基于分布式计算的what-if分析并行处理策略  

Parallel processing strategy for what-if analysis based on distributed computing

作　　者：郑雪梅[1,2] 陈梅[1,2] 李晖[1,2] 

机构地区：[1]贵州大学计算机科学与技术学院,贵州贵阳550025 [2]贵州大学贵州省先进计算与医疗信息服务工程实验室,贵州贵阳550025

出　　处：《微型机与应用》2016年第9期81-84,共4页Microcomputer & Its Applications

基　　金：国家自然科学基金(61462012);贵州省科技计划项目(GY[2014]3018);贵州大学研究生创新基金(研理工2015015)

摘　　要：根据基于OLAP的what-if分析的查询特点,使用分布式并行处理技术解决what-if分析性能较低的问题。以星座模型为基础的what-if分析中,将多维聚集查询分布到不同计算节点进行聚集计算,然后将各个计算节点的聚集计算结果合并输出。该方法根据基于OLAP的what-if分析中其维表远远小于事实表的特性,将事实表中的记录进行水平分片,充分利用各节点计算和I/O处理能力,以解决OLAP查询中计算密集型及I/O消耗过大的难题。在该方法中,随着计算节点数目的增加,其查询时间随之减少,有效地提升了分析效率。According to the query characteristics of OLAP based what-if analysis,using distributed parallel processing technology to solve the problem of low performance of what-if analysis. On the basis of the constellation model,the query with aggregate functions are distributed to each computing node to get aggregate results and final result can be achieved by merging all the aggregate results from multiple computing nodes. The features of OLAP based what-if analysis is that the dimension table is far less than the fact table,using horizontal fragmentation policy to distribute data in multi-node,and making full use of each node computation and I / O processing ability to improve efficiency. In this method,with the increase of the number of computing nodes,the query time is reduced,and the performance is improved significantly.

关 键 词：OLAP WHAT-IF分析 分布式并行处理 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]