5.3. AWS Greengrass Cloud Connectivity

The AM62x can be used as an Edge Gateway or Edge Router to the cloud, whether private or public. AWS is one of the most popular cloud options available. AWS Greengrass is an edge computing platform that can easily be added to the Processor SDK for the AM62x. This addition allows a AM62x based device to be added to AWS as an edge device.

../_images/greengrass-AM62xEdgeGatewayConcept.png

The above diagram shows a concept of an AM62x based device connecting a variety of wireless and wired sensors and control elements to AWS via Greengrass running on Linux provided by the Processor SDK. This short how-to covers how to add AWS Greengrass to the Processor SDK to start building a system like the one above. This guide focuses on getting the AM62x Starter Kit connected to AWS by adding the necessary components to Linux, running Linux on the AM62x Starter Kit, and connecting to AWS to run the provided getting started tutorials.

5.3.1. Necessary Hardware

Here is a list of things needed to run this demo:

  • AM62 Starter Kit (SK-AM62) EVM with the necessary connections outlined in the Quick Start Guide to boot off of a SD Card. This includes a power supply and cable and a USB to serial connection for the Linux console.

  • SD card that can be overwritten and a way to write to it from a PC.

  • A host machine running Linux (preferably an Ubuntu LTS distribution, 18.04 or newer).

  • An ethernet connection for the EVM to connect to AWS.

5.3.2. System Setup

To set up the AM62x-SK, follow the Quick Start Guide to boot the Processor SDK from a SD Card on the AM62 Starter Kit (SK-AM62) EVM.

Make sure the EVM can connect to the internet using an Ethernet or Wireless connection. This connection is needed to talk to the AWS servers with Greengrass in the following steps. Please refer to Setup Network for more information on setting up these network connections on the EVM.

5.3.3. Use Yocto to Build the Corretto JVM

The Processor SDK is built from Yocto/Open Embedded and can easily be extended or modified. In order to run AWS Greengrass, the SDK needs a Java Virtual Machine (JVM). Since AWS recommends Corretto, Yocto can be used to add Corretto to the SDK in order to run Greengrass.

The instructions for building the SDK with Yocto are documented here. Follow the instructions up to Section 1.2.2.2 Build Steps. Before executing the “oe-layertool-setup.sh” instruction, a few changes need to be made to add Corretto to the build.

  • Open the configs/processor-sdk/processor-sdk-08.03.00.19-config.txt file and add the following line right after the one starting with “oe-core”.

    meta-aws,https://github.com/aws/meta-aws.git,dunfell,9979cfa676105cb68cfadfdaeabf044d7c919319,layers=

    Save and close the file.


  • Execute the build instruction, bitbake, like below to only build the Corretto package (ipk) instead of the entire SDK. This command saves quite a bit of time since the entire SDK is not needed at this point:

    MACHINE=am62xx-evm bitbake corretto-11-bin

    This step will take a few minutes to complete as it will download and build Corretto and all of its dependencies.

    The built packages will be in arago-tmp-external-arm-glibc/deploy/ipk/aarch64.


  • Copy the necessary packages listed below to the target device by either copying them to the SD card using a SD card writer, or using the network with something like a SCP command.

    • corretto-11-bin_11.0.12.7.1-r0.0_aarch64.ipk


  • On the target, install the package with the below command:

    opkg install corretto-11-bin_11.0.12.7.1-r0.0_aarch64.ipk

  • Verify the installation is successful:

    java --version

    If java is installed and working, output with the version should be displayed.

5.3.4. Run AWS Greengrass Getting Started Tutorials

After installing the Corretto JVM, the AM62x EVM and SDK now satisfy all the pre-requisites to run AWS Greengrass. The best way to proceed is to leverage AWS’ excellent Getting Started Tutorial. The JVM is already installed by the above steps, so all that needs to be done is to provision the device in AWS and install Greengrass. The tutorial will help get everything set up and connected. It also provides a number of other options to try in order to learn more.