A novel AI-powered defense system, TwinGuard, developed by researchers at the University of Surrey, can identify and mitigate advanced 5G cyber-attacks in less than 100 milliseconds. This system is designed to bolster the security framework of both existing 5G networks and upcoming 6G infrastructures, addressing new vulnerabilities linked to the increasing openness of mobile networks.
The flexible nature of modern 5G systems, while beneficial for cost reductions and upgrades, has inadvertently expanded the attack surface for cybercriminals. TwinGuard utilizes a digital twin model, providing a live virtual representation of a mobile network that updates frequently, combined with reinforcement learning AI to proactively detect and counter threats before they disrupt operations.
In tests simulating real-world 5G scenarios, including a multi-cell Open Radio Access Network and a fully virtual 5G core network, TwinGuard effectively blocked several types of attacks. These included a handover flooding attack and an E2 subscription flooding attack, showcasing its capability to adapt to dynamic threats. Dr. Sotiris Moschoyiannis, the study's lead, stressed the importance of such adaptive security systems in maintaining the integrity of future networks.