A New Approach to Semantically Derive Enterprise Information

Architecture from Business Process Architecture

Mahmood Ahmad and Mohammed Odeh

Software Engineering Research Group, Faculty of Environment and Technology, University of the West of England,

Coldharbour Lane, Bristol, BS16 1QY, U.K.

Keywords: Enterprise Information Architecture, Semantic Information Architecture, Semantic Information Integration,

Semantically Derived Information Architecture, Enterprise Information Management.

Abstract: The design of enterprise information architecture (EIA) is a critical success factor of the information

management capability of an enterprise. The EIA design not only supports business process re-engineering

activity but can also facilitate radical design of new business processes. This paper demonstrates that EIA

can be derived semantically from the business process architecture, hence overcoming the classical

problems of time-consuming interviews for business information analysts by using semantically enriched

BPA to derive information entities and processes with the help of domain and process ontologies. The

cancer care and registration process of a health organisation has been used as a case-study for demonstration

purposes. This approach is currently being transformed into a generic Framework for semantic enterprise

information architecture design with an aim to bridge the gap between business architectures and enterprise

information architectures.

1 INTRODUCTION

Classically, the EIA design activity consisted of

building a map of a firm's information resources and

business functions based on time-consuming

managerial interviews. The subsequent time

requirements led to an increasing lack of interest

from the strategic management in this vital design

effort. However, a consistent emphasis on the

vitality of IA (Information Architecture) design in

relation to a firm's business processes and

information resources has remained the case

(Martin, 1989; Teng and Kettinger, 1995) in the

business information management research

community.

Recent research has identified that the

information management capability of an enterprise

has a significant contribution towards development

of its capabilities in customer management, process

management and performance management (Mithas

et al., 2011; Sauer and Willcocks, 2003). The

Information Management (IM) Capability may be

defined as the ability to provide data and

information within the desired accuracy and time

scale for a given enterprise business process. This

implies that the design of information architecture

(IA) is of pivotal importance for developing the IM

capability. The enhancement implied by this ability

develops customer management (CM) capability that

generates opportunities to develop customer

relationships both as consumers and innovation

partners. The IM capability can also contribute

towards developing and redesigning business

processes for carrying out the activities of an

enterprise, reflecting on the Process Management

(PM) capability of the enterprise, (Mithas et al.,

2011).

The design of enterprise information architecture

may, however, remain irresponsive to requirements

of gathering business analytics data as well as to

changes in business strategy unless it incorporates a

comprehensive analysis of business information into

the EIA design. Business information resources

(business entities and processes) of an enterprise can

have pivotal role in information (data) quality,

security and corporate strategy as the requirements

from these enterprise areas translate into how well

and what information resources are managed in the

enterprise, and to what extent they correspond to the

business process architecture of the enterprise.

This paper puts forward an approach that bridges

this gap between business architecture and enterprise

363

Ahmad M. and Odeh M..

A New Approach to Semantically Derive Enterprise Information Architecture from Business Process Architecture.

DOI: 10.5220/0004561803630369

In Proceedings of the 15th International Conference on Enterprise Information Systems (ICEIS-2013), pages 363-369

ISBN: 978-989-8565-61-7

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

information architecture using semantic

technologies. Semantic information about the

business process architecture can be used to derive

information architecture of the enterprise which is

not only responsive to business information needs

but that can also provide leverage to the design of

new business processes. Section 2 discusses the

classical and contemporary attempts to BPA-

oriented EIA. Section 3 discusses the proposed

approach that semantically derives EIA entities and

processes of the enterprise from BPA ontological

artefacts. Section 4 demonstrates the current work

through a case-study in cancer-care domain while

identifying some adjustments for the parent

framework that conceptualises the BPA

methodology. Section 5 discusses issues in this

approach that we currently face with some possible

remedies, and Section 6 concludes this paper.

2 RELATED WORK

Non-Semantic approaches to IA design include

enterprise data model approach by (Brancheau et al.,

1989), long-range information architecture and

ends/means analysis approaches (Wetherbe and

Davis 1983), critical success factors approach

(Rockart 1979), Information Engineering (IE) based

approaches (Martin, 1989) such as IBM's Business

System Planning (BSP) and strategic data modeling

(SDM). Among these approaches, the enterprise data

model approach and the IE-based approaches were

business process-centric; however, the IE-based

approaches lacked appeal due to the absence of

technological advances of today (Kettinger et al.

