Examkin

Soru

Zorluk: ZorSelecting and Designing Compute and Application Platforms

A digital gaming enterprise is designing the backend architecture on Google Cloud for a new multiplayer title. The backend consists of two main services:

1. Matchmaking & User Profile Service: A stateless containerized HTTP/2 REST API with highly variable traffic pattern that needs to scale instantly from zero to thousands of concurrent requests while minimizing operational overhead.
2. Real-time Game State Engine: A custom C++ stateful server application that requires persistent low-latency raw TCP sockets, direct access to host network interfaces, and specific Linux kernel tuning parameters (`sysctl` network stack modifications).

Which compute platform combination should the cloud architect select to satisfy these requirements while adhering to Google Cloud best practices?

  1. Deploy the Matchmaking & User Profile Service on Cloud Run, and deploy the Real-time Game State Engine on Compute Engine Managed Instance Groups (MIGs).Cevap
  2. B
    Deploy both the Matchmaking & User Profile Service and the Real-time Game State Engine on Google Kubernetes Engine (GKE) Autopilot clusters using custom DaemonSets for kernel modifications.
  3. C
    Deploy the Matchmaking & User Profile Service on Cloud Run, and deploy the Real-time Game State Engine as Cloud Run Jobs.
  4. D
    Deploy the Matchmaking & User Profile Service on Compute Engine VMs with CPU-based autoscaling, and deploy the Real-time Game State Engine on App Engine Standard environment.

Cevap

Deploy the Matchmaking & User Profile Service on Cloud Run, and deploy the Real-time Game State Engine on Compute Engine Managed Instance Groups (MIGs).
The solution selecting Cloud Run for the HTTP microservice and Compute Engine MIGs for the stateful engine correctly aligns platform capabilities with workload constraints. Cloud Run provides zero-to-hero serverless autoscaling for stateless REST APIs with zero infrastructure maintenance. Compute Engine provides root OS access required to modify `sysctl` Linux kernel parameters and handle persistent stateful TCP sockets.

Adım Adım Çözüm

1
Analyze workload requirements for Service 1 (Matchmaking & User Profile API).
The service is stateless, HTTP-based, requires rapid auto-scaling down to zero, and demands minimal operational management.
Cloud Run is Google Cloud's fully managed serverless platform specifically designed for stateless HTTP workloads, offering fast scaling to zero without node administration.
2
Analyze workload requirements for Service 2 (Real-time Game State Engine).
The application is stateful, uses non-HTTP raw TCP sockets, and requires host kernel configuration (`sysctl` network stack tuning).
Serverless platforms (Cloud Run/App Engine) do not permit arbitrary kernel-level tuning or persistent raw TCP connection management. Compute Engine (VMs / MIGs) provides complete OS and kernel-level control.
3
Select the platform combination balancing operational overhead and technical capabilities.
Pair Cloud Run for the stateless API with Compute Engine MIGs for the kernel-tuned stateful engine.
This combination minimizes management effort for standard web services while providing maximum OS flexibility where strictly required.

Anahtar Kavram

Selecting optimal GCP compute platforms based on workload statefulness, protocol requirements, kernel customization needs, and operational overhead limits.
Tahmini Süre:2m 0s
Bu soruyu puanla

Konular

Selecting and Designing Compute and Application Platforms