The IT industry is already in the cloud, and more and more enterprise IT infrastructures will be deployed in public, private, or mixed clouds, and it will be important for the database to be the most important part of the IT infrastructure to be integrated into the cloud. SQL Server 2014 opens up the relationship between the public cloud and the organization, whether it is running inside the enterprise or running on a Microsoft Azure virtual machine, SQL Server 2014 directly supports the deployment of data files and logs to Microsoft Azure Public cloud storage. This allows seamless opening of the boundaries of public and private clouds, as shown in Figure 1.
Figure 1. SQL Server 2014 directly supports the deployment of database files in Azure BLOBs
Why you need to deploy SQL Server data files to Microsoft Azure
Directly deploying data in Azure blob storage can directly lead to benefits such as performance, data migration, data virtualization, high availability, and disaster preparedness. But the main application scenarios and advantages are as follows:
portability : In an azure virtual machine environment, deploying data in an azure blob is easier to migrate, simply separating the database and attaching it to another azure virtual machine without moving the database file itself.
Database Virtualization : In a cloud environment that provides services to tenants, you can combine the database features after SQL Server 2012 to smooth the databases on the higher-load virtual machines to other virtual machines without affecting the normal operation of the virtual machine environment.
high availability and disaster preparedness : Because the database file is now located on a blob storage in Microsoft Azure, you can only attach the database file to another standby, even if the virtual machine itself crashes. The database can be recovered in a very short time and the data itself is unaffected by the virtual machine corruption, thus ensuring high RTO and RPO.
Scalability: Both on azure virtual machines and within the enterprise, stored iops are limited to specific environments, and IOPS can be very high on azure blob storage.
The new interactive module for SQL Server 2014 and Azure blob storage is not just an adaptation mechanism in an existing software environment, but is directly integrated into the SQL Server storage engine. The data access mechanism is shown in Figure 2.
Figure 2. Data access mechanism for database files on Azure blobs
As you can see from Figure 2, the data interaction mechanism no longer requires the virtual machine to interact through the host and IO channels, when the virtual machine can invoke storage directly through the rest API. Better performance than direct access to local storage on virtual machines.