Skip to Content
The subject of data quality is receiving often little attention. Data validation  efforts frequently are reduced to manual checks in the final preparation phase.  Projects with processes and technologies for data validation are still the exception.  However, not just since Sarbanes-Oxley is data quality of critical importance:  How good is business intelligence, if you cannot trust the underlying data? The  opportunities for data contamination should not be underestimated, as the potential  sources are ample: Incomplete or corrupt source data; customizing errors; software  bugs (application or database); and even hardware errors can lead to data inconsistencies.

It is not, that the businesses and implementers are not aware of the challenge.    It is rather a lack of accountability, as well as of methodology and tools (or    the knowledge of those) for validation. There is an obvious lack of standards,    how and what to check. I see three primary reasons for the lack of standards:

     

         

    1. IT landscapes look quite different at every implementation (even when        the same solutions are deployed).
    2.    

    3. Data quality standards vary by each organization.
    4.    

    5. Focus of data criticality changes often and dynamically (e.g. while accuracy        of finance data is always important, the criticality peaks at the time of        financial closing). Given all the above, it is clear that a rigid implementation        of data quality checks would be expensive and inflexible, and its ROI often        questionable.
    6.  

A service-oriented architecture (SOA, or ESA) deals effectively with those    challenges, and hence serves as an excellent infrastructure for data quality    assurance. While not as ‘sexy’ as the average composite application, its    business benefits in this context are immense, and quite obvious. Let me quickly    lay out the building blocks:

  image

     

         

    1. Data check points (‘hotspots’) exposed as web services throughout the        system landscape. In its simplest case, it could be ‘value of purchase order’,        or ‘sales for product xyz in current month’.
    2.    

    3. A central registry for those services.
    4.    

    5. Intelligent (web) services to support data checks. In its simplest case,        a service to compare the value of the same purchase order, from two systems.     
    6.    

    7. A graphical modeling (business process management) tool to deploy and        monitor those services in a very flexible manner.
    8.    

    9. A methodology for deployment.
    10.  

SAP NetWeaver could serve as an excellent platform for realization. BAPI’s,    RFC’s, XML Queries are being exposed as web services. SAP web services, as well    as external web services, are registered in WAS’s UDDI. XI BPM allows the flexible    graphical modeling of processes. Finally, Alert Framework and EP are tools of    choice for result visualization.

To report this post you need to login first.

Be the first to leave a comment

You must be Logged on to comment or reply to a post.

Leave a Reply