Answer: B
From the scenario, we can see that the On-Premises applications` workload is continuous & stable. For these kinds of applications, it is easy to predict upfront capacity. Also, since the application runs continuously, I will benefit by reserving capacity for a certain period of time.
Option A is incorrect. The Pay-as-you-go model will be best used for workloads used for short durations & have unpredictable load. On-demand EC2 instances will be the best fit for this purpose.
Option B is CORRECT. Since there is continuous usage of these applications with a predictable load, it will be best for me to reserve capacity upfront (period of 1 - 3 years) that will provide a substantial discount of 30 - 50% compared to its On Demand counterparts.
Option C is incorrect. Pay less by using more refers to volume discounts provided by AWS for increased usage.
E.g., S3 Standard provides the following storage pricing, also referred to as Tiered-Pricing.
Data Storage
Storage Pricing
First TB / month
$0.025 per GB
Next 450 TB / month
$0.024 per GB
Next 500 TB / month
$0.023 per GB
Option D is incorrect. Pay-per-compute-time refers to the use of serverless architectures like Lambda, where you pay only for the time when the compute resources are running. Unlike EC2 Pay-as-you-go pricing, AWS provisions resources for executing Lambda functions on the fly & removes them immediately after execution. So there is no idle utilization time that needs to be accounted for. Since our scenario consists of long-running applications, this option will be impractical for usage.
References:
https://aws.amazon.com/pricing/#:~:text=AWS%20offers%20you%20a%20pay,utilities%20like%20water%20and%20electricity.
https://dzone.com/articles/the-cost-of-the-cloud-the-ultimate-aws-pricing-gui
https://www.apptio.com/blog/aws-reserved-instances-cost-optimization/
https://d1.awsstatic.com/whitepapers/aws_pricing_overview.pdf