Classification of Load Balancing Optimization Algorithms in Cloud Computing: A Survey Based on Methodology

Abstract

In this paper, a modern classification of load balancing strategies in cloud computing is presented that can help researchers and practitioners to better understand and select the most appropriate algorithms for their specific needs. Our classification covers three main types of load balancing algorithms including static heuristics, dynamic heuristics, and metaheuristic algorithms. We thoroughly evaluate each type of techniques and identify their respective benefits and drawbacks. Static heuristics are easy to implement and require minimal computational resources, but may yield suboptimal allocations due to significant workload variations. Dynamic heuristics are more adaptable but may incur higher overhead costs due to frequent updates. Metaheuristic algorithms offer highly optimized solutions but may be computationally expensive and not always guarantee optimal results. In addition to offering a framework for understanding various load balancing algorithms, our classification can help researchers and practitioners select the most suitable algorithm for their specific needs and constraints. By taking into account factors such as workload variability, resource availability, and computational limitations, users can choose an algorithm that optimally addresses their requirements, while simultaneously minimizing overhead costs and enhancing efficiency.