Examkin

Question

Difficulty: MediumPlanning Compute Engine Resources and Machine Types

A financial services organization is planning the Google Compute Engine infrastructure for a continuous, memory-intensive risk modeling service that operates 24/7. The application requires 6 vCPUs and 48 GB of RAM to maintain optimal throughput, but cannot tolerate unexpected node shutdowns or terminations. The team wants to ensure high availability while minimizing compute costs over a 3-year production deployment period. Which Compute Engine configuration strategy should you recommend?

  1. A
    Provision Spot VMs with a custom machine type of 6 vCPUs and 48 GB of RAM to leverage maximum hourly discounts.
  2. B
    Provision standard N2 instances with 8 vCPUs and reliance on automatic Sustained Use Discounts without purchase commitments.
  3. Provision a custom N2 machine type with 6 vCPUs and 48 GB of RAM, combined with a 3-year Committed Use Discount (CUD) for compute resources.Answer
  4. D
    Migrate the risk modeling backend service to Cloud Functions using memory-allocated trigger execution instances.

Answer

Provision a custom N2 machine type configured with 6 vCPUs and 48 GB of RAM, combined with a 3-year Committed Use Discount (CUD).
Configuring a custom machine type allows exact tailoring of vCPUs (6) and RAM (48 GB) to prevent overprovisioning costs. Combining this with a 3-year Committed Use Discount delivers the maximum possible cost reduction for uninterrupted, continuous production workloads.

Step-by-Step Solution

1
Analyze workload resource requirements and lifecycle constraints
Workload is continuous (24/7), requires explicit non-standard CPU-to-memory ratio (6 vCPUs to 48 GB RAM), and requires high availability (no preemption).
Matching exact workload characteristics prevents capacity overprovisioning and application failures.
2
Determine machine type sizing approach
Select a Custom N2 machine type with 6 vCPUs and 48 GB RAM.
Standard machine types follow rigid ratios (e.g. 1:4 vCPU-to-RAM ratio). Custom machine types avoid paying for unused vCPUs when exact memory targets are needed.
3
Select appropriate pricing optimization mechanism for 24/7 steady-state usage over 3 years
Apply a 3-year Committed Use Discount (CUD).
Committed Use Discounts offer the highest cost savings for steady-state continuous workloads, whereas Spot VMs are risky due to preemption and Sustained Use Discounts provide lower overall savings.

Key Concept

Compute Engine Custom Machine Types and Committed Use Discounts (CUDs)
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic