Examkin

Question

Difficulty: HardAWS Pricing Models

A medical research institute is deploying a genomics platform on Amazon EC2 with three distinct workloads:

1. A public-facing web portal that must run continuously 24/7 with a predictable, constant baseline capacity.
2. A data analysis cluster that processes large genetic datasets. The processing jobs are fault-tolerant, can be paused and resumed, and have no strict completion deadlines.
3. A clinical reporting application used by hospital staff during local business hours (9:00 AM to 5:00 PM, Monday through Friday). This application requires constant availability during these hours and cannot tolerate any service interruptions.

Which combination of Amazon EC2 pricing models is the most cost-effective for these workloads?

  1. A
    Spot Instances for the public-facing web portal, Savings Plans for the data analysis cluster, and On-Demand Instances for the clinical reporting application
  2. B
    On-Demand Instances for all three workloads to ensure the institute avoids operating expenses (OpEx) and relies entirely on capital expenses (CapEx)
  3. Savings Plans for the public-facing web portal, Spot Instances for the data analysis cluster, and On-Demand Instances for the clinical reporting applicationAnswer
  4. D
    Reserved Instances for the public-facing web portal, On-Demand Instances for the data analysis cluster, and Savings Plans for the clinical reporting application

Answer

Savings Plans for the public-facing web portal, Spot Instances for the data analysis cluster, and On-Demand Instances for the clinical reporting application
The public-facing web portal has a continuous, predictable 24/7 load, making a commitment-based model like Savings Plans (or Reserved Instances) the most cost-effective choice. The data analysis cluster runs fault-tolerant, interruptible workloads, which are ideal for Spot Instances to achieve the maximum discount. The clinical reporting application only runs during business hours (40 hours per week, which is approximately 24% utilization). Committing to a 24/7 Savings Plan for this application would be wasteful compared to using On-Demand Instances and shutting them down outside of business hours.

Step-by-Step Solution

1
Evaluate the pricing model for the 24/7 public web portal.
Savings Plans or Reserved Instances are chosen.
Since the portal requires continuous, predictable 24/7 availability, committing to a 1-year or 3-year term yields the highest discount compared to On-Demand rates.
2
Evaluate the pricing model for the data analysis cluster.
Spot Instances are chosen.
Because the processing jobs are fault-tolerant, can be paused and resumed, and have no tight deadlines, they can utilize spare AWS capacity via Spot Instances at up to a 90% discount.
3
Evaluate the pricing model for the clinical reporting application.
On-Demand Instances are chosen.
The application runs only 40 hours a week. Running On-Demand costs less because the utilization is only 40 hours/168 hours≈23.8%40 \text{ hours} / 168 \text{ hours} \approx 23.8\%. Purchasing a 24/7 commitment (Savings Plan or Reserved Instance) for this workload would charge for the remaining 128 unused hours, making it more expensive than paying the On-Demand rate.
4
Combine the decisions into a single optimal recommendation.
Savings Plans for the portal, Spot Instances for the analysis, and On-Demand for the reporting application.
This combination minimizes the total cost of ownership (TCO) while matching the reliability constraints of each workload.

Key Concept

Selecting cost-optimized Amazon EC2 pricing models based on workload utilization, predictability, and fault tolerance.
Rate this question

Topics

AWS Pricing Models

More questions from this topic