Show TOC

InfoCube with Data Persistency in the BWALocate this document in the navigation structure

Concept

An InfoCube with data persistency in the BW Accelerator only saves transaction data on the BWA server and not in the database of the BW system. In other words, the data of the F fact table, E fact table and dimension table is stored in the BW Accelerator. The master data is saved in the BW system and in the BW Accelerator. This means that the data is available to other InfoProviders, as is usually the case.

The InfoCube with data persistency in the BWA offers the following advantages over standard InfoCubes:

  • Avoid data redundancy in the BW system: This allows you to save memory space and reduce the load on the system, which improves database performance.

  • Relocation of data analysis layer: The data analysis layer is relocated from the BW system to the BWA. This results in improved performance for data analysis.

  • You do not need to create aggregates.

The Data Warehouse layer remains in the BW system and provides a stable data basis.

Note

Note that a backup and restore process is not implemented on the system for InfoCubes with data persistency in the BWA.

Structure

The InfoCube definition is saved in the BW system. An F fact table and the package dimension are also created in the BW system. The F fact table is needed to derive the definition of the BWA index. The package dimension is needed to manage requests in the BW system. This is the only table that is filled in the BW system, if transaction data is loaded into the InfoCube.

The fact index of the InfoCube is flat. In other words, the master data IDs are written directly to the fact index without the need for dimension indexes. This can lead to improved performance for data loading. The only exception in this reduced star schema is the package dimension index in the BWA.

This flat fact index offers advantages for modeling: You do not need to create the dimensions according to performance criteria. Instead you can use the dimensions for semantically structuring the InfoObjects.

Integration

You use the data transfer process (DTP) to load data. There is no restriction for any of the possible sources. The DTP writes the data directly to the fact index in the BWA. Here the DTP uses a report that was generated when the InfoCube was created. You can create several DTPs for one InfoCube with data persistency in the BWA. However the DTPs cannot run in parallel due to technical reasons.

You can use the BW Accelerator Monitor for the consistency check.

More information: Monitoring BWA Indexes