Examkin

Soru

Zorluk: OrtaSelecting and Designing Compute and Application Platforms

A financial technology company is designing a new transaction validation pipeline on Google Cloud consisting of two distinct workloads:

1. A stateless HTTP webhook ingestion component that receives payload spikes ranging from 0 to 5,000 requests per second with execution times under 10 seconds per request, requiring minimal operational management.
2. A specialized legacy fraud-analysis service that depends on a custom compiled Linux kernel module for cryptographic hardware acceleration and maintains persistent, long-lived TCP socket connections.

Which TWO compute platform choices align with Google Cloud best practices for operational efficiency and workload compatibility? (Select TWO)

  1. Deploy the HTTP webhook ingestion component to Cloud Run to leverage automatic container scaling to zero and fully managed infrastructure.Cevap
  2. Deploy the legacy fraud-analysis service to Compute Engine Virtual Machines (or Managed Instance Groups) using a custom OS image.Cevap
  3. C
    Deploy the HTTP webhook ingestion component to a Google Kubernetes Engine (GKE) Standard cluster with fixed node pools.
  4. D
    Deploy the legacy fraud-analysis service as a Cloud Run Job scheduled to execute continuously.
  5. E
    Deploy the HTTP webhook ingestion component to App Engine Flexible Environment configured with auto-scaling.

Cevap

The optimal architecture combines Cloud Run for the stateless HTTP webhook ingestion component and Compute Engine Virtual Machines (with custom OS images) for the legacy fraud-analysis service requiring custom kernel modules.
Cloud Run is the optimal serverless container platform for stateless HTTP services with unpredictable traffic spikes, scaling rapidly to zero with zero cluster management. Compute Engine is required for workloads that depend on custom host Linux kernel modules and persistent network daemons, as virtual machines provide total control over the guest OS environment.

Adım Adım Çözüm

1
Analyze requirements for the HTTP webhook ingestion service
The service is stateless, HTTP-based, handles bursty traffic (0 to 5,000 rps), and requires low operational overhead.
Cloud Run is designed specifically for stateless containerized HTTP workloads, scaling seamlessly to zero and managing infrastructure automatically.
2
Analyze requirements for the legacy fraud-analysis service
The service requires a custom Linux kernel module and long-lived TCP socket connections.
Fully managed container services (like Cloud Run or App Engine) do not permit host kernel modifications. Compute Engine allows custom OS images and full root/kernel configuration.

Anahtar Kavram

Matching Google Cloud compute platforms (Cloud Run vs. Compute Engine vs. GKE) to workload statefulness, protocol, kernel requirements, and operational overhead targets.
Bu soruyu puanla

Konular

Selecting and Designing Compute and Application Platforms