On Sun, 23 Sep 2001, Mike Hicks wrote:

> Scott Dier <dieman+tclug at ringworld.org> wrote:
> > 
> > /me wonders if you know the exact plain text and the encrypted text, how
> > 'hard' is it to extract the secret key?

This would fall into the category of a "known plaintext attack." Strong
encryption algorithms are pretty safe from this these days. These sorts of
attacks were crucial in WWII.

> This is certainly what happens when small keys (these days, <64 bits) are
> used.  I wouldn't be surprised if the NSA can crack this size key pretty
> easily
> (within hours or days).
> 
> Most people, when using PGP or GPG, make keys of 1024 or 2048 bits.  This
> is 2^960 (9.7e288) to 2^1984 (? - probably somewhere around 10^600) times
> harder to calculate.

Be careful to distinguish between symmetric and asymmetric cryptography. A
64 bit AES key is still pretty good. Asymmetric algorithms like the kind
used for public key cryptography require much longer keys to be secure.

For anyone who's interested in a nice discussion of the role of cryptography
in history and a good description of public key cryptography I would
recommend Simon Singh's "The Codebreakers." A very entertaining piece of
geek writing.

-Tim

--
Tim Wilson      |   Visit Sibley online:   | Check out:
Henry Sibley HS |  http://www.isd197.org   | http://www.zope.com
W. St. Paul, MN |                          | http://slashdot.org
wilson at visi.com |  <dtml-var pithy_quote>  | http://linux.com