Plan and implement data platform resources
Three Azure SQL products, three levels of control
An application that schedules SQL Server Agent jobs can't keep them on Azure SQL Database, which has no SQL Server Agent and offers elastic jobs instead, while Azure SQL Managed Instance supports SQL Server Agent. The gap comes from how much of SQL Server each of the three Azure SQL products gives you, which this guide calls the level of control: Azure SQL Database sits at the database level and supports most on-premises database-level capabilities, Azure SQL Managed Instance sits at the instance level and supports almost all on-premises instance-level and database-level capabilities, and SQL Server on Azure Virtual Machines (VMs) sits at the operating-system level, with full administrative control over the SQL Server instance and the underlying OS. The same choice decides which migration methods can bring existing SQL Server data in. An online migration keeps picking up the changes made on the source while it runs, and an offline one needs you to stop the workload that modifies source data; Azure SQL Database doesn't run the RESTORE statement, so data reaches it offline as, for example, a BACPAC file (a package of a database's schema and data) imported into a new database or online through transactional replication. A managed instance, by contrast, restores SQL Server backups placed in Azure Blob Storage and can also receive databases over Managed Instance link, which replicates them through a distributed availability group. Extra capacity never changes the level: the databases in an elastic pool share a set of resources on one server, but they are still Azure SQL databases, without SQL Server Agent.
Choose and deploy, configure for scale, then migrate
This domain carries 15–20% of the DP-300 exam, and its three subtopics follow the skills outline's order. Plan and deploy Azure SQL solutions covers the skills outline bullets from "Recommend a database offering based on specific requirements" and "Deploy database offerings on selected platforms" through "Plan for table partitioning", "Recommend a database sharding solution", and "Apply patches and updates for hybrid and infrastructure as a service (IaaS) deployment": you choose a product, plan partitioning or sharding, deploy it, and patch it. Its bullets also name two related products: "Identify use cases for Azure SQL Database in Microsoft Fabric", where SQL database in Fabric uses the same SQL Database Engine as Azure SQL Database, and "Identify use cases for Azure Arc-enabled SQL services", where SQL Server enabled by Azure Arc extends Azure services to SQL Server instances hosted outside of Azure. Configure resources for scale and performance covers one skills outline bullet per product, from "Configure Azure SQL Database for scale and performance" to "Configure SQL Server on Azure Virtual Machines for scale and performance", then "Configure table partitioning" and "Configure data compression": planning partitioning sits with the first subtopic, building it with the second. Each setting resizes one scaling scope (this guide's term for the unit a setting changes): a single database or elastic pool, whose maximum resources you define per database or per group of pooled databases; a managed instance, where all databases share the resources allocated to the instance; a SQL Server VM, where you control the VM size, the number of disks, and their storage configuration; or a table or index, because partitioned tables and indexes and row and page compression apply to SQL Server, Azure SQL Database, and Azure SQL Managed Instance alike. A service-tier change, such as a move into Hyperscale, resizes the scaling scope in place: a new compute instance is created with the requested service tier while the database remains online, though the docs call this a migration to Hyperscale. Plan and implement a migration strategy covers the skills outline bullets from "Evaluate requirements for a migration" and "Evaluate offline or online migration strategies" through "Implement a migration between Azure SQL services", "Implement Azure SQL Managed Instance database copy and move", and "Troubleshoot a migration": you migrate, copy, or move existing databases. Azure Database Migration Service (DMS) migrates online or offline to Azure SQL Managed Instance and to SQL Server on Azure VMs, but only offline to Azure SQL Database, and a VM also accepts SQL Server's own methods, such as backup and restore or log shipping.
Let requirements, not tools, raise the level of control
When two products meet the requirements, what remains is who runs the operations: with Azure SQL Database and Azure SQL Managed Instance you don't manage virtual machines, the operating system, or database software, nor upgrades, high availability, or backups, while on SQL Server on Azure VMs it's up to you to decide when to update the operating system and database software. So each step up in level of control should answer a requirement the lower level can't meet, such as SQL Server Agent jobs for the instance level, or operating-system access for the operating-system level, where SQL Server on Azure VMs is best for migrations and applications that require OS-level access. Those requirements come from an assessment rather than from the migration tool: DMS doesn't assess the compatibility of the source, whereas Azure Migrate assesses SQL Server instances to measure cloud readiness and identify risks and SQL Server enabled by Azure Arc automatically produces an assessment for migration to Azure, before you pick a target.
Level of control, scaling scope, and migration methods for each Azure SQL product
| Decision | Azure SQL Database | Azure SQL Managed Instance | SQL Server on Azure VMs | Drill into |
|---|---|---|---|---|
| Level of control | Database level: most database-level capabilities | Instance level: almost all instance-level and database-level capabilities | Operating-system level: full engine control plus OS access | Plan and deploy Azure SQL solutions |
| Scaling scope | Single database or elastic pool, plus a table or index | Managed instance, whose resources all its databases share, plus a table or index | SQL Server VM (VM size and disks), plus a table or index | Configure resources for scale and performance |
| Migration methods (examples) | No native backup restore: BACPAC import or DMS offline, transactional replication online | Backups restored from Azure Blob Storage, Managed Instance link, or DMS online or offline | DMS online or offline, or SQL Server's own methods such as backup and restore | Plan and implement a migration strategy |