A clustering-based approximation scheme for in-network aggregation over sensor networks

Recently in-network aggregation has gained much attention for effectively reducing energy consumption over sensor networks. In this paper, we present CASA, a clustering-based approximation scheme for in-network aggregation, which significantly reduces the overall energy consumption while maintaining user-specified data quality. We explore two ideas to support the approximation scheme. First, we propose a clustering-based framework to effectively utilize temporal coherency tolerance (tct) in conjunction with in-network aggregation to save communication cost over sensor network. Secondly, we propose an adaptive tct-reallocating technique to further reduce communication cost and maintain load balance. Our experiment results indicate that significant benefits can be achieved by using our CASA approximation scheme.