Question

A retail company is designing a new customer-facing web application on Google Cloud. The business requires that 99% of page loads complete within 200 ms for users across North America and Europe. The application consists of a stateless front-end tier, a REST API tier, and a relational data store. Static assets change infrequently but are requested heavily. The architecture team must design the compute and delivery layers to meet the latency success metric while keeping operational overhead low. Which design best satisfies these business and technical requirements?

Accepted Answer

Serve static assets through Cloud CDN backed by a global external HTTP(S) load balancer, and run the front-end and API on autoscaling managed instance groups in multiple regions closest to users.

Answer

Deploy the front-end and API on regional managed instance groups in a single US region, and rely on Cloud SQL read replicas in Europe to reduce latency.

Answer

Deploy the entire stack, including static assets, on a single Compute Engine VM per region and use DNS round-robin to distribute traffic between regions.

Answer

Host all tiers in one region and enable Premium Tier networking so that Google's global backbone alone brings latency under 200 ms for all users.