IETF
QUIC
quic@jabber.ietf.org
Wednesday, June 7, 2017< ^ >
Matthew Ford has set the subject to: 'QUIC IETF98 WG Meeting 20170330'
Room Configuration
Room Occupants

GMT+0
[01:40:37] spencerdawkins leaves the room
[02:27:36] spencerdawkins joins the room
[02:54:28] spencerdawkins leaves the room
[02:56:51] spencerdawkins joins the room
[03:04:01] ekr joins the room
[03:04:59] spencerdawkins leaves the room
[03:17:06] spencerdawkins joins the room
[03:30:54] spencerdawkins leaves the room
[03:40:53] ekr leaves the room
[04:11:19] spencerdawkins joins the room
[04:49:46] ekr joins the room
[04:52:36] ekr leaves the room
[04:53:42] spencerdawkins leaves the room
[04:55:40] spencerdawkins joins the room
[05:03:05] cabo joins the room
[05:12:02] yann joins the room
[05:36:27] ekr joins the room
[05:37:44] spencerdawkins leaves the room
[05:47:01] spencerdawkins joins the room
[05:47:27] ekr leaves the room
[06:20:57] yann leaves the room
[06:21:43] afrind joins the room
[06:23:23] afrind leaves the room: Stream closed by us: Replaced by new connection (conflict)
[06:23:23] afrind joins the room
[06:31:12] ihlar joins the room
[06:36:25] mcmanus joins the room
[06:37:14] Mirja Kuehlewind joins the room
[06:39:08] Mirja Kuehlewind leaves the room
[06:44:22] spencerdawkins leaves the room
[06:45:27] spencerdawkins joins the room
[06:49:33] gorryf joins the room
[06:49:33] ihlar leaves the room
[06:53:50] ekr joins the room
[06:55:04] Daniel joins the room
[06:55:07] Martin Thomson joins the room
[06:56:40] buck joins the room
[06:57:59] gryning joins the room
[06:59:00] cabo leaves the room
[07:00:07] Martin Thomson leaves the room
[07:00:50] gmbarr joins the room
[07:03:29] Mirja Kuehlewind joins the room
[07:03:39] Mirja Kuehlewind leaves the room
[07:03:41] Martin Thomson joins the room
[07:03:48] Ted.h joins the room
[07:05:46] afrind leaves the room: Stream reset by peer
[07:06:52] janaiyengar joins the room
[07:10:01] herver joins the room
[07:10:56] mnot joins the room
[07:12:36] Mirja Kuehlewind joins the room
[07:14:57] Christian Huitema joins the room
[07:15:44] Ted.h has set the subject to: Day 2, Electric Boogaloo
[07:17:06] <mnot> Do remote folks have sound?
[07:17:13] <Daniel> no...
[07:17:27] <Daniel> now!
[07:17:28] <mnot> now?
[07:19:24] Lars Eggert joins the room
[07:19:31] Brian Trammell joins the room
[07:19:41] Lars Eggert leaves the room
[07:22:13] Lars Eggert joins the room
[07:23:22] <gryning> anyone know the link to the implementation on wiki?
[07:23:24] <gryning> lost it :)
[07:23:44] <gryning> found!
[07:23:57] <Christian Huitema> Can you copy it here?
[07:24:12] <gryning> https://github.com/quicwg/base-drafts/wiki/Implementations
[07:25:06] yann joins the room
[07:25:47] spencerdawkins leaves the room
[07:26:09] chi.jiun.su joins the room
[07:26:15] afrind joins the room
[07:26:50] Martin Thomson leaves the room
[07:27:14] ranjeeth joins the room
[07:28:39] Martin Thomson joins the room
[07:29:01] Martin Thomson joins the room
[07:29:22] janaiyengar leaves the room
[07:30:17] Martin Thomson leaves the room
[07:30:19] Roni Even joins the room
[07:31:41] ranjeeth leaves the room: Disconnected: Replaced by new connection
[07:31:42] ranjeeth joins the room
[07:32:12] ihlar joins the room
[07:33:46] <gryning> morning all
[07:34:17] Sean Turner joins the room
[07:34:21] <Ted.h> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit?usp=sharing
[07:34:42] spencerdawkins joins the room
[07:35:03] mikebishop joins the room
[07:35:42] fielding joins the room
[07:38:57] <spencerdawkins> By the way, the audio is about an order of magnitude better today. Thank you for the effort in fixing that.
