IETF
taps
taps@jabber.ietf.org
Wednesday, November 7, 2018< ^ >
Kyle Rose has set the subject to: TAPS @ IETF-101 London
Room Configuration
Room Occupants

GMT+0
[08:24:13] Meetecho joins the room
[08:27:19] hsFGg1P8 joins the room
[08:28:24] Aaron Falk joins the room
[08:28:38] Aaron Falk has set the subject to: TAPS @ IETF-103 Bangkok
[08:35:09] Brian Trammell joins the room
[08:35:13] Simone Ferlin joins the room
[08:35:15] Philipp Tiesel joins the room
[08:37:45] <Brian Trammell> 'morning
[08:37:47] <Brian Trammell> mics are hot btw
[08:38:17] <Brian Trammell> (sorry to hear about your ear, aaron :) )
[08:38:17] <Philipp Tiesel> Morning
[08:39:45] <Aaron Falk> Hi Brian!
[08:39:54] <Aaron Falk> At least I can still read
[08:40:04] <Aaron Falk> gotta baby yet?
[08:40:11] Jonathan Lennox joins the room
[08:40:21] <Aaron Falk> I assume not
[08:41:21] ted.h joins the room
[08:43:23] <Brian Trammell> nope, he does not appear to be in any hurry at all
[08:43:43] <Philipp Tiesel> Remote
[08:43:47] Kyle Rose joins the room
[08:44:32] <Kyle Rose> I am Jabber scribe. Please prefix your comment with "mic:" if you want me to repeat it to the room.
[08:49:14] <Philipp Tiesel> @gorry: +1
[08:49:41] <Jonathan Lennox> Brian, you're overdriving your mic
[08:49:46] <Kyle Rose> meetecho: we're getting some static from our remote
[08:50:19] taps joins the room
[08:50:20] <Kyle Rose> Yeah, sounds like clipping, but it's unclear if it's at the source or not
[08:50:31] <Meetecho> Let me check
[08:50:48] Simon Pietro Romano joins the room
[08:51:35] <Brian Trammell> that might actually be my voice, which is clipping due to a cold
[08:51:58] <Kyle Rose> if that were caused by your voice, I'd recommend consulting a doctor or a mortician
[08:54:09] <Kyle Rose> meetecho: any ideas?
[08:54:54] <Meetecho> He's clipping a bit in Meetecho as well, but definitely audible
[08:55:01] <Meetecho> Not sure how bad that is in the room
[08:55:07] <Jonathan Lennox> It's pretty bad in the room
[08:55:10] <Meetecho> We've lowered the remotes volume in room, please let us know if that helps
[08:55:13] <Jonathan Lennox> I mean, understandable, but annoying
[08:55:55] <Kyle Rose> The undesirable noise is quite loud, but the remote voice isn't. So I'm afraid lowering the volume will just make the remote too soft without addressing the problem.
[08:56:32] jules joins the room
[08:57:09] <Kyle Rose> 2. WGLC kickoff for draft-ietf-taps-transport-security - Chris (5m)
[08:57:29] <Brian Trammell> no idea how to set mic gain in meetecho / firefox
[08:57:45] <Brian Trammell> if someone can give me a pointer i can turn it down in the source
[08:57:59] <Meetecho> Your gain seems fine, but anyway the browser does automatic gain control
[08:58:05] <Brian Trammell> otherwise i can try a headset
[08:58:13] <Brian Trammell> ok
[08:58:17] <Kyle Rose> Can't hurt to try. (The audacity of hope.)
[08:58:23] <Meetecho> If you have a different headset you can try, yep
[08:58:30] <Jonathan Lennox> Brian: you may have OS-level gain controls
[08:58:41] <Meetecho> You may have to rejoin, not sure if the new mic would be taken right away
[09:00:08] Emile Stephan joins the room
[09:01:07] Michael Scharf joins the room
[09:01:35] <Brian Trammell> ok, turned down the gain at the system level
[09:03:31] <Kyle Rose> 3. API Mappings (30m) - Tommy
[09:05:01] taps leaves the room
[09:05:21] Michael Tüxen joins the room
[09:09:05] <Brian Trammell> ...and all of the models for idempotent data are wrong...
[09:09:12] tom@jabber.tdr.uni-due.de joins the room
[09:09:38] <Kyle Rose> …and "idempotent" isn't a sufficient criteria, because delete is idempotent but not safe for 0-RTT
[09:13:49] tom@jabber.tdr.uni-due.de leaves the room
[09:13:57] tom@jabber.tdr.uni-due.de joins the room
[09:14:39] tom@jabber.tdr.uni-due.de leaves the room
[09:14:44] tj joins the room
[09:14:58] Michael Tüxen leaves the room
[09:15:13] <Brian Trammell> so that's an interesting point, Kyle -- we say "idempotent" everywhere but we should really call it "0rtt-safe"
[09:15:22] <Brian Trammell> yaaaaaaay terminology
[09:15:49] <Kyle Rose> This has been one of my pet peeves for years. I could have sworn I said it at the mic at some point. I know I've made the same point in architecture discussions at Akamai for years.
