Діагностика несправностей конфігурації VLAN та VTP за використання великих мовних моделей

Abstract

In the presented work, the effectiveness of large language models (LLM) as a tool for automated diagnostics of typical faults in configuring virtual local area networks (VLANs) and their centralized management protocol (VTP) was studied. A simulation of critical errors in a computer network was carried out, a structured text prompt was formed for different LLMs, and the responses generated by different models were analyzed. Methods for improving the quality of diagnostic recommendations obtained from neural networks for their application in corporate networks built on the basis of Cisco switches were considered.