Re: Minimum length of secret key

Michel Bouissou <[email protected]> Sat, 24 Jul 2004 13:37:36 +0200
Newsgroups gmane.mail.spam.srs.general
Organization Me, Myself and I
Message-ID <[email protected]>
Le samedi 24 Juillet 2004 13:17, James Couzens a écrit :
> >
> > What is minimum recommended length of the secret key used to generate the
> > cryptographic hash so that the secret key cannot be guesses by
> > cryptographic analysis of email addresses modified by SRS?
>
> 512 bits for SHA-160, and ideally they would be all truly random too.

If I understand well, the default length of the SRS generated hash is 4 
characters, base-64 encoded, which makes "24 bits of security" according to 
http://www.libsrs2.org/srs/srs.pdf , §4.2

This actually looks quite short to me, especially as $4.1 of the same document 
states that the hash comparison will pass with only a warning if the case of 
the hash would have been modified. I wonder if this doesn't reduce the 
security of the hash to about 20 bits for a 4-characters hash...

I understood (maybe I'm mistaken) that the secret should be made of characters 
from the base-64 set. So if we would want a 512-bit equivalent secret, we 
would need a secret made of 86 base-64 characters ;-)

-- 
Michel Bouissou <[email protected]> OpenPGP ID 0xDDE8AC6E