Why choose Windows Azure to host SQL?

Author information: This article is a joint work by Madhan Arumugam and Guy Bowerman of the SQL Server Cloud infrastructure team.

Where SQL data is hosted, Windows Azure gives you two choices, SQL Server on the VM (hereinafter referred to as SQL/VM) and Windows Azure SQL database (hereinafter referred to as SQL DB). Both SQL/VM and SQL DB are built from a single code base and have their own lengths in the cloud deployment. This article will draw on the previous blog SQL Server in Windows Azure Virtual Machine vs. SQL Database and provide a simple decision tree to help you choose the best SQL for your application or service on Windows Azure Database Solutions. Windows Azure SQL DB has been in production for several years as a Microsoft platform, a service (PaaS). SQL Server on Windows Azure Virtual Machine provides an integrated database server instance of Microsoft infrastructure as a service (IaaS) as a set of pre-built platform images. The public preview currently provides Windows Azure Virtual Machine.

The focus of this article is on SQL Engine, not BI and reporting. You will then explore in detail how to make choices in bi and reporting.

Why Azure is chosen to host SQL?

Windows Azure is highly integrated across the Microsoft stack, such as applications such as SharePoint and Active Directory, while consolidating the licensing of different deployment modules through unified development experience. These integrations go through a lot of tests that work with Microsoft products. It also provides strong interoperability. In addition to supporting a variety of operating systems including Linux, Windows Azure VMs is also based on the standard VHD format, making it easy to import and export them in Azure environments.

Such integration, standardization, and interoperability make Azure an ideal cloud environment for running SQL Server, as well as providing a cohesive experience for Microsoft assets from a traditional indoor platform to a private cloud and a public cloud.

Ongoing innovation in SQL/VM and SQL db not only lowers the overall TCO of managing existing applications or services, but also provides an efficient platform for a new generation of cloud applications or services. In addition, it can be combined with existing and new applications on the same managed service, and can also assemble indoor applications and networks as a hybrid model to provide great flexibility.

Choosing the Right Database solution

Microsoft database products are available in a physical, private cloud environment, a third-party managed private cloud environment, and a public cloud. It provides a range of SQL Server products to make it a coherent and flexible open data platform.

Each product can be differentiated by your level of control over the infrastructure, and the cost effectiveness of the integration and automation of the database. There are, of course, other aspects to differentiate, such as compatibility, which we will detail in the decision tree below.

This article focuses on the comparison between SQL/VM and SQL DB and provides a way to select database services, according to which you can choose the database services that match the cost efficiencies and controls you expect.

Decision Tree

In this decision tree, consider the following specific decision points for the application or service:

New or existing applications or services: typically migrating existing applications ("Upgrades and Transformations") requires simulating indoor application operations more than new applications, and new applications are generally not necessary. Application or Service requirements: This decision point checks for any compatibility, network, or SQL Server security requirements that may need to be in the VM. For more information, see SQL Server in Windows Azure Virtual Machine vs. sql database and MSDN article for a complete list of SQL DB and SQL/VM feature compatibility differences guidelines and Limitations (Windows Azure SQL Database). Large-scale economic (RE) architecture/zoning will: This decision point checks whether to use SQL DB extension technology such as federations in Windows Azure SQL database to achieve flexibility and expansion requirements. To take full advantage of the benefits of SQL DB's massive economies and resiliency, a certain degree of restructuring is often required. Database size: The SQL database version is limited in size and is currently a single Azure SQL db with a maximum of GB. However SQL/VM, the current maximum VM size is extralarge (8 CPU kernel, GB RAM, and up to TB of disk space). VM size for more information see MSDN article How to Configure Virtual Machine sizes.

Other considerations

In addition to these high-level decision points, there are a number of lower-level design standards and size expectations that can also affect database solutions, such as a new application that may require a special feature rather than SQL DB. The framework used by this decision tree simply provides a simple way to help you quickly evaluate your application requirements.

An important consideration is the SLA provided by IaaS and the PaaS database solution. For example, all VM storage in Iaas,windows Azure has SLAs. When Windows Azure VM is available, a 99.95% of the Windows Azure Vm-level SLA for two or more instances typically is available in one set. There is a database-level availability SLA in SQL db. For more details, see http://www.windowsazure.com/en-us/support/legal/sla/. Operations that are not covered by the SLA require the user to do it themselves.

Applications or services do not need to choose between IaaS and PaaS. The hybrid application pattern, which includes both PaaS and IaaS services (for example, SQL DB and SQL/VM on Azure), represents the reality that application development and deployment allows applications to run components on VMS, which require compatibility and control when migrating data from within the user BI or applications, and use databases in SQL db to minimize DBA and database costs.

Conclusion

Windows Azure SQL database and SQL Server running on Windows Azure virtual machines are good ways to implement different database application requirements.

In summary, if the following conditions are met, select SQL/VM:

you need to be fully compatible with your internal SQL Server. You want to implement the existing application with minimal changes. You need infrastructure layer isolation. Your application or service is designed for the Schedule database table (AKA extension).

Select SQL DB If the following conditions are true:

you are creating a net cloud based on your application. You want to reduce the cost of database management and focus on the application layer. You need the isolation of the database layer. Your application or service is designed for "unplanned and unrestricted database tables" (through extended resiliency).