From: Michael Paul Johnson <mpj@netcom.com>
Date: Sat, 6 Jan 1996 08:52:10 +0800
To: cypherpunks@toad.com
Subject: Ruby Block Cipher Mark 5
Message-ID: <Pine.SUN.3.91.960105090605.14176B-100000@netcom19>
MIME-Version: 1.0
Content-Type: text/plain


-----BEGIN PGP SIGNED MESSAGE-----

The Ruby Block Cipher, Mark 4, had some problems with slow avalanche with a
worst-case input, as well as some really amateurish reference code bugs that
I apologize for.  The good news is that these things have (I hope) all been
corrected in the Ruby Block Cipher, Mark 5.

The biggest change to the algorithm is that the "family key" is no longer
simply added (modulo 2) to the input blocks, but is used in add and multiply
steps to eliminate the old worst case avalanche patterns.

The Ruby Block Cipher is not a general block cipher in that it cannot be used
in Electronic Codebook (ECB) mode.  It is a cryptographic hash function with
a block size of only 64 bits. Of course, 64 bits is too short for a
cryptographic hash function intended for digital signature use, but it is
just fine for a quick block cipher.

This may be a good reference for those folks who want a quick & easy
encryption algorithm that need not withstand nuclear attack but can provide
something better than common weak encryption methods in use in the software
industry.  The small amount of code, fast operation, and lack of the need for
a lengthy key setup time are definite advantages where computing resources
are at a premium.  On the other hand, fast key setup time substantially
reduces the cost of a brute force attack on the key, so use of the full 64
bits of the key is essential.

Your comments and suggestions on this rather strange little cipher are
welcome and encouraged.  I'm particularly interested in any ideas as to how
many rounds (the STRENGTH constant in the source code) are appropriate for
well-balanced security.

Information on the Ruby Block Cipher is available as

ftp://ftp.csn.net/mpj/public/ruby_m5.ps.gz or
ftp://ftp.csn.net/mpj/public/ruby_m5.rtf.gz

and, if you are in the USA or Canada, a reference implementation and a sample
file encryption program with free source code is in

ftp://ftp.csn.net/mpj/I_will_not_export/crypto_???????/mpj/ruby_m5.zip
where the ??????? is revealed in
ftp://ftp.csn.net/mpj/README

ruby_m5.zip is also available on the Colorado Catacombs BBS at 303-772-1062.

I urge caution in using such a new cipher in actual applications, but if you
feel the need to, you might consider asking me if any known weaknesses have
been reported before you do.

Note: this is not a product for sale (it is free and probably worth at least
as much as you pay for it).  It is also not a prepublication (it is THE
publication in electronic form with no paper publication anticpated in the
near future).

                  ___________________________________________________________
                 |                                                           |
 |\  /| |        | Michael Paul Johnson  Colorado Catacombs BBS 303-772-1062 |
 | \/ |o|        | PO Box 1151, Longmont CO 80502-1151 USA   Jesus is alive! |
 |    | | /  _   | mpj@csn.net aka mpj@netcom.com m.p.johnson@ieee.org       |
 |    |||/  /_\  | ftp://ftp.csn.net/mpj/README.MPJ          CIS: 71331,2332 |
 |    |||\  (    | http://www.csn.net/~mpj                 -. --- ----- .... |
 |    ||| \ \_/  | PGPprint=F2 5E A1 C1 A6 CF EF 71  12 1F 91 92 6A ED AE A9 |
                 |___________________________________________________________|



-----BEGIN PGP SIGNATURE-----
Version: 2.7.1

iQCVAgUBMO1VgvX0zg8FAL9FAQECzQP/fD9dLLeixfZAtKXM2tDXrGgrashiqEsn
jU9ohnNsou9MMU+PUmNj8RJCRHSXy4HMskV5BhnILUYaSI5ztJjZYvhBcIbBcR8J
ecGl5++iaj4vRAb1vs32Y6LVsQm7hsMvy0byaszOWUKDpn+ZJrFCwMiKgD50ecXW
y+XlpkdOhiA=
=FQec
-----END PGP SIGNATURE-----
Please include my address on followups, since I don't read all mail on this
wonderful (but high volume) list.  Thanks! mpj@netcom.com