IETF
rtcweb@jabber.ietf.org
Friday, March 12, 2021< ^ >
martin.thomson has set the subject to: RTCWEB WG http://tools.ietf.org/wg/rtcweb/agenda
Room Configuration
Room Occupants

GMT+0
[06:39:14] zulipbot leaves the room: Disconnected: closed
[06:39:19] zulipbot joins the room
[06:41:15] zulipbot leaves the room: Disconnected: closed
[06:41:27] zulipbot joins the room
[06:48:28] zulipbot leaves the room: Disconnected: closed
[06:48:42] zulipbot joins the room
[06:49:31] zulipbot leaves the room: Disconnected: closed
[06:49:48] zulipbot joins the room
[06:51:03] Glen (AMS IT) joins the room
[06:51:06] Glen (AMS IT) leaves the room
[06:58:48] zulipbot leaves the room: Disconnected: closed
[07:10:52] zulipbot joins the room
[07:11:41] glen joins the room
[07:11:45] glen leaves the room
[07:29:15] zulipbot leaves the room: Disconnected: closed
[07:29:21] zulipbot joins the room
[11:47:39] Meetecho joins the room
[11:50:03] Lorenzo Miniero_web_127 joins the room
[11:50:03] Spencer Dawkins_web_363 joins the room
[11:51:17] Alessandro Amirante_web_378 joins the room
[11:53:43] Spencer Dawkins_web_363 leaves the room
[11:53:58] Spencer Dawkins_web_109 joins the room
[11:54:03] Ted Hardie_web_860 joins the room
[11:54:17] Sean Turner_web_362 joins the room
[11:55:06] Cullen Jennings_web_828 joins the room
[11:56:12] Renan Krishna_web_555 joins the room
[11:56:40] Lorenzo Miniero_web_127 leaves the room
[11:56:44] Lorenzo Miniero_web_865 joins the room
[11:56:45] Alex Huang Feng_web_157 joins the room
[11:57:50] Magnus Westerlund_web_259 joins the room
[11:58:28] Jingrong Xie_web_764 joins the room
[11:58:31] <Cullen Jennings_web_828> why do the slides say make sure your video is off ?
[11:59:09] Jonathan Lennox_web_478 joins the room
[11:59:47] hta joins the room
[11:59:52] Eve Schooler_web_555 joins the room
[12:00:02] Harald Alvestrand_web_243 joins the room
[12:00:07] Jean Mahoney_web_621 joins the room
[12:00:24] Murray Kucherawy_web_501 joins the room
[12:00:25] <Cullen Jennings_web_828> thanks
[12:00:34] Mike English_web_814 joins the room
[12:00:35] Eric Rescorla_web_500 joins the room
[12:00:57] Alan Ford_web_438 joins the room
[12:00:58] Justin Uberti_web_257 joins the room
[12:01:07] Juliana Guerra_web_326 joins the room
[12:01:08] Roman Shpount_web_294 joins the room
[12:01:13] सचिन गर्ग_web_560 joins the room
[12:01:15] Bernie Hoeneisen_web_271 joins the room
[12:01:20] Steve Donovan_web_529 joins the room
[12:01:22] Bernard Aboba_web_542 joins the room
[12:01:43] Spencer Dawkins_web_109 leaves the room
[12:01:49] Spencer Dawkins_web_462 joins the room
[12:01:50] Martin Thomson_web_969 joins the room
[12:01:58] <Eric Rescorla_web_500> Oh fine
[12:02:03] Stan Srednyak_web_494 joins the room
[12:02:14] <Eric Rescorla_web_500> Let me see if I can pull up CODIMD
[12:02:18] Timothy Panton_web_296 joins the room
[12:02:24] Kenneth Murchison_web_553 joins the room
[12:02:31] <Sean Turner_web_362> v
[12:02:33] <Sean Turner_web_362> https://codimd.ietf.org/notes-ietf-110-rtcweb#
[12:02:48] Herman Ramos_web_938 joins the room
[12:02:49] Roland Schott_web_967 joins the room
[12:03:12] <Eric Rescorla_web_500> annoyingly it just has a spinner when it's in-frame
[12:03:18] Bastiaan Wissingh_web_327 joins the room
[12:03:39] James Gruessing_web_638 joins the room
[12:04:16] Suhas Nandakumar_web_854 joins the room
[12:04:27] <Eric Rescorla_web_500> meetecho: yeah, the section you type in just has a spinner
[12:04:49] <Meetecho> Eric: do you mean codimd?
[12:04:51] Eve Schooler_web_555 leaves the room
[12:04:56] Eve Schooler_web_405 joins the room
[12:04:59] Ali Begen_web_143 joins the room
[12:05:02] Geng-Da Tsai_web_935 joins the room
[12:05:09] <Eric Rescorla_web_500> yes.
