IETF
rtcweb@jabber.ietf.org
Thursday, 17 November 2011< ^ >
martin.thomson has set the subject to: RTCWEB WG http://tools.ietf.org/wg/rtcweb/agenda
Room Configuration

GMT+0
[00:27:08] Bran, Cary joins the room
[00:30:50] Simon Romano joins the room
[00:31:44] MeetechoAudio joins the room
[00:31:47] Matthew Kaufman joins the room
[00:35:28] Matthew Kaufman leaves the room
[00:35:57] EKR joins the room
[00:36:45] pm joins the room
[00:49:11] Matthew Kaufman joins the room
[00:49:40] <Simon Romano> Presentation stopped
[00:50:47] EKR leaves the room
[00:50:51] Simon Romano leaves the room
[00:51:32] Simon Romano joins the room
[00:52:24] m&m joins the room
[00:52:56] jdrosen joins the room
[00:53:00] Cullen Jennings joins the room
[00:53:11] <jdrosen> hi all. anyone else having problems with meetecho?
[00:53:35] burn joins the room
[00:53:43] EKR joins the room
[00:54:32] Lorenzo Miniero joins the room
[00:54:32] EKRm joins the room
[00:54:51] <m&m> is EKR in witness protection?
[00:54:54] Jean Mahoney joins the room
[00:55:05] <Simon Romano> Presentation stopped
[00:55:06] Jonathan Rosenberg joins the room
[00:55:27] <Simon Romano> Slide 1: Proposed WebRTC Security Architecture
[00:56:44] Tim Panton joins the room
[00:56:46] Ethan Hugg joins the room
[00:56:58] Randell Jesup joins the room
[00:57:45] <Randell Jesup> I'm here
[00:58:11] Hendrik Scholz joins the room
[00:58:18] Hadriel Kaplan joins the room
[00:58:22] <m&m> those remote…please preface your questions with MIC:
[00:58:32] tsuichi joins the room
[00:58:36] <Simon Romano> mEETECHO IS @ taipei1.conf.meetecho.com/WebLite/login.jsp?ietf=RTCWEB
[00:58:37] <Randell Jesup> I may need someone to do a fair bit of repeating for me when Data is being presented as I'm a primary author
[00:58:48] Neil Stratford joins the room
[00:58:59] simo.veikkolainen joins the room
[00:59:08] Alessandro Amirante joins the room
[00:59:09] <Lorenzo Miniero> http://www.meetecho.com/ietf82/rtcweb
[00:59:17] lef_jp joins the room
[00:59:23] Jan Skoglund joins the room
[00:59:27] <Hadriel Kaplan> Randell: m&m is your remote mic, as well as not melting in your hands
[00:59:38] <m&m> (-:
[00:59:46] Colman Ho joins the room
[00:59:57] suzukisn joins the room
[01:00:17] Paul Hoffman joins the room
[01:00:46] Ralph Giles joins the room
[01:00:51] <Simon Romano> mEETECHO AUDIO STREAM IS @ rtsp://taipei1.conf.meetecho.com/3330018.sdp
[01:01:01] patrick Mourot joins the room
[01:01:13] Barry Leiba joins the room
[01:01:29] <jdrosen> hmm, anyone else have a big volume difference between ted and ekr?
[01:01:48] Rifaat joins the room
[01:01:57] <Simon Romano> Slide 2: Trust Model
[01:03:00] <Simon Romano> Slide 3: Authenticated Entities
[01:03:23] coopdanger joins the room
[01:03:29] jesup joins the room
[01:04:15] <Simon Romano> Slide 4: Unauthenticated Entities
[01:05:02] <Simon Romano> Slide 5: Basic Design Principle: As good a job as
[01:05:09] Jonathan Lennox joins the room
[01:05:28] tterribe leaves the room
[01:06:29] hardie joins the room
[01:06:33] Rifaat leaves the room
[01:07:09] rillian joins the room
[01:07:14] <Simon Romano> Slide 6: Overall Topology
[01:08:00] gmc joins the room
[01:08:26] <Simon Romano> Slide 7: Call Flow (I)
[01:10:26] Kepeng joins the room
[01:11:59] tandonghui joins the room
[01:12:04] <Simon Romano> Slide 8: Call Flow (II)
[01:13:20] <Simon Romano> Slide 9: Permissions Models
[01:13:27] <Simon Romano> Slide 8: Call Flow (II)
[01:13:40] Jim McEachern joins the room
[01:14:18] Youngwan So joins the room
[01:14:36] <Simon Romano> Slide 9: Permissions Models
[01:14:42] <Simon Romano> Slide 8: Call Flow (II)
[01:14:52] Linyi Tian joins the room
[01:15:19] <Simon Romano> Slide 9: Permissions Models
[01:15:46] Youngwan So leaves the room
[01:17:51] Lorenzo Miniero leaves the room
[01:18:04] mph joins the room
[01:18:39] <Simon Romano> Slide 10: Permissions API
[01:18:47] <Simon Romano> Slide 9: Permissions Models
[01:19:05] mph leaves the room
[01:19:44] <Simon Romano> Slide 10: Permissions API
[01:20:12] ben joins the room
[01:20:17] suhasHere joins the room
[01:20:19] ben is now known as Ben’s iPad
[01:20:51] Ben’s iPad leaves the room
[01:21:26] ben@nostrum.com joins the room
[01:21:34] Lorenzo Miniero joins the room
[01:22:03] <Simon Romano> Slide 11: Who “owns” the permissions”
