AWS Lambda is a serverless, event-driven compute service that runs your function code without you having to manage or scale servers. It provides an API to trigger the execution of your functions and ensures that the function are executed in parallel, regardless of scale. It provides additional capabilities for your functions such as logging and monitoring. Lambda functions are essentially stateless, trigger-based code execution. For example, the streaming events could be landing on S3 or DynamoDB, or they could be an API Gateway call, or even a scheduled job. These functions can access services inside or outside your VPC.

Compared to provisioning and managing EC2 instances yourself, using the Lambda service means no servers or instances to manage and it comes with built-in scaling and a fixed cost model. However, as AWS Lambda is a relatively new service, you will have limited experience with it, and you will be limited to using CloudWatch for monitoring (at this time). Additionally, there is a 6 MB data limit on Lambda functions, and debugging the logs can be a very time-consuming task.