Multi-layered Enterprise Modelling and Its Challenges in Business

and IT Alignment

Tomasz Kaczmarek

, Ulf Seigerroth

and Nikolay Shilov

Poznan University of Economic, Department of Information Systems, Os. B. Chrobrego, 14/28 60-681, Poznan, Poland

School of Engineering at Jönköping University, P.O. Box 1026, 55111 Jönköping, Sweden

St.Petersburg Institute for Informatics and Automation of the Russian Academy of Sciences,

39, 14 Line, 199178, St. Petersburg, Russia

Keywords: Business and IT Alignment, Enterprise Modelling, Multi-layered Framework, Ontology Support.

Abstract: In this article we have presented different challenges that need to be handled during enterprise modelling if

we want to achieve business and IT alignment. The over all research question in the paper is: What are the

challenges within and between the levels in a multi-layered model for business and IT alignment? In

relation to this research question we have addressed what consequences these challenges will have on our

view on and use of enterprise models as a mean to achieve business and IT alignment. The aim with the

paper is also to give guidance about how to deal with these challenges. Our recommendations for how to

deal with these challenges are summarised in five conclusions about, 1) the use and support from

ontologies, 2) change as driving force during modelling (AS-IS – TO-BE). 3) the dependencies within and

between layers must be clarified, 4) create balance between degree of formalism, degree of details, and

degree of accuracy, and 5) guidelines and best practices about how to deal with these types challenges

during enterprise modelling to create business and IT alignment.

1 INTRODUCTION

A model is usually an abstraction and a

representation of a phenomenon for a certain

purpose (Matthews, 2007) and modelling is the

activity of creating and manifesting the model as a

tangible artefact. Models are used for different

purposes and in the context of this paper we focus

on models and modelling as a mean to deal with the

gap between organizational context and technology.

We are therefore concerned with how to deal with

different challenges related to models and business

and IT alignment. Based on this we have identified

the following challenges:

• How to deal with different abstraction levels of

an enterprise?

• How to deal with the transformation from AS-IS

to TO-BE models?

• How to deal with different aspects (focal areas)

of an enterprise, both within and between

different abstraction levels?

• How to deal with the degree of formalism on

different abstraction levels?

• How to deal with degree of details on different

abstraction levels?

• How to deal with degree of accuracy on

different abstraction levels?

• How to deal with the mutual traceability

between different abstraction levels?

2 CONCEPTUAL FOUNDATION

Enterprise models are to be regarded as tangible

descriptions of patterns addressing different aspects

and constructs of an enterprise where people are

acting often supported by artefacts, within and

between enterprises (Lundqvist et al., 2011).

To arrive at business aligned IS/IT solutions we

need to understand and to be able to handle the

complexity that exists in terms of different aspects

or conceptual domains of an enterprise (Langefors,

1973); (Vernadat, 2002). Lankhorst (2005)

exemplify these multiple enterprise aspects with five

heterogeneous architectural domains (Information

architecture, Process architecture, Product

257

Kaczmarek T., Seigerroth U. and Shilov N..

Multi-layered Enterprise Modelling and Its Challenges in Business and IT Alignment.

DOI: 10.5220/0003972002570260

In Proceedings of the 14th International Conference on Enterprise Information Systems (ICEIS-2012), pages 257-260

ISBN: 978-989-8565-12-9

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

architecture, Application architecture and Technical

architecture), which are related to each other and

need to be integrated in the alignment process.

One way to deal with business and IT alignment

in this context is through a three level framework

(Seigerroth, 2011), which also has served as a

foundation for this paper, see figure below.

Figure 1: Levels in enterprise architecture (Seigerroth,

2011).

This conceptualization depicts conceptual areas

that we need to deal with in the area of business and

IT alignment. In this conceptualization enterprise

architecture is divided into three levels: 1) strategies

& business models, 2) processes & practices, and

3) IS/IT structures and how they are related to each

other. Enterprise modelling is also depicted in the

figure where enterprise models will serve as the

prime tool for addressing business and IT alignment

during transformation (AS-IS to TO-BE). This

conceptualization conveys our belief in the power of

enterprise modelling and enterprise models as

artefacts for coherent descriptions, evaluation, and

design of the three levels in Figure 1 above.

Ontological models is in this context used to

solve heterogeneity problems of models and

different enterprise elements. It provides the

common semantics for interoperability, information

reuse & sharing between disparate modelling

methods, paradigms, languages and software tools

(Uschold and Grüninger, 1996). Ontology

engineering is based on an ontology hierarchy.

