科研之友
微信
新浪微博
Facebook
分享链接
摘要
Current researches on reliability of storage systems mainly focus on data redundancy methods but rarely pay attention to storage capacity allocation for maximizing system reliability return. In order to overcome the deficiencies of the existing storage resource allocation strategies, a concept called priority density was firstly introduced and was defined as the ratio of file priority to file size. Based on this, a nonlinear integer programming model was proposed. Finally, an optimal resource allocation scheme which could maximize the system reliability return was figured out. Experimental results indicate that this resource scheme compared favorably to the greedy algorithm can attain a much higher system reliability return under the same amount of given storage capacity.
