Good Day All,
I would like to share my BPC 10.0 on HANA implementation experience which we have developed the business logic via HANA objects.
Our approach has two main advantages;
- Overall significant decrease on development time
- Huge performance increase
The initial phase of the project was implemented in traditional way which was developing the business logic through optimizing the ABAP codes and parallel processes to maximize the performance and decrease the runtime process of packages. The runtime process was taking too much time due to huge amount of data.
The 2nd phase is implemented with the similar concept and additional simulation logic to calculate results based on changed assumptions, on the fly master data creation and other functional logics specific to the client’s requirement. This phase we decided to change the approach of developing the business logic and calculations by using HANA objects.
The flow illustrated as below;
Performance increased by 20 to 100+ times.
In our calculation layer which is handled on HANA, we process all calculations, security check and model update. The approach to implement in HANA is more complicated. Nevertheless, when we compare between ABAP and HANA the package runtime was significantly reduced and performance increased by 20 to 100+ times.
My first BPC on HANA implementation was in 2012. I have implemented several projects using different ways but this was my first implementation fully utilised HANA Objects. I have faced many problems, errors, memory issues, cursor performance problems but all of them are fixed by using different approach on calculation logic with very late night hours :).
I had different results while using ROW and COLUMN type tables based on my scenario as below;
Test is done with a calculation that generates and inserts 26 million records into a table.
and during testing i found a hidden cat in my package log 🙂