Examkin

Question

Difficulty: MediumPlanning Compute Engine Resources and Machine Types

A cloud engineer is designing Google Cloud compute infrastructure for a healthcare organization with two distinct application requirements:

1. Workload 1: An asynchronous, fault-tolerant batch image processing pipeline that can resume cleanly if an instance is terminated unexpectedly.
2. Workload 2: A 24/7 mission-critical relational database with high RAM requirements that demands uninterrupted execution and high availability.

Which TWO deployment and machine type strategies should the engineer select to meet performance requirements while minimizing costs? (Select TWO.)

  1. Provision Spot VMs for Workload 1 to significantly reduce compute expenses while accommodating instance preemptions.Answer
  2. Provision memory-optimized standard VMs with Committed Use Discounts (CUDs) for Workload 2.Answer
  3. C
    Provision Spot VMs for Workload 2 to lower the cost of memory-intensive database compute instances.
  4. D
    Deploy Workload 1 on Cloud Functions to run long-running batch jobs requiring custom third-party binary libraries and multi-hour execution windows.

Answer

The architect should select Spot VMs for the fault-tolerant batch processing workload and memory-optimized standard VM instances with Committed Use Discounts for the 24/7 mission-critical database.
Spot VMs provide significant cost reductions for fault-tolerant, asynchronous batch jobs that tolerate preemption. For steady-state 24/7 high-memory database workloads, memory-optimized standard VMs secured with Committed Use Discounts (CUDs) provide guaranteed uptime alongside maximum cost optimization.

Step-by-Step Solution

1
Analyze Workload 1 operational characteristics
Identified as fault-tolerant, stateless, and asynchronous batch processing.
Fault-tolerant batch jobs can withstand abrupt node terminations without data loss, making them perfect candidates for deep-discounted Spot VMs.
2
Analyze Workload 2 operational characteristics
Identified as a continuous, 24/7 mission-critical high-memory database.
Stateful databases require predictable uptime and continuous memory capacity. They should use standard (non-preemptible) memory-optimized instances coupled with Committed Use Discounts (CUDs) for predictable 24/7 workloads.
3
Evaluate and eliminate non-viable compute options
Eliminated Spot VMs for databases (risk of unannounced outage) and Cloud Functions for heavy batch jobs (timeout and binary limits).
Matching workload SLA requirements with appropriate compute families prevents service outages and compute framework mismatches.

Key Concept

Planning Compute Engine machine types, Spot VM suitability, and discount strategies based on workload SLA and resource profile
Rate this question

Topics

Planning Compute Engine Resources and Machine Types

More questions from this topic