Modifikasi Algoritma Whale Optimization (WOA) untuk Peningkatan Performa Load Balancing pada Lingkungan Simulasi CloudSim Plus
HAFIZH ABIYANIQBAL HARAHAP, Dr.techn. Khabib Mustofa, S.Si., M.Kom.
2024 | Skripsi | ILMU KOMPUTER
Cloud computing is one of the high-demand computing paradigms, making the performance of load balancer components in cloud infrastructure a crucial factor. In the past decade, research trends in load balancing algorithms have shown an inclination towards the use of meta-heuristic optimization methods such as Ant-Colony Optimization (ACO), Particle-Swarm Optimization (PSO), and similar approaches. On the other hand, the utilization of the Whale Optimization Algorithm (WOA) in the context of cloud load balancing is still limited in terms of research quantity, providing ample room for optimization of WOA implementation specifically in this domain.
This research proposes a modified Whale Optimization Algorithm (MWOA)-based load balancer model in the population initialization phase, considering solutions from previous cases. This modification aims to enhance task allocation performance based on Quality-of-Service (QoS) aspects such as makespan, response time, DOI, and VM utilization, while reducing iteration redundancy in similar task allocation cases to optimize convergence. The implementation is carried out on the CloudSim Plus simulation framework to test and evaluate the proposed model’s performance against the classic WOA.
Hypothesis testing results between the proposed modified WOA model and the classic WOA demonstrate a significant improvement in QoS performance, with a notable reduction in makespan by 0.175 seconds and an absolute increase in mean VM utilization by 1.7%. However, the modification has yet to prove effective in enhancing convergence speed without impacting performance degradation. These conclusions are drawn based on the outcomes of 150 simulation experiments across 5 varied cloud environment configuration scenarios.
Kata Kunci : Whale Optimization Algorithm, optimasi meta-heuristik, load balancing, CloudSim, komputasi cloud