An Energy Efficient Filtering Approach to In-Network Join Processing in Sensor Network Databases

Abstract

In processing join queries in a sensor network it is important to minimize communication cost, which is the main consumer of battery power, among sensor nodes. Shipping only relevant data involved in join result to the sensor nodes that is responsible for the final join means reduction in the size of data transmitted and hence lowers battery consumption. In this paper, we present algorithm SBJ which uses bloom filter and algorithm RFB which uses bit vector to filter unnecessary data during join processing in a sensor database. We compare our algorithms with the existing synopsis join (SNJ) algorithm in terms of the communication cost associated with processing BEJ (Binary Equi-Join) queries. The query performance results show that both the SBJ and the RFB algorithms outperform the SNJ algorithm. In addition, the RFB algorithm generally performs better than the SBJ algorithm.