The short answer
Azure Synapse Analytics Dedicated SQL Pool (formerly Azure SQL Data Warehouse) uses a massively parallel processing architecture designed around a fixed physical foundation: 60 distributions.
+-------------------------------------------------+
| Control Node |
| Receives client SQL, optimizes MPP execution |
+-------------------------------------------------+
|
+----------------+----------------+
| |
+----------------+ +----------------+
| Compute Node | | Compute Node |
| Distributions | | Distributions |
| 1 through 30 | | 31 through 60 |
+----------------+ +----------------+
|
+-------------------------------------------------+
| Azure Blob / ADLS Gen2 Storage |
| 60 underlying storage locations on Azure |
+-------------------------------------------------+The Control Node
The control node is the front door of the Dedicated SQL Pool. It accepts client connections, parses standard T-SQL, coordinates parallel execution plans, and dispatches query fragments to compute nodes. It does not store user data.
Compute nodes and DWU scaling
The compute power of a Dedicated SQL Pool is measured in Data Warehouse Units (DWUs), ranging from DW100c to DW30000c. Regardless of your DWU setting, Synapse always divides table data into exactly 60 logical storage buckets.
What changes when you scale DWUs is how many compute nodes own those 60 distributions:
- At DW100c (smallest size), a single compute node processes all 60 distributions sequentially.
- At DW500c, 5 compute nodes process 12 distributions each in parallel.
- At DW6000c, 60 compute nodes process 1 distribution each simultaneously.
The Storage layer
Compute nodes do not store permanent data on local disks. Data resides in Azure Storage. When you pause a Dedicated SQL Pool, compute nodes are released and billing for compute drops to zero, while your storage remains intact.