[07:41:37] <mikebishop> Interestingly, I can't see that anything has changed w.r.t. mic position, the reverb baffles, etc.  I'm curious to know what brought the improvement.
[07:41:43] <mnot> nothing has changed
[07:41:47] <mnot> huh
[07:42:02] <Daniel> to me the sound experience is basically the same as yday
[07:42:29] janaiyengar joins the room
[07:42:31] <Roni Even> the audio is the same as far as I can tell.
[07:42:35] lucaspardue joins the room
[07:47:09] holdrege joins the room
[07:48:07] <spencerdawkins> May all of QUIC's challenges fix themselves just that way :D
But I can follow what's being said much better.
[07:48:39] <mnot> It could be that we didn’t properly select the mic yesterday, and it fell back to the camera mic. but I’m pretty sure we selected it...
[07:49:20] <Martin Thomson> won't this look like a broken path if the client goes away
[07:50:29] lucaspardue leaves the room
[07:50:43] Magnus Westerlund joins the room
[07:54:23] lucaspardue joins the room
[07:58:32] <lucaspardue> would there be any resistance to modifying the QUIC implementation wiki to a format more like the H2 one (https://github.com/http2/http2-spec/wiki/Implementations)?
[07:59:15] <mnot> Don’t see why not.
[07:59:31] <Daniel> I'd like that
[08:01:32] <lucaspardue> ok I'll take a stab at building the template, and then missing/incorrect details can be updated
[08:02:18] <Ted.h> But this semantic has changed with the "lost state" semantic
[08:05:21] lucaspardue leaves the room
[08:09:11] gorryf leaves the room
[08:11:49] ihlar joins the room
[08:11:49] ihlar leaves the room
[08:18:38] lucaspardue joins the room
[08:19:04] buck leaves the room
[08:20:33] <Lars Eggert> ekr joe ted mirja
[08:20:48] <Mirja Kuehlewind> you forgot joan (?)
[08:20:58] <Ted.h> Jo Kulik, not Joe
[08:21:04] <Lars Eggert> sorry!
[08:21:06] <Lars Eggert> bad acoustics
[08:21:12] <Lars Eggert> ekr jo ted mirja
[08:21:17] holdrege leaves the room
[08:21:32] holdrege joins the room
[08:21:33] lucaspardue leaves the room
[08:21:52] dragana joins the room
[08:23:23] <Lars Eggert> jo ted mirja
[08:23:34] <Lars Eggert> jo ted mirja brian
[08:24:31] <Lars Eggert> ted mirja brian ekr
[08:25:10] <Lars Eggert> ted mirja brian ekr jana igor
[08:25:52] <Lars Eggert> ted mirja brian ekr jana igor sub.
[08:26:04] melinda.shore@xmpp.is joins the room
[08:26:15] <mnot> ted mirja brian ekr jana igor subodh coffee
[08:26:25] <mnot> or coffee
[08:26:30] <mnot> covfefe
[08:26:51] <janaiyengar> I've got covfefe.
[08:27:02] buck joins the room
[08:27:49] <buck> wouldn't the secret derived from the cert private key make sense?
[08:28:10] <Lars Eggert> brian ekr jana igor subodh
[08:28:59] <buck> I think many of the scalability concerns are the same for certs, no?
[08:29:32] holdrege leaves the room
[08:31:14] <Lars Eggert> jana igor subodh
[08:32:14] <Martin Thomson> I need to think about Mirja's suggestion regarding Reset, ORLY, YA RLY exchange
[08:33:18] <Lars Eggert> igor subodh
[08:35:01] <Lars Eggert> igor subodh brian
[08:35:12] <Lars Eggert> igor subodh brian jo
[08:36:34] <Martin Thomson> There is an easy fix for igor's problem: a cycle ends when you detect that the other side is sending the same packet over and over
[08:36:54] <janaiyengar> Buck: I believe you can derive the secret from the cert private key, but you may want to rotate this more frequently? Not sure what the scalability concerns are
[08:37:05] <Ted.h> It won't be exactly the same, though, right as the packet number will change.  
