Improving the Performance of Spatial Index Based on Structured Data Representation

The spatial index takes very important role on accelerating the spatial query processing. The index structure for the spatial space is made from two processes. The first one called filtering uses abstract figure and extracts candidates which may involve user wanted data quickly. And the next, refinement distills those candidate using exact data representation taken from the databases. The refinement process needs very large I/O, so removing unnecessary candidates which do not meet the user\u27s requirements is the key to gain better performance. At first we point out the problem of R^*-tree which prevents the index from filtering more candidates off. Then, we propose new technique to operate on that problem and show its performance with simulations