基于混合策略的低成本云存储方案

发布时间：2018-03-14 01:44

本文选题：云计算　切入点：云存储　出处：《计算机应用》2014年10期 　论文类型：期刊论文

【摘要】：为了保证云存储系统数据的高可用性、降低数据存储成本和带宽成本、缩短数据对象的访问时间,提出一种称为缓存大小自适应确定(CAROM)的新方案。CAROM结合传统的基于缓存策略的方法和纠错码方法来提高云文件系统的弹性和效率。另外,为了在缓存大小及其效益间实现平衡,提出一种基于总体成本凸函数特性的自适应方法来实现缓存大小的自适应选择。在基于现实世界文件系统数据的性能评估中,CAROM方案的存储成本和带宽成本分别比复制策略和纠错码策略下降60%和43%,同时访问延时与复制策略相当。结果表明,CAROM方案在支持当前云文件系统语义一致性的同时,兼具带宽成本低、存储成本低和访问成本低等特性。
[Abstract]:In order to ensure the high availability of data in cloud storage system, reduce the cost of data storage and bandwidth, and shorten the access time of data objects, A new scheme called cache size adaptive determination (CAROM) is proposed to improve the flexibility and efficiency of the cloud file system. In order to balance the cache size and its benefits, a new scheme called CAROM is proposed to improve the flexibility and efficiency of the cloud file system, which is based on the traditional cache strategy and error-correcting code method. An adaptive method based on the characteristics of total cost convex function is proposed to realize the adaptive selection of cache size. In the performance evaluation of real-world file system data, the storage cost and bandwidth cost of the CAROM scheme are compared. The replication and error-correcting policies decrease by 60% and 43, and the access latency is equivalent to the replication strategy. The results show that the current cloud file system semantic consistency is supported by the CAROM scheme. It has the characteristics of low bandwidth cost, low storage cost and low access cost.
【作者单位】：西北大学信息科学与技术学院;汉中职业技术学院机电工程系;清华大学深圳研究生院;
【基金】：国家自然科学基金资助项目(61371138/F010403)
【分类号】：TP333

【参考文献】