I was asked to append some information on scaling a database system. Planning and design of a database system should include methods for scaling  - upwards and outwards. The requirement to scale could be due to various factors , examples are; (a) change of focus on data gathering , which leads to different size of datasets b)business acquisition  which requires integration with other systems.

Some features Ive found useful in SQL Server for scaling are:

-Partitioned Views

-Service Broker

-Services Oriented Database Architecture(SODA)

-Compression

-Data Dependant Routing (DDR)

-Table Partitioning

A Partioned View splits data from a larger table into smaller tables (member tables).The partitioning occurs via a range(s) of values in a defined column horizontally.

Service Broker allows loosely held tasks to work together  as an asynchronous application.

SODA  Large centralised databases are still key  but think of  multiple large databases that are repositories for different classes of data

Compression  used with Read Only file groups  , useful in databases with ultiple filegroups

DDR  application intelligence#development is focused on middle tier , to decide on which databases to route to. Note: this is different from views across multiple databases.

Table partitioning (& Index Partioning)  Spread tables and indices across multiple file groups in partitions .This allows flexibility in spreading across multiple devices. Useful in assisting archiving strategy

Ref:Jack Vamvas (http://www.sqlserver-dba.com)