Adaptive Data Storage for Querying in Wireless Sensor Networks

In this paper, an adaptive storage method which can dynamically adopt suitable storage mode based on the computation of potential costs according to statistics of event data rate and query rate is proposed. We envision queries for event data following a biased distribution, i.e. Zipf-like distribution, just like the situation in real file sharing systems. In our proposed method, system can switch the storage mode between external storage and data-centric storage adaptively for different types of event data according to their popularity. Simulation results demonstrate that our proposed method is a feasible and energy-efficient data storage method.