Azure SQL Database Service Tiers

Last updated: October 26, 2023

Azure SQL Database offers a variety of service tiers designed to meet different performance, scalability, and cost requirements. Understanding these tiers is crucial for selecting the optimal configuration for your application. This document provides a comprehensive overview of the available service tiers and their key characteristics.

Note: Azure SQL Database has transitioned to a vCore-based purchasing model. The DTU-based model is still available for existing customers, but new deployments should consider the vCore model for its flexibility and advanced features.

vCore-based Purchasing Model

The vCore model provides more flexibility and transparency. You choose compute (vCores, memory) and storage resources independently. It offers three distinct service tiers:

General Purpose

The General Purpose tier is designed for most common workloads, offering a balanced combination of price and performance. It provides a managed database experience with a focus on ease of use and cost-effectiveness.

Ideal for typical business applications.
Uses Azure Premium Storage for low latency.
Provides read-only replicas for offloading read workloads.
Supports geo-replication for disaster recovery.

Business Critical

The Business Critical tier is designed for mission-critical applications that require the highest levels of performance, availability, and low latency. It offers advanced features for demanding workloads.

Provides the lowest I/O latency.
Includes built-in high availability and disaster recovery with multiple replicas.
Optimized for transactional workloads and demanding databases.
Offers features like in-memory online transaction processing (OLTP) for enhanced performance.

Hyperscale

The Hyperscale tier is a highly scalable service tier designed for large databases with potentially exponential growth. It offers independent scaling of compute and storage, allowing for very large database sizes and high transaction rates.

Supports database sizes up to 100 TB.
Provides rapid database restores and high availability.
Scales compute and storage independently.
Cost-effective for large datasets.

DTU-based Purchasing Model (Legacy)

The DTU (Database Transaction Unit) model simplifies management by abstracting underlying resources into a bundled unit. While still supported, the vCore model is recommended for new deployments.

Service Tier	Performance Level	DTUs	Max Data Size	Description
Basic	Low	5 - 10	2 GB	Entry-level for small applications.
Standard	S0 - S12	10 - 300	250 GB	Balanced performance for most common workloads.
Premium	P1 - P15	125 - 4000	500 GB - 1 TB	High performance for demanding applications.

Choosing the Right Service Tier

The selection of a service tier depends on several factors:

Performance Requirements: What is the expected transaction volume and latency tolerance?
Availability Needs: What level of uptime is required, and what are the disaster recovery RPO/RTO targets?
Database Size: How large is your database expected to grow?
Budget: What is the allocated budget for database infrastructure?

Recommendation: For new deployments, it is highly recommended to use the vCore-based purchasing model. The General Purpose tier is a good starting point for most applications, while Business Critical and Hyperscale are suitable for more demanding scenarios.

Key Considerations

DTUs vs. vCores: Understand the differences and choose the model that best suits your needs. The vCore model offers more granular control and better price performance for many workloads.
Performance Benchmarking: Test your application's performance against different service tiers to identify the optimal configuration.
Scalability: Plan for future growth and select a tier that can scale with your application.

For detailed specifications, pricing, and comparison guides, please refer to the official Azure SQL Database pricing page and the Microsoft Docs.