A case study of networked sensors by simulations and experiments

Abstract

In this paper we first introduce sensor networks then we illustrate the problems of performance evaluation in such networks. We compare the two evaluation methods, simulation and experiments. We show how the assumptions and modellings in simulators impact the accuracy of simulation results, with our experience on evaluating flooding protocols in sensor network. We give some suggestions on whether a simulation or an experiment should be carried out while evaluating a sensor network protocol.