适合异构数据的数据库引擎设计和实现
首发时间:2012-02-09
摘要:在网络信息化和大数据时代,如何高效地管理各类不同数据格式的数据文件是件很重要的事情。由于具有高性能的并行数据库不具有扩展性、容错性,而基于MapReduce的系统在处理结构化数据时正好相反,性能不足。本文通过修改并行数据库的引擎,在保证高性能的前提下,引入扩展性和容错性。实验结果表明,新的混合的并行数据库可以在性能,扩展性和容错性之间做到很好的平衡。
关键词: 并行数据库 大数据 高性能 扩展性 容错性 MapReduce
For information in English, please click here
Heterogeneous Data Structure Based Database Engine Designment and Implement
Abstract:In the erea of internet and big data , it is a vital important thing to how to manage different data effectively. Usually, data should be not only scaled dynamically, but be analysed fastly and effectively. Tranditianal parallel Database has high performance but lack scalability and fault tolerance. MapReduce-based systems are just on the contrast when structured data are processed. This paper hopes to introduce scalability and fault tolerance into parallel database by modifying its engine on the premise of guaranteeing high performance. Experimental results show that the new hybrid parallel database has a good balance between performance, scalability and fault tolerance.
Keywords: parallel Database big data high performance scalability fault tolerance MapReduce
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
适合异构数据的数据库引擎设计和实现
评论
全部评论0/1000