[12:05:12] Simon Romano_web_634 joins the room
[12:05:15] <Eric Rescorla_web_500> When I opened it in a new window it worked
[12:05:42] Florent Castelli_web_391 joins the room
[12:05:50] ekr@jabber.org joins the room
[12:05:52] <Meetecho> It may have to do with cross-site cookies: it's known not to work if you're running meetecho in a private/incognito window for instance
[12:05:53] Geng-Da Tsai_web_935 leaves the room
[12:06:27] Roland Jesske_web_161 joins the room
[12:06:51] सचिन गर्ग_web_560 leaves the room
[12:07:07] <ekr@jabber.org> Hmm.... Well I do have Firefox tracking protection on, but (1) the UI loads and (2) i don't see any errors in the browser console
[12:07:36] <ekr@jabber.org> And TP doesn't block non-trackers, at least theoretically
[12:08:01] Sergio Garcia Murillo_web_886 joins the room
[12:08:31] Christer Holmberg_web_151 joins the room
[12:09:21] Frode Kileng_web_512 joins the room
[12:09:56] Roland Jesske_web_161 leaves the room
[12:10:00] Roland Jesske_web_511 joins the room
[12:10:28] Sergio Garcia Murillo_web_886 leaves the room
[12:10:39] Sergio Garcia Murillo_web_699 joins the room
[12:10:44] <ekr@jabber.org> So I do think we should consider this a potential bug in the site
[12:10:51] Takio Yamaoka_web_340 joins the room
[12:11:08] <ekr@jabber.org> though perhaps not something that Meetecho needs to concern itself with
[12:11:56] <Meetecho> Ekr: we'll try to debug your specific scenario when the event is over, to see if we can get codimd to work even in those scenarios. Thanks for the heads up!
[12:12:02] <ekr@jabber.org> thx
[12:12:25] <ekr@jabber.org> Happy to do some debugging. Like I said, I don't see anything in the browser console with errors, but maybe i missed it
[12:12:58] Murray Kucherawy_web_501 leaves the room
[12:13:04] Murray Kucherawy_web_290 joins the room
[12:13:09] Bron Gondwana_web_595 joins the room
[12:17:50] Youngkwon Lim_web_652 joins the room
[12:20:43] <Lorenzo Miniero_web_865> Bernard: it can lead to issues actually, e.g., we had some similar issues interacting with gstreamer's webrtcbin because of that. You don't always have a JavaScript app where you can tweak things.
[12:23:50] Chris Wendt_web_591 joins the room
[12:23:53] Meetecho Robot_web_913 joins the room
[12:25:37] Meetecho Robot_web_913 leaves the room
[12:25:41] Meetecho Robot_web_558 joins the room
[12:26:31] Meetecho Robot_web_558 leaves the room
[12:27:18] Herman Ramos_web_938 leaves the room
[12:27:38] Murray Kucherawy_web_290 leaves the room
[12:27:42] Murray Kucherawy_web_759 joins the room
[12:27:46] Eve Schooler_web_405 leaves the room
[12:28:03] Eve Schooler_web_376 joins the room
[12:28:42] Eve Schooler_web_376 leaves the room
[12:31:05] Lorenzo Miniero_web_865 leaves the room
[12:31:09] Lorenzo Miniero_web_325 joins the room
[12:31:26] <Jonathan Lennox_web_478> Does Firefox also exhibit the same behavior as libwebrtc?  (Not sure whether this behavior comes from the part of libwebrtc Firefox uses or not.)
[12:31:48] <ekr@jabber.org> No, itdoes not
[12:32:08] <Jonathan Lennox_web_478> Which behavior does it implement?
[12:32:13] <ekr@jabber.org> The JSEP one, I believe
[12:34:19] <ekr@jabber.org> @Jonathan Lennox: the answer to your more general question is that the only pieces we share with libwebrtc are about the media proper
[12:36:44] <Cullen Jennings_web_828> Thanks tim
[12:39:46] Bron Gondwana_web_595 leaves the room
[12:41:18] <Lorenzo Miniero_web_325> Roman: you need to click the mic icon to talk
[12:46:48] <hta> Datachannel apps are probably 100% written to webrtc, and likely all do bundle.
[12:53:54] yangrulan_web_398 joins the room
[12:54:01] <hta> I like must-bundle :-)
[12:56:45] yangrulan_web_398 leaves the room
[12:58:41] <ekr@jabber.org> must-bundle-safe
[13:05:14] Martin Thomson_web_969 leaves the room
[13:06:53] <ekr@jabber.org> hey chairs, are we done with this topic?
[13:07:26] <Ted Hardie_web_860> @ekr Roman is the last speaker, as far as the queue goes.
