An Adaptive Swarm Optimization Technique for Load Balancing and Task Scheduling in Cloud Computing

Abstract

In recent era of growing technology, cloud computing has attracted various industries and researchers due to it significant nature in parallel and distributed computing systems. This technology has grown drastically in various real-time applications such as medical field, health organizations, multi-media applications etc. cloud computing urge to provide the better quality of service for clients or users. However, increasing demand of application leads to the imbalance in cloud resources and causes huge power consumption in data centers. Due to multiple tasks imbalance occurs in cloud platforms which degrade the quality of service for clients. Task scheduling is considered as an important aspect which can be used for load balancing in cloud systems. This helps to allocate best available resources to complete the task by considering various other parameters such as computation time, scalability, makespan, throughput etc. However, various techniques have been proposed for task scheduling in cloud computing but these techniques still suffer from various issues such as makespan, execution time etc. To address this issue, an adaptive swarm optimization approach is presented for heterogeneous virtual machine systems. Proposed approach addresses the multi-objective problem by developing a probabilistic model resulting in optimization and convergence rate improvement. During task execution in cloud computing, if any VM is overloaded then the task is removed from that VM and proposed adaptive swarm optimization technique is used to find any other available optimal resource for task completion. An extensive simulation is performed along with comparative analysis. Experimental study shows that proposed approach outperforms when compared with state of-art technique for load balancing in cloud computing systems.