WEAK1 -- A Poor Man's Data Encryption Algorithm

Mok-Kong Shen (mok-kong.shen@stud.uni-muenchen.de)
Thu, 02 Jul 1998 11:00:06 +0100

• Next message: Ben Laurie: "Re: safety of blindly truncating hashes (Re: TEA)"
• Previous message: Hamdi Tounsi: "blowfish's key len: using 576-bit keys"

Recently I have proposed elsewhere a mixture of stream and block
encoding. A full implementation of the scheme in Fortran 90 is now
available at the URL:

     http://www.stud.uni-muenchen.de/~mok-kong.shen/#paper6

The basic algorithm employed has a user choosable variable number of
rounds, with each round performing the following on a record of 160
hexadecimal digits (80 bytes of information):
  a. Pseudo-random permutation of the hex digits.
  b. Transformation of the hex digits with pseudo-randomly selected
     substitution tables.
  c. Circular shift of the words (groups of 8 hexs) by pseudo-
     randomly determined number of bit positions.
  d. Addition modulo 2^32 of the words with words of pseudo-random
     bits obtained from a multiple seed PRNG.

The pseudo-random number generation is done through activating one
and the same PRNG (in our implementation the one by Park and Miller)
using a number of seeds in round robin fashion. The resulting output,
consisting of a regularly interleaved sequence of values corresponding
to the different seeds, is shuffled by the method of Bays and Durham.
This renders it difficult to infer the ensemble of seeds from the
sequence outgoing from the buffer of shuffling, thus enabling an in
principle arbitrarily long effective key to be securely used.

Critiques, comments and suggestions for improvements are sincerely
solicited.

M. K. Shen

• Next message: Ben Laurie: "Re: safety of blindly truncating hashes (Re: TEA)"
• Previous message: Hamdi Tounsi: "blowfish's key len: using 576-bit keys"