Creating and Scheduling a Reorganization Schema 
Use
You can create your own reorganization schema to define the period of time over which performance data should be retained in the performance database. After this period, the data is flagged for deletion and deleted when the reorganization is next run. You can also determine whether or not to retain data after it is flagged for deletion, in case you want to access this data at a later date.

SAP's default reorganization schema is SAP_DEFAULTSCHEMA. The performance database is reorganized once a week, when the basic data that is reported to the monitoring architecture is aggregated and summarized into weekly data. The basic data is then deleted from the database, so that only the compressed data remains.
To create an reorganization schema, you must execute two programs,
RSALPREORG and RSALPFPOLICY to populate the tables that contain the schema itself. The reorganization schema is run as a background job, which the system administrator also needs to schedule.
The system automatically schedules the background job. You can check its progress by calling Transaction SM37 or by choosing System ® Own jobs.
Procedures
To create a reorganization schema

Program
RSALPREORG populates the table that stores the system ID, MTE class and the schema ID. This enables you to assign a different reorganization schema to different MTE classes in different systems.
You may still want to access this data even after it has been flagged for deletion. In this case, enter the number of additional days, weeks, quarters or years that the data should be retained. For example, if you specify that the performance data should be kept for 15 days but deleted after an additional 5 days, the data is deleted after 20 days.
To schedule the reorganization schema as a background job

You now have to schedule when the performance database reorganization will take place. The reorganization is scheduled as a background job. You can either define a time and date for the job, or change the date and time that has already been defined.
There are two steps to this background job: the first is to define the frequency of the data collection, the second is to define the date and time of the reorganization.
See also:
Background Processing