[08:37:06] <Lars Eggert> brian jo
[08:38:00] <Lars Eggert> brian jo jana
[08:38:15] holdrege joins the room
[08:39:36] <Lars Eggert> jo jana
[08:39:42] <buck> I think managing certs is already a problem large scale deployments have to manage, so consolidating with that makes it sort of free.
[08:40:13] <Lars Eggert> covfefe!
[08:42:18] <gorryf> Hum for not hidden -
[08:42:38] <gorryf> ... at least until we understand more.
[08:44:31] janaiyengar leaves the room
[08:44:33] yann leaves the room
[08:44:34] gmbarr leaves the room
[08:44:34] ekr leaves the room
[08:44:42] ekr joins the room
[08:44:43] ekr leaves the room
[08:44:45] Lars Eggert leaves the room
[08:44:46] herver leaves the room
[08:45:08] Martin Thomson leaves the room
[08:46:23] Mirja Kuehlewind leaves the room
[08:46:25] mcmanus leaves the room
[08:46:51] Brian Trammell leaves the room
[08:47:21] mikebishop leaves the room
[08:47:43] Ted.h leaves the room
[08:49:34] ihlar leaves the room
[08:53:54] buck leaves the room
[08:56:06] fielding leaves the room
[08:57:07] gmbarr joins the room
[08:57:15] yann joins the room
[09:01:11] Mirja Kuehlewind joins the room
[09:01:27] Brian Trammell joins the room
[09:01:49] Martin Thomson joins the room
[09:02:34] Ted.h joins the room
[09:02:37] ranjeeth leaves the room: Disconnected: Broken pipe
[09:02:46] Lars Eggert joins the room
[09:02:57] ranjeeth joins the room
[09:03:04] buck joins the room
[09:03:08] dragana leaves the room: Disconnected: No route to host
[09:03:30] mcmanus joins the room
[09:03:47] janaiyengar joins the room
[09:03:55] ihlar joins the room
[09:05:52] dragana joins the room
[09:08:00] mikebishop joins the room
[09:09:12] buck leaves the room
[09:12:10] fielding joins the room
[09:18:11] <mnot> Can remote folks hear Martin OK?
[09:18:59] <Roni Even> yes
[09:19:01] <Daniel> yeps
[09:24:32] <melinda.shore@xmpp.is> The audio generally seems better today
[09:24:52] lucaspardue joins the room
[09:25:18] <Lars Eggert> igor brian
[09:26:15] <Lars Eggert> brian
[09:26:29] <Lars Eggert> brian christian jana
[09:26:42] buck joins the room
[09:28:10] <gryning> is the protocol expecting to specify an upper bound on idle timeouts?
[09:28:47] <Lars Eggert> jana
[09:29:00] <buck> does the application need to tell the transport to close, as opposed to closing all its streams?
[09:29:14] herver joins the room
[09:30:05] <gryning> i don't know whether to have recervations; not enough known aobut timeouts, implementation overheads etc. Spend a lot of years fighting agianst time_wait context irl
[09:30:56] gryning is now known as craigt
[09:31:11] <Lars Eggert> subodh brian
[09:31:39] <Lars Eggert> subodh brian mark
[09:31:50] <Lars Eggert> subodh brian buck
[09:34:07] holdrege leaves the room
[09:35:39] holdrege joins the room
[09:36:56] <Lars Eggert> link to minutes: https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:37:09] <Lars Eggert> queue is all the way at the end, you can add yourselves ther
[09:37:10] <Lars Eggert> e
[09:37:17] <Lars Eggert> instead of waitign to catch my attention
[09:37:24] <Lars Eggert> (or in addition)
[09:38:10] <mikebishop> Link again?  Sorry -- I keep losing the history when my client closes.
