gcp-iot-provision

0.2.7 • Public • Published

Google Cloud Function for IoT Device Provisioning

This Cloud Function allows you to provision and synchronize a balena device with Google Cloud IoT Core in a secure and automated way via an HTTP endpoint. The Cloud Function may be called by a balena device, as seen in the cloud-relay example.

Method Action
POST Provisions a balena device with IoT Core. First the function verifies the device UUID with balenaCloud. Then it creates a public/private key pair and adds the device to the registry. Finally the function sets balena device environment variables for these entities.
DELETE Removes a balena device from the IoT Core registry and removes the balena device environment variable for the private key. Essentially reverses the actions from provisioning with HTTP POST.

Device Environment Variables

Once the Cloud function has provisioned the device with GCP, it sets balena device environment variables as described below, which allow the device to connect to IoT Core. Some aspects of the variable values are common across devices, and are collected from the cloud function deployment.

Variable Value
GCP_CLIENT_PATH <registry-path>/devices/<device-id>

<registry-path> is derived from the project ID, GCP region, and registry ID
<device-id> is derived from the balena UUID for the device
GCP_DATA_TOPIC_ROOT /devices/<device-id>
GCP_PROJECT_ID Google Cloud project ID
GCP_PRIVATE_KEY Private key in PEM format, base64 encoded to eliminate line wrapping

Setup and Testing

GCP setup

The Cloud Function interacts with Google Cloud IoT Core via client code operating with service account credentials. You must setup a Google Cloud project with an IoT Core registry. The service account must have the Cloud IoT Provisioner role to manage device records in the IoT Core registry. See the IoT Core documentation for more background.

Workspace setup

Clone this repo

$ git clone https://github.com/balena-io-examples/gcp-iot-provision

The sections below show how to test the Cloud Function on a local test server and deploy to Cloud Functions. In either case you must provide the environment variables in the table below as instructed for the test/deployment.

Key Value
BALENA_API_KEY for use of balena API; found in balenaCloud dashboard at: account -> Preferences -> Access tokens
GCP_PROJECT_ID Google Cloud project ID, like my-project-000000
GCP_REGION Google Cloud region for registry, like us-central1
GCP_REGISTRY_ID Google Cloud registry ID you provided to create the registry
GCP_SERVICE_ACCOUNT base64 encoding of the JSON formatted GCP service account credentials provided by Google when you created the service account. Example below, assuming the credentials JSON is contained in a file.

cat <credentials.txt> | base64 -w 0

HTTP API

The HTTP endpoint expects a request containing a JSON body with the attributes below. Use POST to add a device to the cloud registry, DELETE to remove.

Attribute Value
uuid UUID of device
balena_service (optional) Name of service container on balena device that uses provisioned key and certificate, for example cloud-relay. If defined, creates service level variables; otherwise creates device level variables. Service level variables are more secure.

Test locally

The Google Functions Framework is a convenient tool for local testing. First, start a local HTTP server (docs reference) using a script like below.

export BALENA_API_KEY=<...>
... <other environment variables from table above>
export GCP_SERVICE_ACCOUNT=<...>

npx @google-cloud/functions-framework --target=provision

Next, use curl to send an HTTP request to the local server to provision a device. See the HTTP API section above for body contents.

curl -X POST http://localhost:8080 -H "Content-Type:application/json" \
   -d '{ "uuid": "<device-uuid>", "balena_service": "<service-name>" }'

After a successful POST, you should see the device appear in your IoT Core registry and GCP_CLIENT_PATH, GCP_DATA_TOPIC_ROOT, GCP_PRIVATE_KEY, and GCP_PROJECT_ID variables appear in balenaCloud for the device. After a successful DELETE, those variables disappear.

Deploy

To deploy to Cloud Functions, use the command below. See the command documentation for the format of yaml-file, which contains the variables from the table in the Development setup section above.

gcloud functions deploy provision --runtime=nodejs14 --trigger-http \
   --env-vars-file=<yaml-file> --allow-unauthenticated \
   --service-account=<name>@<xxxx>.iam.gserviceaccount.com

The result is a Cloud Function like below. Notice the TRIGGER tab, which provides the URL for the function.

Alt text

Test the Cloud Function

To test the function, use a command like below, where the URL is from the TRIGGER tab in the console. See the HTTP API section above for body contents.

curl -X POST https://<region>-<projectID>.cloudfunctions.net/provision \
   -H "Content-Type:application/json" \
   -d '{ "uuid": "<device-uuid>", "balena_service": "<service-name>" }'

After a successful POST, you should see the device appear in your IoT Core registry and GCP_CLIENT_PATH, GCP_DATA_TOPIC_ROOT, GCP_PRIVATE_KEY, and GCP_PROJECT_ID variables appear in balenaCloud for the device. After a successful DELETE, those variables disappear.

Package Sidebar

Install

npm i gcp-iot-provision

Weekly Downloads

13

Version

0.2.7

License

Apache-2.0

Unpacked Size

137 kB

Total Files

9

Last publish

Collaborators

  • balena.io