1996; Teng and Kettinger, 1995). More recently, the

EIA is seen as a part of the overall enterprise

architecture (EA) of an enterprise and is also

mentioned as data architecture. Examples of these

approaches include Zachman's Information Systems

Architecture (ISA) (Zachman, 1987; Sowa and

Zachman, 1992), the Architecture Development

Model (ADM) by TOGAF (TOGAF, 2009) and the

CEiSAR model (CEiSAR, 2008).

Semantic approaches for Enterprise Architecture

(EA) include the TOronto Virtual Enterprise

(TOVE), (Fox et al., 1995), however, no semantic

approach to generate a data (or information)

architecture of an enterprise is process-centric.

Genre and Ontologies Based Information

Architecture Framework (GOBIAF) by (Kilpelinen,

2007) is based on information need interviews and

not based on the knowledge of business entities and

processes. The Field-Actions approach by (Pascot et

al., 2011) uses field actions for incorporating

business process information and uses HL7 ontology

but it lacks derivation of information architecture of

an enterprise.

3 THE PROPOSED APPROACH

3.1 EIA Semantic Derivation

The approach presented in this paper derives the

EIA elements from the fundamental elements of

business process architecture of an enterprise

modelled using the Riva BPA methodology (Ould

2005). The Riva methodology concentrates on the

business of the enterprise and collects essential

business entities (EBEs) without which the

enterprise will cease to perform its function. It may

also collect additional information which is related

to a particular way by which an enterprise may have

decided to run its business (called designed business

entities or DBEs). It extracts from these business

entities a set of units of work (UoWs), each of which

leads to a business process at the operational (case

processes – CPs), managerial (case management

processes – CMPs) and strategic (case strategy

processes – CSPs) levels respectively. It must be

noted here that the CMP and CSP processes are not

carried out by business managers or enterprise

strategists, rather their names indicate the very

nature of the tasks they carry out for a particular

EBE of UoW. The CMP and CSP, however, can be

used by business managers and / or CIOs to induce

changes in BPA for the corresponding entities

corresponding to any new decisions made at the

enterprise level. This needs to be carried out using

enterprise information systems which rely on the

EIA, hence highlighting the need for the EIA

elements to be directly derivable from the BPA.

3.2 The BPAOnt Ontology

The elements of Riva BPA methodology by (Ould,

2005) were conceptualised into the BPAOnt

ontology (Yousef et al., 2009) in their

BPAOntoSOA Framework. This ontological

representation of BPA can be named as semantically

enriched BPA. The BPAOnt ontology was

developed in OWL (W3C-OWL, 2004) and contains

BPA concepts like EBE (essential business entity),

UOW (units of work), Riva_Relations (within

UOWs), CP (case process) and CMP (case

management process), and relationships among

these concepts which makes a good starting point for

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

364

the design of the EIA. The starting point for the

BPAOntoSOA Framework are crude BP models of

an enterprise modelled using Role-Activity

Diagrams (RADs) or more recent BP modelling

languages like BPMN (Yousef et al., 2009). The

business information that causes the development of

BPA of an enterprise should rather originate from

some other sources, in our view, such as business

documents etc. However, the BPAOnt ontological

concepts developed through the BPAOntoSOA

Framework are considered enough for our approach

to start deriving the EIA elements.

As the effort of developing the semantically

enriched BPA is a one-off activity for an enterprise

corresponding to the developed business process

architecture, and needs only minor adjustment

corresponding to business change, an Enterprise

Information Architecture that holds direct additional

knowledge of business processes helps improving

the automation of the EIA design process. Thus, it

reduces the time requirements for interviews and

questionnaires in the sense that the knowledge of

business entities and processes is already captured

through a semantically enriched BPA. However, this

time-saving is more relaizable once the process of

semantically enriched BPA development is

automated by either accessing machine-readable

business process models and workflows or by using

natural language processing techniques to analyse

business documents and extract business process

architectural elements.

3.3 EIA Derivation

The enterprise IA contains design elements that are

derivable from the BPA and are one step closer to

the systems level because EIA elements represent

explicit concepts that can be programmed in an

object-oriented environment to develop a business

information system that can not only capture events

in the enterprise, but can also respond to those

events. The EIAOnt ontology has been designed in

OWL (W3C-OWL, 2004) using the approach by

(Noy and McGuiness 2001) and based on IA

concepts by (Gilchrist and Mahon, 2004) and

(Evernden and Evernden, 2003; Brancheau et al.,

1989), and forms a major component of the

BPAOntoEIA Framework. Fundamental EIA

elements are EIA entities and EIA processes which

are conceptualised as InformationEntity and

EIAProcess concepts respectively in the EIAOnt

Ontology. The InformationEntity concept is

based on the Bung-Wand-Weber’s ontological

model of information that provides ontological basis

for both concrete (physical) and conceptual (non-

physical) entities (Wand and Weber, 1990).

Although this type of entity classification provides a

sound ontological basis for specification and

semantic representation of EIA entities, yet this

classification may not be essential for designing and

implementing enterprise information system based

on these entities. This framework consists of a three-

step approach for EIA derivation:

Figure 1: The proposed approach for Deriving EIA from Enterprise BPA.

ANewApproachtoSemanticallyDeriveEnterpriseInformationArchitecturefromBusinessProcessArchitecture

365

1) The first step of this approach derives initial set

of EIA entities and processes from the EBE and

process instances of the BPAOnt ontology by

instantiating the BPAOntoSOA Framework for a

particular enterprise. The semantic derivation of

the EIA entities includes which EBEs qualify to

become EIA entities and classifies each of them

into concrete and conceptual entities. The

semantic derivation also includes derivation of

EIA processes from process concepts of BPA,

and also captures the associated relationships

among process instances, which are taxonomic

(whole-part) and / or non-taxonomic.

2) The second step of this derivation uses the

instances of the EIA information entity and

process concepts of the EIAOnt ontology to

search for related concepts in external domain

ontologies. This also includes identifying the

taxonomic and non-taxonomic relationships

among new and existing case-study entities and

processes. The search for related entities and

processes may also result in formation of new

external domain ontologies or updating enriching

external ontologies through a structured process

of searching and cataloguing EIA information

entities and processes.

3) Finally, more complex EIA elements such as

EIA traceability matrices, EIA diagrams (such as

Figure 2: RAD Model of the CCR Patient General

Reception sub-process.

Information flow diagrams and / or Entity-

Relationship diagrams) and EIA information

views are to be derived. As the name suggests,

the EIA traceability matrices provide traceability

information for information entities (IEs)

corresponding to EBEs in BPA, IEs vs EIA

Processes, and the like to ensure that all EIA

elements are traceable to generate complete

information flow diagrams for an EIA process.

4 THE CANCER CARE

CASE-STUDY

We demonstrate our new approach by applying it to

one sub-process of the Cancer Care and Registration

(CCR) process of King Abdullah Cancer Centre in

Jordan. The CCR case-study was used by (Yousef et

al., 2009) in her research for identifying services

corresponding to the BPA elements. The CCR case-

study includes the sub-processes of the Patient

General Reception, Hospital Registration, Cancer

Detection, Cancer Treatment and Patient Follow-up.

Of these, we use Patient General Reception sub-

process that models the process of a patient’s

general reception at the Cancer Care Centre. Figure

2 shows the RAD (Ould, 2005) model of Patient

General Process.

4.1 CCR BPA Elements

The EBEs and UoWs generated by the instantiated

by BPAOntoSOA Framework are listed in Table 1.

The output of CCR BPA is described as BPAOnt-

CCR in Figure 1. The units of work are listed in

bold. Corresponding to every UoW listed, the Riva

BPA methodology generates an instance of CP

concept and one instance of CMP concept. Before

identifying EIA elements, we propose to complete

the BPAOntoSOA framework by adding an instance

of the CSP concept for every unit of work.

