An end-to-end stochastic delay bound analysis in sensor networks

Sensor Networks are deployed in many fields for its convenience and feasibility, while the limited energy and the uncertain topology bring a difficult problem to provide QoS guarantees in real applications. There are less widely valid and accepted methodology for modeling and estimating the performance of sensor networks. Some approaches have been proposed to provide a worst-case performance analysis based on the deterministic network calculus. However, the worst-case analysis wastes a lot of network resources and its performance bounds cannot be tight enough. In this paper, a stochastic end-to-end delay bound is presented. Compared to the node-by-node bound, we proved that the calculation of the stochastic bound is more simple. We also show that this stochastic delay bound is more tight than the deterministic bound. Further, for the SBB-form bounding functions, an analytical solution for our stochastic end-to-end delay bound is given, which can predict the performance in deployments of sensor networks well.