A financial services company runs a high-throughput processing application on AWS. The app… — AWS Certified Solutions Architect - Associate

A financial services company runs a high-throughput processing application on AWS. The application has two primary compute requirements: a web-tier backend running on Amazon ECS using AWS Fargate that requires a continuous, predictable baseline of $64$ vCPUs and $256$ GB RAM 24/7, scaling up during peak business hours; and a batch-tier risk analysis job that runs nightly for $6$ hours, requiring up to $1,000$ vCPUs of highly parallelized, fault-tolerant compute that checkpoints its state to Amazon DynamoDB. The application's database layer runs on an Amazon Aurora PostgreSQL DB cluster. Which two options should the solutions architect select to design a hosting and purchasing strategy that minimizes costs while meeting these requirements?

Purchase a Compute Savings Plan to cover the continuous 24/7 baseline of the Amazon ECS tasks running on AWS Fargate.Answer
Deploy the batch processing tier on Amazon EC2 Spot Instances using an EC2 Fleet with a capacity-optimized allocation strategy.Answer
C
Purchase an EC2 Instance Savings Plan to lower the cost of the Amazon Aurora PostgreSQL database instances.
D
Re-architect the 6-hour batch processing workload to run on AWS Lambda functions to leverage scale-to-zero billing.
E
Purchase an EC2 Instance Savings Plan to cover the 24/7 baseline usage of the Amazon ECS tasks running on AWS Fargate.

Answer

The correct strategy is to purchase a Compute Savings Plan for the steady-state Fargate tasks and deploy the fault-tolerant batch tier on Amazon EC2 Spot Instances.

Purchasing a Compute Savings Plan covers the continuous 24/7 baseline of the Amazon ECS tasks running on AWS Fargate. Deploying the batch processing tier on Amazon EC2 Spot Instances with a capacity-optimized allocation strategy minimizes compute costs for a parallelized, fault-tolerant workload that checkpoints its state.

Step-by-Step Solution

Analyze the baseline compute requirement for the 24/7 ECS web tier.

The web tier runs on AWS Fargate continuously. Fargate compute costs are eligible for Compute Savings Plans, but not EC2 Instance Savings Plans.

This determines the correct discount type for the Fargate baseline usage.

Analyze the batch processing workload's characteristics.

The batch job runs for 6 hours, is highly parallelized, fault-tolerant, and checkpoints its state. This fits the profile for Spot Instances.

This selects the most cost-effective compute pricing model for transient, interruption-tolerant workloads.

Evaluate the database layer cost-optimization options.

Database costs for Amazon Aurora cannot be covered by EC2 Instance Savings Plans or Compute Savings Plans.

This rules out incorrect purchasing models that attempt to apply compute-specific Savings Plans to database services.

Key Concept

Selecting and combining cost-effective compute options (Compute Savings Plans, Spot Instances) based on workload stability, duration, tolerance to interruption, and service scope constraints.

Estimated Time:3m 0s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics