Bigtable作为Google云计算的一项关键技术,在需要海量的存储要求的Google地图、GoogleEarth、Gmail、Youtube等上面得到了成功的应用。Bigtable是基于GFS和Chubby开发的分布式存储系统,能够处理Google中海量繁杂的数据类型,也能够将不同应用的数据分布地存储到数千台服务器上。文章介绍了Bigtable的数据模型、设计和实现,并引入了随机Petri网对Bigtable系统的可信性进行模拟和量化分析,提出了云计算环境下Key/value存储系统的发展趋势,并从理论上得出Bigtable系统的高可用性和高可靠性。
As a key technique of Google Cloud Computing, Bigtable has achieved big success implemented in many Google applications such as Google Earth, Gmail and YouTube. Based on GFS and Chubby, Bigtable is a distributed storage system for managing large scale structured data across thousands of commodity servers. In this article, we will briefly introduce data model and design of Bigtable, then we modeling Bigtable using Stochastic Petri Net (SPN) and analyze it’s dependability put forward a cloud computing environment key / value storage system development trend, and draws a conclusion that Bigtable system with high availability and high reliability.