虚拟化数据库系统的资源调度技术分析word论文.docx

Abstract
puting makes puting and storage possible, thus data-intensive applications are increasingly deployed in the cloud. As the cloud technology es widely used, the model of Database as a Service has increasingly receive more attention. In cloud database, in order to improve the resource utilization and ensure the quality of service, it is significant to achieve resource well scheduled and load balanced.
The thesis studies the resource scheduling technology of Sharing DBMS Database Virtualization Platform, and considers the quality of service(QoS, Quality of Service), influence of resource integration, node status and so on. Then we define the QoS, and propose a global design of resource scheduling in database virtualization. We Design a resource integration strategy, which takes account of system status and QoS plete resource initial integration. When dealing with tenants’ requests, we extract the useful pattern to evaluate user’s behavior and node status based on the QoS. Then give a node load evaluate method, and monitor the system in real-time in order to get the node status and find whether the QoS is violated. And optimize the system based on data migration when it is poor. Meanwhile, implement a dynamic schedule strategy, and get the free nodes according to tenant’s database distribution and node evaluation results in order to redistribute the requests. Then allocate VMs based on the QoS and workload plete the schedule tasks in real-time.
Finally, we have functional tests parison tests for the resource scheduling system implemented. The results show that the resource scheduling schema can integrate system resources effectively, schedule service dynamically, and adjust the system by integrating resource again. That is, the goals we set at the beginning have been achieved.
Key words:Database Virtualization,Resource Scheduling,Resource Integration,QoS
目录
摘要..................................................................................................