DESIGN AND IMPLEMENTATION OF A HYBRID LOAD BALANCING ALGORITHM FOR OPTIMIZED SERVER RESPONSE AND RESOURCE UTILIZATION

Abstract

Distributed servers need good load balancing since changing client requests and resources can cause response times to vary and performance to degrade. Round-Robin methods have been used for a long time, however they don't account for server workload changes, and static approaches may overload servers and waste resources. This work creates and uses a load-balancing system that combines Round-Robin, Least Connections, and Weighted Response Time to solve these issues. The purpose is to reduce reaction time, allow workload changes, and improve server efficiency. The proposed method includes creating a hybrid load balancer that uses a fitness-based selection mechanism based on natural selection to keep an eye on backend servers and do regular health checks. This system checks servers on the fly based on current connections and average response times. This makes sure that requests are sent to the fastest server. We tested how clients and servers interact in a controlled setting with three servers and two clients. GUI-based monitoring was used to see the state at all times. The hybrid approach does better than traditional Round-Robin techniques by lowering error rates, increasing throughput, and shortening the average response time. The system can also grow as needed and handle errors, as broken servers are quickly taken out of the pool to keep service running smoothly. These results show that hybrid load balancing is a good way to improve distributed server systems. This makes it a strong and flexible choice for real-time applications that need reliable and consistent performance.