A Trust-Based Offloading for Mobile M2M Communications

Nowadays mobile Machine-to-Machine (M2M) communications are widely used in the IoT applications. In this context, a large number of devices are integrated making a distributed and heterogeneous network. This technology brings a great challenge into reliability, trust management and quality of services. In such a context, the goal of our paper is to propose a trust management system for mobile M2M communications. Since nodes may lie about their profiles and can perform attacks such as bad mouthing, self-promoting and on-off attacks, our solution relies only on direct observations collected from previous interactions. The reinforcement learning and a feedback system have been used to enable the initiator node to rate the trustworthy nature of other nodes. We evaluated the performance of our model using a real test-bed implementation in existence of different attacks. Results show that our model outperforms existing literature in terms of energy consumption and execution time.