A Relative Load Balancing (RLB) Method for Efficient Multi-Server Request Handling
Ameer Akram Mousa, Mahdi S. Almhanna · Engineering Technology & Applied Science Research · 2025
Load balancing is a critical technique in distributed systems and network infrastructures, designed to efficiently distribute incoming workloads or network traffic across multiple servers or resources. By dynamically allocating tasks, load balancing maximizes resource utilization, minimizes response times, and prevents any single server from becoming a bottleneck. This paper focuses on building a Relative Load Balancing (RLB) system between server groups to process client requests and avoid problems in static and dynamic load balancing systems, such as round robin and fewer connections. The requests were divided proportionally according to the capacity of a group of root servers. A system for processing text images and extracting data was developed to implement the principle of relative load balancing between servers, controlled by a master server called the backbone. This system handles the testing and measurement mechanism for the RLB request processing. The results showed that the proposed RLB system achieved a throughput of 101,000 B/s, a PDR of 97 %, average latency of 600 ms, and average response time of 2 s, with RAM usage of 2100 MB, main memory usage of 52% and CPU usage of 50%. In addition, the total processing time (Makespan) was 10 s, with a loss of time in waiting of 1 s and a gain of 8 s.