Examkin

Soru

Zorluk: OrtaCost-Optimized Compute Selection and Purchasing Strategies

A software company is redesigning the compute architecture for its multi-tenant application to optimize costs. The application runs on Amazon ECS on AWS Fargate and consists of two main workloads:

1. A suite of web APIs that runs continuously (24/724/7) to handle customer requests. This workload has a highly predictable baseline of 1010 ECS tasks, with occasional scaling peaks.
2. A background processing system that pulls video transcoding jobs from an Amazon SQS queue. The jobs are stateless, can be interrupted and resumed without issues, and run sporadically throughout the day.

Which combination of compute purchasing strategies will meet these requirements most cost-effectively? (Choose two.)

  1. Purchase a Compute Savings Plan to cover the baseline usage of the continuously running ECS tasks.Cevap
  2. Run the background video transcoding ECS tasks on Fargate Spot.Cevap
  3. C
    Purchase a Compute Savings Plan to cover both the ECS tasks and the application's Amazon Aurora PostgreSQL database compute costs.
  4. D
    Migrate the background video transcoding jobs to AWS Lambda and configure them to run continuously as long-running processes.
  5. E
    Purchase EC2 Instance Savings Plans to apply discounts to the ECS tasks running on AWS Fargate.

Cevap

The most cost-effective compute purchasing strategy combines a Compute Savings Plan for the baseline web API ECS tasks and Fargate Spot for the stateless, interruptible background video transcoding tasks.
Purchasing a Compute Savings Plan is the most cost-effective way to cover the continuously running Fargate tasks, as it provides a discount in exchange for a commitment to a consistent amount of compute usage. Running the stateless, queue-driven video transcoding tasks on Fargate Spot leverages spare capacity at a substantial discount, which is perfect since the workload can be interrupted and resumed via the Amazon SQS queue.

Adım Adım Çözüm

1
Analyze the baseline API workload requirements.
The API workload runs 24/724/7 continuously and has a predictable baseline of 1010 tasks, meaning it represents a stable, commit-friendly compute load.
Stable, predictable compute workloads are ideal candidates for commit-based discounts like Savings Plans.
2
Evaluate Savings Plans options for Fargate compute.
Only Compute Savings Plans apply discounts to AWS Fargate. EC2 Instance Savings Plans do not cover Fargate usage.
Compute Savings Plans offer up to 66% discount on Fargate, Lambda, and EC2 instances.
3
Analyze the background video transcoding workload requirements.
The transcoding jobs are stateless, interruptible, pull from SQS, and run sporadically.
Workloads that can tolerate interruption and run asynchronously are ideal for Spot compute models.
4
Select the appropriate Spot purchasing model for Fargate container tasks.
Fargate Spot is selected as it runs containers at a significantly reduced rate (up to 70% off standard Fargate pricing) while tolerating tasks being terminated with a 2-minute warning.
Fargate Spot matches ECS tasks on Fargate directly and offers the highest savings for interruptible containerized workloads.

Anahtar Kavram

Selecting the optimal compute purchasing model (Savings Plans vs. Spot Instances) for containerized ECS Fargate workloads based on predictability and tolerance to interruption.
Bu soruyu puanla

Konular

Cost-Optimized Compute Selection and Purchasing Strategies