Research Data Storage Index Mechanism Massive GML Space Ambient Cloud

GML as a encoding specification of geographic information,is being consummate constantlyand be widely used in the area of data interchange、data integration、data storage、datamanagement、Mobile GIS and LBS,therefore, A large number of GML spatial data Constantlyemerging. on the other hand,with the field of geographic information solving problem scalely、complicately, Need higher requirements to the storage and calculation of the computing platform.Traditional single-machine environment encountered a lot of bottlenecks in process mass GMLspace data.in resent years, Rising of cloud computing technology in academia,providing a newway for GML spatial data efficient management. How to store、manage and apply the hugeamounts of GML spatial data has became a hot issues currently.Oranizating spatial dataproperly、 designing efficient spatial indexing is the basic of spatial query and spatialanalysis.so,the paper research on the huge numbers of GML spatial data storage and indexingmechanism in cloud computing has important significance.The paper is based on the current GML storage theory, Research on GML spatial datastorage in cloud computing environment and the theory of indexing building. Use Hadoop opensource cloud computing platform to implement and test efficiency,the Mainly research contentincludes the following several aspects.1）Deeply research on GML spatial data model, Characteristics of the data and comparingwith the traditional spatial data.deeply research on theory and technology related to Cloudcomputing, laying the theoretical basis for GML spatial data stored and management in cloudcomputing environment.2）Use Hadoop open source cloud computing platform,deeply study its HDFS、MapReduceparallel programming framework and the architecture in the distributed file system of the columndatabase of HBase,aiming at the semi-structured characteristics of GML spatial data, puttingforward a kind of GML spatial data storage in the HBase data model. To test the performance ofHBase distributed storage database, the paper made test on the speed of data writing in HBASEunder the different numbers of nodes of Cluster environment,test proving,with the increasing ofthe node number,the data writting speed speeding up.and did the experiment with the tradionaldatabase,turn out that,in the case of a large amount of data, HBase parallel storage of GMLspatial data, which compared to the traditional spatial database has greater performanceadvantages.3）Research on exploring GML spatial data parallel spatial indexing in the cloud computing environment: this paper combining with the existing spatial data partitioning strategy and parallelspatial index algorithm, putting forward a suitable GML spatial data partitioning strategy anddesigning a multi-level parallel spatial index structure based on it.finally, carried out a simplespatial query experiment, under the scheme of different query boundaries and different number ofnodes to test the query efficiency of index.it turned out to be that the proposed parallel indexsystem has good performance in terms of spatial query.