Subject: Workshop on ENTERPRISE-INTEGRATION PRINCIPLES AND FUNDAMENTALS (includes SEMANTIC ISSUES)
On the ICEIMT Initiative
The United States and the European Commission jointly support an international initiative on consensus building in Enterprise Integration. NIST in the USA and the CIMOSA Association in Europe have organizied a series of workshops preceding the ICEIMT'97 conference. Each workshop concentrated on a different topic relevant to enterprise integration with the aim to increase consensus on issues pertaining to the workshop topic. The results of the workshops will be presented at the conference and published in commercially available proceedings.
ICEIMT'97 Objectives: The purpose of the ICEIMT is to develop an international pre-normative consensus among both suppliers and users on a synthesized set of technical issues sufficient to provide a common context for the discussion and comparison of EI technology alternatives. Further, we tried to enumerate the various approaches and characterize them so as to reach consensus on where they overlap, complement, and conflict. Where they conflict, we tried to develop approaches for resolution, potentially leading to collaborative research demonstrations. The workshop reports identify problems and gaps in the current technologies that point to new research. Where applicable, the report refers to relevant standardization efforts aiming on a coherent set of standards supporting enterprise integration.
Workshop 4 Objectives: ENTERPRISE-INTEGRATION APPLICATIONS
Workshop 4 Themes:
Assess the needs in EI and the many solutions currently proposed. This is an introduction to, and definition of, the problems, challenges, barriers, and strategies that exist to resolve them. We can standardize the protocols and interconnections all we want but until we fix the semantic problem we will not be able to communicate sufficiently to accomplish the degree of functionality we need.
Related Themes:
Workshop 4 Format:
June 16: Familiarization--Participants presentations, Development of Expectations
June 17: Direction--Open brainstorming on Alternatives (Workgroups)
June 18: Commitment--Consolidate Workgroup Results, Formulate and adopt plans of action.
Abstract
The focus of the work was on the user oriented semantic issues. The outcome of the work was the extension to the ENV 12 204 standard, which describes enterprise modelling constructs and their inter-relationships, and included the identification of modelling constructs to address human communication. Specifically the extension to the existing standard involved the identification of two additional constructs - human role and business goals and objectives. To further progress knowledge within the domain of Enterprise Integration and Modelling a set of project proposals have resulted from the workshop.
With the advent of the virtual enterprise the common understanding of the nature and composition of enterprise operations across organisational boundaries becomes a prerequisite to initiate and support business relationships. Models of business processes are useful tools to explicitly represent and understand enterprise operations.
Following a description of the foundations of current standards and a classification of current knowledge, the paper describes the basis of the identification of constructs to facilitate the modelling of the communication perspective. Prior to the presentation of future research themes for Enterprise Integration and Modelling, the main conclusions of the work are highlighted.
Proposals
Manufacturing performance is a key issue in business today. Many dimensions of manufacturing performance need to be addressed simultaneously which require new tools and approaches to assist with the desired improvements. Enterprise integration and Enterprise Modelling, as part of the integration process, are key issues to be addressed in the near future. It is indeed perceptible that the industrial use of enterprise modelling languages and tools is far from state-of-the-art achievement. There is therefore an urgent need to bridge the existing gap between actual industrial needs and the research work which has been carried out.
To address this problem, there is a strong belief that an industrial assessment of process models it needed. The industrialists view of the enterprise integration and modelling process has to be further analysed to ensure the operational validity of the research work into standards, reference architectures, business process libraries, and especially modelling languages with appropriate constructs and semantics. We recommend that the ICEIMT'97 initiative should be followed-up with different projects that have a strong emphasis on user needs. These projects should focus on:
Within the scope of enterprise modelling languages, additional efforts have to be made to integrate human aspects and organisational roles with the existing modelling knowledge. To ensure the operational validity of the work these efforts should be progressed by the construction of demonstrator projects using real operational data.
J. Goossenaerts, M. Gruninger, J.G. Nell, M. Petit (editor), and F Vernadat.
Abstract
The paper provides a summary of the discussions held in a working group during ICEIMT Workshop 4 (Enterprise Integration Principles and Fundamentals). The topic of the working group was Formal Semantics of Enterprise Models. In the paper, we suggest that enterprise integration is currently hindered by shortcomings of enterprise modelling languages, that a better language should be defined and that its semantics should be defined in terms of a formal theory. The paper concludes with two proposals for research and standardisation projects.
