Enterprise Modelling and Integration

From Fact Modelling to Enterprise Interoperability

Francois B. Vemadat
EC/EUROSTA T. Luxemburg & LGIPM. ENIM/University ofMetz, France,
Francois. [email protected]

Abstract: Enterprise Modelling and Integration has evolved over the last decades from
entity-relationship and activity modelling to object and flow modelling as well
as from pier-to-pier system integration to inter-organisational exchanges ena-
bling various forms of electronic commerce. The next challenge is Enterprise
lnteroperability, i.e. seamless integration in terms of service and knowledge
sharing. The paper discusses modelling and integration issues to progress to-
wards Enterprise lnteroperability and shows how the CIMOSA architecture
can be revised to host these emerging techniques and standards.

1 INTRODUCTION

Enterprise Modelling (EM) is the art of extemalising enterprise knowl-

edge, which adds value to the enterprise, be it a single enterprise, a private or
government organisation, or a networked enterprise (e.g. extended enter-
prise, virtual enterprise or smart organisation). Enterprise Integration (EI)
deals with facilitating information flows, systems interoperability and
knowledge sharing among any kind of organisation. Enterprise Interopera-
bility, as one of the many facets ofEI, provides two or more business entities
(of the same organisation or from different organisations and irrespective of
their location) with the facility to exchange or share information (wherever it
is and at any time) and to use functionalities of one another in a distributed
and heterogeneous environment (Kosanke, Nell, 1997, OAG. OAGIS, 2001,
Petrie 1992, Vemadat, 1996).
With the emergence of A2A (application-to-application) and X2X tech-
nologies in business (B2B: business-to-business, B2C: business-to-customer,

The original version of this chapter was revised: The copyright line was incorrect. This has been
corrected. The Erratum to this chapter is available at DOI: 10.1007/978-0-387-35621-1_43

K. Kosanke et al. (eds.), Enterprise Inter- and Intra-Organizational Integration

© IFIP International Federation for Information Processing 2003
26 Vernadat, F.B.

C2C: customer-to-customer ... ) as well as in governments (G2B: govern-

ment-to-business, G2C: government-to-citizen, G2G: government-to-
government, G2N: government-to-non government organisations), there is a
need for sound and efficient methods and tools to design and operate effi-
cient integrated systems made of autonomous units.
In this context, EM provides a semantic unification space at the corporate
level where shared concepts can be properly defined, mapped to one another
and widely communicated in the form of enterprise models (Goranson,
1992).
This position paper first briefly reviews the current state of EM and EI
and then probes the future in terms of their evolution before indicating how
the CIMOSA framework can be revised to cope with these evolutions.

2 ENTERPRISE MODELLING& ENGINEERING

What it is: Enterprise Modelling is concerned with representing the

structure, organisation and behaviour of a business entity, be it a single or
networked organisation, to analyse, (re-)engineer and optimise its operations
to make it more efficient. Enterprise Modelling is a crucial step both in En-
terprise Engineering and Enterprise Integration programmes (Vernadat.
1996).
Enterprise Engineering (EE) is concerned with designing or redesigning
business entities. It concerns all activities, except enterprise operation, in-
volved in the enterprise life cycle, i.e. mission identification, strategy defini-
tion, requirements definition, conceptual design, implementation description,
installation, maintenance and continuous improvement as defined in PERA
and GERAM (IFAC-IFIP Task Force, 1999, Williams 1992). It mostly con-
centrates on engineering and optimising business processes in terms of their
related flows (materials, information/decision and control), resources (hu-
man agents, technical agents, roles and skills) as well as time and cost as-
pects. EM techniques for EE should therefore support at least representation
and analysis of function, information, resource and organisation aspects of
an enterprise (AMICE, 1993, IFAC-IFIP Task Force, 1999, Vernadat. 1996).
As advocated in the Zachman Framework (Sowa, Zachman, 1992), the
objective of EM is to define the six perspectives of what, how, where, who,
when and why of the Enterprise Model, System Model, Technology Model
and Component level of an enterprise. The what defines entities and relation-
ships of the business entity, the how defines the functions and processes per-
formed, the where defines the network of locations and links of entities and
agents, the who defines agents and their roles, the when defines time aspects
Enterprise Modelling and Integration 27

