How to run Lambda functions on AWS Greengrass

AWS Greengrass is a form of edge computing service that extends the cloud’s functionality to your IoT devices by allowing data collection and analysis closer to its point of origin. This is accomplished by executing AWS Lambda functions locally on the IoT device itself, while still using the cloud for management and analytics. Today, we will learn how to leverage AWS Greengrass to run simple lambda functions on an IoT device.

How does this help a business? Well to start with, using AWS Greengrass you are now able to respond to locally generated events in near real time. With Greengrass, you can program your IoT devices to locally process and filter data and only transmit the important chunks back to AWS for analysis. This also has a direct impact on the costs as well as the amount of data transmitted back to the cloud.

Here are the core components of AWS Greengrass:

• Greengrass Core (GGC) software: The Greengrass Core software is a packaged module that consists of a runtime to allow executions of Lambda functions, locally. It also contains an internal message broker and a deployment agent that periodically notifies the AWS Greengrass service about the device’s configuration, state, available updates, and so on. The software also ensures that the connection between the device and the IoT service is secure with the help of keys and certificates.
• Greengrass groups: A Greengrass group is a collection of Greengrass Core settings and definitions that are used to manage one or more Greengrass-backed IoT devices. The groups internally comprise a few other components, namely:
  • Greengrass group definition: A collection of information about your Greengrass group
  • Device definition: A collection of IoT devices that are a part of a Greengrass group
  • Greengrass group settings: Contains connection as well as configuration information along with the necessary IAM Roles required for interacting with other AWS services
  • Greengrass Core: The IoT device itself
• Lambda functions: A list of Lambda functions that can be deployed to the Greengrass Core.
• Subscriptions: A collection of a message source, a message target and an MQTT topic to transmit the messages. The source or targets can be either the IoT service, a Lambda function or even the IoT device itself.
• Greengrass Core SDK: Greengrass also provides an SDK which you can use to write and run Lambda functions on Greengrass Core devices. The SDK currently supports Java 8, Python 2.7, and Node.js 6.10.

With this key information in mind, let’s go ahead and deploy our very own Greengrass Core on an IoT device.

Running Lambda functions on AWS Greengrass

With the Greengrass Core software up and running on your IoT device, we can now go ahead and run a simple Lambda function on it! For this particular section, we will be leveraging an AWS Lambda blueprint that prints a simple Hello World message:

1. To get started, first, we will need to create our Lambda function. From the AWS Management Console, filter out the Lambda service using the Filter option or alternatively, select this URL: https://console.aws.amazon.com/lambda/home.

2. On the AWS Lambda console landing page, select the Create function option to get started.
3. Since we are going to be leveraging an existing function blueprint for this use case, select the Blueprints option provided on the Create function page.
4. Use the filter to find a blueprint with the name greengrass-hello-world. There are two templates present to date that match this name, one function is based on Python while the other is based on Node.js. For this particular section, select the greengrass-hello-world Python function and click on Configure to proceed.
5. Fill out the required details for the new function, such as a Name followed by a valid Role. For this section, go ahead and select the Create new role from template option. Provide a suitable Role name and finally, from the Policy templates drop-down list, select the AWS IoT Button Permissions role.
6. Once completed, click on Create function to complete the function’s creation process. But before you move on to associating this function with your AWS Greengrass, you will also need to create a new version out of this function. Select the Publish new version option from the Actions tab.
7. Provide a suitable Version description text and click on Publish once done. Your function is now ready for AWS Greengrass.
8. Now, head back to the AWS IoT dashboard and select the newly deployed Greengrass group from the Groups option present on the navigation pane.
9. From the Greengrass group page, select the Lambdas option from the navigation pane followed by the Add Lambda option, as shown in the following screenshot:

10. On the Add a Lambda to your Greengrass group, you can choose to either Create a new Lambda function or Use an existing Lambda function as well. Since we have already created our function, select the Use existing function option.
11. In the next page, select your Greengrass Lambda function and click Next to proceed. Finally, select the version of the deployed function and click on Finish once done.
12. To finish things, we will need to create a new subscription between the Lambda function (source) and the AWS IoT service (destination). Select the Subscriptions option from the same Greengrass group page, as shown. Click on Add Subscription to proceed:

13. On the Select your source and target page, select the newly deployed Lambda function as the source, followed by the IoT cloud as the target. Click on Next once done. You can provide an Optional topic filter as well, to filter messages published on the messaging queue. In this case, we have provided a simple hello/world as the filter for this scenario. Click on Finish once done to complete the subscription configuration.

With all the pieces in place, it’s now time to deploy our Lambda function over to the Greengrass Core. To do so, select the Deployments option and from the Actions drop-down list, select the Deploy option, as shown in the following screenshot:

The deployment takes a few seconds to complete. Once done, verify the status of the deployment by viewing the Status column. The Status should show Successfully completed.

With the function now deployed, test the setup by using the MQTT client provided by AWS IoT, as done before. Remember to enter the same hello/world topic name in the subscription topic field and click on Publish to topic once done. If all goes well, you should receive a custom Hello World message from the Greengrass Core as depicted in the following screenshot:

This was just a high-level view of what you can achieve with Greengrass and Lambda. You can leverage Lambda for performing all kinds of preprocessing on data on your IoT device itself, thus saving a tremendous amount of time, as well as costs.

With this, we come to the end of this post. Stay tuned for our next post where we will look at ways to effectively monitor IoT devices.

We leveraged AWS Greengrass and Lambda to develop a cost-effective and speedy solution. You read an excerpt from the book AWS Administration – The Definitive Guide – Second Edition written by Yohan Wadia.  Whether you are a seasoned system admin or a rookie, this book will help you learn all the skills you need to work with the AWS cloud.