Content Mutual

How do you know if something is really random? When testing Random Number Generators (RNGs) for randomness you quickly run into the “is that really random” dilemma. As pattern seeking creatures, we naturally look for patterns on an ongoing basis. Like many things, randomness varies. Most of the time, pseudo-random numbers generators (PRNGs) are basically creating randomness with lots of non-random processes.

“Computer Composition With Lines” 1964 This work closely mimics the painting “Composition With Lines” by Piet Mondrian[7]

How do you really know if something is random? Is “111111” random? Or “0100100101001011110111010”? It's hard to tell. Why? Because the moment you start to define the parameters of randomness you lose your randomness.

The best method that scientists and programmers have come up with is to create a series of tests. There exists a multitude of tests and many test suites. Popular tests include NIST RNG testing suite, Dieharder test suite [8], Knuth test, Crypt-x test, and the list goes on. RNGs come in all shapes and sizes, so you can imagine it would be difficult to orchestrate testing for all the different types of random generators.

There are two phases to test the random number generator process. First you need a source of entropy[1] that is impossible to guess like the weather. Second you need a deterministic algorithm to expand the seed into a multitude of sequences for keys and the like. Testing usually starts with checking...

Log-in to view the full article

The World's 1st Quality Content Exchange

Provable Randomness: How To Test RNGs

Our platform enables you to manage and monetize your copyrighted works the way you want.