A .NET Core Lambda function needs to interact with DynamoDB and S3 while minimizing deployment package size and reducing invocation latency. Which approach meets these goals?

Including only the necessary AWS SDK for .NET modules (DynamoDB and S3) directly addresses both goals. This minimizes the deployment package size, which in turn reduces the time it takes to upload the function and extract it for execution, thus lowering invocation latency. Increasing memory allocation might reduce invocation latency by providing more CPU, but it doesn't minimize package size. Including the full AWS SDK for .NET unnecessarily increases package size and latency. Downloading the SDK at runtime from S3 would significantly increase invocation latency due to the network overhead and download time for every cold start.