Master data – Everyone Needs it, but No-one Wants to Maintain it (2)

Part one
SAP MDM creates a complete environment for the integration of business processes in heterogeneous system landscapes. The solution also supports cross-company cooperation by ensuring the consistency of data – regardless of the location or application manufacturer. SAP MDM can be implemented step by step. Customers do not have to forfeit all of their previous work towards data harmonization in one “big bang”. This reduces downtime and interruptions in day-to-day work to a minimum.

The components of SAP MDM
The components of SAP MDM

The main components correspond to the Enterprise Service Architecture (ESA), which enables existing IT investments to be integrated and reused. The integration and application platform SAP NetWeaver makes it possible to integrate existing partial solutions based on Web services and SOAP into a new, total solution.
SAP MDM consists of the SAP Exchange Infrastructure, the Content Integrator, the Master Data Server and SAP MDM adapters for connecting SAP solutions and applications from other providers.

Bringing master data objects to their destination

The integration solution SAP Exchange Infrastructure (SAP XI) enables the process-controlled exchange of data between any applications in a heterogeneous IT environment. SAP XI goes beyond the conventional, purely technical EAI (Enterprise Application Integration) approach, because in addition to the technical integration information it also manages the cross-system business semantics, such as business scenarios, interfaces, or transformation and distribution rules.
In SAP MDM, SAP XI simplifies the integration of the master data. SAP MDM uses SAP XI during the distribution of data in order to transfer business objects to predefined, local applications. While SAP MDM is responsible for the definition of business objects and their ongoing maintenance, SAP XI offers the required functions for routing, queuing, and converting the data. These ensure that the objects reach the right destination.
SAP XI and SAP MDM are linked with each other by means of publish-and-subscribe mechanisms. SAP MDM uses open protocols and standards such as HTTP, XML and XSL to send the master data earmarked for distribution to SAP XI (publishing). A central routing model within SAP XI stores information about which systems in the IT landscape request access to this master data (subscription). When new master data is received, SAP XI assigns the relevant values to the objects, which are identified by a different key in each system. In addition, the solution converts semantically identical data types that have a different syntax or structure – for example as a result of different date formats such as or mm.dd.yyyy. Once the data has been prepared in this way, SAP XI distributes it to the relevant target systems. This is done with a queuing mechanism which ensures that the messages are sent correctly and just once.

Object comparison using unique characteristics

The Content Integrator (CI) links the master data objects in different systems by means of qualifying attributes that are used across the whole enterprise. This enables data objects to be compared using user-defined rules. For the “business partner” object, for example, a hit list based on the sales tax ID could be defined. In this case, the Content Integrator would find all objects with the same sales tax ID, and the hit list would initially be restricted to all business partners that belong together from a legal perspective. On the basis of this, additional rules could be defined that limit the hit list further. The Content Integrator can compensate for minor spelling or typing errors because it supports phonetic algorithms: individual short descriptions are broken down into phonemes, which then serve as the basis for the comparison (for Release 3.0). In this way, the Content Integrator identifies master data objects that belong together in the various applications, and removes duplicates.
The ID assignments that result from this comparison are used to distribute the data and simplify the creation of cross-system reports and analyses. The Content Integrator also contains an enterprise-wide search function which can be used to determine in which application a particular master data object is located.

The master data server is the heart of SAP MDM

The master data server (MDS) is the heart of SAP MDM. As the central processing unit, this master data server manages the master data in the whole enterprise. The MDS is tightly integrated with the Content Integrator and SAP XI. The information that arises in the Content Integrator about object relationships is automatically stored on the MDS. The SAP Exchange Infrastructure is the central data platform via which objects are loaded to the MDS or distributed from the MDS to the local applications. Data is processed on the MDS alone, whereas it can be created and maintained both on the server and in the connected applications. SAP divides the tasks and functions of the MDS into the object layer, the service layer, and the processing layer:

  • The object layer describes the master data objects in flexible, extensible form, and forwards them for further processing to the Content Integrator and to SAP XI. It provides predefined object types, which can be added to in specific applications.
  • The service layer offers generic services and procedures for managing master data. These include object creation, change and status management, queries, routing maintenance, authorization, workflow and cooperative data cleanup. If required, these services are also available as Web services.
  • The processing layer controls the distribution of the master data in tight integration with SAP XI. In addition to the conventional distribution of individual objects, the objects can also be put together from groups and distributed in accordance with the business context defined in the MDS. The publish-and-subscribe mechanisms in SAP XI are also equipped with intelligent subscription functions, which send the data objects to the local applications on the basis of predefined business contexts.

Adapters create connection points for applications

In a heterogeneous IT landscape, master data can only be managed correctly if every application is connected by means of an SAP MDM adapter. This adapter receives all the generated master data and unifies it in a common data format which allows SAP MDM to work with all business-critical applications – regardless of their manufacturer.
As the master data attributes can be adjusted to particular requirements at local level, SAP MDM adapters are able to transfer business-relevant data from the local application back to SAP MDM with the help of bi-directional assignment functions. This ensures the consistency of the data objects. The assignment functions work both with the search and extraction functions at local level and with the SAP MDM services for context searches and technical distribution.
For SAP systems, SAP MDM will receive the appropriate adapters with the full functional scope. Tools and a development framework are available to cover the range of non-SAP and legacy systems. In addition, SAP is currently collaborating with a number of partners towards embedding a comprehensive adapter system within SAP MDM.
SAP MDM supports open standards such as SOAP, HTTP, XSD, WSDL or BPEL4WS at all levels. In addition, every component of SAP MDM is scaleable, as it is based on the SAP Web Application Server and therefore several instances can be installed.
Comprehensive master data maintenance with SAP MDM lowers the cost of managing master data on the one hand, and on the other leads to a reduction in the amount of inconsistent or old data – which in turn reduces the need for costly troubleshooting. Initial experiences by customers have shown, for example, that the cleaning up of duplicate data records generated the highest costs. If duplicates need to be cleaned up with the existing transaction history, the duplicate data record cannot simply be deleted, as value-flow-relevant information is linked to the master data object. In individual cases, data cleanup can cost around 1000 Euro per duplicate. In addition, a company benefits in a multitude of ways from the consistent data: business processes can be carried out more quickly, cross-company collaboration is made easier, and analyses are based on accurate information.

Tobias Wittebrock
Tobias Wittebrock