| Modeling and Simulation (M&S) provides a way to cost effectively play the "what-if" game, while facilitating specification, design and analysis of computer and communication networks. With the implementation of the actual network comes the question "How representative is the data obtained from simulations when compared to the actual network experimental data?".;The scope of the dissertation encompasses development of a methodology for comparison between simulation outputs and empirical measurements in the area of Wireless Sensor Networks (WSN). The methodology is based on hypothesis testing and statistical evaluation of traffic performance indicators.;The approach entails the study of a WSN platform, its performance parameters and their measurements, with least intrusion. Subsequently, modeling and simulation of the WSN platform is accomplished. For several scenarios of interest comparisons of simulation and experiment data are performed. A "Comparator Tool" is developed to facilitate the reproducibility and repeatability of model verification.;The comparison results between performance simulation data and empirical data demonstrate to the modeler and model user whether the model is a good representation of the actual real-life network being modeled. The "Comparator Tool" documents and captures the Hypothesis Testing Based methodology while producing comparison results in seconds.;A subset of the performance parameters such as bandwidth, end-to-end delay, path loss, jitter, throughput, utilization or packet loss, and others, which are the basis for which networks are graded on, are successfully compared for a WSN platform and its simulation model. |
