Generative AI–powered graph for network digital twin (TLC202)

Transforming Network Operations with AI-Powered Digital Twins

Introducing the Challenge

  • Orange is a global telecommunications operator with operations in 26 countries, serving over 300 million customers.
  • Orange's network consists of over 1 million network elements across various domains like transport, core, and radio access networks.
  • Managing this complex, heterogeneous, and rapidly evolving network is a significant challenge, requiring new approaches to detect anomalies, perform root cause analysis, and automate remediation.

Limitations of Traditional Approaches

  • Reliance on expert rules and static correlation-based techniques for anomaly detection and root cause analysis is inefficient and difficult to maintain.
  • Siloed management of network domains (transport, core, radio access) hinders end-to-end visibility and understanding of network dependencies.
  • The vast volume and variety of network data (e.g., 1 trillion call detail records per day) make it difficult to extract meaningful insights.

Transforming Network Operations with AI and Graph Technologies

  1. Graph Modeling of the Network:

    • Representing the network as a heterogeneous and temporal graph, capturing the interconnections between different network elements.
    • Enabling a holistic view of the network, breaking down siloes between domains.
    • Preserving the temporal evolution of the network topology and associated KPIs/alarms.
  2. Graph-powered Analytics and Machine Learning:

    • Leveraging graph neural networks to perform predictive analytics, anomaly detection, and root cause analysis, taking into account network dependencies.
    • Utilizing graph analytics techniques to identify critical network elements, cluster nodes by traffic patterns, and uncover hidden insights.
    • Integrating these graph-based capabilities into the network operations workflows.
  3. Generative AI for Intelligent Reporting:

    • Connecting the graph-based insights to a generative AI model to automatically generate detailed root cause analysis reports.
    • Enabling network operators to quickly understand the root cause of incidents and take appropriate actions.
    • Bridging the gap between the complex network data and the needs of the operational teams.

Proof of Concept and Benefits

  • Orange and AWS collaborated to implement this approach in a regional network, demonstrating significant improvements in the time to root cause analysis, from hours to just seconds.
  • The solution provides a comprehensive, end-to-end view of the network, breaking down traditional siloes and enabling cross-domain troubleshooting.
  • The generative AI-powered reporting streamlines the incident management process, making it more efficient and accessible for the network operations team.

Next Steps and Future Outlook

  • Scaling the solution to cover the full network of Orange, across multiple countries and network domains.
  • Continued refinement of the graph-based machine learning models to enhance the accuracy and robustness of the root cause analysis.
  • Exploring further automation and self-healing capabilities to move towards the vision of autonomous networks.

Your Digital Journey deserves a great story.

Build one with us.

Cookies Icon

These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.

If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.

Talk to us