Implicit Channel Coordination to Tackle Starvation Attacks in 5G and Wi-Fi Coexistence Systems

Due to the scarcity of spectrum bands, 5G and Wi-Fi systems are embracing coexistence in the unlicensed 5 and 6 GHz bands to support high data rate demands and growing number of users. To provide fair and effective coexistence in shared frequency bands, both technologies rely on carrier sensing. However, differences in sensing thresholds creates an unfair advantage for 5G nodes who access the shared wireless medium more aggressively and degrade the data rate and latency of Wi-Fi nodes. We show in this paper that an adversary who intends to deny service to Wi-Fi can stealthily exploit this unfairness to drastically reduce the spectrum occupancy and data rate of Wi-Fi nodes. We then propose a novel implicit channel coordination (ICC) approach to mitigate the attack and improve sharing fairness. In ICC, Wi-Fi nodes influence 5G gNB into choosing a precoding matrix that nearly nullifies downlink signals at Wi-Fi nodes. We demonstrate our starvation attack on a USRP testbed and further evaluate our proposed ICC approach using simulations. We show that ICC doubles the data rate of a Wi-Fi network subject to an active attack.