Discovering Business Models for Software Process Management
An Approach for Integrating Time and Resource Perspectives from Legacy
Information Systems
C. Arevalo, I. Ramos and M. J. Escalona
Department of Computer Languages and Systems, University of Seville, Avda. Reina Mercedes s/n, 41012 Seville, Spain
Keywords: Software Process Management, BPMN, Interoperability, Legacy Systems, Reverse Engineering, Model-
Driven Engineering.
Abstract: Business Process Management (BPM) is becoming the modern core to support business in all type of
organizations and software business is not an exception. Software companies are often involved in important
and complex collaborative projects carried out by many stakeholders. Each actor (customers, suppliers or
government instances, among others) works with individual and shared processes. Everyone needs dynamic
and evolving approaches for managing their software projects lifecycle. Nevertheless, many companies still
use systems that are out of the scope of BPM for planning and control projects and managing enterprise
content (Enterprise Content Management, ECM) as well as all kinds of resources (ERP). Somehow systems
include scattered artifacts that are related to BPM perspectives: control and data flow, time, resource and case,
for example. It is aimed to get interoperable BPM models from these classical Legacy Information Systems
(LIS). Model-Driven Engineering (MDE) allows going from application code to higher-level of abstraction
models. Particularly, there are standards and proposals for reverse engineering LIS. This paper illustrates LIS
cases for software project planning and ECM, looking at time and resource perspectives. To conclude, we
will propose a MDE-based approach for taking out business models in the context of software process
management.
1 INTRODUCTION
Software companies, as all companies operating in
the current world, are involved in complex changes
due to new globalization rules, such as joint ventures,
mergers and acquisitions, as well as in new
collaboration models among stakeholders such as
outsourcing, offshoring or near shoring, among
others. Today, most of these organizations take into
consideration the Business Process Management
paradigm (BPM) (Van der Aalst, 2004) (Dumas et
al., 2013) (Weske, 2012) to support their software
business processes that is, managing all kind of
software project lifecycles. Such processes are more
unpredictable (Ruiz-González et al., 2004) than
other production processes, since they are
continuously changing and tightly tied to
communication among single people, working teams
protocols and companies contracts. These
organizations also use models, systems, standards
and best practices in the market. In consequence,
they must adapt their knowledge to the requirements
of a global and changing environment, by using
multiple kinds of sophisticated and automated tools
at once, such as planning and control projects,
documents and content management, together with
all kinds of resources (human, technological,
economical or material, for instance) to gain
competitive advantage in the market. These systems
constantly evolve for supporting new paradigms,
architectures, methodologies and languages, as well
as different databases. They should be tailor-made
systems or customizable market products (like
ECMs and ERPs) with different languages, data
structures and details of each business process.
Managing all these complex organizational factors
simultaneously is seemed as a big challenge. Even
though most of these Legacy Information Systems
(LIS) (Bisbal et al., 1999) are not BPM-oriented, we
need to use this new approach for viewing them as
new business models composed of a set of integrated
processes. A BPM model will offer interoperability
among different systems. Geraci (Geraci, 1991)
defines this term as “Ability of a system or a product
353
Arevalo C., Ramos I. and Escalona M..
Discovering Business Models for Software Process Management - An Approach for Integrating Time and Resource Perspectives from Legacy
Information Systems.
DOI: 10.5220/0005454903530359
In Proceedings of the 17th International Conference on Enterprise Information Systems (ICEIS-2015), pages 353-359
ISBN: 978-989-758-096-3
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)
to work with other systems or products without
special effort on the part of the customer.
Interoperability is possible by the implementation of
standards”. Each organization will take advantage
of this ability to get higher flexibility in its internal
business process and will share it with other actors in
big projects, keeping strong independence from the
architecture of information systems.
The Process Mining Manifesto (Van der Aalst et
al., 2012) identifies different perspectives in a BPM
system: control-flow, data-flow, organization and
resource, case and time perspectives. If we had a set
of LIS and we aimed to represent an organization’s
these unwritten business processes, we could
elaborate a conventional modeling business through
interviews, document, information flows or business
rules, among other tools. Now, if possible, we would
like to consider business models derived from
structures and rules already existing in each LIS. In
software projects context, experts often use systems
for planning and control (Hansen and Hansen, 2014),
(Lang, 2010) and collaborative tools for ECM
(Shariff, 2013), since both contain data structures
and rules that are closely linked to business processes
time (Pmbok, 2013) and resource perspectives.
