Impact at a Glance
- 13 active researchers using MX-PDK O-RAN
- 32 supported use cases*
- 25 supported network deployments*
The Digital Connectivity & Innovation Centre (DCIC) at the University of the West of Scotland’s (UWS) Ayr Campus is an innovation hub focused on 5G and future digital connectivity solutions. Established with UK government funding, its mission is to support the adoption of 5G/6G technologies among SMEs. DCIC is part of the UK’s ten innovation regions designed to accelerate the adoption of emerging technologies. The Centre has a particular focus on supporting advanced manufacturing through private 5G networks across western Scotland.
Collaboration between UWS and BubbleRAN started in 2024 when UWS researchers needed to innovate on the network without spending months rebuilding infrastructure. The partnership began with BubbleRAN’s pre-configured deployment blueprints that significantly simplified the outroll of the initial network setup. What started as a faster way to deploy infrastructure over time evolved into a fully trained team with the confidence and capability to work across the entire network stack. The telco expertise BubbleRAN engineering teams transmitted significantly helped upgrade lasting in-house expertise at UWS.
DCIC wanted researchers to be able to manipulate the network itself: to deploy different network functions, experiment with slicing, integrate edge applications, work with xApps and collect detailed network telemetry. Building all of that manually from separate open-source components would have been certainly possible, but the integration,configuration and ongoing maintenance can consume a significant amount of research time.
*These are wider DCIC and Ayrshire programme figures illustrating the scale of the environment within which the BubbleRAN platform is being used
Challenge of interoperability
At the same time, DCIC had another important requirement: interoperability. They were building an environment involving commercial private 5G, open-source software-based solutions, UWS infrastructure, edge computing and equipment from multiple vendors. They therefore needed something sufficiently open and flexible to become part of that wider architecture rather than another isolated testbed.
“BubbleRAN helped us establish that programmable research layer while still allowing us to work with technologies such as OpenAirInterface, FlexRIC and Kubernetes. That gave our researchers much greater freedom to concentrate on experimentation and innovation rather than repeatedly rebuilding the underlying network environment. The platform reduces repetitive infrastructure work, while still giving researchers access to the underlying network and the freedom to innovate on top of it.“ says Pablo Salva-Garcia, Lecturer and Network Lead at UWS DCIC.
DCIC teams particularly appreciate BubbleRAN’s hands-on support, with engineers visiting the premises, practical training sessions, effective collaboration, and ongoing guidance. The teams have maintained a direct communication channel and high responsiveness is appreciated for research environments where the configuration naturally evolves over time.
Key element of the Open RAN infrastructure
Today, BubbleRAN MX-PDK O-RAN forms a key part of the Open RAN infrastructure at Digital Connectivity and Innovation Centre. The Centre has two complementary 5G environments. One is an enterprise-grade standalone private 5G network, developed with Vodafone and Ericsson, which is particularly suited to production-oriented and industrial deployments. The other is a more open and programmable research environment based on BubbleRAN MX-PDK O-RAN.
The BubbleRAN platform gives DCIC a cloud-native, disaggregated O-RAN environment where they can work across the CU, DU, radio units, 5G Core, edge computing and network-control functions. It is a platform not only for current 5G research, but also for experimenting with technologies technologies and architectural concepts relevant to 5G+ and future 6G networks.
“We want to go beyond simply using connectivity: we want to understand, programme and innovate across the network itself. The platform was specifically designed to support future developments. We see MX-PDK O-RAN as shared research infrastructure rather than simply equipment for a single telecoms research team.“ Says Pablo Salva-Garcia, Lecturer and Network Lead UWS DCIC.
At DCIC premises BubbleRAN MX-PDK O-RAN is used in different ways. First, it allows the demonstration of the benefits of 5G+ technologies to businesses, industries and other stakeholders across the Ayrshire region. This enables teams to reply to a particular operational problem, identified by industry partners and, and then work on its connectivity requirements, and develop a tailored demonstration. This enforces the important link between research and industry.
Second, it is a research platform where academics and research students experiment with network slicing, xApps, edge computing, AI-driven network management, UAV connectivity and increasingly autonomous network operations. Experiments undertaken at the Centre have supported work on areas such as industrial 5G, UAV surveillance,network slicing and AI-driven network operations, including research presented through international conferences and published in recognised journals.
Third, BubbleRAN MX-PDK O-RAN is also an important learning and teaching environment at DCIC. Because the solution is based on a programmable and accessible ecosystem, teams can develop practical workshops and learning activities that allow the next generation of students and researchers to gain hands-on knowledge of technologies that would otherwise be difficult to access in a conventional university laboratory. As the solution continuously evolves with regular releases, also the user community continues to grow within the Centre.