[09:38:24] <Ted.h> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:38:26] <mnot> link to minutes:
[09:38:26] <mnot> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:38:59] <Mirja Kuehlewind> mike: is this a privacy feature?
[09:40:06] <mikebishop> No idea the programmer's motivation.  I'm also not sure about the feature that apparently hides links, since both Ted and Mark's messages appear to be empty on my end.  :-(
[09:40:33] <craigt> font colour?
[09:40:58] <Sean Turner> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:40:59] <Brian Trammell> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:41:19] <Mirja Kuehlewind> might be more secure to make sure that you can see any bogus links… ;-)
[09:41:20] <holdrege> Should there be a shutdown timer to bound the overall time for a shutdown sequence?
[09:41:28] <mikebishop> No, when I highlight, there's actually nothing there.  Sean and Brian's messages also blank.
[09:41:47] <Mirja Kuehlewind> docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:41:51] <Mirja Kuehlewind> did this work?
[09:42:35] <Ted.h> https://docs.google.com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:42:52] <Ted.h> Darn I tried to paste as plain text and it re-converted.
[09:43:41] <Mirja Kuehlewind> docs(dot)google(dot)com/document/d/1Vm7vqMPxlCEuPnbrzIeE4Div9l3H1-rYxQqGX6A7MiI/edit#
[09:43:44] <craigt> 0rrt isn't computationally neutral either...so don't want to garbage collect to often
[09:43:55] <craigt> +o
[09:44:22] <mikebishop> *That* works.  Thanks, Mirja!
[09:44:23] <Martin Thomson> people are ratholing on the 0-RTT thing, that was put forward as a way to repair the condition where you end up disagreeing about connection state
[09:44:27] mikebishop leaves the room
[09:45:07] <Daniel> hey, I was prepared to paste a QR code with the link =)
[09:45:35] mikebishop joins the room
[09:54:44] <Martin Thomson> JUST DON'T use 0-RTT for repair
[09:55:11] <Martin Thomson> 0-RTT carries an indeterminate amount of state over, it's got all the aforementioned limitations
[09:55:23] <Martin Thomson> it's no substitute for a real continuation of a connection
[09:56:22] <buck> could use the stateful reject for this?
[09:56:33] <buck> stateless
[09:59:21] gmbarr leaves the room
[09:59:27] janaiyengar leaves the room
[09:59:37] <craigt> Is there currently one idel timeout to rule them all? applying to both application mapping and transport idle timeout?
[09:59:43] Ted.h leaves the room
[09:59:47] Ted.h joins the room
[10:00:07] Martin Thomson leaves the room
[10:00:15] buck leaves the room
[10:00:18] Christian Huitema leaves the room
[10:00:23] Ted.h leaves the room
[10:00:35] ihlar leaves the room
[10:00:37] craigt leaves the room: Disconnected: closed
[10:00:41] Lars Eggert leaves the room
[10:00:55] yann leaves the room
[10:00:57] mnot leaves the room
[10:01:01] yann joins the room
[10:01:16] Brian Trammell leaves the room
[10:01:22] Magnus Westerlund leaves the room
[10:01:25] herver leaves the room
[10:01:44] yann leaves the room
[10:01:48] yann joins the room
[10:02:16] holdrege leaves the room
[10:02:31] Mirja Kuehlewind leaves the room
[10:02:32] yann leaves the room
[10:02:37] yann joins the room
[10:03:21] yann leaves the room
[10:05:25] mcmanus leaves the room
[10:05:55] mcmanus joins the room
[10:11:36] lucaspardue leaves the room: Disconnected: No route to host
[10:15:57] dragana leaves the room: Disconnected: No route to host
[10:18:21] holdrege joins the room
[10:20:13] ranjeeth leaves the room: Disconnected: closed
[10:21:58] Sean Turner leaves the room
[10:24:03] cabo joins the room
[10:36:25] mcmanus leaves the room
[10:38:03] yann joins the room
[10:42:30] holdrege leaves the room
[10:47:28] yann leaves the room
[10:47:30] yann joins the room
[10:48:13] yann leaves the room
[10:48:18] yann joins the room
[10:48:39] fielding leaves the room
[10:49:01] yann leaves the room
[10:49:05] yann joins the room
[10:49:49] yann leaves the room
[10:49:53] yann joins the room
[10:50:37] yann leaves the room
[10:50:42] yann joins the room
[10:51:25] yann leaves the room
[10:56:23] mcmanus joins the room
[10:59:09] yann joins the room
[11:01:11] gmbarr joins the room
[11:03:32] Ted.h joins the room
[11:05:15] Mirja Kuehlewind joins the room
[11:05:17] Sean Turner joins the room
[11:05:46] Sean Turner leaves the room
[11:06:43] Daniel leaves the room
[11:06:57] Daniel joins the room
[11:07:02] Sean Turner joins the room
[11:10:34] mnot joins the room
[11:11:09] <mnot> Remote folks: Can you hear us?