Firstly, we will work with these systems, but we
could extend the study scope to many systems with
other perspectives. These LIS include events, tasks,
resources or resource groups, documents, code, and
business rules, such as rules of time for planning and
controlling tasks and rules of resource allocation,
approval tasks or workflows. Such elements enable
getting, through reverse engineering, core business
processes that subsequently the expert can examine
and polish.
Figure 1: Extracting Business Models.
Figure 1 shows the scenario for taking out a target
BPMN model from these LIS source. LIS persistence
layer usually involves relational database structures
and constitutes the most stable view of business
along processes lifecycle. We are interested in
database artifacts linked to a process time and
resource perspectives. For this aim, this article
suggests using the Model-Driven Engineering
Paradigm (MDE) (Schmidt, 2006).
The paper is structured in different sections as
follows: a) related work, b) taxonomy for time and
resource allocation rules, c) our MDE-based
approach for reverse engineering LIS and d)
conclusions and future work.
2 RELATED WORK
We have explored standards and research work
regarding Model-Driven Architecture (MDA),
software process management, time and resource
perspectives in BPMN and reverse engineering LIS
databases as follows:
Firstly, MDA (MDA, OMG 2011) is the major
representative of the MDE paradigm that proposes
different level of abstraction to gain independency
and interoperability (Elveszeter and Berre, 2010)
among different views of a universe of discourse. It
defines Platform Specific Models (PSM), Platform
Independent Models (PIM) and Computer
Independent Models (CIM), where PSM is close to
the application code and CIM is found at business
expert level. Meta Object Facility (MOF, OMG
2011) introduces the metamodel concept and Query
and View Transformation (QVT, OMG 2011) is a
language for mapping artifacts of different
abstractions.
Regarding Model-Driven Development (MDD)
(Mellor et al., 2003) and other methodologies, some
studies, such us (Valderas and Pelechano, 2011)
present a relevant current global view about MDE
methodologies. It is not the aim of this paper to
depict all of them, but we illustrate our work with
Navigational Development Technique (NDT)
(Escalona et al., 2008), which is supported by a set
of tool-case grouped in NDT-Suite (García-García et
al., 2012) and NDTQ-Framework. This methodology
has being successfully used in several software
development projects such as (Cutilla et al., 2011),
(Escalona et al., 2013) and (Salido et al., 2014),
which enable us to have a good set of data and
examples.
Secondly, (García-Borgoñón et al., 2014) carries
out a Systematic Literature Review (SLR) for
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
354
Software Process Management that covers software
process modeling languages, methods, standards and
best practices. They briefly conclude that Business
Process Model and Notation (BPMN OMG, 2011) is
the selected technology, among other approaches, to
model processes, because it offers simplicity,
standardization and support for execution processes.
Many UML-based approaches have been developed,
but there are actually two main languages: Software
& Systems Process Engineering Metamodel
Specification 2.0 (SPEM, OMG 2008), which
provides a language for software methodologies and
(ISO/IEC 24744, 2007), which offers guidelines to
identify process models as well as improve
consistency and uniformity when defining these
standards.
BPMN and MDD are becoming the key elements
of the aforementioned standards for software
management. Business Process Definition
Metamodel (BPDM, OMG 2008) provides
interoperability by means of different tools, so we
will work with a metamodel for business processes
based on BPMN, UML activity diagrams (UML,
OMG 2011) and BPDM.
Thirdly, with respect to time perspective in
BPMN, researchers agree on its weakness to show
this dimension. (Flores and Sepúlveda, 2011)
propose patterns to express time rules included in a
Gantt chart. They work with BPMN 1.2 supported
artifacts. Besides, (Gagné and Trudel, 2008) focus on
Allen's Interval Algebra, extending BPMN diagrams
with decorators in Time-BPMN (Gagné and Trudel,
2009), where the start and end task events and
associations among them are semantically redefined
to display these constraint types. (Awad et al., 2009)
recommend a set of Workflow Resource Patterns
(WRP) over processes, working with an extension of
BPMN metamodel for the resource perspective and
OCL rules as expression of allocation rules.
Finally, LIS maintenance can be too expensive
because these systems are not always well
documented. Additionally, if they are highly
complex, they are not easy to replace for a new
system (Bisbal et al., 1999), (Heuvel, 2006). OMG
Architecture-Driven Modernization (ADM) (Ulrich
and Newcomb, 2010) is an approach based on MDA
that identifies metamodels and processes for LIS
modernization. It is also important within MDA to
review the Information Management Model (IMM,
OMG 2006) as a proposal of interoperability that,
among others, establishes metamodel views for
tables, keys, routines and triggers (Eisenberg, 1996)
of a SQL relational database (Melton and Simon,
2002). We have also revised, among other sources,
the work by (Boronat et al. 2005) and (Pérez-Castillo
et al., 2011 and 2012) for relational database schema
reverse engineering (the latter uses the concept of
process archaeology and code mining) and by
(Izquierdo et al., 2008 and 2012) where its main
example is a DSL tool named Grammar to Model
Language (Gra2Mol), which acts as a bridge going
from grammar-ware to model-ware. These studies
are based on an ADM-approach and use basic
relational SQL database metamodels for extracting
tables and keys, such as UML classes and
associations. Other studies, for example by (Demuth
et al., 1999 and 2001) and (Arevalo et al., 2014), also
take out ECA rules hardcoded as triggers (Eisenberg,
1996) for enriching UML class diagrams with Object
Constraint Language (OCL, OMG 2014) assertions.
3 TIME AND RESOURCE RULES
There are several interesting characteristics derived
from databases to define our BPM, as it is mentioned
in subsequent work. Basically, this paper mainly
focus on time and resource perspectives, although it
enables the taxonomy for business rules we are
searching in each LIS. However, other kind of rules
may appear when considering other perspectives, but
they are out of the scope of this analysis. We just aim
to find Time Constraints and Resource Allocation
Constraints. Figure 2 shows a rule taxonomy.
Figure 2: Business rules taxonomy.
As Time Constraints concerns, it defines rules over a
task, a process or a sub-process or precedence
dependencies between pairs of tasks, processes and
sub-processes. They are shown as Temporal
Constraints (TC) and Temporal Dependencies (TD),
and in turn, (TC) is subdivided into (a) General
Constraints, (b) Inflexible TC and (c) Flexible TC.
DiscoveringBusinessModelsforSoftwareProcessManagement-AnApproachforIntegratingTimeandResource
PerspectivesfromLegacyInformationSystems
355
All kinds of rules, except General Constraints, are
founded in (Pmbok, 2013), (Flores and Sepúlveda,
2011) and (Gagné and Trudel, 2008 and 2009).
(a) Firstly, General Constraints provide
chronological rules between the start and finish
events of an activity. Furthermore, time-interval
must be in the scope of parent activity (process or
sub-process).
(b) Secondly, Inflexible TC fix the start and finish
task events: ‘MSO: Must Start On’ and ‘MFO: Must
Finish On’.
(c) Last, Flexible TC enable a floating start and
finish task events; thus ‘ASAP: As Soon As Possible’,
‘ALAP: As Last As Possible’, ‘SNET: Start Not
Earlier Than’, ‘SNLT: Start Not Later Than’,
‘FNET: Finish Not Earlier Than’ and ‘FNLT: Finish
Not Later Than’ allow advancing or delaying a task,
whenever the critical path of the total process is not
affected.
TD constraints govern previous relations
between events of a preceding task and events of a
current task. Thus, the four possible combinations
are: ‘SS: Start to Start’, ‘SF: Start to Finish’, ‘FS:
Finish to Start’ and ‘FF: Finish to Finish’.
