VegaIndexer: A Distributed composite index scheme for big spatio-temporal sensor data on cloud

Abstract

With the prevalence of data-intensive geospatial applications, massive spatio-temporal sensor data are obtained and the big data have posed grand challenges on existing index methods based on spatial databases due to their intrinsic poor scalability and retrieval efficiency. Motivated by the deficiencies, in this paper, we propose a distributed composite spatio-temporal index scheme called VegaIndexer for efficiently answering queries from large collections of space-time sensor data. Firstly, we present a distributed spatio-temporal indexing architecture based on cloud platform which consists of global index and local index. Moreover, we propose Multi-version Distributed enhanced R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">+</sup> (MDR <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">+</sup> ) tree algorithm for accelerating data retrieval and spatio-temporal query efficiency. Furthermore, we design a MapReduce-based parallel processing approach of batch constructing indices for big spatiotemporal sensor data. In addition, we implement VegaIndexer middleware on top of the leading cloud platform, i.e., Hadoop and associated NoSQL database. The experimental experiments show that VegaIndexer outperforms the index methods of typical spatial databases.