针对静态物化视图集动态适应能力的不足,提出一种动态cache优化算法DCO(dynamic cacheoptimization).它在保持静态算法获取最优物化集能力的基础上,将cache机制直观、快速的动态特性结合进来,以提高数据仓库的动态自适应性能.在cache机制具体实现中提出了一种新颖的空间申请方法,可以充分利用系统剩余空间提高查询响应性能.实验结果在表明算法有效、可行的同时,也显示出该算法可以在一定程度上克服静态物化集存在的空间-性能饱和效应(space-performance saturation effect,简称SPSE),使通过增加物化空间进一步提高数据仓库对查询的响应速度成为可能.
Because the static materialized views lack of better response performance for dynamic query, an optimized algorithm DCO (dynamic cache optimization) is proposed, which generates a dynamic materialized views set to cooperate with the existing static materialized view set by cache. With the assistance of the additional materialized views, the dynamic adaptability and response capability to query increase greatly. Meanwhile, a novel method of allocating the space is presented to provide the alternative for realizing the dynamic cache, and then the free space of system can be used efficiently to store more materialized views for improving the response capability. Experimental results indicate the efficiency and feasibility of DCO, and also show that DCO can overcome the SPSE (space-performance saturation effect) of materialized views in some degree.