With respect to Resource Allocation Constraints
(Awad et al., 2009) propose a metamodel to capture
resource and case perspectives of a business process
and also the OCL specification of nearly thirty
Workflow Resource Patterns (WRP). Nevertheless,
we will find simple patterns like (WRP01) Direct
Allocation, (WRP02) Role Allocation and (WRP03)
Authorization Allocation in software systems such as
project planning tools and ECMs. More complex
allocation patterns are associated with case
perspective in a process that needs references to
individual task instances.
4 EXTRACTING BUSINESS
MODELS
Our approach stands as a model-driven solution. If
we consider time and resource perspectives to enrich
our heuristic process, we will enhance the
representation of business processes in comparison
with the models obtained by means of typical ADM
approaches (see Section 2). However, in order to
illustrate the grade of abstraction and make it
understandable, we present our approach in the
context of MDA working at different MDA levels of
abstraction: (i) Target PIM BPMN models with time
and resource perspectives and (ii) Source PSM
Relational Models including table structures,
constraints and rules at lower level, but related to the
same perspectives. A specific heuristics with QVT
transformations will allow applying reverse
engineering to turn PSM models into PIM models.
Nonetheless, we will need specific metamodels at
both levels.
4.1 BPMN Metamodel Extensions
BPMN metamodel extensions are necessary to
capture time and resource semantics, and later on,
other perspectives may be included with new
extensions. We only work with BPMN 2.0 and OCL
2.4 standard specifications, therefore transforming
models into executable code will be easier.
Figure 3: PIM Packages.
Figure 3 shows two packages: BPMN for standard
selected artifacts and Extension for new time and
resource-related classes.
Figure 4: BPMN metamodel extensions.
In Figure 4 Process and Activity have dates (real and
scheduled) associated to start and finish task events;
an activity may have TCs and TDs with preceding
activities; a TD may have a delay interval with
respect to its predecessor: lead (-) or lag (+).
Resource and Role support the basic allocations rules
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
356
(WRP01, WRP02 and WRP03) with OCL rules
(Awad et al., 2009). The latter is associated with the
Lane class to determine the performer of an Activity.
4.2 Heuristics from PSM to PIM
We will use the PSM metamodels capabilities
relational database proposed, among others by
(Arevalo et al., 2014) and (IMM, OMG 2006), to
take out table schemas and constraints. The
metamodel uses two packages, Generic Abstract
Syntax Metamodel (GASTM) and Specific Abstract
Syntax Metamodel (SASTM), to capture enough
behavior from each concrete LIS relational model,
i.e.: Oracle, MS-SQL*Server, PostgreSQL and
MySQL, for instance, that have their own procedural
languages to code triggers with PL/SQL, Transact-
SQL, pgSQL and MySQL stored procedures as
variations of ISO SQL/PSM (Eisenberg, 1996).
We suggest a model-to-model (M2M) (MDA
OMG, 2011) method to transform structures from
LIS databases at PSM level into target PIM BPMN
metamodel (see Figure 5). It will:
(i) Select a table from each LIS database that is
linked to time and resource perspectives. Map tables
and foreign keys to metamodel classes and
associations. Map table columns to classes attributes.
Figure 5: Reverse engineering from PSM to PIM.
(ii) Map database constraints to OCL business
rules given by the classes taken out, considering
studies like those by (Demuth et al., 1999 and 2001)
and (Arevalo et al., 2014) for this purpose.
(iii) Infer new classes and OCL constraints to
capture time and resource perspectives given on LIS
systems, taking into account the taxonomy rules
defined in Section 3.
This proposal includes models, metamodels, a
Domain Specific Language (DSL) and heuristic
transformations where: (a) Every model conforms to
its metamodel; (b) Metamodels, both at PSM and
PIM level are defined with a DSL based on UML2
profiles; (c) The procedure transforms artifacts
among different levels of abstraction and it also
merges different source models to get a unique and
global business data object view according to the
referenced PIM metamodel.
5 CONCLUSIONS AND FUTURE
WORK
BPMN is essential for software organizations.
Consequently, this paper proposes a MDE-approach
to find out business models from LIS, such as project
planning and ECM systems. For this aim, we propose
the following points:
(i) Work with time constraints and resource
allocation rules gathered from project planning and
ECM systems. Derive a BPMN PIM view with more
expressive structures and rules that ADM approaches
could apply to generic LIS systems.
