=Paper=
{{Paper
|id=None
|storemode=property
|title=Applying Model-Driven Paradigm: CALIPSOneo Experience
|pdfUrl=https://ceur-ws.org/Vol-1017/Paper4CAiSE_IT2013.pdf
|volume=Vol-1017
|dblpUrl=https://dblp.org/rec/conf/caise/EscalonaGMOV13
}}
==Applying Model-Driven Paradigm: CALIPSOneo Experience==
https://ceur-ws.org/Vol-1017/Paper4CAiSE_IT2013.pdf
Applying model-driven paradigm: CALIPSOneo
experience
M.J.Escalona1, J.A. García-García1, F. Mas2, M. Oliva2, C. del Valle1
1
Department of Computer Languages and Systems. University of Seville. Spain
mjescalona@us.es, julian.garcia@iwt2.org, carmelo@us.es
2
Airbus, Spain {fernando.mas, manuel.oliva}@airbus.com
Abstract: Model-Driven Engineering paradigm is being used by the research
community in the last years, obtaining suitable results. However, there are few
practical experiences in the enterprise field. This paper presents the use of this
paradigm in an aeronautical PLM project named CALIPSOneo currently under
development in Airbus. In this context, NDT methodology was adapted as
methodology in order to be used by the development team. The paper presents
this process and the results that we are getting from the project. Besides, some
relevant learned lessons from the trenches are concluded.
Keywords: Model-Driven Paradigm, Product Life-Cycle Management (PLM),
industrial Digital Mockup (iDMU)
1. Introduction
The model-driven paradigm is a paradigm that combines the power of concepts and
its relations, using abstracts models, in order to offer suitable mechanisms for
software development. This paradigm is focused on concepts and how these concepts
evolved in the life cycle of a product. We can refer to model-driven engineering
(MDE) [1], when this paradigm is applied in the context of engineering. In the last
years, software engineering community has applied MDE in different contexts of
software engineering, getting suitable results. For instances, it was used in the context
of the web engineering, in methodologies such as UWE (UML Web Enginnering) [2]
or WebML (Web Model-Languages) [3]; in the context of software testing, with
approaches such as [4][5]; in software product lines, such as [6]; or even in more
specific contexts like software architecture, with the approach WebSA [7] or in the
context of RIA (Rich Internet Applications) [8] These researches have produced
suitable results that could be considered an inspiration source for the industrial
community. However, very few experiences were reported from the enterprise [9].
This paper presents a real application of MDE in an industrial context. NDT
(Navigational Development Techniques) [10] is a MDE methodology that was
applied in a large number of real projects. Nevertheless, in this situation, it was
adapted and applied in a new context, the aeronautical one, which entails modifying
the methodology and its management policies. The paper presents this version and a
global view of the experience. It analyses learned lessons and future works and
concludes that the application of the model-driven paradigm in aeronautical contexts
can be a relevant area and can offer very successful results.
2. CALIPSOneo
CALIPSOneo (advanCed Aeronautical soLutIons using Plm proceSses & tOols) is an
ambitious project developed in Airbus that has as a main target the definition through
a comprehensive requirements collection process, using PLM (Product Life-Cycle
Management) software existing in the market, that will allow the industrial engineers
to define, simulate, optimize and validate the aeronautical assembly processes in a 3D
virtual environment before they are implemented in a real shop floor. Through the
years, many analyses were performed on PLM and its benefits for Airbus [11]. From
this information, the project team has developed the project requirements
documentation, the requirements management plan, and the requirements traceability
matrix that satisfies the PLM business needs. CALIPSOneo covers the design of a
new PLM methodology to conform to a PLM collaborative design and the required
development of the software that satisfies that concept [11]. CALIPSOneo is
subdivided into three individual subprojects, in order to effectively manage the work
needed to complete it. It will let the project team more effectively manage the
project’s scope as they work on the tasks necessary to complete the project. Each of
the three projects is subdivided in work packages. Tasks are the next level below
work package, which are the minimal entity the project can be structured into.