[01:22:10] <Simon Romano> Slide 10: Permissions API
[01:22:34] victor.pascual.avila joins the room
[01:22:47] <Simon Romano> Slide 11: Who “owns” the permissions”
[01:22:55] Youngwan So joins the room
[01:23:05] <Simon Romano> Slide 10: Permissions API
[01:23:40] tterribe joins the room
[01:23:57] Lorenzo Miniero leaves the room
[01:23:58] tterribe leaves the room
[01:24:06] Youngwan So leaves the room
[01:24:08] Youngwan So joins the room
[01:24:43] tterribe joins the room
[01:24:58] tterribe leaves the room
[01:25:37] Youngwan So leaves the room
[01:25:52] tterribe joins the room
[01:25:59] tterribe leaves the room
[01:26:21] hardie leaves the room: Replaced by new connection
[01:26:22] hardie joins the room
[01:26:50] <m&m> <reminder type='friendly'>if you want me to relay your question, preface it with "MIC:"</reminder>
[01:27:05] <Simon Romano> Slide 11: Who “owns” the permissions”
[01:27:36] tterribe joins the room
[01:27:58] tterribe leaves the room
[01:28:31] <Simon Romano> Slide 12: Permissions UI
[01:30:19] tterribe joins the room
[01:30:27] rlb joins the room
[01:30:29] tterribe leaves the room
[01:32:39] <Simon Romano> Slide 13: Why HTTP origins are a problem
[01:34:24] tterribe joins the room
[01:34:28] tterribe leaves the room
[01:34:29] <Simon Romano> Slide 14: Web Security Issues
[01:34:43] <jesup> mic: note that camera indicator lights are often ignored or not understood by users. trojaned computers have been used to record audio and video from people's homes and blackmail them (recent news story involving hundreds of people, mostly women, mostly in CA). I don't consider hardware lights to be guaranteed to be enough notice, even if present.
[01:35:04] <jesup> inject at a reasonable breakpoint
[01:35:22] Magnus Westerlund joins the room
[01:36:18] <jesup> mic: not-mixed can become mixed after page is loaded, I believe - may be tough to handle
[01:36:20] <rillian> jesup: I think most of those have been still images, where it's easy to miss the camera light flash. (or decide you imagined it)
[01:36:30] <rillian> which might be an argue for chrome advertising permission
[01:36:48] <rillian> also, it's unclear how many of those camera lights can be software controlled
[01:37:33] <rillian> an *argument* for chrome
[01:37:38] <Jonathan Lennox> I note that the permissions model described would make the usage we're doing right now impossible -- EKR couldn't see or hear us whenever Ted switches over to the slides.
[01:37:41] Xavier Marjou joins the room
[01:38:00] Cullen Jennings leaves the room
[01:38:12] <jesup> rillian: this case involved full realtime recordings
[01:38:12] <burn> Jonathan, i think that would be a great mic comment
[01:38:20] <suhasHere> Doesn't having an indicator per page (drawn by the browser) tell user if the camera is on or off
[01:38:23] <rillian> ouch
[01:38:24] <jesup> Good point
[01:38:28] tterribe joins the room
[01:38:28] tterribe leaves the room
[01:38:30] <EKRm> Jonathan, I'm not sure that's actually true. There could be an indicator in the tab.
[01:38:33] <EKRm> I.e., the tab throbs.
[01:39:02] Roy joins the room
[01:39:10] <Jonathan Lennox> I don't see a tab on the screen...all I see is a slide deck and scrollbars.
[01:39:33] Cullen Jennings joins the room
[01:39:43] <Jonathan Lennox> At the mic now to say it
[01:40:14] <jesup> j lennox: I think the camera-active needs to be in 'chrome', and audio/video stay running even if the tab isn't visible. If another *window* overlays the indicator, that's different. But may be hard to know
[01:40:25] Ram Mohan R joins the room
[01:40:49] patrick Mourot leaves the room
[01:41:11] patrick Mourot joins the room
[01:42:13] <Simon Romano> Slide 15: Web Security and State Machine in JS
[01:42:26] <Roy> i used to use a music player. it can add the music title in any window which is in the top.
[01:42:32] tterribe joins the room
[01:42:58] tterribe leaves the room
[01:43:41] <Simon Romano> Slide 16: Communications Consent
[01:43:49] <Simon Romano> Slide 15: Web Security and State Machine in JS
[01:44:08] rlb leaves the room
[01:44:17] hta joins the room
[01:44:25] <Simon Romano> Slide 16: Communications Consent
[01:46:17] <Jonathan Rosenberg> MIC: ekr, why is verification of continuing consent important? what attacks is that addressing?
