Implementation of a Low-Power Intrusion Detection System for IoT
Victor Maina, Dennis Mugambi Kaburu, Tobias Mwalili · 2021 International Conference on Electrical, Computer and Energy Technologies (ICECET) · 2021
Technological advances in internet of things have enabled processing, storage and retrieval of information with minimal resource overhead. However, these battery-powered systems cannot host computationally intensive intrusion detection and prevention systems such as cloud-based machine learning intrusion detectors or blockchain encryption. We propose a low-power security setup for the internet of things. Power consumption is reduced by under clocking the gateway device processor, limiting peripherals and employing efficient network intrusion detection schemes. Security performance of the system is validated by comparing penetration test results for the power-optimized system and a baseline system running on default power configurations. This setup confirms that power savings of up to forty four percent can be achieved without compromising system security.