Research Of Replication Placement Strategy In Cloud Storage Based On Bidding Model

In the past20years, the rapid development of the Internet profoundly affects every enterprise’s operation and people’s daily life. So, cloud computing is an inevitable result of the development of the Internet. Cloud storage is born with the generation of cloud computing. The storage equipments in cloud storage are distributed in all over the world, they are packaged up by an application software as a service, so users can use the service for business visit and data storage.As a new service, cloud storage has a lot of advantages, such as:easy management, conservation of resources, and supporting the mass storage and dynamic expansion and so on. Despite cloud storage has a lot of advantages, the data in the cloud storage which users visit is stored in the clouds, and the physical location of the data is not in control, this will necessarily bring hidden trouble on data safety, reliability and stability. According to the survey, many users worry about the safety, reliability and usability of cloud storage, in this paper, the usability refers to the response speed of cloud storage system for users.If not strengthen the safety and reliability of the cloud storage, the two aspects will become the bottleneck of the development of cloud storage. A key technology of the cloud storage system is the distributed file system, safety of the distributed file system mainly uses the data encryption mechanism to realize, and reliability and usability mainly use the replication technology to realize, so this paper chooses the replication technology of cloud storage as the research objectives to discuss how to strengthen the reliability and availability of cloud storage.The dynamic of Cloud storage environment means the dynamic replication strategy is more suitable, but at the same time the dynamic replication strategy also has caused a lot of new problems. Such as:the number of the replication, the placement of the replication. A good replication place strategy can improve the access speed of users, realize the load balancing, and avoid the replications migration in the late migration as far as possible.This paper put forward a mixed topology structure which combines the advantages of cloud computing and P2P in the third chapter, the file system is similar with GFS, this paper uses a Topsis method for initial replication allocation, and based on this, in the fourth chapter This paper put forward a new replication placement strategy based on bidding model, this strategy is established in the dynamic adjustment stage, the simulation experiment results show that the strategy based on bidding model not only considers the consistency and usability, but also makes a good effect in the load balance. Along with the increase of data, the strengthening of the load is changing quite gentle, this fully explain the advantages of bidding strategies.