Towards zero downtime: Using machine learning to predict network failure in 5G and beyond

A stable network is very important to both providers and their customers, as it increases reliability, improves security helps customers companies save costs. When outages occur, they result in significant downtime financial losses for organizations users. Traditional methods of detecting troubleshooting failures are often reactive time-consuming, whereby administrators rely on traditional such monitoring manual troubleshooting. These not effective preventing failures. In this paper, we propose a machine learning-based approach predict minimize downtime. Network performance observability data from 5G core testbed based Cloud-native Functions (CNFs) used train several supervised learning models, including random forest, gradient boosting regressor, conventional support vector regressor proposed Our experiments analysis show that the model Support Vector Regressor (SVR) produced better results compared other models. short amount time (ten seconds), SVR capable predicting whether failure event will occur or within next ten minutes, with an f1-score more than 0.9. indicate approaches can significantly enhance detection prediction failures, leading zero improved performance.