AI/ML has been a trending topic for a while now, yet the application of these techniques to specific challenges remains obscure. In this session, we unveil the seamless integration of AI/ML into the domain of IP core networks to perform traffic flow forecasting (based on time series analysis).

We will present a demo of a solution that might help in predicting and preemptively addressing potential network congestion events. These issues can be caused by many reasons, and we will demonstrate how to detect those caused by various failure types.

Detection of congestion events for various resiliency levels is the first step. Additionally, we will demonstrate how to efficiently mitigate future congestion problems with the help of analytical tools similar to the one presented during our demo.