Genetic-Feedback Algorithm Based Network Security Policy Framework

As the rapid increase of network security events, policy-based management is more and more widely used to monitor and manage the behavior of network. But the policies are created and deployed by the administrator, and can not automatically adjust without the involvement of human. This paper develops an evolutional network security policy framework based on genetic-feedback algorithm. Based on the historical security events, using genetic algorithm, we can generate a rule base. When a new network event comes, the analyzer judges whether the event is secure or not according to the rule base, and the policy system may give a policy decision too. Obviously, these two results may be different. So the policies can be automatically adjusted refer to the genetic calculated results. A brief overview of genetic algorithm is presented. The proposed framework and the evolution process are discussed in detail.