Examkin

Question

Difficulty: MediumCost-Optimized Compute Selection and Purchasing Strategies

A financial technology company is designing a cost-optimization strategy for its application environment on AWS. The application consists of three main components:

1. A core transactional processing service running continuously 24/7 on Amazon ECS with AWS Fargate.
2. A metadata repository on Amazon DynamoDB that handles highly unpredictable, spiky request volumes throughout the day.
3. An image-processing task that runs for less than 10 seconds per invocation, triggered whenever users upload new documents to Amazon S3.

Which of the following compute selection and purchasing strategies is the most cost-effective for this environment?

  1. Run the ECS tasks on AWS Fargate covered by a Compute Savings Plan; configure Amazon DynamoDB in On-Demand capacity mode; and use AWS Lambda for the short-lived document processing tasks.Answer
  2. B
    Run the ECS tasks on AWS Fargate, configure Amazon DynamoDB in Provisioned capacity mode, and cover all compute and database costs (Fargate, DynamoDB, and Lambda) under a single Compute Savings Plan.
  3. C
    Migrate the continuous 24/7 ECS core transactional processing service to run on AWS Lambda; configure Amazon DynamoDB in On-Demand capacity mode; and purchase a Compute Savings Plan to cover all Lambda functions.
  4. D
    Run the ECS tasks on AWS Fargate covered by a Compute Savings Plan; configure Amazon DynamoDB in Provisioned capacity mode with high fixed read/write capacity units to handle the peaks; and use AWS Lambda for the short-lived tasks.

Answer

Run the ECS tasks on AWS Fargate covered by a Compute Savings Plan; configure Amazon DynamoDB in On-Demand capacity mode; and use AWS Lambda for the short-lived document processing tasks.
The correct strategy combines Compute Savings Plans for the steady-state ECS Fargate tasks, DynamoDB On-Demand capacity mode for the spiky, unpredictable database traffic, and AWS Lambda for the short-lived, event-driven tasks. Compute Savings Plans apply to Fargate and Lambda, offering up to a 66% discount on Fargate workloads. DynamoDB On-Demand avoids over-provisioning costs by charging only for active read/write requests. AWS Lambda is cost-effective for sub-second and low-duration tasks as you only pay when the function is active.

Step-by-Step Solution

1
Analyze the compute requirements for the core transactional processing service running continuously 24/7 on Amazon ECS with AWS Fargate.
Identify that a Compute Savings Plan is the most cost-effective purchasing model for Fargate as it offers up to 66% savings over On-Demand rates for a 1- or 3-year commitment, fitting the steady-state nature of the workload.
Steady-state containerized workloads running continuously benefit significantly from the commitment-based discounts of Compute Savings Plans.
2
Analyze the database requirements for Amazon DynamoDB with highly unpredictable, spiky traffic.
Select On-Demand capacity mode rather than Provisioned capacity mode.
On-Demand capacity mode charging per request is more cost-effective for unpredictable and spiky workloads because it avoids the high cost of provisioning for peak capacity that remains largely idle.
3
Analyze the requirements for the short-lived image-processing tasks running for less than 10 seconds.
Select AWS Lambda as the compute service.
AWS Lambda is designed for short-lived, event-driven functions and automatically scales. Since charges are based on execution duration (rounded to the millisecond) and execution count, it is highly cost-effective compared to maintaining idle EC2 or Fargate instances.

Key Concept

Selecting cost-optimal compute services (ECS Fargate vs Lambda) and aligning them with appropriate purchasing strategies (Compute Savings Plans) and database scaling modes (DynamoDB On-Demand).
Estimated Time:1m 30s
Rate this question

Topics

Cost-Optimized Compute Selection and Purchasing Strategies