IETF
tcpm@jabber.ietf.org
Monday, November 4, 2013< ^ >
jishac has set the subject to: TCP Maintenance and Minor Extensions WG
Room Configuration
Room Occupants

GMT+0
[14:24:11] britram joins the room
[15:13:52] britram leaves the room
[16:14:30] Alexander Zimmermann joins the room
[16:19:59] britram joins the room
[16:40:31] jlcJohn joins the room
[16:42:40] PasiS joins the room
[16:43:30] Michael Scharf joins the room
[16:43:41] <Alexander Zimmermann> if someone has a working audio channel…
[16:43:48] <Alexander Zimmermann> please speak up :-)
[16:43:58] <PasiS> did you hear?
[16:44:04] <Alexander Zimmermann> no
[16:44:16] <PasiS> nothing?
[16:44:16] <Alexander Zimmermann> I get a 404 in VLC
[16:44:26] <PasiS> ok. let's see
[16:45:03] <Michael Scharf> No audio on my end neither
[16:45:32] <PasiS> Wes is looking for someone to fix it
[16:46:17] britram leaves the room
[16:54:45] britram joins the room
[16:55:46] <PasiS> is audio still down?
[16:56:19] <Michael Scharf> Yes
[16:56:24] <PasiS> i'm told people are working on it
[16:59:17] <PasiS> stream starts exactly at 9
[16:59:24] <PasiS> exact timing...
[17:00:27] <Michael Scharf> AUdio up
[17:00:39] <Michael Scharf> Please say something in the mic
[17:00:48] Tim Wicinski joins the room
[17:01:08] <britram> hear that?
[17:01:09] <Alexander Zimmermann> yes
[17:01:14] <Alexander Zimmermann> it works
[17:01:46] <Alexander Zimmermann> uff
[17:01:53] <Michael Scharf> Audio quality is horrible
[17:01:54] <Alexander Zimmermann> the mic is bad
[17:02:44] <jlcJohn> (I really can't make out the words)
[17:03:35] <britram> is that better?
[17:03:35] <Michael Scharf> Much better now
[17:04:04] <britram> (slide 5 - tcp fast open)
[17:06:58] <britram> (Question: how many would like an IANA allocation for a new option code)
[17:07:14] <britram> 10
[17:09:15] dborman42 joins the room
[17:10:39] <dborman42> I’m not getting any streaming audio.  The other IETF streams seem to work, but Regency A doesn’t.
[17:11:06] <britram> anyone else able to hear the audio still?
[17:11:16] <Michael Scharf> Yes, audio works for me
[17:11:19] <dborman42> Ok, it just started working
[17:11:22] <Alexander Zimmermann> I'm able
[17:11:28] <Alexander Zimmermann> i use VLC
[17:11:28] <britram> slide 6 - wg documents
[17:11:40] <Alexander Zimmermann> open http://ietf88streaming.dnsalias.net/ietf/ietf885.m3u in VLC
[17:12:24] <britram> tcp roadmap
[17:12:41] <britram> slide 2 doc history
[17:15:03] <Alexander Zimmermann> Yes
[17:15:22] <Alexander Zimmermann> I'm fine with this
[17:15:59] <Alexander Zimmermann> i will make the split into CC and LR
[17:16:07] <britram> cool, please note this on the list...
[17:16:13] <Michael Scharf> Audio is very low
[17:16:15] <britram> (already moved forward)
[17:16:53] <dborman42> I can barely hear the speaker, please have them use the mikes
[17:17:04] <britram> done, mike was mounted poorly
[17:17:08] <britram> is that better?
[17:17:10] <Alexander Zimmermann> tano
[17:17:12] <Alexander Zimmermann> no
[17:17:18] <Michael Scharf> no
[17:17:51] <britram> mike was off
[17:18:02] <Alexander Zimmermann> still off
[17:18:03] <dborman42> much better!
