Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
An e-commerce company is replacing its monolithic relational database after flash sales caused severe lock contention and latency spikes. The new global architecture must support three distinct workloads:
Which database architecture should the company implement to address these trade-offs?
Cloud SQL is a fully managed relational database service for MySQL, PostgreSQL, and SQL Server. It excels at traditional single-region OLTP workloads that depend on normalized schemas, foreign keys, and ACID transactions.
The application were a regional workload with moderate write throughput, modest data volume fitting on a single instance, and no requirement for global active-active scaling or petabyte-scale analytical processing.
This architecture pairs purpose-built relational, non-relational, and analytical Google Cloud services so each workload uses the optimal consistency and scaling model. Datastore handles high-scale BASE (Basically Available, Soft state, Eventual consistency) operations, Spanner provides globally distributed ACID transactions, and BigQuery executes petabyte-scale analytical queries.
Forcing all workloads into a single database forces painful compromises between strict locking, horizontal scale, and analytical flexibility. Combining Datastore, Spanner, and BigQuery aligns each business requirement with the right balance of transactional integrity, eventual consistency, and query power.
Cloud Bigtable is a petabyte-scale, wide-column NoSQL database designed for high-throughput, low-latency operational and time-series workloads. With multi-cluster routing, every cluster acts as a primary that accepts reads and writes and replicates asynchronously to other clusters.
The entire workload consisted solely of massive-scale key-value lookups, IoT telemetry, or append-only event ingestion where multi-row ACID transactions and complex SQL joins were unnecessary.
Spanner is a globally distributed, horizontally scalable relational database that provides external consistency via TrueTime and Paxos. Under its default serializable isolation, read-write transactions use pessimistic concurrency control by acquiring shared read locks and upgrading them to exclusive write locks at commit time.
Every operation in the system strictly required global ACID serializability with zero tolerance for stale reads and no heavy ad-hoc analytical reporting workload.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated