A Performance Benchmark of Open-Source NMS in Constrained Network Environments

NMS Performance Benchmark Dataset and Analysis Scripts Overview This repository contains the complete dataset and analysis scripts for the paper "The Resource Overhead of Enterprise-Grade Monitoring: A Performance Benchmark of Open-Source NMS in Constrained Network Environments." The dataset and scripts support the replication of benchmarking results for four leading open-source Network Monitoring Systems (NMS): Zabbix, LibreNMS, Nagios Core, and Cacti under resource-constrained conditions (2 vCPU, 2GB RAM). Contents Data Baseline scenarios: Idle resource consumption (CPU, memory, disk I/O, dashboard response time) Scaling scenarios: Performance impact when scaling from 2 to 4 monitored devices Traffic spike scenarios: System responsiveness during simulated network congestion Endurance scenarios: 24-hour continuous monitoring stability analysis Analysis Scripts All scripts are written in Python and include: baseline-analysis.py – Idle resource consumption analysis scaled-analysis.py – Scaling factor analysis spike-analysis.py – Crisis responsiveness analysis endurance-analysis.py – 24-hour stability and regression analysis regression-analysis.py – Linear regression for endurance characteristics Requirements Python 3.8+ pandas, numpy, matplotlib, seaborn, scipy Key Findings Idle Tax: Memory footprints range from 437 MB (Nagios) to 509 MB (LibreNMS) Scalability: Total CPU Work scaling factors range from 5.42x (Cacti) to 54.07x (Zabbix) Response to Traffic Spikes: Dashboard response times amplify 10.0x to 17.9x during network incidents Endurance: Response time degrades consistently across all platforms, with Nagios, Zabbix, and LibreNMS showing identical degradation rates (+15.87 ms/hr) Usage Clone the repository or download the dataset Install required Python packages: pip install pandas numpy matplotlib seaborn scipy Run individual analysis scripts from the repository root Results are saved to the Results/ directory with appropriate subfolders