A Hybrid Algorithm Based on PSO Algorithm and Chi-Squared Distribution for Tasks Consolidation in Cloud Computing Environment

In order to maximize the effectiveness and performance of cloud computing systems, this study focuses on addressing the challenges of workload balancing and resource utilization in cloud scheduling. Workload balancing plays a crucial role in ensuring that computing workloads are evenly distributed across available resources, thereby reducing the likelihood of resource constraints and enhancing system performance. On the other hand, resource utilization aims to utilize processing power, memory, and network bandwidth to their fullest capacity, resulting in improved efficacy and cost-effectiveness of the cloud infrastructure. To tackle these challenges, we propose a novel optimization technique called CHPSO (Chi-squared Particle Swarm Optimization) in this context. The proposed algorithm demonstrates its effectiveness in optimizing resource utilization compared to other algorithms such as PSO (Particle Swarm Optimization) and CS (Cuckoo Search).

I. Introduction

Cloud computing has gained widespread adoption as a computational technology, offering users the ability to perform tasks using a pay-as-you-go model without the need for extensive hardware infrastructure [[!]]. This paradigm provides flexibility in configuring various services and architectures within the cloud environment to suit the specific requirements of applications and users [2]. Users can request and utilize virtual machines (VMs) as computational resources, including computing power, storage, and network bandwidth, based on their needs. However, a significant portion of energy consumption in cloud data centers is attributed to servers. To address this issue, dynamic VM consolidation techniques can be employed to reduce the number of active servers while maintaining service quality. An effective server consolidation framework should identify overloaded and underutilized hosts to optimize resource allocation for VM execution. Once a host is identified as underloaded, it can be shut down, and the VMs can be migrated to other suitable hosts [3].