Adversarial Machine Learning for Flooding Attacks on 5G Radio Access Network Slicing

Network slicing manages network resources as virtual resource blocks (RBs) for the 5G Radio Access Network (RAN). Each communication request comes with quality of experience (QoE) requirements such as throughput and latency/deadline, which can be met by assigning RBs, communication power, and processing power to the request. For a completed request, the achieved reward can be measured by the weight (priority) of this request. The accumulated reward over time should be maximized by allocating resources, e.g., with reinforcement learning. In this paper, we introduce a novel flooding attack on 5G network slicing, where an adversary generates fake network slicing requests to consume the 5G RAN resources that would be otherwise available to real requests. The adversary observes the spectrum and builds a request generation algorithm through reinforcement learning that decides on how to craft fake requests to maximize the reward of fake requests over time, which in turn minimizes the reward of real requests over time. We show that the portion of the reward achieved by real requests may be much less than the reward that would be achieved when there was no attack. We also show that this flooding attack is more effective than other benchmark attacks such as generating random fake requests and fake requests with the minimum resource requirement (lowest QoE requirement). Finally, we evaluate how the attack performance depends on system parameters such as the weight and the generation rate of fake requests, and the channel condition. Our results show that the flooding attack poses a major threat for network slicing in terms of starving 5G RAN resources and denying service to real network slicing requests.