الملاحظات تحرير

Migrate your relational databases to Azure

  • مقالة
  • قراءة خلال 6 دقائق

Tip

This content is an excerpt from the eBook, Modernize existing .NET applications with Azure cloud and Windows Containers, available on .NET Docs or as a free downloadable PDF that can be read offline.

Download PDF

Modernize existing .NET applications with Azure cloud and Windows Containers eBook cover thumbnail.

Vision: Azure offers the most comprehensive database migration.

In Azure, you can migrate your database servers directly to IaaS VMs (pure lift and shift), or you can migrate to Azure SQL Database, for additional benefits. Azure SQL Database offers the managed instance and full database-as-a-service (DBaaS) options. Figure 3-1 shows the multiple relational database migration paths available in Azure.

Database migration paths in Azure

Figure 3-1. Database migration paths in Azure

When to migrate to Azure SQL Database Managed Instance

In most cases, Azure SQL Database Managed Instance will be your best option to consider when you migrate your data to Azure. If you are migrating SQL Server databases and need nearly 100% assurance that you won't need to rearchitect your application or make changes to your data or data access code, choose the Managed Instance feature of Azure SQL Database.

Azure SQL Database Managed Instance is the best option if you have additional requirements for SQL Server instance-level functionality, or isolation requirements beyond the features provided in a standard Azure SQL Database (single database model). This last one is the most PaaS-oriented choice, but it doesn't offer the same features as that of a traditional SQL server. Migration might surface frictions.

For example, an organization that has made deep investments in instance-level SQL Server capabilities would benefit from migrating to SQL Managed Instance. Examples of instance-level SQL Server capabilities include SQL common language runtime (CLR) integration, SQL Server Agent, and cross-database querying. Support for these features is not available in standard Azure SQL Database (a single-database model).

An organization that operates in a highly regulated industry, and which needs to maintain isolation for security purposes, also might benefit from choosing the SQL Managed Instance model.

Managed Instance in Azure SQL Database has the following characteristics:

  • Security isolation through Azure Virtual Network

  • Application surface compatibility, with these features:

    • SQL Server Agent and SQL Server Profiler

    • Cross-database references and queries, SQL CLR, replication, change data capture (CDC), and Service Broker

  • Database sizes up to 35 TB

  • Minimum-downtime migration, with these features:

    • Azure Database Migration Service

    • Native backup and restore, and log shipping

With these capabilities, when you migrate existing application databases to Azure SQL Database, the Managed Instance model offers nearly 100% of the benefits of PaaS for SQL Server. Managed Instance is a SQL Server environment where you continue using instance-level capabilities without changing your application design.

Managed Instance is probably the best fit for enterprises that currently are using SQL Server, and which require flexibility in their network security in the cloud. It's like having a private virtual network for your SQL databases.

When to migrate to Azure SQL Database

As mentioned, the standard Azure SQL Database is a fully managed, relational DBaaS. SQL Database currently manages millions of production databases, across 38 datacenters, around the world. It supports a broad range of applications and workloads, from managing straightforward transactional data, to driving the most data-intensive, mission-critical applications that require advanced data processing at a global scale.

Because of its full PaaS features, better pricing-and ultimately lower cost-you should move to the standard Azure SQL Database as your "by-default choice" if you have an application that uses basic, standard SQL databases, and no additional instance features. SQL Server features like SQL CLR integration, SQL Server Agent, and cross-database querying are not supported in the standard Azure SQL Database. Those features are available only in the Azure SQL Database Managed Instance model.

Azure SQL Database is the only intelligent cloud database service that's built for app developers. It's also the only cloud database service that scales on-the-fly, without downtime, to help you efficiently deliver multitenant apps. Ultimately, Azure SQL Database leaves you more time to innovate, and it accelerates your time to market. You can build secure apps and connect to your SQL database by using the languages and platforms that you prefer.

Azure SQL Database offers the following benefits:

  • Built-in intelligence (machine learning) that learns and adapts to your app

  • On-demand database provisioning

  • A range of offers, for all workloads

  • 99.99% availability SLA, zero maintenance

  • Geo-replication and restore services for data protection

  • Azure SQL Database Point in Time Restore feature

  • Compatibility with SQL Server 2016, including hybrid and migration

The standard Azure SQL Database is closer to PaaS than Azure SQL Database Managed Instance. Prefer the standard Azure SQL Database because you'll get more benefits from a managed cloud. However, Azure SQL Database has some key differences from regular and on-premises SQL Server instances. Depending on your existing application's database requirements, and your enterprise requirements and policies, it might not be the best choice when you are planning your migration to the cloud.

When to move your original RDBMS to a VM (IaaS)

One of your migration options is to move your original relational database management system (RDBMS), including Oracle, IBM DB2, MySQL, PostgreSQL, or SQL Server, to a similar server that's running on an Azure VM. If you have existing applications that require the fastest migration to the cloud with minimal changes, or no changes at all, a direct migration to IaaS in the cloud might be a fair option. It might not be the best way to take advantage of all the cloud's benefits, but it's probably the fastest initial path.

Currently, Microsoft Azure supports up to 331 different database servers deployed as IaaS VMs. These include popular RDBMS like SQL Server, Oracle, MySQL, PostgreSQL, and IBM DB2, and many other NoSQL databases like MongoDB, Cassandra, DataStax, MariaDB, and Cloudera.

Note

Although moving your RDBMS to an Azure VM might be the fastest way to migrate your data to the cloud (because it is IaaS), this approach requires a significant investment in your IT teams (database administrators and IT pros). Enterprise teams need to be able to set up and manage high availability, disaster recovery, and patching for SQL Server. This context also needs a customized environment, with full administrative rights.

When to migrate to SQL Server as a VM (IaaS)

There might be a few cases where you still need to migrate to SQL Server as a regular VM. An example scenario is if you need to use SQL Server Reporting Services. In most cases, though, Azure SQL Database Managed Instance can provide everything you need to migrate from on-premises SQL servers, so migration to a SQL Server VM should be your last resort to try.

Use Azure Database Migration Service to migrate your relational databases to Azure

You can use Azure Database Migration Service to migrate relational databases like SQL Server, Oracle, and MySQL to Azure, whether your target database is Azure SQL Database, Azure SQL Database Managed Instance, or SQL Server on an Azure VM.

The automated workflow, with assessment reporting, guides you through the changes you need to make before you migrate the database. When you are ready, the service migrates the source database to Azure.

Whenever you change an original RDBMS, you might need to retest. You also might need to change the SQL sentences or Object-Relational Mapping (ORM) code in your application, depending on testing results.

If you have any other database (for example, IBM DB2) and you opt for a lift and shift approach, you might want to continue using those databases as IaaS VMs in Azure, unless you are willing to perform a more complex data migration. A more complex data migration will require additional effort because you'd be migrating to a different database type with the new schema and different programming libraries.

Additional resources

Previous Next