Improving Load Balancing Efficiency in Cloud Data Centers Through Hybrid Grey Wolf with Cat Swarm Optimization

• Published in: Informatica (Ljubljana) Vol. 49; no. 28; pp. 121 - 135
• Main Authors: Ullah, Arif, Razak, Siti Fatimah Abdul, Mrhari, Amine, Yogarayan, Sumendra, Kazmi, Syeda Hina Mazhar
• Format: Journal Article
• Language: English
• Published: Ljubljana Slovenian Society Informatika / Slovensko drustvo Informatika 01.08.2025
• Subjects: Algorithms; Cloud computing; Computer centers; Cost control; Data centers; Internet; Load; Machine learning; Optimization; Performance degradation; Resource management; Resource utilization; Scheduling; Software services; Support vector machines; Virtual environments; Workloads
• ISSN: 0350-5596; 1854-3871; 1854-3871
• DOI: 10.31449/inf.v49i28.8860

An effective resource management strategy that anticipates server resource utilization and appropriately distributes the load is recommended in order to address these problems and enhance data center performance. By reducing the number of servers in use, facilitating virtual machine migrations, and optimizing resource utilization, it helps save power. To reduce the likelihood of service level agreement (SLA) violations and performance degradation caused by either oveloded or under loaded servers and virtual machines. Resources for software applications can now be dynamically altered as needed thanks to the growth of cloud computing. Since better resource consumption can lead to increased scalability as well as significant cost and energy savings, effective resource management is crucial in cloud computing. The flexibility of cloud resources allows clients to dynamically increase and decrease their resource demands over time. However, predefined virtual machine sizes and variable resource requirements result in underutilization of resources, load imbalances, and high power consumption. The goal of this research is to develop a hybrid technique by combining Grey Wolf with algorithms. The hybridization processes take place in the Grey Wolf portion, when the Cat Swarm initialization process takes the place of the startup phase. The virtual machine (VM) section's data selection is enhanced by this substitution. The Grey Wolf and Cat Swarm algorithms are two examples of optimization algorithms. The evaluation criteria that are used are makespan, throughput, degree of imbalance, and turnaround time with degree of imbalance. The recommended approach outperforms alternative algorithms in each of these metrics. The proposed hybrid strategy resulted in 0.3% increase overall performance. Potential directions for future research include testing the proposed approach in larger and more complex data distribution in cloud data centers.