Design data storage solutions

1. A finance team wants predictable performance with the ability to choose hardware generation and control license costs. Which Azure SQL purchasing model best fits their needs and why?

• A.Choose the vCore-based model to pick compute generation and scale storage independently while applying licensing benefits.✓ Correct
• B.Choose the DTU-based model to mix compute and storage into a single unit that maps cleanly to CPU cores.
• C.Choose the DTU-based model because it enables reserved instances and customer-managed licensing discounts.
• D.Choose either model because both allow CPU generation selection and granular IOPS configuration.

2. A trading app needs <3 ms storage latency, high IOPS, and fast failover for mission-critical OLTP. Which Azure SQL Database service tier should you choose?

• A.Select Business Critical to use local SSD with low-latency storage and synchronous replicas for fast failover.✓ Correct
• B.Select General Purpose to use remote storage with read scale-out to meet strict sub-3 ms latency goals.
• C.Select Hyperscale to gain unlimited storage growth and log service built for analytical batch workloads.
• D.Select General Purpose with tempdb optimization because it always matches SSD-local storage latency.

3. Your company owns SQL Server Enterprise licenses with Software Assurance. How can you reduce Azure SQL Database costs for long-running production workloads?

• A.Apply Azure Hybrid Benefit to vCore databases to reuse existing licenses and lower compute charges.✓ Correct
• B.Switch to DTU purchasing to automatically convert owned cores into discounted database DTUs.
• C.Move to serverless compute so licensing is not required and all usage becomes free of charge.
• D.Enable dev/test pricing on production resources because it allows SA-licensed workloads in any environment.

4. A single-tenant database experiences CPU spikes during seasonal peaks. You want minimal app changes but higher headroom. Which scaling approach should you try first?

• A.Scale up vertically to a larger compute size to increase CPU and memory without changing the data model.✓ Correct
• B.Scale out horizontally by sharding to multiple databases to eliminate all hot partitions with zero app work.
• C.Introduce eventual consistency via queues to avoid any need for scaling the database tier at all times.
• D.Migrate to Hyperscale since it always removes CPU bottlenecks regardless of access patterns.

5. A development database is idle for hours daily. You want compute to stop when idle and resume on demand. What configuration enables this outcome?

• A.Use Serverless with min and max vCores and set an autopause delay so compute stops after idle time.✓ Correct
• B.Use Provisioned compute with scale to zero enabled, which halts compute after five minutes of inactivity.
• C.Use Hyperscale tier because its log service automatically suspends compute during idle windows.
• D.Use Elastic Pools since pools always pause member databases when utilization is below threshold.