Research On Cloud Storage Simulator And Resource Management Strategies

Cloud storage services primarily offers internet-based online storage services toordinary users and enterprise businesses. Its features are flexible, scalable, pay-as-you-go,cheap, reliable and easy to use and manage; However, to realize the many advantages ofcloud storage services there is still facing many practical problems, especially how to deployand manage the storage resources under complex network environment has become theprimary task of research and developing the current cloud storage systems. therefore the studyof resource scheduling strategies under cloud storage environment has important theoreticaland practical significance.As data-intensive cloud computing tasks, cloud storage services share a lot of similarpoints with cloud computing services, but there has not yet been appeared on the market aneffective simulation tool target for cloud storage environment. Therefore, undertaking cloudstorage simulator related research and development work has also been a imperative.Based on this, this dissertation focuses on the research of cloud storage simulator and theproblem of resource management strategies under cloud storage environment, undertakes aseries of in-depth research, the research work is divided into two main parts:(I) Studied the task scheduling and resource allocation strategy under cloud storageenvironments, the author developed a CloudSim based cloud storage system simulator namedCloudStorageSim. As the same with CloudSim, CloudStorageSim used the event-drivensimulation engine JavaSim, but developed its own data center and virtual machine specific fordata storage. Within the data center, the storage requests is divided into user registration,virtual machine creation, data upload, data download etc. In addition, CloudStorageSim notonly can be used to simulate PaaS, SaaS layer data storage user requests, I t can also simulatethe resource allocation strategies in IaaS layer.(II) Achieved the strategy for cloud storage resource management underCloudStorageSim environment. This paper proves by experiments that CloudStorageSim caneffectively simulate the execution of a variety of data storage related and storage requirementrelated requests under cloud storage environments. As to Proactive models which can predictthe workload, experiments proved that under a relatively gentle load fluctuation, which iseasier to predict, this scheduling mode can pre-start idle storage node and virtual machinemigration work, which can ease the load pressure of the system. Finally, this paper alsointroduces the concept of hybrid cloud environment, namely: to allow the outsourcing ofstorage resources among different private cloud and small cloud storage service provider to resolve the problem of load fluctuations.The author developed a cloud storage simulating system named CloudStorageSim andproposed resource management strategies based on CloudStorageSim environment, providedsome new research ideas and methods for the research of resources scheduling policies underreal cloud storage environments.