[01:46:36] tterribe joins the room
[01:46:44] rlb joins the room
[01:46:59] tterribe leaves the room
[01:48:05] <EKRm> JDR: I can address this verbally, but consider what happens if I visit your internet cafe, set up a high bandwidth channel and then take off. I've now spammed you.
[01:49:18] <m&m> Jonathan Rosenberg: do you still want this asked?
[01:49:25] <EKRm> it's not a huge deal.
[01:49:42] <m&m> will assume yes sans notice
[01:50:22] <Jonathan Rosenberg> @ekr I think from this discussion its worth you clarifying the threats
[01:50:36] Roy leaves the room
[01:50:40] tterribe joins the room
[01:50:58] tterribe leaves the room
[01:51:42] <Simon Romano> Slide 17: IP Location Privacy
[01:51:43] Jan Skoglund leaves the room
[01:53:07] Roy joins the room
[01:54:11] hardie leaves the room
[01:54:44] tterribe joins the room
[01:54:59] tterribe leaves the room
[01:56:33] patrick Mourot leaves the room
[01:58:37] <jdrosen> MIC: This has to be at the site level; since there may be other authorization checks (like the caller is in my buddy list) which obviate the need to do ICE before call answer
[01:58:38] Linyi Tian leaves the room
[01:58:46] spromano781 joins the room
[01:58:48] tterribe joins the room
[01:58:59] tterribe leaves the room
[01:59:24] <rillian> jdrosen: or at least it must be possible deny ice at the site level
[01:59:57] <jdrosen> yes - my point is, this is not a browser configuration or browser chrome issue - its site level. API needs to allow site to do either
[02:00:09] <EKRm> well, JDR, I agree with that.
[02:01:22] Lorenzo Miniero joins the room
[02:02:50] <jesup> mic: make sure you cover the case of the battered-wife hiding from husband
[02:02:52] tterribe joins the room
[02:02:59] tterribe leaves the room
[02:03:23] <jesup> which was the point of my pushing the "force TURN-only until answered)
[02:03:35] gonzalo joins the room
[02:03:59] <Simon Romano> Slide 18: Communications Security: Implementation
[02:06:47] <jesup> mic: how does this avoid MiTM bid-down attacks?
[02:06:56] tterribe joins the room
[02:06:58] <jesup> That's why I'm interested in MUST USE
[02:06:59] tterribe leaves the room
[02:07:23] <EKRm> Jesup: Absolutely, that's an issue here.
[02:07:28] <EKRm> You know what I would prefr.
[02:07:32] <EKRm> And it is MUST USE.
[02:07:33] <Cullen Jennings> I'm in the Q behind Mat
[02:07:35] Lorenzo Miniero leaves the room
[02:08:34] hta leaves the room
[02:08:41] tomtaylor joins the room
[02:09:45] <jesup> Agreed about it being a good marketing point (and a good-feeling for usres)
[02:10:05] <m&m> jesup: do still want me to relay?
[02:10:25] <jesup> Yes, please - I want the discussion of bid-down
[02:10:31] <m&m> /noted
[02:11:13] tterribe joins the room
[02:11:18] <rlb> since the line has been cut, i'll do my rant in here
[02:11:29] tterribe leaves the room
[02:11:45] t01 joins the room
[02:11:49] <jesup> not for the line: inability to avoid bid-down is a critical flaw
[02:11:54] <rlb> the overriding point here is that the benefit of MUST use is illusory unless you have authentication that goes along with this
[02:12:16] <rlb> Hadriel notes the marketing benefit from Skype's security model, but that's not what we're talking about here
[02:12:30] <jesup> You're depending on the user to understand what the "insecure!" warning in the UI means - if they see it
[02:12:46] <rlb> Skype doesn't just have encrypted channels, they're also authenticated; the same is not true of RTCWEB
[02:13:04] <Matthew Kaufman> they could be
[02:13:08] t01 leaves the room
[02:13:14] <Matthew Kaufman> see my previous comment
[02:13:17] <Jonathan Lennox> That was the identity service proposal, no?
[02:13:19] suhasHere leaves the room
[02:13:22] <Simon Romano> Slide 19: Communications Security: API Requirement
[02:13:27] <rlb> so even if you get some brownie points for having encryption always turned on, you're still completely open to MitM
[02:13:37] <Cullen Jennings> I was going to say at the mic that I ahve a hard time deciding where to come down on this withotu understanding the rest of security model. There is not point in making this stronger than the weekest link. If we allow SDES over HTTP, not cleear that mandating SRTP helps much
[02:13:38] <rlb> Kaufman: Which one?
[02:13:44] <rillian> The user agent can also cache credentials, like the ssh model
[02:13:44] <jesup> SDES opens you to MITM by the website
[02:13:59] <Matthew Kaufman> my first comment. about using the fingerprint from the DTLS + server side auth.
[02:14:14] Jan Skoglund joins the room
[02:14:23] <Simon Romano> Slide 20: Communications Security: UI [based on dr
[02:14:23] <Matthew Kaufman> SDES doesn't have PFS. that's a much better issue than MITM by the website, etc.
