Nature inspired preemptive task scheduling for load balancing in cloud datacenter

Abstract

Scheduling of the tasks in the cloud computing is assigning tasks to the particular machine. Load balancing of preemptive independent tasks on the virtual machine (VMs) is an important aspect of the task scheduling in clouds. Load balancing aims to optimize resource, minimize response time, maximize throughput and avoid overload of any of the resources. If the load balancing is not properly done, then it may lead to the condition where some virtual machines may get overloaded while others may become underloaded or even idle. Due to unbalanced load, the tasks are not completed in time and as a result the user satisfaction is not achieved. The proposed algorithm almost reduces makespan through preemptive task scheduling which follows the foraging behavior of honey bees. This algorithm considers the priority of the tasks and their expected remaining completion time with an aim to maximize throughput and minimize latency. Our approach improves the response time of users by effective utilization of available resources.