I know little about RNGs -- but wouldn't it be practical to run about 3 RNGs with different algorithms in parallel, and add their results? Surely that would be a lot harder to attack than any one RNG.
Ditto if you start with the RNG output and apply even a simple transformation to it. After all, if you start with a pseudo-random-number and apply to it an isomorphism from the valid number set to itself, you should wind up with something that's still pseudo-random.
The old crypto literature has lots of ways of combining RNGs to make better ones (for some values of "better"). You're generally better off just designing one with known properties.
I was thinking of it as a way of protecting against design flaws. Similarly, I think -- and Ed Snowden recently said the same thing -- that for max security one should wrap things in a sequence of different kinds of cryptography.
And by the way, not all parts have to be individually robust.
Random numbers don't always need to be cryptographically secure. In situations like games you might just want to use a function with low resource requirements.
If you want a cryptographically secure random generation you should use SecRandomCopyBytes(3) which I think comes from an isolated hardware random number generator and should be tough to monitor through software.
Disclaimer: I'm not sure if what I'm saying is true
4 comments
[ 3.3 ms ] story [ 25.0 ms ] threadDitto if you start with the RNG output and apply even a simple transformation to it. After all, if you start with a pseudo-random-number and apply to it an isomorphism from the valid number set to itself, you should wind up with something that's still pseudo-random.
And by the way, not all parts have to be individually robust.
If you want a cryptographically secure random generation you should use SecRandomCopyBytes(3) which I think comes from an isolated hardware random number generator and should be tough to monitor through software.
Disclaimer: I'm not sure if what I'm saying is true