Black-Box Attack Using Neuro-Identifier

In this book, the cryptanalysis problem of unknown cipher system is presented as a Black-Box, nonlinear, system identification problem, which assumes no priori knowledge about the cipher system except its input and output. Black-Box model is constructed using Artificial Neural Networks to attack the target cipher system, with the addition of building an equivalent Neuro-Model for the target cipher system. In cryptanalysis terminology, it is a known-plaintext attack. The constructed Neuro-Identifier has been used in two different approaches; the objective of the first approach is to determine the enciphering key, which is a total break and represent the ultimate performance of any cryptanalysis method. The cryptanalysis objectives are achieved in this approach. The aim of the second approach is to build an equivalent Neuro-Model for the target cipher system, which presents a new direction in cryptology. The constructed Neuro-Model can be regarded as an equivalent model for the real target system, which is a valuable product in many cases where the real system is not available.