MDE FOR BPM
A Systematic Review
Jose Manuel Perez
UCLM-Soluziona Research and Development Institute, Ronda de Toledo s/n, 13005, Ciudad Real, Spain
Francisco Ruiz, Mario Piattini
Alarcos Research Group, University of Castilla-La Mancha, Paseo de la Universidad, 4, 13071Ciudad Real, Spain
Keywords: Business process management, M
odel driven engineering, Model driven architecture, Systematic review.
Abstract: Due to the rapid change in the business processes of organizations, Business Process Management (BPM)
has come into being. BPM helps business analysts to manage all concerns related to business processes, but
the gap between these analysts and people who build the applications is still large. The organization’s value
chain changes very rapidly; to modify simultaneously the systems that support the business management
process is impossible. MDE (Model Driven Engineering) is a good support for transferring these business
process changes to the systems that implement these processes. Thus, by using any MDE approach, such as
MDA, the alignment between business people and software engineering should be improved. To discover
the different proposals that exist in this area, a systematic review was performed. As a result, the OMG’s
Business Process Definition Metamodel (BPDM) has been identified as the standard that will be the key for
the application of MDA for BPM.
1 INTRODUCTION
There is a need for today’s business to create and
modify value chains rapidly. This brings about
continuous growth and change in business processes.
The goal of Business Process Management (BPM) is
to help business people to manage these changes.
Business process management is defined as the
cap
ability to discover, design, deploy, execute,
interact, operate, optimize and analyze process in a
way that is complete, doing it at the business design
level and not at the technical implementation level
(Smith, et al., 2002).
BPM offers numerous benefits to organizations
suc
h as improving the speed of business, giving
increased customer satisfaction, process integrity
and accountability. It promotes process optimization,
at the same time eliminating unnecessary tasks. It
also includes customers and partners alike in the
business processes and provides organizational
agility.
BPM represents a “third wave” in business
pr
ocess engineering. The first wave was guided by
process papers that reorganized human activity. The
second wave focused on reengineering of business
processes and the use of Enterprise Resource
Planning (ERP). The third wave centers on formal
business process models and the ability to modify
and combine those models so as to align business
process with organizational needs (Frankel, 2003).
BPM starts with proce
ss modeling. Process
modeling is a business-driven exercise in which
current and proposed process flows are documented
in detail, linked to quantifiable performance metrics,
and optimized through simulation analysis.
Standards for process modeling languages are the
key to the attaining of BPM’s goal as well as in
achieving the platform independence of the process
models. Platform independence is one of the
principles on which Model Driven Engineering
(MDE) is based. The combination of both concepts,
MDE and BPM, is the target of this systematic
review.
MDE was conceived in an effort to solve several
p
roblems that have arisen in the last decade. On one
hand, the growth of platform complexity, there being
thousands of classes and methods with very
complicated dependencies. On the other hand, we
118
Manuel Perez J., Ruiz F. and Piattini M. (2006).
MDE FOR BPM - A Systematic Review.
In Proceedings of the First International Conference on Software and Data Technologies, pages 118-124
DOI: 10.5220/0001310901180124
Copyright
c
SciTePress
can observe the continuous technological evolution
of the systems, forcing programmers to modify the
system code every time a new requirement is given.
In the MDE paradigm, every concept must be
modeled. Thus, any change in the system must be
shown in the model that represents that system. To
model the systems, MDE proposes using Domain-
Specific Modeling Languages (DSML). By means of
these languages, different modeling notations for
each kind of system are achieved. Thus, the software
engineer has specific tools for modeling all kind of
systems.
Another important concept in MDE is model
transformation. By transforming models, the
evolution of the systems is facilitated. A model
could be transformed to another model or to a XML
specification as well as to the source code that
implements the model functionality.
The OMG group has developed Model Driven
Architecture (MDA) as an example of MDE. MDA
emerged with the established idea of separating the
business logic specification of a system from the
platform specific details in which the system is
implemented (Miller, et al. 2003).
MDA adds some concepts to the MDE
philosophy. MDA defines three level of abstraction.
The Computational Independent Model (CIM), the
Platform Independent Model (PIM) and the Platform
Specific Model (PSM).
The key technology in MDA is MOF, as it is as
in the definition of metamodels, which are MOF
instances (figure 1) (Bézivin, 2003). The
transformations among these models are the basis of
MDA philosophy.
Figure 1: MOF metamodels structure (Bézivin, 2003).
The structure of this paper is as follows. In
section 2, systematic reviews are introduced. In
section 3, the carrying out of the review is shown in
part, presenting the selection of studies and the
classification of these. The information analysis is
described in section 4 by summarizing the different
authors’ proposals about the MDE for BPM
application. Section 5 presents the conclusions
extracted from the systematic review along with
future work, taking into account the different views
found.
2 SYSTEMATIC REVIEWS
A systematic review of the literature is a means of
identifying, evaluating and interpreting all available
research relevant to a particular research question, or
topic area, or phenomenon of interest (Kitchenham,
2004).
Systematic review is a scientific methodology
that can be used to integrate empirical research on
software engineering (Travassos, 2005).
Some of the characteristics that make the above
methodology different from a conventional review
are that a systematic review starts by defining a
review protocol that specifies the research question,
along with the methods and the criteria to drive the
review. Added to all this, a systematic review is
based on a search strategy that aims to detect as
much relevant literature as possible. Moreover,
performing a systematic review is needed in order to
document the whole search strategy so that another
researcher can replicate the same review with
identical results.
There are three main phases that organize the
different stages of the review process.
The phase called “planning the review” has as its
purpose to identify the need for this study and to see
through the development of a review protocol. A
researcher may need a systematic review to be able
to draw more general conclusions about a
phenomenon or as a prelude to further research
activities.
The protocol specifies the methods that will be
used to undertake a specific systematic review. A
pre-defined protocol is needed to avoid the
possibility of researcher bias. Without a protocol, the
selection of individual studies might possibly be
driven by the expectations of the researcher.
When the whole planning is done, the review can
start. This is the second phase, called “conducting
the review”. This phase lies in the identification of
research, the selection of primary studies, the quality
assessment study, data extraction and monitoring,
together with data synthesis.
Firstly, the researcher must search the documents
by using the strings specified in the protocol. When
a first potential set of primary studies is obtained,
the researcher must perform a selection by assessing
MDE FOR BPM - A Systematic Review
119
Table 1: Studies Selection.
Author, date Study name Source
Roser and Bauer (2005)
A Categorization of Collaborative
Business Process Modeling
Techniques
IEEE Digital Library
Zeng, et al. (2005)
Model-Driven Business Performance
Management
IEEE Digital Library
Pfadenhauer, et al (2005)
Comparison of Two Distinctive
Model Driven Web Service
Orchestration Proposals
IEEE Digital Library
Rosen (2004) SOA, BPM and MDA ACM Digital Library
Frankel (2005) BPMI and OMG: The BPM Merger Business Process Trends
Harmon (2004)
The OMG's Model Driven
Architecture and BPM
Business Process Trends
Frankel (2003)
BPM and MDA: The Rise of Model-
Driven Enterprise Systems
Business Process Trends
Smith (2003)
BPM and MDA: Competitor,
Alternatives or Complementary
Business Process Trends
Kano, et al. (2005)
Analysis and simulation of business
solutions in a service-oriented
architecture
Wiley Digital Library
MEGA & Standard Bodies (2004)
Business Process Modeling and
Standardization
bpmg.org
the studies’ actual relevance. Quality assessment
must be done over the selected studies. As the result
of assessing the information quality, according to the
criteria defined in the protocol, a new set of studies
is generated.
Finally, the data synthesis provides researchers
with the results of the systematic review. The
synthesis may be either quantitative or descriptive.
The last phase lies in the communication of the
results. Usually the systematic review is reported in
at least two formats: In a technical report or in a
section of a PhD thesis as well as in a journal or
conference paper.
3 REVIEW RESULTS
This section presents the selected works in the
searches performed in the digital libraries, journals
and internet sites related to the issue in hand.
Moreover, a classification of studies is given. This
has used aspects which are of relevance to the goal
of the review as a basis for this classification
3.1 Studies Selection
The first step was to search in the predefined
information sources. Those sources are: ACM digital
library, IEEE digital library, Science Direct Digital
Library, Business Source Premier, Wiley
InterScience, www.BPTrends.com, www.bpmg.org.
The result of this search was a first set,
composed of 22 studies. With the aim of tuning the
set of studies, the selection criteria were applied.
The studies had to contain information about the
application of model driven engineering or model
driven architecture in business process management.
The issue of the systematic review is MDE for BPM,
but because MDA is currently so widespread in the
model engineering world, MDA was included in the
selection criteria.
As the result of the application of selection
criteria, the new set of studies was composed of 10
works (Table 1).
3.2 Classification of Studies
The selected studies have been classified according
to several aspects that have been chosen to satisfy
the goal of the systematic review (Table 2).
First of all, the author’s opinion about the issue
of systematic review is the most important aspect to
take into account in classifying the studies. Another
important aspect is whether the study offers a
proposal about the use of CIM, PIM and PSM
(MDA models) within the business process context.
This means that the author suggests a specific
utilization of MDA models, pointing out the possible
modeling standards used in each model. Finally, the
different standards proposed by authors for
modeling business process are also aspects that are
taken into account.
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
120
Table 2: Classification of the selected studies.
Author, date
MDE
for
BPM
Propose
CIM,
PIM, &
PSM
UML
BPML
BPMN
BPDM
BPEL
J2EE
Others
Roser and Bauer (2005) Yes Yes X X X X X ebXML, AIRIS, WS-CDL
Zeng, et al. 2005 Yes No X X
Pfadenhauer, et al. (2005) Yes Partially X X X
Rosen (2004) Yes Yes X X X X
Frankel (2005) Yes No X X X X SBVR
Harmon (2004) Yes Yes X X X X X SBVR
Frankel (2003) Yes No X X X X
Smith (2003) No No X X
Kano, et al. (2005) Yes Yes X X X
MEGA & Standard Bodies (2004) Yes No X X X X XPDL
4 FINDINGS AND ANALYSIS
This systematic review goal is to identify studies
that can provide an approach for the application of
the MDE paradigm to business process management.
Note that from here on in the text, MDA will be the
modeling approach that will always be mentioned,
whereas MDE will not. This is because MDA is the
most widely-seen example of MDE application, and
because all the papers deal specifically with MDA,
and not with MDE in general.
The article “BPM and MDA: Competitors,
Alternatives or Complementary” (Smith, 2003), does
not share the optimism of the rest of the authors. In
Smith’s opinion, BPM and MDA are very different.
He declares that MDA must be used by software
engineers and that BPM must be used by business
people. He also affirms that the latter are not
interested in a new approach for developing more
software, but rather in a design-driven architecture
based on processes and on a business process
management system (BPMS) that interprets such
designs, in the same way that RDBMS interprets a
relational model. Although he does not deny the
possibility that in the future the two philosophies
may work together, at the moment he advocates the
separation of both approaches.
The work “Model-Driven Business Performance
Management” (Zeng, et al., 2005) proposes a
technical approach for developing a complete
application related to the BPM context. This study
presents a relation between the two important
concepts
of this systematic review, using a model-
driven approach to build the solution. The technical
approach is based on the observation metamodel and
its transformations. When the models are
transformed, the approach suggests compiling the
operational aspects of the model and finally
developing a runtime engine that interprets the
model and executes the generated code.
The study ”Comparison of Two Distinctive
Model Driven Web Service Orchestration
Proposals” (Pfadenhauer, et al., 2005) focuses on the
way to generate a set of web services that implement
the organization business processes. By applying the
MDA approach, and using some of the business
process standards, the final solution is generated.
This document mentions the BPDM standard as the
MDA BPM connection.
The article “Analysis and simulation of business
solutions in a service-oriented architecture” (M.
Kano, 2005), offers a four-layer model architecture,
in which the first two layers, when viewed together,
are similar to the CIM layer in MDA from the
business point of view rather than from the software
MDE FOR BPM - A Systematic Review
121
Figure 2: Use of OMG BPDM (Harmon, 2004).
system point of view. The last two layers correspond
directly to the MDA PIM and PSM layers. By
separating the independent platform concerns of a
solution from the specific platform concerns and
their associated code by means of MDA, the reuse of
solution components is supported. Furthermore, the
system is more flexible and adaptable to the changes
in business requirements.
