Research On Dynamic Resource Scheduling For Cloud Computing Based On QoS

With the rapid development of electronic equipment, technology products, big data and artificial intelligence in recent years, people's demand for computing power is rising. At the same time, the rapid increase in hardware computing power has also brought the price of the server, so a shared computer processing resources and data, the concept of cloud computing is proposed for the theoretical study and business applications.Cloud computing is a kind of resource service mode provided by the way of Internet access, like water and electricity, with the obvious characteristics of on-demand supply. Cloud computing is based on virtualization technology , cloud computing data center virtualization resources can be divided into virtual resources such as computing,networking, storage and other services to users, the user has a natural sensitivity for the quality of service QoS. As for the management and scheduling of these virtualized resources, they are the key to guarantee the quality of QoS services, the academic and industry are studying the management and scheduling of virtual resources under the premise of guaranteeing QoS from multiple perspectives.QoS is playing an increasingly important role in IaaS-based cloud computing, especially for private cloud. The limitation of physical resources of private cloud requires a new service model to dynamically schedule the available resources for better QoS. This paper proposes a service model framework for dynamic scheduling of cloud computing resources based on QoS guarantee for laaS private cloud. The model framework can implement different resource scheduling tasks by loading different policies. Then, the experimental prototype is designed for load balancing, automatic scaling, automatic recovery and fault tolerance.In this paper, three experiments are designed to verify the above four QoS indicators, which are elastic model experiment, automatic recovery model experiment and fault-tolerant model experiment. The elastic model experiment realizes the elastic scalability of virtual machine cluster and nodes failover in cluster, meanwhile an automatic environment deploy script is provided. The automatic recovery model about the virtual machine evacuates in the case of physical machine failure is achieved. Fault-tolerant model experiment to achieve the main server and backup server, dual-system hot backup succeeds as well. The results of the experiments confirm the effectiveness and efficiency of the QoS service model.