TASK SCHEDULING IN CLOUD COMPUTING

Abstract

Cloud computing is a way through which a user can access all types of resources from the location that is different from the location where his system presents. Basically cloud computing is a combination of grid computing, utility computing (Complete package of computer resources that can be used as a metered service) and autonomic computing (ability to do self management). Job scheduling is one of the most necessary tasks in a cloud computing environment because cost is associated with every resource based upon the time that is used by the user.

Generally such type of task scheduling algorithm that gives optimized solution in cloud computing. Also, the utilization of resources must be efficient, for that purpose, we have to build such type of job scheduling algorithm that does better utilization of resources. There have been various varieties of scheduling algorithm applied in a distributed computing system. By making certain changes, we can apply them in a cloud computing environment. The main goal of the task scheduling algorithm is to maximum resource allocation, gain high computation and the best system throughput. In our proposed task scheduling algorithm, tasks are submitted to the virtual machines by applying particle swarm optimization technique through which maximum number of tasks can be submitted which proves best utilization of resources.

In our algorithm, we vary the number of tasks for execution and virtual machine count for execution. This algorithm is implemented in JAVA platform using eclipse IDE. In our algorithm, we have few constraints that migration of tasks from one virtual machine to other is possible, arrival time of all the tasks are considered to be same. The swarm intelligence helps in convergence of solution finding in the big search space more quickly. Execution of the algorithm results an optimized schedule which provides maximum resource allocation possible in short span of time.