[09:16:11] <Brian Trammell> by delete I presume you mean HTTP DELETE?
[09:16:20] <Brian Trammell> (wait, why is that not 0rtt-safe?)
[09:16:28] <Kyle Rose> Well, REST delete in general.
[09:16:34] <Kyle Rose> DELETE /foo
[09:16:37] <Kyle Rose> PUT /foo
[09:16:39] <Kyle Rose> DELETE /foo
[09:17:04] <Brian Trammell> ahhhhh
[09:17:07] <Kyle Rose> i.e., replay the DELETE after the PUT
[09:17:31] <Kyle Rose> It's like PUTting an empty object
[09:17:42] <Brian Trammell> okay got it
[09:18:02] <Brian Trammell> 0RTT isn't just repeatable, it breaks ordering
[09:18:10] <Kyle Rose> right
[09:18:25] terrebyte joins the room
[09:18:26] <ted.h> @Kyle you can make DELETE safe by using e-tags, though, can't you?  If the DELETE doesn't apply to the right e-tag, it fails?
[09:18:28] <Kyle Rose> 0-RTT is really only safe for read-only requests.
[09:18:44] <Brian Trammell> so you could do this with the interface now by having two message property conditions need to be met before 0rtt is okay:
[09:18:52] <Kyle Rose> Yes, you can use other mechanisms to make it safe. I'm just saying that by itself it isn't safe.
[09:18:54] <Brian Trammell> !ordered + idempotent
[09:18:57] <Kyle Rose> You can make POST safe the same way
[09:19:14] <Brian Trammell> and e-tag at the application layer allows the application to set !ordered
[09:19:16] <ted.h> @Kyle  Right.  The WebDav folks spent a lot of time on all of that, if I remember right.
[09:19:50] <Kyle Rose> To be clear, my pet peeve is about the use of the term "idempotent" where that isn't a sufficient constraint
[09:20:11] <Aaron Falk> MEETECHO: remote audio is very faint now
[09:20:25] <Meetecho> Better now?
[09:20:35] <Aaron Falk> dunno.he stopped talking.
[09:20:45] <Kyle Rose> $ dig dunno.he
[09:21:11] <Philipp Tiesel> shoud I turn up my hardware gain?
[09:21:35] <Kyle Rose> No, I think Meetecho turned down the remote volume because of the problems we were having with Brian's audio
[09:21:39] tj leaves the room
[09:23:09] Sukumal Kitisin joins the room
[09:23:13] <Meetecho> Ack, please ping me again when he gets the floor again so that we can monitor
[09:23:20] <Kyle Rose> will do.
[09:31:01] Sukumal Kitisin leaves the room
[09:31:59] tj joins the room
[09:40:29] <Kyle Rose> 4. draft-ietf-taps-{arch,impl,interface} topics (20m)  -- Michael
[09:40:40] <Kyle Rose> Only 45 minutes of agenda to fit into 30 minutes. Race to the finish!
[09:40:56] tj leaves the room
[09:41:04] tj joins the room
[09:43:48] <Kyle Rose> meetecho: Brian's on
[09:43:50] <Meetecho> Is Brian's audio fine?
[09:43:53] <Kyle Rose> He sounds fine.
[09:43:59] <Aaron Falk> audio is good
[09:44:02] <Meetecho> (y)
[09:44:05] <Kyle Rose> I hear none of the clipping I heard before
[09:44:12] <Kyle Rose> And the volume is good
[09:44:29] <Meetecho> If Brian changed the headset, that's what may have fixed it then
[09:44:38] <Kyle Rose> No headset I can hear
[09:44:45] <Kyle Rose> He just said he turned his gain down
[09:44:56] <Meetecho> Ah got it
[09:44:56] <Kyle Rose> s/hear/see/
[09:45:11] <Brian Trammell> +1 mirja
[09:48:00] <Brian Trammell> i like cost prefs, but i want to reconcile it with PvD/interface selection
[09:48:10] <Brian Trammell> and i don't see how that works
[09:48:15] <Kyle Rose> How is a send/receive rate that is non-numeric actionable at this layer? Conversational over AC3 is different from conversational over Opus
[09:48:58] <Kyle Rose> meetecho: Philipp is coming on now
[09:49:26] <Kyle Rose> Sounds mostly fine. A little low, but understandable
[09:51:07] <Kyle Rose> Time check: 19 minutes remain
[09:54:10] <Philipp Tiesel> mic: going under minimum can raise a soft error
[09:55:09] <Aaron Falk> @phillip: can we do your talk on registries in Prague?
