Intelligent Fault Root Cause Localization for 5G Message Platform
Ying Tong, Xiang Dong Jia · 2024
5G message, as an emerging communication technology leveraging 5G networks, provides users with convenient and secure rich media services. The robust maintenance of the 5G message platform is fundamental for delivering high-quality services. Efficiently identifying the root cause of service malfunctions is a pressing concern. This paper presents a machine learning-based multi-service fault root cause localization method. Firstly, it utilizes business key performance indicators (KPIs) to swiftly detect various service faults. Secondly, it employs TF -IDF and word2vec for extracting alert features and applies clustering algorithms to consolidate alert logs within the fault cycle. Lastly, the consolidated alert logs are correlated with business operations using network topology and service transmission chains to pinpoint the underlying root cause of the fault. Experimental results demonstrate that this approach effectively reduces troubleshooting time and achieves rapid root cause localization.