An Incremental Learning Assurance Approach for Intent Based Networking Enabled Data Centers
A network assurance approach for intent-based data centers that uses incremental learning to forecast resource utilization while the infrastructure keeps changing.
- Authors
- Steve Lévesque, Xiaoang Zheng, John Violos, Aris Leivadeas (ÉTS), Matthias Falkner (Cisco Systems, Ottawa)
- Venue
- IEEE IISA 2024, Chania, Greece
- Year
- 2024
- Pages
- 1–4
- Keywords
- Intent Based Networking, Network Assurance, Machine Learning, Network Automation
- Status
- Published
Overview
Intent Based Networking (IBN) aims for closed-loop automation that leads toward a self-governed network. Network assurance is a core part of it: it should detect performance drift automatically and suggest a remedy. In this paper, my co-authors and I bring this idea to the data center.
Approach
We propose an IBN-enabled assurance approach that predicts key resource utilization metrics, so that network administrators can take corrective action proactively when needed. The model learns incrementally, so it can keep up with the dynamic nature of data center infrastructures.
Results
Experiments on real-world datasets show that the approach forecasts resource utilization accurately, even in highly dynamic environments.
Publication details
Presented at the International Conference on Information, Intelligence, Systems & Applications (IISA 2024), held July 17–19, 2024, in Chania, Crete, Greece, and published by IEEE. The work brings together researchers from ÉTS and Cisco Systems.