SIMULATED CHARACTERIZATION OF 5G NETWORK RESILIENCE TO ADVERSARIAL ATTACK

Adversarial attack is an attack tactics employed by cybercriminals to deceive modern network security systems and penetrate networks for cyber attack. While several studies focused on known threats, there has been gap in their inability to detect unknown threat features, but this weakness can only be identified through characterization of a testbed network. The aim of this study is simulated characterization of 5G network resilience to adversarial attack. The testbeds are the Networks of Nigerian immigration services passport offices at Awka and Enugu, both in Nigeria. The testbed characterization involved simulation of diverse adversarial attack vectors on the network. The parameters considered for data collection are latency, throughput, packet loss, Mean Time to Response (MTTR), error rate and Mean Time to Detect (MTTD). The results obtained for the Enugu network showed that throughput, latency, packet loss, MTTR, error rate and MTTD recorded values of 68.87Mbps, 87.09ms, 7.9%, 393ms, 3.89, and97.10ms respectively. Similarly, the Awka network recorded a throughput of 90.26Mbps, latency of 39.45ms, packet loss of 1.65%, MTTR of 183ms, error rate of 0.685, and MTTD of 35.6ms. Overall, these results implied that during the penetration test on the two network facilities, the quality of service was affected as it degrades, showing that the traditional security solution was not sensitive to adversarial threat features. To address this problem, this paper recommends development of adaptive adversarial attack detection system and then integrating to the testbed for the improved security against adversarial attack.