面向OLAP的高效海量数据存储技术研究与实现

The Study and Implementation of an Effective Massive Data Storage Technique to OLAP

文章在分析了OLAP大规模海量多维数据特点的基础上,针对当前一般规模关系数据库处理海量多维数据能力的不足,通过使用数据库分区技术和并行磁盘组技术,设计并实现了一种高效的、面向OLAP应用的数据存储管理结构。并针对传统查询优化器对于大规模海量数据统计信息更新不及时的问题,设计了一种基于分区技术的统计方法。实验证明,该设计可以对百TB级的多维海量文本数据进行有效的管理。

On the basis on analysis of features of massive multi-dimension data and limitation of traditional relational database on handling massive multi-dimension, we design and implement an effective data storage management structure on OLAP by partition and parallel disk group technique. Focus on problem of statistics of query optimizer on update, we design a statistics method based on partition technique. As a result, we proved that multi-dimension massive text data on hundreds of TB level could be efficiently managed by single RAC through this solution.