Analysis of Multilayer Selfish Attacks in Internet of Things Sensor Networks Using an Extended Bianchi Throughput Model

Improving network performance has become an active area of research in light of recent advances in Internet of Things sensor networks. This performance depends on configurable parameters at the transmitting node, including the contention window (CW), arbitration interframe space number (AIFSN), packet generation rate, and transmit (Tx) power. However, these parameters can be exploited by malicious nodes, leading to selfish attacks, which monopolize network resources. Conventional studies on selfish attacks primarily focus on manipulating parameters within a single protocol layer and have not sufficiently considered selfish attacks that simultaneously manipulate multiple parameters across different layers. Therefore, this article proposes a closed-form performance analysis framework based on the Bianchi throughput model to analyze the impact of selfish attacks involving multilayer parameter manipulations on network performance. The framework quantitatively evaluates the effects of such manipulations by incorporating the influence of the packet generation rate, AIFSN, and Tx power into an extended Bianchi throughput model. The performance of the proposed model is validated by evaluating various combinations of layer-specific parameter manipulations and quantitatively analyzing the resulting performance variations between selfish and normal nodes. Numerical results show that the throughput of a selfish node becomes only 2.790 times higher than that of a normal node when it manipulates only the packet generation rate. When the CW, AIFSN, and packet generation rate are manipulated simultaneously, its throughput increases to 10.744 times that of a normal node. When only the Tx power is manipulated, the throughput becomes 3.540 times higher. Finally, when the CW, AIFSN, packet generation rate, and Tx power are manipulated simultaneously, the throughput of the selfish node becomes 13.702 times higher than that of a normal node, severely compromising the fairness of network resource allocation.