Comparative Analysis of Network Troubleshooting Tools in Linux Environments with AI- assisted Root Cause Analysis

Sammanfattning

Network troubleshooting is important for maintaining service availability and operational stability within IT environments. This project evaluates Linux-based troubleshooting tools across simulated network-related scenarios involving connectivity issues, service failures, port availability problems, and DNS resolution failures. Troubleshooting outputs generated using ping, curl, nmap, and tcpdump were further analyzed through AI-assisted root cause analysis (RCA) using the TinyLlama model hosted locally through Ollama. The results showed that ping and curl provided fast troubleshooting feedback and high usability, while nmap and tcpdump offered greater diagnostic visibility. The tools curl, nmap, and tcpdump demonstrated high troubleshooting accuracy across the simulated scenarios. The AI-assisted RCA workflow generated contextual troubleshooting suggestions, although some responses were broader than the actual fault condition and required human validation. The study demonstrates how traditional Linux troubleshooting tools and AI-assisted interpretation approaches may be combined to support troubleshooting activities. The findings suggest that locally hosted AI-assisted RCA can provide useful troubleshooting support when used together with traditional troubleshooting methods.