基于连接依赖信息的分布式连接查询优化算法

摘要

The characteristics of site dependency algorithm and fragment duplication algorithm in the distributed database are analyzed. The multi⁃join query optimization algorithm based on join dependency information is put forward,in which the join dependency information is used to logically decide whether the join query based on multi⁃site depends on the site to avoid un⁃necessary communication cost,the fragment duplication is used to redistribute the site data to ensure the local join processing to meet the demands of site dependency,and the locality of SQL application and high parallelism of multithreading among different sites are used to reduce the costs of network communication and local calculation. The effectiveness of the algorithm is verified by experimental results.%分析分布式数据库中站点依赖算法和片段复制算法的特性，提出基于连接依赖信息的多连接查询优化算法。该算法中，连接依赖信息用于逻辑判定基于多个站点的连接查询是否对站点依赖，以避免不必要的通信代价；片段复制用于重新分布站点数据，确保局部连接处理满足站点依赖；利用SQL应用的本地性和站点间多线程的高度并行性以缩减网络通信代价和局部计算代价。实验结果证明了该算法的有效性。