IETF
EDM
edm@jabber.ietf.org
Thursday, July 28, 2022< ^ >
Room Configuration
Room Occupants

GMT+0
[12:56:27] <zulipbot> (Dhruv Dhody) are slides sync to edm track?
[12:57:12] <zulipbot> (Dhruv Dhody) 😁
[12:57:25] <zulipbot> (Tommy Pauly) Hehe
[12:57:25] <zulipbot> (Tommy Pauly) I'm open to trying
[12:58:10] <zulipbot> (Dhruv Dhody) Hehe! we could soon be heading towards -- could this presentation be a ticktok
[12:59:15] Yoshiro Yoneya joins the room
[12:59:40] <zulipbot> (Dhruv Dhody) Audio is good
[12:59:53] <zulipbot> (Tommy Pauly) Great
[13:00:06] <zulipbot> (Tommy Pauly) Thanks!
[13:00:06] <zulipbot> (Gorry Fairhurst) I hear you from scotland :-)
[13:01:35] <zulipbot> (Brian Trammell) ntz ntz ntz ntz
[13:03:50] <zulipbot> (Martin Thomson) I can't believe that this has been running for 2 years.
[13:04:19] <zulipbot> (Martin Thomson) RFC 6709, for instance
[13:22:34] <zulipbot> (Antoine Fressancourt) Error on intentionally malformed messages rather than silent dropping is a highway to DDoS
[13:23:16] <zulipbot> (Antoine Fressancourt) There is a need to put the burden on the sender of malformed message rather than on the receiver
[13:25:47] <zulipbot> (Martin Thomson) @**Brian Trammell** is being mixed by Daft Punk
[13:26:00] <zulipbot> (Martin Thomson) in real time
[13:26:36] <zulipbot> (Dave Thaler) Brian's audio is choppy :(
[13:27:27] <zulipbot> (Alessandro Amirante) Bad audio is not Brian's fault, there's a networking issue at the venue the NOC is sorting out
[13:27:49] <zulipbot> (Dave Thaler) @**Antoine Fressancourt** agree, I put myself back in queue to say that (and related points to distinguish when each is right)
[13:28:56] <zulipbot> (Alessandro Amirante) it should be fixed now.
[13:29:58] <zulipbot> (Brian Trammell) choppy audio: just trying to keep the edm theme. audio was good after I dropped video though, yes?
[13:30:38] <zulipbot> (Tommy Pauly) Correct!
[13:30:51] <zulipbot> (Antoine Fressancourt) Accepting options you don’t understand in case you are not the intended receiver is key to ipv6 extensibility, and if you are the receiver then drop silently
[13:33:52] <zulipbot> (Antoine Fressancourt) @**Dave Thaler** agree with your connection / connectionless distinction
[13:36:07] <zulipbot> (Antoine Fressancourt) Agree on the fact that protocol development lifecycle has not changed that much when chips or ASICS are involved
[13:37:25] dkg joins the room
[13:37:36] <zulipbot> (Alessandro Amirante) @_**Brian Trammell|781** [said](https://zulip.ietf.org/#narrow/stream/329-edm/topic/jabber/near/28740):
```quote
choppy audio: just trying to keep the edm theme. audio was good after I dropped video though, yes?
```
yes but it was a coincidence. the NOC worked around a bad route at the same time you stopped your video, and so things improved
[13:37:57] <zulipbot> (Deborah Brungard) In case my audio is bad, +++1 Wes
[13:38:49] <zulipbot> (Uri Blumenthal) Audio quality is horrible, bordering on unintelligible. Audio stream bandwidth indicated as 9-10kbps, not a whole lot, compared to 279kbps video.
[13:39:38] <zulipbot> (Uri Blumenthal) Stopping video feed did not improve audio.
[13:39:57] <zulipbot> (Deborah Brungard) Apple may be able to survive crashes, I’d say that’s a corner case, not most
[13:41:27] <zulipbot> (Tommy Pauly) @Deborah, what do you mean? I don't follow. When was it saying that something should crash?
[13:42:14] <zulipbot> (Antoine Fressancourt) Having the whole protocol ecosystem agree is ok when you have 2 or 3 implementations, not in a completely fragmented space
[13:43:54] <zulipbot> (Uri Blumenthal) Video dropped to 127kbps. Audio at 7-8kbps. Quality - worse than in 1990-ties and 9600bps modems. Not very flattering for IETF.
[13:44:35] <zulipbot> (Uri Blumenthal) Following discussion is impossible now - audio too choppy.
[13:45:25] <zulipbot> (Deborah Brungard) @Tommy, Brian and others were saying protocol deployment has changed, that today do earlier and crashes are ok. Maybe for an Apple deployment, not for in general.
[13:45:38] <zulipbot> (Martin Thomson) I don't know where this idea that this was intended as a bludgeon for "bad protocols" or "bad deployments" came from, but that was never intended and I can't find any way to interpret the text that way.
[13:46:17] <zulipbot> (Brian Trammell) (out of the queue because my comment reduced to "I agree with MT")
[13:46:43] <zulipbot> (Martin Thomson) I realize that this idea that updates are necessary is something that I will stand by.
[13:46:56] <zulipbot> (Brian Trammell) I think the doc should be explicit about that
[13:47:09] <zulipbot> (Tommy Pauly) @Deborah I'm not sure why you'd call out Apple here? But I think that was just saying when you are bringing up a new implementation test in a hackathon, your test code can crash so you see your bugs.
[13:47:09] <zulipbot> (Brian Trammell) "If you're on the Internet and you don't have an immune system, you're going to have a bad time" is... true
[13:47:22] <zulipbot> (Tommy Pauly) @Brian, agreed
[13:47:35] <zulipbot> (Martin Thomson) @**Brian Trammell** yeah, that's something that I think is almost a distraction from this document.
[13:47:48] <zulipbot> (Brian Trammell) doc should probably state its assumptions about the properties of the lifecycles it applies to
[13:48:24] <zulipbot> (Martin Thomson) It's true, but also this document only needs to acknowledge the range of ways in which protocol actors might change (or not)
[13:48:25] <zulipbot> (Dave Thaler) Even logging errors in some cases can be used for an attack (against disk space, or bandwidth on the backend network if logs are network messages), so must be a consideration
[13:48:37] <zulipbot> (Brian Trammell) and there probably needs to be another doc that explores the robustness of protocols where you can't do that.
[13:48:40] <dkg> deborah?  you're on queue
[13:49:33] <zulipbot> (Brian Trammell) +1 dave
[13:52:35] <zulipbot> (Brian Trammell) this discussion has a lot of people making assumptions that their operational envelopes are the same as everyone else's...
[13:53:25] <zulipbot> (Colin Perkins) It should have a dissent, like legal opinions
[13:53:38] <zulipbot> (Brian Trammell) (for example, I don't operate in a world where bugfix releases and upgrade releases are different, and for the most point deployments are version-heterogeneous. But I get that a lot of organizations have a different process.)
[13:54:16] <zulipbot> (Deborah Brungard) @Tommy, totally agree, demo code. Production ready is very different.
[13:54:51] <zulipbot> (Brian Trammell) +1 mirja, short is good
[13:56:00] <dkg> MT: updates might be both necessary and hard (maybe impossible)
[13:56:23] <zulipbot> (Tommy Pauly) Indeed
[13:57:02] <zulipbot> (Brian Trammell) great to see y'all, have a good thursday!
[14:00:25] Yoshiro Yoneya leaves the room
[16:23:39] dkg leaves the room