Three levels of ontologies have been pointed out

(e.g., Guarino, 1998).

The top-level ontology is the "shared ontology"

for domain independent representation of the

problem set. This type of ontology is needed to

describe an abstract model using common

knowledge model representation. Such ontologies

describe very general concepts like time, matter,

object, action, etc., which are independent of a

particular problem or domain.

Domain ontologies and task ontologies describe

vocabularies related to a generic domain (like

production, or automobiles) by specializing the

terms introduced in the top-level ontology.

Task Ontologies are oriented to solving specific

problems and consist of ontologies of tasks and

methods. A task ontology includes all the concepts

necessary to describe the inferential process, from

the very abstract concepts related to the inference

scheme to more specialized concepts specific for

single methods.

Application Ontologies describe concepts depending

both on a particular domain and task, which are

often specializations of both related ontologies. The

purposes of an application ontology is (1) remove

the gap between domain ontologies and task

ontologies joining them into one application

ontology; (2) enable domain experts to use the

language which is applied in application they work

with, and which may differ from the language used

in the domain and task ontologies.

3 DISCUSSION:

CONSEQUENCES ON

ENTERPRISE MODELLING

AND ENTERPRISE MODELS

Degree of Formalism has to do with how formal the

notation rules for a certain type of model are. There

exist modelling notations that span from very formal

machine interpretable languages to very informal

rich pictures.

Selecting formalism has profound consequences

as it defines the expressivity and thus sets

boundaries to what can, and what cannot be

expressed in the model. The difficult question to

answer before starting the modelling is: what will be

left behind as a result of the decision regarding the

formalism to model in, and whether it will be vital

for the task or not. This question essentially can't be

answered for a given case (as one has only vague

idea of what is to be modelled before engaging into

the exercise), it can only be answered based on

previous modelling experiences.

Degree of Detail is about deciding how many things

we put into a model at different layers of enterprise

modelling in order to describe a certain situation.

There is a certain tension between the formality,

and the number of details of the model required by

IT perspective, and the usefulness of the model to

the business perspective. On the other hand the great

number of details (often technical ones) does not

concern modellers at the business level.

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

258

Unfortunately, the details have to be added

somehow anyway during the transformation of

design models into the real IT system. There are two

problems connected with that - the process of adding

the details is often unspecified, and second, the

modelling tools are not designed to support this. It is

usually unclear who adds the necessary details.

Formalisms in models are most of the time

inflexible - it is often not possible to model in a

given formalism with “less” detail and “more”

detail, or if it is possible, then essentially separate

models have to be created without explicit

connection between them. For example it is possible

to model a business process in a general way (just

specify a single chain of tasks to be performed in the

organization) or in a detailed way (specifying

conversations or message exchange protocols

between process participants as well as message

exchange formats and semantics), but there is little

connection between the general and detailed models

of the same process. Consequently it is usually

impossible to abstract from details in a systematic

way and see a simplified model.

Accuracy of the View is a challenge of selecting the

right point of view when modelling.

Multi aspect representations stems from the

inherent characteristic of different focal areas. First

it is hard to keep the multi aspects in sync. Models

evolve over time and maintaining multi aspect

models is even harder than maintaining a single

view one. Second, it seems, that when implementing

an IT solution it is hard to sustain all the aspects and

specific views on the modelled objects or situations

that has to be selected. This problem result in

information systems that usually adopt one view (or

approach, or method), which from the business point

of view makes them less flexible, constrained and

eventually may lead to the need to change the

system. If we accept this, then the question is how to

make educated decisions on which view would suit

best the goals set for an enterprise, the processes that

should be supported or the strategy to be

implemented. There is a deficiency in many models

and modelling approaches, because they don't give

straightforward answers to what will be the impact

on other parts of the model if we select certain

approach or view on some distant part of the model.

Change and Model Dependencies refer to the fact

that modelling usually is done in a constantly

changing environment.

Regarding the dynamic aspect of modelling - the

transition from AS-IS to TO-BE, the need for

transition may arise in two ways. Either the business

drives the need for change in the IT/IS layer, or that

the technology drives the change - which gives new

opportunities or creates certain restrictions for the

processes that are supported by IT solutions.

Assuming that we use modelling as approach to the

problem, and all the layers have their respective

models, the model of TO-BE differs from the model

of AS-IS. It is usually not a problem to identify the

differences in the models of a single layer. The

question is how the differences should propagate to

other layers. It is not entirely sure at this stage

whether such clues can be delivered automatically.