[11:11:14] holdrege joins the room
[11:12:42] <mnot> Anybody out there?
[11:13:52] <Daniel> we hear you
[11:13:56] <mnot> great
[11:14:08] ihlar joins the room
[11:14:08] <melinda.shore@xmpp.is> yup
[11:16:57] ranjeeth joins the room
[11:18:29] Lars Eggert joins the room
[11:18:32] herver joins the room
[11:18:37] Martin Thomson joins the room
[11:18:57] Brian Trammell joins the room
[11:19:11] magnus joins the room
[11:19:35] gryning joins the room
[11:19:40] buck joins the room
[11:20:02] fielding joins the room
[11:20:26] Christian Huitema joins the room
[11:20:28] janaiyengar joins the room
[11:20:47] magnus joins the room
[11:25:27] buck leaves the room
[11:25:31] buck joins the room
[11:29:51] buck leaves the room
[11:36:13] holdrege leaves the room
[11:36:26] holdrege joins the room
[11:36:53] magnus leaves the room
[11:44:51] lucaspardue joins the room
[11:45:44] buck joins the room
[11:46:55] melinda.shore@xmpp.is leaves the room
[11:48:23] <Martin Thomson> I just reviewed the sections of these documents: the timeout option is exactly this design, the other includes text on slowloris
[11:53:05] buck leaves the room
[11:54:07] buck joins the room
[11:56:39] <gryning> ": The idle timeout is a value in seconds that is encoded as an unsigned 16-bit integer. The maximum value is 600 seconds (10 minutes).
"
[11:56:46] gryning is now known as craigt
[12:01:00] buck leaves the room
[12:01:45] buck joins the room
[12:08:56] holdrege leaves the room
[12:11:49] holdrege joins the room
[12:12:10] <buck> nit:  stream frame for already closed frame should be silently dropped I think.
[12:12:39] <buck> s/closed frame/closed stream/
[12:12:42] cabo leaves the room
[12:20:43] <Brian Trammell> this is why switch statements in C fall through btw
[12:22:53] buck leaves the room
[12:24:54] buck joins the room
[12:31:44] <Lars Eggert> maybe instead of byte offset into the packet that caused an error, we should encode draft name and line number that was violated :-)
[12:31:54] <Brian Trammell> :D
[12:32:09] <Brian Trammell> no draft, no error for you
[12:32:13] <Lars Eggert> we got the codepintes
[12:32:26] <mnot> I don’t buy the i18n argument; this isn’t user-visible, and the protocol documentation itself is in english.
[12:32:44] <Brian Trammell> + ichi mnot
[12:32:46] <Lars Eggert> @mnot you get to handle the discuss then :-)
[12:32:54] <mnot> deal
[12:33:36] <Lars Eggert> the ops guys will make the argument that ops people need to understand these strings
[12:34:17] <holdrege> Ops people all over the world can read English
[12:35:00] <Ted.h> @holdrege 說真的嗎?
[12:36:09] <holdrege> Baidu translate. :)
[12:37:06] <holdrege> Speaking English is hard, but reading it is mandatory even in China.
