大数据分析的分布式MOLAP技术

大数据的规模效应给数据存储、管理以及数据分析带来了极大的挑战，学界和业界广泛采用分布式文件系统和MapReduce编程模型来应对这一挑战.提出了大数据环境中一种基于Hadoop分布式文件系统（HDFS）和MapReduce编程模型的分布式MOLAP技术，称为DOLAP（distributed OLAP）.DOLAP采用一种特殊的多维模型完成维和度量的映射；采用维编码和遍历算法实现维层次上的上卷下钻操作；采用数据分块和线性化算法将维和度量保存在分布式文件系统中；采用数据块选择算法优化OLAP的性能；采用MapReduce编程模型实现OLAP操作.描述了DOLAP在科学数据分析的应用案例，并与主流的非关系数据库系统进行性能对比.实验结果表明，尽管数据装载性能略显不足，但DOLAP的性能要优于基于HBase，Hive，HadoopDB，OLAP4Cloud等主流非关系数据库系统实现的OLAP性能.

Distributed MOLAP Technique for Big Data Analysis

To address the new challenges that big data has brought on data storage, management and analysis, distributed file systems and MapReduce programming model have been widely adopted in both industry and academia. This paper proposes a distributed MOLAP technique, named DOLAP (distributed OLAP), based on Hadoop distributed file system (HDFS) and MapReduce program model. DOLAP adopts the specified multidimensional model to map the dimensions and the measures. It comprises the dimension coding and traverse algorithm to achieve the roll up operation on dimension hierarchy, the partition and linearization algorithm to store dimensions and measures, the chunk selection strategy to optimize OLAP performance, and MapReduce to execute OLAP. In addition, the paper describes the application case of the scientific data analysis and compares DOLAP performance with other dominate non-relational data management systems. Experimental results show that huge dominance in OLAP performance of the DOLAP technique over an acceptable performance lose in data loading.