Anomaly Detection > Implementing Anomaly Detection
Implementing Anomaly Detection
Deployment Methods
Anomaly Detection is implemented using both Analytics Server and ThingWorx to monitor streaming data and detect possible anomalous values. Beginning in the 9.0 release, Anomaly Detection functionality can be deployed in two ways:
• Standalone – Deployed on a single ThingWorx server, where a new anomaly monitor is generated each time an anomaly alert is added to, or updated on, a Thing property.
• Cluster Mode – Deployed on a cluster of ThingWorx servers, where anomaly monitor calculations are handled outside of ThingWorx so that each server in the cluster can send property updates to a central location.
The standalone deployment does not require any additional components, beyond Analytics Server and ThingWorx. However, the cluster mode deployment requires installation of multiple Property Transform components available via the Platform Analytics installer. For more information about Anomaly Detection in a ThingWorx cluster, see the following information:
• The Server Based Third Party Applications section of the Platform Analytics release matrix available on the PTC Release Advisor. For more information, see Using Release Advisor.
Anomaly Detection Configuration
Anomaly Detection is enabled by default in ThingWorx. However, the following procedures might be useful to configure and run the functionality for your specific environment:
• Prepare ThingWorx for Anomaly Detection by creating a remote Thing to represent your edge device and configuring properties that are bound to the edge device.
• Create an anomaly alert to launch the anomaly detection process for a specific property.
• Migrate existing anomaly models that are already bound to Thing properties for the purpose of Anomaly Detection.
Was this helpful?