实时云计算数据库——数据立方

[摘要] 基于快速发展的并行数据库技术、云计算MapReduce技术及其混合技术，分析了这些技术的优缺点，对并行计算架构、分布式存储系统之上的索引以及其他方面进行了研究，提出了一种被称为数据立方的大数据处理系统。通过与大数据处理系统Hive和HadoopDB的对比实验表明，数据立方的大数据处理系统在入库、查询、并发、扩展等多方面有明显的优势。

[关键词] 云计算；实时；大数据；并行计算

[Abstract] In this paper, we discuss parallel database technology, MapReduce for cloud computing, and hybrid (parallel and MapReduce) technology. We discuss the advantages and disadvantages of all these technologies. We discuss parallel architecture and indexing on distributed storage system. We also discuss other aspects of big-data processing technology and propose a big-data processing system called Datacube. Datacube ios shown to have advantages over Hive and HadoopDB in terms of in query, concurrency, and expansibility.

[Keywords] cloud computing; real-time; large-data; parallel computing