
Embedded AI

Sensor Operations
• Problem: Malicious insiders and malware use covert outbound connections (DNS, HTTP, custom protocols) to exfiltrate data without detection.
• Solution: Exfiltration Sensor monitors packet flows across all interfaces, tracks frequency/duration, compares against trusted host lists, and triggers alerts for abnormal patterns.
• Why Different: Embeds ONNX AI models for exfiltration scoring, anomaly detection, adaptive thresholds, and agentic AI updates. Supports vendor pattern recognition, syslog export, and JSON payloads.
• Syslog Output Fields: Timestamp, Sitename, Host, Service, Source IP, Destination IP, Packet Count, Packet Rate, Duration, Flags, Vendor, Threat Score, Anomaly Flag, Recommended Action.
Host Resource Requirements
-
CPU: < 1%
-
Memory(RAM) < 50MB
-
Disk: < 1MB per day log writes
-
Network: < 1KB/s syslog traffic






