think L4S is in a state where it can be deployed
[08:37:47] Ana Custura joins the room
[08:37:57] Bob Hinden leaves the room
[08:38:02] Bob Hinden joins the room
[08:38:09] <Mirja Kühlewind> the intended status is experimental to get large scale experience going
[08:38:13] Tommy Pauly leaves the room
[08:38:19] Tommy Pauly joins the room
[08:38:38] <Jonathan Morton> I think WGLC at least has to wait for l4s-ops to be completed
[08:39:23] <Mirja Kühlewind> I disagree, let's get things moving
[08:39:26] <Richard Scheffenegger> @jana +1
[08:39:32] Huanan Chen leaves the room
[08:39:40] <Richard Scheffenegger> especially the arch draft doesn't have any dependencies
[08:39:43] <Sebastian Moeller> DualQ does not equitably share between its two classes in a number of relevant cases and requires active fudges in the protocol to work as expected and promises.That simply is not a robust and reliable
[08:39:43] <Koen Schepper> +1 Only the real experiment can bring more engagement and experience
[08:39:57] Kazuho Oku leaves the room
[08:40:24] <Colin Perkins> Who's running the queue?
[08:40:30] <Greg White> +1 to starting WGLC and to starting the experiment
[08:40:59] <Ingemar Johansson> The opts draft is not required before a WGLC of the L4S, on the contrary, I would see that the ops draft will in part be a result of the experiments
[08:41:03] 藤﨑 智宏 leaves the room
[08:41:33] <Carsten Bormann> The purpose of an experimental RFC is to agree what exactly the experiment is
[08:41:48] <Sebastian Moeller> sorry robust and reliable design. Making the endpoints responsible for safety and equitable sharing is an odd choice, since the same endpoints will profit from unequal sharing.
[08:42:08] <Andrew McGregor> The ops draft is required... but it shouldn't be last called until after the experiment is done, because it will become the ops document for the eventual standard.
[08:43:05] <Ingemar Johansson> +1
[08:43:06] Donald Eastlake leaves the room
[08:43:53] Martin Duke leaves the room
[08:44:12] <Jonathan Morton> there is also a risk that deploying the experiment will "eat" the ECT(1) codepoint and prevent it from being used in a different way later
[08:44:34] <Jonathan Morton> that is why I want to see more on l4s-ops first
[08:44:51] <dschinazi> You can avoid that issue by having an end date in the experiment code
[08:46:17] Tommy Pauly leaves the room
[08:47:04] <Jonathan Morton> frankly, I am convinced that the basic idea of redefining CE was the fundamental mistake
[08:47:17] Magnus Westerlund leaves the room
[08:47:19] <Jonathan Morton> you can't fix that with a magic CC
[08:47:20] <Gorry Fairhurst> I think Jonathan is correct, the intention is that TSVWG will assign ECT(1) for this, until the experiment is discontinued.
[08:47:30] <Mirja Kühlewind> we discussed the question if RFC3168 marking is deployed. If the question is no, L4S deployment is much simpler. If those who have deployed RFC3168 are the first once to update to L4S that fine as well.
[08:47:30] Sebastian Moeller leaves the room
[08:47:33] Sebastian Moeller joins the room
[08:47:35] Stephen McQuistin leaves the room
[08:47:41] <Peter Heist> can we get experimental results protected with a dscp first, then commit to ECT(1) later?
[08:47:44] <Mirja Kühlewind> if the answer is no I meant
[08:47:45] Stephen McQuistin joins the room
[08:48:21] <Gorry Fairhurst> @Peter - I think you are revisiting an earlier disucssion, we've decided on using ECT(1)
[08:49:42] <Bob Briscoe> For open questions, open issues, etc that have been collected to define the experiment, see https://tools.ietf.org/html/draft-ietf-tsvwg-ecn-l4s-id-12#section-6
[08:49:51] <Colin Perkins> I don't see how that data can be found without trial deployments
[08:49:58] <Jana Iyengar> +1 to Colin
[08:50:03] <Ingemar Johansson> +1
[08:50:07] <dschinazi> +1 to Cloin
[08:50:08] <Mirja Kühlewind> Not sure how Sebastian defines safe?
[08:50:16] <dschinazi> +1 to Colin (sorry typo)
[08:50:32] <Ram Ranganathan> +1 to Colin.
[08:50:35] <Colin Perkins> Cloin is not the worst thing I've been called :)
[08:50:41] <dschinazi> :D
[08:50:44] Francois Ortolan leaves the room
[08:50:52] Greg Wood leaves the room
[08:52:01] <Mirja Kühlewind> RTT dependency is a CC experimentation question (also for non L4S conform CC) but not something that needs to address in the draft we have in tsvwg
[08:52:02] <Jonathan Morton> just in lab tests, we can easily find cases where L4S smacks AIMD traffic into the floor
[08:52:34] Eric Orth leaves the room
[08:52:40] <Jonathan Morton> don't need to deploy on a real network to find those problems
[08:52:40] <Ingemar Johansson> +1 That RTT dependency problem was a problem in the RMCAT standardisation too
[08:52:49] <Mirja Kühlewind> let's deprecate RFC3168 marking and deploy L4S! :-)
[08:52:57] Mike English leaves the room
[08:53:04] <Gorry Fairhurst> +1 Mirja - I think we have seen thoughts about control-loops and RTT depenendency before in CC discussion, and we need to keep working on that.
[08:53:11] <Jana Iyengar> Chairs -- do you want to ask for a show of hands on this?
[08:53:28] <Gorry Fairhurst> whaty is this....
[08:53:49] <cabo> So, in summary, the purpose of this experiment is to kill Classic and force people to work on L4S.  Surprisingly, I like the idea
[08:53:54] <Jonathan Morton> may I remind everyone that Challenger happened because people said "let's go, the risks are one in a million" despite evidence to the contrary
[08:53:54] Bob Hinden leaves the room
[08:54:39] <Peter Heist> list of devices with fq_codel deployed, signaling with RFC3168 by default: https://github.com/heistp/l4s-tests/#deployments-of-fq_codel
[08:55:00] <Mirja Kühlewind> fq is fine
[08:55:02] Mike Kosek leaves the room
[08:55:11] <Peter Heist> with tunnels, it is not
[08:55:33] <Jana Iyengar> @Jonathan: I .... don't think the comparison of L4S to the Challenger is meaningful
[08:55:38] <Jonathan Morton> incidentally, Jake's data is coloured by the number of endpoints that request ECN negotiation; it's still not default in Windows IIRC
[08:55:55] <Gorry Fairhurst> @Peter: If these are new, could they change to reflect IETF consensus when this is called?
[08:56:11] <Jana Iyengar> It wasn't integrated -- RED was a separate bit :-)
[08:56:13] Alissa Cooper leaves the room
[08:56:17] Alessandro Amirante leaves the room
[08:56:21] Frode Sorensen leaves the room
[08:56:49] Frode Kileng leaves the room
[08:56:52] <Peter Heist> @Gorry: i'm sure they could be changed, but since many are already deployed, that makes it less practical
[08:57:10] Christian Huitema leaves the room
[08:57:13] <Meetecho> Will you notify us on the chat when the session is over? We're focusing on the plenary right now
[08:57:14] <Peter Heist> they might be the source of the marking that is seen
[08:57:33] <Gorry Fairhurst> @meetecho yes - thanks
[08:57:34] <Jonathan Morton> I think RFC-3168 AQMs will also end up squashing Not-ECT traffic if they share with L4S traffic, because they will still signal with drop
[08:57:50] Mark McFadden leaves the room
[08:57:51] John Border leaves the room
[08:57:53] <Ingemar Johansson> We (Ericsson) have a large interest in experiments with L4S.
[08:57:55] Paolo Saviano joins the room
[08:58:00] Chris Lemmons leaves the room
[08:58:18] Spencer Dawkins leaves the room
[08:58:29] Colin Perkins leaves the room
[08:59:28] <Ingemar Johansson> or is it "big" ? .. my english grammar sucks
[08:59:44] <Mirja Kühlewind> Is the concern about tunnels about VPNs?
[08:59:45] David Lawrence leaves the room
[08:59:57] <Mirja Kühlewind> (given most fq-codel is in home/wifi networks)
[09:00:16] Renan Krishna leaves the room
[09:00:21] Yoshifumi Atarashi leaves the room
[09:00:33] <Sebastian Moeller> I am against, not ready yet.
[09:00:48] <Philip Eardley> Are you doing a poll?
[09:00:56] <Vidhi Goel> waiting for the poll
[09:01:02] <Mirja Kühlewind> I think I can review (depending on the exact timing of course)
[09:01:12] <Peter Heist> any time the tunneled traffic shares the same 5-tuple. most VPNs are one example
[09:01:28] David Schinazi leaves the room
[09:01:31] <Sebastian Moeller> I would review, but unless major changes happen I object strongly.
