An online gaming enterprise is designing the infrastructure for a global multiplayer matchmaking service on Google Cloud. The frontend microservices are lightweight and stateless, requiring dynamic scaling to handle sudden traffic spikes with minimal operational overhead. The data tier manages active player session states and transactional matchmaking queues, requiring strong global consistency and high availability across multiple regions to ensure zero data loss during regional outages. Which architecture best satisfies these technical and availability requirements while minimizing operational complexity?
- Deploy the stateless microservices on Cloud Run across multiple regions, and use a multi-region Cloud Spanner instance for the transactional player session data tier.Answer
- BDeploy the stateless microservices on a single zonal Google Kubernetes Engine (GKE) cluster, and use a multi-region Cloud Spanner instance for the transactional player session data tier.
- CDeploy the stateless microservices on Cloud Run across multiple regions, and use Cloud SQL for PostgreSQL with cross-region read replicas for the transactional player session data tier.
- DDeploy the stateless microservices on Cloud Run across multiple regions, and rely on a primary Cloud SQL instance backed up continuously to Cloud Storage for disaster recovery.
Answer
Deploy the stateless microservices on Cloud Run across multiple regions, and use a multi-region Cloud Spanner instance for the transactional player session data tier.
Combining multi-region Cloud Run deployments with a multi-region Cloud Spanner database fulfills all constraints. Cloud Run automatically handles bursty traffic for stateless HTTP services across regions without managing cluster infrastructure. Cloud Spanner delivers globally distributed, strongly consistent SQL transactions with synchronous multi-region replication, eliminating regional single points of failure and meeting high availability targets.
Step-by-Step Solution
Key Concept
Designing High Availability Cloud Architectures using Serverless Compute and Multi-Region Spanner
Estimated Time:1m 30s