Behavior analysis to automate direct and indirect local monitoring of the health of Internet of Things devices

摘要

The present disclosure relates generally to behavioral analysis for automating the direct and / or indirect monitoring of the health of Internet of Things (IoT) devices. In particular, normal behavior associated with IoT devices in the local IoT network may be modeled, thereby comparing the behavior observed in the IoT device with the modeled normal behavior and being observed in the IoT device. It may be determined whether the behavior is normal or abnormal. Thus, in a distributed IoT environment, a higher performance “analyzer” device collects behaviors observed locally in other (eg, simpler) “observer” devices and analyzes behavior across the distributed IoT environment. To detect anomalies that could potentially indicate malicious attacks, malfunctions, or other problems that require customer service and / or further attention. In addition, a fully functional device may perform (local) on-device behavior analysis to detect abnormal conditions without sending the locally observed behavior to another aggregator device and / or analyzer device.