Research On Distributed Spatial Index Algorithm With R*

With the development of spatial information technology, it has been applied to all aspects of society widely, and penetrated into social life. As times goes on, more and more spatial data was survived by spatial information system. So how to retrieval and mining the spatial information efficiently, so the information will not lose value with the passage of time, but will be more and more high value of the data with the passage of time, thus leading to the development of spatial database technology based on the calculation of spatial information is more and more important. With the development of computer hardware technology, single equipment’s computing performance encountered it’s bottlenecks. The distributed computing to improve computing efficiency of data per unit time is an excellent solution in current data calculation. Especially in processing the massive spatial data. In the distributed computing platform of Hadoop, to promote the further development of distributed computing. The Hadoop distributed platform with its query tool Hive and distributed development model-- MapReduce based on these together constitute a good distributed computing scheme.Therefore, how to overcome the calculation of massive spatial data in a distributed platform of high performance Hadoop computational complexity in calculation of space caused by the expansion has become the key. In this paper, starting from the spatial data index theory to realize the distributed index based on MapReduce to realize the high efficiency of massive spatial information processing with distributed development model in MapReduce.The results of this paper are as follows:(1) Based on the current popular algorithm and distributed spatial index. Focus on the analysis of the advantages and disadvantages of R* algorithm. Study on the Hadoop distributed computing framework, focuses on the analysis of the distributed computing framework and calculation principle under the framework of MapReduce.(2) Analysis of high performance R* in spatial index in the query, and its shortcomings can not be applied in the distributed system, puts forward a kind of improved R* algorithm-- DSR*(Distributed Spatial R*) algorithm.(3) In the distributed computing platform Hadoop Sptial, through the experiment, verify the DSR* performance of the algorithm in the Hadoop platform based on spatial index is higher.(4) In MapReduce, using the improved DSR* algorithm, to achieve the possibility of achieving high performance distributed spatial index in the Hadoop platform, and its application in the massive spatial data search.