The emerging model of cloud computing to manage distributed large-scale infrastructure resources is becoming very popular. Entities in both industry and academia are investing a huge time and effort to investigate and develop its features. As more and more applications are becoming data-intensive, one of the most important issues for this promising model is data management and to efficiently handle storage. Cloud storage needs are basically of two kinds, storage support for Virtual Machine (VMs) images and storage support for cloud applications data. In this paper, we propose a cloud storage architecture for Both VMs images and application data. This architecture is designed to allow the system to be self-adaptive and to provide better scalability with a good quality of service. This is achieved by modeling the global behavior of the underlying storage system and acting accordingly. Our design is going to be implemented to provide storage service for OpenNebula cloud platform.