AI service assurance for the era of cloud-based network architecture

Commercial context

As telecoms infrastructure and its capabilities continue to develop apace, so too do customer expectations of the services they can deliver. The growing customer demand for more sophisticated network services often means adding new complexity to the network – and increasing the resources required to maintain it in the process. As conventional IT and network operations converge with newer, cloud-based architecture, CSPs need to reevaluate the method by which they maintain service assurance – which is otherwise currently designed for the outgoing era of static infrastructure.

The Revolutionizing service assurance through AI powered, intent-based systems for continuity and customer satisfaction Catalyst aims to solve this challenge. The solution identified is to use AI-enabled operations to enable CSPs to manage and operate advanced hybrid networks in an efficient, scalable and sustainable way. The solution uses AI-driven operations to assist CSPs in effectively managing advanced hybrid networks. Through real-time modelling, comprehensive cross-domain visibility is achieved, enabling automatic correlation of real-time data across applications, infrastructure, and networks, the system accelerates problem identification and resolution processes. It substantially decreases operational metrics like MTTI and MTTR, paving the way for autonomous operations with predictive capabilities and automated resolutions.

The solution

The solution is being tested against multiple use cases – the principal application being an automated operational process that supports delivery of a cloud-native VoIP service. To demonstrate machine learning capabilities, the project team has chosen a particularly challenging failure case in this application. In this scenario, a telephony application server experiences degradation in performance, due to a shared network infrastructure switch port going offline. In this instance, the relationships between service applications, cloud (Kubernetes), infrastructure and network layer are automatically captured, modelled and used to automatically and accurately correlate service, infrastructure and network problems together, triggering an automated root cause analysis process.

Application and wider value

By building automations to optimize network performance, the need for manual intervention is reduced, as is the occurrence and impact of human error. As Marton Sabli, Head of IT Production Architecture & Roadmap at Telia confirms, “the Catalyst is succeeding in its objective to increase operations performance and efficiency, reduce OPEX and improve customer experience for cloud-based services using AIOps and advanced topology modelling.”

In a telecoms era defined by ever-evolving and complex networks, it’s high time the industry switched to an operation management system which is both more cost effective and provides superior resolution times. In demonstrating the success of automated problem detection, correlation and root-cause analysis, the industry can replace the manual problem identification and diagnostic processes – and craft a sustainable alternative to be scaled across the globe.