[02:14:24] <rlb> Kaufman: but you still need some sort of identity binding
[02:14:40] <jesup> I strongly oppose SDES unless someone can convince me it doesn't open a big MiTM hole (to the website)
[02:15:00] <jesup> And I agree with matthew
[02:15:05] tterribe joins the room
[02:15:19] <Matthew Kaufman> @rlb: facebook can provide that (for instance)
[02:15:29] tterribe leaves the room
[02:15:42] <rlb> Kaufman: right, but unless that's universal, there's no point to making SRTP universal
[02:15:44] <Cullen Jennings> @jesup, but questions will be are you equally apposed to all other things that would allow a MiTM by the website
[02:15:44] <jesup> Yes, an no user understands this
[02:15:50] <Simon Romano> Slide 21: Example IdP Interaction: BrowserId
[02:15:51] <jesup> s/an/and/
[02:16:08] <rlb> on a more meta level, a lot of this seems like we're over-reaching what we need for interop; all these "MUST use" directives could be implemented by policy; they're not necessary for interop
[02:16:11] <Matthew Kaufman> @rlb: i disagree with that as well, for several reasons that i'd rather take to the list and not just put in chat.
[02:16:14] patrick mourot joins the room
[02:16:16] tterribe joins the room
[02:16:28] <rlb> Kaufman: But htat would require me to join the rtcweb list :)
[02:16:29] tterribe leaves the room
[02:16:48] <m&m> rlb: man up! (-:
[02:16:49] <Matthew Kaufman> rlb: you could just email me directly. matthew.kaufman@skype.net. i'll debate anyone.
[02:17:10] <jesup> rlb: take a speed-reading course ;-)
[02:17:34] <Matthew Kaufman> (note that my normal IETF email address isn't that, and gets lower priority)
[02:17:45] <rlb> jesup: speed filter to /dev/null :)
[02:17:48] <Hadriel Kaplan> Cullen: I'm ok with only allowing SDES if JS is HTTPS
[02:18:02] <Simon Romano> Slide 22: Example ROAP OFFER with BrowserID
[02:18:07] <Simon Romano> Slide 21: Example IdP Interaction: BrowserId
[02:18:08] <Matthew Kaufman> @Hadriel: why is lack of PFS ok?
[02:18:18] Paul Hoffman leaves the room
[02:18:24] <Hadriel Kaplan> for the reasons I said last night :)
[02:18:44] patrick mourot leaves the room
[02:18:45] <Matthew Kaufman> @Hadriel: put those reasons on the list
[02:18:57] <rlb> going into PFS seems *way* too fine-grained for this WG
[02:19:11] Paul Hoffman joins the room
[02:19:15] <jesup> hadriel: I'm not, because per the threat model I don't trust the service providers in all cases
[02:19:22] <Matthew Kaufman> @rlb: disagree. given the choice between IETF and W3C, i think it belongs here.
[02:19:37] <jesup> hadriel: I was refering to SDES over https
[02:19:43] <rlb> kaufman: how is it not a completely general DTLS question?
[02:19:48] patrick mourot joins the room
[02:19:54] <Hadriel Kaplan> if you really don't trust them, DTLS-SRTP won't help either
[02:19:55] <rlb> (agree it should be ietf and not w3c)
[02:20:37] <Hadriel Kaplan> rlb: what other WG should it go into?
[02:20:38] <Simon Romano> Slide 22: Example ROAP OFFER with BrowserID
[02:20:47] <Matthew Kaufman> rlb: we're providing a "use these parts of the toolkit" recommendation. the toolkit has DTLS-SRTP and SDES-keyed SRTP. one of those can (and often does) have PFS. the other cannot.
[02:20:55] <Jonathan Lennox> BTW, does anyone know whether EKT breaks PFS for DTLS-SRTP?
[02:21:18] <Cullen Jennings> @lenox, my understanding is that it does not
[02:21:25] <rlb> hadriel: tls?
[02:21:27] juberti@gmail.com/Meebo joins the room
[02:21:34] patrick mourot leaves the room
[02:22:00] <rlb> kaufman: oh, ok, i would be completely fine with requiring dtls-srtp. i thought we were talking about like which ciphersuites you would have to use, which is not cool
[02:22:12] <Jonathan Lennox> Cullen: it was describing non-PFS as "server-provided keys" that triggered my worry...maybe that's just an inprecise description of the problem.
[02:22:57] <Matthew Kaufman> @rlb: well, again, we're recommended which parts of the toolkit. recommending for/against some of the ciphersuites is a good idea. for instance, MUST use NULL cipher isn't very helpful.
[02:24:02] <rlb> kaufman: but still, requiring specific ciphersuites is usually not done,
[02:24:30] <m&m> anything from the room to the mic?
[02:24:41] <rlb> otherwise, we wouldn't need ciphersuite negotiation in tls
[02:26:18] Cullen Jennings leaves the room
[02:26:37] <Simon Romano> Slide 23: Example JSEP Transport Info with Browser
[02:26:55] sal joins the room
[02:27:15] <Simon Romano> Slide 24: Generic Third-Party Identity Assertions
[02:28:01] <Jonathan Lennox> Hm, 5479 claims that EKT has "No PFS" (A.5.2) with no further elaboration.
