Examkin

Question

Difficulty: HardPlanning Compute Engine Resources and Machine Types

A media logistics company is planning its Google Cloud Compute Engine resource strategy for two distinct components of an enterprise video pipeline:

1. A memory-bound metadata indexing service that requires a specific non-standard resource ratio of 16 vCPUs to 128 GB of RAM (1:81:8 vCPU-to-memory ratio) and must run continuously 24/7 with strict zero-downtime SLAs.
2. A stateless, highly parallel batch video encoding job that processes queue messages, is fully fault-tolerant, and can automatically resume execution from checkpoints if interrupted.

Which provisioning strategy aligns with Google Cloud recommended practices to optimize costs while satisfying all operational requirements?

  1. Configure a custom N2 machine type with 16 vCPUs and 128 GB RAM coupled with Committed Use Discounts for the indexing service, and use Spot Virtual Machines in a Managed Instance Group for the batch video encoding workload.Answer
  2. B
    Provision Spot Virtual Machines for the continuous metadata indexing service to minimize hourly instance costs, and deploy memory-optimized M2 instances backed by Committed Use Discounts for the stateless batch encoding workload.
  3. C
    Deploy the metadata indexing service to Cloud Functions to achieve automatic scaling, and provision standard N2 Compute Engine instances with Committed Use Discounts for the stateless batch encoding workload.
  4. D
    Provision M-series memory-optimized instances without discounts for both workloads to simplify administrative overhead, relying on automatic Sustained Use Discounts to optimize overall compute spending.

Answer

Configure a custom N2 machine type with 16 vCPUs and 128 GB RAM coupled with Committed Use Discounts for the indexing service, and use Spot Virtual Machines in a Managed Instance Group for the batch video encoding workload.
The correct strategy combines custom machine sizing with targeted pricing models based on workload fault tolerance and steady-state patterns. Custom N2 machine types allow specifying exact vCPU and RAM ratios when standard machine types would cause paid over-provisioning. Pairing continuous 24/7 workloads with Committed Use Discounts (CUDs) secures deep contract discounts, while using Spot VMs inside a Managed Instance Group for stateless, checkpointable batch jobs yields the lowest possible compute cost for interruptible tasks.

Step-by-Step Solution

1
Analyze the workload characteristics of the metadata indexing service.
The service requires a custom memory ratio (1 vCPU to 8 GB RAM, totalling 16 vCPUs and 128 GB RAM) and runs 24/7 continuously with high availability requirements.
Standard machine types (like N2 standard with 1:4 vCPU-to-memory ratio) would force over-provisioning of vCPUs to achieve 128 GB RAM. A Custom Machine Type allows exact resource allocation, while Committed Use Discounts (CUDs) provide up to 57% savings for steady-state 24/7 usage.
2
Analyze the workload characteristics of the batch video encoding job.
The encoding job is stateless, fault-tolerant, queue-driven, and supports checkpointing.
Workloads that tolerate interruptions are prime candidates for Spot VMs, which provide significant cost savings (up to 60–91% compared to standard pricing) when managed via Managed Instance Groups (MIGs).
3
Synthesize the optimal Compute Engine architectural plan.
Combine Custom N2 Machine Types + CUDs for Workload 1, and Spot VMs + MIGs for Workload 2.
This combination minimizes waste from fixed machine predefined ratios and maximizes financial optimization across both steady-state and fault-tolerant compute components.

Key Concept

Compute Engine Resource Planning, Custom Machine Types, Committed Use Discounts (CUDs), and Spot VM Suitability
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic