Secure multiparty computation of DNF

Homomorphism based multiparty computation techniques are studied in this paper as they have several advantages over the other multiparty computation schemes. A new homomorphism based multiparty computation technique is proposed to evaluate functions in DNF form. The new technique exploits homomorphism of a certain sealing function to evaluate a function in DNF. The new technique has two advantages over the existing homomorphism based multiparty computation schemes. Firstly, it supports any input format. Secondly, a general method to reduce any function to DNFs is proposed in this paper. With this method, functions like the famous millionaire problem can be reduced to DNFs and efficiently evaluated. Security of the new scheme is formally defined in the static active adversary model and proved in a new simulation model.