Study On Task Scheduling Strategy Based On QoS Constraints In Cloud Environment

With the rapid development and popularization of the network, cloud computing industry appeared. With the feature of low price and easy implementation, cloud computing got an explosive growth in every field. Nevertheless, different users have diversified demand for resources which led to the complexity of scheduling process in cloud computing. Task scheduling as a key issue for cloud computing, its scheduling efficiency has an impact on user satisfaction and efficient of resources, operating costs.At present, a variety of task scheduling algorithms of cloud computing appeared.Some researchers introduced optimal scheduling models and particular algorithms into cloud computing. These algorithms have the pursuit of the optimal time to complete the task, but not having taken account of the satisfaction of the cloud service providers and users. Cloud computing as a new model of economic calculation, its goal is that it can cost less resource and earn more customer satisfaction. The thesis focuses on scheduling algorithm and scheduling model and it had a great significance for making the task scheduling strategy from the perspectives of providers and users.In the thesis, the specific research contents is described as follows:The thesis proposes a task scheduling model based on QoS. In the model: For heterogeneous and distributed resources in the cloud environment, they will be split into multiple virtual machines by virtualization, and they will be gathered and managed.Meanwhile, in order to reduce the matching scope of tasks and resources, a method of division of virtual resource is proposed. In order to obtain the higher matching satisfaction,the thesis proposes a task sorting algorithm based on the matching satisfaction.According to user's different QoS requirements, the thesis proposes a task scheduling algorithm based on the constraint of QoS and the comprehensive matching satisfaction,then realizes the task scheduling strategy based on the constraint of QoS.Finally, the thesis uses Cloud Sim to verify the feasibility of the algorithm. The experiment shows that the scheduling strategy proposed by the thesis could better meet the demands of users and service providers.