摘要
云技术实现了异构硬件资源的有效整合,为向用户提供各种形式的服务奠定了技术基础.同时,由于使用的硬件资源性能、可靠性等差异巨大,所以如何使系统提供高效、高可靠性的服务是该领域研究的热点.本文对常用的可靠性和性能保证策略进行分析比较,探讨了系统能够容忍固定数目个硬件同时失效的必要条件,提出了一种既能保证服务效率,又能最大限度容忍硬件失效的内容存储机制(REST).该机制保证了在系统中同时出现故障的设备数不大于M时,系统仍能正常提供服务.最后,通过理论分析和仿真试验,我们验证了REST在存储效率、容失效能力等方面具有较大的优势.
Cloud computing technology incorporates asymmetric hardware equipments and establishes the basis of providing constant and efficient service to the users .In this paper ,intrigued from the redundancy backup strategy ,we design a cloud storage mechanism which can tolerate hardware or software failures so as to ensure the reliability of the content cloud system while dis -tributing the contents efficiently .Adopting this mechanism ,the system can still operate smoothly in the presence of a fixed number of hardware or software failures ,and if the number of failures is larger than that ,the mechanism can protect the system from mas-sive breakdowns .Afterwards ,through experiments we compare several kinds of storage technologies ,and validate that our technol-ogy takes advantages in storage efficiency and fault tolerance .
出处
《电子学报》
EI
CAS
CSCD
北大核心
2014年第4期633-639,共7页
Acta Electronica Sinica
基金
国家973重点基础研究规划(No.2010CB328105
No.2011CB302703)
国家自然科学基金重点项目(No.60932003
No.61020106002)
国家自然科学基金(No.61071065
No.11171368)
关键词
计算机网络
内容云
可靠性
computer networks
content cloud
reliability