Evaluating cloud computing scheduling algorithms under different environment and scenarios

Scheduling in cloud computing is an essential requirement to execute the tasks given as per the requirements. The scheduler orders the tasks according to the steps mentioned under the scheduling algorithms. The scheduling algorithms need to ensure that the cloud is functioning to the optimum efficiency as well as the tasks are given equal weightage. The way of execution of a scheduling algorithm differs from one another and hence a system, if established, to compare these scheduling algorithms will help the cloud service providers to work better and in a more efficient way as per the user requirements. In fact, it is very important to have a standard comparison system amongst the scheduling algorithms, as per now, there is no particular research done to compare and contrast the scheduling algorithm. This paper develops a set of environments and scenarios in order to differentiate the functioning of the scheduling algorithms. The following scheduling algorithms are studied: First Come First Serve, Shortest Job First, Opportunistic Load Balancing, and Generalized Priority Algorithm. They were put to test under the various environment and scenarios and were evaluated based on parameters such as Cost and Makespan. Based on these studies, an intelligent tool was developed which takes in the user preferences, such as the number of jobs, their preference of time metrics and cost metrics and suggests the scheduler about the number of VMs, and the Scheduling algorithm which should be used to accommodate the user requirements.