In this paper, we present a method to create a safe arithmetic that can be used to obfuscate implementations that require operations over commutative groups. The method is based on the structure of the endomorphisms of certain extensions of the original commutative group. The endomorphisms of a commutative group are non-commutative (in general), thus we can use a non-commutative group to emulate the arithmetic of a commutative one. The techniques presented in this paper are very flexible and the programmer has a wide variety of options to obfuscate the algorithms. The system can be parameterized using conjugations, thus it is possible to generate a different arithmetic for each instance of the program with a change in the security parameters, even in cases in which this number is huge (for example, in IoT applications). The security of this method is based not only on the difficulty of the conjugacy search problem (in a harder version because only partial information about the groups is known by the attacker), but also in a number of extra options that can be chosen by the programmer. The paper explains the general method, analyzes its algebraic properties and provides detailed examples based on the vector spaces over F 2 and XOR operators.
Keywords: IoT Security; cloud computation; code obfuscation; fog computation; group-based cryptology; non-commutative cryptology; white box cryptology.