A multiplayer game model to detect insiders in wireless sensor networks

Area of Science:

• Computer Science
• Network Security
• Game Theory

Background:

• Insider threats pose significant risks to Wireless Sensor Networks (WSNs) due to attackers' privileged access.
• Compromised nodes can disrupt normal network operations, necessitating rapid detection and mitigation strategies.

Purpose of the Study:

• To propose a novel game theory model, the Game of Wireless Sensor Networks (GoWiSeN), for detecting and mitigating insider attacks in WSNs.
• To develop a framework that utilizes local and global intrusion detection systems to identify and isolate malicious nodes.

Main Methods:

• Formulated an imperfect information, non-cooperative game theory model assuming rational players.
• Integrated Local Intrusion Detection Systems (LIDSs) communicating with a Global Intrusion Detection System (GIDS).
• Utilized extensive form game representation and von Neumann-Morgenstern utility functions to quantify outcomes and payoffs.

Main Results:

• The GoWiSeN model was solved by locating Nash Equilibria (NE) in both pure and mixed strategies.
• Experimental evaluations on real network datasets demonstrated the model's efficiency in detecting and isolating compromised nodes.
• Simulations included various Intrusion Detection System (IDS) capabilities and specific insider attack scenarios.

Conclusions:

• The proposed GoWiSeN model effectively addresses insider attacks in WSNs.
• The game theory approach provides a robust framework for real-time threat detection and response in WSNs.
• The findings offer practical insights into securing WSNs against sophisticated insider threats.