When Microsoft released the new SQL Server 2008 R2 product line, it also introduced a new edition of SQL Server: SQL Server 2008 R2 Parallel Data Warehouse. The Parallel Data Warehouse edition is quite different from other editions of SQL Server. Here are answers to the most frequently asked questions about the Parallel Data Warehouse edition.

1. What is SQL Server 2008 R2 Parallel Data Warehouse?

Formerly code-named Madison, SQL Server 2008 R2 Parallel Data Warehouse is essentially the SQL Server equivalent to the Windows Server High Performance Computing (HPC) edition. The Parallel Data Warehouse edition will utilize Massively Parallel Processing (MPP) scale-out architecture to support very large data warehouses. It’s decidedly not a low-end offering—Microsoft is licensing it at $57,489 per processor.

2. How is it different from other editions of SQL Server?

Unlike the other editions of SQL Server which are software-only products, the Parallel Data Warehouse edition is a combination hardware and software offering. You can’t just buy SQL Server 2008 R2 Parallel Data Warehouse edition from Microsoft and install it on one of your own severs as you can with the other editions of SQL Server. Instead, you purchase SQL Server 2008 R2 Parallel Data Warehouse from one of Microsoft’s partners, which are HP, Dell, IBM, and Bull. Each of the different partners provides a variety of different sizes of system configurations.

3. Is the Parallel Data Warehouse edition the same thing as the Fast Track Data Warehouse?

No. The Fast Track Data Warehouse and the Parallel Data Warehouse edition are different. SQL Server Fast Track Data Warehouse is a set of reference hardware configurations that are available from several different Microsoft partners. These hardware configurations have been optimized by each of the vendors to provide excellent out-of-the-box performance for data warehousing implementations. However, Fast Track Data Warehouse system configurations are standard Symmetrical Multi Processing (SMP) systems that run the SQL Server Standard, Enterprise, or Datacenter editions.

4. Who is it targeted toward?

It’s targeted toward organizations that have very, very large data warehousing requirements. Solutions from the different partners are designed to support data warehouses ranging from 10TBs to 1+ PBs. The Parallel Data Warehouse edition is designed for businesses whose data warehousing requirements push the limits of a standalone SQL Server implementation.

5. How is it architected?

With the other editions of SQL Server, queries are processed within one physical instance of a database using that system’s CPUs, memory, and storage. The Parallel Data Warehouse edition’s “appliance” consists of a Control Node and multiple Compute Nodes. Tables are partitioned across multiple Compute Nodes where each node has its own CPU, memory, and storage. Since the data is distributed across multiple Compute Nodes, the Parallel Data Warehouse edition is able to process queries in parallel where each node is acting on its own data. This enables the queries for these large tables to be completed much faster than processing them on a single SMP system. To add additional scalability to a Parallel Data Warehouse system, you simply add more Compute Nodes.

You can find out more about the SQL Server 2008 R2 Parallel Data Warehouse at Getting Started with Parallel Data Warehouse and also from the Microsoft Parallel Data Warehouse website.