On Predicting Service-oriented Network Slices Performances in 5G: A Federated Learning Approach
Bouziane Brik, Adlen Ksentini
Abstract
To achieve the vision of Zero Touch Management (ZSM) of network slices in 5G, it is important to monitor and predict the performances of the running network slices, or their Key Performance Indicator (KPI). KPIs are usually monitored, but also with the advance of Machine Learning (ML) techniques are predicted, aiming at proactively reacting to any service degradation of running network slices. While network- and computation-oriented KPIs can be easily monitored and predicted, service-oriented KPIs are difficult to obtain due to the privacy issue, as they disclose critical information on the performance of services. To tackle this issue, in this paper, we propose to use a new ML technique, known as Federated Learning (FL), which consists of keeping raw data where it is generated, while sending only users' local trained models to the centralized entity for aggregation. Hence, making FL as an adequate candidate to be used for predicting slices' service-oriented KPIs.