[09:01:34] <Mirja Kühlewind> you can turn off ECN for VPN traffic is that is a real concern
[09:01:45] <Carsten Bormann> review: 18:2 (out of 52)
[09:01:58] Esko Dijk leaves the room
[09:02:02] <Ingemar Johansson> Did not see where to say yes. add 1
[09:02:04] Stephen McQuistin leaves the room
[09:02:12] Lucas Pardue leaves the room
[09:02:13] Mihail Yanev leaves the room
[09:02:27] <Peter Heist> i'm not sure if most home users with those fq_codel instances would know how to do that
[09:02:35] <Carsten Bormann> Bob: make a video
[09:02:38] frodek leaves the room
[09:02:46] Lars Eggert leaves the room
[09:02:47] Lv Yunping leaves the room
[09:02:48] Kazuhisa Matsuzono leaves the room
[09:02:51] Eric Kinnear leaves the room
[09:02:55] Ram Ranganathan leaves the room
[09:02:55] Philip Eardley leaves the room
[09:02:55] Mirja Kühlewind leaves the room
[09:02:59] <Ingemar Johansson> Bye
[09:02:59] Igor Lubashev leaves the room
[09:03:00] Julien Maisonneuve leaves the room
[09:03:01] <David Black> @Meetecho - we're now done
[09:03:01] Carsten Bormann leaves the room
[09:03:01] Joakim Misund leaves the room
[09:03:03] Michael Scharf leaves the room
[09:03:04] Vidhi Goel leaves the room
[09:03:04] Sergey Fomin leaves the room
[09:03:04] <Tobia Castaldi> ok
[09:03:04] Xavier de Foy leaves the room
[09:03:05] Koen Schepper leaves the room
[09:03:05] Richard Scheffenegger leaves the room
[09:03:06] Peter Heist leaves the room
[09:03:06] Ingemar Johansson leaves the room
[09:03:06] Hannu Flinck leaves the room
[09:03:07] <Tobia Castaldi> thanks
[09:03:07] Mihail Zverev leaves the room
[09:03:07] Andrew McGregor leaves the room
[09:03:09] Jana Iyengar leaves the room
[09:03:10] dschinazi leaves the room
[09:03:10] <Meetecho> Ack
[09:03:13] Meetecho leaves the room
[09:03:13] David Black leaves the room
[09:03:14] Chi-Jiun Su leaves the room
[09:03:14] Sebastian Moeller leaves the room
[09:03:15] Greg White leaves the room
[09:03:16] Daniel Havey leaves the room
[09:03:16] Juhamatti Kuusisaari leaves the room
[09:03:18] Jake Holland leaves the room
[09:03:19] Stuart Cheshire leaves the room
[09:03:19] Kazuaki Ueda leaves the room
[09:03:21] Jingrong Xie leaves the room
[09:03:24] dahavey leaves the room
[09:03:28] Wesley Eddy leaves the room
[09:03:29] Jörg Deutschmann leaves the room
[09:03:35] Steffen Klassert leaves the room
[09:03:35] Jonathan Morton leaves the room
[09:03:38] <Gorry Fairhurst> @meetecho - THANKS!
[09:03:42] Markus Amend leaves the room
[09:04:33] Gorry Fairhurst leaves the room
[09:05:11] Zaheduzzaman Sarker leaves the room
[09:05:25] Marcus Ihlar leaves the room
[09:05:52] Mohit Tahiliani leaves the room
[09:06:19] Hirochika Asai leaves the room
[09:07:27] Juan Cerezo leaves the room
[09:07:49] Zhen Cao leaves the room
[09:16:48] Jiao Kang leaves the room
[09:20:30] Ana Custura leaves the room: Disconnected: BOSH client silent for over 60 seconds
[09:26:10] Paolo Saviano leaves the room
[09:27:05] Paolo Saviano joins the room
[09:34:45] Paolo Saviano leaves the room
[10:30:34] irl joins the room
[10:32:14] irl leaves the room
[10:34:50] spencerdawkins leaves the room
[10:45:51] Christian Huitema * joins the room
[10:54:20] cabo leaves the room
[10:54:31] cabo joins the room
[11:00:47] Bob Briscoe leaves the room
[11:07:59] Igor Gashinsky leaves the room
[11:07:59] Tobia Castaldi leaves the room
[11:32:23] gorryf leaves the room
[12:03:22] gorryf joins the room
[12:07:11] cabo joins the room
[12:07:21] cabo leaves the room
[12:07:21] cabo leaves the room
[12:12:33] cabo joins the room
[12:12:51] cabo leaves the room
[12:26:11] Ana Custura joins the room
[12:40:27] Ana Custura leaves the room: Disconnected: BOSH client silent for over 60 seconds
[12:49:20] cabo joins the room
[12:59:50] Qin W joins the room
[13:03:13] Qin W leaves the room
[13:18:03] gorryf leaves the room
[13:18:33] Christian Huitema * leaves the room
[13:28:53] cabo leaves the room
[13:35:50] Ana Custura joins the room
[14:06:05] Ana Custura leaves the room: Disconnected: BOSH client silent for over 60 seconds
[14:11:28] Ana Custura joins the room
[14:13:20] Ana Custura leaves the room: Disconnected: BOSH client silent for over 60 seconds
[15:08:15] Qin W joins the room
[15:09:13] Qin W leaves the room
[16:19:54] gorryf joins the room
[17:59:49] gorryf leaves the room
[18:31:16] gorryf joins the room
[18:35:50] cabo joins the room
[18:58:55] gorryf leaves the room
[22:54:31] cabo leaves the room
[23:16:53] cabo joins the room