Introduction

Partitioning is used to enable dividing up large quantities of data into smaller parts, to manage them easier. Tables and indexes can be partitioned so parts of them can be imported, exported, and reorganized independently of one another. Each of these partitions can be assigned to a certain table space. Another advantage is the possibility to observe runtime behavior for the access to the partitioned tables. Depending on the query in some circumstances no access to the entire table is necessary but only to parts of it. Database partitioning is a two-step process:

1. Assign Key

   The IntegrationService assigns a partition key to every document, folder entry or AuditTrail event stored in the system. This partition key can have at most 50 characters and is constant over the lifetime of the object. The configuration for this step is described in chapter Partitioning configuration.

2. Map Key

   The database maps this partition key to a physical partition where the object will be stored. Multiple partition keys may be mapped to the same physical partition. This step is configured directly in the database via the ImageMaster Upgrade Tool, see chapter Partitioning strategy for more details.

Note the following:

• Support of partitioning should be considered before installation because later partitioning requires to copy the database.

• It is recommended using table partitions which are smaller than 2GB (for Oracle instances).

• Always check beforehand, if a license for partitioning by the database provider is required, and if so, make sure that this is available for your environment.