作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2013, Vol. 39 ›› Issue (4): 32-35. doi: 10.3969/j.issn.1000-3428.2013.04.008

• 先进计算与数据处理 • 上一篇    下一篇

基于固态硬盘的云存储分布式缓存策略

李东阳,刘 鹏,丁 科,田浪军   

  1. (解放军理工大学指挥自动化学院,南京 210007)
  • 收稿日期:2012-06-11 出版日期:2013-04-15 发布日期:2013-04-12
  • 作者简介:李东阳(1985-),男,硕士、CCF会员,主研方向:分布式系统,网格计算;刘 鹏,教授、博士生导师;丁 科,讲师、博士;田浪军,硕士
  • 基金资助:
    工业和信息化部电子信息产业发展基金资助项目“云安全存储系统关键技术研究”(KYHXRJ021101)

Distributed Cache Strategy in Cloud Storage Based on Solid State Disk

LI Dong-yang, LIU Peng, DING Ke, TIAN Lang-jun   

  1. (Institute of Command Automation, University of Science and Technology of PLA, Nanjing 210007, China)
  • Received:2012-06-11 Online:2013-04-15 Published:2013-04-12

摘要: 为满足海量数据存储的需求,提出一种基于低功耗、高性能固态硬盘的云存储系统分布式缓存策略。该策略对不同存储介质的硬盘虚拟化,将热点访问数据的缓存与存储相结合,实现在不同存储介质之间的热点数据迁移,解决热点元数据的访问一致性与存储服务器的动态负载均衡问题。工作负载压力测试结果表明,该策略可使云存储系统的读峰值速率最高提升约86%,并且能提高存储服务器的吞吐量。

关键词: 云存储, 固态硬盘, 分布式缓存, 热点数据, 数据迁移, 负载均衡

Abstract: In order to meet the needs of mass data storage, this paper proposes a distributed cache strategy in cloud storage based on Solid State Disk(SSD) and low power consumption. This strategy achieves the virtualization among different storage mediums, combines the cache and storage of the hot spot data together, and migrates the hot spot data among different storage mediums. The strategy also resolves the problems of the consistency of the hot spot metadata and the dynamic load balancing among the storage servers. Load pressure test indicates that the distributed cache strategy in cloud storage system based on SSD can improve the read peak rate by about 86%, and improves the system throughput effectively.

Key words: cloud storage, Solid State Disk(SSD), distributed cache, hot spot data, data migration, load balancing

中图分类号: