RAID: Root Cause Anomaly Identification and Diagnosis

Abstract

Wi-Fi networks are widely used for modern connectivity but remain vulnerable to impairments such as bandwidth fluctuations, interference, packet loss and latency spikes. These challenges make it difficult to support latency-sensitive applications like Cloud Virtual Reality (Cloud VR), which offloads intensive computation to remote servers to reduce local hardware requirements but demands high throughput and ultra-low latency. Consequently, Wi-Fi network degradations can severely impact the Quality of Experience (QoE) of such applications. Traditional Root Cause Diagnosis (RCD) approaches rely on expert-defined rules or supervised ML (Machine Learning) models that require extensive labeled datasets. This dependence on manual labeling makes them costly, time-consuming, and impractical for real-world Wi-Fi diagnostics. To overcome these limitations, we introduce RAID (Root cause Anomaly Identification and Diagnosis), a two-stage ML framework that diagnoses Wi-Fi performance issues using time series KPIs collected directly from the Wi-Fi access point, with Cloud VR serving as a use case. RAID combines contrastive learning-based anomaly detection with a lightweight classifier to categorize network impairments. We evaluate RAID, with a real-world Cloud VR use case, in a testbed using NVIDIA CloudXR and a Meta Quest 2, collecting Wi-Fi performance metrics on the access point, under controlled conditions. Results demonstrate that RAID outperforms existing RCD methods, achieving high accuracy even with minimal labeled data. Compared to conventional supervised and self-supervised time series models, RAID offers a scalable, real-time solution with a good trade-off between training efficiency and inference speed, making it well-suited for practical deployment in dynamic Wi-Fi network environments.