The design of a fast inverse module in AES

Almost all kinds of data can be transmitted through a network or by wireless. In order to protect the content or to keep privacy, the information system should be equipped with encryption and robustness techniques. Many of these techniques and applications are based on the arithmetic of the finite field, for example, AES (Advanced Encryption Standard) and RS (Reed Solomon) codes. In the operation of computation in AES, several steps have to use inverse and multiplication functions, which are analyzed as the critical path. The Rijndael proposal for AES (see Daemen, J. and Rijmen, V., "AES Proposal: Rijndael", Document Version 2, 1999) suggests using look-up tables to replace complex computation modules. Each table is neither of suitable size nor available for various applications, such as the choices of different irreducible polynomials. This paper presents the analysis of a new algorithm for computing inverses in GF(2^m) on the standard basis. The architecture of the inverse module can use parallel and pipelined techniques to speed up the AES. This method reduces the complexity of the critical modules and promotes the overall performance. Those modules are particularly simple, flexible and expandable; therefore, the algorithm is naturally suitable for VLSI implementation