Understanding the Security Implications in O-RAN with Abusive Adversaries

Open-Radio Access Network (O-RAN) is considered the next scalable solution, which aims to devolve the network into Near-real-time RIC and Non-real-time RIC to have far more flexibility in services with adaptable components. This disaggregation, however, will have broader security implications, primarily arising because of the use of legacy systems in the new architecture. Current threat models take a lighter tone towards the evaluation of security measures. Thus, strict adversarial methods must be adopted, which can consider scenarios of cyber-vandalism in such networks. Based on this ideology, the article presents security implications posed by abusive adversaries in the offloading procedures. This methodology provides a viewpoint on how an adversary forms predictive methods on when to attack the system, which is followed by mitigation mechanisms for the network to avoid it from happening. The work is based on the Markov Decision Process (MDP) and a Fuzzy Inference System (FIS), which uses Synthetic Data Augmentation for Tabular Data (SMOTE) to generate a set of metrics that can offer a high probability of attack in the transition mode to the adversary. The implications are presented using a synthetic dataset created on the backbone of the simulated scenario in NS3 and followed by mitigation strategies.