Load Balancing in Cloud Computing Model Using Meta Heuristic Algorithms

Abstract

Large-scale homogeneous cloud computing environments offer the promise of access to an enormous quantity of computing resources at a comparatively less price. It is an empowering and constructive paradigm for both supplier and receiver in miscellaneous opportunities of activities. One of the principle issues that these runtime frameworks need to address is dynamic load balancing that guarantees no virtual machine in the system is underutilized or overloaded with work.
This thesis deals with the issue of acquiring least makespan for unpredictable and autonomous tasks on homogeneous cloud computing systems and offers a mathematical model for load balancing in the Cloud. It concentrates on meta-heuristic nature inspired algorithms that can be utilized for load balancing throughout the execution of unpredictable tasks. This thesis gives distinctive consideration to evolutionary methodologies, named as Honey Bee Foraging Behavior and Particle Swarm Optimization (PSO) that has been functional to search an optimal solution of load balancing and directing to minimize the total makespan of the virtual machines executing autonomous tasks, and then assessed by comparing with the traditional approaches in the experiments. The achieved output demonstrate that our suggested algorithm Improved Honey Bee and TMBLB-PSO produces superior solutions than the existing approaches.