按需服务是云存储系统有别于传统系统的最显著特征之一。它改变现有产业布局和应用模式，对信息技术发展具有深远的影响。云存储系统的特点是高性价比、高伸缩性和高可用性。其中，高伸缩性是实现云存储系统按需服务的最基本保证。随着云存储系统用户数量增加，规模更大，结构更复杂的发展，系统可伸缩性和可用性面临巨大的挑战。. 本项目针对目前云存储系统伸缩性研究中资源配置策略粒度过大，对存储资源重视不足、评估验证困难、对系统性能影响较大等问题，研究基于非主从式异构的云存储服务平台兼顾性能的伸缩性分析和评估模型以及实现和优化等问题。内容包括:1)文件级的伸缩性管理优化方法。2)测试与计算相结合的系统极限伸缩性的评估与验证方法。3)面向多用户基于预测的预反应式存储资源配置策略。4)基于多维目标的动态伸缩性模型。. 研究成果将为提高云存储系统利用率，高效实现、评估和动态管理系统伸缩性提供科学的理论指导和技术支持。

On-demand service is the key differentia of cloud storage system comparing to the traditional systems, in which it will change the existing industrial layout and application model, and bring deep impact to the development of information technology. The key feature of cloud storage system includes highly cost-effective, highly elasticity and highly availability. Among them the highly elasticity is the foundation for on-demand provisioning of cloud storage system. As the user number rapid raising, the system volume highly expanding, system elasticity and availability are facing huge challenges...This study focuses on the analysis, the evaluation model and the optimization of elasticity to the non-master-slave heterogeneous cloud storage system based on system performance guarantee, in which to addressing the issues in current research, such as the oversize of resource placement granularity, ignoring of the resource types and evaluating and verifying difficulties. The study topic includes: a)the optimization methods for file-level elasticity management. b)the evaluation and testing methods for system limitation in elasticity. c)the multi-tenant-oriented reactive storage resource allocation policy based on forecasting. d)the dynamic elasticity model based on multi-objective...The results of this study will offer theoretical guidance and technical support for the utilization optimization, the evaluation and the elasticity management of cloud storage system.