On 26 Mar 1998, James Youngman wrote:
> The trouble with this is that it places a strong relationship between
> successive passwords. This means that the breaking of one password
> can be fatal; the knowledge of one password allows you to break the
What's even worse, the bad guy has only to obtain your random number seed
and he can launch a nice dictionary attack consisting of only a few
thousand passwords and he can break every password that ever used this
random system.
> passwords of all the following users trivially. I've seen this done
> to deduce the passwords of an entire year of undergraduates [in fact
It has also been used to deduce all the passwords of a notable government
agency. Rather than run the risk of their users picking "bad" guessable
passwords, they produced a list of passwords using some means and assigned
every user one of these so-called "good" passwords. The hackers promptly
obtained the algorithm used to generate the passwords and cracked them
all.
> If you must select random passwords, then please use a
> (cryptographically) stronger source of randomness, for example
You probably will gain more security by shadowing the password file than
you will by insisting on any sort of deterministically selected passwords.
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
