
Embedded AI

Sensor Operations
• Problem: Network performance issues (latency spikes, link flaps, ISP routing anomalies) often remain hidden until end users complain.
• Solution: tRouteSensor runs scheduled traceroutes, stores hop RTTs in a local SQLite DB, purges daily, and streams hop and anomaly data to syslog.
• Why Different: Embeds ONNX AI models for hop -level anomaly scoring, adaptive thresholds, and 30 -minute AI summaries. Supports autonomous model updates, registry - driven config, and syslog with site tagging.
• Syslog Output Fields: Timestamp, Sitename, Host, Service, Hop #, IP, RTT (ms), Destination, Anomaly Score, Anomaly Level, Issue Type, Recommended Action.
Host Resource Requirements
-
CPU: < 1%
-
Memory(RAM) < 50MB
-
Disk: < 1MB per day log writes
-
Network: < 1KB/s syslog traffic




