Research On Big Vector Data’s High Performance Computing Model And Key Technologies

In recent years, as the collection means getting plentiful and fine management getting intensified of spatial data in China, and the geographic information technology is widely used in variety of fields such as land, forestry, transportation, marine and water conservancy, high-precision, ultra coverage massive vector data growing explosively, the era of big vector data is coming. Quantitative change leads to qualitative change, confront with the demand of storage, management, analysis and computing over one hundred million scale large-scale vector data, the existing GIS parallel computing theory, method and capacity of commercial spatial data processing platform are nearing the limit, geographic information theory and technology facing serious challenges.This paper focuses on the spatial computing of big vector data, to explore the architecture of big vector data’s high performance computing model in terms of spatial cognition, representation model, storage structure, organization and indexing, analysis and computing. Establish spatial computing theoretical model based on the data vitalization theory, design the Key technology and methods of big vector data’s high performance computing. This paper aims at establishing a set of big vector data’s high performance computing model, organizing, computing and showing one hundred million scale large-scale vector data efficiently, providing an effective and comprehensive solution of big vector data’s research and applications. Specific contents are as follows:(1) Investigate the connotations, methods and platform framework of big vector data computing. Summarize the methodology of big vector data computing. Discuss relevant research status, architecture, connotations and methods of cloud GIS and big vector data computing. Summarize the research of efficient data storage management and parallel spatial algorithms on big vector data. Explore the GIS integration framework of big vector data computing.(2) Establish big vector data’s high performance computing model. In hexahedron cognitive model, bring the data vitalization theory into spatial computing, propose the concept model of vector vitalization cell, data vitalization pool, vitalization behavior, the vitalization process, etc., to abstract and present vector data tuple model, multi-state storage and index, high-performance computing paradigm. Then establish big vector data’s high performance computing model.(3) Study the key technologies of big vector data’s high performance computing. For big vector data indexing, analysis and visualization three fundamental issues, in big vector data’s high performance computing model, based on independent super step parallel computing architecture, design and implement high-performance algorithms of in-memory distributed SDGR*-tree index buliding, two-way and multi-way spatial join, and hybrid vector tile pyramid construction. Then evaluate and analyze the performance with the real large-scale vector data.(4) Construct the national basic land data integration and management system. Based on open source projects, establish an integration and management system include massive basic land data’s cloud storage, high-performance spatial computing, data integration and browsing and other business application. Achieve the ability of nationwide one hundred million scale land vector data’s multi-form storage, high-performance spatial analysis and multi-scale viewing.Research and experiment results show that the big vector data’s high performance computing model and key technology in this paper, can effectively combined cloud computing, memory computing and other high-performance computing technologies with spatial computing theory, to achieve high-performance spatial index, spatial analysis and tiled map visualization on big vector data. Finally, through the establishment of a national basic land data integration and management system, the study provides technical support for "Big Data and Big Analysis" of national land, which has important scientific and practical significance.