A novel framework for collaborative intrusion detection for M2M networks

Abstract

The proliferation of sensor devices has introduced exciting possibilities such as the Internet of Things (IoT). Machine to Machine (M2M) communication underpins efficient interactions within such infrastructures. The resource constraints and ad-hoc nature of these networks have significant implications for security in general and with respect to intrusion detection in particular. Consequently, contemporary solutions mandating a stable infrastructure are inadequate to fulfill these defining characteristics of M2M networks. In this paper, we present COLIDE (COLlaborative Intrusion Detection Engine) a novel framework for effective intrusion detection in the M2M networks without incurring high energy and communication cost on the participating host and edge nodes. The framework is envisioned to address challenges such as flexibility, resource constraints, and the collaborative nature of the M2M networks. The paper presents a detailed system description along with its formal and empirical evaluation using Contiki OS. Our evaluation for different communication scenarios demonstrates that the proposed approach has limited overhead in terms of energy utilization and memory consumption.