The ShieldIOT AI-based network anomaly detection solution analyzes device communications to identify the first signs of cyber-attacks and operational anomalies. Deployed at the IoT network operator / service provider cloud, the agent-less solution is highly scalable, protecting massive multi-application IoT networks against various threats including DDoS, unknown malware, ransomware, communication hijacking, node takeover and replication, replay attacks, data falsification and many more. Features:

• Unmatched Accuracy. ShieldIOT Corset-based AI anomaly detection is x5 to x20 more accurate compared to leading industry solutions
• Real-Time Detection. From one week to one minute. No speed or accuracy trade-offs. Detect the first signs of attack before damage takes place.​
• Actionable Alerts. Attack type, duration, affected devices, score, severity, trigger features… Enabling automated or non-expert threat mitigation
• Unlimited Scalabilty. Transforms huge volumes of big data into small data sets. Agent-less solution with no device or network changes

How it works The ShieldIOT AI network security is based on coreset real time streaming anomaly detection. The solution analyzes network traffic and additional data sources, to detect the first signs of cyber-attacks, while providing unmatched out-of-the-box event forensic to facilitate threat mitigation actions.

1. Data Processing. Data filtering, null completion, data sources joining, data enrichment, feature engineering
2. AI Engine. Coreset computing to reduce the data, followed by running a set of AI and machine learning detection models.
3. Anomaly Classification. Automatically generate event type, by combining patented labeling process with domain expertise