4.2 The CCR EIA Elements

The BPA elements generated by BPA (Table 1)

form the basis for the EIA entities and processes in

BPAOntoEIA Framework. The semantic derivation

process for EIA entities includes SWRL rules that

identify which EBEs qualify to become EIA entities.

The EBEs, which are semantically derived as EIA

entities, are classified as concrete or conceptual

entities. This basic classification is useful because it

can facilitate the decision-making processes within

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

366

Table 1: BPA elements for the Patient General Reception

sub-process in the CCR case-study.

EBEs or UoWs

Patient General Reception

Receptionist (General)

Patient

Medical Records

Appointment

Patient File

Emergency Unit

Cancer detection unit

Database

Patient details

Case Processes (CPs)

Handle Patient general reception

Handle a patient medical record

Case Management Processes (CMPs)

Manage the flow of Patient general reception

Manage the flow of patient medical record

business information system, e.g. supply chain and

delivery of a printed book or an electronic book

(ebook) version and decide upon cost of delivery

accordingly. In the CCR context, EMERGENCY

UNIT is a physical entity and has a location,

whereas DATABASE is a conceptual entity.

The EIA derivation function generates the

CRUD (Create, Read, Update and Delete) processes

for every EIA entity. The CRUD processes are sub-

concepts of EIAProcess concept in EIAOnt

Ontology. The CPs and CMPs also form an initial

set of EIA (non-CRUD) processes. Once

relationships between these concepts are established

and traceability among these elements is determined,

the EIA design function then moves on to search in

external domain and process ontologies, using

automated ontology search processes to look for

related entities and processes. This is possible only

when domain specific knowledge for a particular

business exists, e.g. cancer care domain knowledge

for CCR case-study. If there is no domain specific

knowledge, the EIA design process can then develop

new domain knowledge as its by-product. The

search may result in significant increase in the

number of EIA entities and processes. The

traceability for these newly found EIA elements

should establish many-to-many relationships

between EIA entities and the initial set of EIA

entities which were originally EBEs in BPA. Many-

to-many relationships also exist between EIA

processes and the EIA entities they access, use and /

or modify. Table 2 lists the set of EIA entities and

processes in the case-study sub-process after

searching for related entity and process concepts in

the NCI Thesaurus (Ceusters et al., 2005) and the

Medical Ontology by Advance Genome Clinical

Trials (ACGT) project (Cocos et al., 2008). We have

noted that entities in these ontologies are not

classified into concrete and conceptual entities and

we therefore recommend constructing a new

ontology using this classification for EIA entities.

The complete EIA for CCR is referred to an

EIAOnt-CCR in Figure 1.

Table 2: Count of EIA elements derived from BPA for

Patient General Reception sub-process in CCR case-study

after look-up in ACGT Medical Ontology for entities.

EIA Element Count

EIA Entities

Entities derived from BPA

Entities searched from

domain ontologies

Concrete entities

Conceptual entities

EIA Processes

CRUD Processes

IE Management (IEMP)

Processes

IE Strategy Process (IESP)

EIA Traceabilit

Matrices

In this table, the IEMP processes are the

processes for CMPs in the BPA that manage CPs,

and the IESP processes correspond to CSPs.

5 DISCUSSION

The derivation process of these EIA elements has

highlighted a number of important issues which can

be significant for a complete and correct design of

an EIA. Firstly, we note that this approach

significantly depends upon the correctness and

completeness of the Riva BPA elements identified

by instantiating (Yousef et al., 2009)’s

BPAontoSOA framework. The starting point of the

BPAOntoSOA framework is, however, the business

process models of the case-study enterprise that

were originally developed through on-site interviews

in a previous research (Aburub, 2006). We suggest

that the input for BPA development needs to be

business documents and business use cases and not

necessarily BP models. This, however, should not

affect the correctness and validity of EIA as the

ANewApproachtoSemanticallyDeriveEnterpriseInformationArchitecturefromBusinessProcessArchitecture

367

needed BPA elements are instances of the BPAOnt

concepts, and hence it will generate an EIA

corresponding to these BPA elements.

Secondly, although this approach causes EIA

design to be heavily dependent upon the business

process architecture, yet this becomes a strength and

not a weakness of the approach because our

proposed EIA is more business process-aware and it

is more responsive to business process change if it

includes a change management mechanism that

tracks any changes in BPA and makes corresponding

adjustments to EIA architectural elements. The

changes in BPA translate into changes in Riva BPA

business entities, processes or relations between its

units of work. The change management mechanism

of the EIA should capture these changes and initiate

EIA ‘change processes’ to assess the impact of these

changes and implement them. Thus, future

information requirements, that are not yet expressed

in process models will need to emerge from a review

of BPA models (driven by strategic management),

followed by change in EIA. This limitation or

dependence should be seen as an opportunity to

review business process architecture. Future

information requirements that do not need change in

process models may be met through change

management processes acting independently of

BPA.

Thirdly, the EIA derivation process needs to

identify whole-part relationships among the EIA

entities derived from BPA and also for the entities

searched from external domain ontologies. This

includes identifying which EIA entities are merely

multiple instances of another entity and whether this

parent entity needs to be included as an EIA entity

or not. We may call this refactoring of EIA entities.

For instance, there are 10 RECEPTIONIST EBE

instances in the BPA. This leads the information

architect to define one RECEPTIONIST instance

with a variable place of deployment. Furthermore,

the RECEPTIONIST entity may be a sub-entity of a

PERSON entity of which PATIENT is another sub-

entity. The whole-part relationship may be added to

the information about EIA entities using OWL

properties and SWRL rules (SWRL, 2004).

Fourthly, in order to ensure the correctness of

EIA derivation approach, human input from

information architect (IA) may be essential at certain

stages of EIA derivation. For example, the IA’s

input may be required when refactoring of EIA

entities and processes is carried out. This may be

carried out through special-purpose dialogue boxes,

which may render the above derivation process

semi-automated rather than being fully automated,

hence further work for further evolution of the

current work.

Finally, limitations of such an approach to derive

EIA from BPA emerge from those of BPA

methodology. This approach is critically dependent

upon the Riva methodology as the underlying BPA

methodology. This is because the Riva BPA

methodology is systematic and focuses on business

entities and process in a way that brainstorms all

entities and units of work that lead to processes, thus

identifying the business components along its

natural fault-lines and hence providing a

comprehensive (initial) set of EIA entities and

processes.

6 CONCLUSIONS

We have presented a new generic approach for

semantically deriving the Enterprise Information

Architecture of an enterprise from its Business

Process Architecture. We have also demonstrated

results of this derivation using a real and validated

case study based on Cancer Care and Registration in

Jordan, namely the CCR case-study. Moreover, we

have identified some shortcomings in the current

BPAOntoSOA framework algorithm with respect to

the extraction of EBEs from business processes and

hence the reflection on the inclusion of case strategy

process (CSP) in the BPAOnt ontology for a given

BPA case. Currently, this approach is limited to the

derivation of the fundamental EIA elements such as

EIA entities and processes, and the traceability

matrices that ensure forward / backward traceability

from/to these elements. This approach is to be

extended to generate more advanced EIA elements

such as information views and information flow

diagrams while exploiting the dynamic relations

within the BPA’s units of work and the traceability

of EIA entities and EIA processes that access these

entities. Further research includes the generalisation

of this approach to a validated Semantic Framework,

with maximum automaticity, for deriving the design

of an enterprise information architecture from the

given business process architecture of that

enterprise.

REFERENCES

Aburub, F. A. F., 2006. A Business Process Improvement

Methodology Based On Process Modelling, Applied

To The Healthcare Sector, Phd Thesis, Bristol Institute

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

368

of Technology (BIT), University of The West of

England.

Brancheau, J. C., Schuster, L. and March, S.T., 1989.

Building and Implementing an Information

Architecture. SIGMIS Database, 20(2), Pp. 9-17.

Ceisar, 2008. Enterprise Modeling, White Paper,

Downloaded From URL: Www.Ceisar.Org, [July 6,

2011].

Ceusters, W., Smith, B. And Goldberg, L., 2005. A

Terminological And Ontological Analysis of the NCI

Thesaurus. Methods of Information in Medicine, 44,

PP. 498-507.

