QoS based dynamic task scheduling in IaaS cloud

Abstract

Cloud Computing is a type of parallel and distributed system consisting of a collection of interconnected and virtualized computers that are dynamically provisioned and presented as one or more unified computing resources between the service provider and consumers. In cloud computing, many tasks are to be executed by the available services to achieve better performance, minimum total time for completion, shortest response time, utilization of resources etc. Because of these different intentions, we need to propose a scheduling algorithm to perform appropriate allocation map of tasks on resources. In existing system task scheduling algorithm have been designed based on priority and total completion time in cloud computing. The task scheduling algorithm first computes the priority of the tasks based on the inputs of the users and then sorts the tasks by priority. Second, this algorithm calculates the minimum completion time of all the tasks on different resources and schedules onto a resources accordingly. The drawbacks in existing system are, it does not effectively use the idle resources. In this paper we proposed a dynamic scheduling algorithm that efficiently uses the idle time of resources from monitoring the task timing information on resources. The multi-dimensional cost matrix table is developed based on execution time, CPU usage of each tasks and current CPU usage of resources and also we have extended the deadline time value using min-max policies to complete the tasks within a earlier time period. In this paper, we have considered deadline, idle time and reliability as QoS parameters for scheduling.