Exercise 8.

1 - AWS IoT Greengrass

In Exercise 1.1, your Car Things were connecting directly with AWS IoT. What if
those Car Things were only able to communicate on a closed network? What if you
didn't need to send all of that telemetry data to AWS and would prefer to do some
processing locally in your network and only send pieces of the data to AWS? The
service that would help with this and much more is AWS IoT Greengrass.

You will use a similar scenario as Exercise 1.1, but instead of having both Car Things
send their data to AWS IoT Core, they will send their data to Greengrass Core. The
devices will communicate the same way using the MQTT protocol using Certificates,
but the endpoint for the connection and the Certificate Authority (CA) will be the
endpoint of Greengrass Core running in your Cloud9 environment and its Certificate
Authority.

Although in this case, Greengrass Core will be running in the same server as Car1
and Car2, they would normally be separate. For the devices to know how to connect
to Greengrass Core within the network, you can use the Greengrass Discovery API
which will return all the different connectivity options (IP and Port) of your
Greengrass Core.

To do processing of the data, you will use a Lambda function that you will deploy on
your Greengrass Core. That's right, Lambda will be running on your server (your
Cloud9 environment). In this case, the Lambda function will simply take the data that
is sent by each of the devices (on the lab/telemetry Topic) and re-publish it to a
different IoT Topic: lab/greengrass/telemetry.

You will configure Subscriptions within your Greengrass Core to manage how the
data is flowing. By default, there is no data that can go anywhere, even between the
devices connected on your Greengrass Core. To allow that, Subscriptions are made
to say what the data flow should be. In this case, you will configure that all the
telemetry data sent by the Cars should be sent to your Lambda function running on
Greengrass Core and that the data re-published by that Lambda function should be
sent to the AWS IoT Cloud.

Finally, you will be able to subscribe to the IoT Topic that your Lambda function is
publishing to (lab/greengrass/telemetry) using the AWS IoT MQTT Client in the AWS
Management Console to validate the flow of the data.

The diagram below shows the resources and data flow that you will create in this
exercise.
 Figure 1 This exercise assumes that the resources from Exercise 1.1 haven't been deleted. If you have deleted those
components, you will need to start with Exercise 1.1 again before continuing.

1. Create Greengrass Group

In this section, you will create a Greengrass Group. When doing so, you will use a
wizard that will create all of the different resources required to make that
Greengrass Group work: Core, Certificate and Key Pairs and Policy.

1.1 Create Greengrass Group, Core, Certificate and Policy

1. In the AWS Management Console, click Services, and then click IoT Greengrass to go to
the IoT Greengrass console. It is in the same console as IoT Core.

2. Make sure you are in the same Region as the one you used in Exercise 1.1. It should
be Frankfurt, Ireland, N. Virginia, Ohio, Oregon or Tokyo. You can validate that by going
to the Cloud9 service and looking for the IoTOnAWS environment. If you don't see it, you
aren't in the right region.

3. Click Create a Group.

4. If this is your first time using Greengrass, you will received the message Greengrass needs
your permission to access other services. AWS IoT Greengrass works with other AWS
services, such as AWS IoT and AWS Lambda. Greengrass needs your permission to access
these services and read and write data on your behalf. Click Grant permission.
5. Click Use default creation to automatically create a Greengrass Group, an IAM Role for
Greengrass to access Lambda and other services, a Greengrass Core and a Certificate and
key pair for the Core. Any of the two buttons named Use default creation works.

6. In the Group Name field, enter labIoTGGGroup.

7. Click Next.

8. Leave the Name field as is labIoTGGGroup_Core.

9. Click Next.

10. In the next screen, you will see the list of actions that this default creation wizard will create.
It first creates the Group. The Greengrass Group requires a Core. This Core is similar to an
IoT Thing. In fact, it will be listed in the Things list. Similar to a Thing, the Core requires a
Certificate, a Private Key and a Public Key. For the Authorization piece, the Certificate will
require a Policy. All of those are created for you.

11. Click Create Group and Core.

12. Click Download these resources as a tar.gz and save the file to your computer.

13. Click Finish.

14. There is one feature that we don't need for this Greengrass group which was enabled by
default: stream manager. This is a feature that makes it easier and more reliable to
transfer high-volume IoT data to the AWS Cloud. It was added automatically by the default
creation and adds an extra requirement to have Java installed. To make things easier, you
will remove this feature.

