Soru

Zorluk: ZorCreating Conceptual, Logical, and Physical Architectures

A global online gaming company is designing a multi-tier leaderboard and matchmaking architecture on Google Cloud. The conceptual design requires high-concurrency event ingestion, low-latency stateful matchmaking logic, and ACID-compliant transactional persistence for player inventory management across regions. The enterprise architecture team is translating these conceptual requirements into a logical tier structure before finalizing physical GCP resource allocations. Which architecture design correctly maps the conceptual tiers to logical component boundaries and concrete physical GCP services while adhering to Google Cloud best practices?

  1. A
    Map event ingestion to Cloud Pub/Sub, stateful matchmaking to Cloud Run stateless containers backed by Cloud Memoryrange, and transaction persistence to Cloud SQL for PostgreSQL configured with cross-region read replicas.
  2. Map event ingestion to Cloud Pub/Sub, stateful matchmaking logic to GKE clusters with StatefulSets, and inventory transaction persistence to Cloud Spanner across a multi-region instance configuration.Cevap
  3. C
    Map event ingestion to HA VPN endpoints, stateful matchmaking logic to Cloud Functions, and inventory transaction persistence to Cloud Bigtable with multi-cluster replication.
  4. D
    Map event ingestion to Cloud Dataflow, stateful matchmaking to a Compute Engine Managed Instance Group with CPU-based autoscaling, and inventory transaction persistence to Cloud Storage dual-region buckets.

Cevap

The correct architecture maps event ingestion conceptually to Cloud Pub/Sub, stateful session handling logically to Google Kubernetes Engine (GKE) StatefulSets, and global multi-region transactional persistence physically to Cloud Spanner.
The solution correctly translates high-level conceptual capabilities (scalable ingestion, stateful compute, global transactional database) into logical service boundaries and selects the optimal physical GCP products (Cloud Pub/Sub, GKE StatefulSets, Cloud Spanner). Cloud Spanner provides the global ACID transactions required for multi-region inventory consistency, while GKE StatefulSets support stateful matchmaking sessions.

Adım Adım Çözüm

1
Analyze conceptual requirements for data persistence.
Identified the need for globally distributed, multi-region ACID transactional writes for player inventory.
Standard relational databases like Cloud SQL offer read scalability via replicas but cannot accept multi-region synchronous ACID writes. Cloud Spanner is required.
2
Evaluate compute execution model requirements.
Selected GKE StatefulSets for complex, stateful matchmaking session logic.
Matchmaking requires persistent in-memory state and long-lived peer-to-peer session management, which serverless platforms like Cloud Run or Cloud Functions cannot natively maintain.
3
Verify event ingestion component selection.
Selected Cloud Pub/Sub as the asynchronous streaming buffer.
Cloud Pub/Sub decouples real-time telemetry and matchmaking request ingestion at massive global scale.

Anahtar Kavram

Mapping Architectural Abstraction Layers (Conceptual to Logical to Physical)
Bu soruyu puanla