Gateway Management Module (GMM)

Create a dataflow

An EFM Dataflow enables you to define policies that control which data goes to which apps at which time.

The Dataflow Editor is a visual programming environment used to create dataflow models. A dataflow model is composed of blocks that do assigned tasks when requirements are met (a dataflow model is not a set of commands that are executed in sequence).

The following steps show a sample scenario of how to subscribe to a "System" DSlink metric CPU Usage, and then aggregate that into a moving 30 second average. See the Dataflow Editor User Guide for detailed instructions.

  1. Create a dataflow:

    1. Open the dataflow webpage at https://BrokerIP:8443/dataflow.html.

    2. Navigate to Data > downstream > dataflow.

    3. Right-click dataflow.

    4. Select Create Dataflow to create a new dataflow called cpuMovingAverage.

    5. Click Invoke.

    6. Click the icon right after the dataflow name to open this data flow.

      EFM

  2. Subscribe to the CPU Usage metric:

    1. Find the System link node in the data tree and select it.

    2. Drag the the CPU Usage metric to the canvas.

      EFM

  3. Create a moving 30 second window of your CPU utilization:

    1. Search for "realtime recorder" in the Blocks section. Then drag and drop it onto the dataflow canvas.

    2. Click the + on the recorder block to add the “name 0” and “value 0” properties (to remove "nameX" and "valueX", click the - ).

    3. Bind the “value” property from your subscription to the “value 0” property of the recorder block.

    4. Enter cpu in the “name 0” property of the recorder block.

    5. Set the “bufferSize” property to 0.

    6. Set the “timeWindow” property to 30. You now have a moving 30 second window of your CPU utilization.

      EFM

  4. Drag the Aggregation block onto the canvas.

  5. Bind the “output” property from the Realtime Recorder to the “input” property on the Aggregation block.

  6. Enter cpu in the “column” property for the Aggregation block.

  7. Select Average in the “method” property of the Aggregation block.

    EFM

  8. In the data tree, right-click the data node.

  9. Left-click the Publish action and drag it to your canvas.

  10. Bind the “output” property from the Aggregation block to the “Value” property on the “Publish” block.

  11. In the “Path” property of the “Publish” block, enter /data/cpuMovingAverage.

  12. Set the “autorun” property of the “Publish” block to true.

    EFM

  13. Now the CPU Average value is being published. This enables any EFM node/broker that connects to this broker to have access to the transformed data point under "/data" path.

Related documentation
See the Dataflow Editor User Guide for detailed instructions.

Next