Researchers from the School of Informatics and Microsoft Research have developed a new open-source system that could make large-scale mobile network experimentation significantly more affordable and accessible. The work was presented at ACM MobiCom 2026, one of the leading international conferences in mobile computing and networking, in a paper titled Chronos: Scalable and Cost-Effective Cloud-based RAN Emulation System. The project was led by Ujjwal Pawar and Andrew Ferguson, both PhD students in the School of Informatics advised by Prof. Mahesh Marina, alongside Yufeng Xia from the School of Informatics, and Xenofon Foukas and Bozidar Radunovic from Microsoft Research. Making large-scale mobile network research more accessible Modern mobile networks increasingly rely on software and artificial intelligence to manage performance, reliability and security. Before new technologies can be deployed, researchers need realistic environments in which to test them at scale. For many researchers, this presents a problem. Access to large-scale mobile network infrastructure is expensive and often restricted to telecom operators or organisations with substantial resources. Existing software-based alternatives are more affordable but typically struggle to accurately model more than a few hundred mobile base stations. This can make it difficult for researchers developing new approaches to areas such as network optimisation, energy efficiency, security and AI-driven network management. Simulating a city-scale mobile network The team's new system, Chronos, offers a different approach. Instead of requiring specialised hardware, Chronos runs real mobile network software in the cloud. By taking advantage of cloud computing's pay-as-you-go model, researchers can conduct large experiments without the cost of building and maintaining dedicated infrastructure. The researchers demonstrated Chronos emulating a network containing 10,000 mobile devices and 10,000 base stations. According to the team, this is more than two orders of magnitude larger than existing software-based alternatives can typically support. This could give universities, startups and smaller organisations access to experiments that were previously out of reach. Our motivation behind Chronos was to build a platform where researchers could experiment with city-scale mobile networks, essentially creating a high-fidelity digital twin of a city like Edinburgh or London. Ujjwal Pawar School of Informatics, The University of Edinburgh Running real mobile network software in the cloud Running mobile network software in a cloud environment is far from straightforward. Radio Access Network (RAN) software normally operates on highly specialised hardware and follows strict timing requirements measured in fractions of a millisecond. Delays caused by shared cloud infrastructure can result in unrealistic behaviour or system failures. Chronos addresses this challenge by virtualising time and carefully synchronising processes across the network. From the software's perspective, it behaves as though it is running on dedicated infrastructure while actually running in the cloud. The team's experiments showed that the system could scale to thousands of devices while closely matching the behaviour observed on real 5G testbeds. The cloud was a natural choice because it allows us to scale computing resources with the size of the emulated network. However, making this work in practice was challenging, as the multi-tenant nature of cloud infrastructure offers no guarantees on computing or network performance, which are critical for real-time systems like RAN.By combining our understanding of RAN design with systems-level research, we developed a way to virtualise time and shield the emulation from these unpredictable delays. Ujjwal Pawar School of Informatics, The University of Edinburgh Open source for the wider community The researchers hope Chronos will make large-scale mobile network experimentation accessible to a wider research community. The project is being released as open source, allowing researchers, startups and organisations to use, modify and build upon the system. By reducing the cost and complexity of experimentation, the team hope Chronos will support future innovation in software-driven and AI-enabled mobile networks. Chronos can now emulate mobile networks at a scale two orders of magnitude greater than previous solutions, bringing us closer to making city-scale mobile network digital twins a practical and accessible research tool. Ujjwal Pawar School of Informatics, The University of Edinburgh Read more Chronos: Scalable and Cost-Effective Cloud-based RAN Emulation System This article was published on Friday 6 November 2026
