GCP Professional Data Engineer Practice Question

Your enterprise is adopting a data-mesh strategy on Google Cloud. Each business domain must own its analytical storage (BigQuery datasets and Cloud Storage buckets), while a central platform team must enforce consistent policies for metadata, data quality, and row-level security without impeding domain autonomy. Which architecture best meets these requirements?

Create one Google Cloud project per domain where teams manage their own BigQuery datasets and Cloud Storage buckets, then register all assets in a centrally managed Dataplex lake that applies attribute-based access controls and data-quality rules across those projects.
Store all domain data in a single BigQuery project owned by the platform team, restrict access with authorized views, and catalog datasets with Dataproc Metastore.
Move each domain's data to Firestore collections replicated globally, protect access with VPC Service Controls, and share metadata through spreadsheets maintained by the platform team.
Deploy one multi-region Bigtable instance shared by every domain, manage table access using service account keys, and rely on Cloud Logging entries for metadata discovery.

Report Issue

Answer Description

Creating a separate Google Cloud project for every business domain gives each team full ownership of its storage resources, satisfying a core data-mesh principle of decentralized data ownership. Registering those cross-project assets in a single Dataplex lake lets the platform team apply attribute-based access controls, data quality checks, and a unified metadata catalog from one control plane. Dataplex policies are enforced in place, so data remains in each domain's project, preserving autonomy while centralizing governance. The alternative designs either centralize all data in one project (violating domain ownership) or rely on services that lack Dataplex's federated governance and fine-grained security features, so they cannot deliver both requirements simultaneously.

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.