Research On Dynamic Resource Scheduling Approach For SBS Application In Cloud Environment

With the maturation and development of Web technology, Service Based System (SBS) has become an important software development model. In SBS, services deployed on different servers or sets of servers have different QoS, because the services’performance is affected by the capacity of server hardware resource. In order to maintain the performance of SBS, dynamic resource scheduling for the SBS has become a hot topic in current research. Especially, with the development of the Internet technology, cloud computing, an important resource usage patterns, provides effective support for dynamic resource scheduling for SBS now.The current SBS application resource scheduling approaches have many drawbacks. Research on SBS application resource scheduling problem in cloud computing are rare. In cloud computing, Services can occupy the limited private cloud resources provided by the providers of SBS application and the unlimited public cloud resources provided by different providers on the Internet, at the same time, service can also occupy hybrid cloud resources composed by both kinds of the above cloud resources. This requires considering under the condition of unlimited cloud resources and limited resources scheduling cloud resources for SBS application appropriately. In addition, because the user’s accessing behavior is uncertainty and dynamic, the performance of SBS application with user access behavior is dynamically changed. In order to ensure the performance of SBS applications, a dynamic resource scheduling approach is needed.For the problem of resource scheduling for SBS application in cloud environment, this thesis constructs a performance evaluation model of SBS application. Based on this model, the approach of dynamic resource scheduling for SBS application is proposed. To solve the initial resource scheduling problem for SBS application, this thesis proposed a model of initial resource scheduling problem which satisfies SBS application performance constraints, and proposed a initial resource scheduling algorithm for SBS application with both unlimited and limited resource respectively. This algorithm ensures the performance of SBS application by initial resource scheduling. To solve the problem of dynamic resource scheduling of SBS application, this thesis proposed the dynamic scheduling adjustment algorithms with limited resource respectively. These algorithms can increase the throughput of SBS application while the resource is limited. Finally, the thesis designed a series of experiments. By comparing with existing resource scheduling algorithms, the effectiveness of the resource scheduling approach and algorithm proposed in this thesis is proved.