Syndication versus federation – what is the better choice to google your enterprise data?

Large organizations are really challenged with data governance and the need to deal with information silos. Specialized enterprise management applications from different domains like enterprise architecture (EA), business process management (BPM), governance risk and compliance (GRC), project portfolio management (PPM) or enterprise resource planning (ERP) do not make life easier when it comes down to leverage data from all these sources. Usually, the control and exchange of data fails by misinterpretation and weak implementations of interchange specification standards. So basically every organization has to reinvent the wheel by developing and customizing point-to-point interfaces.

And in addition, the struggle of organizations is complicated by the fact that application vendors implement specific subsets or enhancements of descriptive standards. The data representation is typically described by specific frameworks, meta-models and modeling languages. In the EA domain TOGAF® and its modeling language ArchiMate® are prominent; BPMN™ is very important for BPM. However, the specifications may not be precise enough to avoid misinterpretations and vendors have for sure an ax to grind with the design of their own applications.

Information is more important than ever before to find the right decisions in the global market. We can see this from hypes like big data. Data analytics and “googling” the own enterprise data helps gaining a more holistic view and finding the right pieces of data.

There are mainly two concepts in a wider scope of enterprise architecture management to solve this challenge of data governance: data syndication and the federation repository approach.

Syndication is a web-based technology which makes content available for other systems. Examples are the popular Atom or RDF feeds to share content among different systems like web sites and feed readers. In a wider context of EA, the Asset Description Metadata Schema (ADMS; released under ISA Open Metadata License) is a standardized vocabulary developed by the European Union (EU) ISA program. EA work products should remain in the specific tool and be made available solely by syndication. The main benefit is that there is no trouble regarding data exchange if the involved tools support the new ADMS standard for sharing enterprise meta-data.

On the other hand, the federation repository approach facilitates the reuse of EA work products and supports governance processes without requiring content management systems to support a syndication standard. The “universal” federation repository stores all different data sources in a single-point-of-truth, a so-called information warehouse. In addition to the data storage, a federation repository has to manage the taxonomy and ontology similar to the ADMS standard.

Tools like BPM-Xchange® are important to bridge the different concepts and standards and leverage data from all different sources to have a fast and highly functional access to central enterprise data.

A standard positioned in-between syndication and a federal repository are the Open Services for Lifecycle Collaboration (OSLC) specifications, supported by many organizations. IBM® for example supports and implements OSLC in their IBM® Rational® tool family and the IBM® Jazz™ platform. Among others, OSLC supports the application lifecycle management (ALM) by meta-models and exchange formats which can be implemented by tool vendors to support and consume data related to architecture, requirements, quality or change management. The OSLC specifications are well-defined, implying that by its fixed meta-model the scope is more limited (fixed meta-model) than ADMS which has a higher degree of descriptive freedom of meta-models and semantics.

BPM-X is researching and developing concepts to leverage these approaches and support standards like ADMS or OSLC to govern enterprise data in large enterprise management application landscapes.

To be continued…


Latest Comments

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join other 25 subscribers

Archive

© 2017 TransWare AG is a registered trademark.