[13:08:25] Youngkwon Lim_web_652 leaves the room
[13:11:55] <Jonathan Lennox_web_478> Would we want this to be a full bundle-bis, or a patch document?
[13:15:06] <ekr@jabber.org> LIVING DOCUMENTS!
[13:15:49] Jana Iyengar_web_309 joins the room
[13:16:30] <Timothy Panton_web_296> At the risk of diving into the weeds, can we avoid using max-bundle-safe and use something less negative ?must-bundle  perhaps?
[13:16:56] <Christer Holmberg_web_151> I suggested must-bundle in a private discussion
[13:17:00] <Cullen Jennings_web_828> Huge thanks to Christer for doing the work on this.
[13:17:05] <Sean Turner_web_362> The important point about this is that there is no opposition ... great!
[13:17:18] <Suhas Nandakumar_web_854> +1
[13:17:19] <Sean Turner_web_362> @Christer yes thanks
[13:17:20] <Cullen Jennings_web_828> +1 on must-bundle as label
[13:17:34] <Mike English_web_814> +1 on must-bundle
[13:17:36] Abhijan Bhattacharyya_web_674 joins the room
[13:17:44] <Sean Turner_web_362> max-bundle-damnit
[13:17:45] <Justin Uberti_web_257> mux-bandle
[13:17:58] <Jonathan Lennox_web_478> mock-bindle
[13:18:04] <Mike English_web_814> :satisfied:
[13:18:20] <Suhas Nandakumar_web_854> Bundle-bundle
[13:18:31] <Jonathan Lennox_web_478> bundle-mcbundleface
[13:18:41] <Sean Turner_web_362> ^ FTW
[13:18:46] Abhijan Bhattacharyya_web_674 leaves the room
[13:18:52] <Mike English_web_814> bundle-bundle-bundle-bundle-mush...
[13:19:30] Behcet Sarikaya_web_918 joins the room
[13:19:40] <ekr@jabber.org> I  just said that I think some combination of Cullen, Justin, and myself can do the JSEP update
[13:19:59] <Cullen Jennings_web_828> Glad to help on JSEP.
[13:20:24] <Sean Turner_web_362> Many thanks Justin for presenting this so clearly
[13:20:25] <Jonathan Lennox_web_478> I suspect having the existing editors do the revs of each document is gonna make the editing process a lot easier.
[13:20:52] <Justin Uberti_web_257> agreed
[13:21:05] <Justin Uberti_web_257> and hopefully avoiding us being thawed out in 2050
[13:22:19] <Magnus Westerlund_web_259> Do a full update!
[13:22:37] <Spencer Dawkins_web_462> @Justin - that argues strongly for cremation if you author key IETF specifications ... :grinning:
[13:24:31] <Magnus Westerlund_web_259> I would note that doing a write up for the IETF last call with it being an update is not that problematic. The main issue that exist is that the review teams don't see the write up in the IETF last call.
[13:24:38] <Sean Turner_web_362> @Christer - sorry 'bout that
[13:25:19] <Cullen Jennings_web_828> We could note on front of draft that could be removed by RFC Editor which explained what was going on
[13:29:09] James Gruessing_web_638 leaves the room
[13:30:21] Suhas Nandakumar_web_854 leaves the room
[13:30:44] <Justin Uberti_web_257> doesn't insertable streams affect this?
[13:31:06] <Justin Uberti_web_257> (wrt packet content, perhaps?)
[13:31:09] BEHCET SARIKAYA joins the room
[13:31:40] <ekr@jabber.org> Yeah, probably mkes it worse
[13:31:58] <Sean Turner_web_362> Note Justin in queue
[13:33:23] <Magnus Westerlund_web_259> But can you run SRTP or DTLS in non-encrypted mode? Otherwise you have to counter the encryption to cause a plain text message that matches your intended message. So for WebRTC does this problem exist outside of TURN?
[13:33:27] <ekr@jabber.org> I am +1 on the "short RFC" version
[13:33:48] <ekr@jabber.org> Magnus: yes, because the SRTP encryption is predictable
[13:35:07] Cheng Li_web_310 joins the room
[13:35:08] <Jonathan Lennox_web_478> Aren't the fragmented UDP packets marked as fragmented at the IP layer?  NAT ALGs are ignoring IP fragmentation?
[13:37:13] <hta> Yes, they are :-(
[13:37:16] <Justin Uberti_web_257> @Magnus I think this could happen with ICE usernames too
[13:38:29] <Roman Shpount_web_294> This looks like NAT Router security issue not RTCWeb issue
[13:40:07] Chris Wendt_web_591 leaves the room
[13:40:18] <Justin Uberti_web_257> the safe port list probably solves most of these issues. But can this list be applied to ICE candidates?
[13:40:40] <Jonathan Lennox_web_478> That might be an interesting thing to do telemetry on.
