Examkin

Question

Difficulty: Very hardAWS Pricing Models

A financial services company is optimizing the deployment costs for two newly migrated workloads on Amazon EC2. The first workload is a real-time trading API that experiences sudden, sharp spikes in demand during market hours. The API cannot tolerate any service interruptions or latency increases, and the capacity requirement during non-trading hours is negligible. The second workload is a batch risk-modeling simulation that runs nightly, requires hundreds of vCPUs, is designed to tolerate interruptions by saving state, and must be completed at the lowest possible cost.

Which of the following EC2 pricing models should the company combine to achieve the most cost-effective solution for both workloads? (Select TWO.)

  1. On-Demand Instances to handle the sudden, unpredictable scaling demands of the trading APIAnswer
  2. Spot Instances for the fault-tolerant, nightly batch risk-modeling simulationAnswer
  3. C
    Dedicated Instances to guarantee hardware isolation for the nightly batch risk-modeling simulation
  4. D
    Reserved Instances to cover the peak capacity needs of the trading API spikes
  5. E
    Savings Plans to cover the compute usage of the nightly batch risk-modeling simulation

Answer

The company should combine On-Demand Instances to handle the sudden, unpredictable scaling demands of the trading API and Spot Instances for the fault-tolerant, nightly batch risk-modeling simulation.
The correct combination consists of On-Demand Instances and Spot Instances. On-Demand Instances provide uninterrupted compute capability without long-term commitments, which matches the trading API's strict SLAs and unpredictable spike patterns. Spot Instances offer the deepest discounts (up to 90%) for the nightly batch risk-modeling simulation, which is designed to tolerate interruptions by saving state.

Step-by-Step Solution

1
Analyze the trading API workload requirements.
The trading API requires immediate capacity during market hours, experiences unpredictable spikes, cannot tolerate interruptions, and has near-zero usage off-hours.
Identifying these traits helps eliminate options that involve long-term commitments for idle capacity or interruptible models.
2
Select the optimal pricing model for the trading API.
On-Demand Instances are chosen because they offer zero risk of interruption and charge only for active compute time with no long-term commitment.
This guarantees performance SLAs while keeping costs aligned to actual spiky usage patterns.
3
Analyze the nightly batch simulation workload requirements.
The nightly batch simulation requires high compute capacity (hundreds of vCPUs), runs for a short duration nightly, is designed to save state and tolerate interruptions, and prioritizes lowest cost.
These characteristics point toward leveraging AWS spare capacity discount models.
4
Select the optimal pricing model for the batch simulation.
Spot Instances are chosen because they offer the lowest cost (up to 90% discount) for interruptible, fault-tolerant batch workloads.
This minimizes the cost of massive parallel compute while accommodating the potential for instance reclamation.

Key Concept

Selecting the most cost-effective EC2 pricing models based on workload predictability, tolerance to interruption, and duration.
Rate this question

Topics

AWS Pricing Models

More questions from this topic