Examkin

Soru

Zorluk: ZorCost-Optimized Compute Selection and Purchasing Strategies

A media company is migrating its video processing pipeline to AWS. The architecture consists of the following components:

1. An orchestration and management web portal that runs continuously 24/7 on Amazon ECS Fargate, requiring a stable baseline of compute resources.
2. A batch video transcoding workload that runs containerized tasks on Amazon ECS Fargate. Individual transcoding jobs can run for several hours, are designed to be fault-tolerant, and can be retried if interrupted.
3. A metadata backend database running on an Amazon Aurora PostgreSQL DB cluster that must be available 24/7.

Which combination of configurations and purchasing strategies is the MOST cost-effective to meet these requirements?

  1. A
    Purchase a single, large Compute Savings Plan to cover the compute costs of the orchestration service, the batch video transcoding workload, and the Amazon Aurora database cluster.
  2. B
    Migrate the batch video transcoding workload to run on AWS Lambda, purchase a Compute Savings Plan for the orchestration service, and purchase Amazon Aurora Reserved DB Instances for the database cluster.
  3. Purchase a Compute Savings Plan to cover the baseline ECS Fargate orchestration service, run the batch video transcoding workload on Amazon ECS Fargate Spot instances, and purchase Amazon Aurora Reserved DB Instances for the database cluster.Cevap
  4. D
    Purchase a Compute Savings Plan for the orchestration service, run the batch video transcoding workload on Amazon ECS Fargate Spot instances while routing all video data transfers from private tasks to Amazon S3 via NAT Gateways, and run the Aurora database cluster on Demand.

Cevap

Purchase a Compute Savings Plan to cover the baseline ECS Fargate orchestration service, run the batch video transcoding workload on Amazon ECS Fargate Spot instances, and purchase Amazon Aurora Reserved DB Instances for the database cluster.
The option recommending a Compute Savings Plan for the 24/7 orchestration service, ECS Fargate Spot for the batch video transcoding workload, and Amazon Aurora Reserved DB Instances for the database cluster is correct. It properly identifies that the predictable ECS Fargate orchestration service benefits from a Compute Savings Plan commitment, the interruptible batch processing tier is best suited for the low-cost Fargate Spot tier, and the 24/7 database cluster is optimized using DB-specific Reserved Instances.

Adım Adım Çözüm

1
Analyze the compute characteristics of the 24/7 orchestration service.
Determine that since it runs continuously on Amazon ECS Fargate, it requires a committed compute purchasing model.
A Compute Savings Plan provides up to a 66% discount on ECS Fargate usage in exchange for a 1-year or 3-year commitment, offering the best cost savings for a stable compute baseline.
2
Evaluate the requirements and constraints of the batch video transcoding workload.
Select ECS Fargate Spot instances rather than On-Demand or AWS Lambda.
Transcoding jobs run for hours, ruling out AWS Lambda due to its 15-minute maximum timeout. Because the jobs are fault-tolerant and retryable, Fargate Spot provides discounts up to 70% compared to On-Demand with the trade-off of potential interruption.
3
Analyze the backend database workload.
Select Amazon Aurora Reserved DB Instances for the 24/7 Aurora PostgreSQL cluster.
Database instances that run continuously 24/7 are not covered by Compute Savings Plans, but purchasing Reserved DB Instances provides significant discounts for the database tier.

Anahtar Kavram

Matching different AWS workloads with the most cost-effective compute purchasing models (Compute Savings Plans, Fargate Spot, and Reserved DB Instances) based on execution duration, predictability, and fault tolerance.
Tahmini Süre:2m 30s
Bu soruyu puanla

Konular

Cost-Optimized Compute Selection and Purchasing Strategies