GCP Professional Data Engineer Practice Question

Your company operates e-commerce sites on three continents. Every customer order must be written to the operational database in under 50 ms and be immediately visible in all regions with 99.999 % availability. Peak traffic reaches 2 million writes per second. Business analysts need to query fresh order data in BigQuery within five minutes of commit without adding complex operations. Which Google Cloud design best satisfies these requirements?

Write orders to Cloud Bigtable and export SSTables nightly to Cloud Storage, then import them into BigQuery with scheduled queries.
Store orders in a multi-region Cloud Spanner instance and use Cloud Spanner change streams processed by Dataflow to write updates to BigQuery in near real time.
Store orders in Cloud SQL for PostgreSQL with cross-region read replicas and schedule hourly BigQuery Data Transfer Service loads from the primary instance.
Stream orders directly into a multi-regional BigQuery table and let analysts query the same table for operational and analytical workloads.

Report Issue

Answer Description

A multi-region Cloud Spanner instance is purpose-built for globally distributed transactional workloads that demand horizontal write scaling, five-nines availability, and synchronous, strongly consistent replication across regions-all of which match the order-processing requirements. Cloud Spanner change streams can capture mutations as they occur and feed them into a Dataflow pipeline that writes to BigQuery, making new records queryable within minutes while keeping administration minimal.

Cloud SQL cannot scale to millions of writes per second and its cross-region replication is asynchronous. Writing orders directly to BigQuery provides a serverless destination but lacks sub-50 ms transactional inserts and strong consistency, and streaming data can take minutes-or longer-to become visible. Bigtable offers high throughput but does not provide multiregion strong consistency, ACID transactions, or SQL support. Therefore, using Cloud Spanner with change streams to Dataflow and BigQuery is the only option that meets all performance, consistency, and operational goals.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.