[02:28:24] <Simon Romano> Slide 25: Questions?
[02:28:40] Cullen Jennings joins the room
[02:29:15] <rlb> who's at the mic?
[02:29:15] cabo joins the room
[02:30:05] <Roy> someone from ALU
[02:30:07] <m&m> I couldn't catch it, sorry /-:
[02:30:36] <m&m> any questions from the room?
[02:30:51] <m&m> the line is currently 3 deep
[02:31:53] tomtaylor leaves the room
[02:32:52] <rlb> chairs: could we please not have scary huge ekr?
[02:33:22] <Cullen Jennings> The huge scare EKR is here to help people make their comments breif and to the point
[02:33:54] ben@nostrum.com leaves the room
[02:33:55] <Cullen Jennings> (and not sure how to make him small on teds computer )
[02:33:59] <m&m> I feel like I'm talking to some supervillian (-:
[02:34:06] <Cullen Jennings> you are
[02:34:14] <m&m> but this makes it much more explicit
[02:34:16] <burn> can we zoom in on his head?
[02:34:27] <rlb> wouldn't be quite so bad if the lighting were better
[02:34:35] <burn> with a caption that says "Dr. Evil"
[02:34:53] <juberti@gmail.com/Meebo> http://www.youtube.com/watch?v=OYecfV3ubP8
[02:35:23] <rlb> juberti: +1
[02:35:36] <rlb> if only i had brought my hammer
[02:35:44] <jesup> hum
[02:35:52] <jdrosen> i read it
[02:35:57] <jesup> hum
[02:36:01] <EKR> i wrote it but did not read it.
[02:36:11] spromano781 leaves the room
[02:36:17] <EKR> I concur with that view as well.
[02:36:50] <EKR> (i.e., independence)
[02:38:14] spromano781 joins the room
[02:38:14] <m&m> anything from the room?
[02:38:23] <EKR> hum the way I hum.
[02:39:00] <Jonathan Lennox> EKR: presumably this identity stuff is going to need to be mapped to SIP at some point, for federation ... have you thought about that at all?
[02:39:10] Pen joins the room
[02:39:13] <jesup> hum
[02:39:20] sftcd joins the room
[02:39:21] <EKR> Jonathan: about 70% as much as I've thought about the generic identity piece.
[02:39:22] <jdrosen> i understand
[02:39:34] <EKR> hum
[02:39:35] <jesup> hum
[02:39:48] <jdrosen> not ready
[02:40:06] <rlb> not ready
[02:40:15] <rlb> in the room, but in the back, so don't think i was heard
[02:40:38] <jdrosen> MIC: I think the identity work is way too early to be adopted as a working item.
[02:40:39] <Simon Romano> Presentation stopped
[02:41:42] <Simon Romano> Slide 1: RTC-Web Codec
[02:41:50] <jdrosen> well, that was clear as mud about identity work but ok
[02:41:53] <Hadriel Kaplan> JDR: we didn't take the identity part
[02:41:57] <Jonathan Lennox> EKR: and it wouldn't be possible to map to 4474, right?
[02:42:04] <Hadriel Kaplan> ya i agree
[02:42:22] <EKR> JLennox: you mean without a cryptographic transformation?
[02:42:36] <EKR> probably not
[02:42:50] <EKR> But I think you could use the same mechanism to handle 4474
[02:43:01] <Jonathan Rosenberg> 4474bis ekr?
[02:43:06] <Simon Romano> Slide 2: 01 - Submission Info
[02:43:18] <EKR> I.e., you could build a 4474 signer/verifier within this framework
[02:43:23] <EKR> 4474.1?
[02:43:37] <Simon Romano> Slide 3: Media Processing
[02:43:54] Ralph Giles leaves the room
[02:44:47] <EKR> gave I mentioned I have now started boozing
[02:45:03] Tim Panton leaves the room
[02:45:06] <Jonathan Lennox> You should have been boozing while you were a giant head
[02:45:28] Ralph Giles joins the room
[02:45:38] <jesup> justin++
[02:45:39] <EKRm> I do plan to improve the IdP stuff quite a bit.
[02:45:51] <EKRm> In the nearish future
[02:46:10] <Simon Romano> Slide 4: Audio Codec
[02:46:31] Paul Hoffman leaves the room
[02:47:49] <EKRm> We probably should require the fixed length stuff, yeah
[02:47:54] <EKRm> gah
[02:49:46] <jesup> DTMF is all about interop
[02:50:55] <Simon Romano> Slide 5: Video Codec
[02:51:08] <Jonathan Lennox> Does 4733 imply 2198?
[02:51:44] hta joins the room
[02:51:47] <Ethan Hugg> Only in Browsers?
[02:53:29] <m&m> "*Browser means devices implementing User Agent behavior for W3C HTML5 or WebRTC"
[02:58:16] <m&m> the line is 4 deep … speak now if you want mic channeling
[03:00:06] spromano781 leaves the room
[03:02:54] <gmc> were other codecs discussed, such as theora or H261?
