Abstract: To the calculation problem to comprehensive load of cloud storage system nodes in the data distribution strategy and system response time, this paper provides a load balancing algorithm to cloud storage system, and validates the algorithm. The algorithm based on the Analytic Hierarchy Process（AHP） can calculate the comprehensive load of storage nodes from the available storage space, available CPU, available memory and access frequency and dispatch data access request. The verification results show, by adjusting weight of different indexes, the algorithm can meet the different application requirements and optimize cloud storage system performance, especially be suitable for high parallel access and big data storage.

Key words: cloud storage, load balancing, Analytic Hierarchy Process（AHP）, big data, comprehensive evaluation, file system

摘要： 针对云存储系统节点在数据分布策略和系统响应时间方面的综合负载计算问题，提出了一种云存储系统的负载均衡算法，并对该算法进行了验证。算法基于层次分析法（Analytic Hierarchy Process，AHP），通过建立综合评估指标体系，从可用存储空间、可用CPU、可用内存和访问热度四个方面，计算各个存储节点的综合负载，并据此对数据存取进行均衡调度。验证结果表明，通过调整不同指标的权重，算法能够很好地满足不同的应用需求，同时，该算法能够很好地反应各节点的综合负载，实现云存储系统整机性能的优化，尤其适用于一些高并发的大数据存储。

关键词: 云存储, 负载均衡, AHP层次分析法, 大数据, 综合评估, 文件系统