Misconfiguration in O-RAN: Analysis of the impact of AI/ML

TL;DR

This paper investigates misconfiguration in O-RAN as a critical risk arising from openness and disaggregation in 5G/6G networks. It provides a structured analysis of misconfiguration modalities across integration/operation, SDN/NFV, and AI/ML, linking concrete examples to direct performance and indirect security threats. The authors survey AI/ML-based detection approaches—encompassing active and passive monitoring, formal verification, offline modeling, and Network Digital Twin—and illustrate them with a case study on conflicting xApps. The work highlights both opportunities and challenges for AI/ML-driven misconfiguration detection in O-RAN, emphasizing the need for standardization, rigorous verification, and realistic testbeds to harness O-RAN’s potential without compromising security or performance.

Abstract

User demand on network communication infrastructure has never been greater with applications such as extended reality, holographic telepresence, and wireless brain-computer interfaces challenging current networking capabilities. Open RAN (O-RAN) is critical to supporting new and anticipated uses of 6G and beyond. It promotes openness and standardisation, increased flexibility through the disaggregation of Radio Access Network (RAN) components, supports programmability, flexibility, and scalability with technologies such as Software-Defined Networking (SDN), Network Function Virtualization (NFV), and cloud, and brings automation through the RAN Intelligent Controller (RIC). Furthermore, the use of xApps, rApps, and Artificial Intelligence/Machine Learning (AI/ML) within the RIC enables efficient management of complex RAN operations. However, due to the open nature of O-RAN and its support for heterogeneous systems, the possibility of misconfiguration problems becomes critical. In this paper, we present a thorough analysis of the potential misconfiguration issues in O-RAN with respect to integration and operation, the use of SDN and NFV, and, specifically, the use of AI/ML. The opportunity for AI/ML to be used to identify these misconfigurations is investigated. A case study is presented to illustrate the direct impact on the end user of conflicting policies amongst xApps along with a potential AI/ML-based solution to this problem. This research presents a first analysis of the impact of AI/ML on misconfiguration challenges in O-RAN.

Figures (5)

• Figure 1: O-RAN architecture presented by the O-RAN Alliance and 3GPP ORANWG12023Architecture (the RAN is connected to the 5GC through the NG interface).
• Figure 2: Deployment options of AI/ML models in O-RAN. In gray boxes: Minimally or unexplored areas.
• Figure 3: Conflicting MLB and MRO. The dataset presented in Adamczyk2023Conflict has been used to illustrate the issue of ping-pong handovers (HOs) between two apps with conflicting objectives. For clarity, the KPIs of 2 gNBs are shown (the original dataset contains data for 19 gNBs). The MLB maintains the balance of the load on the gNBs (top plot), while the MRO maintains the RLFs close to zero (middle plot). The interaction of these xApps causes multiple ping-pong handovers as illustrated in the bottom plot.
• Figure 4: Managing conflicts between xApps: detection and mitigation using AI/ML techniques. Shaded components have been incorporated into the original Near-RT-RIC architecture of the O-RAN WG3 ORANWG32023RICARCH, which include the information database for xApp actions (xNIB) and the xApps KPIMON, AD, CD, and CM.
• Figure 5: Method for detection and mitigation of conflict xApps. Shaded components may require AI/ML techniques.