Chapter 3. Optimizing IoT Computing Using AWS Greengrass

Some companies probably have security policies that do not allow some internal devices, such as computers and IoT devices, to connect to the internet directly. To solve these issues, they provide proxy or bridge servers to accommodate connectivity between internet devices and internet networks. For IoT cases, we can build a gateway server to serve those requests. In this chapter, we will learn how IoT devices connect to AWS and perform computation locally. We will explore and implement AWS Greengrass to serve all IoT devices to perform AWS computations. We will deploy AWS Greengrass and use Raspberry Pi 3 as AWS Greengrass Core.

The following is a list of topics that we will cover in this chapter:

• Introducing AWS Greengrass
• Exploring supported IoT devices for AWS Greengrass
• Deploying AWS Greengrass on Raspberry Pi 3
• Accessing AWS Greengrass
• Building IoT projects with AWS Greengrass

Introducing AWS Greengrass

AWS Greengrass is one of the AWS services that enables you to build the AWS IoT server locally. This solution has an advantage, because our IoT devices should not connect to AWS IoT directly.

AWS Greengrass acts as a bridge or gateway for connected IoT devices. It offers AWS computation, such as AWS Lambda. We can deploy our own service over AWS Lambda on AWS Greengrass. AWS Greengrass can be implemented on a computer or device-based ARM CPU. All devices that are used for AWS Greengrass should be registered to AWS IoT. IoT devices have access to AWS Greengrass and AWS Greengrass needs to be configured in one group so that both the IoT devices and AWS Greengrass can communicate with each other.

A general overview of AWS Greengrass is depicted in the following figure. Each IoT device can connect to AWS Greengrass to perform computation without connecting to the AWS server. IoT devices and AWS Greengrass can only communicate within the Greengrass Group:

AWS Greengrass...

Exploring supported IoT devices for AWS Greengrass

To run AWS Greengrass on a machine, we should install AWS Greengrass Core.  This software can run on x86_64, ARMv7, and AArch64 (ARMv8) architecture platforms. It also supports Ubuntu 14.04 LTS, Jessie Kernel 4.1/4.4, and other Linux distributions based on Kernel 4.4 or later versions.

The list of supported IoT boards for AWS Greengrass can be found at https://aws.amazon.com/greengrass/faqs/. This book will use Raspberry Pi 3 for targeting AWS Greengrass.

Deploying AWS Greengrass on Raspberry Pi 3

In this section, we will deploy AWS Greengrass on Raspberry Pi 3 Model B. To implement AWS Greengrass deployment, we have the Raspberry Pi 3 board and an active AWS account. There are several steps to deploy AWS Greengrass on Raspberry Pi 3. Perform the following steps:

1. Prepare Raspberry Pi 3
2. Configure Raspberry Pi 3 and other IoT devices to AWS IoT
3. Install AWS Greengrass

These steps will be explained in the following sections.

Preparing Raspberry Pi 3

The Raspberry Pi 3 board can run on several OSes. A list of supported OSes for Raspberry Pi 3 can be found at https://www.raspberrypi.org/downloads/. To work with AWS Greengrass Core, it is recommended that you install the latest Raspbian Linux OS for Raspberry Pi, which can be found at https://www.raspberrypi.org/downloads/raspbian/. For a refresher on installation on Raspberry Pi 3, I recommend you follow the instructions mentioned at https://www.raspberrypi.org/documentation/installation/installing-images...

Accessing AWS Greengrass

To access the AWS Greengrass server from IoT devices or other applications, we apply AWS SDK as usual. We only change our target server to the AWS Greengrass machine.

We need to build AWS Lambda into the AWS Greengrass Core machine. Then, we can access it from our IoT devices or computers. We will explore more about AWS Greengrass and IoT device interaction in the next chapter.  

Building IoT projects with AWS Greengrass

In this section, we will build a simple scenario to access AWS Greengrass. We will build an AWS Lambda application into AWS Greengrass. Then, we will access it from the client application. Our scenario is illustrated in the following figure:

To implement this demo, we will perform the following tasks:

1. Prepare libraries on AWS Greengrass Core
2. Create AWS Lambda
3. Deploy AWS Lambda to AWS Greengrass
4. Test

Each task will be explained in the following sections.

Preparing runtime libraries on AWS Greengrass Core

The first task to ensure our AWS Lambda runs well on AWS Greengrass Core is to prepare all requirement modules/libraries. In general, we can define three modules/libraries that should be installed on the AWS Greengrass Core machine. These modules/libraries are as follows:

• AWS IoT SDK
• AWS Greengrass SDK
• External modules/libraries

AWS IoT SDK and AWS Greengrass SDK should be installed on your AWS Greengrass Core machine. You can select a runtime platform, such...

Summary

We have learned how to deploy AWS Greengrass on the Raspberry Pi 3 board and make connectivity among IoT devices. We have also built an IoT application to connect AWS Greengrass on Raspberry Pi 3.

In the next chapter, we will learn how to build AWS Lambda on AWS Greengrass and interact with IoT devices.