Insecurities in the Linux /dev/random

Re:Insecurities in the Linux /dev/random by syzo on Monday October 14, 2013 @03:31PM (#6549116)

you can also check that (1,1), (1,2), (2,1), and (2,2) all happen with expected frequencies (i.e. roughly 25%), where the two numbers in the pair are just two numbers in a row in the sequence. You would have gotten 100% (1,2) and 0% for the rest, so that's suspicious.

same for 3-tuples, 4-tuples, ..., and n-tuples.

There's probably other, more sophisticated tests as well, but I'm not an expert.

[ Parent | on HN ]

Re:Insecurities in the Linux /dev/random by buo on Monday October 14, 2013 @03:40PM (#6549173)

I think you missed my point, which is this: the first sentence in the paper's abstract is wrong.

[ Parent | on HN ]

Re:Insecurities in the Linux /dev/random by syzo on Monday October 14, 2013 @03:44PM (#6549202)

My examples with pairs and n-tuples are also distributions.

[ Parent | on HN ]

Re:Insecurities in the Linux /dev/random by buo on Monday October 14, 2013 @04:09PM (#6549320)

That's a good point, and that may well be what the authors meant.

Since all PRNGs are periodic (at least all I now of), the definition eventually breaks down. I can see how it can be made to work with a bit of extra formalism, though.

Thanks.

[ Parent | on HN ]

Re:Insecurities in the Linux /dev/random by jameshart on Monday October 14, 2013 @06:28PM (#6550050)

Any finite state PRNG has to be periodic. Good luck implementing an infinite state one.

[ Parent | on HN ]