tip: web requests should not be measured in Hz [hertz] as that is only used for periodic frequencies, which random events (like requests hitting a web server) are not!

[damonwakes@mastodon.sdf.org]

@ben @moof @4censord @sophie Simpsons did it.

[drmikepj@mastodon.social]

@tsukkitsune @ben @moof @4censord @sophie

It's not currently nuclear power stations, but it is true that the Urenco nuclear fuel processing plant at Capenhurst in Cheshire has a regular rhythmic sound that's always sounding to indicate that everything is okay. https://www.thetimes.com/sunday-times-100-tech/hardware-profile/article/uranium-firm-finds-path-to-enrichment-jl7bbp88xw3 - archived here https://archive.ph/2nK2H

[pozorvlak@mathstodon.xyz]

@sekomi @sophie I can't find a link for this, but I'm sure I've seen John Graham-Cumming quoting Cloudflare's log rate in megahertz. And I'd argue that at those frequencies the distinction between Hz and Bq is academic!

[wlukewindsor@c.im]

@philcowans @moof @4censord @sophie https://dl.acm.org/doi/abs/10.1145/142750.142754 or https://scholar.google.com/scholar?oi=bibs&hl=en&cluster=9406488945496633684#d=gs_qabs&t=1777293873015&u=%23p%3DVDH3s9uPioIJ

[johnefrancis@cosocial.ca]

@ben @moof @4censord @sophie core melted into a problematic noncritical puddle

[stylus@social.afront.org]

@lain_7 @moof @4censord @sophie

Using s=0.59c as the signal propagation rate of copper, Δt=10°C, and 17.6ppm/°C as the coefficient of expansion of copper, a 1km transmission line would see an additional 1ns delay when heated. (100ns for 100km and so on)

So clearly with plenty of wire and a dedicated measurement setup you could measure it.

(I was unable to find whether there's a temperature dependence to signal propagation rate in copper conductor nor do I have a good intuition as to why there should or shouldn't be. however this paper reviews two additional temperature-dependent factors:
https://www.haystack.mit.edu/wp-content/uploads/2020/07/memo_mark-5_067.pdf
such as phase delay due to changes in impedance)

Measuring it incidentally using network traffic seems harder. On less than 10m of cable (& 1 intervening network switch) my ping time standard deviation is 31us on 1000 packets. You'd need many measurements (and a very stable local timebase, not just a standard crystal oscillator) to find a nanosecond-level variation in transmission time through statistics on ping times. would love to hear more about whether & how this has actually been measured!