15. Click on the Greengrass group labIoTGGGroup.

16. Click Settings.

17. Next to Stream manager, click Edit.

18. Click Disable.

19. Click Save.

20. Click Cores within the Greengrass group. You can see that the labIoTGGGroup_Core
Core is associated to this Group.

21. Expand Manage and click Things. You can see that a new Thing has been created
called labIoTGGGroup_Core.

22. Click labIoTGGGroup_Core.

23. Click Security. You can see that a Certificate is associated to this Thing.

24. Click on the certificate.

25. Click Policies. You can see that an IoT Policy is associated to this Certificate.

26. Click on the labIoTGGGroup_Core-... policy. You can see the IoT Policy that was
automatically created.

As you can see, a Greengrass Group Core is like a normal IoT Thing similar to car1.
It has a certificate and a policy. You now have an Greengrass Group and a Core
ready to be connected to.

2. Setup Greengrass Core on Cloud9 and start it

In this section, you will configure a few settings on the Cloud9 environment to be
ready to install Greengrass. Those settings are a user and a group under which
Lambda will run, hardlink and softlink protection and mounting Linux control groups.

You will then install Greengrass by downloading it from source and extracting it in
the root (/). You will also extract all of the files from the archive that you downloaded
in the previous section.

Finally, you will start Greengrass.

2.1 Start Cloud9

Your Cloud9 environment may have shut down at this point as it's supposed to
automatically shutdown after 30 minutes. To restart it, follow these steps:

1. In the AWS Management Console, click Services, and then click Cloud9 to go to the
Cloud9 console.

2. You should see a list of environments. If you don't, click on the hamburger menu icon (the
three parallel lines) near the top left of the screen and click on Your environments.

3. Click the Open IDE button in the IoTOnAWS card.

4. It may take a minute for your environment to start.

1. Add the user ggc_user and ggc_group to the Cloud9 instance under which the
Lambda functions running in Greengrass will run under. In the
Cloud9 terminal enter the following commands:

sudo adduser --system ggc_user

sudo groupadd --system ggc_group

2. Greengrass requires that the security on the instance be improved by enabling

hardlink and softlink protection on the operating system. In the
Cloud9 terminal enter the following commands:

echo 'fs.protected_hardlinks = 1' | sudo tee -a /etc/sysctl.d/00-defaults.conf

echo 'fs.protected_symlinks = 1' | sudo tee -a /etc/sysctl.d/00-defaults.conf

sudo sysctl --system

3. Extract and run the following script to mount Linux control groups (cgroups). This is
an AWS IoT Greengrass dependency. In the Cloud9 terminal enter the following
commands:

cd /tmp

curl https://raw.githubusercontent.com/tianon/cgroupfs-mount/951c38ee8d802
330454bdede20d85ec1c0f8d312/cgroupfs-mount > cgroupfs-mount.sh

chmod +x cgroupfs-mount.sh

sudo bash ./cgroupfs-mount.sh

2.3 Install Greengrass

Download Greengrass and extract it. In the Cloud9 terminal enter the following
commands:

cd /tmp
 wget https://d1onfpft10uf5o.cloudfront.net/greengrass-core/downloads/1.10
.2/greengrass-linux-x86-64-1.10.2.tar.gz

sudo tar -xzf greengrass-linux-x86-64-1.10.2.tar.gz -C /

2.4 Upload and extract your Greengrass Group configuration

Upload the ...-setup.tar.gz file that you downloaded in section 1.

1. Click on the folder IoTOnAWS in the directory tree on the left in Cloud9.

2. Click File > Upload Local Files....

3. Click Select files.

4. Select the ...-setup.tar.gz file in your file system that you downloaded in section 1.1.

5. Click on the x icon to close the Upload Files window.

6. Extract the ...-setup.tar.gz archive. This will place the Certificate and Private Key of
your Core in the /greengrass/certs folder. It will also place the config.json file
containing all the information to connect to the AWS IoT Thing Core you created
earlier in the /greengrass/config folder. In the Cloud9 terminal enter the following
commands:

cd /tmp

mv ~/environment/*-setup.tar.gz setup.tar.gz

sudo tar -xzf setup.tar.gz -C /greengrass

7. Place the AWS IoT Root Certificate Authority in the /greengrass/certs folder. In the
Cloud9 terminal enter the following commands:

cd /greengrass/certs/

sudo wget -O root.ca.pem

https://www.amazontrust.com/repository/AmazonRootCA1.pem
 2.5 Start Greengrass
1. Start Greengrass by executing the following commands in the Cloud9 terminal:

cd /greengrass/ggc/core/

sudo ./greengrassd start

You should see the following output:

Setting up greengrass daemon

Validating hardlink/softlink protection

Waiting for up to 40s for Daemon to start

Greengrass successfully started with PID: .....

