Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Thu May 27 1999 - 23:44:22
Mike Rosing (eresrch@msn.fullfeed.com)
Mon, 26 Apr 1999 20:58:57 -0500 (CDT)
On Mon, 26 Apr 1999, Mj Dascombe wrote:
> I am a first year student at Manchest University. I have been following
> the posts of the analysis the UNIX random function. Has it ever been
> suggested that seperate hardware be used to generate the random bits
> required for key generation. I created such a device using a FLIPFLOP
> random bit circuit and a clock, attached to pin 2 of an RS232, wich
> provided a terminal program with random characters. Would these characters
> not be random beyond question?
No, you have to ask lots of questions. Run up about 10 MB and feed it
to DIEHARD and see how it performs. If you get a batch of all 1.0000
or 0.0000 probabilities chances are good your hardware generator has
a problem. You can also try Diaphony from pLab which only needs 128K
to do multidimensional correlation analysis. You want those values to
be in the range .998...1.02 or there are problems. MUST has also been
discussed here but I'm not sure exactly what it does. Do a web search
for all these things, I'm sure you'll get something that will help show
you how hard it is to build a hardware RNG :-)
Patience, persistence, truth,
Dr. mike
The following archive was created by hippie-mail 7.98617-22 on Thu May 27 1999 - 23:44:22