基于Google Bigtable的海量数据存储探索

Big Data Storage Research Based on Google Bigtable

随着网络技术的快速发展,存储网络中的海量数据已经超越了传统关系型数据库的负载能力.如何存储海量数据,以及如何基于海量数据提供高效的数据查询的能力,使得程序的使用者能够得到及时的回应等诸多问题是Google等网络服务供应商们所亟需解决的挑战.为了解决这些问题,Google研发了Google文件系统(Google File System,GFS)、Bigtable以及很多其他相关的技术和算法.本文介绍了Google Bigtable的数据模型,并且详细解释了Bigtable是如何提供可扩展性,如何提供高效率的读和写操作,以及Bigtable是如何控制并发事务的.读者通过阅读可以更加深刻地理解Bigtable的技术架构.

Along with the rapid development of internet, the massive data stored in Internet has exceeded the storing capacity of traditional relational databases. Questions, like how to store big data, and how to provide efficient querying abilities in the context of big data so that application users can receive instant responses and etc, are the emergent issues waiting to be solved. To deal with these challenges, Google develops GFS （ Google File System）, Bigtable and many other technologies and algorithms. In this paper, I will explain the data model of Google Bigtable, and describe how Bigtable supports scalability, how Bigtable supports high performance of reading and writing, and how Bigtable controls the concurrent transactions and so on. Contribution of this paper is that readers can gain a deeper insight into Bigtable architecture through reviewing the examples given in each section.