How fast computers have made ciphers unbreakable
From: Mxsmanic (mxsmanic_at_hotmail.com)
Date: Tue, 09 Nov 2004 06:15:53 +0100
I have a theory, and I'd like to know if anyone shares it. I assert
that the increasing speed of computers has made or will make most good
ciphers effectively unbreakable. The days of codebreaking by methods
such as exhaustive key searches or other computationally intensive
techniques are gone. In the absence of a truly glaring defect in a
cipher, it is likely to be unbreakable today (given a reasonable key
length). My logic is as follows:
1. Computers continue to improve in performance (computational
2. Faster computers allow longer keys to be used with ciphers.
3. As key lengths increase, the computing power required to legitimately
use most ciphers (i.e., for encryption and decryption with the correct
key) increases much more slowly than the computing power required to
illegitimately crack them (by exhaustive key searches, factorization, or
4. The rate of increase in the ratio between the computing power
required for a legitimate use of a cipher and that required for cracking
of the cipher increases much more rapidly than the ratio between the
computing power of readily affordable desktop computers and the biggest,
fastest supercomputers on the planet.
5. It follows, then, that as computing power increases, the length of
the longest practical keys also increases, and a point is rapidly
reached at which it is still very practical to use a cipher, but it is
no longer possible to crack it through any type of brute force, because
the computing power required for the latter, when used against a cipher
that is practical to implement on an ordinary desktop computer, is
greater than anything currently available anywhere (or likely to become
available anywhere in the future).
Thus, all good ciphers are or will become unbreakable as computer speeds
increase. Of course, many ciphers do not have variable-length keys, but
as key lengths can be practically increased, new ciphers can be
developed that use longer keys.
So, is this reasoning sound? Why or why not? I assume someone else has
examined this issue; are there any places on the Web where it is
discussed in detail? I don't know if it has been discussed in sci.crypt
-- Transpose hotmail and mxsmanic in my e-mail address to reach me directly.