Dave, what's the updated presentation?
", "time": "2023-07-28T16:30:17Z"}, {"author": "Lorenzo Miniero", "text": "I see 7 decks on the materials, and 7 imported ones
", "time": "2023-07-28T16:30:27Z"}, {"author": "Dave Plonka", "text": "Hi Lorenzo. This one was uploaded about an hour ago this morning (again):
\nExploring the Cookieverse: A Multi-Perspective Analysis of Web Cookies - Ali Rasaii
Ah then it's already there: we automatically import decks 15 minutes before the session starts, so one hour before was fine :)
", "time": "2023-07-28T16:32:11Z"}, {"author": "Dave Plonka", "text": "Thanks!
", "time": "2023-07-28T16:32:17Z"}, {"author": "Behcet Sarikaya", "text": "Folks, if you looked at the notes page for maprg, https://notes.ietf.org/notes-ietf-117-maprg it says IRTF maprg agenda for IETF-116 (San Francisco)
\n:smile:
The relative increase for wikipedia being higher than instragram is a bit couner intuitive--is it because the Instragram total load time is so high, given the large number of assets?
", "time": "2023-07-28T17:07:49Z"}, {"author": "Anthony Somerset", "text": "i suspect so
", "time": "2023-07-28T17:08:22Z"}, {"author": "Mirja K\u00fchlewind", "text": "fixed the notes and agenda. Thanks!
", "time": "2023-07-28T17:15:32Z"}, {"author": "Mirja K\u00fchlewind", "text": "Ted, you want me to rely that or take it on chat after live Q&A?
", "time": "2023-07-28T17:16:20Z"}, {"author": "Ted Hardie", "text": "It's okay, I have some other questions and comments (e.g. that there were serious concerns about the kind of coalescing they will test), but I will reach out to Jayasree directly.
", "time": "2023-07-28T17:17:19Z"}, {"author": "Andrew Campling", "text": "On resolverless DNS, here is a post by Geoff Huston of APNIC - https://blog.apnic.net/2022/05/17/the-path-to-resolverless-dns/
", "time": "2023-07-28T17:17:45Z"}, {"author": "Kazuho Oku", "text": "It sounded me that Chrome does not use 0-RTT when doing DoH over HTTP/3, is that correct?
", "time": "2023-07-28T17:20:14Z"}, {"author": "Anthony Somerset", "text": "i am struggling with audio clarity
", "time": "2023-07-28T17:25:21Z"}, {"author": "Eduard V", "text": "me too
", "time": "2023-07-28T17:25:30Z"}, {"author": "Anthony Somerset", "text": "its almost like distortion
", "time": "2023-07-28T17:25:49Z"}, {"author": "Jayasree Sengupta", "text": "@Ted Hardie, yes you're correct. because instagram is more resource-full and render time-intensive, hence connection setup has less influence on the total PLT
", "time": "2023-07-28T17:25:55Z"}, {"author": "Anthony Somerset", "text": "thats noise cancellation or auto gain control gone wrong
", "time": "2023-07-28T17:26:37Z"}, {"author": "Eduard V", "text": "yes, too much automation
", "time": "2023-07-28T17:27:01Z"}, {"author": "Andrew Campling", "text": "@Eric K - it would be interesting to have your thoughts at some point on how the designated resolver concept can comply with GDPR etc when the user has set a specific DNS resolver.
", "time": "2023-07-28T17:28:05Z"}, {"author": "Jayasree Sengupta", "text": "@Kazuho Oku, yes correct. By default, it works as DoH + H3 1-RTT
", "time": "2023-07-28T17:28:41Z"}, {"author": "Marten Seemann", "text": "Still can't understand a word
", "time": "2023-07-28T17:28:55Z"}, {"author": "Spencer Dawkins", "text": "Ted Hardie said:
\n\n\nIt's okay, I have some other questions and comments (e.g. that there were serious concerns about the kind of coalescing they will test), but I will reach out to Jayasree directly.
\n
Yes, exactly, and coalescing is something that's being included in https://datatracker.ietf.org/doc/draft-ietf-avtcore-rtp-over-quic/, and doing the same sort of research there would help us do a better job of explaining trade-offs.
\nI also reached out to Jayasree :grinning_face_with_smiling_eyes:
", "time": "2023-07-28T17:29:20Z"}, {"author": "Spencer Dawkins", "text": "My apologies to one and all for drawing a blank at the mike - I had just clicked on another of Jayasree's papers and took several seconds to context-switch! :upside_down:
", "time": "2023-07-28T17:32:29Z"}, {"author": "Geoff Huston", "text": "sadly, this talk is still hard to discern due to rather poor audio quality. I see the auto transcript is also struggling.
", "time": "2023-07-28T17:33:21Z"}, {"author": "Spencer Dawkins", "text": "Geoff Huston said:
\n\n\nsadly, this talk is still hard to discern due to rather poor audio quality. I see the auto transcript is also struggling.
\n
I am doing better with slides than I would be without them. I wonder what Meetecho is seeing - they have better instrumentation than I do.
", "time": "2023-07-28T17:35:16Z"}, {"author": "Lorenzo Miniero", "text": "Not much we can do unfortunately! I see Ali is sharing the screen instead of using the preloaded slide deck: if it's a bandwidth issues, the deck would have helped leave more room for audio
", "time": "2023-07-28T17:36:27Z"}, {"author": "Mirja K\u00fchlewind", "text": "I'm sorry. Not sure there much we can do now...
", "time": "2023-07-28T17:37:56Z"}, {"author": "Spencer Dawkins", "text": "Lorenzo Miniero said:
\n\n\nNot much we can do unfortunately! I see Ali is sharing the screen instead of using the preloaded slide deck: if it's a bandwidth issues, the deck would have helped leave more room for audio
\n
That's still a good observation - I'm using preloaded slides, when I'm chairing anything at IETF (HotRFC), but that's more about us ttrying to move from presenter to presenter with minimal delay.
