AN IMPROVED DEFICIT WEIGHTED ROUND ROBIN ALGORITHM FOR EFFICIENT SERVER LOAD BALANCING

Load balancing is the process of improving the performance of a parallel and distributed system through a redistribution of load among the processors. The Round Robin (RR) algorithm, known for its simplicity and fairness, distributes requests evenly across servers in a cyclic manner [1] [5]. However, RR does not account for server capacity or dynamic workload conditions, often leading to inefficient resource allocation and potential server overloading. In this paper presents an enhanced approach based on the Dynamic Weighted Round Robin (DWRR) algorithm. Unlike the traditional RR, DWRR dynamically adjusts the weight of each server based on its current performance metrics, such as CPU load, memory usage, or response time. This enables more intelligent and adaptive request distribution, ensuring that more capable servers handle greater loads while preventing weaker servers from becoming bottlenecks. The main purpose of this paper is to reduced response time, and more balanced load distribution compared to standard RR.