Research On User-oriented Software Defined Cloud Storage Resource Allocation Strategy

With the continuous promotion and application of cloud computing, the generation and processing of massive data, making the field of cloud storage more and more attention. Cloud storage inherits the characteristics of cloud computing, reliability, availability, easy to expand, on-demand use on-demand payment. Cloud storage is an important part of cloud computing, using infrastructure virtualization, grid computing, distributed storage and other technologies to provide a reliable data storage services.The traditional cloud storage service is mainly for users or enterprises to upload data to a cloud storage service providers. So there will be some risks,such as cloud storage service providers data locking, data leakage and other security risks. On the other hand, a cloud storage service providers can provide limited services, not flexible enough and so on. Moreover, with the rise of software defined network, the formation and selection of the network is more flexible, and the concept of cloud storage based on software definition has gradually become a hot research field.In this paper, the traditional cloud storage is researched, mainly from the cloud storage architecture and cloud storage strategy. Then analyzed the advantages of existing cloud storage system and the current problems. Then, in order to solve these problems, we propose a software defined structure of cloud storage, usage scenarios and processes of the users uploading files under this framework. The system helps the users to define the storage strategy to save the upload time and storage cost before the users upload the files. The specific work and innovation of this paper are as follows:(1) On the basis of studying the traditional cloud storage architecture,combining with the concept of software definition network and the concept of multi-cloud data center, a user-oriented software defined cloud storage architecture is proposed. This architecture includes multiple cloud storage providers and software defined controllers. Then, according to the process of storing the files and the pain points, the model of the user storing files is proposed. In order to help users to develop saving time and storage cost savings strategy, we take the bandwidth and price of various cloud service providers into account. Then measure storage strategy performance through the specific data.(2) The objective formula proposed in this paper is a nonlinear programming problem. The traditional solution is very complex. In this study,the objective function is transformed into a graph theory by mathematical modeling. By using the algorithm of minimum cost flow in graph theory, the minimum storage cost corresponding to each upload time is obtained by controlling the variables. The relationship between the upload time and the objective function is established. The optimal solution is found by simulation,and then the optimal solution is transformed into the strategy of storing the file.(3) Compare the proposed storage strategy with other storage strategies and carry out experiments from three aspects: the whole performance, the upload time and the storage cost. In order to be universal in parameter setting,random data of normal distribution is adopted, and the average effect is obtained by a plurality of experiments. The comparison shows that the proposed user storage strategy can perform well in the three aspects of the whole performance, the upload time and the storage cost, and can effectively save the upload time and storage cost for users.