Spatial Join Query Based On Distributed In-Memory

With the continuous improvement of spatial information acquisition ability,the magnanimity and complexity of spatial data are becoming more and more prominent.Traditional spatial databases are faced with the challenge of managing and storing the growing spatial data,followed by the problems of analysis and processing of spatial data.The complexity of space operations in spatial join query,which is a common spatial analysis operation,is to make large-scale spatial join query becomes a technical difficulty.With the development of computer distributed cluster,distributed memory computing framework provides a new possibility for the analysis and processing of large-scale spatial data.Efficient processing of massive and complex spatial data spatial join query algorithm for the characteristics of distributed memory structure has become an important research direction.Based on the characteristics of distributed memory computing framework,this paper establishes the cost model of spatial connection query in the distributed memory environment.The model can be used to guide spatial join query optimization.This paper designs the key-value spatial data model,which is suitable for spatial data analysis of distributed memory environment.The model has good memory adaptability and good scalability for other spatial databases.Based on Spark distributed memory computing platform,the algorithm of spatial join aggregation and spatial overlay analysis is optimized.In the process of spatial join aggregate algorithm,this paper constructs the multi-level R-tree index,which is based on Hilbert R-tree.And then,this algorithm greatly improvs the efficiency of spatial join aggregation query by using the cache and broadcast in Spark.In the process of spatial overlay analysis algorithm,the mathematical model of spatial overlay analysis is constructed.This algorithm greatly improvs the efficiency of spatial overlay analysis,in which a hybrid grid index based on distributed spatial grid index is constructed.