Fragmentation guidelines

One of the most frequent causes of poor performance in relational database systems is contention for data that resides on a single I/O device. Proper fragmentation of high-use tables can significantly reduce I/O contention. These topics discuss the performance considerations that are involved when you use table fragmentation.

The database server supports table fragmentation (also partitioning), which allows you to store data from a single table on multiple disk devices. Fragmentation enables you to define groups of rows or index keys within a table according to some algorithm or scheme. You can store each group or fragment (also referred to as a partition) in a separate dbspace associated with a specific physical disk.

For information about fragmentation and parallel execution, see Parallel database query (PDQ).

For an introduction to fragmentation concepts and methods, see the HCL OneDB™ Database Design and Implementation Guide. For information about the SQL statements that manage fragments, see the HCL OneDB Guide to SQL: Syntax.