PolyBase Accesses Azure Blob Storage

The PolyBase technology in SQL Server 2016 accesses unstructured data stored in Azure Blob Storage. The illustration shows how a SQL Server 2016 cluster (top) can use PolyBase to process T-SQL queries to unstructured data in Blob Storage (bottom). The PolyBase engine runs on the head node of a PolyBase Group of servers, where it coordinates queries that are distributed to compute nodes.

Blob Storage data schemas must be created by administrators (left) who use new SQL commands to create table schemas that represent the data source. The schemas are saved as database-level external tables that users can query with T-SQL (upper left). This limits the use of Blob Storage to data with known formats, such as historical data stored in delimited text files.

Querying Blob Storage data starts with the head node, which parses queries, generates query plans, and distributes the workload to the compute nodes (center). Compute nodes access the Blob Storage data (bottom) directly and manage the movement of data. Results can also be joined with traditional SQL Server relational tables (top right) in real time.