Hacker News new | past | comments | ask | show | jobs | submit login
Quantum Random Number Generator as free webservice (hu-berlin.de)
38 points by fs111 on Nov 27, 2010 | hide | past | favorite | 9 comments

John Walker's Fourmilab has been providing quantum random bits for 14 years already: http://www.fourmilab.ch/hotbits/

I can't really see a use for this. For non-cryptographic purposes such as Monte Carlo simulations, it's hard to imagine an application where a good conventional pseudo-random generator such as a Mersenne twister wouldn't be perfectly adequate. Even for cryptographic purposes, a crypto-standard PRNG seeded from a good entropy source (/dev/random) is usually secure enough, and for the few seriously high-security applications where that wouldn't be enough, nobody would trust a source that wasn't under their physical control.

There are two (plus one) market niches that exist, which another random number service (random.org) fills fairly well.

1.) People who don't know anything at all about random number generation, and just want a web form where they can press the button and get a random number.

You hardly need a hardware random number generator for this, or any kind of serious randomness.

2.) Acting as a neutral, trusted third party, for contests, etc. Dooce just did a big Microsoft-sponsored Kinect giveaway PR circus bullshit event, and they used random.org to pick the winners.

Unlike 1, there might be actual money in this, and is one of the few, amusing, businesses where the more money you charge, the higher the perceived utility.

3.) Providing random numbers as an advertisement for your fine line of hardware random number generators. Here it doesn't matter how much money you make, you just want people to buy the hardware that made them. Oddly enough, none of the random number services (and there are quite a few) do this, for some inexplicable reason. There's not even an argument-from-proprietary technology, since HRNGs are supposed to generate perfectly random noise, and there's no way an attacker could stage a replay attack.

I'm currently building a high-quality zener diode-based RNG. There's a dearth of good information on the internet about these things, everyone seems to offer a schematic but their explanations don't really show any sort of understanding of the process.

In any case, I think this design qualifies as "quantum" because zener diodes at 5.6V and less work through quantum electron tunneling ;)

This isn't useful for cryptography (how can I trust the random data to be random?), but for certain computational tasks this could indeed be quite useful (think Monte Carlo simulations).

I've never encountered a Monte Carlo algorithm that required random numbers of this quality. That doesn't mean they don't exist though.

This is really cool.

But is there anything it’s useful for as it stands? Or is it better thought of as an experiment in distribution?

Assuming the many-worlds interpretation of quantum mechanics is true, it's useful for making decisions.

Make your decision based on quantum-generated random numbers, and you're guaranteed that an alternative-universe counterpart of yourself made the opposite decision. Thus if your decision turns out to be the wrong one you can at least console yourself that there's another version of yourself who didn't make the same mistake.

I don't see any use for cryptographic applications, except perhaps as a common reference string.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact