SQL Managed Instance Adds Limited Data Virtualization

• Azure SQL Managed Instance has added data virtualization features based on SQL Server’s PolyBase capability.
• The features can help query and join external data with SQL data, providing near real-time access and removing some integration requirements.
• However, Managed Instance data virtualization can only query Azure storage, which makes it less useful than using PolyBase with SQL Server.

Azure SQL Managed Instance released data virtualization features based on SQL Server’s PolyBase capabilities. Data virtualization features query external data in-place rather than copying it into the service; for example, an analytic application might use data virtualization to perform some calculations directly on data in a firm’s enterprise resource planning  (ERP) system, rather than copying the data into a data mart first. This makes it more likely the data is up to date, reduces the workload on SQL Managed Instance, and reduces the amount of data transferred to the service. Data virtualization features also help simplify querying non-SQL data, such as unstructured and document-based data. Support for data virtualization could provide an option for moving some SQL Server workloads to Managed Instance if the workload can fit within Managed Instance’s limited implementation of PolyBase. The service only allows querying specific data formats stored in Azure Data Lake Storage (ADLS) and Azure Blob Storage, which is substantially fewer options than SQL Server provides. (For more details, see the chart “Comparing PolyBase in SQL Server and Azure”.)