Following this structure, CALIPSOneo consists of three subprojects: MARS
(autoMAted shop-flooR documentation updating System), PROTEUS (PROcess
sTructure gEneration and Use) and ELARA (gEneraLization to assembly oriented
authoring Augmented ReAlity). These subprojects are independent and teams
involved in each of them also differ. However, subprojects have to be coordinated and
they have to be correctly integrated because they have common actors who demand
common functionality. Thus, the main objective ELARA must achieve is to develop a
system which should provide workers in assembling tasks, with augmented reality
technology, still valid on any airframe, using the 3D information contained in the
iDMU (industrial Digital Mock Up) [13] or information coming from MARS. The
final result will be an industrial prototype to be used in the assembly process for the
FanCowl product in the A320Neo program. PROTEUS is responsible for the Product,
Process and Resources structure definition and the their interrelations to create an
iDMU and MARS is responsible for exploiting the iDMU in order to obtain
automatically the documentation needed at shop floor for the airplane assembly. Thus,
despite each of them have different functions and aims; it is required to guarantee that
they are incorporated, connected and well-integrated.
3. Background
3.1 NDT-Navigational Development Techniques
Initially, NDT was defined as a MDE methodology focused on requirements and
analysis processing. At the beginning, it dealt with defining a set of formal
metamodels for the requirements and analysis phases. In addition, NDT defined a set
of derivation rules, stated with the standard QVT (Query View Transformations)[14],
which generated analysis models from requirements model. The main goal of the
Requirements phase of NDT is to build the catalogue of requirements containing the
needs of the system to be developed. It is divided into a series of activities: capture,
definition and validation of requirements. Requirements can also be classified
according to their nature: information storage requirements, functional requirements,
actor requirements, interaction requirements and non-functional requirements. NDT
defines derivation rules to generate the analysis phase models, once the requirements
specification phase has been completed and the catalogue of system requirements has
been drafted and validated. Figure 1 shows this idea through the stereotype
«QVTTransformation». The transition between the requirements and the analysis
model is standardized and automated. It is based on QVT transformations, which
translate the concepts of requirements metamodels into the first versions of the
analysis models. These models are known in NDT as basic models of analysis. For
example, the basic conceptual model of analysis is obtained from the storage
requirements defined during the requirements phase. Thereafter, the team of analysts
can transform these basic models to enrich and complete the final model of analysis.
Figure 1 Transformations from Requirements to Analysis
As this process is not automatic, the expertise of an analyst is required.
Transformations are represented in Figure 1 through the stereotype «NDTSupport».
NDT controls these transformations by means of a set of defined rules and heuristics,
to ensure consistency between requirements and analysis models. This idea, which
was only applied in the requirements and analysis phases, was extended in the last
years and, currently, this context supports the complete life cycle: viability study,
requirements, analysis, design, implementation, maintenance and testing under
different life cycles, such as classical, agile or the one based on prototypes1. To sum
up, NDT offers an environment conducive to Web systems development, completely
covering the software development life cycle. Recently, NDT has been applied to
many practical environments where it has succeeded due to the application of
transformations among models, which has reduced the development time.
1
You can get more information about NDT full life cycle in www.iwt2.org
3.2 NDT-Suite
The application of MDE and, particularly, the application of transformations among
models, may become monotonous and very expensive if there are no software tools
that automate the process. Therefore, NDT has defined a set of supporting tools called
NDT-Suite 2 [15] to meet this need. Currently, the suite of NDT comprises the
following main free Java tools: (a) NDT-Profile is a specific profile for NDT,
developed using Enterprise Architect. NDT-Profile offers the chance of having all the
artifacts that define NDT easily and quickly as they are integrated into a tool called
Enterprise Architect [16]. (b) NDT-Quality is a tool that automates most of the
methodological review of a project developed with NDT-Profile. It checks both, the
quality of using NDT methodology in each phase of software life cycle and the
quality of traceability of MDE rules of NDT. (c) NDT-Driver, allows transformation
can be automatically applied every phase of the life-cycle.(d) NDT-Prototype is a tool
designed to automatically generate a set of XHTML prototypes from the navigation
models, described in the analysis phase, of a project developed with NDT-Profile. (e)
NDT-Merge uses the comparison among metamodels to identify syntactic and
semantic inconsistencies among different versions of the same requirements
catalogue. In addition, NDT-Suite has more tools: NDT-Report, NDT-Glossary,
NDT-Checker or NDT-Counter. You can see the purpose of these tools on IWT2
website.
