This is a White Paper that was approved by the UDEF Project of The Open Group in May 2011. It describes the approach to distributed UDEF that is currently envisaged by the UDEF project. It is not an approved Open Group Technical Standard.
The Universal Data Element Framework (UDEF) is a standard way of indexing enterprise information that can produce big cost savings and is a primary focus area of the Open Group Semantic Interoperability Work Group. The UDEF was first published in 2006, and subsequently further developed into a tool for semantic interoperability. The central paradigm of UDEF is that it enables an enterprise to describe the meaning of its data elements, independent from the corporate taxonomy and independent from any ontology, and independent of language and representation in a uniform way, such that automated comparison and matching of independent UDEF descriptions of data from different enterprises and different applications is possible.
UDEF consists of a set of object classes by which all concepts can be described in terms of their relation to the describing enterprise, and a set of properties defined according to the UN/CEFACT ebXML standard. UDEF is an implementation of the ISO 11179 metadata standard. UDEF is currently available in English, Dutch and French. Additional languages are being developed.
The UDEF Project is a project of the Semantic Interoperability Work Group of The Open Group. Its aim is to ensure that the appropriate technical, legal, and commercial frameworks are in place for the UDEF to become established as the universally-used classification system for data element concepts. The UDEF Registry Switching Workgroup is a subgroup of the UDEF project that aims to extend the usability of UDEF by defining the architecture for a distributed version of UDEF.
UDEF is the root source of understanding the meaning of any Data Element concept. UDEF however is limited in scope due to:
The distributed version of UDEF introduced in this white paper aims to extend the usability of UDEF by overcoming the above limitations.
Distributed data element trees are a generic solution approach to overcome this limitation. With distributed data element trees the ownership of UDEF-style vocabularies is extended to multiple organizations beyond The Open Group. Distributed data element trees are a distributed set of trees with correspondingly distributed management responsibility that are designed as a consistent and collaborative whole. Each distributed data element tree is managed by a registrar other than The Open Group, but in a way that is consistent with the management of UDEF, such that each distributed data element tree can be combined with UDEF trees in variety of ways. Distributed data element trees are collected in distributed registries that are related to the master UDEF registry in some way, and that can be used in parallel. However not all of the requirements and restrictions for the core UDEF also apply to each distributed data element tree. By extension of UDEF with distributed data element trees the combined set of trees:
It is the vision of The Open Group that the distributed version of UDEF will consist of the core UDEF as it is known today, and a collection of descendent data element trees that will be created to overcome the limitations in the usability of the core UDEF. Below is an overview of different types of distributed descendent data element trees that have been conceived. Each type specifically addresses a subset of the limitations of UDEF.
Types of Descendent Data Element Trees
With distributed UDEF it is possible to construct a much wider range of data element concepts than could be constructed by UDEF alone, in particular:
Distributed trees can be recognized from UDEF trees by the root labeling convention. Whereas all UDEF object and property trees use a numeric as root, distributed trees use as root a numeric with a letter prefixed. The letter is dependent on the type of distributed tree. For each distributed tree type a different prefix letter is used. Distributed trees are uniquely identified for each type of distributed registry with tree identifiers assigned within the context of that type of distribution.
In addition to the above a registry switching mechanism is defined that allows data element concepts to be defined that use a distributed tree as an extension to a UDEF-tree from a certain connection point. This registry switching concept allows objects to be defined by a combination of a UDEF object tree with a distributed object tree, and/or properties to be defined by a combination of a UDEF property tree with a distributed property tree.
The attaching or detaching of a distributed tree to a UDEF-tree is a well understood semantic conversion and is in fact semantically equivalent with specification and abstraction of the generic enterprise context, respectively.
The role of Distributed UDEF in comparison with traditional UDEF can be represented by the Information Continuum, analogous to the Architecture Continuum of TOGAF.
The Information Continuum
Distributed UDEF is a more comprehensive implementation of ISO 11179, but each registry except the master UDEF registry is a more specific registry as defined in ISO 11179.
|UDEF - Universal Data Element Framework.||This represents a standard to implement part of semantic interoperability by means of a set of trees that convey arbitrary parts of meaning and are structured in trees. The UDEF trees are managed by The Open Group|
|SDEV - Satellite Data Element Vocabularies.||These are additional sets of UDEF trees managed by accredited registrars that convey arbitrary parts of meaning restricted to a specific knowledge domain that has been assigned to the accredited registrar by The Open Group|