[13:40:51] Geng-Da Tsai_web_323 joins the room
[13:41:02] <hta> I think we're currently applying the blocked port list to ICE candidates, but I have to test to make sure.
[13:44:07] <Sean Turner_web_362> I uploaded the slides
[13:44:12] <Jonathan Lennox_web_478> Only 443
[13:44:17] <hta> and 80
[13:44:24] Geng-Da Tsai_web_323 leaves the room
[13:44:25] <Ted Hardie_web_860> Roman, we cut the cue.
[13:44:25] <Alessandro Amirante_web_378> what about UDP 53?
[13:44:37] <Cullen Jennings_web_828> Just for FYI, we use port 53 in some cases
[13:45:05] <hta> Cullen, Cisco filed a bug on port 80 TURN servers. Can you file one on port 53?
[13:45:08] <Lorenzo Miniero_web_325> Yeah, 80, 443 and 53 are quite common for TURN
[13:45:38] <Alessandro Amirante_web_378> same here. 53 is kinda often let through firewalls
[13:45:44] <Justin Uberti_web_257> libwebrtc tries to avoid ever fragmenting a packet
[13:46:01] <Jonathan Lennox_web_478> To the previous point about SRTP, once you know the keys and can control the plaintext, you can absolutely generate the ciphertext of your choice, as long as you can structure things such that the auth tag doesn't interfere with your attack.
[13:46:04] <Cullen Jennings_web_828> @hta - glad to work with you on whatver needed here
[13:46:16] <Jonathan Lennox_web_478> Insertable streams, or maybe G.711 too.
[13:46:56] Alan Ford_web_438 leaves the room
[13:47:10] <Justin Uberti_web_257> right. so ICE-TCP + mss tweaking + insertable streams would allow you to make any packet you want
[13:47:10] Behcet Sarikaya_web_918 leaves the room
[13:47:10] <ekr@jabber.org> Thanks Justin for taking the lead on this
[13:47:13] Eric Rescorla_web_500 leaves the room
[13:47:14] Sean Turner_web_362 leaves the room
[13:47:14] <Cullen Jennings_web_828> These are best chairs ever - clearly getting stuff done fast
[13:47:14] Juliana Guerra_web_326 leaves the room
[13:47:15] Roman Shpount_web_294 leaves the room
[13:47:16] Roland Schott_web_967 leaves the room
[13:47:17] Timothy Panton_web_296 leaves the room
[13:47:20] Steve Donovan_web_529 leaves the room
[13:47:20] Takio Yamaoka_web_340 leaves the room
[13:47:21] Alex Huang Feng_web_157 leaves the room
[13:47:23] Jean Mahoney_web_621 leaves the room
[13:47:24] Bastiaan Wissingh_web_327 leaves the room
[13:47:25] Harald Alvestrand_web_243 leaves the room
[13:47:25] BEHCET SARIKAYA leaves the room
[13:47:27] <Justin Uberti_web_257> adios all
[13:47:34] Murray Kucherawy_web_759 leaves the room
[13:47:35] Justin Uberti_web_257 leaves the room
[13:47:35] Christer Holmberg_web_151 leaves the room
[13:47:36] Mike English_web_814 leaves the room
[13:47:51] Lorenzo Miniero_web_325 leaves the room
[13:47:59] Renan Krishna_web_555 leaves the room
[13:48:02] Florent Castelli_web_391 leaves the room
[13:48:03] Magnus Westerlund_web_259 leaves the room
[13:48:08] Alessandro Amirante_web_378 leaves the room
[13:48:08] Ted Hardie_web_860 leaves the room
[13:48:08] Cullen Jennings_web_828 leaves the room
[13:48:08] Jingrong Xie_web_764 leaves the room
[13:48:08] Bernie Hoeneisen_web_271 leaves the room
[13:48:08] Bernard Aboba_web_542 leaves the room
[13:48:08] Spencer Dawkins_web_462 leaves the room
[13:48:08] Jonathan Lennox_web_478 leaves the room
[13:48:08] Stan Srednyak_web_494 leaves the room
[13:48:08] Kenneth Murchison_web_553 leaves the room
[13:48:08] Ali Begen_web_143 leaves the room
[13:48:08] Simon Romano_web_634 leaves the room
[13:48:08] Frode Kileng_web_512 leaves the room
[13:48:08] Roland Jesske_web_511 leaves the room
[13:48:08] Sergio Garcia Murillo_web_699 leaves the room
[13:48:08] Jana Iyengar_web_309 leaves the room
[13:48:08] Cheng Li_web_310 leaves the room
[13:48:11] <hta> the attacker would have to extract the symmetric key at the attacker end, so cannot do this during the setup phase
[13:50:02] Meetecho leaves the room
[14:09:55] ekr@jabber.org leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!