We study architectures and technologies for deploying and orchestrating programmable networks and distributed computing systems at the edge of the Internet. Our research combines analytical models, optimisation, AI, software platforms, and experiments.
Moving computation closer to people and devices. We design stateful serverless platforms, computation offloading, and adaptive service placement across the IoT–edge–cloud continuum. Our work connects application requirements to the resources available at the network edge: where to run functions, how to manage state, and how to balance latency, reliability, and energy consumption.
Making 5G/6G systems responsive to application needs. We investigate new architectures and control paradigms for disaggregated, software-defined 5G/6G systems. Our goal is to coordinate communication, computing, and services across heterogeneous infrastructures, enabling autonomous, intent-driven orchestration that adapts to mobility, changing workloads, and application requirements while meeting performance, reliability, and energy-efficiency targets.
Applying machine learning and reinforcement learning to optimise resource allocation, service placement, traffic routing, and network monitoring. We use AI to make networked systems more adaptive, efficient, and responsive. Our work includes lightweight inference in programmable data planes and sim-to-real learning from radio signals for localisation and environment mapping. We also investigate how multi-agent systems and agentic AI can support intent-based orchestration and coordinated decisions across network and computing resources.
Coordinating networks and edge computing to deliver predictable latency and reliable services for extended reality (XR), metaverse applications, and industrial systems. We study joint rendering-server placement, routing, scheduling, and flow prioritisation across 5G, mobile edge computing, and time-sensitive networks. Our models connect end-to-end delay and resource sharing to user experience, including XR-assisted collaboration and productivity in manufacturing. We also consider avatar management in networked virtual environments, including aspects of avatar behaviour, diversity, and identity.
Designing networking and computing systems that adapt to the tasks of connected and autonomous vehicles, with cooperative perception, computation offloading, and reliable communication in dynamic environments. We investigate task-aware orchestration across vehicles, drones, and roadside infrastructure. Our work explores how 5G, Wi-Fi, millimetre-wave, visible-light, and vehicle-to-everything links can work together to support mobility and changing mission requirements.
Building and using testbeds to evaluate programmable networks, edge platforms, and service orchestration under realistic conditions, with an emphasis on reproducible experiments.
Explore our research infrastructures →