An enterprise is migrating a containerized microservices application to AWS. The architect… — AWS Certified Solutions Architect - Associate

An enterprise is migrating a containerized microservices application to AWS. The architecture consists of a steady-state backend API running on Amazon Elastic Container Service (Amazon ECS) on AWS Fargate, a relational database running on Amazon RDS for PostgreSQL, and a data warehouse hosted on Amazon Redshift. The compute requirements are as follows:

- The ECS backend API requires a constant baseline of $80\text{ vCPU}$ and $320\text{ GiB}$ of memory running 24/7.
- A nightly batch analytics job runs on ECS Fargate for 3 hours, requiring an additional $120\text{ vCPU}$ and $480\text{ GiB}$ of memory. The batch job is fault-tolerant and can resume if interrupted.
- The RDS database requires a constant baseline of $16\text{ vCPU}$ and $64\text{ GiB}$ of memory running 24/7.
- The Amazon Redshift cluster runs continuously to serve near-real-time business intelligence dashboards.

The enterprise wants to minimize costs over a 3-year period by committing to a purchasing strategy, while minimizing operational overhead.

Which purchasing strategy will achieve the lowest cost for these requirements?

Commit to a 3-year Compute Savings Plan for the baseline ECS Fargate compute. Purchase 3-year Reserved Instances for both the RDS PostgreSQL database and the Amazon Redshift cluster. Run the nightly batch jobs using Fargate Spot.Answer
B
Commit to a 3-year Compute Savings Plan with a commitment level sized to cover the baseline ECS Fargate compute, the RDS PostgreSQL database, and the Amazon Redshift cluster. Run the nightly batch jobs using Fargate Spot.
C
Commit to a 3-year Compute Savings Plan to cover the baseline ECS Fargate compute. Purchase a 3-year Reserved Instance for the RDS PostgreSQL database. Commit to a 3-year EC2 Instance Savings Plan for the Amazon Redshift cluster, and run the nightly batch jobs using Fargate Spot.
D
Commit to a 3-year Compute Savings Plan sized to cover the baseline ECS Fargate compute, the RDS PostgreSQL database, and the nightly batch jobs. Purchase a 3-year Reserved Node for the Amazon Redshift cluster.

Answer

Commit to a 3-year Compute Savings Plan for the baseline ECS Fargate compute, purchase 3-year Reserved Instances for both the RDS PostgreSQL database and the Amazon Redshift cluster, and run the nightly batch jobs using Fargate Spot.

The correct strategy combines a 3-year Compute Savings Plan for AWS Fargate baseline compute, 3-year Reserved Instances/Nodes for RDS and Redshift (since they are out of scope for Compute Savings Plans), and Fargate Spot for the fault-tolerant nightly batch job to achieve the maximum discount without commitment waste.

Step-by-Step Solution

Identify the steady-state and transient compute workloads.

Steady-state workloads: ECS Fargate API (24/7), RDS PostgreSQL (24/7), and Amazon Redshift (24/7). Transient workload: Nightly batch analytics job on ECS Fargate (3 hours/day).

Separating workloads helps determine where long-term commitment models apply versus flexible/spot models.

Select the correct purchasing model for the steady-state container compute.

A 3-year Compute Savings Plan is selected to cover the baseline ECS Fargate compute (

80\text{ vCPU}

and

320\text{ GiB}

of memory).

Compute Savings Plans apply to AWS Fargate, EC2, and Lambda, providing the maximum discount for continuous container workloads.

Determine the purchasing model for the database and data warehouse tiers.

Purchase a 3-year RDS Reserved Instance for the PostgreSQL database and a 3-year Redshift Reserved Node for the Redshift cluster.

Compute Savings Plans do not apply to RDS or Redshift. Dedicated Reserved Instances/Nodes must be purchased to reduce costs for these continuous database and data warehouse workloads.

Select the optimal purchasing model for the transient, fault-tolerant batch job.

Configure the nightly batch job to run on ECS Fargate Spot.

Since the batch job is fault-tolerant and runs for only 3 hours a day, Fargate Spot provides up to a 70% discount compared to On-Demand rates without incurring the 24/7 commitment costs of a Savings Plan.

Key Concept

Selecting cost-optimized compute purchasing strategies based on workload duration, service compatibility, and fault tolerance.

Estimated Time:3m 0s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics