Examkin

Question

Difficulty: MediumPlanning Compute Engine Resources and Machine Types

A healthcare organization is planning its Google Compute Engine architecture for two distinct workloads: a core patient search indexing engine that requires continuous uninterrupted availability and fixed memory capacity, and a nightly batch job that processes large volumes of anonymized medical images in parallel with stateless, fault-tolerant tasks. The organization wants to optimize compute infrastructure costs while maintaining strict availability requirements for critical operations. Which compute provisioning strategy should the cloud engineer recommend?

  1. Provision standard or memory-optimized Compute Engine instances with Committed Use Discounts (CUDs) for the core search indexing engine, and use Spot VMs managed by an instance group for the nightly batch processing pipeline.Answer
  2. B
    Provision Spot VMs for both the core patient search indexing engine and the nightly batch image processing pipeline to maximize cost savings across all Compute Engine resources.
  3. C
    Purchase Committed Use Discounts for Spot VMs running the batch processing pipeline, and rely on Sustained Use Discounts for the search indexing engine.
  4. D
    Deploy the core patient search indexing engine to Cloud Functions to avoid VM management, and use standard non-discounted Compute Engine instances for the batch image processing pipeline.

Answer

Provision standard or memory-optimized Compute Engine instances with Committed Use Discounts for the uninterrupted core search indexing engine, and leverage Spot VMs for the fault-tolerant nightly batch processing pipeline.
The correct strategy pairs baseline, continuous workloads with Committed Use Discounts (CUDs) to guarantee availability and reduce costs, while using Spot VMs for stateless batch processing to maximize savings on interruptible workloads.

Step-by-Step Solution

1
Analyze workload 1 requirements
The core patient search indexing engine requires 24/7 continuous uptime and predictable capacity.
Baseline stateful or critical workloads with predictable usage benefit most from Committed Use Discounts (1-year or 3-year commitments).
2
Analyze workload 2 requirements
The nightly medical image processing pipeline is stateless, highly parallelizable, and fault-tolerant.
Stateless batch processing that can withstand instance preemptions is the ideal candidate for Spot VMs, reducing compute costs by up to 60-91%.
3
Match provisioning strategies to workload characteristics
Combining Committed Use Discounts for baseline VM instances with Spot VMs for batch workloads optimizes total cost of ownership without violating availability SLAs.
Ensures high availability for critical services while taking advantage of deep discounts for interruptible workloads.

Key Concept

Selecting compute machine types and discount models (CUDs vs. Spot VMs) based on workload interruptibility and availability requirements.
Estimated Time:1m 30s
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic