Comparative Analysis of Intrusion Detection Systems for Internet of Things
Nithya Nedungadi, Akshitha K Subran, Sriram Sankaran · 2023
The expansion of IoT has led to an abundance of interconnected devices that are susceptible to cyber risks, including malware, ransomware, denial-of-service attacks, and data breaches. Conventional security measures are frequently inadequate in handling these notorious threats, giving rise to the importance of Intrusion Detection Systems (IDS) as a vital means of identifying and mitigating them. The comparative analysis of IDS involves evaluating and comparing the performance of different IDS solutions based on a set of criteria, such as detection accuracy, resource consumption, and ease of deployment and management. Several network service providers offer packet analysis services, yet selecting the optimal IDS approach is a challenging and crucial decision due to its substantial impact on performance and energy consumption for resource-constrained IoT devices. IDS techniques have been well examined for this purpose in terms of performance, but the evaluation of energy usage is typically unexplored. In this work, we comparatively analyse the signature-based and anomaly-based Intrusion Detection Systems using the CIC IoT dataset. Our experiments conducted using Raspberry Pi and an external source meter reveal that Anomaly-based IDS uses 12.5% more peak power while utilizing only 8% of CPU resources, as compared to the 10% utilization of a Signature-based IDS which achieves a higher accuracy of 95.52%. The insights from the study can be used to help organisations make informed decisions about selecting and deploying the most appropriate IDS for their IoT environments.