Sure. I think Intel's RDRAND, using thermal noise (or whatever) makes more sense, because you don't have to hook up some antenna -- I'm no physicist but my guess is it gets substantively affected by atmospheric noise.
Because pseudo random is good for most purposes, and there is enough chaos in the machine to seed cryptographic random numbers. So why add extra manufacturing cost.
Comments
Having a radio antenna feeding information into the computer has its own negative security implications.
Hold on, what of wifi and Bluetooth?
In a highly secure environment you don't use WiFi or Bluetooth.
The OP is talking about random numbers in general I believe, not in highly secure environments only.
Sure. I think Intel's RDRAND, using thermal noise (or whatever) makes more sense, because you don't have to hook up some antenna -- I'm no physicist but my guess is it gets substantively affected by atmospheric noise.
We're not generating random numbers with them.
"Having a radio antenna feeding information into the computer has its own negative security implications."
My point is, if that is true, then wifi and bluetooth would have those same negative implications regardless of what you are using them for.
Because pseudo random is good for most purposes, and there is enough chaos in the machine to seed cryptographic random numbers. So why add extra manufacturing cost.
Probably because interference (intentional or otherwise) at similar wavelengths could seriously skew the "random" numbers. This obviously isn't good.
Apparently they already do this though
https://en.wikipedia.org/wiki/Random_number_generation#%22Tr...