Building trees based on aggregation efficiency in sensor networks

Abstract

Sensor network protocols must minimize energy consumption due to their resource-constrained nature. Large amounts of redundant data are produced by the sensors in such networks; however, sending unnecessary data wastes energy. One common technique used to reduce the amount of data in sensor networks is data aggregation. Therefore, we consider the impact and cost of data aggregation in sensor networks to achieve energy-efficient operation. We propose a new notion of energy efficiency that can be used to decide where aggregation points in the network should be placed. The optimal choice of these points is determined by the aggregation efficiency, which determines the amount of data reduction, and the cost in terms of energy to perform the aggregation. We present our aggregation tree algorithm “Oceanus” that produces energy-efficient aggregation trees by taking into account both of these factors. Our evaluation shows that Oceanus provides higher energy efficiency compared to existing solutions.