Implementation and tests of clustering over vertically distributed data without sharing their values

This paper describes an experimental implementation of a secure distributed k-Clustering algorithm which applies the homomorphic encryption scheme for preserving privacy of data. We show in details technical aspects of the homomorphic encryption implementation and using it in a distributed application. The implementation enables us to evaluate experimentally the efficiency of the clustering algorithm, indicate its limitations and potential bottlenecks. Finally, we provide some conclusions and propose further work in this area.