[17:18:05] <Alexander Zimmermann> yes
[17:18:16] <dborman42> but still a bit soft
[17:18:20] <britram> slide - updates since draft-02
[17:18:28] <Michael Scharf> Better, indeed
[17:21:26] <britram> slide - issues to address
[17:27:53] <Alexander Zimmermann> you can see the line burst in Lars slides
[17:28:15] <Alexander Zimmermann> but Lars is still in the airplane…
[17:31:08] Juan-Pedro Cerezo Martin joins the room
[17:32:09] <Alexander Zimmermann> Agree with Jana
[17:32:35] <Michael Scharf> Q: Is the algorithm with pacing and w/o pacing indeed the same?
[17:33:45] <PasiS> sorry, we already moved on… take it to the list
[17:34:09] <britram> slide - updates to draft (1)
[17:34:17] <britram> (RTO restart)
[17:35:05] Juan-Pedro Cerezo Martin leaves the room
[17:35:14] <britram> (2)
[17:35:31] <britram> (3)
[17:35:55] <britram> (Implementation)
[17:41:54] <Alexander Zimmermann> do we have some data if we really have here an issue?
[17:43:25] <Alexander Zimmermann> the core case
[17:43:47] fgont joins the room
[17:44:20] <britram> would getting that data answer your question?
[17:44:50] <britram> (from yuchung, presuming he regenerated it?)
[17:45:07] <Alexander Zimmermann> yes
[17:47:23] <fgont> folks, where are we with the tcpm agenda?
[17:47:32] <britram> RTO restart
[17:47:43] <fgont> I'm n 6man 8two documents to present), but need to sneak into tcpm to present one there
[17:47:47] <Michael Scharf> I heard from the WG that the RFC has to change. It is a WG document
[17:48:03] <Michael Scharf> s/RFC/draft/
[17:48:06] <britram> (discussion post presentation)
[17:48:34] <Michael Scharf> OK, then the audio lag is significant right now
[17:48:56] <britram> michael, your comment is on rto restart yes?
[17:49:11] <Michael Scharf> Yes, my audio is still on rto restart
[17:49:31] <britram> ok -- we’re still there, shall I +1 the current comments for you?
[17:50:11] <britram> (the we need a change aspect thereof)
[17:50:11] <Michael Scharf> maybe you can post from time to tome who is commenting - this allows us to judge the lag
[17:50:17] <britram> ok
[17:50:24] <britram> that was gorry
[17:50:29] <britram> here is stuart
[17:50:40] <britram> still using 1 sec rto
[17:50:56] <britram> stuart done
[17:51:01] <britram> yuchung speaking
[17:51:08] <Michael Scharf> at least 30s lag
[17:51:17] <britram> ouch
[17:51:58] <britram> not much we can do about that
[17:52:06] <dborman42> I seem to be more in sync.  yungchung “try to figure out..”
[17:52:20] <britram> anna speaking now
[17:52:52] <britram> matt speaking now -
[17:53:01] <britram> “advantage of being experimental”
[17:54:29] <britram> richard presenting accecn
[17:55:01] <Michael Scharf> >30sec lag :(
[17:55:04] <dborman42> ugh, my audio stream just rebuffered.
[17:57:30] <britram> requirements slide
[17:58:27] <dborman42> My audio is 5-10 seconds behind
[17:58:46] <britram> questions on this draft?
[18:06:47] albakri joins the room
[18:07:47] Tim Wicinski leaves the room
[18:13:33] CMCC-PC joins the room
[18:14:52] CMCC-PC leaves the room
[18:19:30] <Alexander Zimmermann> Do also want to obsolete 1122?
[18:19:55] <dborman42> Can’t obsolete 1122, it’s more than just TCP
[18:20:09] <Alexander Zimmermann> i know …
[18:22:06] <Alexander Zimmermann> D. D. Clark. Window and Acknowledgement Strategy in TCP. RFC 813. Juni 1982. url: http://www.ietf.org/rfc/rfc813.txt.
[18:22:14] <Alexander Zimmermann> Silly Window Syndrom
[18:22:49] <Alexander Zimmermann> Algo is in RFC1122
[18:31:13] <Alexander Zimmermann> Go ahead Wes!
