Enhancing cloud security: harnessing bayesian game theory for a dynamic defense mechanism

Security challenges in complex information technologies continue to grow and diversify. To improve network security, many researchers have explored the game theoretic approach as a hopeful modeling tool. Knowing that the attacker can take advantage of vulnerabilities and explore existing weaknesses in the network configuration to gain access to the system for a successful attack, our objective is to benefit from virtual machines' migration as a moving target defense technique and honeypot as a deceiving technique to increase the attack surface's dynamicity. This paper presents a game-theoretic framework for modeling attack-defense interaction. A model based on incomplete information game and attack graph is developed. Our main findings reveal in which case migration of virtual machines should be established in a architecture where a honeypot is deployed and identify the potential attack paths based on system security parameters. This provides network administrators with the ability to find unsecure nodes, avoid negative externality and more precisely inefficient migrations which impact the quality of service.