Examkin

Soru

Zorluk: OrtaCost-Optimized Compute Selection and Purchasing Strategies

A company is designing a batch processing solution on AWS that runs containerized simulation jobs on a daily schedule for 4 hours. The jobs are stateless, fault-tolerant, and designed to checkpoint their progress so they can resume if interrupted. The simulation results must be stored in an Amazon RDS database, while the raw input data is temporarily stored in Amazon S3 for 10 days before being deleted. Which strategy is the most cost-effective to meet these requirements?

  1. Deploy the simulation jobs on Amazon ECS using Amazon EC2 Spot Instances, and store the raw input data in Amazon S3 Standard.Cevap
  2. B
    Deploy the simulation jobs on AWS Lambda functions to eliminate idle compute costs, and store the raw input data in Amazon S3 Standard.
  3. C
    Deploy the simulation jobs on Amazon ECS using AWS Fargate, purchase a Compute Savings Plan to cover both the Fargate tasks and the Amazon RDS database, and store the raw input data in Amazon S3 Standard.
  4. D
    Deploy the simulation jobs on Amazon ECS using Amazon EC2 Spot Instances, and store the raw input data in Amazon S3 Standard-Infrequent Access (S3 Standard-IA).

Cevap

Deploy the simulation jobs on Amazon ECS using Amazon EC2 Spot Instances, and store the raw input data in Amazon S3 Standard.
Deploying the simulation jobs on Amazon ECS using Amazon EC2 Spot Instances is the most cost-effective compute strategy because Spot Instances offer up to a 90% discount and are designed for stateless, checkpoint-enabled workloads. Additionally, storing raw data in Amazon S3 Standard is correct because it has no minimum storage duration charge, making it cheaper than S3 Standard-IA for data deleted after 10 days.

Adım Adım Çözüm

1
Evaluate the compute requirements and purchasing strategies.
Identify that the containerized workload runs for a short duration (4 hours daily) and is stateless and fault-tolerant. This makes EC2 Spot Instances on Amazon ECS the most cost-effective compute choice compared to On-Demand, Fargate, or Lambda.
Spot Instances provide maximum savings for interruptible workloads, and the 4-hour daily duration does not justify long-term commitments like Savings Plans for compute.
2
Assess the storage tiering for the 10-day retention period.
Determine that S3 Standard must be used instead of S3 Standard-IA.
S3 Standard-IA charges a minimum of 30 days of storage. Deleting the raw data after 10 days incurs a cost penalty for the unused 20 days, making S3 Standard more cost-effective for short-lived datasets.
3
Validate the coverage of purchasing discounts and service limits.
Confirm that Compute Savings Plans do not extend to Amazon RDS, and AWS Lambda cannot support continuous 4-hour executions due to its 15-minute timeout.
Eliminating options that assume incorrect discount scopes or violate service limitations ensures a technically sound and cost-optimized architecture.

Anahtar Kavram

Selecting cost-optimized compute purchasing models and storage classes based on workload duration, fault tolerance, and data retention policies.
Bu soruyu puanla

Konular

Cost-Optimized Compute Selection and Purchasing Strategies