[18:31:32] <Alexander Zimmermann> +1
[18:31:40] <dborman42> +1
[18:31:58] <britram> i don’t have that many hands but i tried :)
[18:32:28] Christoph Paasch joins the room
[18:37:32] wesley.m.eddy joins the room
[18:50:37] ijjarvin joins the room
[18:51:41] <dborman42> Comment for mike: The biggest issue of PAWS vs RTT is that PAWS assumes TS in every packet, and if you want to use TS for RTO measurement on only some packets, it’ll cause issues for PAWS.  That’s why late negotiation of TS means sending them in every packet after that.
[18:52:01] <dborman42> The big challenge is backwards compatibility
[18:53:21] <dborman42> (this is David Borman)
[18:53:46] martin.stiemerling joins the room
[18:58:26] martin.stiemerling leaves the room
[18:58:51] swb joins the room
[18:58:54] fgont leaves the room
[19:08:46] swb leaves the room
[19:09:27] <Michael Scharf> Q: Regarding the bufferbloat causing RTT inflation: What about pacing with some RTTmin estimator? Would that be doable in that scheme?
[19:19:20] fgont joins the room
[19:21:12] <dborman42> Comment for mike at appropriate point:  The MUST is to support simultaneous opens and closes.  TCP self connects are an easy way to test that, but they aren’t the MUST, so I
[19:21:24] martin.stiemerling joins the room
[19:21:24] <dborman42> So I’m fine with changing self connects to a SHOULD
[19:21:37] <britram> ok, queueing...
[19:22:43] <britram> ok, he made the point...
[19:22:59] <Alexander Zimmermann> If we adopt this draft then we can incorporate it directly into 793bis, no?
[19:23:15] <britram> seems like it to me, will ask
[19:23:17] <dborman42> Please say that as an author, I’m okay with with changing the MUST to a SHOULD.
[19:23:27] <britram> ah ok
[19:23:38] <Alexander Zimmermann> without generate another RFC
[19:28:23] <wesley.m.eddy> Alex: IMHO, this can be done very quickly and can hold the description of why it was done, like the draft does now
[19:28:36] <wesley.m.eddy> 793bis will take much longer and we won't want to include all the analysis
[19:30:11] britram leaves the room
[19:30:12] <dborman42> Comment to Matt’s comment: The intention is that this document fixes the problem of crossing ZWP where the updated ACK and window information isn’t accepted, we should make sure the document says that, because that is the intention.
[19:30:34] <Alexander Zimmermann> @Wes: fine
[19:31:05] britram joins the room
[19:31:20] albakri leaves the room
[19:33:20] britram leaves the room
[19:34:46] <dborman42> We can do an adoption call on the mailing list.
[19:38:42] Christoph Paasch leaves the room
[19:39:47] wesley.m.eddy leaves the room
[19:39:51] Michael Scharf leaves the room
[19:39:52] britram joins the room
[19:47:17] PasiS leaves the room
[19:50:00] Alexander Zimmermann leaves the room
[19:50:42] britram leaves the room
[19:51:21] martin.stiemerling leaves the room
[19:55:03] ijjarvin leaves the room
[19:56:29] PasiS joins the room
[20:45:47] PasiS leaves the room
[20:49:49] martin.stiemerling joins the room
[21:08:17] martin.stiemerling leaves the room
[21:08:27] britram joins the room
[21:16:39] britram leaves the room
[21:17:49] fgont leaves the room
[21:18:15] fgont joins the room
[21:19:48] britram joins the room
[21:32:43] dborman42 leaves the room
[21:51:06] britram leaves the room
[21:52:46] fgont leaves the room
[22:13:28] britram joins the room
[22:29:46] britram leaves the room
[22:45:32] wesley.m.eddy joins the room
[22:46:11] wesley.m.eddy leaves the room
[23:01:00] britram joins the room
[23:40:37] britram leaves the room
[23:41:14] britram joins the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!