• A non-relational and multi-model database service.
• Improved DB account availability: Data is distributed to other regions without affecting the data consistency across various areas where the data replicates.
• Fits any web application, mobile application, gaming, or IoT application that requires processing, reading, and writing a massive amount of data.
• Users can scale the throughput and storage resources elastically and independently across Azure regions. Thus, it enjoys a fast, less than 10ms latency for both reads and writes workloads due to using local data caching and automatically indexes all incoming data.
• Azure Cosmos DB can access the stored data using the API that fits the user’s application requirements from the supported SQL, MongoDB, Cassandra, Tables, or Gremlin, and the appropriate SDK from the .NET, Java, Node.js, Python, and Xamarin-supported SDK.
• Easy Migration: A user can migrate the application to Cosmos DB without significant changes.
• It provides five well-defined consistency options: Strong, bounded staleness, session, consistent prefix, and eventual, which offers complete flexibility and a low cost-to-performance ratio.
• Throughput: It pro allows the ability to scale the read and write operations and add or remove any regions to respond directly to unexpected workload spikes.
• Security: The stored data is always secured using data encryption at rest and in motion and configuring the row-level authorization.

• Easy Migration: A user takes a backup from the on-premises SQL Server, saves it in an Azure storage account, and restores it to your Azure SQL Managed Instance.
• Compatible with SQL Server and Linked Server
• It allows hosting up to 100 databases in the same SQL Server instance, thus overcoming the cross-database querying limitation of the Azure SQL Database.
• Supports the common language runtime (CLR) feature
• Provides access to the system databases, allows the use of SQL Agent jobs, and provides the ability to manually take copy-only backups of the databases to Azure storage.
• Dedicated computing and storage services
• Supports maximum database size of 8TB
• High Availability: 99.99 to 99.995%
• Low price

Service Tiers

The General Purpose and the Business-Critical

Business-Critical

Supports the In-Memory OLTP, readable secondary replicas, includes more memory per core
offers lower storage latency by using direct-attached storage.

Features

• Single Fully Managed PaaS Service
• It is more accessible and cost-effective to store the application-deployed processed data.
• Enables developers to focus on database design and performance optimization.
• The developer chooses the database name and a few options to get the database up and running within a few minutes!
• Dynamically Scale Resources: The provider offers several pricing tiers to scale the resources up and down by tuning the performance tier on the fly to meet current workload requirements and at convenience.
• Lower Price: Easy auto-scale and auto-close capabilities lower the costs of the development and testing environments of the Azure SQL Database. The database charging pauses after an hour of inactivity but with a higher per-hour or vCore pricing cost than the other Azure SQL Database deployment models.
• Less Administrative Effort: The provider Azure implements, manages, and maintains the underlying infrastructure required to host the database, patching the operating system and SQL Server with the latest security updates. Furthermore, it contains all backup, restore, or retention policies. Hence, the developer need not worry about spending time on the same.
• Automatic database Index tuning feature: It helps to identify the expensive queries, automatically forcing the last good performing execution plan and adding and removing indexes based on the workload requirements, using built-in intelligence and advanced heuristics mechanisms.
• High Availability & Uptime Guarantee: 99.99% per cent, a customer never loses data due to a failure. The database has no single point of failure, with the high availability and disaster recovery solutions built into that PaaS platform.
• Fully Managed Automatic Backup: It takes weekly full backups, twice-daily differential backups, and log jams every 5-10 minutes.
• The database deployed to an SQL Server instance hosted in an Azure Virtual machine supports the same features and structure the Microsoft SQL Server supports.
• High Performance, Monitoring, Database scoped configuration and authentication.

Resource charging models

Database Transaction Unit (DTU) model

A customer pays one fixed price for computing or IO/ memory and data storage.

vCore model

It allows separate charges for computing and a particular order for your storage, which provides more flexibility in managing your costs than with DTU.

Deployment models

Single database
Elastic pools
Hyper-scale

Single database

• The simplest deployment model of Azure SQL Database, managed and scaled individually, with its dedicated resources.
• A user creates a logical server, deploys the database to that server, and connects to that database directly.

Elastic Pools model

• A database deployed to an elastic pool shares the same resources with other multiple databases deployed to the same collection. Thus reducing the overall cost of hosting these databases in Azure and allowing the scaling process to be performed at the elastic pool level.
• This model fits unstable workloads with non-concurrent or not-frequent spikes in the workload.

The Azure SQL Database Hyperscale model

• A serverless model allows overcoming the Azure SQL DataDatabase’sstorage limitation.
• It can expand 100 TB and beyond storage for each database using an advanced scaling technique by adding compute nodes as the data sizes grow, with the additional cost only per terabyte for the repository.