DYSTO - A Dynamic Storage Model for Wireless Sensor Networks

Abstract

One of the main problems in wireless sensor networks (WSN) is the energy consumption needed for storing and querying sensed data, since sensors are devices with very limited resources and battery. One of the main factors that has an impact on energy consumption is the data repository location, that is, the location where the sensor data is stored. In this paper we propose DYSTO, an in-network and dynamic data storage, in which the location of the repository is chosen based on information collected over the network. Intuitively, DYSTO chooses to store sensed data close to where it is most frequently needed: close to the query entry point when the query rate is high, and close to data sources for coping with high sensed data production rates. Moreover, the query load is analyzed in order to keep values that are frequently queried together in the same repository. Our ultimate goal is to reduce the power consumption of the network based on an adaptive approach for repository selection, and a user-defined update policy based on data thresholds. DYSTO has been implemented on NS2 network simulator and our experimental results show that it has less power consumption than similar approaches.