A company hosts a two-tier application on Amazon EC2 instances and Amazon RDS.

This solution meets the requirements of a two-tier application that has a variable demand based on the time of day and must be available at all times, while minimizing the overall cost. EC2 Reserved Instances can provide significant savings compared to On-Demand Instances for the baseline level of usage, and they can guarantee capacity reservation when needed. EC2 Spot Instances can provide up to 90% savings compared to On-Demand Instances for any additional capacity that the application needs during peak hours. Spot Instances are suitable for stateless applications that can tolerate interruptions and can be replaced by other instances. Stopping the RDS database when it is not in use can reduce the cost of running the database tier.

Option A is incorrect because using all EC2 Spot Instances can affect the availability of the application if there are not enough spare capacity or if the Spot price exceeds the maximum price. Stopping the RDS database when it is not in use can reduce the cost of running the database tier, but it can also affect the availability of the application. Option B is incorrect because purchasing EC2 Instance Savings Plans to cover five EC2 instances can lock in a fixed amount of compute usage per hour, which may not match the actual usage pattern of the application. Purchasing an RDS Reserved DB Instance can provide savings for the database tier, but it does not allow stopping the database when it is not in use. Option D is incorrect becausepurchasing EC2 Instance Savings Plans to cover two EC2 instances can lock in a fixed amount of compute usage per hour, which may not match the actual usage pattern of the application. Using up to three additional EC2 On-Demand Instances as needed can incur higher costs than using Spot Instances.