Examkin

Question

Difficulty: HardSelecting and Designing Compute and Application Platforms

A global supply chain organization is architecting a new containerized route-optimization service on Google Cloud. The service receives stateless HTTP requests from mobile dispatch devices, executes short-lived algorithms requiring less than 15 seconds per request, experiences dramatic traffic fluctuations ranging from 0 to over 50,000 requests per minute during peak operational hours, and must scale down to zero during inactive periods to minimize costs. The organization enforces strict security policies prohibiting unmanaged server infrastructure, and the devops team aims to eliminate Kubernetes control plane management overhead while ensuring minimal cost for idle resources. Which compute platform design should you recommend?

  1. Deploy the containerized service to Cloud Run, configuring container concurrency and scaling limits to automatically scale instances down to zero when idle.Answer
  2. B
    Provision a Google Kubernetes Engine (GKE) Standard regional cluster with Cluster Autoscaler and Horizontal Pod Autoscaler (HPA) enabled to manage the workload deployment.
  3. C
    Deploy the application onto a Compute Engine Managed Instance Group (MIG) behind an External HTTP(S) Load Balancer, configured to auto-scale strictly based on target CPU utilization.
  4. D
    Purchase 3-year Committed Use Discounts (CUDs) for fixed-capacity Compute Engine virtual machine instances to host the container runtime environments continuously.

Answer

Deploy the containerized service to Cloud Run with automatic scaling down to zero when idle.
Cloud Run is the optimal Google Cloud platform for containerized, stateless HTTP workloads with bursty request patterns. It abstracts away all server infrastructure and cluster management while providing seamless automatic scaling from zero to tens of thousands of requests per minute, charging strictly for resources consumed during request execution.

Step-by-Step Solution

1
Analyze workload characteristics and constraints
The application is stateless, containerized, handles short-lived HTTP requests (<15 seconds), experiences extreme traffic bursts (0 to 50,000 requests/min), and requires scaling to zero.
Identifying statefulness, protocol, latency, and scaling requirements dictates the optimal compute platform selection.
2
Evaluate Google Cloud compute platform options against operational overhead limits
Cloud Run handles arbitrary containers via stateless HTTP, provides automated scaling to zero, charges per millisecond of execution, and requires zero control plane management.
Eliminating server and cluster management aligns directly with the requirement to minimize DevOps operational overhead.
3
Compare against Kubernetes and VM-based alternatives
GKE Standard introduces master node management overhead and static node costs. Compute Engine MIGs cannot scale to true zero efficiently and incur baseline OS maintenance.
Selecting GKE or Compute Engine MIGs represents compute misallocation for stateless bursty microservices.

Key Concept

Selecting Serverless Container Compute Platforms based on Workload Characteristics
Rate this question

Topics

Selecting and Designing Compute and Application Platforms