AWS Healthcare Greengrass Workshop

Introduction

Greengrass is a software that allows you to extend AWS cloud capabilities to local devices, making it possible to collect data and securely communicate with local network and AWS Cloud. Fits very well for IoT solutions as gateway that could handle communication with Bluetooth, Zigbee devices, for example.

Greengrass Group, Core and Devices

Greengrass Group: a set of Cores and Devices
Greengrass Core: a set of lambdas and devices, runs greengrass daemon
Greengrass Device: small devices, no greengrass daemon

Table of Content

Configure your Raspberry Pi
Create Lambda Function
Create Greengrass Group
Deploy Lambda to your Raspberry Pi
Monitoring and Processing AWS IoT Messages
Greengrass Controlling Bluetooth Blood Pressure Sensor

1.Configure Permissions and Raspberry Pi

Before we start using Greengrass we need to configure our Raspberry PI AWS credentials for awscli and also be sure you have access to your Pi using SSH.

Step #1: Open IAM Console to create an access key

Step #2: Click in your AWS user

Step #3: Click "Security Credentials"

Step #4: Click "Create Access Key"

Step #5: Copy the Access Key ID and Secret Access

You can also download a CSV file with your keys!

Step #6: Access your Raspberry Pi using SSH and configure awscli

Open ssh: ssh pi@RASP-IP-ADDRESSS
Type: aws configure
Enter your keys
Choose us-east-1 as region
Choose json as output

Step #7: Return to IAM Console and select your user

Step #8: Click "Add permissions"

Step #9: Click "Attach existing policies directly"

Step #10: Choose "AdministratorAccess"

Step #11: Choose "AmazonDynamoDBFullAccess"

Step #12: Choose "AWSGreengrassFullAccess"

Step #13: Choose "AWSIoTFullAccess" and click "Next Review"

Step #14: Click "Add permissions" and done!

2.Create a Lambda Function

Now we are going to create a Lambda function that we will deploy to our Raspberry Pi later. This Lambda will be sending simulated blood pressure data.

Step #1: Open Lambda Console and click "Create a function"

Step #2: Select "Author from scratch"

Name: healthcare-data-generator
Runtime: Python 2.7
Existing Role: lambda_basic_execution

Step #3: Scroll down to Function code panel

Download to your machine
Select "Upload a .ZIP file" in "Code entry type" combo

Step #4: Click "Upload" and select the .ZIP file downloaded in the previous step and then click "Save"

Step #5: Select "Edit code line" in "Code entry type" to see the Lambda code

Step #6: Click "Actions -> Publish new version"

Step #7: Name it "v1" and done!

3.Create Greengrass Group

It's time to create a Greengrass Group that represents our Raspberry Pi as a Core Device.

Step #1: Open Greengrass console and click "Get started" to create a group

Step #2: Select "Use easy creation"

Step #3: Name your group as "healthcare"

Step #4: Accept the suggested core name

Step #5: Click "Create group and core"

Step #6: Click "Download these resoures as .tar.gz"

Step #7: Download Greengrass security resources to your machine

Step #8: Click "Finish"

Step #9: Access your Raspberry Pi and type this command to create a fresh Greengrass core

Step #10: Type "cd greengrass"

Step #11: Upload your .tar.gz security resources file and uncompress into Greengrass directory

Step #12: Type these commands to copy the root.ca.pem certificatre

Step #13: Now you can start your Greengrass core!

4.Deploy Lambda to your Raspberry Pi

Let's deploy our Lambda function to run locally in our Raspberry Pi.

Step #1: Open Greengrass console, select "healthcare" group and click "Lambdas"

Step #2: Select "Use existing Lambda"

Step #3: Select "helthcare-data-generator" Lambda

Step #4: Select the published version

Step #5: Click "Edit configuration"

Step #6: Change Lambda Life-cycle to Long-lived function

Step #7: Your configuration should be like this

Step #8: Back to healthcare group and click "Subscriptions"

Step #12: Select "Lambdas -> healthcare-data-genertor" as Source

Step #13: Selecr "IoT Cloud" as Target

Step #14: With this subscrition our Lambda will be able to reach any topic in the Cloud

Step #15: Click finish

Step #16: Now let's deploy by clicking "Actions -> Deploy" menu

Step #17: Click "Automatic detection"

Step #18: Grant the required permissions

Step #19: In Group Role click "Add role" link

Step #20: Choose "Greengrass_ServiceRole" and click "Save"

Step #21: Now let's do the final deploy, click "Actions -> Deploy" again

Step #22: You should see "Successfully completed message"

5.Monitoring and Processing AWS IoT Messages

Now it's time to monitor AWS IoT to check for incoming MQTT messages from your Raspberry Pi.

Step #1: Open AWS IoT Console

Step #2: Click "Test" and then subscribe to healthcare/data topic

Step #3: You should start seeing health data generated by Greengrass / Raspi

Step #4: Now let's start processing this data

Open Lambda console and create a new Lambda
Name: Healthcare-Data-JSONToDynamoDB

Step #5: Use the following code

Step #6: Open IoT Console and click "Act"

Step #7: Click "Create a rule"

This rule will trigger our lambda function for each message received on healthcare/data topic
Name: HealthData

Step #8: As topic filter, type healthcare/data

Step #9: Add an action and select Lambda

Step #10: Select your lambda function and done!

After you finish this, your healthcare data will be stored on DynamoDB, let's check it!

Step #11: Open DynamoDB Console

Step #12: Click "Tables", choose the "patient" table and click "Items"

If you can see the data coming to the data, congrats, evertything is working!

6.Greengrass Controlling Bluetooth Blood Pressure Sensor

Now that we simulated the blood pressure and heart beat data, we will progress to a real example using a Bluetooth Low-energy Blood Pressure monitor. For this case we are using Libelium / MySignals that provides an open platform for e-helath hardware and software development.