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.
A company runs a Cloud Spanner database in a single US region alongside its write-heavy order-processing services. Expanding traffic in Europe and Asia has caused high read latency for users browsing the product catalog.
The team must update the architecture to satisfy these requirements:
Which scaling and deployment strategy should the team implement?
A custom regional Spanner configuration builds upon a standard single-region deployment—which maintains three read-write voting replicas across zones in one region—by adding optional read-only replicas in distant geographic regions. These read-only replicas asynchronously receive replicated data to scale read traffic locally without participating in transactional write quorums.
Unlike a full three-continent multi-region topology, a custom regional configuration with read-only replicas balances cost, write performance, and global read latency. It delivers localized read scaling where eventual staleness is acceptable while keeping synchronous write consensus strictly within the US region.
Running a self-managed database on Compute Engine VMs across three continents with asynchronous bidirectional replication allows each region to accept reads and writes locally while asynchronously propagating changes to the other regions.
The workload required a specialized third-party engine unavailable as a managed service, could tolerate eventual consistency for all writes, and prioritized local write latency across every continent over strong transactional consistency.
Replicating self-managed databases across regions is error-prone and expensive to maintain. When strong consistency is required for writes and low-latency stale reads are only needed for specific queries, a managed distributed database provides superior reliability at lower operational cost.
A dual-region Spanner configuration replicates data synchronously across two regions within a single country with a witness replica for automated failover, raising availability to 99.999%. Adding compute nodes scales the instance's processing throughput (queries per second) horizontally within those US regions.
Users were concentrated exclusively in a single country, the company required 99.999% availability with strict data residency compliance, and the primary bottleneck was query throughput rather than intercontinental latency.
Architects often confuse horizontal compute scaling (adding nodes for higher QPS) with geographic read scaling (placing read-only replicas near distant user bases). Only geographic placement reduces propagation delay.
A three-continent multi-region configuration distributes read-write voting replicas and witness replicas across multiple US regions alongside read-only replicas in Europe and Asia, offering up to 99.999% availability. Configuring queries to use strong reads instructs Spanner to guarantee that every read reflects all transactions committed prior to the start of the read.
The application required 99.999% multi-region availability and resilience against an entire regional outage, and catalog queries in Europe and Asia were explicitly configured to use stale reads instead of strong reads.
Using strong reads from distant continents negates the latency benefit of local read replicas because Spanner must ensure global linearizability by consulting the leader region.
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