This content is part of the Essential Guide: An enterprise guide to big data in cloud computing

When should I use Azure SQL Data Warehouse?

Amazon Redshift is the market leader in cloud data warehousing services. But what does Azure SQL Data Warehouse bring to the table?

Data warehouses in the cloud are gaining popularity, especially as cloud vendors offer data warehousing services...

at lower costs. And while Amazon Redshift dominates the cloud data warehouse market, Microsoft Azure offers a competitive platform that businesses should also consider.

Microsoft Azure SQL Data Warehouse, currently in preview, builds on the Microsoft SQL Server platform and should be familiar to organizations that work with Microsoft T-SQL and Power BI. The data warehouse service uses a columnar data store, so it is optimized for the queries typically found in business intelligence applications. Amazon Redshift, which is based on PostgreSQL, also uses columnar data storage.

But how else does SQL Data Warehouse compare to Redshift?

The biggest difference between Azure SQL Data Warehouse and Amazon Redshift is the decoupling of storage and compute resources. When creating a Redshift cluster, users choose machine instances with the amount of storage needed. If there are extended periods when users do not need the cluster, they can shut it down. But first, users must export the data warehouse contents to Amazon Simple Storage Service. Then, they must reload the data when creating another cluster.

By comparison, in Azure SQL Data Warehouse, users scale compute resources and storage independently. If they are not running queries, they can pause compute resources while data stays intact on cloud storage.

Another valuable feature in SQL Data Warehouse is PolyBase, which lets users query across nonrelational sources. Developers and analysts can pull data from multiple sources, such as Azure's blob storage, Hadoop HDFS and files in delimited and ORC format.

Azure SQL Data Warehouse pricing consists of separate charges for compute and storage resources. Compute resources are billed in Data Warehouse Units -- a Microsoft metric based on query performance related to scanning, loading and copying data -- which costs $0.70 per hour or approximately $520 per month. Storage billing is based on Azure's standard storage rates.

Although SQL Data Warehouse is in preview mode, it has the potential to attract committed Microsoft shops and data warehouse users who want to separately manage storage and compute. In addition, the service's ability to query over both relational and nonrelational data stores appeals to organizations working with applications that require NoSQL back ends, as well as those working with big data platforms like Hadoop.

Next Steps

An in-depth looks at Amazon Web Services

Is Amazon Redshift the right data warehouse for you?

Migration issues with cloud data warehousing

Should you store your data in the cloud?

Your alternatives to on-premises data warehouses

Dig Deeper on Google and other public cloud providers