and the schedule of events, and the why defines the strategy of the enter-
prise.
What needs to be modelled: The following aspects are concerned
(AMICE, 1993, IFAC-IFIP Task Force. 1999).
- Function aspects: functional domains, triggering events, business
processes (or control flows), enterprise activities (or process steps)
- Information aspects: enterprise objects, object relationships (semantic
and user-defined links), object flows, object states
- Resource aspects: doers (human and technical agents), resource com-
ponents, resource capabilities and/or competencies, roles
- Organisation aspects: organisation units, organisation cells (or deci-
sion centres), responsibilities, authorities
- Temporal and causal constraints
These are the usual modelling constructs found in prominent EM lan-
guages (ARIS, CIMOSA, GRAI, IDEF, IEM ... ) as reviewed in (Vemadat.
1996).
What for: The enterprise models must provide abstract representations
of the things of the organisation being analysed with enough precision and in
a way which lends itself to computer processing to support:
- Enterprise Reengineering I Process Improvement (establishing the
business-process map, simplifying and reorganising some processes,
optimising use of resources, simulating enterprise behaviour)
- Workflow design and management (to automate critical processes)
- Tuning enterprise performances (mostly in terms of costs and delays
but also quality, reactivity and responsiveness)
- Management decision support ("what if' scenarios, simulating
planned situations, forecasting, etc.)
- Enterprise integration (i.e. seamless exchange across the system to
provide the right information at the right place at the right time)
Enterprise Knowledge Management: Enterprise modelling is a form of
enterprise knowledge representation method in the sense that it captures,
represents and capitalises basic facts and knowledge about the way the en-
terprise is structured, organised and operated (mostly surface knowledge).
According to G. Mentzas, Enterprise Knowledge Management (Tiwana,
2000) is a new discipline for enabling individuals, teams and the entire or-
ganisation to collectively and systematically create, share and apply corpo-
rate knowledge to better achieve organisational efficiency, responsiveness,
competency and innovation. Thus, there is a need to also address deep
knowledge.
Within an enterprise, knowledge is exhibited at various levels. It is in the
mind of people (individual level), within team structures (team level), encap-
28 Vernadat, F.B.

sulated in business processes and rules (organisational level) and linked to

inter-organisational interactions (environment level).
Knowledge is usually classified as tacit (formalised as a theory or ex-
pressed in a structured language/notation) or implicit (individual feeling or
known by humans but not formalised in a theory or in a structured model).
Nonaka has proposed a cyclic model of knowledge emergence and con-
solidation within an organisation (Tiwana, 2000). The model is a cycle made
of four steps: socialisation (tacit know-how becomes shared know-how),
externalisation (shared tacit know-how becomes codified knowledge), com-
bination (codified knowledge becomes enterprise knowledge), and internali-
sation (enterprise knowledge becomes individual tacit know-how).
Evolution of Enterprise Modelling and future issues: The origins of
Enterprise Modelling can be set back to the mid-70's when several dia-
grammatic methods were proposed for information system analysis and
software development. The early methods can be qualified as fact modelling
methods in the sense that little or poor semantics of the enterprise was cap-
tured. Pivot concepts taken into account were the concepts of enterprise enti-
ties, relationships among entities and activities made of sub-activities. The
models produced only represent static facts. Pioneering methods are the en-
tity-relationship model of P.P.S. Chen and the SADT method of D.T. Ross,
also known as IDEFo, (Vemadat, 1996).
They were soon followed in the 80's by flow-charting methods combin-
ing ideas of the two previous ones but in addition depicting the flow of proc-
essing activities (SSAD by Gane and Sarson, Yourdon's notation,
DeMarco's notation, MERISE in French spheres) (Martin, McClure, 1985).
For CIM, IDEF and GRAI methods appeared (Vemadat, 1996). Time as-
pects were missing in such models.
At the same period, a lot of more fundamental work was carried out on
(1) semantic models (e.g. extended entity-relationship model, semantic net-
works, frames, binary model) to capture more of the semantics of data or for
knowledge representation, and (2) onforma/ models to analyse system be-
haviours (e.g. Petri nets, timed Petri nets, coloured Petri nets, state-charts).
The 90's have been dominated by two complementary trends, which have
seriously impacted and boosted EM: business process (BP) modelling and
object-oriented (00) modelling. BP modelling focuses on business processes
and related concepts: events, activities, roles, resources and object flows.
Many of the common EM tools and approaches have emerged from this
trend (CIMOSA, IDEF3, ARIS, IEM and the workflow technology). 00
modelling focuses on the abstract concept of objects and brings structuring
modelling principles, e.g. object uniqueness, property inheritance, aggrega-
tion mechanisms, and reusability. The prominent method in the field is UML
(Unified Modelling Language), which has become an OMG and ISO stan-
Enterprise Modelling and Integration 29