[03:03:13] <m&m> gmc: do you want that asked at the mic?
[03:04:18] <rillian> gmc: they were discussed briefly; the general argument against was that they wouldn't providesufficiently compelling quality and hurt adoption
[03:04:25] tobcast joins the room
[03:04:32] <rillian> I'm not clear why that doesn't apply to 8 kHz PCMA
[03:04:41] <rillian> mjpeg is another option
[03:04:47] <gmc> are you aware of the MPEG internet coding initiative? it should provide a suitable codec too
[03:05:13] <rillian> gmc: that's the MPEG RF group?
[03:05:35] <Jonathan Lennox> rillian: for PCMU/PCMA, SIP/PSTN interop is the compelling factor I think.
[03:05:41] <Roy> that may need quite a long time, comparring with the deadline 2012 march
[03:05:47] <gmc> it aims at providing a video codec with a quality equivalent to H264 baseline, under type I licensing
[03:05:54] <rillian> (also the same vendors who haven't deployed VP8 haven't deployed theora)
[03:05:55] <gmc> no timeline is somewhat aline
[03:06:16] <gmc> is VP8 going to meet the march timeline?
[03:06:39] <gmc> i think the same conditions should be apply to any codecs that are in competition
[03:06:46] <rillian> gmc: what's type I licensing?
[03:06:55] <m&m> <reminder>If you want your comment brought to the mic, please preface with MIC:</reminder>
[03:06:57] <Roy> agree. so why not using h.264 which has been deployed so much
[03:07:30] <Roy> even it is not free, but VP8 may not be able to be free.
[03:07:56] <gmc> rillian: http://www.itu.int/dms_pub/itu-t/oth/04/04/T04040000020002PDFE.pdf
[03:08:07] <rillian> gmc: I also don't know what you mean by "is VP8 going to be meet the march timeline"
[03:08:07] <gmc> type 1 is option 1
[03:08:20] <rillian> what requirement is it not currently meeting?
[03:08:30] <rillian> the current comment aside, anyway
[03:09:07] Michael Graves joins the room
[03:09:25] <gmc> roy: agree, conditions should be the same to all candidates
[03:09:35] <Simon Romano> Slide 5: Video Codec
[03:09:39] <rillian> gmc: right, that would be great to have
[03:10:48] <Jonathan Lennox> rillian: VP8's RTP payload is still a work-in-progress, not yet even in WGLC as far as I recall.
[03:11:04] <rillian> yes, that's an issue
[03:11:10] <Jonathan Lennox> Neither is Opus's, for that matter (still not even a WG item in that case).
[03:11:55] <rillian> there is an implementation of the vp8 draft, however, in the webrtc.org codebase
[03:12:19] <hta> Jonathan: VP8 bitstream RFC was published on Friday. Pushing the RTP payload RFC hasn't been a priority. Do you need it soon?
[03:12:44] <gmc> but not clear how stable is VP8 to meet a march deadline as the new version has just been published
[03:12:48] <Jonathan Lennox> Not sooner than the rest of WebRTC, I guess -- as long as it's not later.
[03:12:52] <jesup> Roy: problematic for a free browser (Firefox) to distribute a royalty-bearing codec (even ignoring mission/etc issues)
[03:12:56] <m&m> line closed
[03:13:33] <Jonathan Lennox> Assuming it's the payload chosen, that is.
[03:13:34] <rillian> gmc: it's not unreasonable for drafts in one group to depend on drafts in another
[03:13:55] <Roy> jesup, I understand this issue. but mandating VP8 doesn't help I think
[03:14:14] <gmc> well, from a non browser perspective you want something stable
[03:14:30] <rillian> and the codec itself have been stable for a year and half
[03:14:34] <Simon Romano> Presentation stopped
[03:14:37] <rillian> and there are hardware implementations available
[03:15:06] <gmc> H264 has been stable and implemented for year. There was a lot of optimisation
[03:15:12] <gmc> that does not exist yet for VP8
[03:16:15] <rillian> sure that's why it's preferred in the draft, if the licensing issues can be solved
[03:16:28] <EKR> I appear to be humming that we would adopt it
[03:17:04] <gmc> rillian: same guaranty should apply to VP8, how is that going to be guarantied?
[03:17:20] <rillian> what guarantee do you mean?
[03:18:07] <Simon Romano> Presentation stopped
[03:18:21] <gmc> would the licensing issue/risk be resolve for VP8 by march?