The work “A Categorization of Collaborative
Business Process Modeling Technique” (Roser, et
al. 2005), provides a proposal for applying MDA
within the collaborative business process
framework. Collaborative business processes are
performed among different enterprises, which could
have different business process development
methodologies. Therefore, the creation of a common
framework in which the organizations could
communicate to each other in terms of business
process would be ideal. The authors have spoken
about MDA as the common framework for
integrating business process from different
organizations. They propose to create the business
process CIMs, PIMs and PSMs in every
organization, by using their own model language for
each kind of model. These model languages must be
MOF metamodels. Thus, transformations among
metamodels can be done. The communication
among the enterprises in terms of business process
will be done by means of the common CIMs, PIMs
and PSMs. These common models are written by
using a common metamodel (one for each kind of
model) and contain a view for the models of each
organization from their CIMs, PIMs and PSMs.
Thus, the common framework is well-known for all
the organizations.
The study “Business Process Modeling and
Standardization” (MEGA & Standard Bodies, 2004),
is a review concerning all of the standards existing
around business process, from languages to
modeling notations. It provides a whole view of the
state of standards (as it stood on September 2004),
as well as their coverage within the BPM context.
Moreover, it reports on the capacities of versions of
new standards that are about to come out.
The study “SOA, BPM and MDA” (Rosen,
2004) does not offer a specific proposal for using
MDE within some business process management
areas, but provides an abstract vision about the role
that both MDE and BPM play. The article points out
how MDA can help business process automation,
reuse and maintenance.
The two works by Frankel selected in the
systematic review, concerning MDA and BPM,
(Frankel, 2003 and Frankel, 2005), point to the use
of MDA as the methodology that guides business
process design, implementation, maintenance and
management. Frankel’s theory is that BPM joined to
MDA is stronger than BPM alone, and MDA
together with BPM is stronger than MDA alone.
Moreover, he gives a wide classification of the
different business process standards that currently
exist. He aims at the aligning of the business process
modeling notation (BPMN) with the OMG
metamodel BPDM. This would provide portability
utility by means of the XMI format and the power of
the MDA transformations, in line with the well-
known BPMN standard. Although Frankel is
optimistic about the application of MDA in BPM, he
also warns us about the wide gap that exists between
the abstraction represented by a business process
model and the specific models that represent the
implementation of the business process.
The study “The OMG's Model Driven
Architecture and BPM” (Harmon, 2004), has as its
goal the use of MDA within the BPM. Harmon puts
BPDM at the centre of business process modeling
(Figure 2). The rest of business process modeling
standards should be transformed directly to BPDM,
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
122
even BPMN. He proposes a way to use the different
kinds of MDA models (CIM, PIM and PSM) for
business process design and implementation. Thus,
CIM will be specified in terms of business process
by using BPDM; the business rules by means of
business rules metamodel (BRM). These models are
used by business analysts. PIM are a transformation
from previous CIM, specified in a software system
metamodel, for example UML. These models are
used by software architects. Finally, PSM are built
by transforming PIM to the platform specific
language in which the business process will be
implemented, for example the J2EE UML Profile.
5 CONCLUSIONS AND FUTURE
WORK
The systematic review performed provides a
complete view of the proposals and opinions
existing in the recent literature about MDE paradigm
application in business process management.
Most of the works found point to the use of
model driven engineering as a valid approach for
business process management. There are proposals
for the use of MDA in the context of collaborative
business process management, where the model
driven plays the role of integration standard and
allows different organizations to cooperate from a
business process point of view. It is also suggested,
on the other hand, that MDA is the methodology that
drives the organization business process design,
implementation, maintenance and management.
Although most authors are in favor of the use of
MDE in business process management, there is
some rejection of this idea, throwing into relief how
far apart both concepts are, and how difficult it is to
obtain cooperation to achieve better results.
