Re: dev random and urandom
On Sun, 28 Oct 2001, James Morrison wrote:
> So far I have located two different (u)random translators.
> The one at http://mel.interspace.dk/~sune/hurd/entropy.html
> and the one from marcus at
> ftp://alpha.gnu.org/gnu/hurd/contrib/marcus/random.tar.gz
> These translators seem to gather entropy in different ways. I'm
> looking for a way to test them to see any advantages or disadvantages
> to either.
From my meagre knowledge of randomness, the best way to test if
things are good random generators is to query them until you get
repeats. Of course, best may not be most efficient, or most practical.
*grin*
Maybe you'd like to look at the code to see if they are doing
anything to dilute randomness? Maybe if rounding occurs, or if a
function is applied to the entropy (then, you have to look at this
function and try to figure out whether it treats all entropy
equally.) Generally, the more sources of entropy, the better (because
then you can keep your pool full. IIRC, Linux gathers entropy from
keyboard and mouse movements, although there are patches for serial
communications as well as network traffic collectors.
Simon
Reply to: