AWS Lambda Reference Architecture for creating an Internet of Things (IoT) Backend. You can build backends using AWS Lambda and Amazon Kinesis for IoT device data telemetry and analysis. The architecture described in this diagram can be created with a CloudFormation template.
The template does the following:
-
Creates a Kinesis Stream
-
Creates an S3 Bucket named <stackname>-eventarchive-<awsaccountid>
-
Creates a DynamoDB table named <stackname>-SensorData
-
Creates Lambda Function 1 (<stackname>-IoTS3EventProcessor) which receives records from Kinesis and writes objects to the S3 Bucket
-
Creates Lambda Function 2 (<stackname>-IoTDDBEventProcessor) which receives records from Kinesis and writes records to the DynamoDB table and CloudWatch
-
Creates an IAM Role and Policy to allow the event processing Lambda functions to write to CloudWatch Logs, read from the Kinesis Streams, write to the S3 bucket and DynamoDB table
-
Creates Lambda Function 3 (<stackname>-IoTAPI) which accepts synchronous calls from clients and reads historical data from the DynamoDB table
-
Creates an IAM Role and Policy to allow Lambda Function 3 to write to CloudWatch Logs and query the DynamoDB table
-
Creates an IAM user with credentials and policy containing permissions to invoke Lambda Function 3 and put records into the Kinesis stream
In addition, the CloudFormation template uses a Lambda function as a CloudFormation custom resource to create map the Kinesis stream as an input event source to Lambda Function 1 and 2.
The services configured by the CloudFormation template can be tested with the HTML page testpage.html, which acts in the same way as a simple device. It can submit events to the Kinesis stream and make synchronous calls against Lambda Function 3 to retrieve event data from DynamoDB.
In a production scenario it would be the devices or device gateways that make the calls against the AWS services.
Important: As the CloudFormation stack name is used in the name of the S3 bucket, it must only contain lowercase letters. Please use lowercase letters when entering the stack name. The provided CloudFormation template retreives its Lambda code from a bucket in the us-east-1 region. To launch this sample in another region, please modify the template and upload the Lambda code to a bucket in that region.
Step 1 – Create a CloudFormation Stack with the template using a lowercase name for the stack. The template and Lambda functions are prepared to run in US East 1 - but can be used in other regions if you package and place the Lambda functions in the S3 in that region and change the bucket name in the parameters.
Step 2 – Open the testpage.html file in a text editor and insert the configuration parameters which you can get from the Outputs tab of the CloudFormation template once it’s created. Save the file once the configuration has been changed.
Step 3 – Open the testpage.html in a web browser and try submitting a value for a sensor ID using the Submit button. You can input any values you like into the text fields.
Step 4 – View the contents of the S3 bucket, which will have one object. Lambda Function 2 sets the key of the object to a today’s date and the sequence number of the Kinesis event.
Step 5 – Explore the DynamoDB table. The table will contain a record with the data submitted.
Step 6 – Using the Query button, make a query for the same sensor ID as you submitted an event for. The previously stored data is retrieved and displayed in the output field.
To remove all created resources, delete the CloudFormation stack.
Note: Deletion of the S3 bucket will fail unless all files in the bucket are removed before the stack is deleted.
This reference architecture sample is licensed under Apache 2.0.