[1203] in cryptography@c2.net mail archive
Re: Better DES challenge update
daemon@ATHENA.MIT.EDU (Phil Karn)
Fri Jul 11 10:32:15 1997
Date: Thu, 10 Jul 1997 21:49:59 -0700 (PDT)
From: Phil Karn <karn@qualcomm.com>
To: daw@cs.berkeley.edu
CC: cryptography@c2.net
In-reply-to: <5phkh0$8l8@joseph.cs.berkeley.edu> (daw@cs.berkeley.edu)
>An alternate (or additional) theory: popcount is extremely useful
>for implementing LSFRs, so perhaps the NSA wants to do lots of LSFR
>analysis on their Crays...
Not just LFSRs, but also things like convolutional encoders and other
useful communication devices.
I have always held the theory that the NSA has better things to do
with their supercomputers than conduct brute-force key searches on
ciphers that are better conducted with cheaper special purpose
machines.
Supercomputers, especially those with good floating point units, make
excellent general purpose digital signal processors. Remember, they
have to get their ciphertext somehow before they can attack it. And
much of this ciphertext comes out of thin air.
Phil
