I'm currently coding an encryption algorithm based on Reversible Cellular Automata.

The original idea is from Stephen Wolfram.

What I do is :

loading the file into bitfield form.

The key is 1024bit and give me 128 type of transformations appliable on the bitfield (from the 256 possible elementary cellular automata rules). I apply this 128 transfo according to the key. Then I got two bitfields necessary to decypher the file.

The size of the file is doubled, so (we need two steps of transfo in order to reverse the transformation. I added a dependance on the step back in order to have a reversible transfo). This is the main issue.

Second issue is that is a symetric encryption.

 

The number of combinations giving a 1024bit key is :

256^128

WONDERFUL isn't it ?

 

The algorithm doesn't need math algebric formula, just 8 axioms per rule, just like the 256 elementary cellular automata of Wolfram.

 

The main info is that we have :

256^( (nb of bit of the key)/8 ) possibilities

 

So the nb of possibilities involved in order to break the algorithm is growing exponentially with the length of the key.

 

I have done the bitfield management code, the reversible cellular automata code, now I'm coding the clean application.

 

Anyone interested ?

Interesting project. Post the download when you get done. I can't wait to try it out. Is it more secure than AES or DES?