Examkin

Question

Difficulty: Very hardPlanning Compute Engine Resources and Machine Types

An enterprise financial technology organization is designing the Compute Engine infrastructure for a compliance auditing system that runs two distinct workload components:

1. Workload 1: An overnight audit reconciliation process that is stateless, fully fault-tolerant, saves state progress to Cloud Storage every 10 minutes, and requires maximum cost optimization.
2. Workload 2: A core continuous transaction validation service that runs uninterrupted 24 hours a day, 365 days a year, cannot tolerate unexpected instance terminations, and requires a specific memory-to-vCPU ratio of 29 GB memory and 4 vCPUs to comply with third-party software license boundaries.

Which TWO Compute Engine resource planning and pricing decisions should you implement to satisfy both technical constraints and cost efficiency? (Select TWO choices)

  1. Provision Spot Virtual Machines (VMs) for Workload 1 to minimize operational cost for checkpointed, fault-tolerant batch processing.Answer
  2. Configure a Custom Machine Type with 4 vCPUs and 29 GB memory backed by a Committed Use Discount (CUD) for Workload 2.Answer
  3. C
    Provision Spot Virtual Machines (VMs) for Workload 2 to lower the continuous execution cost of the transaction validation service.
  4. D
    Rely exclusively on automatic Sustained Use Discounts (SUDs) instead of Committed Use Discounts (CUDs) for Workload 2 to achieve the maximum possible baseline discount.

Answer

Deploy Workload 1 using Spot Virtual Machines (VMs) and configure Workload 2 using a Custom Machine Type (4 vCPUs, 29 GB RAM) combined with a Committed Use Discount (CUD).
Spot VMs provide deep cost discounts for stateless, checkpointed batch applications that can gracefully handle instance preemption. For continuous 24/7 workloads with non-standard resource constraints, Custom Machine Types allow exact allocation of vCPUs and memory to satisfy licensing limits, and Committed Use Discounts (CUDs) maximize cost reduction for predictable, long-term operational commitments.

Step-by-Step Solution

1
Analyze Workload 1 requirements (fault-tolerant, batch, checkpointed, cost-sensitive).
Identify that Spot VMs offer maximum discount for interruptible, fault-tolerant workloads.
Because progress is checkpointed every 10 minutes, preemption does not cause loss of work, making Spot VMs ideal.
2
Analyze Workload 2 resource constraints (non-standard 4 vCPUs / 29 GB RAM ratio).
Select a Custom Machine Type to match exact licensing boundaries without paying for unused predefined resources.
Predefined machine types (e.g., n2-standard-4 with 16 GB or n2-highmem-4 with 32 GB) either under-provision RAM or force over-payment for unnecessary capacity.
3
Analyze Workload 2 availability and billing requirements (24/7 continuous runtime, no unexpected terminations allowed).
Pair the Custom Machine Type with a 1-year or 3-year Committed Use Discount (CUD).
CUDs deliver substantial price reductions for predictable, steady-state enterprise capacity without risking preemption.

Key Concept

Selecting Compute Engine provisioning models (Spot vs. On-Demand/CUD) and machine specifications (Custom vs. Predefined) based on fault tolerance, non-standard memory-to-vCPU ratios, and runtime predictability.
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic