Stochastic data transformation boxes for information security applications

Abstract

Stochastic methods are commonly referred to as methods which are directly or indirectly based on using a pseudo-random number generator (PRNG). In some cases, stochastic methods are the only possible mechanism of protecting information from an active adversary. In this paper we examine a construction of R-boxes, which are a generalization of S-boxes, classical structural elements of cryptographic primitives of hashing, block and stream encryption. R-boxes are in fact stochastic adders, i.e. adders with an unpredictable operating result, which depends on the key table H. A distinguishing feature of R-boxes is their efficient software and hardware implementation.