An Incremental Learning Assurance Approach for Intent Based Networking Enabled Data Centers

Intent Based Networking (IBN) is a novel paradigm that aims to create a closed-loop automation towards building a self-governed network. An inherent part of IBN is the network assurance, which tries to automatically detect any performance drift and propose a remedy solution. Following this trend, in this paper we propose a novel IBN -enabled network assurance approach in a data center setting, in order to predict key resource utilization metrics and to help network administrators to proactively take corrective actions when needed. Specifically, we propose an incremental learning approach for data center assurance, designed to accommodate the dynamic nature of infrastructures. Experimental findings obtained from real-world datasets substantiate the effectiveness of our approach, demonstrating its ability to accurately forecast resource utilization even amidst highly dynamic environments.