Analyzing, modeling and predicting organizational effects in a distribuited sensor network
© The Brazilian Computer Society 2006
The organizational design of a distributed system defines how entities act and interact to achieve local and global objectives. We describe how a system employing different types of organizational techniques has been used to address the challenges posed by a distributed sensor network environment. The high-level, multi-agent architecture of this realworld system is given in detail, and we provide empirical results demonstrating the effects the organization has on the system’s performance across several different metrics. As with any design, the particular approach that is employed makes trade-offs, some of which are obvious and some more subtle. The presence of such trade-offs motivates the need for a better understanding of precisely how the organization influences large and small-scale behaviors. To address this need, we first demonstrate how a collection of analytic models can be developed to predict such effects. This experience is then used to ground the presentation of a more comprehensive, domain-independent organizational modeling language called ODML. The structure and capabilities of ODML are explained through the construction of a unified model of our sensor network organization. We then show that this model provides an accurate prediction of the original empirical results.