IETF
rtcweb@jabber.ietf.org
Wednesday, May 21, 2014< ^ >
martin.thomson has set the subject to: RTCWEB WG http://tools.ietf.org/wg/rtcweb/agenda
Room Configuration
Room Occupants

GMT+0
[00:49:38] Thiago Marinello leaves the room: Machine going to sleep
[02:07:11] hta joins the room
[02:30:06] hta leaves the room
[03:57:54] spencerdawkins joins the room
[04:03:52] spencerdawkins joins the room
[05:46:01] chr1sw3ndt leaves the room
[05:46:15] chr1sw3ndt joins the room
[05:48:28] chr1sw3ndt leaves the room
[06:02:25] chr1sw3ndt joins the room
[07:48:34] spencerdawkins leaves the room
[07:48:34] spencerdawkins joins the room
[09:47:34] hta joins the room
[10:28:16] hta leaves the room
[10:51:34] spencerdawkins leaves the room
[11:29:27] spencerdawkins leaves the room
[12:03:33] spencerdawkins joins the room
[12:20:43] coopdanger joins the room
[12:21:15] hta joins the room
[12:22:12] Magnus Westerlund joins the room
[12:26:08] coopdanger leaves the room
[12:26:08] spencerdawkins joins the room
[12:26:16] coopdanger joins the room
[12:27:20] burn joins the room
[12:27:35] Cullen Jennings joins the room
[12:28:06] <Cullen Jennings> HI all - we will be starting soon
[12:29:29] <jesup> hta: did you see Tuexen's email?
[12:31:23] <jesup> I'm thinking have the values come from the W3 side, since data-channel is purposely more generic than WebRTC.  clue might not use 4 levels, etc
[12:32:41] <jesup> hta: that's also what we'd agreed to earlier per meeting notes and emails
[12:35:24] <hta> jesup: I did see it - if I can get 4 levels from SCTP with the required characteristics, I can specify 4 levels in the W3C API, and put the mapping to the SCTP priority numbers in the transport document (which is IETF). I think that's the right places to have it.
[12:38:55] <hta> I think it was actually Cullen who invented the 4 values as part of the DSCP-codepoint document.....
[12:40:03] <hta> jesup, will you hold tuexen's feet to the fire to type up the RR priority scheduler definition?
[12:41:02] <jesup> sure
[12:42:19] chr1sw3ndt leaves the room
[12:42:32] mary.h.barnes joins the room
[12:45:35] jlcJohn joins the room
[12:47:56] <jesup> q+
[12:49:54] <spencerdawkins> Is "MGW on fire" just a reference to the processing load on the MGW? Sorry for the newbie question.
[12:50:07] <hta> if the scaling problem you're worried about is scaling to 1, what is the problem?
[12:50:39] <hta> spencer, yes, it's from a presentation at the Paris IETF where the presenter was arguing for EKT to give end-to-end encryption between SDES and DTLS-SRTP users.
[12:51:06] <Cullen Jennings> @spencer - yes. But the point of Dan's slide was certinaly not we should remove security
[12:51:33] <spencerdawkins> mta and cullen: thanks!
[12:53:27] chr1sw3ndt joins the room
[12:58:01] <hta> a hardware offload that handles your unencrypted data .... gives me a warm and fuzzy feeling.
[12:58:10] <hta> how long do we have to spend on this before we say "no"?
[13:07:29] rbarnes joins the room
[13:15:11] Ted.H joins the room
[13:17:06] Jonathan Lennox joins the room
[13:19:12] ekr@ecotroph.net joins the room
[13:20:07] <Jonathan Lennox> md5?  Really?
[13:20:25] <Ted.H> Do you want that reflected to the room?
[13:20:25] <ekr@ecotroph.net> In an example? Seems reasonable
[13:20:27] <rbarnes > jonathan: my reaction, exactly
[13:20:33] <Jonathan Lennox> Ted: no, just snarking.
[13:21:13] <Cullen Jennings> Martin trolls the AD :-)
[13:22:30] <rbarnes > http://xkcd.com/221/
[13:28:36] <ekr@ecotroph.net> https://github.com/rtcweb-wg/security-arch
[13:29:54] <rbarnes > jesup: fyi, you have been giving us a thumbs-up for like the last 30min.  we appreciate your support :)
[13:32:36] <jesup> rbarnes : I video-muted.  Maybe I should give a frown and freeze it for you....
[13:33:02] <ekr@ecotroph.net> jesup: look solemnt
[13:33:14] <rbarnes > security is serious business
[13:33:48] <jesup> that better?
[13:34:02] <rbarnes > now it's just black
[13:34:18] <jesup> File a bug on Justin
[13:36:50] <jesup> hta: the 4-value table: should that be in the establishment spec (data-protocol (DCEP)) instead, and leave data-channel just state the values are fed into ndata's weighted-fair-queuing?
[13:47:54] <jesup> hta: ping ^
[13:52:13] rbarnes leaves the room
[13:56:53] <jesup> hta: (or anyone else):  If we split the table over to data-protocol (DCEP establishment protocol), it could look like this:  Priority: 2 bytes (unsigned integer): The priority of the channel as described in [I-D.ietf-rtcweb-data-channel]. The higher the number, the lower the priority.  For use in WebRTC, the values used should be one of 128, 256, 512 or 1024.
[13:57:44] rbarnes joins the room
[13:57:50] <jesup> This keeps the protocols generic, but sets the values to be used for webrtc.
[13:58:22] <Ted.H> Jesup:  That appears to lose the weighted fair queueing statmenet
[13:58:27] <Ted.H> er, statement
[13:58:35] <rbarnes > i'm a little confused on what this extension would say.  is it a flag that says "don't give this to JS"?  an origin tag? something else?
[13:58:39] <hta> jesup: that works for me. either the table or -transport- has to say that 128 is "extra high", 256 is "high", 512 is "normal" and 1024 is "below normal" (or the other way round).
[13:58:40] <jesup> Ted.H: that change would still be in data-channel
[13:59:36] <jesup> Ted.H: so, the previous text I proposed for data-channel, and this for data-protocol (establishment protocol)
[13:59:37] <Ted.H> Okay, as long as we know that "extra high"'s relationship to "high" is determined by the weighted fair queuing method, I think we're fine.
[13:59:44] <jesup> ok
[13:59:52] <jesup> I'll propose that to the list today
[14:00:20] <Ted.H> By spinning the draft? (hint hint)
[14:00:39] <jesup> Sure.
[14:00:54] <Ted.H> Thanks!
[14:14:16] jesup raises hand for both
[14:15:23] <jesup> hta: The priority of the channel as described in [I-D.ietf-rtcweb-data-channel]. The higher the number, the lower the priority.  For use in WebRTC, the values used should be one of 128 ("extra high"), 256 ("high"), 512 ("normal") or 1024 ("below normal").
[14:15:51] <hta> works for me!
[14:17:15] rbarnes leaves the room
[14:35:19] rbarnes joins the room
[14:52:40] <Jonathan Lennox> When would there be TCP at all?
[14:52:51] <Jonathan Lennox> Oh, n/m
[15:01:53] <spencerdawkins> FWIW, I'm available for the rest of the day, including dinner, if anyone wants to chat about TSV-ish topics in more detail. Just let me know ...
[15:02:16] Jonathan Lennox leaves the room
[15:03:02] Jonathan Lennox joins the room
[15:03:04] chr1sw3ndt leaves the room
[15:10:38] mary.h.barnes leaves the room
[15:27:10] Ted.H leaves the room
[15:38:35] Jonathan Lennox leaves the room
[16:00:49] coopdanger leaves the room
[16:02:11] ekr@ecotroph.net leaves the room
[16:19:42] rbarnes leaves the room
[16:36:13] ekr@ecotroph.net joins the room
[16:36:28] rbarnes joins the room
[16:41:26] rbarnes leaves the room
[17:02:06] coopdanger joins the room
[17:02:57] rbarnes joins the room
[17:32:25] Cullen Jennings leaves the room
[17:33:22] jlcJohn leaves the room
[19:16:20] rbarnes leaves the room
[19:20:23] rbarnes joins the room
[19:44:33] sean.turner@jabber.psg.com joins the room
[19:58:34] sean.turner@jabber.psg.com leaves the room
[20:04:11] rbarnes leaves the room
[20:08:25] coopdanger leaves the room
[20:11:33] Magnus Westerlund leaves the room: I'm happy Miranda IM user. Get it at http://miranda-im.org/.
[20:18:17] hta leaves the room
[20:35:23] burn leaves the room
[20:35:23] burn joins the room
[20:35:23] burn leaves the room
[20:37:55] ekr@ecotroph.net leaves the room
[20:38:23] ekr@ecotroph.net joins the room
[20:44:26] ekr@ecotroph.net leaves the room
[20:49:43] spencerdawkins leaves the room
[20:52:44] spencerdawkins leaves the room
[21:02:38] coopdanger joins the room
[22:37:54] rbarnes joins the room
[22:38:00] rbarnes leaves the room
[22:52:16] spencerdawkins joins the room
[23:00:05] spencerdawkins joins the room
[23:19:17] spencerdawkins joins the room
[23:20:22] spencerdawkins leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!