[03:18:23] <Simon Romano> Slide 1: WebRTC Data Channels
[03:18:42] <Simon Romano> Slide 2: Uses
[03:18:42] Paul Hoffman joins the room
[03:19:01] rlb leaves the room
[03:19:06] <Simon Romano> Slide 3: Data Channel Requirements
[03:19:31] <rillian> gmc: I don't know what risk you're referring to. The only organization which claims IPR on VP8 granted a royalty free license
[03:19:36] victor.pascual.avila leaves the room
[03:19:55] <rillian> the patent system does not provide absolute assurances
[03:20:06] <rillian> which is why the topic is generally off topic here
[03:20:10] <Simon Romano> Slide 4: Options
[03:20:24] <jesup> m&m: I'll be speaking as an author if no one minds to answer questions
[03:20:49] <m&m> got it
[03:21:27] gmc leaves the room
[03:21:45] <Simon Romano> Slide 5: Pseudo-TCP-over-UDP (reliable) DCCP (u
[03:22:15] <rillian> we do hope :)
[03:23:08] <Simon Romano> Slide 6: SCTP-DTLS-(ICE)-UDP or DTLS-SCTP-(ICE)-U
[03:23:15] burn leaves the room
[03:23:54] <EKR> FWIW, I don't think that DTLS over SCTP is that great an option
[03:24:05] <EKR> SCTP over DTLS, otoh seems very reasonable
[03:24:33] <Paul Hoffman> To the mic?
[03:24:42] <m&m> EKR: I'm not going to relay it without the MIC: preface
[03:24:44] <EKR> Not necessary yet.
[03:24:51] <EKR> m&m: thanks. I understnad.
[03:24:51] <m&m> grazie
[03:25:04] <jesup> ekr: I agree
[03:25:21] <Simon Romano> Slide 5: Pseudo-TCP-over-UDP (reliable) DCCP (u
[03:25:45] <rillian> EKR: I agree as well
[03:25:50] Barry Leiba leaves the room
[03:26:04] <EKR> I feel like the big question about SCTP over DTLS is the maturity of the user-land SCTP stack
[03:26:25] <EKR> (or any SCTP)
[03:26:47] <jesup> mic: There are more flows in tcp/DCCP than in sctp, even without full mergeing of CC flows
[03:26:51] <Simon Romano> Slide 6: SCTP-DTLS-(ICE)-UDP or DTLS-SCTP-(ICE)-U
[03:26:55] <Simon Romano> Slide 7: SCTP-DTLS-(ICE)-UDP vs DTLS-SCTP-(ICE)-U
[03:27:31] <Simon Romano> Slide 8: Open issues
[03:28:09] <Simon Romano> Slide 7: SCTP-DTLS-(ICE)-UDP vs DTLS-SCTP-(ICE)-U
[03:28:21] <Simon Romano> Slide 8: Open issues
[03:28:44] <jesup> userland SCTP update in progress, linux/mac available, windows soon
[03:29:00] <m&m> jesup: mic that?
[03:29:07] <Simon Romano> Slide 9: Questions/Discussion
[03:29:07] <EKR> jesup: that's good news.
[03:29:43] <hta> I have downloaded and compiled the userland sctp. haven't yet managed to run a test, though.
[03:29:54] <EKR> Hey, if it compiles it works
[03:30:00] <rillian> hta: some of the test programs ran for me!
[03:30:07] <EKR> Ship it!!!!
[03:30:17] <hta> I managed to get all the way to "address already in use" :-)
[03:30:31] <EKR> 2 per second!
[03:30:59] Paul Hoffman leaves the room
[03:31:02] <jdrosen> nooooo.... not more transport protocols!
[03:31:21] <jesup> kills a lot of our use-cases... or hurts them.
[03:31:32] <jesup> wrong thing?
[03:32:02] <rillian> only in the sense that all human efforts are flawed
[03:32:09] Bran, Cary leaves the room
[03:32:27] <EKR> Is CLUE in the same room?
[03:32:32] <Simon Romano> That's all folks!
[03:32:39] <Simon Romano> See you soon on Meetecho ;-)
[03:32:46] sal leaves the room
[03:32:49] <Jonathan Lennox> No, it's downstairs I believe
[03:32:59] suzukisn leaves the room
[03:32:59] Hadriel Kaplan leaves the room
[03:32:59] coopdanger leaves the room
[03:33:00] <jesup> Can we hum for consensus on SCTP vs tcp/dccp?
[03:33:08] <jesup> too late?
[03:33:12] <m&m> yeah
[03:33:13] <EKR> I would like to attend the CLUE thing. Can someone see about pinging me in?
[03:33:13] gonzalo leaves the room
[03:33:14] sftcd leaves the room
[03:33:14] Hendrik Scholz leaves the room
[03:33:15] Ram Mohan R leaves the room
[03:33:16] <jesup> ok
[03:33:17] <EKR> s/pinging/bringing/
[03:33:18] simo.veikkolainen leaves the room
[03:33:18] <Jonathan Lennox> CLUE thing is in 101D
[03:33:19] <jesup> on the list then
[03:33:20] Jean Mahoney leaves the room
[03:33:21] <m&m> I htink i got a evil glare from Ted
[03:33:26] <rillian> jesup: indeed
[03:33:32] <jesup> no problem
[03:33:33] <EKR> or is it going to have a working stream
[03:33:38] m&m leaves the room
[03:33:40] <rillian> thanks everyone
[03:33:45] tandonghui leaves the room
[03:33:45] <Jonathan Lennox> Dunno...
[03:33:45] Cullen Jennings leaves the room
[03:33:45] <jesup> thanks!
[03:33:47] hta leaves the room
[03:33:50] <Simon Romano> Presentation stopped
[03:33:51] <Jonathan Lennox> EKR: will enquire
[03:33:51] <EKR> OK, I'm listening to it.