[12:37:06] <holdrege> And most engineers
[12:37:06] <holdrege> For ops people
[12:37:17] <Roni Even> this is application problem easy to solve by having a file with the stringsin other languages
[12:38:05] <Ted.h> @That turns the string into a pointer, used to look up strings.  That's the equivalent of an error code, just a prolix one.
[12:38:48] <Ted.h> Sorry, that should have been @Roni.
[12:39:10] ranjeeth leaves the room
[12:39:12] ranjeeth joins the room
[12:40:50] holdrege leaves the room
[12:42:02] <Roni Even> BTW: if in RFC there can be nonascii text https://tools.ietf.org/html/draft-iab-rfc-nonascii-02
[12:42:41] <Roni Even> then it can be done anywhere
[12:42:59] holdrege joins the room
[12:49:09] <craigt> emacs or vi?
[12:50:28] <lucaspardue> https://vim-adventures.com/
[12:51:01] Ted.h leaves the room
[12:51:26] ted.h joins the room
[12:51:44] ted.h leaves the room
[12:52:25] janaiyengar leaves the room
[12:52:47] gmbarr leaves the room
[12:52:55] <Roni Even> how long the break?
[12:52:58] Lars Eggert leaves the room
[12:53:00] yann leaves the room
[12:53:04] yann joins the room
[12:53:05] <herver> 10 minutes break
[12:53:13] <Roni Even> thx
[12:53:16] Martin Thomson leaves the room
[12:53:20] herver leaves the room
[12:53:45] yann leaves the room
[12:53:50] yann joins the room
[12:54:15] craigt leaves the room: Disconnected: closed
[12:54:33] yann leaves the room
[12:54:37] yann joins the room
[12:55:21] yann leaves the room
[12:55:26] yann joins the room
[12:55:40] buck leaves the room
[12:56:09] yann leaves the room
[12:56:14] yann joins the room
[12:56:57] yann leaves the room
[12:57:02] yann joins the room
[12:57:45] yann leaves the room
[12:59:50] gmbarr joins the room
[13:00:47] yann joins the room
[13:01:55] Mirja Kuehlewind leaves the room
[13:04:06] ranjeeth leaves the room: Disconnected: timeout during writing
[13:04:23] ranjeeth joins the room
[13:09:42] Mirja Kuehlewind joins the room
[13:11:32] ihlar joins the room
[13:11:33] ihlar leaves the room
[13:12:23] holdrege leaves the room
[13:12:29] ihlar leaves the room
[13:14:26] fielding leaves the room
[13:15:37] Lars Eggert joins the room
[13:15:50] holdrege joins the room
[13:16:02] gryning joins the room
[13:16:04] ted.h joins the room
[13:16:13] herver joins the room
[13:16:22] ranjeeth leaves the room
[13:16:24] ranjeeth joins the room
[13:16:33] janaiyengar joins the room
[13:16:41] Martin Thomson joins the room
[13:16:51] ihlar joins the room
[13:18:00] buck joins the room
[13:24:12] gryning is now known as craigt
[13:36:34] <Roni Even> But are we going to manadte one congestion control and not allowed others?
[13:37:42] <Roni Even> I think there was also a discussion about marking for better AQM in chicago
[13:38:58] <gorryf> Yes -and that may be a really good reason for exposing numbers in packets headers to allow operators to understand how the AQMs are marking and dropping - since present methods depend on RTT.
[13:39:26] dragana joins the room
[13:41:39] Mirja Kuehlewind leaves the room
[13:44:43] Mirja Kuehlewind joins the room
[13:45:57] <gorryf> EKR’s rate of transition is challenging to interpret across the room Mic!
[13:50:32] ihlar joins the room
[13:50:50] ihlar leaves the room
[13:51:29] ihlar leaves the room
[13:59:15] <gorryf> Inserting a path RTT is much less useful than a timestamp - for example, it does not allow measurement of the latency upstream or downstream of a segment:-).
[14:00:19] buck leaves the room
[14:00:25] <craigt> indeed; server->client tends to care about the absolute. Path cares about segment or hop by hop...
