A Comparative Analysis of Metaheuristic Techniques for High Availability Systems

In the ever-evolving technological landscape, ensuring high system availability has become a paramount concern. This research paper focuses on cloud computing, a domain witnessing exponential growth and emerging as a critical use case for high-availability systems. To fulfil the criteria, many services in cloud infrastructures should be combined, relying on the user's demands. Central to this study is load balancing, an integral element in harnessing the full potential of heterogeneous computing systems. In cloud environments, dynamic management of load balancing is crucial. This study explores how virtual machines can effectively remap resources in response to fluctuating loads dynamically, optimizing overall network performance. The core of this research involves an in-depth analysis of several metaheuristic algorithms applied to load balancing in cloud computing. These include Genetic Algorithm, Particle Swarm Optimization, Ant Colony Optimization, Artificial Bee Colony, and Grey Wolf Optimization. Utilizing CloudAnalyst, the study conducts a comparative analysis of these techniques, focusing on key performance metrics such as Total Response Time (TRT) and Data Center Processing Time (DCPT). The findings of this research offer insights into the varying behaviors of these algorithms under different cloud configurations and user retention levels. The ultimate aim is to pave the way for developing innovative load-balancing strategies in cloud computing. By providing a comprehensive evaluation of existing metaheuristic methods, this paper contributes to advancing high-availability systems, underscoring the importance of tailored solutions in the dynamic realm of cloud technology.