Mathematical aspects of evaluating the security of block ciphers with respect to linear and differential cryptanalysis.

Currently, the only method known to achieve absolute security, is the One-Time Pad, which uses impractically long keys. For all modern purposes we use encryption methods that have no proof of absolute security. Instead, we can prove that certain classes of attacks do not work or that they would require a very large amount of computations or certain extra information. The study of the security of encryption methods against various attacks is called cryptanalysis. In this project, we will extend the field of cryptanalysis by doing a study of two important types of attacks: linear attacks and differential attacks. We will extend the existing theory by improving the mathematical framework that is used to model these attacks. We will develop and extend tools to automate the security analysis of existing encryption methods and to improve the accuracy of the security analysis. Furthermore, we will look at new design techniques regarding the needs of modern communication.