摘要
在云计算的催生下,基于不同网络拓扑结构和数据传输机制的分布式云存储系统应运而生,系统特性不同,应用场合也不尽相同。本文从是否存在元数据中心节点的角度出发,以有元数据节点的MooseFS和无元数据节点的GlusterFS为代表,全面分析两种系统架构、工作机制、系统特点等,并对两种系统进行不同文件类型的读写测试,给出在现有硬件基础上选择适合的文件系统的方法,从而使云存储系统数据处理性能达到最优。
Due to the improvement of cloud computing, distributed cloud storage systems based on different network topologies and data transmission mechanisms are arising. The different characteristics of systems lead to the different applications. This paper, using GlusterFS with no metadata node and MooseFS with a metadata nodes as representative, shows a comprehensive analysis of the system architectures, working mechanisms, system features, etc., distinguished from whether the system contains a metadata node .Based on the ability test, this paper then gives the solutions of how to choose the most appropriate file system on the basis of existing hardware, which makes the cloud storage system achieve optimal performance.
出处
《电子设计工程》
2015年第9期34-38,共5页
Electronic Design Engineering
