Cryptanalysis by Solving Equation Systems

The need for secure digital communication has increased a lot during the last decade. Various algorithms are used to achieve the security needed, for example encryption algorithms, message authentication codes and digital signatures. It is important tha t these algorithms are thoroughly investigated, as the consequences of holes in their security will be big. Any encryption algorithm can be described by a non-linear equation system, and the security of the algorithm relies on the assumption that this sy stem can not be efficiently solved. Some methods have recently been proposed for solving such systems, but so far without any major results. This project will approach the problem from a different angle, and test some promising new ideas for solving the se kinds of systems. These ideas are motivated by coding theory, where similar techniques are used with great success in the decoding of error-correcting codes. The results of this research will give insight into this kind of attack on computer security , and may uncover serious security flaws in some encryption algorithms. The algorithms identified as weak should be taken out of use and replaced by stronger ones before their weaknesses are exploited.