This article is more than 1 year old

Geneva boffins make light work of random numbers

Take one photon and fake it 'til you make it

“How good is your random number generator?” is a pretty ticklish question in cryptography that a bunch of Swiss quantum bods have set out to answer.

The history of crypto is littered with examples of buggy random number generators, so the group at the University of Geneva have set out to create a self-testing quantum RNG that can report on the quality of its randomness.

Their paper, at Physical Review Letters (abstract here, Arxiv pre-print version here), explains that the work is designed to provide “a continuous estimate of the output entropy” with a minimum of assumptions about the devices, and without a “detailed model” of how they work.

The protocol they've developed provides a way to “certify randomness from a pair of incompatible quantum measurements”, the paper states.

Because the two measurements can't influence each other, the amount of “quantum randomness can be quantified directly from the data, and can be separated from other sources of randomness such as fluctuations due to technical imperfections,” they claim.

Uni Geneva Quantum RNG

Choose your weapon, fire your photon, and get a quantum number: schematic of the Uni of Geneva QRNG

Here's what they do: the user selects how a photon is to be prepared and how it is to be measured; that measurement results in an estimated probability distribution; the “witness value” provides boundaries for the entropy, and the random bit string is extracted from the raw data.

The researchers argue that their approach gets around two problems with current quantum random number generators: device-dependent key generation (used in commercial quantum key distribution kit), which demands that the kit at both ends is tested; and current device-independent protocols, which are very slow.

By separating randomness occurring because of technical noise (unwanted) from the quantum noise (which they want), and that accelerates the device-independent approach.

Not a huge amount faster: while device-dependent schemes generate more than a million random bits per second, the University of Geneva scheme only outputs 23 bits per second.

However, the upside of the approach is that you don't need to know that nobody's trying to tamper with the random number generator, because as the receiver of the data, you can test the randomness of what you're offered.

In the setup described in the paper, the optics that generate and test photons are sensitive to temperature, so by turning off the air conditioner the researchers could observe rising thermal noise, which their implementation of the protocol could compensate for. ®

More about

TIP US OFF

Send us news


Other stories you might like