INESC TEC aims to make the dynamic and unpredictable conditions of 5G network testbeds reproducible 

Researchers at INESC TEC have developed a solution that makes it possible to repeat and reproduce any experimental scenario using a simulator. 

INESC TEC created digital twins of 5G network testing infrastructures (testbeds) that are able to accurately repeat and reproduce past scenarios and experiments -even when the testbeds themselves or the experimental conditions change. 

This ability to precisely reproduce complex experimental conditions can significantly advance real-world simulations. Wireless experiments are typically difficult to repeat under identical conditions, particularly those involving mobile terminals. 

There are several reasons why experiments cannot always be replicated: the testing platform may be unavailable (offline or busy running other experiments) or completely inaccessible when it has been modified (i.e., when experimental conditions change due to alterations in topology or hardware), or even decommissioned. 

The newly developed solution enables past experimental conditions and their corresponding results to be repeated and reproduced in simulation. This advance was achieved within the scope of the REPLICA – Replicable Cellular Networking Experiments using ns-3 project, one of the winners – and the top-ranked proposal – of the second open call of the European project 6G-SANDBOX (SNS-JU). The 6G-SANDBOX initiative is developing a comprehensive and modular tool for the European experimentation ecosystem, in line with the guidelines defined by the EU Joint Undertaking for 5G and 6G networks and services. 

This raises an important question: what if the same execution conditions of the 6G-SANDBOX could be easily made available to researchers, allowing any experimental scenario to be readily repeatable and reproducible using the ns-3 simulator? 

This simulator plays a key role: ns-3 is essentially a virtual networking laboratory – a free, open-source computer network simulator widely used in communications research and development (R&D). 

“By reproducing signal quality in ns-3, it was possible to obtain simulation results that are much closer to real-world behaviour than those achieved using purely simulated models. This work builds on more than seven years of research on wireless network digital twins, which had been extensively validated for Wi-Fi in projects such as SIMBED and SIMBED+, and is now being applied to 5G networks for the first time,” explained Hélder Fontes, researcher at INESC TEC. 

INESC TEC has a vast experience with aerial and maritime testbeds and has authored many scientific publications on simulation and experimentation, while holding a recognised position within the ns-3 community. 

The institution has also been developing a trace-based simulation approach that allows past physical-layer traces – such as received signal strength (RSS) – to be recorded and replayed in the ns-3 network simulator. These traces can also be used to train machine-learning-based digital models tailored to the environment of each testbed, enabling the creation of more versatile and robust digital twins of wireless networks. Until now, this approach had been validated for Wi-Fi. Through REPLICA, within the 6G-SANDBOX framework, it has now been tested and validated for 5G/6G networks.