[1524] in bugtraq

home help back first fref pref prev next nref lref last post

Re: passwd hashing algorithm

daemon@ATHENA.MIT.EDU (Timothy Newsham)
Tue Apr 18 18:23:00 1995

From: newsham@aloha.net (Timothy Newsham)
To: J.S.Peatfield@damtp.cam.ac.uk (Jon Peatfield)
Date: Mon, 17 Apr 1995 08:39:48 -1000 (HST)
Cc: don@paranoia.com, bugtraq@fc.net, jp107@damtp.cam.ac.uk
In-Reply-To: <m0s09zT-0000o5C%kro.amtp.cam.ac.uk@damtp.cam.ac.uk> from "Jon Peatfield" at Apr 15, 95 04:42:30 pm

> Too fast, it still allows dictionary attacks rather easily (yes I know that 
> users should choose good passwords, but some won't).
> 
> md5^500 (500 rounds of md5), or however many takes about 0.5 seconds on a fast 

The hashing should be computationally adjusted and should be adjusted
on each box to be barely tolerable.  There should also be a salt
value of course.  An attacker shouldnt be allowed to precompute
md5^(big num) and later do the (actual num - big num) md5's for
your particular system.

>   -- Jon


home help back first fref pref prev next nref lref last post