The power of the HANA-optimized Business Content (SAP BW/4HANA) – Example FI-AA
In this blog post I want to demonstrate you the power of the SAP HANA-optimized Business Content and how it can help to speed up your SAP BW/4HANA implementation.
For this, let us have a closer look together into the business content area of Fixed Asset Accounting – as an example.
Providing Business Content as a part of the EDW solutions has been a differentiator for SAP since the early releases of SAP Business Warehouse (SAP BW). Business Content enables quick and cost-effective implementation as it provides a model that can be used as a guideline during implementation.
The delivered standard data models are intended as a foundation for designing, modeling, and building your EDW, and while often adapted to customer needs, they provide a jump start that facilitates faster and more consistent delivery of EDW projects.
The SAP BW/4HANA Content was specifically designed to leverage the SAP HANA capabilities in SAP BW/4HANA following a consistent EDW reference architecture called LSA++ (Layered Scalable Architecture).
The power of the HANA-optimized Content compared to the classic SAP BW business content in the area of Fixed Asset Accounting
Below graphic shows the ‘classic’ BI Content data flow for the component FI-AA (Fixed Asset Accounting) – as it exists for SAP BW:
Let’s compare it now with the data flow provided by the HANA-optimized business content that best runs in a SAP BW/4HANA system:
At a first glance you might already see the immense simplification:
- Less persistencies
- Less data load processes
- Less dependencies of data loads
- Only ONE InfoProvider for reporting
This not only makes the solution easier to understand and to use. It also drastically reduces the load times inside the SAP BW system (what it means in minutes, we will see later in this blog).
Thanks to some features that were injected into the data model, the report performance was further improved – and this is in addition to the performance boost you will get out of the box by just having HANA as the underlying data base.
To get a better understanding, here a detailed list of the changes and benefits compared to the classic version of the content
Comparison of data load duration
Finally let us compare now what the data model simplification means in regards to the data load times inside the data warehouse system.
Before this, it is important to point out that Fixed Asset Accounting is all about calculating depreciation and appreciation values, and while some of the calculations are happening in the source (SAP ERP or S/4HANA), still a bigger part of the calculation logic is done inside the datawarehouse as part of a SAP BW Transformation (start routine). This makes FI-AA to one of the application components with the highest load times inside the SAP BW system – at least once a month during the period-end closing processes. Therefore reducing the FI-AA data load time might be key to manage the overall load cycle inside the given time window.
Now let us have a closer look at the load times. We have measured it in systems without any performance tuning and based on the same amount of data that we received from the extractors using the same SAP ERP source system.
0FI_AA_11: ~5 million records
0FI_AA_12: ~12 million records.
BW system (Non-HANA) with classic content model
First we processed the data load in a non-HANA SAP BW system into the classic BI content model.
Here the entire load cycle inside BW (!) took 1 hour 21 minutes.
BW on HANA system with classic model
Let’s see what is happening when we only migrate the database to HANA without any functional changes to the model. As part of the data load process there are some steps that are obsolete now when running the same data loads in a BW system on HANA, such as processes to calculate indexes, statistics and roll up of aggregates. Additionally the activation job is now being processed in HANA which reduces the activation time.
At this point we are still using the classic model, but just by changing the underlying data base by HANA, we are getting already a great reduction of the total load time by almost 25 minutes down to 57 minutes.
BW/4HANA with the HANA-optimized content (BW/4HANA Content)
Now we are loading the same amount of data into the new HANA-optimized business content. Here we are experiencing a further massive reduction of the data load duration by further 38 minutes down to now 20 minutes.
The total reduction of the load time was from 1 hours 21 minutes down to 20 minutes. And this includes both type of optimization, the push down of data load processes into HANA and the lighter HANA-optimized data model.
It is important to understand that we did not sacrifice any functionality nor flexibility in the reporting layer. The opposite is true. The report consumer has now more flexibility e.g. drill down to account level for all metrics – something that could not be provided in the classic business content due to the use of aggregates.
I hope this blog post could give you an idea what HANA-optimized content means and its benefits – and maybe also what to take into account when implementing your own data models to fully leverage the enormous capabilities of SAP BW/4HANA.
thanks for the blog post... and what is the use of infosource in bw4hana? do we really need it?
InfoSources in BW/4HANA are not a must but are extremely helpful.
The different use cases for InfoSources are nicely summarized in the SAP BW/4HANA Online Help "Recommendations for Using InfoSources".
The SAP BW/4HANA content data flows for transactional data use one InfoSource to allow connection of multiple source systems while having the business rule in one common transformation between InfoSource and the target DSO.
Advantage: In the case that you would need to change the logic in the business rule, you would need to do so only once.
When the DSO objects 0FIA_DS11 and 0FIA_DS12 can be combined in HANA -Optimized content , but why not the same can be followed in the classice approach ? is it because that the transactions DSO data is in-memory , so the annual value can be calculated from the same DSO on the fly in HANA -Optimized content ?
yes, that's exactly the reason. In the classic content the values dervied from the FI-Transactions were calculated as annual values based on the cumulated values previous years plus transactions of the current year in the transformation between 0FIA_DS11 and 0FIA_DS12.
With HANA this kind of calculation (mainly aggregation of records) is quick and therefore it is now possible to virtualize this calculation. It is done by a set of restricted and calculated key figures used in the analytical queries (https://help.sap.com/viewer/06e872f914a44d77b6c692b0273ca400/2.0.4/en-US/5539c96c1eb04185b56a9eae2e4d38a3.html).
Technically you could have implemented the same approach in BW on any DB - but performance would have been not satisfying at all.
Does this answer your question?
All the best for your project(s), Andreas