dard and has supplemented OMT (Object Modelling Technique) (ISO/IEC

DIS 19501-1, 2000).
Current modelling tools are quite good at modelling structured business
processes, i.e. deterministic sequences of activities with related object flows
and associated resources (e.g. ARIS Tool Set, FirstSTEP, etc.). However,
they need to be extended in several ways. Among these, we can cite:
- Socio-organisational aspects: More research work and extensions to
commercial tools are required in terms of modelling human roles, in-
dividual and collective competencies, decision centres. To this end, a
competency model has recently been validated in industry and pro-
posed to extend CIMOSA constructs (Berio, Vemadat, 1999, Harzal-
lah, Vemadat, 2002).
- Weakly structured workflow: Structured business process and work-
flow system implementations tend to rigidify the enterprise, i.e. to
automate processes in an inflexible way. Modem tools should be able
to cope with weakly or ill-structured processes, i.e. processes for
which the exact control flow sequence is not fully known. Three es-
sential constructs have been proposed to this end but not yet imple-
mented in commercial tools: AND construct (the process step is made
of n activities that must all be done but the execution order of which
will be decided at run-time), XOR construct (there are n activities in
the process step but only one will be executed, the choice of which
will be decided at run-time), and the OR construct (k among n activi-
ties will be done in the process step at run-time but the selection will
be decided at run-time) (Berio, Vernadat, 1999). Another interesting
problem concerns the modelling of the decision knowledge associated
to each case, which is also a research issue (El Mhamedi, et al, 2000).
- Inter-organisational Interaction and Co-ordination aspects: The
modelling of networked organisations and supply chains requires that
new constructs be proposed to cope with such structures.
- EM ontologies: Because there are different ways of representing the
same concepts, there is the need to have an ontology of enterprise
modelling concepts (specialised by industrial sectors, application do-
mains, tools, and so on) (ACM, 2002). Examples of such ontologies
for enterprise modelling are the TOVE ontology (Fox, Groninger,
1998) or the ontology for PSL (Process Specification Language)
(Schelenoff, et al, 2000). The UEML (Unified Enterprise Modelling
Language) initiative of the IFAC-IFIP Task Force on Enterprise Inte-
gration is another one (Vemadat, 2001 ). EM ontologies have a crucial
role to play to make Enterprise Interoperability a reality in the next
decades.
30 Vernadat, F.B.

