在总结了现有并行数据库实现模型的基础上,基于“半重写变换”模型实现了一个并行数据库系统的原型。通过对数据划分/重划分、并行选择、并行排序、并行连接等关键操作的实验分析,指出了“半重写变换”模型存在的缺陷,并提出了一种混合式的改进模型。从理论上说,在机群架构下实现并行数据库系统,这种混合模型较单一模型更有优势。
Based on the analysis of existing implementation models of parallel database system, a prototyped parallel database system implemented on the basis of"Semi-rewrite and Transfer" (SRT) model is described. The experiments focused on some key operations such as data partitioning/re-partitioning, parallel selection, parallel sorting, parallel join revealed that the SRT model had some limitations. A hybrid model is proposed, which is considered as a better method for implementing parallel database systems on cluster architecture than single models.