AWS Certified Solutions Architect – Professional — Question 579

A company is migrating applications from on premises to the AWS Cloud. These applications power the company's internal web forms. These web forms collect data for specific events several times each quarter. The web forms use simple SQL statements to save the data to a local relational database.
Data collection occurs for each event, and the on-premises servers are idle most of the time. The company needs to minimize the amount of idle infrastructure that supports the web forms.
Which solution will meet these requirements?

Answer options

• A. Use Amazon EC2 Image Builder to create AMIs for the legacy servers. Use the AMIs to provision EC2 instances to recreate the applications in the AWS Cloud. Place an Application Load Balancer (ALB) in front of the EC2 instances. Use Amazon Route 53 to point the DNS names of the web forms to the ALB.
• B. Create one Amazon DynamoDB table to store data for all the data input. Use the application form name as the table key to distinguish data items. Create an Amazon Kinesis data stream to receive the data input and store the input in DynamoDB. Use Amazon Route 53 to point the DNS names of the web forms to the Kinesis data stream's endpoint.
• C. Create Docker images for each server of the legacy web form applications. Create an Amazon Elastic Container Service (Amazon EC2) cluster on AWS Fargate. Place an Application Load Balancer in front of the ECS cluster. Use Fargate task storage to store the web form data.
• D. Provision an Amazon Aurora Serverless cluster. Build multiple schemas for each web form's data storage. Use Amazon API Gateway and an AWS Lambda function to recreate the data input forms. Use Amazon Route 53 to point the DNS names of the web forms to their corresponding API Gateway endpoint.

Correct answer: D

Explanation

Option D is the ideal solution because Amazon API Gateway, AWS Lambda, and Amazon Aurora Serverless are fully managed serverless services that scale down to zero when there is no traffic, effectively eliminating idle infrastructure costs while maintaining support for the application's SQL queries. Options A and C rely on persistent compute resources (EC2 and Fargate behind an ALB) which would continuously incur costs even when idle. Option B is incorrect because DynamoDB is a NoSQL database, which is incompatible with the application's existing relational SQL queries, and Kinesis is not built to host web forms.