Kaspersky Machine Learning for Anomaly Detection

Configuring the Stream Processor service

December 6, 2023

ID 248002

The Stream Processor service gathers real-time telemetry data (input stream) received from the monitored asset at arbitrary points in time and converts this data to a UTG (output stream). Based on the accumulated data, the Stream Processor service determines the values of tags in the output data stream. After converting data into an output stream, the Stream Processor service forwards this data to the ML model for processing.

When converting incoming telemetry data, the Stream Processor service accounts for potential data losses (for example, if the network of the monitored asset temporarily goes down) and processes observations that were received in Kaspersky MLAD too early or too late. In these cases, the Stream Processor service generates default incidents and/or forwards default tag values to the output data stream.

The Stream Processor service can also compute derivative tags based on incoming telemetry data (for example, to calculate the moving average or average value of a group of tags).

The Stream Processor service configuration file for uploading is provided by Kaspersky specialists or a certified integrator.

System administrators can configure the Stream Processor service.

To configure the Stream Processor service:

  1. In the lower-left corner of the page, click the Main menu button.

    You will be taken to the administrator menu.

  2. Select System parametersStream Processor.
  3. In the Fixed-interval sequence frequency (sec) field, specify the period (in seconds) for which the Stream Processor service will process incoming telemetry data.
  4. Using the Browse button under the Configuration file setting, add a file that contains configuration settings for the Stream Processor service.

    To delete the configuration file for the Stream Processor service, click Clear (). To save the configuration file on your computer, click the Download icon ().

  5. Click the Save button.

Did you find this article helpful?
What can we do better?
Thank you for your feedback! You're helping us improve.
Thank you for your feedback! You're helping us improve.