Examkin

Soru

Zorluk: OrtaPlanning and Assessing Cloud Compute Solutions

A logistics company needs to run a weekly 12-hour batch processing job to re-calculate global shipping routes. The workload is containerized, fully fault-tolerant, and designed to automatically save state to Cloud Storage and resume from checkpoints if interrupted. The operations team wants to minimize compute costs while ensuring that any terminated compute capacity is automatically replaced until the job finishes. Which compute configuration should the cloud engineer recommend to meet these requirements at the lowest cost?

  1. Deploy a Compute Engine Managed Instance Group (MIG) configured with Spot VMs.Cevap
  2. B
    Deploy the job to a Cloud Functions function triggered by Cloud Scheduler to run continuously for 12 hours.
  3. C
    Provision standalone Compute Engine instances backed by a 3-year Committed Use Discount (CUD).
  4. D
    Provision On-Demand instances in a GKE cluster with static node sizing to prevent instance termination.

Cevap

Deploying a Compute Engine Managed Instance Group (MIG) configured with Spot VMs is the optimal solution because Spot VMs offer up to 60-91% discounts for fault-tolerant batch workloads, and the Managed Instance Group automatically maintains capacity by replacing preempted instances.
The option recommending a Compute Engine Managed Instance Group (MIG) configured with Spot VMs is correct because Spot VMs offer extreme cost savings for workloads that tolerate preemption. Since the batch application saves state to Cloud Storage and can resume from checkpoints, instance preemption will not cause data loss. The MIG ensures high availability by automatically recreating instances when GCP reclaims Spot capacity.

Adım Adım Çözüm

1
Analyze workload characteristics and constraints.
The job runs intermittently (12 hours per week), is stateless/checkpoint-capable, and tolerates interruptions.
Fault-tolerant batch workloads can leverage discounted preemptible capacity.
2
Evaluate pricing models for intermittent fault-tolerant workloads.
Spot VMs provide maximum cost savings for preemptible workloads, whereas Committed Use Discounts require continuous steady-state usage.
Selecting Spot VMs yields significant cost reductions over standard or committed rates.
3
Determine the management mechanism for replacement capacity.
A Managed Instance Group (MIG) maintains target size and automatically creates replacement instances when preemption occurs.
Automating instance replacement fulfills the operational goal without manual intervention.

Anahtar Kavram

Selecting cost-effective compute solutions using Spot VMs and Managed Instance Groups for fault-tolerant batch workloads.
Bu soruyu puanla

Konular

Planning and Assessing Cloud Compute Solutions

Bu konudan daha fazla soru