TELECOM LAB PROJECT
Integrated E2E network-wide management
This project adopts an end-to-end (E2E) cross-network approach to revolutionize the management of cellular networks. It addresses:
Radio layer, encompassing interactions between base stations and user terminals.
Virtualized resources and configurations of RAN elements.
The project leverages advanced ML techniques to estimate end-user quality indicators under varying network conditions, resource availability, and contextual factors. These predictive insights will inform sophisticated AI mechanisms designed to optimize E2E quality, ensuring superior user experiences and efficient resource utilization.
A key focus of the project is the development of a comprehensive failure management system that provides a holistic view of the network. This system integrates insights from the radio link and virtualized RAN resources to enhance network resilience.
To achieve this, the project will design novel algorithms and AI mechanisms capable of embracing an E2E cross-network vision. These solutions will build upon and integrate the metrics and methodologies defined by previous initiatives, delivering a robust framework for optimization and failure management in next-generation cellular networks.
USE CASE
With the rise of decentralized networks, and AI-based network control, the ability to detect problems within the O-RAN network has become even more challenging task.
Picture this scenario: a group of users have a video chat and in the middle of a video conversation with friends from another country, their video quality suddenly drops. At this point, we’re unsure if it’s a network problem due to interference, wrong radio resource configuration, downtime due to bad handover management, high load of a gNB or maybe even an issue with the media relay server.
There’s great interest in offering the operators more in-depth insights about what is happening in the radio access network by detection and mitigation of failures that are specific for O-RAN networks.
INNOVATION DIRECTION
Our innovation in failure detection and mitigation focuses on a deep analysis of how various applications operate in real-world conditions, including their communication patterns with users and servers and their ability to adapt to changing network conditions.
By thoroughly understanding the characteristics of application-generated network traffic, we gain valuable independent insights into app behavior and performance. These insights enable us to develop advanced tools, such as real-time Quality of Experience (QoE) estimators, which can pinpoint the root causes of network faults.
This comprehensive approach not only enhances our ability to detect failures quickly but also facilitates effective mitigation strategies, ensuring seamless application performance and robust network reliability.