Efficient Load Balancing Using Enhanced Dragonfly with Firefly Optimization Over Cloud Computing Environment

Abstract

The large number of users of the cloud has resulted in a rapid increase in the number of requests for tasks, and scheduling tasks efficiently and load balancing them across non-homogeneous virtual machines (VMs) has become an urgent matter. This is a problem that is inherently NP-hard, especially when the Objective is to optimize multiple Objectives, such as makespan, execution cost, scheduling time, and resource utilization. Current meta-heuristic methods, such as the Particle Swarm Optimization (PSO), are computationally expensive and do not converge to global optimality when applied in isolation. To overcome these shortcomings, this paper proposes an Enhanced Dragonfly-Firefly Optimization (EDFO) algorithm, coupled with an Improved Advanced Encryption Standard (IAES) mechanism for safe and effective scheduling of cloud tasks. The EDFO model was an effort to merge the global search capability of the Dragonfly algorithm and the local refinement strength of the Firefly algorithm. The multi-objective function is formulated to minimize the maximum difference in VM completion times and to achieve even workloads. The results of the proposed EDFO-IAES are experimentally evaluated and shown to be much more effective than currently used methods such as DFGA and IBPSO-LBS. In particular, the imbalance level decreases to 0.45, the average storage usage increases to about 94%, and makepan is minimized by nearly a quarter across the workload variabilities. Besides, incorporating IAES increases data security without imposing a substantial computational burden. All in all, the suggested framework is more efficient, scalable, and secure, which is why it can be considered a viable solution for a real-life cloud computing setup.