联网审计系统中海量数据的存储与管理策略

分析了联网审计应用系统的特点,提出了一种适用于计算机联网审计系统的海量数据存储与管理策略,并着重对该系统的基本构成、物理设计与逻辑设计进行了描述.针对本系统中所采用的集群并行数据库系统,还提出了计算机联网审计系统中并行事务处理操作的设计原则,并给出了具体的操作案例.

A Massive Data Storage and Management Strategy for Online Computer-Assisted Audit System

With the development of information technology,an online computer-assisted audit system needs to be established.The audit system would store massive data which was retrieved from variable databases and has different structures.This paper presents a massive data storage and management tactic for online computer-assisted audit system,including system's structure,physical and logical storage design.In the system,a share-nothing cluster is adopted and data is distributed to different nodes of cluster according to table's size.Then,this paper analyzed the executive time of a SQL statement,from user committing to server returning executive results.Based on it,the actual performance of different partition method is compared.The theoretical(analysis) and actual testing results demonstrate that the small size data is suitable to store single node,while the massive size data is more suitable to store several nodes of share-nothing clusters.For providing better performance,some optimal storage methods of massive audit data are provided,including data cluster and join index technology.At last,this paper gives three methods which support parallel transaction process on online computer-assisted audit system using parallel database,including no-limited standard parallel query algorithm,limited parallel query algorithm and rebuilding of parallel tables and indexes.