Defense against network attacks using game theory

As Internet has become ubiquitous, the risk posed by network attacks has greatly increased. Network attacks have been used to achieve a wide gamut of objectives ranging from overloading a website to accessing classified data. Effective defense against such attacks is a critical research area. In this thesis, we demonstrate how game theory can be used to devise effective defense systems. We utilize game theory for defense systems in two scenarios in this thesis. The first scenario is that of the attacker carrying out a Distributed Denial of Service (DDoS) attack. The second scenario involves the attacker possessing the ability to carry out a number of different attacks such as Denial of Service (DoS), Dictionary attacks and Portscans. An important restriction imposed in repeated complete-information games is that each player has complete knowledge of the adversary???s payoffs. This assumption is unrealistic when the adversaries are the defense system and the attacker. We employ a Fictitious-Play approach in order to remove this restriction