[03:33:54] <EKR> thanks!
[03:33:55] <Jonathan Lennox> Also which jabber room is appropriate
[03:33:56] <jesup> thanks m&m
[03:33:58] <jdrosen> bye
[03:33:59] tsuichi leaves the room
[03:34:01] jdrosen leaves the room
[03:34:14] <jesup> Yeah, any way to take part in Clue?
[03:34:30] Pen leaves the room
[03:34:37] Kepeng leaves the room
[03:34:46] Magnus Westerlund leaves the room: I'm happy Miranda IM user. Get it at http://miranda-im.org/.
[03:34:56] Xavier Marjou leaves the room
[03:34:57] Alessandro Amirante leaves the room
[03:34:59] Matthew Kaufman leaves the room
[03:35:06] Neil Stratford leaves the room
[03:35:06] Jonathan Rosenberg leaves the room
[03:35:08] Simon Romano leaves the room
[03:35:23] <jesup> ekr: can you let me know if there's a way to follow clue?
[03:35:29] Colman Ho leaves the room
[03:35:51] Ethan Hugg leaves the room
[03:35:58] <jesup> ekr: just the 101b stream?
[03:36:31] tsuichi joins the room
[03:39:29] Michael Graves leaves the room
[03:39:29] EKRm leaves the room
[03:39:29] Ralph Giles leaves the room
[03:39:29] Jan Skoglund leaves the room
[03:39:29] MeetechoAudio leaves the room
[03:39:29] Randell Jesup leaves the room
[03:39:29] tobcast leaves the room
[03:39:30] Jim McEachern leaves the room
[03:39:59] Jonathan Lennox leaves the room
[03:40:01] Jonathan Lennox joins the room
[03:42:44] Roy leaves the room
[03:42:47] <EKR> I think it's D
[03:42:47] <EKR> isn't it?
[03:43:03] <EKR> I'm hearing background noise in 1010d
[03:43:23] Colman Ho joins the room
[03:43:35] cabo leaves the room
[03:44:22] <Jonathan Lennox> Mary says the CLUE Jabber room will be appropriate for this ad-hoc
[03:44:33] <Jonathan Lennox> And if you just heard Ted talking you're on the right audio stream
[03:44:34] juberti@gmail.com/Meebo leaves the room
[03:45:47] tsuichi leaves the room
[03:47:34] <EKR> OK, I can hear
[03:47:37] marshall joins the room
[03:47:51] <marshall> Hi - CLUE / RTCWEB ad hoc
[03:48:05] <marshall> discussion and jabber scribing should be in clue jabber list
[03:48:16] <EKR> no meetecho? Can someone send slids?
[03:48:30] <Jonathan Lennox> They're on the meeting materials page under CLUE
[03:49:02] victor.pascual.avila joins the room
[03:49:08] pm leaves the room
[03:50:00] Matthew Kaufman joins the room
[03:50:25] <EKR> ack
[03:51:30] lef_jp leaves the room
[03:52:07] burn joins the room
[04:07:08] victor.pascual.avila leaves the room
[04:10:09] hta joins the room
[04:10:21] Hadriel Kaplan joins the room
[04:14:45] <hta> so what's the right chatroom for the adhoc?
[04:15:28] <Jonathan Lennox> CLUE
[04:30:00] Hadriel Kaplan leaves the room
[04:36:30] Colman Ho leaves the room
[04:38:08] <marshall> Come and join the confusion
[04:39:49] Colman Ho joins the room
[04:46:49] Jean Mahoney joins the room
[04:53:06] hta leaves the room
[04:53:24] Jonathan Lennox leaves the room: Computer went to sleep
[04:53:25] burn leaves the room
[04:57:55] marshall leaves the room
[04:58:08] Paul Hoffman joins the room
[04:59:30] rlb joins the room
[05:00:00] Matthew Kaufman leaves the room
[05:00:32] Matthew Kaufman joins the room
[05:02:50] lef_jp joins the room
[05:03:30] rlb leaves the room
[05:07:00] Matthew Kaufman leaves the room
[05:12:44] Paul Hoffman leaves the room
[05:14:15] hta joins the room
[05:14:27] hta leaves the room
[05:17:18] Jean Mahoney leaves the room
[05:29:28] Hadriel Kaplan joins the room
[05:44:28] Colman Ho leaves the room
[06:16:01] lef_jp leaves the room
[06:33:32] Bran, Cary joins the room
[06:34:43] Bran, Cary leaves the room
[06:53:12] rillian leaves the room
[07:02:31] Hadriel Kaplan leaves the room
[07:24:16] Hadriel Kaplan joins the room
[08:20:22] jesup leaves the room
[08:22:38] hta joins the room
[08:55:54] Hadriel Kaplan leaves the room
[09:30:27] burn joins the room
[09:31:03] burn leaves the room
[09:31:19] hta leaves the room
[09:32:15] EKR leaves the room
[10:02:36] hta joins the room
[10:46:21] hta leaves the room
[13:38:33] hta joins the room
[14:03:52] hta leaves the room