You have successfully setup Greengrass on your Cloud9 instance and connected it
to AWS IoT Greengrass.

3. Create the Greengrass Lambda function

In this section, you will create a Lambda function. When creating a Lambda function,
the Lambda service requires a Role to be passed. However, as this function will run
on the Greengrass Core, this Role will never be used so you will use the default Role
creation from Lambda. Greengrass can only use an alias or a publish version of your
Lambda function (immutable versions). As the last step, you will publish your
Lambda function.

This function has been developed in Python 2.7 especially because that's the version
installed on Cloud9. It could have been developed in any of the supported language
of Greengrass. At the time of writing this class, these were: Python 3.7, Java 8,
Node.js 6.10, Node.js 8.10, Node.js 12.x and C/C++.

1. In the AWS Management Console, click Services, and then click Lambda to go to the
Lambda console.

2. Click Create a function.

3. For Function name, enter labIoTGreengrassLambda.

4. For Runtime, select Python 2.7.

5. Expand Choose or create an execution role under Permissions.

6. Select the radio button next to Create a new role from AWS policy templates.

7. For Role name, enter labIoTGGLambdaRole. Leave the Policy templates empty as
you are creating a Role with no specific permissions other than logging.

8. Click Create function.

9. Download the code of the Lambda function to your computer by clicking here.

https://aws-tc-largeobjects.s3.amazonaws.com/OTP-AWS_D5-2019/v1.0/code/exercise-
3.2-greengrassLambda-python.zip

10. In the Function code section, click Actions and select Upload a .zip file.

11. Click Upload, select the exercise-3.2-greengrassLambda-python.zip you just

downloaded and click Save.

12. The code has been be uploaded and Lambda will show the code in the browser. Feel free
to review the code. In summary, it ensures that the data is coming from car1 or car2. It
then publishes using the Greengrass IoT client the same data that was received on
the lab/greengrass/telemetry IoT Topic.

13. At the top of the page, click Actions > Publish new version.

14. Click Publish.

15. At the top of the page, click Actions > Create alias.

16. Under name, enter prod.

17. Click Save.

You now have a Lambda function that's published to version 1 and an alias
called prod that points to it. Greengrass can only point to published version of
Lambda as well as aliases. However, the best practice is to use the alias as it's much
easier to change the version the alias in Lambda points to than having to redo all
subscriptions after they are created when they point to a specific Lambda version.
 4. Add Cars, Lambda and Subscriptions to Greengrass and
Deploy
In this section, you will add the 2 Car Things from Exercise 1.1 to your Greengrass
Group. You will also associate the Lambda function you created to your
Greengrass Group so it can be deployed to the Greengrass Core in Cloud9.

You will then create 3 Subscriptions to allow the data sent from the cars to be sent
to Lambda. As well as for data sent by Lambda to be sent to AWS IoT Cloud.

 car1 data sent on the lab/telemetry IoT Topic is sent to the Lambda function

 car2 data sent on the lab/telemetry IoT Topic is sent to the Lambda function

 Lambda data sent on the lab/greengrass/telemetry IoT Topic is sent to AWS IoT Cloud

Finally, you will deploy the Greengrass configuration to your Greengrass Core
running in Cloud9.

4.1 Cars

1. In the AWS Management Console, click Services, and then click IoT Greengrass to go to
the IoT Greengrass console.

2. Click Groups in the left menu.

3. Click on the labIoTGGGroup link.

4. Click Devices.

5. Click Add Device.

6. Click Select an IoT Thing.

7. Click the radio button next to car1.

8. Click Finish. The car1 Thing has now been added.

9. Click Add Device.

10. Click Select an IoT Thing.

11. Click the radio button next to car2.

12. Click Finish. The car2 Thing has now been added.
 4.2 Lambda

1. Click Lambdas.

2. Click Add Lambda.

3. Click Use existing Lambda.

4. Click the radio button next to labIoTGreengrassLambda.

5. Click Next.

6. Click the radio button next to Alias: prod.

7. Click Finish.

4.3 Subscriptions
In this section, you will configure the subscriptions for the data flow between car1,
car2, Lambda and AWS IoT Cloud.

4.3.1 car1 to Lambda

You will add the Subscription from car1 to Lambda for the lab/telemetry Topic.

1. Click Subscriptions.

2. Click Add Subscription.

3. Under Select a source, click Select.

4. Click Devices.

5. Click car1.

6. Under Select a target, click Select.

7. Click Lambdas.

8. Click labIoTGreengrassLambda.

9. Click Next.

10. For Topic filter, enter lab/telemetry.

11. Click Next.

12. Click Finish.

 4.3.2 car2 to Lambda
You will add the Subscription from car2 to Lambda for the lab/telemetry Topic.

