Autonomous networks have become the goal of operators globally, using fewer people for higher levels of network change and activity. Even if not yet widely implemented, closing the loop and intent-driven networks have become a foundational idea for a stepwise improvement of network-wide autonomy.

The rapid emergence of agentic AI – intelligent agents, empowered to make network-changing decisions in real time – makes it essential that operators have the correct model for understanding autonomous networks and operations.

“Needing a model for network-wide decision making”

This paper gives a revised framework to deliver a highly autonomous network. It advocates moving the focus of closed loop from one of control to one based on the management and optimization of decision-making across network and service lifecycles. This will be based on three foundations:

• Digital twins of the network providing an accurate reference network model (simulated and planned future, present and historic past) on which all network autonomy processes rest.
• Inventory combined with network observability (operational data)
• Strong intent-driven orchestration.

TABLE OF CONTENTS

Executive Summary

How agentic AI changes the network autonomy equation

The origins of the report

Autonomous Network Reality

The autonomous network digital twin

Managing decision-making risk

Practical network-wide autonomy

Conclusion – Strategic Imperatives for Autonomy

Annex – Network autonomy control loop challenges

About VIAVI Solutions

About Inmanta