期刊文献+

小规模集群文件系统中两级元数据服务器的设计与实现

Design and Implementation of Two-level Metadata Server in Small-scale Cluster File System
下载PDF
导出
摘要 在集群文件系统中,元数据服务器是整个系统正常运转的核心,它的可靠性和性能是设计系统时需要着重考虑的问题之一.本文设计了一个具有高可靠性、高性能的两级元数据服务器系统,兼顾了集中式元数据管理和分布式元数据管理的优点.系统中高级元数据服务器负责维护文件系统全局的目录结构和管理整个文件系统的命名空间,双元数据服务器负责维护文件元数据的分布信息,并采用了马尔可夫回报模型对两级元数据服务器系统进行了可靠性分析.实验数据表明,具有两级元数据服务器的集群文件系统能提供高吞吐量. The metadata server is the heart of the regular operation system in cluster file system. So the reliability and performance is one of the important problems when being designed. Presented our design of a two-level metadata server system, which behaves high reliability and performance. The merits both centralized management and distributed management are considered simultaneously in our design. In this system, the advaneed-metadata server is responsible for maintaining a global file directory and the whole file namespaee. The double-metadata server is responsible for file metadata allocation. And this paper uses the Markov return model to analyze the reliability of the two-level metadata server. The experiment data indicates that the design can provide high throughput.
出处 《小型微型计算机系统》 CSCD 北大核心 2007年第6期1026-1030,共5页 Journal of Chinese Computer Systems
基金 河北省产业化基金项目(020501)资助
关键词 可靠性 两级元数据服务器 马尔可夫回报模型 reliability two-level metadata server Markov return model performance
  • 相关文献

参考文献3

二级参考文献33

  • 1Gunter Bolch, Stefan Greiner, Hermann de Meer, et al. Queueing Networks and Markov Chains[M]. John Wiley & Sons,1998.
  • 2Qin Xin, Ethan L Miller. Reliability Mechanisms for Very Large Storage Systems[A].20th IEEE / 11th NASA Goddard Conf on Mass Storage Systems and Technologies[C].2003.
  • 3Peter J Braam. The Lustre Storage Architecture[M]. Cluster File Systems,Inc,2002.
  • 4Scott A Brandt, Ethan L Miller, Darrell D E,et al. Efficient Metadata Management in Large Distributed Storage Systems[A].The 20th IEEE / 11th NASA Goddard Conf on Mass Storage Systems and Technologies[C].2003.
  • 5A Scott, L BrandtEthan, D E MillerDarrell,et al. Efficient Metadata Management in Large Distributed Storage Systems [A].20th IEEE / 11th NASA Goddard Conf on Mass Storage Systems and Technologies[C]. 2003.
  • 6Peter J Braam.The Lustre Storage Architecture[M]. Cluster File Systems Inc, 2002.
  • 7Brian Pawlowski, Spencer Shepler, Carl Beame, et al. The NFS Version 4 Protocol[EB/OL]. http://www. citi. umich.edu/projects/nfsv4, 2002-05.
  • 8Zhou Feng,Jin Chao,Wu Yinghui,et al. TODS: Cluster Object Storage Platform Designed for Scalable Services [EB/OL]. http://www. cs. berkeley. edu/-2f/papers/zf-jc-todsica3pp. pdf, 2000-05.
  • 9Qin Xin, Ethan L Miller. Reliability Mechanisms for Very Large Storage Systems[A]. 20th IEEE / 11th NASA Goddard Conf on Mass Storage Systems and Technologies[C].2003.
  • 10J. Menon, D. A. Pease, R. Rees, et al. IBM storage tank-A heterogeneous scalable SAN file system. IBM Systems Journal,2003, 42(2): 250~267

共引文献12

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部