Growing crystals to generate random numbers:
A team at the University of Glasgow has developed a novel way to generate random numbers by using the randomness inherent in crystal growth. In their paper published in the journal Matter, the group describes using chemistry to generate random numbers for use in other applications.
Generating random numbers has always been a tricky problem for computer engineers because computers were designed to be as predictable as possible. But random numbers are required in a wide variety of applications in virtually every scientific field. One of the more pressing applications is data encryption—most existing schemes rely on the constant generation of random numbers. Without randomness, computers designed to crack encryption can soon spot a pattern, making it relatively easy to crack the encryption code. In this new effort, the researchers have turned to a real-world process shown to be more random than pseudo-random number generators—a chemical reaction by which a material begins to crystallize.
[...] The researchers tested their random number generator with a common encryption application that ordinarily uses a conventional generator. They encrypted the word "crystal." They then used a common encryption cracking system to crack the system. They found that the cracking system had more difficulty deciphering their word when it was encrypted by their crystal random generator than with conventional pseudo random number generators.
Physics team uses pixel sensitivity of smartphone as a random generator for encryption
More information:
Edward C. Lee et al. A Crystallization Robot for Generating True Random Numbers Based on Stochastic Chemical Processes, Matter (2020).
DOI: 10.1016/j.matt.2020.01.024
Journal information: Matter
(Score: 4, Insightful) by stormwyrm on Saturday February 22 2020, @02:47AM
I built my own random number generator that uses avalanche noise from a couple of Zener diodes hooked up to a differential amplifier to cancel out interference. I get random bits from it that pass just about every statistical test that I can throw at them at a rate of perhaps 100 kbps, mostly limited by the GPIO interface data rate of the Orange Pi Zero it lives on. Better circuits built on the same principles can get multi-megabits, though they tend to become hard to audit because of that. I skimmed the actual paper, and it seems like a ridiculous Rube Goldberg contraption [literatumonline.com] for doing essentially the same thing as my circuit, using a robot for mixing the chemicals for crystallisation and running a camera to look at the crystal growth patterns. With all that, their system gets something like 25 kbps. The only advantage I can think of is that their system is essentially immune to electromagnetic/RF interference, but with judicious use of shielding and differential amplifiers it is possible to minimise that for electronic circuitry. They say that it might be possible to miniaturise their set up, but given the complexity of the setup I doubt it will be possible to build something like it as cheap or as small as one built entirely out of solid-state electronics.
Numquam ponenda est pluralitas sine necessitate.