On Thu, 21 Apr 2005 11:32 pm, Rod Taylor wrote:
> On Thu, 2005-04-21 at 11:06 +0200, Tino Wildenhain wrote:
>
> One advantage of a random salt would be that the username can be changed
> without having to reset the password at the same time.
And the main advantage is that if you have an account with the same username
on another database server, it makes attacks twice as hard. (Assuming those
two database servers, or their backups are compromised)
Note that salts aren't supposed to be secret (well, no more so than the
password hash), they just need to be unique.
Of course, if you're going to add a salt (which requires changing the protocol
and the storage format), you may aswell move away from md5 (which has been
considered deprecated for quite some time... longer than it's been used in
PostgreSQL), and move towards something more secure, such as blowfish crypt
(or an AES adaptation of it?).
> On Thu, 2005-04-21 at 11:06 +0200, Tino Wildenhain wrote:
>
> One advantage of a random salt would be that the username can be changed
> without having to reset the password at the same time.
And the main advantage is that if you have an account with the same username
on another database server, it makes attacks twice as hard. (Assuming those
two database servers, or their backups are compromised)
Note that salts aren't supposed to be secret (well, no more so than the
password hash), they just need to be unique.
Of course, if you're going to add a salt (which requires changing the protocol
and the storage format), you may aswell move away from md5 (which has been
considered deprecated for quite some time... longer than it's been used in
PostgreSQL), and move towards something more secure, such as blowfish crypt
(or an AES adaptation of it?).
[ reply ]