", "time": "2023-07-28T17:39:36Z"}, {"author": "Behcet Sarikaya", "text": ":+1:
", "time": "2023-07-28T17:41:21Z"}, {"author": "Ted Hardie", "text": "It would be interesting to survey managed browser managers as well.
", "time": "2023-07-28T17:48:18Z"}, {"author": "James Gruessing", "text": "Ted: managed browser managers? Corp managed deployments you mean or something else?
", "time": "2023-07-28T17:49:23Z"}, {"author": "Ted Hardie", "text": "Yes, managed deployments by corporations or schools.
", "time": "2023-07-28T17:50:38Z"}, {"author": "Ted Hardie", "text": "(Part of the interest would be to see if school-provided systems had significantly different defaults from bring your own device configurations.)
", "time": "2023-07-28T17:51:50Z"}, {"author": "James Gruessing", "text": "So measuring sysadmin awareness and knowledge (in conjunction with the internal politics of their organisation) instead of the end user employee/student/etc?
", "time": "2023-07-28T17:53:00Z"}, {"author": "Ted Hardie", "text": "Not instead of, but as well (so that they could be compared).
", "time": "2023-07-28T17:54:06Z"}, {"author": "Andrew Campling", "text": "It\u2019s no surprise that most people have not heard of DNS, don\u2019t know what it does
", "time": "2023-07-28T17:55:05Z"}, {"author": "Spencer Dawkins", "text": "Ted Hardie said:
\n\n\nNot instead of, but as well (so that they could be compared).
\n
+1. That's a different quesion that's also worth answering.
", "time": "2023-07-28T17:56:01Z"}, {"author": "James Gruessing", "text": "Ted: Knowing if the deployments have multi-horizon DNS, regulatory compliance requirements (FINRA/HIPAA/etc) would be hugely important variables too.
", "time": "2023-07-28T17:56:24Z"}, {"author": "Andrew Campling", "text": "Using terminology like \u201csecure\u201d or \u201cprivate\u201d can be problematic. For example, DoH does not protect you from privacy violations by the resolver operator, nor are your queries protected from it. Also, someone in a jurisdiction covered by GDPR looses privacy protections if the resolver operator resides in, for example, the USA.
", "time": "2023-07-28T18:01:14Z"}, {"author": "Brian Dickson", "text": "Also the orthogonal question on whether the other \"security\", DNSSEC, is supported by DNSSEC validation by either the browser or the resolver would be good to investigate.
", "time": "2023-07-28T18:02:37Z"}, {"author": "Spencer Dawkins", "text": "(Spencer quietly checking his browser settings).
\nHuh. my Chrome says \"Custom\", which is true, but doesn't tell me that I don't need to google for public DNS servers, but it does give you meaningful choices if you click on that box. Maybe \"Choose a secure DNS provider\"?
", "time": "2023-07-28T18:04:11Z"}, {"author": "Andrew Campling", "text": "@Spencer, secure by any particular definition? Firefox did something similar to encourage use of DoH but the dialogue box ignored your current resolver choice which could be more secure than the ones being recommended (eg malware filtering, solid privacy policy etc).
", "time": "2023-07-28T18:07:09Z"}, {"author": "Andrew Campling", "text": "For a GDPR-compliant privacy policy for resolver operators see https://europeanresolverpolicy.com
", "time": "2023-07-28T18:08:02Z"}, {"author": "Ted Hardie", "text": "@meetecho the speaker's gain is very high; is it possible to manage it a bit?
", "time": "2023-07-28T18:08:21Z"}, {"author": "Dave Plonka", "text": "Note that w/the technical difficulties we have very little, if any time, for questions and comments through rest of meeting, so please consider commenting here and/or contacting the contributors directly to follow-up on details.
\nOf course we will have the queue open if we think there is time to be fair to these two final contributors.
Ted Hardie said:
\n\n\n@meetecho the speaker's gain is very high; is it possible to manage it a bit?
\n
Heading there
", "time": "2023-07-28T18:09:07Z"}, {"author": "Lorenzo Miniero", "text": "Better now?
", "time": "2023-07-28T18:11:50Z"}, {"author": "Ted Hardie", "text": "A bit, yes. Thank you.
", "time": "2023-07-28T18:12:22Z"}, {"author": "Lorenzo Miniero", "text": "We can lower another bit if you think it's still too loud
", "time": "2023-07-28T18:13:11Z"}, {"author": "Ted Hardie", "text": "It's okay
", "time": "2023-07-28T18:13:47Z"}, {"author": "Spencer Dawkins", "text": "Andrew Campling said:
\n\n\n@Spencer, secure by any particular definition? Firefox did something similar to encourage use of DoH but the dialogue box ignored your current resolver choice which could be more secure than the ones being recommended (eg malware filtering, solid privacy policy etc).
\n
Yeah, that's the problem. We get people trained to choose DNS over HTTP, just in time to start telling them to choose a better mechanism. I doubt seriously that I'm designing the Chrome configuration UI, but there are other UIs (for other settings) that give you four choices, with one saying \"recommended\", and that seems like an improvement.
\nI also note that I see user-level choice of servers for other things, like Ubuntu mirrors, and arrange the choices based on measurements from my perspective. That also might not solve all the problems, but might also be an imprement.
\nThank you for coming to my TED Talk. :smile:
", "time": "2023-07-28T18:18:40Z"}, {"author": "Andrew Campling", "text": "The DNS Research Federation is gathering some pretty interesting information too - see https://dnsrf.org
", "time": "2023-07-28T18:31:48Z"}]