For SEM BPS, we generally do data loads to pull data from SAP ERP into BW objects to get most recent data.This is required to data validation while writting into the cube. It sometimes becomes cumbursome due to errors in data loads,Load Monitoring, lapse time for data load etc.
We can avoid data lods into BW purpose by directly validating the records using a RFC connection.
Any data which needs validation say Company Code and Cost Centre can be directly done by respective data maintained in SAP ERP through a RFC Function Module. According to the result of the validation, the data will be written into the cube or rejected.
This will enhance the purpose in following way:
- The validation will always be up-to-date as we don’t have to wait for data loads to complete
- Significant amount of time is lost while rectifying the loads
- Sinificant amount of time will be saved as there won’t be any development efforts towards data extraction and modelling
- BW monitoring efforts are not required
- We don’t need to write separte exit Characteristics relationship to validate data/Derive data
- With no Characteristics Relationship, performance will increase
There are few drawbacks in this type of design:
- All validation will require RFC FM development
- Needs expert level of ABAP development
But lookin at the advantages where each time new records are created in SAP ERP at the same instance, BPS application is ready for validation and execution. It doesn’t wait for any kind of BW loads.
According to the requirement design architects can decide what kind of approach to take.