A legacy application must be migrated from an on-premises data center to AWS because of hardware capacity limits. The application runs 24/7 and its database storage will continue to grow over time. What is the MOST cost-effective migration approach?

The most cost-effective approach for a 24/7 application with growing data is to use Reserved Instances (RIs) for both compute and database. EC2 RIs offer significant discounts (up to 72%) compared to On-Demand pricing for instances committed for 1 or 3 years, making them ideal for a continuously running application. Similarly, Amazon Aurora Reserved Instances provide substantial savings for database capacity. Migrating the application to EC2 Spot Instances is not suitable because Spot Instances can be interrupted with short notice, which is unacceptable for a 24/7 production application. Migrating data storage to Amazon S3 is incorrect because S3 is object storage, not a relational database service suitable for a legacy application's database. Using Amazon RDS On-Demand Instances for the database would be more expensive than Aurora RIs for a long-running workload. Finally, EC2 On-Demand Instances for the application layer would also be less cost-effective than EC2 RIs for a 24/7 application.