A Meta-Heuristic based work Scheduling and Resource Optimization Framework for Cloud Virtual Machines

Abstract

Resource allocation and cloudlet scheduling in cloud computing are challenging issues, particularly when a medium to large number of tasks are involved. The concurrent execution of multiple cloudlets on available resources is important to satisfy user expectations and maximize system performance. Dealing with load balancing patterns and minimizing makespan holds significant importance in this context. To a large extent, task scheduling remains a difficult problem despite advances in cloud computing.Task scheduling is the assignment of user tasks to Virtual Machines (VMs) in such a way as to reduce turnaround time and increase resource utilization. The NP-hard problem complexity of (O(mn)) makes scheduling (n) tasks on (m) devices a challenging problem. The wide range of solutions explored in the task scheduling process highlights the need for algorithms that can return optimal solutions within polynomial runtime.To tackle these issues, this paper presents the Cluster-based Energy Efficient Q-Fruitfly Algorithm (CEQFA) schema for cloud computing systems. Utilizing clustering, the proposed approach groups cloud Virtual Machines (VMs) according to their workloads. It then employs a Q-Fruitfly Algorithm to optimize the assignment of VMs to physical machines. The results show that CEQFA can reduce computation time and energy consumption by up to 6%, while satisfying the Quality of Service (QoS) requirements of cloud users..