[09:56:27] ted.h leaves the room
[09:56:44] <Kyle Rose> Philipp: is that still relevant?
[09:56:59] <Philipp Tiesel> @aaron: unsure - but can do it in 3-5min
[09:57:22] <Brian Trammell> interim
[09:57:31] <Philipp Tiesel> better
[09:57:42] <Brian Trammell> (as in, let's try to schedule one soon)
[09:57:50] <Kyle Rose> Yeah, I agree: let's get an interim in the next two weeks
[09:58:00] <Kyle Rose> We wanted a 2 hour slot, and only got 1:30.
[09:59:00] <Aaron Falk> but we get friday to get massages
[10:01:33] <Brian Trammell> +1 mirja -- all the timeouts seem protocol specific at this point
[10:02:11] <Philipp Tiesel> but we may want to have an abstract "map to the best timeout available" knob
[10:03:55] Michael Tüxen joins the room
[10:03:56] <Philipp Tiesel> but that can also go one level up
[10:04:21] <Brian Trammell> philipp: i'd want to do the homework to make sure that there is an obvious deterministic way to interpret it
[10:04:51] <Brian Trammell> it is not clear to me that there is, i'd want to have a look at the tcp option in detail first
[10:05:04] <Kyle Rose> Time check: 5 minutes remain
[10:05:54] <Philipp Tiesel> brian: makes sense – so neither remove below nor move up
[10:07:44] <Brian Trammell> hand
[10:07:54] <Philipp Tiesel> hand
[10:08:09] <Brian Trammell> (depends on the new kid of course, will try to make it)
[10:08:15] jules leaves the room
[10:08:29] <Kyle Rose> Finished early! /s
[10:08:29] Jonathan Lennox leaves the room
[10:08:31] Michael Tüxen leaves the room
[10:08:32] tj leaves the room
[10:08:35] Kyle Rose leaves the room
[10:08:36] Jonathan Lennox joins the room
[10:08:39] Aaron Falk leaves the room
[10:08:45] Simon Pietro Romano leaves the room
[10:08:54] Michael Tüxen joins the room
[10:08:56] <Philipp Tiesel> brian: mine is in the next room – worked so far ;-)
[10:09:05] Simone Ferlin leaves the room
[10:09:05] Brian Trammell leaves the room
[10:09:05] Michael Tüxen leaves the room
[10:09:05] Philipp Tiesel leaves the room
[10:09:05] Michael Scharf leaves the room
[10:09:06] Emile Stephan leaves the room
[10:11:07] Meetecho leaves the room
[10:14:00] ted.h joins the room
[10:16:52] Aaron Falk joins the room
[10:17:41] ted.h leaves the room
[10:18:43] Aaron Falk leaves the room
[10:21:16] Aaron Falk joins the room
[10:22:23] Jonathan Lennox leaves the room
[10:23:21] Aaron Falk leaves the room
[10:30:36] ted.h joins the room
[10:30:59] Jonathan Lennox joins the room
[10:32:30] ted.h leaves the room
[10:32:34] tj joins the room
[10:33:02] tj leaves the room: Replaced by new connection
[10:33:03] tj joins the room
[10:33:13] Jonathan Lennox leaves the room
[10:35:35] jules joins the room
[10:39:49] tj leaves the room
[10:47:51] taps joins the room
[10:47:55] taps leaves the room
[10:48:10] taps joins the room
[10:48:15] taps leaves the room
[10:48:20] taps joins the room
[10:48:25] taps leaves the room
[10:49:02] taps joins the room
[10:49:06] taps leaves the room
[10:49:17] taps joins the room
[10:49:26] taps leaves the room
[10:49:39] taps joins the room
[10:49:49] taps leaves the room
[10:50:18] taps joins the room
[10:50:27] taps leaves the room
[10:50:33] taps joins the room
[10:50:42] taps leaves the room
[10:50:52] taps joins the room
[10:51:01] taps leaves the room
[10:51:07] taps joins the room
[10:51:16] taps leaves the room
[12:37:05] jules leaves the room
[12:38:42] jules joins the room
[13:11:41] terrebyte leaves the room
[13:11:43] terrebyte joins the room
[13:18:52] jules leaves the room
[13:46:05] terrebyte leaves the room
[13:46:07] terrebyte joins the room
[13:56:24] terrebyte leaves the room
[14:07:21] terrebyte joins the room
[14:16:29] terrebyte leaves the room
[14:16:33] terrebyte joins the room
[14:25:56] terrebyte leaves the room
[14:26:03] terrebyte joins the room
[14:51:34] jules joins the room
[17:01:54] jules leaves the room
[20:52:58] terrebyte leaves the room
[20:53:01] terrebyte joins the room
[22:48:44] jules joins the room
[22:54:41] jules leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!