
Embedded AI

Sensor Operations
• Problem: Critical business services can suffer performance degradation, availability loss, DoS -like saturation, or resource exhaustion —often before any alert is raised.
• Solution: VectorScan continuously probes critical service endpoints/ports, measures latency and availability, and emits concise cycle summaries that reveal faults, performance hotspots, and potential security incidents (e.g., DoS or saturated resources).
• Why Different: Combines high -frequency active probing with ONNX - powered anomaly detection and adaptive thresholds; AI is folded into aggregated summaries (minimizing noise) and supports autonomous model updates and registry -driven tuning.
• Syslog Output Fields: Timestamp, Sitename, Host, Service, Type (POLL/REDISCOVER), MeasuredCount, ActiveNow, Avg_ms, P95_ms, TopSlowEndpoints, WentDown, Recovered, AI_AnomaliesCount, AI_HighCount, AI_TopIssues (ip:port:score:issue).
Host Resource Requirements
-
CPU: < 1%
-
Memory(RAM) < 50MB
-
Disk: < 1MB per day log writes
-
Network: < 1KB/s syslog traffic




