Development and Evaluation of a Game Theory-Based Security Framework for Attack Detection and Mitigation in Fog Computing
Development and Evaluation of a Game Theory-Based Security Framework for Attack Detection and Mitigation in Fog Computing
Sina Pakdel1
1) Graduate of Computer Engineering, Islamic Azad University, Borujerd Branch, Borujerd, Iran.
IOI Code
:
ABEB-ACAGEZ
Publication :
4th International Conference on Recent Advances in Engineering, Innovation and Technology - Belgium(eitconf.com)
Abstract :
The proliferation of the Internet of Things (IoT) and the growing demand for real-time processing have made fog computing an important paradigm for distributed computing. However, the distributed nature of fog nodes and their resource constraints pose significant challenges to securing these environments. This study aims to investigate security enhancement in fog computing through the application of game theory by modeling the interaction between an attacker and a defensive system and evaluating mechanisms for attack detection and mitigation. In the theoretical analysis, the players strategies and the outcomes of their strategic interactions are examined using the concept of Nash equilibrium. For the performance evaluation, the network is simulated using OMNeT++. The simulation topology consists of 500 IoT nodes, 25 fog nodes, and six cloud servers. Network performance is compared under three scenarios: without security mechanisms, with encryption, and with both encryption and the deployment of intrusion detection nodes. The evaluation considers service latency, energy consumption, and attack detection time. The results indicate that the average latency reaches 2.6655 seconds when intrusion detection nodes and encryption are deployed, compared with 2.6268 seconds in the absence of these mechanisms. Accordingly, the additional latency overhead introduced by the security mechanisms is estimated at approximately 38.7 milliseconds. The elapsed time until attack detection ranges from 2.4810 to 2.5303 seconds in the simulation, while the operational detection time, measured from the beginning of the inspection session, is reported as 0.5 seconds. Furthermore, the game-theoretic model yields two Nash equilibria: allowing communication when no malicious behavior is present and terminating the connection when an attack attempt is detected. These findings highlight the importance of jointly evaluating security effectiveness, latency, and energy consumption when designing defensive mechanisms for fog computing environments.
Keywords : fog computing network security game theory nash equilibrium intrusion detection internet of things (iot)