Research on Network Configuration Verification Based on Association Analysis

Abstract

This paper studies the application of association analysis in the scenario of massive network configuration verification, and puts forward a kind of network configuration anomaly detection method and system based on association analysis. We creatively use the weak association rules in association analysis to detect configuration anomaly. And we can generate a configuration anomaly verification model through training the processed configuration data, which is applied to scan the massive configuration data and output configuration anomaly results. At the same time, we constructed a network configuration verification system based on ZTE's AI platform and verified the effectiveness of the algorithm and model by using the massive realistic configuration data collected from the existing networks. The experimental results show that the precision and recall of the proposed network configuration verification system are above 80%.