Learn about PI Connector for MQTT Sparkplug operations

Each PI connector is designed for a specific data source. PI Connector for MQTT Sparkplug queries the data source to discover assets, relations between assets, time-series measurements, and relations between assets and time-series measurements that the data source contains. The connector creates generalized streams to convey the assets, relationships, and time-series measurements to one or more PI Connector Relay hosts.

The connector output streams are generalized in the sense that they are independent of the databases or historians that store data from the streams. For more information about connectors in general, refer to the Overview of PI Connectors user guide.

The following figure shows the data flow from a data source to the PI System.

PI connector architecture diagram

In this figure, the connector communicates with PI Connector Relay over an encrypted TCP connection, which allows it to be on a different computer.

Although the figure shows only one PI Connector Relay host, a connector can replicate its output streams to multiple PI AF hosts. Similarly, a PI Connector Relay can support multiple PI AF servers and Data Archive servers, including collectives.

The connector does not communicate directly to any type of historian or database. Streams from the connector are strictly one-way to the PI Connector Relay host, which means that it cannot obtain any information from either PI AF or Data Archive.

Both the connector and PI Connector Relay communicate with PI Data Collection Manager, which is used to specify all settings for the configuration. For additional information about PI Connector Relay and PI Data Collection Manager, refer to PI Connector Administration user guide.

Asset creation and data collection

From the response, the connector creates the PI AF element hierarchy and PI points in Data Archive. Each device is represented as an element in PI AF. Therefore, an AF data reference attribute is created for each measurement under the corresponding device.

The PI points' time stamps are based on the "timestamp" field in the metric. If there is no "timestamp" field, then it uses the "timestamp" field of the message. If there is no "timestamp" field on the message, the current time of the machine that the connector is running on is used. Aliasing of metric names is supported.

Note: For more information on the point source of the PI points created, see Learn about data source configuration settings for MQTT Sparkplug.