Re: Brute-forcing DES

[mpd@netcom.com (Mike Duvos)]

"Peter Trei" <trei@process.com> writes:

 > Sadly, after further calculation, I'm not so sure if it's
 > doable just yet.

...

 > The fastest general purpose, freely available des
 > implementation I'm aware of is libdes. by Eric Young. With
 > this, I can do a set_key in 15.8 us, and an ecb_encrypt in
 > 95 us/block. That adds up to about 9,000 keytests/sec (this
 > is on a 90 MHz P5, running NT).

What you really want to do to sweep the DES keyspace is to
"schedule" the input and output block you are testing, performing
any static operations, and do only enough computation to see that
a given key fails.  Special purpose assembler to do this
particular function would probably run faster than any algorithm
which could also be employed to encrypt data.

 > What will make this brute doable, if not now, then in the
 > near future?

 > 1. Faster Processors

 > 2. More processors.

 > 3. More interest

4. Better code.

This is actually a problem I plan to analyze someday. Looking at
single DES as a function of the key bits with the input and
output fixed.  This can be viewed as a boolean function, whose
result depends upon whether the given key works to map the input
onto the output.  Viewing this function as a composition of
single bit operations and optimizing it would perhaps lead to
insights on how best to compute it on a typical 32 bit CPU with
the usual collection of operations.

A messy little project, but probably one worth doing if I get
some free time.

Single DES is certainly ripe for a spectacular public failure.  A
little analytic work could bring breaking it within range of
available computing power.  If you are going to use regular
encryption code to brute force the keyspace, then it probably is
just a tad beyond reach at this point.

--
     Mike Duvos         $    PGP 2.6 Public Key available     $
     mpd@netcom.com     $    via Finger.                      $