Data Warehousing in an SAP-Centric Environment

March 1, 2010 by Rajesh Mohanasundaram 

Source :http://www.ironsidegroup.com/

 

SAP-R/3 is one of the most common ERP applications for manufacturing industries.  The data warehouse for such installations has been Oracle.  The biggest pain point for Oracle implementations has been getting the data out of SAP-R/3, which involves complex processes and expert domain knowledge.  With the introduction of SAP-BW, the data warehouse suite, ’Getting the data out’ of SAP-R/3 requires little effort.  Over years, SAP-BW has matured to handle large data volumes and meet the service level agreements for large enterprises.  The question then arises, especially for those organizations with an Oracle data warehouse, if moving to SAP-BW makes long term sense. There are several factors to be considered, the core being organizational commitment and long term vision.  One of the factors is understanding the differences between SAP-BW and Oracle within the context of an enterprise data warehouse.

Advantages of SAP-BW

SAP-BW simplifies data integration with SAP-R/3.  Standard extractors are available by subject area to get the data out of SAP-R/3 at the click of a button.  The standard extractors can also be copied and customized, if required, to suit the business needs.  This solves the biggest problem that many companies struggle with – getting the data out of SAP-R/3. Many concepts required in the data warehouse environment are automated.  Slowly changing dimensions can be configured in a few clicks.  Hierarchy management is simplified especially with recursive hierarchies.  Currency and Unit conversion are managed well.  Re-usability of objects is inherent due to info-objects – structures that store data.  An info-object, particularly a master data type called characteristic info-object, allows encapsulation of attributes and hierarchies associated with the entity.  They can be reused anywhere including in other info-objects.  The data and properties are not replicated so data redundancies can be greatly avoided. Tuning options are available including the hardware performance booster called BW Accelerator.

Shortcomings of SAP-BW

The core of SAP-BW is an ABAP engine like any other SAP product.  While it is configurable and data movement within the system is easy to implement, manual access of data is difficult.  Quick and dirty data analysis that was possible in Oracle using SQL becomes a time consuming task. This also means that data cleansing and data profiling efforts are complex. If the data sources are disparate (non-SAP) then SAP-BW offers no specific advantage since the data transfer is done via flat files or ETL tools, which is no different from Oracle. Talented SAP-BW resources are hard to find.  Even with consulting companies an expert is hard to get so there is an inherent skepticism among the IT executives for such initiatives.

Advantages of Oracle

Oracle has been a proven platform for data warehouse implementations for many years.  The optimizer is the best in its class.  Large data volumes can be handled with tuning options like partitioning, bit-map indexes, and materialized views with query re-write. The governance can be controlled at a system level. For example:  tables can be created only by a DBA so developers cannot create tables to satisfy only their needs.  This prevents data proliferation.  In contrast, in SAP-BW, the developers can create info-objects leaving the doors wide open for data proliferation – a huge problem in many SAP-BW implementations. Because of the ease of data access, data cleansing and integration operations are robust, especially with disparate data sources.  Tools specific for these purposes are widely available for Oracle platform. Resources for Oracle are readily available at cost effective rates – a key factor in resource planning.

Shortcomings of Oracle

Fundamental concepts like surrogate keys, slowly changing dimensions, and hierarchy management must be implemented manually.  There is a risk in losing consistency during development.  There are no standard extractors to get the data out of SAP-R/3.  It is dependent on ETL tools like Informatica/Infosphere for data loading.  This is not a major concern but it means additional specialist resources are required for ETL development. A successful implementation of SAP-BW relies heavily on strong data architecture, governance and project management teams with solid experience in this area.  As a result,  relearning is required to operate in an SAP-BW environment, especially if the resources are from the Oracle data warehouse domain.  By taking into consideration all the above factors, SAP-BW with its rich tools, can be turned into a powerful enterprise business intelligence platform.   The overall vision should consider the data quality.  Should there be data quality and integration issues then Oracle offers many advantages over SAP-BW.  There is no perfect answer between choosing Oracle or SAP-BW for data warehouse solution – both have their share of advantages and shortcomings. Understanding them within the context of an organization’s commitment and long term vision are key factors in making the decision.