Research On The Storage Of Landuse Vector Data And Spatial Index Based On Main Memory Database

Land use data is the basis and factual of land resources information society services.lt has the large amount of data and complex types. The existing land use management systemstore and manage data with disk resident database, and it can meet the requirementsbasically. But theperformance of the system is no thatsatisfactory on the statistics, query and analysis due to I/O of the disk.In order to meet the demands of real-time, this paper will analyze the existing technology of main memory database and the characteristics of land use vector data, then attempts to apply Redis to store land use vector data and design a spatial index for it.The study focuses on the following three aspects.First, the paper design a vector database and the metadata storage structure base on Redis. The different storage structure is proposed for point features, line features and area features because of their different spatial characteristics. For area features, their spatial information are stored by a series of coordinates or by some known line features according to whether its boundary cari be constituted by the known line features.After studying the structure and characteristics of spatial index R+-tree and main memory database, this paper propose a spatial index which called CR+-tree and design the structure of CR+-tree and storage structure of different nodes types. The space occupied by MBR information in main memory is reduced by compressing MBR information. It has good cache sensitivity and can perform well on main memory database.Finally the paper design and build a simple land use management system base on Redis. After compared the performance of spatial query and statistical summary with those based on traditional database product, the following conclusions can be drawn that the main memory database have a better performance.