Friday 28 March 2025
The quest for a more intelligent, automated, and scalable network has been ongoing for years, with researchers and engineers working tirelessly to develop solutions that can adapt to the ever-evolving demands of modern communication. In recent times, the Open Radio Access Network (O-RAN) Alliance has made significant strides in this regard, introducing the concept of near-real-time RAN Intelligent Controllers (near-RT RICs) that enable intelligent, software-defined management of Radio Access Networks (RANs). However, despite these advancements, a unified AI-driven control framework across both wireless and optical networks remains an elusive goal.
Enter the modular and integrated AI control framework proposed by researchers at Trinity College Dublin. This comprehensive design aims to bridge the gap between radio access networks and transport networks, enabling decentralized AI-driven control that can self-configure, monitor, and repair at scale. The framework is built upon the principles established by O-RAN near-RT RIC controllers but extends this concept into the optical domain.
The proposed framework consists of five key components: an AI engine, a registry, two message brokers, and a protocol translation module. This modular design allows for flexible implementation tailored to the specific needs of various network nodes and domains. The AI engine provides an environment for running AI control applications, which can monitor network performance, predict capacity requirements, and adjust resource allocation accordingly.
One of the most significant contributions of this framework is its ability to integrate AI capabilities into Passive Optical Networks (PONs). This enhances functionality by enabling seamless coordination and resource management between optical and wireless networks. By decoupling publishers and subscribers through a publish-subscribe pattern, the system ensures efficient communication and control across different segments of the network.
The proposed framework also enables decentralized collaboration between multiple controllers, allowing for seamless interaction and coordination between different AI controllers. This is achieved through inter-AI message brokers that facilitate east-west communication between different AI controllers.
In practical terms, this framework has significant implications for the development of intelligent, zero-touch networks. For instance, it can be used to predict capacity requirements in real-time, anticipating RAN usage and reassigning Optical Network Units (ONUs) and other ONUs to different channels to optimize Quality of Service (QoS). This would enhance wavelength management and distribute RAN traffic across channels, meeting high-priority, low-latency needs.
The proposed framework is a significant step toward the realization of intelligent, automated networks capable of self-configuration, monitoring, and repair at scale.
Cite this article: “Modular AI Control Framework for Intelligent Networks”, The Science Archive, 2025.
Ai-Driven Control Framework, Open Radio Access Network, Ran Intelligent Controllers, Near-Rt Rics, Software-Defined Management, Decentralized Ai-Driven Control, Modular Design, Passive Optical Networks, Publish-Subscribe Pattern, Inter-Ai Message Brokers,