Cocos, C., Brochhausen, M., Bonsma, E. And Martin, L.,

2008. Design Principles of the ACGT Master

Ontology: Examples and Discussion. Deliverable 7.7.

Madrid, Spain: UPM.

Evernden, R. And Evernden, E., 2003. Information First:

Integrating Knowledge and Information Architecture

for Business Advantange. Oxford, UK: Elsevier

Butterworth Heinemann.

Fox, M. S., Barbeceanu, M.and Gruninger, M., 1995. An

Organisation Ontology for Enterprise Modelling:

Preliminary Concepts for Linking Structure and

Behaviour. Computers in Industry, 29, Pp. 123-134.

Gilchrist, A. and Mahon, B., Eds, 2004. Information

Architecture: Designing Information Environment for

Purpose. London, UK: Facet Publishing.

Kilpelinen, T., 2007. Genre And Ontologies Based

Business Information Architecture Framework

(Gobiaf), University of Jyvaskyla.

Martin, J., 1989. Information Engineering: Introduction.

Usa: Englewood Cliffs, New Jersey: Prentice-Hall.

Mithas, S., Ramasubbu, N. and Sambamurthy, V., 2011.

How Information Management Capability Influences

Firm Performance. Mis Quarterly, 35(1), Pp. 475-500.

Noy, N. and Mcguiness, D. L., 2001. Ontology

Deelopment 101: A Guide to Creating Your First

Ontology. SMI-2001-0880.

Ould, M., 2005. Business Process Management: A

Rigorous Approach. UK: British Computer Society.

Pascot, D., Bouslama, F. and Mellouli, S., 2011.

Architecturing Large Integrated Complex Information

Systems: An Application to Healthcare. Knowledge

and Information Systems, 27(1), PP. 115-140.

Rockart, J. F., 1979. Chief Executives Define their own

Data Needs. Harvard Business Review, 57(2), PP. 81-

89.

Sauer, C. And Willcocks, L., 2003. Establishing The

Business of The Future: The Role Of Organisational

Architecture And Informational Technologies.

European Management Journal, 21(4), Pp. 497-508.

Sowa, J. F. And Zachman, J.A., 1992. Extending And

Formalizing The Framework For Information Systems

Architecture. IBM Systems Journal, 31(4), PP. 590-

616.

Swrl, 2004. A Semantic Web Rule Language Combining

OWL and RuleML, URL: Http://www.w3.Org/

Submission/Swrl/, [November 10, 2010].

Teng, J. T. C. and Kettinger, W. J., 1995. Business

Process Redesign and Information Architecture:

Exploring the Relationships. DATABASE Advances,

26(1), PP. 30-42.

TOGAF, 2009-Last Update, The Open Group Architecture

Framework (TOGAF). Available: http://www.

opengroup.org/architecture/togaf9-doc/arch/ [February

25, 2010].

W3C-OWL, February 10, 2004, 2004-Last update, OWL

Web Ontology Language Overview. Available: http://

www.w3.org/tr/owl-features/ [July 1, 2009, 2009].

Wand, Y. And Weber, R., 1990. An Ontological Model of

an Information System. Software Engineering, IEEE

Transactions on, 16; 16(11), PP. 1282-1292.

Wetherbe, J. C. and Davis, G. B., 1983. Developing A

Long-Range Information Architecture, Proceedings of

National Computer Conference, May 16-19 1983,

ACM, Pp. 261-269.

Yousef, R. M., Odeh, M., Coward, D. And Sharieh, A.,

2009. Bpaontosoa: A Generic Framework To Derive

Software Service Oriented Models From Business

Process Architectures, Second International

Conference on Applications of Digital Information

and Web Technologies, IADIWT '09 2009, Pp. 50-55.

Yousef, R. M., Odeh, M., Coward, D. and Sharieh, A.,

2009. Translating Rad Business Process Models Into

Bpmn Models: A Semi-Formal Approach.

International Journal of Web Applications (IJWA).

Zachman, J. A., 1987. A Framework for Information

Systems Architecture. IBM Systems Journal, 26(3),

PP. 276-292.

ANewApproachtoSemanticallyDeriveEnterpriseInformationArchitecturefromBusinessProcessArchitecture

369