Examkin

Question

Difficulty: MediumPlanning Compute Engine Resources and Machine Types

A logistics enterprise is planning to deploy an enterprise resource planning (ERP) database on Google Cloud Compute Engine. The workload requires a constant 16 vCPUs and 128 GB of RAM to run continuously 24 hours a day, 7 days a week, with strict high availability SLA requirements. The finance team requires maximum cost optimization for this steady-state workload over a 3-year horizon. Which compute configuration and purchasing strategy should you recommend?

  1. Provision a standard Compute Engine instance matching the vCPU and RAM requirements, and purchase a 3-year Committed Use Discount for the workload.Answer
  2. B
    Provision Spot VMs with custom machine types (16 vCPUs, 128 GB RAM) to minimize compute costs for the continuous database workload.
  3. C
    Deploy the database workload using Cloud Run container instances with cpu-always-allocated enabled to eliminate VM management overhead.
  4. D
    Provision standard Compute Engine instances and rely solely on automatic Sustained Use Discounts to achieve maximum cost reduction.

Answer

Provision a standard Compute Engine instance matching the vCPU and RAM requirements, and purchase a 3-year Committed Use Discount for the workload.
The correct response combines standard Compute Engine virtual machines with a 3-year Committed Use Discount. Persistent enterprise databases with strict SLA requirements require reliable non-preemptible VMs. When resource consumption is predictable and continuous over a multi-year period, purchasing a Committed Use Discount yields significantly higher savings than on-demand rates or automatic Sustained Use Discounts.

Step-by-Step Solution

1
Evaluate workload availability and persistence requirements
The ERP database requires 24/7 uptime and high availability with a strict SLA, ruling out preemptible/Spot instances.
Spot VMs can be stopped at any time by GCP, which violates high-availability stateful database SLAs.
2
Determine appropriate compute architecture (VM vs Serverless)
Select Compute Engine Virtual Machines over serverless platforms.
Enterprise relational databases require stateful disk storage and dedicated compute resources, which fit Compute Engine better than Cloud Run.
3
Select optimal pricing model for 3-year steady-state baseline demand
Choose 3-year Committed Use Discounts (CUDs) over Sustained Use Discounts (SUDs).
CUDs offer the maximum percentage discount for predictable 24/7 workloads over a multi-year timeframe.

Key Concept

Planning Compute Engine Workloads with Committed Use Discounts vs Spot VMs
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic