IMPLEMENTATION OF LOAD BALANCING ON WIFI CORNER FOR RT/RW NET SERVER REDUNDANCY

Abstract

As community-based networks increasingly rely on the internet, providing stable, reliable, and uninterrupted service has become essential. However, public internet hubs like "WiFi Corner" often suffer from network overloads and frequent disruptions due to the limited capacity of local RT/RW Net servers. This bottleneck degrades service quality, spikes latency, and increases the risk of downtime, ultimately disrupting user activities. To address this issue, this study implements a load-balancing technique at WiFi Corner to improve server redundancy and ensure continuous uptime. We utilized the Per Connection Classifier (PCC) method on MikroTik RouterOS, which dynamically distributes traffic based on active connections while providing an automatic failover mechanism if an ISP link drops. To evaluate performance, we compared three distinct scenarios: A single ISP setup, Dual ISPs without load balancing, and Dual ISPs optimized with PCC load balancing and failover. The performance metrics analyzed include throughput, delay, packet loss, jitter, and overall system uptime. The results demonstrate that implementing PCC load balancing successfully doubled the throughput compared to the baseline setup. Additionally, delay and jitter were significantly reduced, and packet loss dropped below 2%. Thanks to the integrated failover mechanism, network uptime reached nearly 100%. Ultimately, this study offers a cost-effective and highly efficient technical solution to boost the reliability and sustainability of community-based RT/RW Net internet services.