ADAPTIVE, FEEDBACK-DRIVEN TASK SCHEDULING AND LOAD BALANCING: A SUPERIOR APPROACH FOR MODERN CLOUD COMPUTING SYSTEMS

Abstract: Cloud computing has emerged as the foundation of modern IT services, providing scalable, on-demand resources via virtualization. Effective resource management in these settings needs not only skilled load balancing but also smart job scheduling algorithms to guarantee fairness, reduce delays, and improve utilization.

Static algorithms that are commonly used for the process of task scheduling and load balancing are simple and easy to implement, but they cannot adapt to any kind of changes in conditions. As a result, there is an inefficient use of valuable resources, increased response time, and reduced throughput in a large-scale cloud setting. This research aims to compare and test traditional algorithms against bio-inspired algorithms. Cloud Sim is used to simulate an extensive cloud environment, including five VMs and fifty cloudlets. Four algorithms, such as Round Robin, Weighted Round Robin, Honeybee Foraging, and Ant Colony Optimization, were implemented under similar conditions. The performance was measured by calculating the average response time, throughput, and utilization.

It was found that bio-inspired algorithms greatly surpassed traditional algorithms. Ant Colony Optimization was able to achieve the minimum response time, maximum throughput, and optimal utilization. Honeybee algorithm displayed remarkable adaptation ability, whereas classic algorithms appeared to be insufficient for changing circumstances. Therefore, adaptive algorithms are superior to static algorithms in contemporary cloud computing settings.

Keywords: Cloud Computing, Static Algorithm, Dynamic Algorithm, Load Balancing, Task Scheduling, Virtual Machine.