earticle

영어

Existing works mainly focus on indexes in a single server or server-client structure. Unfortunately, such works fail to provide efficiency since performance bottleneck is introduced. This paper designs a two-layered index structure to prune search space among computing nodes for query processing. The index reduces the number of involved computing nodes while query processing, and improves I/O efficiency inside a single server. The initiation method and maintenance methods are proposed for the index, together with optimization strategies for improving query throuphput. This thesis designs the point query algorithm, the range query algorithm and the kNN query algorithm for cloud systems, including distributed algorithms among computing nodes, and optimization strategies inside a single computing node.