3 ENTERPRISE INTEGRATION
Enterprise Integration: Since the early 90's, EI has drastically evolved
from specialised communication protocols (e.g. MAP, TOP, field-buses),
diverse dedicated standard data exchange formats (e.g. IGES, STEP, EDII
EDIFACT, HTML. .. ) and complex monolithic integration infrastructures for
distributed computing environments (e.g., OSF/DCE in the Unix world, OLE
and DCOM in the MS Windows world and OMG/CORBA in the 00 world)
proposed at that time (Vernadat, 2001 ). Regarding Enterprise Application
Integration (EAI), the state of the art is now to use Message-Oriented Mid-
dleware (MOM) (either in stateless or state-full mode as well as in synchro-
nous or asynchronous mode) on top of computer networks compatible with
TCP/IP (Linthicum, 2000). The middleware must provide sufficient scalabil-
ity, security, integrity and reliability capabilities. Messages are more and
more in the form of HTML and XML documents. The most recent trend is to
switch to Java programming (JSP, EJB) and apply the J2EE (Java to Enter-
prise Edition and Execution) principles to build integrated collaborative sys-
tems.
On top of these, large applications are implemented according to the 3-
tier client-server architecture using the web architecture and a standard pro-
tocol (HTTP). A client user can access the application on his/her PC via
HTTP using a standard HTML browser. The request is sent to a web server,
which concentrates all requests and passes the request to the application
server (AS). The AS processes the request using its local database server.
A new trend for the development of application servers is to build them
as a set of remote services accessible via the web, called web services. The
client does not need to know where they are located on the web but can re-
quest their use at any time. Services need to be declared via WSDL (Web
Service Description Language) and registered in a common web repository,
called UDDI
Concerning message exchange, the trend is to make wide use of XML
(eXtensible Mark-up Language) (XML, http) to neutralise data because of
the ability ofXML to separate the logic of documents as well as data format-
ting from data itself. This means that well-known data exchange formats
used in industry (e.g. EDI, STEP, etc.) will soon have to be reworked in the
light of XML (e.g. cXML, ebXML. .. ).
Finally, concerning transport of messages, new protocols are being pro-
posed including SOAP (Simple Object Access Protocol) (http://), RosettaNet
(http://), Bolero.net (http://), Biztalk (http://) among others.
Towards Enterprise Interoperability: Broadly speaking, interoperabil-
ity is a measure of the ability of performing interoperation between two or
more different entities (be they pieces of software, processes, systems, or-
Enterprise Modelling and Integration 31

ganisations ... ). Thus, Enterprise Interoperability is concerned with interop-

erability between organisational units or business processes either within a
large distributed enterprise or within a network of enterprises (e.g. supply
chain, extended enterprise or virtual enterprise). The challenge relies in
communication, co-operation and co-ordination of these processes.

4 CIMOSA REVISION

CIMOSA (AMICE, 1993), a pioneering Enterprise Integration architec-

ture designed in the late 80's - early 90's, is made of three main compo-
nents, namely the Modelling Framework (MFW), the Integrating Infrastruc-
ture (liS, made of distributed computer services) and the System Life Cycle
(SLC or deployment methodology). This architecture can be revisited as fol-
lows.
Concerning the EM Modelling Framework, it is proposed to add a model-
ling view to CIMOSA, called Interaction View, to deal with inter-
organisational aspects, mostly interaction and co-ordination mechanisms
between business entities making a networked organisation or supply chain.
Constructs of this modelling view would include (Fig. 1):
- Business Entity, used to define the components (or nodes) of a net-
worked organisation or supply chain. They can represent External
Suppliers, Manufacturing Units, Warehouses, Final Assembly Units,
Distribution Centres and Customers)
- Interface, used to define the corporate competencies and services of-
fered by each Business Entity and the protocol to access them
- Channel, used to define exchange mechanisms between two Business
Entities in terms of frequency, exchange mode, exchange rate, carrier,
exchange cost, availability, reliability and alternatives). Two types of
Channels need to be distinguished: Communication Channels for
data/information exchanges (information flows) and Transportation
Channels for goods exchanges (material flows).
Concerning the Integrating Infrastructure (liS) the recommendation is to
develop liS services as Web services on top of a Message-Oriented Middle-
ware where messages would be encapsulated in XML format and exchanged
in a secured SOAP-like envelope.
Concerning the System Life Cycle, currently CIMOSA uses the life cycle
defined in GERAM and approved by ISO TC 184/SCS (IFAC-IFIP Task
Force, 1999). However, this life cycle has a linear layout, which might con-
fuse the business user because it does not show the principles of Continuous
Process Improvement currently prevailing in industry and based on the Dem-
ing's Wheel philosophy (Deming, 1982). We suggest the adoption of a more
32 Vernadat, F.B.

cyclic view of the SLC, presented on Fig. 2 and based on modern iterative
prototyping methods used in software engineering as well as in system de-
sign and implementation.
lrteali Fu-dial lriaTTiful R:mrce Qgri
VieN 'lieN 'lieN 'lieN 'lieN

as. 81ity EVert Rlstual

trterf..:e as. Ao::. Ccrr'!xJ19:t

(a9ae)
Chn1fj lrtegity
nJes Fe;pcrliitiily

Figure l: Revised CIMOSA MFW Figure 2: Revised CIMOSA SLC

5 CONCLUSION

Enterprise Modelling has evolved over the last three decades from fact
modelling to Knowledge Management while at the same time Enterprise In-
tegration has evolved from computer systems integration and CIM to Enter-
prise Interoperability and e-commerce.
This paper has provided a short overview of the field in terms of where we
stand and what has to be done next. It also proposes an extension of the
CIMOSA framework to host extended principles for Enterprise Modelling
and Integration.

