Completed
SDN-Based Machine Learning IDS
Network Security
Intrusion Detection System for Software-Defined Networks using ML/DL models
Python
TensorFlow
scikit-learn
Mininet
Ryu
CloudLab
Software-Defined Networking (SDN) architectures introduce unique security challenges due to their centralized control plane. This project implements an intelligent Intrusion Detection System using machine learning.
Machine Learning Models
- Random Forest Classifier: Ensemble method for robust baseline detection
- Deep Neural Networks: Multi-layer perceptron for complex pattern recognition
- LSTM Networks: Sequence-based detection for temporal attack patterns
CloudLab Deployment
- Real-time Traffic Monitoring with sub-second detection latency
- Scalable Multi-node Testing across multiple physical hosts
- Reproducible Experiments for benchmarking
Key Results
The system achieved >95% detection accuracy with real-time processing under 100ms latency.