Register a device in AWS IoT Core and read its device shadow

Steps

1. Find your account's IoT Core data endpoint, the address devices connect to, and keep it in a variable for the later steps:

EP=$(aws iot describe-endpoint --endpoint-type iot:Data-ATS --query endpointAddress --output text)
echo $EP

It ends in -ats.iot.us-east-1.amazonaws.com. If CloudShell times out later, run the first line again. 2. Register the sensor in the IoT Core registry, the list of devices IoT Core knows:

aws iot create-thing --thing-name clf-lab22-sensor
aws iot describe-thing --thing-name clf-lab22-sensor --query thingName --output text

The second command prints clf-lab22-sensor. A real sensor would also get a certificate to prove its identity; that is the authentication job, and this lab skips it because nothing physical connects. 3. In the AWS IoT Core console, open Test, then MQTT test client. Under Subscribe to a topic, enter clf/lab22/# as the topic filter and choose Subscribe. The # matches every topic under clf/lab22/. Leave the tab open. 4. Publish a reading to the broker, as the sensor would:

aws iot-data publish --endpoint-url https://$EP --topic clf/lab22/temperature --cli-binary-format raw-in-base64-out --payload '{"temperature": 21}'

Within a second or two the message appears in the MQTT test client under clf/lab22/temperature. That is the brokering job: the publisher and the subscriber never talk to each other directly. A real device would use MQTT; the CLI publishes over HTTPS, which IoT Core also accepts. 5. Record the state the device reports in its device shadow:

aws iot-data update-thing-shadow --endpoint-url https://$EP --thing-name clf-lab22-sensor --cli-binary-format raw-in-base64-out --payload '{"state": {"reported": {"temperature": 21}}}' clf-lab22-out.json

The last argument is the file the response is written to. 6. Now act as the application, while the sensor is imagined to be offline: ask for a different temperature by setting the desired state, then read the whole shadow back:

aws iot-data update-thing-shadow --endpoint-url https://$EP --thing-name clf-lab22-sensor --cli-binary-format raw-in-base64-out --payload '{"state": {"desired": {"temperature": 19}}}' clf-lab22-out.json
aws iot-data get-thing-shadow --endpoint-url https://$EP --thing-name clf-lab22-sensor clf-lab22-shadow.json
jq '.state' clf-lab22-shadow.json

The shadow holds reported 21, desired 19, and a delta of 19: the change still to be made. When the sensor reconnects, it reads the delta and acts on it. That is how an application can set a device's state while the device is offline. 7. In the console, open All devices, then Things, then clf-lab22-sensor, and choose the Device Shadows tab. Open Classic Shadow and see the same reported, desired and delta values. The third job, routing messages to other services with the rules engine (for example, storing readings in Amazon S3), is not built here; it is the step a real project would add next.