O-RAN xApps Conflict Management using Graph Convolutional Networks

TL;DR

The paper addresses conflict management in O-RAN by modeling xApp interactions as a graph problem and predicting conflict types with a Graph Convolutional Network-based framework called GRAPHICA. It introduces binary-state data encoding, a Graph Structure Creator with three subgraphs, a Graph Anomaly Predictor using a two-layer GCN and focal loss, and a Root Cause Analyst to extract responsible xApps. Across highly imbalanced synthetic datasets, GRAPHICA achieves $F1$-scores exceeding $0.98$, with optimal focal-loss parameters yielding near-perfect detection for rare conflicts. The approach enables proactive mitigation by pinpointing root-causes, and it demonstrates robustness and generalizability, with future work targeting predictive maintenance and real-world deployment in multi-vendor O-RAN.

Abstract

The lack of a unified mechanism to coordinate and prioritize the actions of different applications can create three types of conflicts (direct, indirect, and implicit). Conflict management in O-RAN refers to the process of identifying and resolving conflicts between network applications. In our paper, we introduce a novel data-driven GCN-based method called GRAPH-based Intelligent xApp Conflict Prediction and Analysis (GRAPHICA) based on Graph Convolutional Network (GCN). It predicts three types of conflicts (direct, indirect, and implicit) and pinpoints the root causes (xApps). GRAPHICA captures the complex and hidden dependencies among the xApps, controlled parameters, and KPIs in O-RAN to predict possible conflicts. Then, it identifies the root causes (xApps) contributing to the predicted conflicts. The proposed method was tested on highly imbalanced synthesized datasets where conflict instances range from 40% to 10%. The model is tested in a setting that simulates real-world scenarios where conflicts are rare to assess its performance. Experimental results demonstrate a high F1-score over 98% for the synthesized datasets with different levels of class imbalance.

Figures (9)

• Figure 1: Examples of xApps conflicts.
• Figure 2: The proposed framework.
• Figure 3: The chronological sequence of events for conflicts prediction.
• Figure 4: The binary-state data creation module steps.
• Figure 5: The graph structure creator (GSC) module.