SLA-based Multi-task User Request Resource Scheduling And Its Profit Model And Related Algorithms In Cloud Environment

Cloud computing as a new service of business computing model, connected the computers that distributing in different geographical positions with the network. Through virtualization and a series of other technologies, these hardware facilities formed a huge pool of resources. It helps to simplify the management of these hardware and it also provides the resources to customers in a "pay as you go" model. With the continuous development of cloud computing and the population of cloud computing users explosion, it puts forward more stringent requirements to the cloud computing service providers in how to efficiently and quickly to the resources scheduling, distributing and management. After all, with more and more service providers have emerged, the requirements of cloud service of customers are also getting higher and higher. Therefore, how to ensure the quality and improve the satisfaction of users of the service has become an important bargaining chip of the cloud service providers in competing with others.In view of two points above, this article proposed a hybrid cloud service model based on the Service-Level Agreement(SLA) from the perspectives of cloud service providers and the customers. Then this article analyses the profit models in the situation of single task in users' requests according to the roles of different service providers and the formal definitions of the profit models are given. Based on the situation of single task in users' requests, this article given a prolongation of multi tasks in users' requests, a detail analyzation and formal definitions of different service provider roles are given, too.Following the problem of users' requests admission control, this article proposed a priority of users'request setting strategy, and it helps to add the proportion of users' requests that been accepted. And this article proposed a priority of tasks setting strategy and a delay compensation strategy for the possible delay conditions in the process of the sub tasks executioa These tow strategies aimed at cutting down the SLA breach of contract as much as possible in the process of sub tasks execution and ensuring the quality of service and to improve the users'satisfaction of the cloud computing service by making the compensation to the users when it occurrences the SLA breach.In view of the point of the resource scheduling, this article proposed a resource scheduling algorithm. This algorithm based on tow resource scheduling strategies that aimed at maximizing profit by minimizing cost and improving the user satisfaction level.At last, this article designed several detail simulation experiments and comparative experiments to exam the validity and the function of the algorithm proposed above. The results of the experiments showed that the algorithm can reduce the usage of resources by improving the resources utilization, and it also helped the resource providers to create much more profits than the comparative algorithms to a certain extent. And it also proved that the research of this article has achieved the original purpose.