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.
Your organization is migrating an 8 TB SQL Server online analytical processing (OLAP) data warehouse to Compute Engine. Baseline metrics from the current environment show the following requirements:
tempdbWhich Compute Engine machine family and storage tier combination should you select to achieve performance parity?
The compute-optimized machine family (C3 or C4 series) prioritizes highest performance per CPU core, while SSD Persistent Disk (pd-ssd) provides traditional network-attached SSD block storage and Hyperdisk Extreme delivers ultra-high configurable IOPS.
pd-ssd, throughput and IOPS scale linearly with disk capacity and vCPU count, forcing you to overprovision storage beyond 8 TB to guarantee 1,500 MB/s of sustained throughput.C3 and C4 machines are modern series designed for Hyperdisk rather than Persistent Disk, and using Hyperdisk Extreme for ephemeral tempdb incurs unnecessary cost for durability and extreme IOPS that Local SSD handles faster and cheaper.C3 and C4 VMs have lower memory-to-core ratios than M3 and M4 series.pd-ssd scaling: Achieving high sequential bandwidth on Persistent Disk requires inflating volume size or vCPU count.The workload were a CPU-bound online transaction processing (OLTP) database executing thousands of small, concurrent transactions per second that required maximum durable IOPS on data and log volumes rather than large in-memory OLAP caching.
The general-purpose E2 machine series provides cost-effective compute with a balanced CPU-to-memory profile, while Balanced Persistent Disk (pd-balanced) offers economical SSD-backed storage and Regional Persistent Disk synchronously replicates block data across two zones in the same region.
E2 instances lack the high memory-to-vCPU ratios required to hold large analytical tables and execution plans in memory, leaving the buffer pool bottleneck unresolved.pd-balanced ties throughput limits to disk size and VM vCPU count and has lower maximum performance ceilings than Hyperdisk, making 1,500 MB/s difficult or costly to achieve.tempdb: Regional Persistent Disk performs synchronous cross-zone replication on every write, adding network write latency and cost to temporary tempdb operations that need zero durability.E2 and pd-balanced are optimized for low cost rather than high-throughput data warehousing.pd-balanced delivers lower bandwidth per gigabyte than pd-ssd or Hyperdisk Balanced.tempdb.You were deploying a development, testing, or staging SQL Server environment where minimizing monthly infrastructure costs takes priority over query throughput and in-memory caching performance.
This option correctly identifies memory-optimized (M3 or M4) VMs for in-memory caching, but pairs them with Hyperdisk Extreme for data volumes and Filestore Regional (a managed NFS file storage service replicated across three zones) for tempdb.
tempdb: Filestore is network-attached file storage designed for shared files over NFS, not low-latency local block storage. Placing tempdb on a regionally replicated NFS mount introduces severe network and synchronization latency during heavy query spills.tempdb writes.You needed Hyperdisk Extreme for an extreme-IOPS financial OLTP database and Filestore Regional separately to host shared application configuration files or common utilities across a multi-VM web tier—not for an OLAP data warehouse's tempdb.
This solution pairs memory-optimized Compute Engine VMs (M3 or M4 series) with Hyperdisk Balanced for durable database files and Local SSD for ephemeral temporary workspace storage. Together, these components align compute memory ratios and block storage performance directly with the demands of large-scale OLAP workloads.
M3 and M4 machine series offer large memory-to-vCPU ratios, allowing SQL Server to cache massive datasets, indexes, and query execution plans directly in RAM rather than repeatedly reading from disk.tempdb spills: Local SSD is physically attached to the host server and provides the lowest latency and highest throughput for temporary sorting and hash-join data that does not require persistence across VM stops.tempdb to Local SSD prevents temporary writes from competing with data-file I/O.Unlike general-purpose or compute-optimized VMs, memory-optimized instances directly resolve buffer pool exhaustion. Pairing them with Hyperdisk Balanced and Local SSD delivers balanced cost and peak analytical throughput without the excessive IOPS pricing of Hyperdisk Extreme or the capacity-coupled scaling limits of Persistent Disk.
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