Business process modeling standards become the
key issue for the MDA application in the context of
BPM. These standards must be metamodels, which
are instances of meta-metamodel MOF. OMG
propose the business process definition metamodel
(BPDM) as the standard for business process
modeling, which has no final version yet (OMG,
2003). BPDM is a semantic description of the
logical relations among several elements of any
business process description. It is not a notation. Its
advantage is that it is a MOF metamodel. Thus, any
other notation language, such as BPMN, can be
transformed to BPDM. As BPDM is a MOF
metamodel, this can be transported via XMI to any
business process tool that knows such a metamodel.
The companies only have to define MDA
transformations from the BPDM metamodel to
executable languages like J2EE or BPEL.
BPMN is the notation standard most frequently
used to define business process at a high level. So
some authors are quite adamant in their assertions
that the next version of BPDM will take on the
BPMN standard. Thus, any high level BPMN model
will be able to be shared via XMI and transformed to
follow the MDA methodology.
In future research, we will monitor the evolution
of BPDM and its convergence with the BPMN
standard. We will propose a QVT transformation
from BPMN to BPDM, as well as from BPDM to a
web services metamodel. To do this, the model
management framework MOMENT will be used
(Boronat, et al. 2005).
ACKNOWLEDGEMENTS
This research has been supported by the project
FAMOSO, partially funded by Ministerio de
Industria, Turismo y Comercio, FIT-340000-2005-
161 Plan Nacional de Investigación Científica,
Desarrollo e Innovación Tecnológica 2004-2007 and
“Fondo Europeo de Desarrollo Regional (FEDER)”,
European Union.
REFERENCES
Smith, H., 2002. The emergence of Business Process
Management. CSC’s Research Services.
Frankel, D. S., 2003. BPM and MDA. The rise of model-
driven enterprise systems. Business Process Trends.
Miller, J., Mukerji, J., 2003. MDA Guide Version 1.0.1,
OMG.
Bézivin, J., 2003. MDA: From hype to hope, and reality.
In UML’03. 6th International Conference. ATLAS
Group.
Kitchenham, B., 2004. Procedures for performing
systematic reviews. NICTA Joint Technical Report.
Travassos, G. H., Biolchini, J., Gomes, P., Cruz, A. C.,
2005. Systematic review in software engineering.
Technical Report.
Smith, H., 2003. BPM and MDA: Competitors,
alternatives or complementary. Business Process
Trends.
Zeng, L., Lei, H., Dikun, M., Chang, H., Bhaskaran, K.,
2005. Mode-driven business performance
management. In ICEBE’05. Proceedings of the 2005
IEEE International Conference on e-Business
Engineering. IEEE.
Pfadenhauer, K., Dustdar, S., Kittl, B., 2005. Comparision
of two distinctive model driven web service
MDE FOR BPM - A Systematic Review
123
orchestration proposals. In CECW’05. Proceedings of
the 2005 Seventh IEEE International Conference on
E-Commerce Technology Workshops. IEEE.
Kano, M., Koide, A., Liu, T., Ramachandran, B., 2005.
Analysis and simulation of business solutions in a
service oriented architecture. IBM Systems Journals,
Vol 44, No 4.
Roser, S., Bauer, B., 2005. A categorization of
collaborative business process modeling techniques. .
In CECW’05. Proceedings of the 2005 Seventh IEEE
International Conference on E-Commerce Technology
Workshops. IEEE.
MEGA & Standard Bodies, 2004. Business process
modeling and standarization. Bpmg.org
Rosen, M. 2004. SOA, BPM and MDA. CTO Azora
Technologies.
Frankel, D. S., 2005. BPMI and OMG: The BPM Merger.
Business Process Trends.
Harmon, P., 2004. The OMG’s model driven architecture
and BPM. Business Process Trends.
OMG, 2003. Business Process Definition Metamodel,
Request for Proposal, OMG document, 2003-01-06,
http://www.omg.org/docs/bei/03-01-06.pdf
Boronat, A., Carsí, J. A., Gómez, A., Ramos, I., 2005.
Utilización de Maude desde Eclipse Modeling
Framework para la gestión de modelos. Departament
de Sistemes Informatics i Computacio. Universitat
Politécnica de Valencia.
ICSOFT 2006 - INTERNATIONAL CONFERENCE ON SOFTWARE AND DATA TECHNOLOGIES
124