GCP on SAP HANA successfully tested in the test laboratory of IM&C

Architecture overview and Proof of Concept

GCP concept

The GCP (Group Costing & Profitability) Engine is an SAP-integrated add-on of the IM&C GmbH, which is able to map global value chains and make them accessible. The use case is to provide cost and revenue structures across systems, company codes and controlling areas in a central system for creating a group cost estimate. In addition to SAP ERP systems also data from non-SAP systems can be integrated.

The integration of data into a central system is of great interest especially for industrial corporations, which organize themselves with their global network of production sites in a production network. The system then typically operates on a global BW system, but without having to necessarily use BI functionality. A GCP content for the use and the integration of group earnings into a consolidated local profitability analysis of the customer is available when needed.

HANA main architectures

There are basically two main architectures at the moment in which a HANA database is used.
Integrated scenarios map applications directly on the HANA DB and take advantage of the new DB technology. In the long run, the most important lever is this scenario. Therefore, however, the base support needs to be provided to access, for example, objects in the HANA DB that promise the biggest performance lever (e.g. Application Function Library or Calculation Views in HANA) directly from an ABAP program.
The so-called side-by-side scenario takes advantage of the higher access speed in a HANA DB in a way that original data is first replicated by a replication mechanism in the HANA DB, from where they can be read much faster. Database triggers are used for replication, because they can replicate data in a signifigantly more performant way than e.g. via RFC or ODBC techniques.
For replication SAP offers the SAP System Landscape Transformation (SLT) server, which supports the trigger-based replication. This method is also available as an add-on for a source system (SAP ERP). At the present time and even in the medium term, the side-by-side scenario is the only option, which can be implemented without HANA-specific coding adjustments.

GCP powered by SAP HANA in the test laboratory of IM&C

GCP triggers the extraction of necessary data (calculation data, transaction data, etc.) from the central system. Extractors in the source systems then provide the data via RFC. With the help of the HANA Application Accelerator, it is now possible to redirect database selections to any other database, e.g. a HANA DB, without program changes. The HANA Application Accelerator is implemented by an SWT2DB add-on.

In the test laboratory of IM&C GmbH, a system landscape for the verification of the described scenario was created. The architecture scheme is described below. The HANA DB is operated as a side-by-side component.

A HANA Appliance (DELL) with all necessary software components was used.
The software version SPS05 was used. The appropriate version of HANA Studio was Revision 48. The BW system was set up as BW on HANA with SAP base level EhP 1 on Netweaver 7.30. Here, the core components for the central GCP system were installed.
The ERP system was set up as a new IDES system (base level 7.31) with Kernel 7.21. In addition to the SWT2DB add-on, both the SLT add-on for source systems and the SLT Replication Server add-on were installed on the source system. The GCP plug-in, which is responsible for data extraction from source systems, was additionally installed.
When placing the secondary database connection for the SLT Replication Server, appropriate entries have been generated in the DB connection table in the ERP source system. These entries were also used by the SWT2DB add-on to redirect data selections to the HANA DB in a defined manner.

The configuration options of the add-on provide the possibility for a context, consisting of table names, main program and job name, to create scenarios that can be activated separately.
It could be demonstrated that the extraction step in GCP can be redirected directly to the HANA DB without program change. Moreover, it was validated that the current GCP release is already running on a BW on HANA system.

Advantages of a “GCP powered by SAP HANA" architecture

GCP typically extracts data at the close of a period, which is then evaluated.
As the data is also transaction data (e.g. sales, consumption, goods receipts, stock transfer, cancellations, etc.), the data acquisition also requires a substantial part of the consolidation process in GCP.

A HANA DB is considerably faster than databases that are not completely stored in the main memory. Even without further optimizations it could be demonstrated that those parts of GCP are getting faster, which are involved with database selection and communication.

Trough the trigger-based replication a replication of data in almost real time is possible. Any loss of speed through double RFC communication is compensated very quickly, especially for very large amounts of data.
In the short run, advantages of this architecture are the following:

There is no code change necessary for achieving significant performance gains.
Customers with a HANA license acquire a license for the SLT server at the same time. Therefore the scenario is a standard replication scenario.
The architecture also supports system landscapes with multiple ERPs.
The basis for the evolution of GCP towards integrated scenarios ("Native HANA") is early set.

Together with the technical advantages there are also direct benefits the business. Through the former availability of group-wide information, group-wide business decisions are faster available, which in turn is the basis for a long-term competitive advantage.
In the medium and long term, there are other possibilities, such as ad hoc reporting and the design of new simulation capabilities. Further performance improvements can also be expected in the applications of GCP subsequent to the extraction.

Outlook and next steps

Based on the status quo, in which GCP is already executable on a central system with HANA DB, further release plans are carried out. It will be shown in the near future, which parts of the SAP Roadmap will be already implemented in 2013. A new ABAP version, which supports the direct use of HANA-specific constructs, is expected to bring not only further performance gains in a side-by-side scenario, but also especially in integrated scenarios where, for example an ERP system is already running on a HANA DB.

If you would like to download this paper, please click here