Spatial indexing for massively update intensive applications

Abstract

Supporting high update throughput is important to provide real-time location information for many applications, which involve moving objects, such as security, military, and environmental monitoring. We present an R-tree-based index structure with a family of update strategies for supporting high update throughput. It groups updates in the update buffer, and bulk-applies them into the R-tree. Extensive empirical studies have shown that it outperforms existing techniques by 2-5 times. In our experimental environment, moreover, it is able to provide the real-time update capability, while its competitors fail to do so. (C) 2012 Elsevier Inc. All rights reserved.