3.3 NDTQ-Framework
In the last years, NDT has evolved again and now, in order to offer a suitable and a
global solution for the real application of NDT, a global framework named NDTQ-
Framework 3 was developed. NDTQ-Framework comprises a set of processes
involving development processes, management processes, quality processes, testing
processes and security processes. This environment is based on different reference
models like CMMi (Capability Maturity Model Integration) [17] and ITIL
(Information Technology Infrastructure Library) [0], and its application in real
projects are certificated under different standards like ISO 27001, ISO 9001:2008,
UNE EN 16602 and ISO 14000. This paper does not aim to present NDTQ-
Framework in detail, but you can download more information from IWT2 website.
However, for this paper is necessary to know that the initial management processes
defined by NDTQ-Framework have to be adapted to support the development and the
management of CALIPSOneo.
4. Adapting NDT and management rules
NDT has been applied to several real projects, since it was used in CALIPSOneo.
However, CALIPSOneo provides a new environment with very relevant aspects: (1)
CALIPSOneo is, in fact, a set of three projects. NDT and its management rules have
to be adapted for improving and supporting integration and coordination among these
projects. (2) CALIPSOneo is a special project for NDT application because, in fact, it
2
In IWT2’s website (www.iwt2.org), you can read more about NDT-Suite and its tools.
3
In IWT2’s website (www.iwt2.org), you can read more about NDTQ-Framework.
produces very “few” software. With CALIPSOneo, we are going to obtain a product
that mainly has to cover both a set of requirements and a set of policies to be applied
to PLM. This project is more centered in requirements and work policies than in code.
In this sense, it offers a suitable environment to confirm whether MDE can support
this kind of development. (3) CALIPSOneo has a very heterogeneous team project. In
fact, the team project is distributed around different cities in Spain. This situation
even complicates coordination policies and communication among the different
teams. Finally, several aspects were modified in order to cope with all the
aforementioned points
Figure 2 NDT-Profile for CALIPSOneo
4.1 Managing the coordination
NDT-Profile, NDT-Quality and NDT-Driver were adapted to coordinate these
three teams. Besides, supporting this change involved some management processes in
NDTQ-Framework. Firstly, NDT-Profile was modified extending the original profile
to offer a “project composed by subprojects”. Figure 2 represents the main interface
of NDT-Profile for CALIPSOneo. On the right margin, we can see three different
folders, each of them for one subproject. In the figure, only Elara is opened and the
rest remains the same. In fact, subfolders included in each subprojects constitute the
same folders that we could find in the original NDT-Profile. In this sense, the original
metamodel of NDT and its profile were extended to support the possibility of a
“project composed by projects”. The inclusion of this new aggregation in the
metamodel, caused the necessity of extended NDT-Suite and its tools to support this
new association correctly. In contrast, if we do not offer management rules to teach
our teams how to work in liaison, modifying the profile and tools will not be enough.
Thus, teams in each project work under some specific coordination rules, which are
later explained in section 4.3.
4.2 Managing “Software” products
There are more experiences related to code generation in the context of MDE. In fact,
there is a historical interest in the code among the software community. However, in
CALIPSOneo a full PLM is being developed and in this context, the “code” is no so
relevant. In the context of CALIPSOneo, requirements and analysis models acquire a
very relevant role because they are the key to guarantee the quality of the results. As a
matter of fact, despite we have to develop code, the most important point is how this
new code and PLM tools will work together to get the final PLM. In this sense, NDT-
Profile is offering a very suitable common repository for teams. In the project, we
have an ECM4-based documental repository for storing each document, which stores
each document, PLM policy or software piece of document integrated in NDT-Profile
and referenced via “links”. In addition, different automatic traceability matrixes were
generated in the profile to connect each artifact between phases. Thus, NDT-Profile
integrates the great deal of knowledge of the project.
4.3 Managing “management”
As previously mentioned, these technical mechanisms are not useful without
providing suitable mechanisms for managing the project. In this sense, a set of
management processes were defined in the context of CALIPSOneo to guarantee that
we can manage MDE mechanisms as well as the quality of results.
Thus, as MDE concerns, teams work with the same Enterprise Architect File.
Nevertheless, as they work in different periods and each subproject has different
advances, we have designed a “master file” that is only managed by the quality
management team. Approximately every month, this team gathers all the work
executed by Elara, Proteus and Mars teams in the master file. This involves three
different steps:
1. Firstly, a quality review with the new version of NDT-Quality is executed. In this
step, inconsistencies among models or problems in projects interconnections are
identified. Besides, NDT-Quality checks traceability matrixes in order to
guarantee that subprojects keep consistency between phases and artifacts
2. Secondly, if the teams develop a new phase, NDT-Driver will be executed in
order to “translate” the acquired knowledge along in phases. For instance, in
March 2013, the teams finished the requirements phase in the third integration of
CALIPSOneo. The quality team used NDT-Driver and generated analysis models
following the idea presented in Figure 1. Thus, teams received in this third
integration a first version of the analysis with the automatic generation of their
requirements knowledge thanks to transformations. These transformations not
only generated analysis models, but they also produced traceability matrixes to
save the origin of each artifact in the analysis in relation to requirements.
3. Finally, a new baseline of the master file is generated and stored in Alfresco to
use at different teams’ disposal.
This mechanism of integration is quite useful because every team, even, everyone
in the project, can work independently. However, using the stored connection among
elements generated from NDT-Driver transformations and baselines will help us
easily manage each team’s work. Thus, the integration process does not take the
quality team more than one hour per month. Obviously, if inconsistencies are
identified in step 1, this time will be extended, as the quality team will have to work
with teams to face them.
4
ECM is acronym of Enterprise Contents Management and within CALIPSOneo, the Alfresco
solution has been used. You can find more information about Alfresco in its website.
5. Learned lessons
As previously introduced in the abstract, CALIPSOneo is a real project where MDE
can show all its strengths. Nowadays, although the project is in the analysis phase and
there are long future ways to go, we can offer some initial learned lessons. The first
one is about MDE. The context of the model-driven paradigm is not always easy to
understand by project teams. In fact, concepts like metamodels, transformations and
some others, can be too abstract for the enterprise context. In contrast, if we enrich
this context with suitable tools, such as NDT-Profile or NDT-Driver, they can become
powerful for companies, since they reduce the cost of the project and the number of
errors and inconsistences between the different phases. In MDE, knowledge is
“transferred” from phases and transformations guarantee this transfer, becoming a
powerful resource for companies, if carried out automatically.
Additionally, MDE is not exclusively used in code generation. We have the intention
to demonstrate with CALIPSOneo that we are using it in other phases, being very
profitable in environments not mainly oriented towards code.
Nevertheless, it is clear that MDE, mainly in big projects with heterogeneous and
distributed teams, requires management rules like every software project, which must
ensure the quality of results.
6. Final conclusions and future work
This paper has pointed to how a MDE paradigm is being used in the context of a
real project named CALIPSOneo. As it was presented, CALIPSOneo gives us a very
interesting opportunity to practically measure the power of this paradigm in a context
that offers special characteristics, such as multiple and heterogeneous work teams
without experience or knowledge in MDE, or a context not specifically oriented
towards software. This has let us demonstrate the power of this paradigm in different
contexts and explore new ways to improve our research in this line.
Thus, we are working in a new version of NDT-Merge, which allows us to
automatically detect errors in general, not only those dealing with quality that we
usually identify with NDT-Quality. We are working in the effective comparison of
different instances of requirements definition to find out syntactic and semantic
problems among different versions of requirements.
