Home |
Previous Page | Next Page Planning Table and Index Fragmentation >
In This Chapter
Fragmentation
of tables and indexes across coservers and disks is at the heart
of good DSS query and OLTP transaction
performance. Appropriate fragmentation can provide the following
benefits:
- Improved parallel processing for DSS applications as well as
improved serial plans for OLTP applications
- Fast data loading and updating
- Reduced I/O contention for data
- Faster simple select, update, and delete transactions
Important:
Plan and test fragmentation schemes carefully before you implement
them in a production database. To change the fragmentation scheme
of a table, you need to unload and reload all table data. You might
also need to revise applications that access the table.
If you are migrating an existing database from a different database
server, for best performance evaluate and adjust your database schema
to take advantage of unique XPS fragmentation methods before you
load data into the new database server.
This manual does not provide information about constructing a
database schema. For basic information about database design, refer
to the IBM Informix: Database Design and Implementation Guide.
This chapter provides information about the following topics:
- Identifying fragmentation goals
- Designing appropriate fragmentation schemes for permanent tables, temporary
tables, and indexes
- Planning storage spaces for fragmented tables
- Verifying fragmentation efficiency
Home |
[ Top of Page | Previous Page | Next Page | Contents |
Index ]