Examkin

Soru

Zorluk: ZorPlanning Compute Engine Resources and Machine Types

A financial analytics company is designing its Google Cloud infrastructure to handle two distinct workload requirements:
1. A batch risk-modeling application that runs nightly for 6 hours, processing fault-tolerant worker tasks that checkpoint state to Cloud Storage every 10 minutes.
2. A core licensing and authentication service that must run continuously 24/7 with strict uptime requirements.

The team wants to optimize infrastructure costs while maintaining compliance with operational requirements. Which Compute Engine provisioning strategy should the cloud engineer recommend?

  1. Provision Spot VMs using Managed Instance Groups for the batch risk-modeling application, and use standard Compute Engine instances with Committed Use Discounts (CUDs) for the 24/7 licensing service.Cevap
  2. B
    Provision Spot VMs for both the batch risk-modeling application and the 24/7 licensing service to maximize overall cost savings across all Compute Engine instances.
  3. C
    Migrate both the batch risk-modeling pipeline and the 24/7 licensing service to GKE Autopilot clusters to eliminate the need for manual VM machine type selection.
  4. D
    Deploy the 6-hour batch risk-modeling application using Cloud Functions triggered by Cloud Scheduler to eliminate Compute Engine virtual machine instance management completely.

Cevap

Use Spot VMs for the fault-tolerant, checkpointed batch processing application, and use standard VM instances backed by Committed Use Discounts for the continuous 24/7 licensing service.
The correct strategy leverages Spot VMs for the batch application because its state is saved every 10 minutes and it can tolerate instance preemption in exchange for high cost discounts. For the 24/7 licensing service, standard instances backed by Committed Use Discounts (CUDs) guarantee baseline availability while securing deep financial savings.

Adım Adım Çözüm

1
Analyze the requirements for Workload 1 (Batch risk-modeling).
Workload 1 runs for 6 hours nightly and regularly checkpoints data, indicating high fault tolerance.
Fault-tolerant batch jobs are ideal candidates for Spot VMs, offering up to 60-91% discount off standard pricing.
2
Analyze the requirements for Workload 2 (Licensing service).
Workload 2 requires continuous 24/7 uptime with high availability.
Steady-state continuous workloads should use standard VMs combined with Committed Use Discounts (CUDs) to maximize savings without risk of preemption.
3
Synthesize the optimal Google Cloud compute strategy.
Combine Spot VMs for batch compute and Committed Use Discounts for baseline 24/7 workloads.
This combination aligns perfectly with Google Cloud recommended practices for cost optimization and workload reliability.

Anahtar Kavram

Compute Engine Workload Planning: Spot VMs vs Committed Use Discounts
Tahmini Süre:2m 0s
Bu soruyu puanla

Konular

Planning Compute Engine Resources and Machine Types

Bu konudan daha fazla soru