Multidimensional Indexing Method And Application Of Spatiotemporal Data For Distributed Storage

As spatial information services gradually transform into spatio-temporal data services,data quality and timeliness are guaranteed.Efficient query and data mining analysis of spatio-temporal big data provides decision-response support for complex scenarios.The current spatio-temporal data management method combined with the big data framework initially solved the data size problem,but the overall research has not fully considered the underlying architecture of the extended database.And in the meanwhile,the data indexing method is not universal.Thus,in this paper,overall architecture combines distributed storage technology with spatio-temporal multi-dimensional index,aiming at establishing efficient and scalable space-time data efficient storage and spatio-temporal index optimization method.Based on the above foundation,a comprehensive scheme of efficient access is constructed to support spatio-temporal data related applications.The main aspects of this paper's research content are as follows:(1)The characteristics of spatio-temporal data and its key issues of efficient storage and parallel processing in distributed environments are analyzed,which based on the distributed computing related technology architecture and data storage characteristics.The HBase spatio-temporal data storage organization model is constructed to reduce the coupling between spatio-temporal data structure and index.Specifically,the storage schema combines data distribution characteristics and the logical organization of hierarchical partitioning.More importantly,it designs a multi-table and multi-index hybrid management data storage mode that provides a highly available data storage structure foundation for efficient querying of spatio-temporal data.(2)The spatio-temporal clustering and full-time characteristics of spatio-temporal data,the feasibility of space-time index is analyzed according to the characteristics of big data technology architecture.Based on the coding and computing ability of S2-Geometry algorithm in geospatial space,this paper constructs a generalized hierarchical grid management spatiotemporal object,and introduces the CompactHilbertlndex algorithm to integrate time information to optimize the spatiotemporal hybrid index value generation method.The indexing method is applied to the data table structure design and optimizes the spatio-temporal range query method through fragmentation sequence analysis to form an overall design scheme for spatiotemporal index in a distributed environment.Finally,through the construction of spatio-temporal big data mining prototype system and the comparison of spatio-temporal index construction efficiency performance,the feasibility and effectiveness of the S2-H3 indexing method are proved,which has been applied to the Ningbo spatio-temporal information cloud platform.The practice shows that the scheme can meet the requirements of efficient management,real-time retrieval and parallel computing of large-scale spatio-temporal data.The method in this paper has good scalability and provides reference for other massive data management modes.