Ikramsyah, Muhammad Aldy
Unknown Affiliation

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Evaluasi Kualitas Layanan dalam Load Balancing NGINX : Studi Perbandingan Algoritma Round Robin dan Least Connection Ikramsyah, Muhammad Aldy; Seta, Henki Bayu; Isnainiyah, Ika Nurlaili; Theresiawati, Theresiawati
Informatik : Jurnal Ilmu Komputer Vol 21 No 3 (2025): Desember 2025
Publisher : Fakultas Ilmu Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52958/iftk.v21i3.10796

Abstract

This study evaluates the Quality of Service (QoS) of the Round-Robin and Least Connection algorithms in load balancing with NGINX. Round Robin outperformed and was consistent across most of the evaluated parameters, including throughput, latency, jitter, and packet loss. Its simplicity stems from the repeated distribution of requests without monitoring server connections, making it ideal for homogeneous traffic and consistent server specifications. However, its performance is poor in dynamic traffic or heterogeneous server architectures. In contrast, Least Connection is more advanced in accommodating dynamic traffic and heterogeneous server environments by allocating load in real-time based on active server connections. Nevertheless, Least Connection showed instability in various tests, especially at high thread counts, where anomalies such as substantial value spikes were recorded. This indicates that it is not suitable for scenarios involving uniform traffic and server specifications. Least Connection is more suitable for complex networks with dynamic traffic and varying server specifications, while Round Robin is recommended for stable workload environments because it provides consistency and simplicity. Round Robin demonstrated superior overall performance, while Least Connection excelled in adaptability, highlighting the need to align algorithm selection with traffic characteristics and server infrastructure. To improve the effectiveness of load balancing strategies, future research should investigate dynamic traffic, higher loads, and diverse environments.