Data Version Management in LSA++

See more in our course BW4HANA – An Introduction

How much data needs to be archived in BW system depends on whether the data should be reproducible and whether ‘single version of truth’ is required. A corporate memory is used to ensure historical data is accessible. This is also applicable for master data of course. Reporting master data which is older is achieved by including time slices in advanced dso.

For global BW systems, LSA++ should have standardized semantic/logical partitioning of transaction data.

Semantic partition means to divide data based on various time zones involved e.g Asia Pacific, Europe and Americas.

Standardized means all advanced DSOs follow the same partition criteria in BW system.

This partitioning can be based on organizational characteristics like sales organization or controlling area.

This semantic or logical partitioning should not change quickly with time and should remain stable.

The parts or process areas created in BW in this way are called as domains in LSA++

To manage these partitioned data flows, infosources play a major role for easy definition and maintenance in transformations.

Using an inbound infosource before the infoprovider and outbound infosource after the infoprovider, we can bundle the logic.

You can also use field based infosources where creation of infoObjects is not required.

The transformation logic is then between the outbound InfoSource of the taarget InfoProvider and the inbound InfoSource of the source InfoProviders.

Note that there will be no transformation between Infosource and infoprovider (See figure below)

Also note that corporate memory is not split in domains.

This partitioning of data and creation of domains fall in data propagation layer in LSA++