(ii) Work with a target metamodel based on
actual versions of BPMN and OCL, allowing
extensions with more resource allocation rules, new
perspectives and a process polishing in liaison with
the business expert.
Nowadays, we are working in OCL formalization
of rules taxonomy. More LIS systems may constitute
sources for extracting and merging process models.
Therefore, we will work with study cases with the
aim of evaluating this approach and going further
with the analysis of more perspectives rather than
time and resource.
ACKNOWLEDGEMENTS
This research has been supported by the MeGUS
project (TIN2013-46928-C3-3-R) of the Ministry of
Economy and Competitiveness, Spain.
REFERENCES
Arevalo, C., Gómez-López, M. T., Reina Quintero, A.M.,
Ramos, I., 2014. An Architecture to Infer Business
Rules from Event Condition Action Rules Implemented
in the Persistence Layer. In R. Pérez-Castillo, & M.
Piattini (Eds.) Uncovering Essential Software Artifacts
DiscoveringBusinessModelsforSoftwareProcessManagement-AnApproachforIntegratingTimeandResource
PerspectivesfromLegacyInformationSystems
357
through Business Process Archeology (pp. 201-221).
Hershey, PA: BSR.
Awad, A., Grosskopf, A., Meyer, A., & Weske, M., 2009.
Enabling resource assignment constraints in BPMN.
Hasso Plattner Institute, Potsdam.
Bisbal, J., Lawless, D., Wu, B., Grimson, J., 1999. Legacy
Information Systems: Issues and Directions. 0740-
7459/99, ©IEEE.
Boronat, A., Carsí, J.A., Ramos, I. 2005. Automatic
Reengineering in MDA Using Rewriting Logic as
Transformation Engine. CSMR 2005: 228-231.
BPDM, OMG 2008. Business Process definition
Metamodel (BPDM), http://www.omg.org/spec/
BPDM/
BPMN, OMG 2011. BPMN V2.0. http://www.omg.
org/spec/BPMN/2.0/
Cutilla, C. R., García-García, J. A., Alba, M.; Escalona,
M.J., Ponce, J., Rodríguez, L., 2011. Aplicación del
paradigma MDE para la generación de pruebas
funcionales; Exp. AQUA-WS. 6ª Conf. Iberic. S.T.I.,
Demuth, B., Hußmann, H., 1999. Using UML/OCL
constraints for relational database design. UML.
doi:10.1007/3-540-46852-8_42.
Demuth, B, Hußmann, H., Loecher,S., 2001. OCL as a
specification language for business rules in database
applications. UML. doi:10.1007/3-540-45441-1_9.
Dumas, M., La Rosa, M., Mendling, J., Reijers, H. A.,
2013. Fundamentals of business process management.
Heidelberg: Springer.
Elveszeter, B., Berre, A.J., 2010. OMG Specifications for
Interoperability. IESA I.C. (p. 31). John Wiley &
Sons.
Eisenberg, A., 1996. New standard for stored procedures
in SQL. ACM SIGMOD Record 25 (4): 81–88.
Escalona, M.J., Aragón, G., 2008. NDT. A Model-Driven
Approach for Web Requirements. IEEE TSE, vol. 34,
no. 3, pp. 377-390.
Escalona, M.J., García-García, J. A., Más, F., Oliva, M.,
Del Valle, C., 2013. Applying model-driven paradigm:
CALIPSOneo experience. CAiSE'13, vol. 1017, pp.
25-32.
Flores, C., Sepúlveda, M., 2011: Temporal Specification of
Business Processes through Project Planning Tools. In
BPMW (pp. 85-96). Springer Berlin Heidelberg.
Gagné, D., Trudel, A., 2008: The Temporal Perspective:
Expressing Temporal Constraints and Dependencies
in Process Models. Process Models in BPM and
Workflow Handbook.
Gagné, D., Trudel, A., 2009. Time-bpmn. In CEC'09 IEEE
Conference on (pp. 361-367).
García-Borgoñón, L, Barcelona, M.A., García-García,
J.A., Alba, M., Escalona, M.J., 2014. Software
Process Modeling Languages: a Systematic Literature
Review. ISS Vol. 56. 103-116.
