Research On Replicas Created Strategies For Cloud Storage

With the rapid development of Internet, especially in recent years, the data grows explosively. Now the people’s needing of processing big data has not been satisfied. The emergence of cloud storage brings gospel to people and meet the people demand for big data processing. However, with the cloud storage emerging, some issues, such as system fault-tolerance, data-read efficiency, and data reliability, were emerging. In order to solve these issues, the data replication technologies are introduced to cloud storage. Although some issues of cloud storage could been solved well, some others were been binged, such as replicas consistency maintenance, load balancing. Therefore it is essential and meaningful to research on replicas created strategies for cloud storage.This paper focus on the following aspects,First of all, something related is introduced detailedly, such as the basic concepts and related characteristics of cloud computing, the basic concepts and development status of cloud storage, the relationship of cloud computing and cloud storage, the relationship of cloud storage and replica management, and the workflow of replica management are introduced. More importantly, development status of replica management is highlighted and explained.Secondly, to solve the issues of replica management in cloud storage, RMSim is designed and implemented.Thirdly, to solve the issues of replica management when a file is been creating, a static replica strategy is designed. To achieve the system load balancing and reduce the network delay, the best nodes are chosen to place the replica according to the real-time system information by this strategy. The experiments show that the strategy do well to achieve the load balancing requirements and reduce network latency.Finally, to solve the issues of replica management in the period of file existing, a dynamic replica creating strategy based on the historical data-read information is designed. The results of experiments show that the strategy can improve the efficiency of data reading, improve performance of cloud storage.