Tasks Scheduling Algorithm Extended in Hybrid Clouds to minimize Costs

Cloud computing is one of the most superior technologies for applications which need numerous resources and high computational power. The main reason of its superiority is the variety of services which are provided by clouds. Task exchanges and cloud interactions are necessary to provide customers with better and more applicable services. Establishing interactions between clouds in some cases such as job scheduling and distributing the tasks between clouds are essential challenges that might be faced in this field. In this paper a method for scheduling and sending tasks to clouds is presented. Our method aims to achieve minimum execution cost and task failure with respect to task deadlines which results in better performance. In this method arrived tasks are sorted based on their deadline. Execution cost for each cloud is repeatedly calculated. Subsequent to these steps the most appropriate cloud is selected and job will be transferred to it. The algorithm checks order and execution of jobs in proper time periods and changes them if necessary. Since our method determines the priority of jobs based on their deadlines, it maximizes task completion. Moreover, it selects the cheapest cloud and places the highest priority on it. As a result it obtains lowest execution and communication cost.