Query Routing Optimization in Sensor Communication Networks

Interest in large-scale sensor networks for both civilian and military applications is burgeoning. The deployment of such networks will require new approaches in resource discovery, query processing and data routing. This paper presents a framework and some analytic results for query satisfaction and data routing in networks consisting of clients, sensors and data filtering/fusion servers. In this model, multiple clients pose queries that are satisfied by processing a set of sensor data streams through a set of filters or fuselets. Fuselets are lightweight data fusion algorithms that can be deployed in a network environment. The queries can have common subexpressions (sub-queries) that should be reused by multiple clients if possible and appropriate. Moreover, effective routing of data streams from sensors to clients requires routing the streams through network nodes that can implement the required filtering/fusion operations. We formulate these problems quantitatively and propose a dynamic programming based solution using sensor-fuselet location and performance tables. The framework is preliminary in that many details and variations are abstracted or ignored. However, at the end of the paper we discuss several directions that can be explored to make these preliminary results more relevant to real scenarios.