Examkin

Soru

Zorluk: ZorSelecting and Designing Compute and Application Platforms

An enterprise retail corporation is modernizing its global order processing ecosystem on Google Cloud. The architecture consists of two core workloads with distinct operational parameters:

1. A stateless HTTP REST microservice that processes user checkout requests. Traffic is extremely bursty, experiencing 50x spikes during seasonal promotions and dropping to zero overnight. The organization mandates minimal operational overhead and zero infrastructure maintenance.
2. A legacy stateful inventory synchronization service that requires custom Linux kernel socket parameter tuning (`sysctl`), direct access to attached persistent disk volumes for local file locking, and fixed internal IP addresses.

Which compute platform architecture should you recommend to meet the requirements of both workloads while minimizing operational complexity?

  1. Deploy the stateless HTTP REST microservice on Cloud Run, and deploy the legacy stateful inventory synchronization service on Compute Engine using stateful Managed Instance Groups (MIGs).Cevap
  2. B
    Deploy both the stateless HTTP REST microservice and the legacy stateful inventory synchronization service as workloads on Google Kubernetes Engine (GKE) Autopilot.
  3. C
    Deploy the stateless HTTP REST microservice on Cloud Functions, and deploy the legacy stateful inventory synchronization service on Cloud Run with mounted Cloud Storage buckets.
  4. D
    Deploy both workloads on unmanaged Compute Engine Virtual Machine instances configured with CPU-based autoscaling policies across multiple zones.

Cevap

Deploy the stateless HTTP REST microservice on Cloud Run, and deploy the legacy stateful inventory synchronization service on Compute Engine using stateful Managed Instance Groups (MIGs).
Deploying the stateless HTTP REST microservice on Cloud Run fulfills the requirement for zero server management overhead while scaling automatically from zero to handle 50x traffic spikes. Deploying the legacy inventory synchronization service on Compute Engine using stateful Managed Instance Groups satisfies all technical constraints of the legacy app, including custom kernel network tuning (`sysctl`), direct POSIX file locking on attached persistent disks, and fixed internal IP address preservation across VM restarts.

Adım Adım Çözüm

1
Analyze the requirements for Workload 1 (Stateless REST API)
Identified requirements: HTTP protocol, bursty traffic with scale-to-zero capability, and minimal operational overhead. Cloud Run is the optimal Google Cloud serverless container platform for stateless HTTP services.
Cloud Run automatically scales instances up and down based on traffic (including scale-to-zero) without requiring VM or cluster lifecycle management.
2
Analyze the requirements for Workload 2 (Legacy Stateful Inventory Daemon)
Identified constraints: Custom kernel tuning (`sysctl`), POSIX persistent disk file locking, and fixed IP addresses. Compute Engine provides full OS and kernel level control.
Fully managed serverless platforms (Cloud Run, Cloud Functions) isolate container environments and do not permit low-level OS/kernel parameter modifications or stateful static IP persistence.
3
Select the optimal compute combination
Combine Cloud Run for the stateless web component and Compute Engine stateful MIGs for the legacy stateful daemon.
This workload segregation ensures each component runs on the compute abstraction best suited to its operational constraints while keeping total operational overhead as low as possible.

Anahtar Kavram

Selecting GCP compute platforms based on workload statefulness, operational overhead limitations, and kernel/OS modification constraints
Bu soruyu puanla

Konular

Selecting and Designing Compute and Application Platforms