Our speed layer is the simplest part of the entire system. With the configuration in place—a single Lambda function to execute whenever a new stream message arrives—the only work we need to do is decode the data from the message payload, calculate the DynamoDB partition key, and write it to DynamoDB. The following code block shows all of this work in the single function, which is processing a single message from AWS Kinesis:
This code is all located in a single handler.py function, which goes against a best practice of splitting up application code and decoupling application logic from the cloud-specific bits. However, this application code is for demonstration purposes, and I can get away with breaking some rules for the sake ...
