Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News SQL Server is the New Hunting Ground in the Cloud

SQL Server is the New Hunting Ground in the Cloud

This item in japanese

You can now run Microsoft SQL Server on Google's Cloud Platform, with a new service in beta. SQL Server has seen a recent focus from the major cloud providers. Amazon added native backups for SQL Server on AWS, and Microsoft built Stretch Database into SQL Server 2016, integrating on-premise databases seamlessly with Azure.

Google's new offering is at the IaaS level, with official public images running SQL Server on Windows Server that you can spin up as VMs on Google Compute Engine. Currently only Web and Standard versions of SQL are supported, although you can choose from SQL Server 2012, 2014 or 2016, and if you have Software Assurance from Microsoft you can bring your own license.

Users have to manage the database and the underlying Windows OS themselves, but Google focused on the performance levels available with their compute offering:

We're increasing the maximum read and write IOPS for SSD-backed Persistent Disk volumes from 15,000 to 25,000.

On Amazon's Relational Database Service (RDS), customers can run SQL Server as a managed, PaaS database, but storage performance is a maximum of 20,000 IOPS. (Microsoft doesn't publish IOPS values for Azure SQL Database, preferring their Database Transaction Unit, "DTU", metric which approximates real-world usage).

Google is positioning the new SQL Server capability as a straight migration path for enterprises currently running their own servers, as Dominic Preuss, Lead Product Manager says in his blog post:

We want enterprises with databases of all sizes and types to experience the best price-performance with the least amount of friction.

But the feature set at the IaaS level is lacking compared to the PaaS competition. Most significantly, to resize a database server on Google Cloud Platform, the VM needs to be taken offline. On Amazon RDS, users can resize a SQL Server database instance on the fly. On Azure SQL Database users can resize on the fly, and can also group multiple databases in an elastic pool to share compute resources.

The Google cloud is ramping up services to compete with the more established suites from Amazon and Microsoft. The existing PaaS database, Google Cloud SQL, is a managed relational database built on MySQL. Adding SQL Server expands Google's database offering. The current IaaS implementation from Google lacks the value-added features of the rival PaaS platforms, but it could be a welcome increase in choice for Microsoft users moving to the cloud.

Rate this Article