Exercise - Configure Azure IoT Edge on NVIDIA embedded hardware

  • 10 minutes

The Azure IoT Edge runtime is what turns a device into an IoT Edge device. The runtime can be deployed on devices as small as a Raspberry Pi or as large as an industrial server. When a device is configured with the IoT Edge runtime, you can deploy containerized workloads based on deployment configurations that are defined your IoT hub.

Complete the following steps on an NVIDIA Jetson embedded device provisioned with JetPack 4.6. You will also need to install the DeepStream 6.0 package for Jetson to the device. We assume that you have completed the steps described in Set up and configure an NVIDIA DeepStream development environment and Introduction to NVIDIA DeepStream Graph Composer with Azure.

  1. To manually provision your NVIDIA Jetson device, you must provide it with a device connection string that you create by registering a new IoT Edge device in your IoT hub.

    Create a new device connection string as described in the Option 1: Register with symmetric keys section in Register an IoT Edge device in IoT Hub.

    You may skip the instructions in the Option 2: Register with X.509 certificates section. Complete the steps to register your device, and be sure to use a descriptive name. For example, if you're using a Jetson Xavier device, use a name like jetson-xavier-01. When you registered your device, familiarize yourself with the process described in View registered devices and retrieve connection strings.

  2. Now that you have obtained a connection string, complete the steps in Install or uninstall Azure IoT Edge for Linux and install the Azure IoT Edge runtime on the NVIDIA Jetson embedded device:

    Complete the steps described in Option 1: Authenticate with symmetric keys. (You can skip the instructions in Option 2: Authenticate with X.509 certificates and Uninstall IoT Edge.)

  3. Verify that all steps have been completed by running the following commands in a terminal on the NVIDIA Jetson embedded device:

    sudo iotedge list

    When the IoT Edge runtime is successfully started for the first time, this command's output should indicate that only the edgeAgent module is running.

  4. Because your containerized DeepStream Graph Composer workload uses an output sink of type EGL (type= 2), before you run the workload from a container, you must grant access to the X11 display environment. Run the following commands from a terminal on the NVIDIA embedded device for access to the X11 services:

    sudo xhost +

When you've verified that IoT Edge has been appropriately installed and configured and you've given container access to X11 services on the device, you can proceed with steps to deploy the deepstream_test4_jetson:v1 module that you published to your container registry.