Optimizing Energy Consumption for IoT Adaptive Security: A Mobility-Based Solution

The Internet of Things (IoT) is transforming communication among devices. The prevalent use of IoT in mobile scenarios introduces vulnerability to attacks, which requires robust security mechanisms. However, the heterogeneity of IoT devices results in diverse security requirements, compounded by varying threat levels across geographical zones. To address this issue, adaptive security solutions that tailor defense mechanisms to the contextual environment (i.e the security requirement of devices and threat level of the environment) are proposed. However, when the threat level increases, the defense mechanism complexity increases. This leads to higher energy consumption and consequently, device failure. To tackle this challenge, we propose in this paper a mobility-based solution to optimize the energy consumption resulting from using adaptive security solutions. Our approach considers the remaining energy of IoT devices, their security requirement and the threat level of the zone in the decision-making process. This task is carried out by an agent trained with Deep Reinforcement Learning. The proposed solution includes Software-Defined Networking and fog computing in order to ensure seamless execution of the security service while the IoT device moves.