中文摘要：

针对当前云存储系统副本策略对商业利益追求的考虑不够充分，构建基于移动agent的代价驱动的云端存储模型，引入商业相关概念并给出定义及计算公式，提出代价驱动的移动agent自适应策略。移动移动agent携带数据及副本以用户良好体验及高“收益”原则进行决策计算及判断在云平台异构节点间移动、复制或销毁，实现商业利益最大化的动态存储。给出移动agent自适应动作的算法实现。实验结果表明，该模型在负载均衡和“收益”方面都优于传统静态副本策略。

英文摘要：

Considering replica strategy of cloud storage solutions, the cloud service confronts challenges to pursue business benefit. This paper presents a cost-driven cloud storage model based on mobile agent, introduces business concepts and expressions, proposes an adaptive mobile agent strategy. Mobile agents with replicas move, copy or destroy in the heterogeneous cloud platform, so the cloud storage can improve load balance and maximize its profits. Last the algorithm is presented. Experiments illustrate that this model is superior to traditional static replica strategy at both load balance and profits.