Azure SQL Data Warehouse

If the stored and analyzed data exceeds the options of the Azure SQL Database for both the storage and the compute capacity, we can consider Azure SQL Data Warehouse as another cloud-based solution for data analysis and data science with SQL Server. With on-premise and cloud, we have several options for the data store and data analysis. The advantage of Azure SQL Data Warehouse is the unlimited storage and massive scalability, which is available with two different configuration models. As you can see in the following diagram, there are several choices for running a database based on the scale needs and possibility to use cloud versus. on-premise solutions:

Azure SQL Data Warehouse uses a massive parallel processing with Azure storage, which offers great performance and scalability for your environment. As with Azure SQL Database, Azure SQL Data Warehouse uses an abstract performance unit named Data Warehouse Unit or DWU for short. DWU is based on three metrics:

• Search/aggregation operation, which is very IO intensive
• Read operation
• Create table as select (CTAS)

The architecture for the Azure Data Warehouse introduces two different layers of processing: control and compute nodes. These are used for work distribution and parallelism:

Applications connect to the control node, which is actually a SQL Server, and this control node splits and delivers the work requests to the compute nodes with massive parallelism.

For migration and data import, you can use several different approaches, one of them being Azure Data Warehouse Migration Utility, which can be used to transfer your on-premise databases or cloud databases to Azure Data Warehouse, including the check and report for any incompatibilities before moving forward: