Genetic Algorithms for Solving Problems of Access Control Design and Reconfiguration in Computer Networks

Abstract

To create solutions for providing the required access control in computer networks it is not sufficient to have only tools and protocols in the network that are needed for it. It is necessary to create corresponding configuration, or scheme, of such tools, which will allow us to satisfy the existing security requirements. At the same time, the problems of creating an access control scheme, as a rule, are NP-complete and require heuristic models for their solving. In this article, we propose a unified approach to creation of control access schemes, based on usage of genetic algorithms. The approach is applied not only to original schemes configuration but to reconfiguration as well. Successful testing of the suggested approach on RBAC, VLAN, and VPN schemes allows us to suppose that it may be applied to other types of access control schemes as well. Experimental testing of suggested genetic algorithms, performed on a specially designed test bed, showed their sufficiently high efficiency.