Enhancing Network Security with Intrusion Detection Systems in IoT Devices

Internet of Things (IoT) devices' surge has dealt a bludgeoning blow to network security, putting a new dimension to this digital space. The growing IoT landscape given by expensive technology has started reminding us of our failure to incorporate security as a core competency. We dedicate this paper to the design of Intrusion Detection Systems (IDS) tailored for the IoT environment with a purpose of improving network security. Traditional IDS methods cannot be applied to IoT networks, due to its high resource demands and limited capability to handle the characteristic traffic pattern in IoT networks. We propose a lightweight IDS by combining machine learning algorithms for the detection of real time anomalous behaviors that indicate an intrusion potential. In other words, the system combines the signature and the anomaly based method detection of both the known threats as well as the new attack patterns. The intended abilities of the framework are to be efficient for IoT devices with limited resources and good detection accuracy. Finally, we demonstrate realization of the proposed IDS through extensive simulations and real world testing of its ability to shield connected IoT networks from a set of sentinels of cyber threats including denial of service (DoS) attacks, data breach and unauthorized access. This work contributes to the state of the art of IoT security, by proposing a secure IoT solution strongly fortified against coming cyber threats.