[erlang-questions] Password generator in Erlang
Zabrane Mickael
zabrane3@REDACTED
Fri Aug 17 09:44:05 CEST 2012
Hi Samuel,
Thanks for these useful insights.
We're not a security software company and our usage of these passwords isn't critical.
Anyway, we want to learn if possible.
> First, random:uniform is not cryptographically secure, which means is
> somewhat predictable. As already mentioned use any other generator
> meant to be secure as the one in crypto or the ssl library.
> Even using a secure pseudrandom generator:
>
> Your first implementation destroys the security, as you are creating a
> seed for each random number an attacker just needs to guess the seed
> sequence, not the pseudorandom sequence. In your case you had a side
> effect of generating collisions, but that was not the worst problem.
Yup. Sverker response was very clear to this regard!
> The second implementation is more secure in that sense, but still the
> original seed is guessable. An attacker can generate possible password
> sequences by bruteforce just tying possible now tuples around the time
> he thinks the real seed was created.
So, how one can generate a secure un-predicatable seeds?
> So, if you want to create passwords difficult to guess, you need at
> least a cryptographically secure PRG, which will give you an
> unpredictable sequence of bytes, and an unguessable seed, which will
> give prevent any attacker from creating the same sequence of bytes
> again an completely break all your passwords.
We also moved to "Tiny Mersenne Twister" (https://github.com/jj1bdx/tinymt-erlang) instead of using
the standard random:uniform since the last Yaws security alert (http://erlang.org/pipermail/erlang-questions/2012-June/067626.html).
Is this sufficient or should we also find a way to generate a unpredicatble seed for it?
Regards,
Zabrane
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20120817/06e222a6/attachment.htm>
More information about the erlang-questions
mailing list