The Research And Implement Of Elastic Scaling In Hybrid Cloud

Since Amazon brought out its public cloud service platform, Amazon Elastic Compute Cloud, in 2006, the technology of cloud computing has been developed rapidly. Its so-called pay-as-you-go pricing model is attractive, and virtualization technology which cloud computing bases on provides massive convenience for enterprise. The famous three products of Google, Google File System, MapReduce and Big Table, became the foundation of cloud’s development. But information security on public cloud was challenged and criticized. Thus, private cloud was proposed. Enterprise started to deploy private cloud solution onto their own infrastructures and finally embraced the benefits of cloud.After public cloud and private cloud, the industry starts to focus on hybrid cloud and is determined that it is the future of cloud computing. However, the existing hybrid cloud solutions demand the user to use their product on both sides. Such monopolism would bring users extra labor costs. They will have to shift their own environments to the demanding cloud environment and then to use the hybrid cloud product. More generally, the natural heterogeneity between clouds makes it nearly impossible to be compatible with others. Immature technology, insufficient communication methods and lack of available and provable managing tools all increase difficulty for the clouds to work with each other.To solve these problems, this paper presents a system called Burstic which will help users to scale their application from private to public. Burstic aims to do little modification to existing environment. It provides autonomy to help user’s application run without excessive intervention. It also provides rich configuration entries to help users adapt to the system.How to decrease the extra cost for adapting to Burstic is the focus of the designation. How to make a suitable scaling plan and make it happen onto the public cloud is the fundamental parts of this paper.To verify the effectiveness of the Burstic system, we generated a workload to the experiment environment and did some tests to evaluate it. It proved that Burstic can help user’s application scale to the public cloud effectively and the scaling plan it made proved to be reasonable.