The platform is also helping teams to strengthen skills development as students and researchers can work with a real cloud-native 5G/O-RAN architecture consisting of two complementary clusters. The partnership has gone beyond the initial plan as the interest in artificial intelligence increased, and DCIC teams expanded with BubbleRAN’s platform’s AI capabilities allowing their teams to expand beyond traditional telecommunications research into AI-based network solutions and autonomous network operations.
Supporting manufacturing industry with real-world use cases
The aerospace sector is particularly important to DCIC and strong links are in place with the Prestwick aerospace ecosystem enabling it to connect advanced digital infrastructure with the wider commercial aviation, aerospace, space and advanced-manufacturing activities taking place in South Ayrshire.
It is important to focus on demonstrating technologies through use cases that are recognisable and relevant to industrial organisations rather than simply demonstrating 5G in isolation. DCIC has already demonstrated an aerospace surveillance scenario including autonomous drones, transmission of live video and network slicing where the BubbleRAN environment was used to create differentiated network services: a slice for the mission-critical drone video, another for applications such as AR/VR training, and best-effort connectivity for background traffic. The experimental architecture was built directly on the BubbleRAN MX-PDK O-RAN platform.
What makes this example demonstration particularly effective is what happens when the network is deliberately congested. Without prioritisation, the end-to-end latency of the drone application increased from around 12 milliseconds to 85 milliseconds. When we activated the URLLC-oriented slice, latency reduced again to below 15 milliseconds, with zero packet loss.
This kind of demonstration transforms network slicing from an abstract telecommunications concept into something that an industrial stakeholder can actually see. The network becomes congested, the video performance deteriorates, the slice is activated, and the mission-critical application continues to operate correctly. For DCIC and BubbleRAN these types of demonstrations help bridge the gap between advanced academic research and an industrial use case. And from an industry-engagement perspective, BubbleRAN and UWS contribute to a much wider DCIC programme that has engaged more than 50 businesses directly and sits within the Scottish Ayrshire innovation ecosystem.
Rather than treating DCIC premises as an isolated laboratory environment, the objective has been to integrate the platform with the Centre’s wider networking, edge-computing, monitoring and 5G infrastructure. That gives the Centre a foundation for a living laboratory where researchers can develop and validate an idea in a controlled environment before evaluating it under increasingly realistic conditions. This is particularly important because many 5G and 6G concepts need to move beyond simulation if they are to be understood in operational environments.
Looking ahead, DCIC and BubbleRAN are expanding their collaboration into emerging areas such as network digital twins and Non-Terrestrial Networks (NTNs). Network digital twins could help Scottish manufacturers test network changes safely before deployment, accelerate fault diagnosis and reduce operational risk. For Scotland’s coasts, NTNs offer another promising direction by extending connectivity where terrestrial coverage can be uneven, supporting ports, ferries, vessels and emergency services with more resilient communications.
“Integrating artificial intelligence into the RAN, testing network configurations, and developing new approaches tailored to advanced manufacturing is becoming increasingly important for identifying industry-specific use cases.” Says Muhammad Zeeshan Shakir, Professor of Digital Technologies and AI at UWS and Director of the DCIC.
Expanding the partnership into future connectivity
Another outcome of successful joint work between UWS and BubbleRAN is a submitted research paper by Doctor Pablo Salva-Garcia titled “State-Grounded Evaluation of LLM Agents for 5G and O-RAN Network Operations”. The work addresses a question that is becoming central to the future of autonomous networks: how do we know whether an LLM-based agent can actually be trusted to operate a real 5G network?
The framework builds on BubbleRAN’s Agentic Toolkit (BAT), extending it with additional safety mechanisms, state validation, and deterministic evaluation methods that enable reliable comparison of different LLMs for autonomous network operations. The research is based on extensive experiments conducted on BubbleRAN infrastructure, where multiple local and cloud LLMs were evaluated across a range of network management tasks. The scale of the study is significant: 18 different models across 1,890 agent runs, using a frozen suite of 35 operational tasks (SLA, observability, safety guardrails…) allowing comparison of not only whether models give good answers, but whether they behave correctly when they are actually interacting with a network.
DCIC and BubbleRAN share a global wider vision of AI-native and increasingly autonomous 5G-Advanced and 6G-ready networks. “If future networks are going to rely on intelligent agents to monitor conditions, diagnose problems and make configuration decisions, then robust and state-aware evaluation becomes a prerequisite for trust. The collaboration with BubbleRAN gives our teams at DCIC a real platform on which we can move that research from theory into operational industry-linked experimentation.” says Muhammad Zeeshan Shakir, Professor of Digital Technologies and AI at UWS and Director of DCIC.
Both DCIC and BubbleRAN are engaged actors of future connectivity. Together they are advancing industrial and academic partners to translate future connectivity technologies into economic and societal benefit. By combining programmable Open RAN infrastructure with AI-driven network research, DCIC and BubbleRAN are helping accelerate Scotland’s journey toward commercially relevant 5G+ innovation.