A Distributed AI System for Improving Single-Domain Network Intent Resolution

Resolving network intents to network service descriptions (NSDs) is a key process in intent-based networking (IBN). Current state-of-the-art systems for single-domain intent resolution operate independently. Agents usually resolve the network intents of their domains without interacting with other agents. However, some aspects of the input network intents will be better interpreted and understood by other agents. We propose to use a distributed multi-agent system to enhance single-domain network intent resolution. Each agent of our proposed system delivers artificial intelligence (AI) operations through a machine-interactive interface. Agents make use of this interface to request other agents the resolution of parts of the input intents. The responses are gathered and a resulting NSD is produced. We demonstrate that adding external intelligence to the operation highly improves the network intent resolution process, reaching 100% accurate resolutions, which is more than 40% improvement over current state-of-the-art solutions.