1. Click Add Subscription.

2. Under Select a source, click Select.

3. Click Devices.

4. Click car2.

5. Under Select a target, click Select.

6. Click Lambdas.

7. Click labIoTGreengrassLambda.

8. Click Next.

9. For Topic filter, enter lab/telemetry.

10. Click Next.

11. Click Finish.

4.3.3 Lambda to AWS IoT Cloud

You will add the Subscription from Lambda to AWS IoT for
the lab/greengrass/telemetry Topic.

1. Click Add Subscription.

2. Under Select a source, click Select.

3. Click Lambdas.

4. Click labIoTGreengrassLambda.

5. Under Select a target, click Select.

6. Click IoT Cloud.

7. Click Next.

8. For Topic filter, enter lab/greengrass/telemetry.

9. Click Next.

10. Click Finish.

 4.4 Deploy the configuration to Greengrass Core

1. Click Actions > Deploy.

2. Click Automatic detection.

3. Click Deployments.

4. Under the name of the group labIoTGGGroup, you will see the status of the deployment.
After a few seconds, you should see a green light with the message Successfully
completed.

Your Greengrass Core on Cloud9 has now been updated with all the configuration
settings you have set. It's time to use it.

5. Download car code, Start the cars and Subscribe to

lab/greengrass/telemetry topic
In this section, you will download the code that connects to Greengrass. The code
isn't that different from Exercise 1.1 other than it needs to query the Greengrass
Discovery API to know how to connect to Greengrass and loop through all the
connectivity options returned and try to connect to each of them until one works.
Then, it's the same as Exercise 1.1 where the code uses MQTT to send telemetry
data on the lab/telemetry IoT Topic.

You will then start both cars which will use the same Certificate and Key used in
Exercise 1.1 to communicate to Greengrass. The Certificate Authority certificate for
connecting to Greengrass Core will be downloaded by using the Greengrass
Discovery API.

Finally, you will use the AWS IoT MQTT Client in the AWS Management Console
to subscribe to the lab/greengrass/telemetry to validate that data is flowing from the
cars, to Lambda and into AWS IoT Cloud.

5.1 Download the car code and install code dependencies

1. Download the code by running the following commands in the Cloud9 terminal.

cd ~/environment
 wget https://aws-tc-largeobjects.s3.amazonaws.com/OTP-AWS_D5-2019/v1.0/co
de/exercise-3.2.js

cp exercise-3.2.js car1/

cp exercise-3.2.js car2/

rm exercise-3.2.js

2. The code uses the request library that needs to be installed. Run the following
command in the Cloud9 terminal to install it.

npm install request

You can ignore any warnings produced.

5.2 Start the cars

1. If you don't have 2 Cloud9 terminal. click the circled + icon that is next to your current
terminal and select New Terminal. You now have 2 different terminals.

2. In the left terminal, execute the following commands to start the code for car1.

cd ~/environment/car1

node exercise-3.2.js

You should see the following:

Data received from Greengrass Discovery API:

...
 }

Trying to connect to Greengrass Core with the address 127.0.0.1 and por
t 8883

Connected to Greengrass Core

Sending car telemetry data to AWS IoT for car1

Sending car telemetry data to AWS IoT for car1

...

Look at the JSON payload returned by the Greengrass Discovery API. You will see all the
different Connectivity options (HostAddress and PortNumber) that the Greengrass Core
running in Cloud9 can be reached at. You can also see the Certificate Authority certificate
for your Greengrass Core.

3. In the right terminal, execute the following commands to start the code for car2.

cd ~/environment/car2

node exercise-3.2.js

You should see the following:

Data received from Greengrass Discovery API:

...

Trying to connect to Greengrass Core with the address 127.0.0.1 and por
t 8883

Connected to Greengrass Core

 Sending car telemetry data to AWS IoT for car2

Sending car telemetry data to AWS IoT for car2

...

5.3 Subscribe to lab/greengrass/telemetry topic

1. In the AWS Management Console, click Services, and then click IoT Core to go to the IoT
Core console.

2. Click Test in the left menu to start the AWS IoT MQTT Client.

3. In the Subscribe topic field, enter lab/greengrass/telemetry.

4. Click Subscribe to topic.

You should see the telemetry data sent by each car every 5 seconds. Car1 and
Car2 are connecting to Greengrass which runs a Lambda function that parses the
data and sends it back to AWS IoT Cloud as you can see it here.

top Greengrass
1. In the Cloud9 terminal, enter the following commands:

cd /greengrass/ggc/core/

sudo ./greengrassd stop