3.2.2. Azure SQL Database Deployment Options

💡 First Principle: Azure SQL Database offers deployment models that trade off isolation, control, and cost—allowing you to optimize for your specific workload pattern. The key insight is that resources can be dedicated (guaranteed but potentially wasteful) or shared (efficient but competitive).

Scenario: A startup launches a new SaaS product. Initially, they need a cheap, small database. As they grow, they may need more performance or want to consolidate multiple customer databases for efficiency.

Single Database

• What It Is: One isolated database with dedicated resources
• Use Case: Simple applications, dedicated workloads
• Scaling: Scale up/down independently

Elastic Pool

• What It Is: Multiple databases sharing a pool of resources (eDTUs or vCores)
• Use Case: SaaS applications with many tenants whose usage peaks at different times
• Benefit: Cost efficiency through resource sharing

Purchasing Models

• DTU-Based: Bundled measure of CPU, memory, and I/O. Simple but less granular control.
• vCore-Based: Choose vCores, memory, and storage independently. More control, easier to compare with on-premises.

Service Tiers

• Basic/Standard/Premium (DTU): Tiered performance levels
• General Purpose (vCore): Balanced compute and storage
• Business Critical (vCore): High performance, built-in HA with local SSD
• Hyperscale (vCore): Up to 100 TB, rapid scale-out, near-instant backups

Visual: SQL Database Deployment Options