Researchers at the University of Glasgow have developed a new method of testing networks that they claim is 25,000 times faster than traditional approaches.
Shenjia Ding, a research student in the university’s School of Computing Science, tested two complex American and European computer networks with 12 and 37 nodes, respectively, using automatically generated digital twins built with machine learning.
The tests included six different types of traffic, including web browsing, video streaming, and file downloads, as well as continuous congestion and background noise to simulate real-world conditions.
The team’s digital twin took 4.78 seconds to test the network speed. In comparison, running the same test using a traditional simulator took 33 hours.
As internet traffic and data volumes grow exponentially, the scientists say their approach has the potential to become a “practical, scalable and cost-effective” approach to network testing and management.
Traditional network testing uses simulators to mimic real-world scenarios and data traffic to test the performance, security, and reliability of computer networks. The researchers used automated machine learning (AutoML) to build the digital twin, which they say not only speeds up the process of building machine learning tools, but also makes them accessible to non-experts with limited machine learning expertise.
“Our results show that testing computer networks using automatically generated digital twins can achieve higher accuracy and significantly faster speed than traditional simulator-based testing,” Ding said. “We demonstrate a very promising alternative to time-consuming manual testing that relies heavily on expert knowledge.”
Testing computer networks using automatically generated digital twins can provide higher accuracy and significantly faster testing than traditional simulator-based testing.
Shenjia Ding, Department of Computing Science, University of Glasgow
Paul Harvey, a co-author of the study and a senior lecturer in the School of Computing Science at the University of Glasgow, is also a co-investigator on TransiT, a collaboration between Heriot-Watt University in Edinburgh, the University of Glasgow and 70 industry partners, funded by the UK Research and Innovation Engineering and Physical Sciences Research Council.
TransiT is seeking to identify the quickest, lowest risk and lowest cost route to decarbonisation in the UK. Harvey believes this research shows how using machine learning to build digital twins can be applied to other network settings, such as transportation.
“As with computing, data volumes are increasing significantly in transportation, and in both cases the pressure on the communication networks that carry all this data is immense,” Harvey explained.
“By proving that machine learning can be used to build digital twins, which is also a time-consuming and labor-intensive task, we highlight the huge potential of this research to also enable testing and optimization of the transportation and other networks we rely on every day.”
He said Ding’s research could assist TransiT in its goal of creating a “digital twin factory” that can automate the production of digital twins, especially for transportation systems.
The researchers will focus on validating digital twin update mechanisms and costs, evaluating their performance in real-time network environments, and conducting comparative studies across different network scenarios.
Ding submits a paper. Automated generation of digital twins for network testing: Multi-topology validationwill explore the use of automated digital twins in network management at the 2026 IEEE International Communications Conference (ICC) in Glasgow later this month.
The paper was co-authored by Paul Harvey and David Flynn from the University of Glasgow.
