Dynamic load balancing mechanism based on cloud storage

Due to extensive coverage and domains of cloud computing-based architecture, most users saves and access highly similar data because of user habits and available resources. Supposing the data stored in the storage devices is of high duplication and similarity, the system administrator cannot guarantee that every node (peer) in the cloud computing system can maintain the best status for data integration. Moreover, when the number of nodes keeps increasing, the number of files will be multiplied, which results to not only the system administrator's heavy workload, but also the waste of hardware resources. Finally, the flexibility and quickness of cloud storage system is gone, and what left is a lot of duplicate and redundant data. Therefore, in order to avoid the system burden caused by duplicate data, we propose novel data center management architecture: Index Name Server (INS), which integrates de-duplication and access point selection optimization techniques to enhance the performance of the cloud storage system. With the attempt to attain load balancing, we use several specific INS parameters to dynamically monitor the parameters, like IP information and busy level index, to avoid the network congestion or the long waiting time during data transmission.