But the even modest support for noting these

changes and the consequences of changes (how did

change in the model X influenced the change in the

model Y) would be certainly beneficial to make

sense of the whole business / IT infrastructure

evolution process. Currently no tools or other

instrumental support exist to record this evolution

even in a single layer, therefore the challenge is even

greater for the multi-layered enterprise modelling.

4 CONCLUSIONS AND

OUTLOOK FOR THE FUTURE

The discussion of challenges and their consequences

to enterprise modelling leads to several conclusions

on how to deal with the challenges or what is still

necessary in enterprise modelling in order for it to be

able to ensure a better business and IT alignment.

It seems that ontologies applied at and between

different levels of enterprise modelling, despite their

inherent problems, would allow to build connections

between these levels and aid in creating coherent

multilevel models spanning across different focal

areas. The challenge in this will be to create a

working relations between the four types of

ontologies that has been presented in this article,

top-level ontology, domain ontology, task ontology,

and application ontology.

Secondly, current modelling approaches lack the

notion of model evolution or change, in the sense

that they do not adopt the change as the main driving

force in modelling. As indicated previously,

modelling transition from AS-IS to TO-BE might be

the key to show the mutual influences between the

business and IT layers of the enterprise. Currently,

there is little correspondence between the different

subsequent versions of the model of a given focal

area, except from the obvious one (e.g. “this object

is the same as in the previous version of the model”).

Since there is no explicit notion of change and

linkage between subsequent model versions, there is

Multi-layeredEnterpriseModellingandItsChallengesinBusinessandITAlignment

259

no way to specify why certain changes were

introduced, and what might be their consequences

on the other parts of the model.

The third conclusion is that changes in the model

on one layer might influence other layers (e.g. a

change in the business layer might influence

necessary technological transitions in the IT layer).

If multilayer, ontologized models are available, and

there is a way to explicitly model changes and their

influence on different model parts, it would be

possible also to model the influences of changes on

different layers of the model. The dependencies

between the layers are the key to aligning the

business with IT and giving them the dynamic

aspect might contribute to better understanding of

the whole enterprise and its evolution over time.

The fourth conclusion is that there is a need to

strike a balance between degree of formalism,

degree of details, and accuracy of the view in

different enterprise models on different levels. In

many cases these dimensions will be determined

based on situated criteria’s and therefore it will be

important to be able to capture such criteria’s and to

translate them into the actual models.

A fifth conclusion is that based on the four

conclusions above there is a need for more

developed guidelines and best practices about how

to deal with these types challenges during enterprise

modelling to create business and IT alignment.

ACKNOWLEDGEMENTS

This paper is due to the project COBIT sponsored by

the Swedish Foundation for International

Cooperation in Research and Higher Education.

REFERENCES

Bradfield, D. J., Gao, J. X., Soltan, H., 2007. A

Metaknowledge Approach to Facilitate Knowledge

Sharing in the Global Product Development Process,

Computer-Aided Design & Applications, 4(1-4) 519-

528.

Guarino N. (1997) Some Organizing Principles for a

Unified Top-level Ontology. Working Notes of AAAI

Spring Symposium on Ontological Engineering,

Stanford. pp 57-63.

Langefors B. (1973) Theoretical Analysis of Information

Systems. Fourth edition, Studentlitteratur, Lund

Lankhorst M. (Eds., 2005) Enterprise Architecture at

Work – Modelling, Communication, and Analysis,

Springer

Matthews M.R. (2007). “Models in science and in science

education: an introduction”, Science & Education

(16:7-8), pp. 647–652.

Lundqvist M., Sandkuhl K., Seigerroth U. (2011)

Modelling Information Demand in an Enterprise

Context: Method, Notation, and Lessons Learned,

International Journal of Information System Modeling

and Design (IJSMD), Vol. 2, Issue 3, pp 75-95, 2011

Seigerroth U. (2011). Enterprise Modelling and Enterprise

Architecture – the constituents of transformation and

alignment of Business and IT, International Journal of

IT/Business Alignment and Governance (IJITBAG),

Vol. 2, Issue 1, pp 16-34, 2011

Uschold, M., Grüninger, M.: Ontologies: Principles,

methods and applications. Knowledge Engineering

Review, 11(2), 93 – 155 (1996)

Vernadat F.B. (2002) Enterprise Modeling and Integration

(EMI): Current Status and Research Perspectives,

Annual Reviews in Control 26 (2002) 15-25

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

260