Adaptive WAN Link Anomaly Detection Using Lightweight Packet-Level Features for Branch-to-HQ Network Stability
Bibliographic record
Abstract
Maintaining stable Wide Area Network (WAN) connectivity between branch offices and centralized headquarters infrastructure is essential for the reliable operation of modern enterprise systems. However, WAN links frequently experience performance degradation caused by congestion, routing instability, and intermittent packet loss, which can significantly disrupt enterprise services such as cloud applications, real-time communications, and data synchronization. This study proposes a lightweight anomaly detection framework designed to monitor branch-to-headquarters WAN links using packet-level telemetry features. The framework utilizes compact statistical indicators derived from packet transmission behaviour, including packet loss ratio, latency deviation, and jitter variance, to characterize network performance conditions. An adaptive anomaly detection model is implemented using a dynamic threshold formulation that adjusts detection boundaries based on the moving average and statistical variance of observed network metrics. The proposed model enables real-time identification of network anomalies while maintaining low computational overhead suitable for deployment on resource-constrained branch routers. Experimental evaluation was conducted using simulated WAN environments representing stable traffic conditions, congestion-induced anomalies, and intermittent packet loss events. The results demonstrate that the lightweight monitoring framework achieves high detection accuracy while maintaining low false-positive rates and reduced detection latency compared with conventional monitoring approaches. The findings indicate that combining packet-level feature engineering with adaptive statistical detection provides an effective and scalable solution for improving WAN stability monitoring in distributed enterprise networks.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".