When you create new environments and models, only a small amount of data exists. Since the amount of data you maintain grows over time, we recommend you periodically run the optimize function to improve performance.
There are two different types of optimization available:
• Light Optimization:
Closes the open request, compresses without Zero-Elimination and indexes the cube, and updates database statistics for the BW InfoCube.
Background (ref SAP 1551154)
- Is a maintenance process similar to the BW InfoCube performance maintenance tasks of index rebuild, statistics rebuilds, and compression?
- This process deletes the infocube indexes and moves records from the F to the E fact tables.
- This process compresses/collapses the records with the same key
- Additional zero suppression is possible (manually adjusted in process chain)
- Updates the infocube statistics
How often should it be run? (ref SAP 1508929)
- There is no set recommendation as to the frequency, however it is generally recommended that these (index/statistics) tasks be executed after the load process to provide satisfactory performance for reporting and input schedule updates.
- From 1508929 “There is no rule of thumb for how often to run optimizations. The need can vary depending on the characteristics of your hardware environment and your application.”
As best practice:
- 1. When new application sets and applications are created run a Full Optimization.
- 2. Lite Optimization doesn’t take the system offline, and can be scheduled during normal business activity (e.g. after a data load)
- 3. Full Optimization needs the system to be offline so they should be run at down-time periods – for example after a month-end close.
Background (ref SAP 1551154)
- This process essentially rebuilds the application in its entirety. The process creates a new application and copies the data to the new application.
- This process is a RESTRUCTURING of the BPC application data model.
• Full Optimization:
Will perform the same operations as Light Optimization, but will also check the NetWeaver BW data model. If the data model can be improved, Full Optimization will do so, and this could take a long time to run (for cubes with large data volumes).
Full Optimization will check whether the data model of the BW Cube is built appropriately. Specifically, it will check:
• Dimension tables have < 20% size of fact table
• As many line item dimensions are used whenever possible If the cube structure can be optimized, then it will:
• Take Environment Offline
• Create a Shadow Cube with Optimal Data Model
• Move data to shadow cube, delete original cube
• Close Open Request, Compress and Index the Cube, Update Database Statistics
• Bring Environment back online.
Because this is technically going to be new cube, any changes that had been made to (or surrounding) the underlying InfoCubes will be lost, which means that:
• Data Transfer Processes (DTPs) to/from the Cube
• Aggregates or Business Intelligence Accelerator (BIA) Indexes will be lost.
Both optimization options require that the environment is offline.
Be careful running a Full Optimization, since significant system resources are used in the process.
You can choose to perform either a light optimization or a full optimization.
It is possible to execute the optimization online or via Data Management Package
Model Optimization in MS Version.
Data is stored in models on the following levels:
• Real-time (Write Back table): corresponds to the most current data sent to the system using the input forms.
• Short-term (FAC2 table): corresponds to the data created or loaded using the Data Manager packages.
• Long-term (FACT table): corresponds to the FACT table, which offers better performance.
You periodically optimize your models to improve system performance. The optimization options are as follows:
Optimization Options:
Optimization Options | Description |
|---|
| Lite Optimize | Clears the real-time data storage and moves data to the short-term data storage. This option does not take the system offline. |
Incremental Optimize | Clears the real time and short term data storage and moves data to the long-term data storage. This option takes the system offline. |
| Full Optimize | Runs an incremental optimization and processes the dimensions. |
Compress Database | Sums multiple identical entries into a single record. |
Index Defragmentation | This option forces a reindex of the database. This option may take a long time. |
For example, you can run Lite Optimize every day, and Incremental Optimize every week.
Caution: When you create a new model, it is best practice to run a full optimization.
Note: You can have an automatic reminder to optimize the model. This reminder appears when you log on to the Planning and Consolidation Administration. You can define the number of records (for example, 50,000 records) to be reached before the reminder appears in the Model Parameters.
References:
http://www.claricent.com/
http://scn.sap.com/
BPC410 SAP Business Objects Planning and Consolidation
BPC430 SAP Business Objects Planning and Consolidation
Thanks and Regards,
Saida Reddy.G
