Low-complexity signature-based malware detection for IoT devices

Abstract

The ominous threat from malware in critical systems has forced system designers to include detection techniques in their systems to ensure a timely response. However, the widely used signature-based techniques implemented to detect the multitude of potential malware in these systems also leads to a large non-functional overhead. Such methods do not lend well to the extremely resource constrained IoT devices. Hence, in this paper, we propose a low complexity signature-based method for IoT devices that only identifies and stores a subset of signatures to detect a group of malware instead of storing a separate signature for every potential malware, as done in the existing work. Experimental results show that the proposed approach can still achieve 100% detection rate while relying on a very low number of signatures for detection.