6 REFERENCES
ACM, (2002), Special issue on Ontology Applications and Design, Communications of the
ACM, 45(2).
AMICE, (1993), CIMOSA: Open System Architecture for CIM, second revised and extended
edition, Springer-Verlag.
Berio, G. Vemadat, F.B. ( 1999), New developments in enterprise modelling using CIMOSA,
Computers in Industry, 40 (2-3): 99-114.
Biztalk (http://www.biztalk.org)
Bolero.net, http://www.bolero.net
Deming, E. W. ( 1982), Quality, Productivity and Competitive Position, The MIT Press.
El Mhamedi, A. Sonntag, M. and Vemadat, F.B. (2000), Enterprise engineering using func-
tional and socio-cognitive models, Engineering Cost and Valuation Analysis.
Fox, M.S. Groninger, M. (1998), Enterprise Modelling, AI Magazine, Fall, 109-121.
Goranson, H. T. ( 1992), Dimensions of Enterprise Integration, in Enterprise Integration
Modeling (Petrie, C. Ed.), The MIT Press, pp. 101-113.
Enterprise Modelling and Integration 33

Harzallah, M. Vernadat, F.B. (2002), IT-based competency modeling and management: From
theory to practice in enterprise engineering and operations, to appear in Computers in In-
dustry.
IFAC-IFIP Task Force, (1999), GERAM: Generalised Enterprise Reference Architecture and
Methodology, Version 1.6, in ISO IS 15704, TC 184 SC5 WGI.
ISO/IEC DIS 19501-1, (2000), Information Technology- Unified Modelling Language
(UML) - Part l: Specification.
Kosanke, K. Nell, J.G. (Eds.), (1997), Enterprise Engineering and Integration: Building In-
ternational Consensus, Springer-Verlag.
Linthicum, D.S. (2000), Enterprise Application Integration, Addison-Wesley.
Martin, J. McClure, C. (1985), Structured Techniques for Computing, Prentice-Hall.
Nonaka, I. A. ( 1994), dynamic theory of organizational knowledge creation, Organization
Science, 5, 4-37.
OAG. OAGIS, (2001), Open Applications Group Integration Specification, Open Application
Group, Inc., Release 7 .2.1, Doc. No. 200 ll 031.
Petrie, C. (Ed.) (1992), Enterprise Integration Modeling, The MIT Press.
RosettaNet, http://www.rosettanet.org
Schelenoff, G. Groninger, M, Tissot, F. Valois, J. Lubell, J., Lee, J. (2000), The Process
Specification Language (PSL): Overview and Version 1.0 Specification, National Institute
of Standards and Technology (NISI), Gaithersburg, MD. USA.
SOAP, http://www.w3.org/TRISOAP
Sowa, J.F., Zachman, J.A. 1992, A logic-based approach to Enterprise Integration, in Enter-
prise Integration Modeling (Petrie, C. Ed.), The MIT Press, pp. 152-163.
Tiwana, A. (2000), The Knowledge Management Toolkit. Practical Techniques for Building a
Knowledge Management System, Prentice-Hall.
Vernadat, F.B. (1996), Enterprise Modeling and Integration: Principles and Applications,
Chapman & Hall.
Vernadat, F.B. (2001), UEML: Towards a Unified Enterprise Modelling Language. Proc.
3eme Conference Francophone de Modelisation et Simulation (MOSIM'Ol), Troyes,
France, 25-27 April, pp. 3-13.
Williams, T.J. (1992), The Purdue Enterprise Reference Architecture, Instrument Society of
America.
XML. http://www.w3.org/TR/1998/REC-xml-19980210.