Agent / (Edge Connector)

Learn how to create, deploy and use agents (formerly Edge Connector).

The agent facilitates the integration of both pre-existing and real-time data from private networks by positioning connectors near relevant data sources or sinks. Upon deployment, it connects to a Heisenware workspace to either supply or receive data.

Technical overview

The agent is an executable that you can create directly in your Heisenware workspace whenever required. During its creation, you will specify the integration capabilities for your agents. It downloads instantly and can be run on most host systems to bridge networks.

The agent establishes an outbound-only connection via port 8883 using MQTTs (encrypted MQTT). You can deploy multiple individual agents as needed. The very same agent can also be used on different host machines.

Setup

Agent setup

System requirements

The agent can be run on the following systems:

  • 64-bit hardware with

    • MacOS

    • Windows

    • Linux

      • There is a separate Alpine version

  • ARM64 on Industrial PC

Agent creation

To open the creation dialogue, click on the cloud symbol with an arrow pointing down within the functions panel.

Creating an agemt
  1. Select the connectors and integrations you would like to include in your agent. A single agent can be used to integrate multiple interfaces and protocols.

  2. Select the target OS the agent shall be deployed on.

  3. Optionally, define a prefix for the agent. Use a prefix if you want to use one specific agent on more than one host machine.

  4. Click submit.

  5. Wait for the agent to be created. Its download starts automatically. Depending on your target OS, it has a different file format.

Using a prefix

Using a prefix allows you to create just one agent and deploy it on several computers. This is extremely useful if you want to control an entire fleet of similar endpoints (e.g., machines).

If you built your agent without specifying a prefix, the corresponding executable will have a built-in unique ID. This has two effects:

  1. You can only run one instance of this specific agent at a time (otherwise the IDs collide).

  2. You can move the agent across computers or within file systems. It will always be recognized as the same one.

If you built your agent with specifying a prefix, the corresponding executable will generate an ID at startup time. This has the following effects:

  1. You can deploy the very same agent at multiple locations and run them simultaneously.

  2. Each agent establishes an individual connection and appears as an own entity in the App Builder.

  3. An agent is only recognized as the exact same one after a restart, as long as it is stored in the same directory (by reading a hidden .hw-agent-id file in this directory).

Deployment and activation

The agent must now be deployed and activated within the network you want to integrate data from.

  1. Save the agent on a device. For example, this can be on an industrial PC in the OT network, a machine controller or a regular laptop. It is crucial though that this device has access to the interface to be integrated, like an OPC UA server in case of OPC UA integration.

  2. Depending on the target OS, simply double-click the file or open a terminal and run the file. On Linux and MacOS it might be necessary to make the file executable.

  3. The agent and its functionality should automatically appear in the functions panel.

  4. Expand it and drag any functions you want to use onto the board.

Available connectors

As mentioned above, the agent is a means of transport for moving connectors into the network of the data source or data sink. If a connector is missing, it can be added to the functionality of the agent on request.

Currently available:

Last updated