Research On The Key Technology Of Service-oriented Spatial Data Management

With the rapid development of information technology,geographic information application has entered the era of network information service.And efficient management of massive spatial data is the foundation of the establishment of a unified dynamically extensible distributed spatial information service center,making massive geographic information storage and sharing possible and providing efficient spatial information services.This paper focuses on the key technologies of service oriented distributed spatial data management:1.This paper summarizes the research development of geographic information application and spatial data management,and analyzes the current situation and development trend of the technology of high-efficiency management of massive spatial data in service-oriented environment.This paper puts forward the basic ideas that the research includes multiple massive spatial data integration management model of service oriented design,improvement and innovation of the key technology which is based on the general distributed file system and database,spatial information file system and spatial database management system.2.Based on the general open source distributed file system,considering the characteristics of spatial data,according to the characteristics of map tiles and Pyramid image data which include low file number,small amount of data,data access with the spatial locality,this paper makes improvement and the construction of massive geodata file system(MGFS).According to the global subdivision model of Pyramid encoding tiles,based on geographical documents the location information stored in sorted order,this paper also makes the adjacent spaces adjacent to the data storage,and makes use of the small file distribution aggregation and indexing technology to provide efficient storage of massive files and concurrent access.3.Based on the massive geodata file system(MGFS),this paper designs the semi structured spatial data storage engine(MGBase).Based on the relation database,this paper puts forward the structure of the distributed spatial database system model and system structure based on "MGFS + RDBMS + MGBase" and establishes a distributed spatial database prototype system of spatial information.Considering the spatial data usage scenarios,this paper designs the Row Key encoding principle of spatial data and spatial data of the physical storage optimizatio,providing a structured,semi-structured and unstructured data of multiple space efficient storage management and concurrent access.4.This paper analyzes the research status of distributed spatial data model and the application of spatial data,and puts forward the requirement of design of massive geodata model.According to the multivariate spatial data of multiple data types,multi scale,multi version and a variety of coordinate system,the corresponding data model are designed based on different data analysis of spatial information service requirements.Based on the global unified subdivision encoding,using hierarchical method by layer+block,this paper constructs multiple spatial data integrated data model in order to achieve efficient distributed management of massive multiple spatial data,providing a large number of concurrent users high performance access to massive amounts of data for the upper space information network service.5.This paper makes analysis of the status quo of spatial data index technology.Combined with the spatial data partitioning strategy and distributed spatial data organization model of Global Subdivision encoding,this paper constructs a distributed spatial data organization model including data sheet,data layer,data block and index in data block,and makes the index concurrency created and high concurrency conditions efficient data access possible.6.As for the demand of the spatial data in parallel storage,index creation process to create an empty layer,global unified metadata information and target ID generatio,based on MapReduce parallel processing model,a new Job generation and submit interface framework and the Joint module is added,which meets the coordination of resources demand between each TaskTracker during Map,Reduce stage,and gets the massive spatial data parallel processing framework.7.This paper carries out the design of prototype system,which includes the experimental process including data storage,data query,updating local content and carries out spatial data loading,concurrent access and concurrent processing,local updating ability test,verify management ability of the distributed massive spatial data.And on this basis,this paper carries out the typical spatial information service experiment based on Surveying and mapping information service system the demand of the construction of Surveying and mapping information and Xinjiang data center.