Partitioning your tables in Microsoft SQL Server

Learn how partitioning your tables can increase SQL Server performance.

This Content Component encountered an error

There are many good reasons for partitioning the data in your database into smaller, separate tables. The reasons include: improved performance, reduced backup windows, better data protection and support for special needs of a particular set of users with their own data set.

Partitioning lets you work with a smaller group of records or locate a data set in an appropriate location or on the best server to suit your needs. Although you essentially partition your data in memory anytime you create a view, that set of records competes for I/O with the larger set of records. So although you can achieve a greater degree of normalization by minimizing the number of tables, there are advantages to be had when you de-normalize your tables.

Most database designers learn quickly that heavily used static data sets are candidates for isolation in a lookup file. Lookup files work best when they are narrow and long (a vertical partition) and especially when they operate against an index. Lookup files work less well for data placed into a wide and shallow table (horizontal partition). Creation of lookup files lets you isolate records that have high transaction traffic, which also lets you protect this data and locate the data in one or more locations. As a rule, heavily accessed data is most susceptible to corruption due to software errors, hardware errors like disk hot spots or user errors.

Heavily accessed data that changes frequently is also a good candidate for partitioning. The loss of efficiency due to the additional overhead associate with extra tables is more than offset by your ability to optimize the operating environment for the data set. It's advantageous to place your most dynamic data on your fastest equipment and in the best protected environment; as well as locating that data set as close to the source of transaction as possible – as you might do for a remote office. When partitioning a database, consider adopting a physical topology that has a hierarchical storage structure.


Barrie Sosinsky is president of consulting company Sosinsky and Associates (Medfield MA). He has written extensively on a variety of computer topics. His company specializes in custom software (database and Web related), training and technical documentation.


This was first published in March 2005

Dig deeper on Microsoft SQL Server Performance Monitoring and Tuning

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchBusinessAnalytics

SearchDataCenter

SearchDataManagement

SearchAWS

SearchOracle

SearchContentManagement

SearchWindowsServer

Close