[16:12:38] schnizlein@jabber.isoc.org joins the room [16:12:55] schnizlein@jabber.isoc.org leaves the room [16:13:58] John S joins the room [16:14:17] John S leaves the room [18:48:57] toby joins the room [19:48:01] toby leaves the room [19:48:04] toby joins the room [19:48:41] toby leaves the room [19:49:54] toby moncaster joins the room [19:55:50] gorryf joins the room [19:57:56] gorryf leaves the room [19:59:38] maciekwojciechowski joins the room [20:03:32] Lars joins the room [20:04:31] Mat Ford joins the room [20:04:46] momose joins the room [20:06:05] wolfgang.beck01 joins the room [20:06:52] Draft-penno practices & recommendations [20:07:06] is the audio ok for those of you listening in? [20:07:14] Yes its fine... [20:07:23] quality is fine [20:07:26] great [20:07:27] iljitsch joins the room [20:08:31] Slide 2 [20:08:34] gorryf joins the room [20:08:47] would be nice if we had the slides available out here in audio-land... [20:08:52] AdamUzelac joins the room [20:09:00] are they not uploaded? [20:09:09] Not according to meeting materials page [20:09:15] shikob joins the room [20:09:17] I haven't found them anywhere. [20:09:21] telling the chairs [20:09:35] Multiple tcp conns - advantages [20:09:45] so far, the slides were just the table of contents of the draft [20:11:13] why would you want to disable SACK??? [20:12:11] I guess you could have an ancient stack ? Otherwise I'd hope it was enabled. [20:12:19] schnizlein@jabber.isoc.org joins the room [20:12:49] I did some measurements, web requests on my server have SACK enabled 97% of the time [20:13:02] ph joins the room [20:13:11] vijay.gurbani joins the room [20:13:17] not far off the figures I've foubnd quoted (90-95%, no idea of source) [20:13:23] ph leaves the room [20:13:32] Slide discusssing hi bw delay links [20:14:53] Slide error resiliency, availability [20:16:19] Slide multle processors in a system [20:17:04] Slide disadvantages of multiple tcp connections [20:17:12] csp joins the room [20:19:41] Slide recommendations [20:22:44] 1 comment 1 question: Comment: ALL the advantages seem to be for the end-host whereas most disadvantages seem to be for the network. Seems to boil down to which do we think is more important... [20:23:36] Question: Are you saying that multiple connections through different bottlenecks are OK? [20:23:39] Lars eggert @ mike [20:24:37] PasiS joins the room [20:30:15] Atarashi Yoshifumi joins the room [20:31:00] It just occurred to me that the title of the WG includes "Background Traffic" which implies data-heavy applications as against interactive applications such as Jabber... [20:33:05] I'd also hoped this was looking at applications that would like not to congest the network and disrupt other uses - to me (I don't know about other) this doesn't necessarily mean it has to be bulk [20:33:42] There is a problem for the network if you end up creating new bottlenecks by spreading your data over multiple new connections... [20:35:16] To toby: I think this would be a good thing to dig into a little. [20:35:25] Lars Eggert @ mike [20:36:46] Gorry - we are already very intersted in this. You could try and pin down Bob for more info... [20:36:59] Presentation draft-shalunov-ledbat-congestion [20:37:27] schnizlein@jabber.isoc.org leaves the room [20:38:56] Lars leaves the room: Replaced by new connection [20:38:56] Lars joins the room [20:38:59] shikob leaves the room [20:39:21] Slide design goals [20:40:51] Slide simplest topology [20:41:02] Slide extra delay [20:41:38] shikob joins the room [20:42:05] resnick joins the room [20:42:06] Slide queuing delay target [20:42:21] resnick leaves the room [20:43:55] Slide need to measure delay [20:44:44] Slide queuing delay estimate [20:47:44] Slide controller [20:50:51] Slide max rampup rate [20:53:25] Slide halve on loss [20:55:35] Slide saverio's alternate [20:55:54] csp leaves the room [20:56:42] danwing joins the room [20:59:06] Slide yield to tcp [21:01:57] I v bejnum @ mike [21:02:58] ...@ mike [21:06:25] Mark handley @ mike [21:07:35] Slide one way delay [21:09:15] Slide route changes [21:12:46] Slide timestamp errors [21:12:56] M handley @ mike [21:16:31] I v bejnum @ mike [21:22:15] shikob leaves the room [21:23:49] Slide noise filtering [21:25:52] I'll be right back [21:26:56] danwing leaves the room [21:26:59] momose leaves the room [21:29:10] momose joins the room [21:29:31] Slide safety and worst case [21:29:39] M handley @mike [21:30:36] Lars Eggert @ mike [21:30:54] Bob Briscoe @ mike [21:36:24] ? @ mike [21:38:07] John daily? @ mike [21:41:33] Bob Briscoe @ mike [21:45:08] Lars Eggert @ mike [21:45:31] Bob Briscoe @ mike [21:45:59] I v Bejnum @ mike [21:46:14] Slide rough pseudocode [21:48:14] AdamUzelac leaves the room [21:48:15] momose leaves the room [21:49:03] momose joins the room [21:50:28] Bob Briscoe @ mike [21:50:58] John D @ mike [21:52:51] Richard Bennet @ mike [21:55:23] mathis@psc.edu joins the room [21:55:27] shikob joins the room [21:55:48] vijay.gurbani leaves the room [21:55:49] momose leaves the room [21:55:50] shikob leaves the room [21:55:50] gorryf leaves the room [21:56:08] Mat Ford leaves the room [21:57:39] Lars Eggert @ mike [21:57:58] Henning schulzrinne @ mike [21:59:23] Lars @ mike [21:59:24] iljitsch leaves the room [22:00:04] Atarashi Yoshifumi leaves the room [22:00:11] I would be prepared to review it and have just read it (as the meeting was going). [22:00:11] Henning @ mike [22:00:24] That was related to first draft BTW [22:02:12] V gurbani @ mike [22:02:28] yes, that was the one... [22:03:14] Lars @ mike [22:03:47] M Handley @ mike [22:04:51] Apologies for misspelled names [22:05:05] maciekwojciechowski leaves the room [22:05:08] Lars leaves the room [22:05:08] (Session ended) [22:05:17] wolfgang.beck01 leaves the room [22:06:20] PasiS leaves the room [22:13:08] mathis@psc.edu leaves the room [22:24:40] Lars joins the room [22:24:50] PasiS joins the room [22:26:05] Lars leaves the room [22:26:13] PasiS leaves the room [23:15:02] toby moncaster leaves the room