Automated Root Cause Analysis of Network Failures in IPMPLS Network Using Machine Learning and Case-Based Reasoning

Modern IPMPLS (Internet Protocol – Multi-Protocol Label Switching) networks form the backbone of enterprise and service provider infrastructures, where uptime and reliability are critical. However, due to the complex and dynamic nature of such networks, diagnosing the root cause of failures remains a significant operational challenge. This project proposes an intelligent, automated framework for root cause analysis using a hybrid approach of Machine Learning (ML) and Case-Based Reasoning (CBR). The system integrates user-friendly web functionalities built with Flask, including dataset upload, model selection, prediction interfaces, and visualization modules.

Upon uploading packet capture or flow-level data, the system preprocesses the input and applies a trained Decision Tree or ensemble model to classify the anomaly level. Each classification is mapped to human-readable insights ranging from benign traffic to confirmed threat levels. Further, the architecture leverages historical cases and associated resolutions to offer contextual root cause insights, thereby reducing downtime and manual intervention. Performance metrics of various models including XGBoost, Random Forest, Logistic Regression, and Stacking Classifiers are also reported to guide algorithm selection. This solution empowers network administrators with faster, more accurate fault identification, enabling proactive remediation strategies in complex IPMPLS environments.

Keywords:
IPMPLS Networks, Root Cause Analysis, Machine Learning, Case-Based Reasoning, Anomaly Detection, Flask Web Application, Decision Tree Classifier, Network Security, Automated Diagnosis, Predictive Analytics