minow@decvax.UUCP (Martin Minow) (04/02/85)
Introduction
This note describes a new cryptographic method (NCM) that has a number
of advantages over many existing systems, such as DES. While it
is similar to DES in that it requires a private (secret) key,
it is more secure than DEX and both encryption and decryption
can be performed with very modest computer resources.
In order to comply with government restrictions, actual algorithms
will not be posted. It should be fairly easy to devise algorithms
from the following description.
NCM transforms plain- to encrypted text by applying two transformations:
encryption proper (the E-box) followed by entropy-reduction (the E'-box).
The Encryption Algorithm
The E-box, encryption proper, uses an algorithm -- well known in
the art -- called "binary addition without carry" which may be
described using the following matrix
0 1 key
+----
data 0 | 0 1
1 | 1 0
I.e., each element of the message data stream is decomposed into
individual bits which are transformed by reference to the same
bit in the key data stream.
Key Data Stream Selection
What distinguishes the NCM from earlier encryption methods is that
the key data stream is bit-for-bit identical to the message data
stream. While this choice of data stream is not totally without
problems, it has the distinct advantage of not requiring that the
message sender memorize a preset key (or from applying the
knowledge of one message key to any other message).
Entropy-Reduction
The output of the E-box is further transformed by an entropy-reduction
algorithm (the E'box) which serves two general purposes: in addition
to improving the security of the encrypted message by preventing the
cryptanalyst from effective application of statistical techniques to
the encrypted message, it serves to minimize communication line
costs and transmission time.
While any entropy-reduction algorithm may be used (such as Huffman
Encoding, or the currently-popular Lempel-Ziv algorithm), it appears
from initial testing that a simple run-length encoding algorithm
will prove most approprate.
Note that, because of the special nature of the algorithm, it
is quite easy to perform encryption using only paper-and pencil
methods.
Unlike other encryption methods, it is important to apply
entropy reduction after encryption for best results. Of
course, it is possible to perform entropy reduction both
before and after encryption and different algorithms may
be used for the two E' stages.
Example
In order to test NCM, an encrypted version of the following
text was given to several experts.
Forescore and seven years ago, our forefathers brought forth
upon this continent a new nation, concieved in liberty and
dedicated to the proposition that all men are created equal.
Now we are engaged in a great civil war, testing whether
that nation, or any nation so concieved and so dedicated
can long endure. We are met on a great battlefield of that
war.
... some text omitted to save space ...
After some consideration, all were able to identify the text
as being Lincoln's Gettysburg Address, however none discovered
that "conceived" was misspelled. Thus, the claim that this
is a secure method is confirmed.
Problems
The problem of key transmission is yet to be resolved. Note,
however, that this is a problem shared with many other algorithms
-- notably DES -- and that many appropriate solutions are available.
Since each unique message has an equally unique key, some method
must be provided to let the recipient know which message has
actually been sent. Again, many appropriate algorithms are known.
Conclusion
A new encryption method having many advantages over existing
algorithms has been described.
Martin Minow
decvax!minow
Posted 1-Apr-1985jbn@wdl1.UUCP (04/03/85)
What ``government restrictions''? Please provide a reference to the U.S. Code or Federal Register. John Nagle
atst@ssc-vax.UUCP (Tom Pace) (04/05/85)
> > This note describes a new cryptographic method (NCM) that has a number > of advantages over many existing systems, such as DES. While it > is similar to DES in that it requires a private (secret) key, > it is more secure than DEX and both encryption and decryption > can be performed with very modest computer resources. > Posted 4/1/85 by Minow Minow's amazing encryption algorithm can be implimented on any UNIX system using the following csh commands: (file to be encrypted is 'fname') %tr -d a-zA-Z '\000' <fname >fname.crypt %rm fname I am still working on the decryption commands. Can anyone help me?
gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (04/06/85)
I assume everyone will recognize this as being a joke.
graner@ut-ngp.UUCP (Nicolas Graner) (04/09/85)
In <594@ssc-vax.UUCP> (by atst@ssc-vax.UUCP (Tom Pace)): > Minow's amazing encryption algorithm [posted 4/1/85] can be implimented > on any UNIX system using the following csh commands: > > (file to be encrypted is 'fname') > > %tr -d a-zA-Z '\000' <fname >fname.crypt > %rm fname > > I am still working on the decryption commands. Can anyone help me? Sure! To get the encrypted version of fname in fname.crypt and the decrypted version in fname.decrypt, use the following commands: %tr -d a-zA-Z '\000' <fname >fname.crypt %tr a-zA-Z a-zA-Z <fname >fname.decrypt %rm fname Nic. {ihnp4,seismo,...}!ut-ngp!graner *If Murphy's law can go wrong, it won't*