An analytical traffic flow model for cluster-based wireless sensor networks

Abstract

An analytical traffic flow model is developed for cluster-based wireless sensor networks. The source-to-sink path is modeled by a number of single-server finite queues linked in tandem. During the process of modeling, both the blocking effect of tandem queuing networks and the impact of connection failure due to limited node power are taken into account. The tandem queuing network is analyzed by decomposing it into individual nodes with modified arrival and service processes and modified queue capacities. The steady-state queue-length distributions of individual nodes are determined iteratively by using the matrix-geometric procedure. A computational algorithm with a global loop from the last node to the first node and local loops in individual nodes is developed to determine the performance metrics, such as source-to-sink throughput (SST) and source-to-sink delay (SSD), of the tandem network.