To get the best possible performance out of any database, it and the applications it supports will occasionally require some sort of modification and tweaking. The process of attempting to improve how well a database functions is commonly referred to as database optimization. This task generally involves breaking the database down into different components and separately analyzing the operation of each one.

The first step in database optimization is taking inventory of everything that is currently contained in the database. This could possibly include recording information such as the number of tables or indexes in the database or how many users or applications touch the database. This phase is known as benchmarking.

Once the database administrator (DBA) has a good starting point for what is normal for his or her system, he or she can start optimizing. Perhaps the best place for the DBA to start is with the data components. For optimal performance, the DBA will want to examine all the tables and indexes and remove any unused or outdated data.

Rapid Retrieval
The primary function of any database is to store and then retrieve data. Thus for optimal database performance the speed at which the information is accessed should also be optimized. This can be accomplished by reducing the number of I/O (input /output) reads or by investing in faster I/O subsystems.