A company runs a custom application on Amazon EC2 On-Demand Instances. The application has frontend nodes that must run 24/7 and backend nodes that run only briefly based on workload. The number of backend nodes varies during the day. The company needs to scale out and in according to workload. Which solution will meet these requirements MOST cost-effectively?

The correct answer is to use Reserved Instances (RIs) for frontend nodes and Spot Instances for backend nodes. RIs are ideal for predictable, long-running workloads like the 24/7 frontend nodes, offering significant cost savings compared to On-Demand instances. Spot Instances are highly cost-effective for fault-tolerant, flexible workloads that can tolerate interruptions, making them perfect for the backend nodes that run briefly and scale based on varying demand. Using Fargate for backend nodes would be more expensive than Spot Instances, as Fargate pricing is based on vCPU and memory usage, while Spot Instances leverage unused EC2 capacity at a much lower price. Using Spot Instances for frontend nodes is not suitable because these nodes must run 24/7 and cannot tolerate interruptions. Using RIs for backend nodes is not cost-effective since their runtime is brief and variable.