UDPI
Real-Time Network Threat Detection with Deep Learning
Neural network-based packet analysis delivering sub-100ms intrusion detection with 97%+ accuracy
How UDPI Protects Your Network
UDPI operates as a silent guardian on your network, analyzing every packet in real-time using machine learning models trained on millions of attack samples.
What UDPI Detects
Performance You Can Measure
Detection Accuracy
| Dataset | Samples | Threat Types | Accuracy |
|---|---|---|---|
| NSL-KDD | 125,973 | DoS, Probe, R2L, U2R | 97.8% |
| CIC-IDS2017 | 2.8M+ | DDoS, PortScan, Web, Botnet | 96.4% |
| CIDDS-001 | 671,241 | PortScan, BruteForce, DoS | 96.8% |
* Shipped UDPI may not use these datasets as training data
Privacy & Security
- Local processing only, no cloud dependency
- IP addresses anonymized in logs
- Configurable data retention policies
- Drop-root privileges after initialization
- SafeTensors format for model loading
Who Uses UDPI
Part of the uWAN Security Platform
UDPI operates as the core detection engine within UWAN, providing real-time threat intelligence
Complementary UWAN Components
- Firewall Integration (automatic rule updates)
- SIEM Correlation (ML confidence scores)
- Threat Intelligence (global attack database)
- Incident Response (automated playbooks)
Extensibility
Getting Started with UDPI
Production Deployment
Install
Download the package for your platform (macOS, Linux x86_64, Linux ARM64)
Configure
Specify network interface and detection preferences in the configuration file
Deploy
Run the service—automatically initializes and begins monitoring
Integrate
Connect alert outputs to your existing SIEM or security workflows
Model Training Data:
- • Synthetic Generated Data (algorithmically generated patterns)
- • Public Domain Datasets (government and research data)
- • Customer Data (your own network traffic)
All models ship with permissive licenses suitable for commercial deployment.
Ready to enhance your network security?
Get in touch with our team to learn more about UDPI and the uWAN security platform.
