Examkin

Question

Difficulty: Very hardPlanning and Assessing Cloud Compute Solutions

Your organization plans to migrate a legacy high-throughput network monitoring containerized application to Google Cloud. The application requires custom Linux sysctl kernel tuning on the host nodes to optimize network buffer sizes and low-level packet processing. Additionally, the workload requires predictable performance without unexpected host terminations, while utilizing Kubernetes container orchestration for deployment and auto-healing. Which Google Cloud compute strategy should you recommend to meet these requirements?

  1. Deploy the application on Google Kubernetes Engine (GKE) Standard using node pools configured with custom node system configurations.Answer
  2. B
    Deploy the application on Google Kubernetes Engine (GKE) Autopilot to minimize node management overhead.
  3. C
    Deploy the application container directly to Cloud Run to eliminate cluster infrastructure management entirely.
  4. D
    Deploy the application on a GKE Standard cluster provisioned entirely with Spot VM instances to minimize operational costs.

Answer

Deploy the application on Google Kubernetes Engine (GKE) Standard using node pools configured with custom node system configurations.
Google Kubernetes Engine (GKE) Standard allows system administrators to supply custom node system configurations (including sysctl parameters) to node pools while supporting standard, non-preemptible Compute Engine instances for predictable uptime.

Step-by-Step Solution

1
Analyze workload constraints and technical requirements.
Identified requirements: Container orchestration, custom host OS kernel tuning (sysctl), and predictable availability without unexpected interruptions.
Evaluating infrastructure flexibility versus management overhead narrows down valid compute options.
2
Evaluate serverless and fully managed container options (Cloud Run and GKE Autopilot).
Cloud Run and GKE Autopilot abstract node management and enforce strict security boundaries that prohibit custom host kernel modifications.
Fully managed environments lock down the host kernel for security and multi-tenancy stability.
3
Evaluate node provisioning options for GKE Standard (Spot VMs vs standard Compute Engine instances).
GKE Standard supports custom node system configurations for sysctl tuning. Standard instances guarantee non-preemptible execution, whereas Spot VMs risk unexpected terminations.
Spot VMs are designed for fault-tolerant batch workloads and violate predictable availability requirements.

Key Concept

Planning Compute Engine and GKE infrastructure based on kernel customization and fault-tolerance trade-offs.
Rate this question

Topics

Planning and Assessing Cloud Compute Solutions

More questions from this topic