Pendekatan Rock Hyrax dan Regret Minimization untuk Optimasi Load Balancing dalam Jaringan Cloud Computing

Abstract

Application server load balancing is a major challenge in improving system performance in the era of cloud computing, especially in web hosting, healthcare, data centers, and customer support services with unpredictable traffic. This research aims to prove that adaptive load balancing can be achieved through the development of ReAdaBalancer, an advanced architecture based on Flask (backend) and Nginx (load balancer) that is capable of distributing traffic in realtime without overloading a single server. Performance analysis was conducted using M/M/s/K queueing theory to simulate various load scenarios, with resource allocation optimization based on real-time monitoring. Simulation testing compared ReAdaBalancer with standard approaches under variable traffic conditions (low, medium, peak). ReAdaBalancer reduces response time by more than 67% and request rejection rates by more than 50% compared to traditional methods. This research has many opportunities for further study. Future improvements may include the application of ReAdaBalancer in distributed multidata center environments, the addition of reinforcement learning for more autonomous decision making, and the study of load balancing strategies that use less energy. These results demonstrate that ReAdaBalancer can be practically implemented to enhance the scalability, stability, and efficiency of application server systems in everyday use.