Traffic Flow Analysis of a Multi-hop Wireless Sensor Network Subject to Node Failure

We develop an analytic model for traffic flow analysis in a multi-hop wireless sensor network subject to node failure through a queueing theoretic framework. The dynamics of traffic flow in the source-to-sink path is modeled by a number of singleserver queues with finite capacity linked in tandem. Both the blocking effect of tandem queues due to limited buffer and the impact of node failure due to limited node power are taken into account during modeling. The tandem-node path is analyzed by decomposing it into individual nodes with modified arrival and service processes as well as queue capacities. An iterative algorithm is developed for evaluating the performance metrics. One metric, i.e., source-to-sink delay, is used for studying QoS (quality of service) control. A simple QoS control scheme is proposed and its effectiveness is validated by simulations.