Examkin

Soru

Zorluk: ZorCost-Optimized Compute Selection and Purchasing Strategies

A financial services company is designing a cost-optimization strategy for its risk analysis platform. The platform consists of two main compute workloads: a containerized orchestration API running continuously on AWS Fargate, and a series of memory-intensive risk simulation workers running on Amazon EC2. The simulation workers are stateless, run in parallel for up to 44 hours at a time, and can be safely interrupted and restarted. The company also runs its primary database on Amazon RDS for PostgreSQL. The company wants to minimize compute costs over the next 33 years.

Which combination of purchasing strategies should a solutions architect recommend to meet these requirements most cost-effectively? (Select TWO.)

  1. Commit to a 33-year Compute Savings Plan to cover the baseline usage of the AWS Fargate API orchestration service.Cevap
  2. B
    Commit to a 33-year Compute Savings Plan to cover the baseline usage of both the AWS Fargate API orchestration service and the Amazon RDS database.
  3. Configure the risk-simulation worker workloads to run on Amazon EC2 Spot Instances.Cevap
  4. D
    Commit to a 33-year EC2 Instance Savings Plan to cover the maximum capacity of the risk-simulation worker instances.
  5. E
    Deploy the risk-simulation worker tasks as AWS Lambda functions using On-Demand capacity to eliminate idle instance costs.

Cevap

To minimize compute costs, the solutions architect should recommend committing to a 33-year Compute Savings Plan for the baseline usage of the AWS Fargate API orchestration service, and configuring the risk-simulation worker workloads to run on Amazon EC2 Spot Instances.
The correct strategy combines a 33-year Compute Savings Plan to cover the continuous, steady-state baseline of the containerized API running on AWS Fargate, and utilizes Amazon EC2 Spot Instances for the risk-simulation workers. Since the simulation workers are stateless, parallel, and can tolerate interruptions, Spot Instances provide the most cost-effective model, offering up to 90% savings. Compute Savings Plans are ideal for Fargate as they automatically apply across Fargate deployments with a long-term commitment.

Adım Adım Çözüm

1
Analyze the operational and architectural characteristics of the workloads.
The Fargate API orchestration service runs continuously (24/724/7), representing a steady-state compute pattern. The risk-simulation workers are stateless, parallel, fault-tolerant, and run for up to 44 hours at a time. The database runs on Amazon RDS.
Identifying workload patterns is the first step in matching them with correct AWS pricing models.
2
Select the optimal purchasing strategy for the steady-state containerized workload.
A Compute Savings Plan is selected for the AWS Fargate API because it provides significant discounts (up to 66%) for a 33-year commitment and automatically applies to Fargate container usage.
Savings Plans offer the best cost-reduction mechanism for predictable, long-term containerized workloads.
3
Select the optimal purchasing strategy for the fault-tolerant batch workloads.
Amazon EC2 Spot Instances are selected for the simulation workers because they offer up to 90% savings over On-Demand and match the stateless, interruptible nature of the 44-hour batch jobs.
Spot Instances provide the highest discount rate for compute resources that do not require high availability guarantees.
4
Evaluate and discard options that violate service limits or purchasing plan scopes.
AWS Lambda is excluded due to the 1515-minute timeout limit. A Compute Savings Plan cannot be applied to Amazon RDS. committing to EC2 Instance Savings Plans for spiky, intermittent compute is cost-ineffective compared to Spot Instances.
Ensures the recommended solution is both architecturally viable and financially optimal.

Anahtar Kavram

Matching AWS compute purchasing models (Savings Plans, Spot Instances) to distinct workload patterns (steady-state vs. stateless batch compute) while recognizing structural service limits (Lambda timeout) and plan scope constraints (RDS exclusion).
Bu soruyu puanla

Konular

Cost-Optimized Compute Selection and Purchasing Strategies