空间数据多服务器群集及高性能实现策略  

Multi-servers Clustering of Spatial Data and Realization Strategies in High Performance

作　　者：严志民[1] 刘仁义[1] 刘南[1] 陆丽珍[1] 

机构地区：[1]浙江大学浙江省GIS重点实验室

出　　处：《遥感学报》2005年第4期421-428,共8页NATIONAL REMOTE SENSING BULLETIN

基　　金：国家自然科学基金(40271087);国家863高科技研究项目(2001AA135180);浙江省自然科学基金(401006)

摘　　要：针对当前海量空间数据网络化管理的特殊要求,分析了改善新一代地理信息系统(GIS)整体性能所需要解决的问题,剖析并引入了群集技术,充分发挥群集技术在高性能运算和高可靠性多服务器架构等方面的技术优势,优化设计了GIS群集多PC服务器体系架构,并结合空间索引分区等常规海量空间数据库管理策略,开发了一套实用化的高效空间定位、多缓冲策略等基于群集的多服务器海量空间数据管理技术。Oracle真正应用群集(RAC)数据库系统在图库管理中的应用,验证了群集空间数据多服务器管理平台的整体性能优越性。Analysis for the need that should be resolved in developing new geographic information system more deeper study and analysis for cluster techniques and Oracle Real Application Cluster (RAC) were taken. To meet special requirement for current management of massive spatial data on net, an optimal multi-level PC servers architecture in GIS with clustering techniques was designed to take full advantages of high performance computing, high reliability and other cluster features. Integrated the reliability, scalability, flexibility and other features of RAC with general spatial database management strategies such as spatial index and partition, remote in-line procedures, mutual-backup of two computers and other strategies, a new type of distributed data management techniques such as high spatial locating, multi-caches strategies cluster based were developed. An application with RAC Servers and Oracle9i application servers (iAs) was tested, which showed the high performance of whole spatial database managing platform, such as high usability, load balance, high data downloading and high efficient spatial data query and anlysis.

关 键 词：计算机群集 地理空间数据 高性能策略 Oracle RAC 

分 类 号：P208[天文地球—地图制图学与地理信息工程] TP392[天文地球—测绘科学与技术]