Implementation and evaluation of I/O efficient range trees

Abstract

Range Tree is a geometrical data structure used to store data in d-dimensions. In order to store larger amount of data in external memory, we need an I/O efficient algorithm. There are many I/O efficient algorithms designed like buffer trees but we need to implement these algorithm designs and evaluate its performance.According to our study, there is no implementation of 2- dimensional I/O efficient range trees available. In this work, we have implemented range trees using buffer technique in 1-dimension and 2-dimensions. Range Tree data structure is used to answer range search queries in an I/O efficient manner. We have evaluated I/O efficient Buff er Range Trees with different set of parameters like buffer size, number of input elements and block size. We did our experiments and evaluation with a randomly generated data set and geographical latitude-longitude data of Open Street Maps.