System Health Evaluation for Enhanced Resilience of Telecom Networks

In our digital society, the resilience of telecommunications networks is crucial to ensuring reliable connectivity. These networks face growing challenges, including accelerated obsolescence and the unavailability of essential components, often amplified by supply chain issues. These risks threaten network stability and can lead to service disruptions if not properly anticipated and managed. Our research focuses on predicting the risk of obsolescence and unavailability of server components, a major concern for the Orange group. However, the obsolescence risk of a server is not solely determined by its age but also by the age of its components, as each part has its own life cycle, which may vary depending on the manufacturer. Our approach involves using graphs to model the dependencies between components and assess the system's overall obsolescence risk. By leveraging AI and predictive analytics, we aim to anticipate the obsolescence or unavailability of individual components, providing crucial insights for proactive management.

Keywords: Obsolescence, Artificial intelligence, Machine learning, Knowledge-graph, Resilience, Telecommunication networks.