[14:01:36] fielding joins the room
[14:01:43] <gorryf> (There’s also a question of “how” the measurement was made - having timestamps lets you observe jitter, min RTT, average RTT, etc - someone giving you a number makes you entirely depend on how the data was collected.)
[14:06:51] <Lars Eggert> we can use the evil bit :-)
[14:13:03] <ted.h> Solution space: timestamps acting as sequence numbers, integrity protected.
[14:13:15] <gorryf> :-)
[14:14:00] <gorryf> You loose the ability to try to measure loss... which could be a “loss” if AQM uses drop and you want to debug that - but you gain already by this.
[14:14:13] <craigt> Forgive my naivety: ...can't IPv6 hop by hop be abused in path for measurement?
[14:14:31] <ted.h> @gryning it's routinely bleached now.
[14:14:48] <craigt> like my nick here
[14:15:34] <Brian Trammell> drop rates for HBH are brutal
[14:15:34] <Brian Trammell> https://tools.ietf.org/html/rfc7872
[14:15:39] <craigt> ted.h: do we have supporting data for that?
[14:15:48] <Brian Trammell> HBH -> "please do not route this packet"
[14:15:55] <gorryf> There is also a proposal - but unsure of model for deployment ??? - (ippm-pdm-6man)
[14:15:57] <craigt> but also, if the operator wants the measurement, then they stop bleaching?
[14:16:06] <Brian Trammell> you can't
[14:16:09] <Brian Trammell> hbh -> slowpath
[14:16:15] <craigt> they would only need to apply the trace within their segment
[14:16:17] <Brian Trammell> you let hbh on your network and it dies
[14:16:31] cabo joins the room
[14:16:46] <Brian Trammell> this is the reason to make measurement happen up in the transport layer
[14:17:00] <craigt> ok, so there are deployment issues he (and some dragons)
[14:28:13] buck joins the room
[14:28:47] <ted.h> Oh merciful queue end, which brings surcease to our labors and eases our pains, we thank you for the opportunity to quaff you bring to us.
[14:29:04] <Brian Trammell> okay, fine, lets grease the measurement header.
[14:29:21] <gorryf> Brian - you need a beverage.
[14:30:36] <Brian Trammell> one beer please.
[14:30:39] <Brian Trammell> grease it.
[14:31:10] <gorryf> Then you will drop it, and it will be a mess - be careful what you request.
[14:32:48] <gorryf> What is Friction? Friction-less beer glasses also seem to be less stable solutions.
[14:34:30] Lars Eggert leaves the room: Disconnected: closed
[14:37:08] Lars Eggert joins the room
[14:38:04] <gorryf> Yes - if we have proposals.
[14:42:12] yann leaves the room
[14:42:13] mikebishop leaves the room
[14:42:13] holdrege leaves the room
[14:42:14] ihlar joins the room
[14:42:15] gmbarr leaves the room
[14:42:15] Mirja Kuehlewind leaves the room
[14:42:17] buck leaves the room
[14:42:17] Lars Eggert leaves the room
[14:42:20] janaiyengar leaves the room
[14:42:22] herver leaves the room
[14:42:26] Christian Huitema leaves the room
[14:42:32] chi.jiun.su leaves the room: offline
[14:42:35] lucaspardue leaves the room
[14:42:42] craigt leaves the room: Disconnected: closed
[14:42:46] fielding leaves the room
[14:42:48] ted.h leaves the room
[14:42:51] ihlar leaves the room
[14:42:52] afrind leaves the room: Stream reset by peer
[14:43:18] mcmanus joins the room
[14:43:26] mcmanus leaves the room
[14:47:14] Martin Thomson leaves the room
[14:47:25] Martin Thomson joins the room
[14:48:36] Martin Thomson leaves the room
[14:49:26] Mirja Kuehlewind joins the room
[14:49:46] Brian Trammell leaves the room
[14:50:02] Mirja Kuehlewind leaves the room
[14:50:34] Mirja Kuehlewind joins the room
[14:50:56] mcmanus joins the room
