Service-Centric Fault Management for 5G and Beyond Wireless Networks

The next generation networks are tailored to support diversity of use cases and services with stress on stringent Quality of Service (QoS) requirement. This necessitates an efficient Fault Management (FM) system. With focus on QoS, only prediction or identification of faults is not sufficient to fix the fault without knowing the impact of fault. To this end, we propose a machine learning framework to focus on service-centric FM. The proposed model analyses the impact of fault by predicting the user data rate and also classifies the faults with multiple priorities. Extensive experiments on real datasets from network demonstrate enhanced performance by capturing the local and global dynamics of network parameters. Our proposed scheme can be deployed by telcos for their Fifth Generation (5G) networks.