The “back” is also another open line. NDT-Driver and NDT-Quality are prepared
to detect changes in previous phases. It means that after the analysis generation,
requirements definition can evolve and they can lose the consistency with previous
analysis artifacts generated. Our tools check these evolutions and present these
inconsistencies to the quality team. Moreover, NDT-Driver is prepared for executing
transformations only in new versions of artifacts, keeping the same form in the
original project. Therefore, we have to keep on working in this line because it would
be very interesting that these evolutions could be detected by NDT-Profile, while the
teams are working on it, for instance, through a warning to the engineer.
To conclude, we would like to insist on the importance of this kind of experience
in the context of our research. The effective possibility to test our tools, approaches
and ideas in real contexts with real projects is a required and essential practice to help
us propose new approaches suitable for the enterprise context.
Acknowledgements
This research has been supported by the Tempros project (TIN2010-20057-C03-02),
and by the project NDTQ-Framework (TIC-5789) of the Junta de Andalucía, Spain
and CALIPSOneo Project.
References
1. Schmidt, D.C. “Model-Driven Engineering,” IEEE Computer, Computer Society, vol. 39,
no. 2, pp. 25-31, 2006.
2. UWE (UML-based Web Engineering). http://uwe.pst.ifi.lmu.de/. Last accesed May 2013.
3. WebML (Web Model Language) . http://www.webml.org. Last accesed May 2013.
4. Gutierrez, J.J., Nebut, C., Escalona, M.J., Mejías, M., Ramos, I. Visualization of use cases
through automatically generated activity diagrams. LNCS 5301. pp. 83-96. 2008.
5. Robles, E., Grigera, J., Rossi, G. Bridging Test and Model-driven Approaches in Web
Engineering. LNCS. 5648 (2009) pp.130-150. 2009.
6. Bertolino, A., Gnesi, S. PLUTO: A Test Methodology for Product Families. LNCS. pp
181-197. 2004.
7. Melia, S., Gómez, J. The webSA approach: applying model driven engineering to web
applications. Journal of Web Engineering 5 - 2, pp 121-149. 2006
8. Robles, E., Escalona, M.J., Rossi, G. Modelling the requirements of rich internet
applications in WebRE. Comunications in Computer and Information Science. 2012.
9. Mohagheghi, P., Gilani, W., Stefanescu, A., Fernandez, M. A. An empirical study of the
state of the practice and acceptance of model-driven engineering in four industrial
cases. Empirical Software Engineering, 18(1), 89-116. 2013.
10. Escalona, M.J., Aragón, G. NDT: A Model-Driven Approach for Web requirements, IEEE
Transactions on Software Engineering. Vol. 34. Nº 3. pp 370-390, 2008.
11. Mas, F., Rios, J., Menendez, J.L., Gomez. A., A process-oriented approach to modeling
the conceptual design of aircraft assembly lines, International Journal of Advanced
Manufacturing Technology 2012, Vol. 62,
12. Mas, F., Menéndez, J.L., Oliva, M., Ríos, J., Collaborative Engineering: an Airbus case
study. 5th Manufacturing Engineering Society International Conference. 2013.
13. Mas, F., Gomez. A., Menendez, J.L., Rios, J., Proposal for the conceptual design of
aeronautical final assembly lines based on the iDMU concept. 10th International
Conference on Product Lifecycle Management. 2013.
14. Query/View/Transformation. www.omg.org/spec/QVT/1.1/. Release 1.1. 2011. Accessed
05/2013.
15. García-García J. A., Alba M., García-Borgoñon L., Escalona M. J.: NDT-Suite: A Model-
Based Suite for the Application of NDT, LNCS 7387, pp. 469–472, (2012)
16. Enterprise Architect. www.sparxsystems.com.au. Accessed in May 2013.
17. M.B. Chrissie, M. Konrad, S. Shrum, “CMMI® for Development: Guidelines for Process
Integration and Product Improvement,” Editorial Pearson Education. 2011.
A. Jong, A. Kolthof, “Fundamentos de ITIL, Volumen 3,” Van Haren Publishing, ISBN
9087530609, 2008.