[14:52:09] ted.h joins the room
[14:54:27] Martin Thomson joins the room
[14:54:33] Brian Trammell joins the room
[14:55:43] ted.h leaves the room
[14:56:35] cabo leaves the room
[14:58:01] magnus leaves the room
[14:58:06] Sean Turner leaves the room
[14:58:49] ranjeeth leaves the room: Disconnected: Broken pipe
[14:59:12] dragana leaves the room: Disconnected: No route to host
[15:00:26] mcmanus leaves the room
[15:05:43] cabo joins the room
[15:11:29] Christian Huitema joins the room
[15:24:27] Brian Trammell leaves the room
[15:25:24] Brian Trammell joins the room
[15:25:26] Lars Eggert joins the room
[15:25:35] Lars Eggert leaves the room
[15:49:13] yann joins the room
[15:59:28] Martin Thomson leaves the room
[16:01:09] Brian Trammell leaves the room
[16:02:10] Mirja Kuehlewind leaves the room
[16:03:17] Mirja Kuehlewind joins the room
[16:05:48] Mirja Kuehlewind leaves the room
[16:08:22] mnot leaves the room
[16:14:26] mcmanus leaves the room
[16:22:31] holdrege joins the room
[16:26:09] yann leaves the room
[16:27:27] yann joins the room
[16:33:05] spencerdawkins leaves the room
[16:35:39] spencerdawkins joins the room
[16:36:15] holdrege leaves the room
[16:36:35] buck joins the room
[16:37:03] buck leaves the room
[16:42:20] mcmanus joins the room
[16:56:56] spencerdawkins leaves the room
[16:56:56] spencerdawkins joins the room
[16:56:56] spencerdawkins leaves the room
[16:58:25] yann leaves the room
[17:00:45] yann joins the room
[17:02:01] yann leaves the room
[17:22:37] Christian Huitema leaves the room: Disconnected: closed
[17:23:31] cabo leaves the room
[17:37:46] cabo joins the room
[17:46:56] mcmanus leaves the room
[18:00:59] holdrege joins the room
[18:10:41] holdrege leaves the room
[18:11:55] holdrege joins the room
[18:20:53] mnot joins the room
[18:27:30] holdrege leaves the room
[19:04:33] gorryf joins the room
[19:27:59] Daniel leaves the room
[19:28:13] Daniel joins the room
[19:33:05] gorryf leaves the room
[19:44:18] Martin Thomson joins the room
[19:44:54] spencerdawkins joins the room
[19:45:03] gorryf joins the room
[19:56:12] yann joins the room
[19:56:46] spencerdawkins leaves the room
[20:02:19] gorryf leaves the room
[20:17:20] Martin Thomson leaves the room
[20:18:17] Martin Thomson joins the room
[20:24:38] janaiyengar joins the room
[20:25:03] janaiyengar leaves the room
[20:34:24] Martin Thomson leaves the room
[20:35:22] Martin Thomson joins the room
[20:48:51] mnot leaves the room: Disconnected: closed
[20:56:07] yann leaves the room
[21:08:37] ranjeeth joins the room
[21:25:29] ranjeeth leaves the room
[21:25:33] ranjeeth joins the room
[21:29:50] mcmanus joins the room
[21:34:58] Martin Thomson leaves the room
[21:40:40] cabo leaves the room
[21:55:08] cabo joins the room
[21:55:19] ranjeeth leaves the room
[21:56:21] ranjeeth joins the room
[22:02:47] ranjeeth leaves the room
[22:10:57] mcmanus leaves the room
[22:29:07] cabo leaves the room: Replaced by new connection
[22:29:11] cabo joins the room
[22:29:20] cabo leaves the room
[22:37:54] cabo joins the room
[22:57:01] janaiyengar joins the room
[22:59:31] janaiyengar leaves the room: Replaced by new connection
[22:59:51] janaiyengar joins the room
[23:07:20] janaiyengar leaves the room: Replaced by new connection
[23:08:00] janaiyengar joins the room
[23:11:09] yann joins the room
[23:12:02] cabo leaves the room
[23:27:23] janaiyengar leaves the room
[23:46:39] yann leaves the room
[23:47:36] yann joins the room
[23:48:07] yann leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!