Specifications

109

C H A P T E R 6

Scalable Data Warehousing

icrosoft SQL Server 2008 R2 Parallel Data Warehouse is an enterprise data ware-

house appliance based on technology originally created by DATAllegro and

acquired by Microsoft in 2008. In the months following the acquisition, Microsoft re-

vamped the product by changing it from a product that used the Linux operating system

and Ingres database technologies to a product based on SQL Server 2008 R2 and the

Windows Server 2008 operating system. SQL Server 2008 Enterprise has many features

supporting scalability and data warehouse performance that Parallel Data Warehouse

uses to its advantage. The combination of SQL Server scalability and performance with

a massively parallel processing (MPP) architecture in Parallel Data Warehouse creates a

powerful new option for hosting a very large data warehouse.

Parallel Data Warehouse Architecture

Parallel Data Warehouse does not install like other editions of SQL Server. Instead, it is

a data warehouse appliance that bundles multiple software and hardware technologies,

including SQL Server, into a platform well suited for a very large data warehouse. A key

characteristic of this platform is the MPP architecture, which enables fast data loads and

high-performance queries. This architecture consists of a multi-rack system, which paral-

lelizes queries across an array of dedicated servers connected by a high-speed network

to deliver results at speeds that are typically faster than possible with a traditional sym-

metric multiprocessing (SMP) architecture.

Data Warehouse Appliances

You purchase a data warehouse appliance as preassembled and precongured integrat-

ed components with all software preinstalled. When you place an order for an appliance

with an authorized vendor, you specify the number of appliance racks that you want to

purchase. The vendor works with you to add options, such as an optional backup node,

and to optimize the system to meet your requirements for faster query performance

and for storage of high data volumes. The vendor then assembles industry-standard

hardware components and loads the operating system, SQL Server, and Parallel Data