Geraci, A., 1991.
IEEE Standard Computer Dictionary.
IEEE Press, Piscataway, NJ, USA.
Hansen, C., Hansen, P. K., 2014. Microsoft Project 2013.
Target Gruppens Forlag.
Heuvel, W., 2006. Aligning Modern Business Processes
and Legacy Systems: A Component-Based
Perspective. CIS, MIT Press.
IMM, OMG 2006. Information management metamodel
(IMM), http://www.omgwiki.org/imm/doku.php.
ISO/IEC 24744, 2007. Software Engineering Metamodel
for Development Methodologies.
Izquierdo, J.L.C., Cuadrado, J. S., Molina, J. G., 2008.
Gra2MoL: A domain specific transformation language
for bridging grammarware to modelware in software
modernization. In Workshop on MDSE.
Izquierdo, J.L.C., García Molina, J., 2012. Extracting
models from source code in software modernization.
Software & Systems Modeling.
García-García, J. A., Ortega, M. A., García-Borgoñón, L.,
& Escalona, M. J., 2012. NDT-Suite: a model-based
suite for the application of NDT. Web Engineering (pp.
469-472). Springer Berlin Heidelberg.
Lang, J. P., 2010. Redmine-Overview.
Mellor, S. J., Clark, A. N., Futagami, T., 2003. Model-
driven development - Guest editor’s introduction.
IEEE Softw. 20, 5, 14–18.
Melton, J., Simon, A.R., 2002. SQL:1999. Morgan
Kaufmann. pp. 541–542. ISBN 978-1-55860-456-8.
MOF, OMG 2011. Meta Object Facility (MOF),
http://www.omg.org/spec/MOF/2.4.1.
OCL, OMG 2014. Object Constraint Language (OCL),
Version 2.4. http://www.omg.org/spec/OCL/2.4/.
Pérez-Castillo, R.; Piattini, M. et al., 2011. MARBLE. A
business process archeology tool. ICSM 2011:578-
581.
Pérez-Castillo R., Piattini, M. et al., 2012. Database
schema elicitation to modernize relational data-bases.
ICEIS.
Pmbok, 2013. A Guide to the Project Management Body of
Knowledge (Pmbok Guide). 5th Edition. Project
Management Institute.
QVT, OMG 2011, Meta Object Facility (MOF) 2.0 QVT.
http://www.omg.org/spec/QVT/1.1/.
Ruiz-González, F., Canfora, G., 2004. Software Process:
Characteristics. T&E., vol. 5, pp. 5-10.
Salido, A., García-García, J.A., Gutiérrez, J.; Ponce, J.,
2014. Tests Management in CALIPSO-neo: A MDE
Solution. Journal of SE&A.
Schmidt, D.C., 2006. Model-Driven Engineering,
Computer, vol.39, no.2, pp.25-31.
Shariff, M., 2013. Alfresco 4 Enterprise Content
Management Implementation. Packt Publishing Ltd.
SPEM, OMG 2008. Software & Systems Process
Engineering Metamodel Specification (SPEM 2.0).
http://www.omg.org/spec/SPEM/2.0.
Ulrich, W. M., & Newcomb, P., 2010. Information Systems
Transformation: Architecture-Driven Modernization
Case Studies. Morgan Kaufmann.
UML, OMG 2011, Unified Modeling Language (UML
2.0). http://www.omg.org/spec/UML/.
Valderas, P., Pelechano, V., 2011. A Survey of
Requirements Specification in Model-Driven
Development of Web Applications. TWEB 5(2):10.
Van der Aalst, W., 2004. Business process management: a
personal view. BPM Journal, vol. 10, no. 2, p. 5.
Van Der Aalst, W. et al., 2012. Process mining manifesto.
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
358
In BPM workshops (pp. 169-194). Springer.
Weske, M., 2012. Business Process Management.
Concepts, Languages, Architectures, 2nd Edition.
Springer.
DiscoveringBusinessModelsforSoftwareProcessManagement-AnApproachforIntegratingTimeandResource
PerspectivesfromLegacyInformationSystems
359