Enterprise integration still remains a goal hard to achieve for most companies. Part of the problem relies in the difficulty for users of integration technologies to adequately model the problem at hand. The group felt that some shortcomings of current Enterprise Modelling Languages (EMLs) are partly responsible for responsibility for this state of affair.
In current practice, a large number of different languages are used for modelling enterprises. Their comparison is a challenging task because each language possibly provides some concepts that others do not, and because apparently equivalent concepts in two different languages can differ slightly in the details. Moreover, each of these languages has its own syntax be it textual or graphical. More importantly, although they have a clearly defined syntax, most languages do not have a clearly defined semantics.
The result of this syntactic and semantic mismatch between EMLs is that models can not be exchanged between existing tools or if they are, they are not properly understood by their non-native tools.
Proposal 1: Define formal semantics of UEML. The goal is to study some mechanisms to define enterprise modelling constructs with a precise semantics and to define the semantics of a complete UEML.
A formalism suitable for the specification of EML semantics needs to be proposed. New concepts for EM have to be formalised and included in the language such as goals, intentions, non-determinism, co-operation mechanisms, speech act performatives, etc. There are several candidates underlying theories that could be used to define EM ontologies and a choice must be made. An ontological approach to EMLs semantic description seems adequate to formally specify UEML constructs. EM construct definition will be provided by project defined in Proposal 2.
The project is to create a formal description technique for specifying semantics of UEML constructs as well as the semantic specification of UEML (to be used by tool developers).
Proposal 2: A Standard UEMLThe goal is to define a complete and generic UEML on the basis of business user requirements and experience. The language would then be proposed for standardisation to serve as a reference in the field of enterprise modelling.
Consensus must emerge from the users' community over the essential set of language constructs for EM and their syntax. Semantics of the constructs should be shared among users and formally stated. A template description along with a graphical notation for construct definition supported by an ontological approach seems adequate for UEML development. A methodology for extracting UEML users' needs and for expressing UEML constructs has to be elaborated (possibly using use cases).
The projects should deliver a standard core UEML definition (probably an extension of ENV 12204), a syntax for defining libraries, possibly several standard libraries for specific application domain and the definition of the semantics of core UEML and these libraries.
P. Bernus (editor), B. Espinasse, M. Fox, H.T. Goranson,
Abstract
The development of a theory of design for virtual organisations or enterprises has been identified and presented in form of a potential research programme. The required results include a theory that explains the dynamic interactions among partners that create the virtual enterprise and a proposal for the representation of the result. Ensuing would be a methodology specialised to the case of the dynamic creation of enterprises using design transactions, typical models for the enterprise engineering process, typical models for virtual enterprises (to be reused in this process), and the potential extension of the enabling enterprise modelling languages and tools.
This group was assigned the task of identifying a major direction for future development in the area of enterprise integration. The group decided to concentrate its efforts on the problems of virtual enterprise design. We meant by virtual enterprise an association of entities, or partners, formed with the view of together satisfying some jointly agreed-on mission.
The interaction among enterprises with the intention to create a new enterprise is an important study object. Companies which understand these interactions well and can apply them efficiently, would have a definite competitive advantage, through being able to adapt to and utilise new opportunities. This is not only a proportionality issue, of course, but the issue of first to market, or the issue of being able to tender for a project for which the ability to tender was not available before.
Proposals
The research would develop a) a design theory of virtual organisations, and based on that theory b) an architecture and methodology for virtual enterprises, which latter would be a specialised case of GERAM components for virtual enterprises.
Project 1: Development of a Design Theory of Virtual Organisations (Virtual Enterprises)
A design theory of organisations covering virtual enterprises would have to describe the dynamics of the design interactions among participants, and have the ability to describe, analyse and predict relevant system level properties of the designed enterprise.
The theory would include:
Project 2: Development of an Architecture and Methodology of Virtual Organisations (Virtual Enterprises)
Based on the above design theory and design representation and Architecture and Methodology of Virtual Organisations would have to be developed. Such an architecture and methodology would consist of components, such as the ones listed below.
Return to: ICEIMT'97 home page.
Edited by: JG Nell, NIST. Updated 1997-September-25.
Send message to: nell@nist.gov, or kosanke@ipa.fhg.de, or both.