On the Construction of Cryptographically Significant Boolean Functions Using Objects in Projective Geometry Spaces

Abstract

Recently, several construction methods of highly nonlinear Boolean functions with relatively good algebraic properties were proposed. These approaches manage in optimizing most of the relevant cryptographic criteria, but not all of them at the same time. Usually, either the nonlinearity bounds are rather loose (though the actual nonlinearity is relatively high) or the functions do not provide a good resistance to fast algebraic cryptanalysis. In this paper, we develop a theoretical framework for using objects in suitable projective geometry spaces for construction of highly nonlinear Boolean functions. This allows us to establish tight bounds on the nonlinearity using simple counting arguments, thus avoiding rather complicated estimates of certain trace sums. Our method generates a class of almost fully optimized functions, that is the functions apart from very high nonlinearity also have the maximum algebraic degree and optimal algebraic immunity. Compared to the classes of functions proposed by Carlet and Feng, Wang , and Zeng , our functions achieve a slightly better nonlinearity which is traded-off against a little worse resistance against fast algebraic attacks. On the other hand, compared to the functions by Tang and Tu and Deng, our nonlinearity is somewhat lower, but the algebraic properties are slightly better.