A Cognitive Approach for Management and Orchestration of Heterogeneous Networks

Today’s world is witnessing exceptional growth in high-end applications such as, mixed reality, tactile Internet, autonomous vehicles, e-sports, and video conferencing etc. The requirements of such applications are large scale, heterogeneous, ubiquitous, and assertive underlying communication network systems. In reality, these applications must be dynamically tuned according to the underlying network resources which are constrained. With significant adaptation of Artificial Intelligence/Machine Learning (AI/ML) techniques, underlying networks’ characteristics can be probed with relevant data collection, analyzed for patterns mapped to network characteristics, deciphered for anomalies, and orchestrated for optimal configurations. In this paper, we introduce a cognitive controller framework for seamless orchestration of enterprise wireless networks. The proposed framework is vendor agnostic and flexible for customization. It is deployable on containers using a Service-Based Architecture (SBA). In this paper, we demonstrate the usability of the framework for a mobility management use-case in an enterprise Wireless Local Area Network (WLAN) scenario. We demonstrate that the proposed framework assists intelligent mobility management in a seamless way with significantly low packet loss and end-to-end jitter as compared to the client-initiated mobility management.