A new metamodel approach of CI/CD applied to

Internet of Things Ecosystem

Badr El Khalyly, Abdessamad Belangour, Mouad Banane, Allae Erraissi
Laboratory of Information Technology and Modeling
Hassan II University, Faculty of sciences Ben M'sik, Casablanca, Morocco
Emails: [email protected], [email protected], [email protected], [email protected]

Abstract— Continuous Integration is a development process A meta-model is a formal definition of a model that helps
where developers and integrators push code into a shared tools to understand it and which facilitates reasoning on its
belonging to a toolchain. This is a kind of practive that permit structure, its semantics, and its use. Meta-modeling, which is
build automation and test automation. This is a primordial the activity of building meta-models, is widely used in the
practise that maintain the continuity of running programs in field of information systems engineering and particularly in
smart agents that constitutes Internet of Things systems. IoT the engineering of models and methods [4,5]. In method
connected objects are new technologies used to connect objects engineering, it is an essential conceptual tool for defining and
to the internet and users. The IoT influences our daily lives in reasoning on new models. In tool engineering, meta-models
several fields such as logistics, health, energy, intelligent
are a formal definition necessary to design and build the tools
vehicles, the smart home, or the smart city, etc. Subsequently,
following the specific characteristics of the different
to edit, verify, transform, and possibly execute instances of
architectures of the Internet of Things, a multitude of IoT these models. Generally, the practice of meta-modeling is to
architectures have been developed and deployed using different specify contemplative metamodels that reflect the static
IoT frameworks. In this paper, we propose a metamodel of the structure of the models, i.e. the concepts and the links
Internet of Things, this proposal aims to standardize existing between these concepts [6]. Depending on the formalism
IoT architectures and those that may be proposed in the future. used, such a meta-model can be supplemented by the
specification of certain constraints. By using UML for
Keywords—Model Driven Engineering, metamodeling, example to build a (meta-) class diagram, it is possible to
Devops, Internet of Thing, Continous Integration, Continous complete this specification with constraints expressed with
Deployment, Continous Delivery. the OCL language. This structure-oriented vision is similar to
the "data" perspective of the reference framework for IS
I. INTRODUCTION engineering models [7]. On the other hand, technological
Continous deployment [1] is the process that follows a advances are becoming more and more important in our
factory to adapt their changes by creating a special build societies. In this context, different cities, cities, and countries
process for their different product. This build process consists around the world are implementing increasingly intelligent
these steps: Fetching dependencies from repos, building services. Becoming a smart city where one becomes aware of
codes, running automated tests (unit, integration, etc.. ) on the protection of the environment and the quality of human
this build, running code analytics, notification (email, sms life, is now the objective of managers of large cities but also
etc..) to developers about possible results (error, success, communities in countries of different continents. This
warnings etc..) and finally packaging the build to next step transition to an intelligent world ensures the sustainable
(Continuous Delivery [2]) As a result of Continous development of our urban environments as well as an
Integration [3], there is: improvement in the quality of life and security of citizens
while ensuring respect for their private lives. The vision of an
• A ready portable product, intelligent world provides that the entire urban space is
• Avoiding of deployment problems, interconnected so that it can interact with the digital world
and therefore the adoption of the Internet of Things (IoT)
• Regularly execution unit tests, code reviews, code- paradigm. Indeed, the IoT makes it possible to interconnect
analytics, bug detection, different objects, machines, sensors that can be carried by
• Good collaboration infrastructure between human beings to offer a new range of services and
developers, applications. By interconnecting different types of objects
using new adapted technologies and by using the capabilities
• Good quality and user experience of product. offered by cloud computing, the IoT makes life easier by
automating certain daily tasks. Similarly, the IoT enables
critical services to be offered remotely such as medical
services using the capabilities of the Internet. Thus, the
impact of IoT on our societies will be significant and will
allow improving the quality of life while reducing the costs
of setting up services. For example, taking charge of health
services through the IoT reduces the burdens on communities
while improving patient follow-up.

Fig. 1. Devops Steps. This article is organized into 8 sections. Section 2 is a

state of the art on the expression of executability in meta-
modeling related work. Section 3 very briefly recalls the
meta-modeling approach and the formalisms used and some

978-1-7281-6921-7/20/$31.00 ©2020 IEEE

basic notions of IoT. Section 4 is the first illustration of our combination of a product meta-model and a process meta-
approach to the Internet of Things metamodel. Section 5 is a model, the product specification is historically the oldest. [18]
proposition of Devops Meta-Model. Section 6 is the second Regarding the process, it is the approach by assembling
illustration of the approach and discusses this proposition, the methodological components that are the most used. In [19],
last section concludes the work and gives some perspectives. the language MEL1, a formal language for the specification
of methods, is proposed. Apart from the structural
II. RELATED WORK
specification of the components, the process aspect is
It is in the field of software engineering that the problem of described in MEL as formal operators whose semantics are
the expression of executability in meta-models has been guaranteed by the underlying mathematical notation.
studied in depth, thanks in particular to the development of This approach by assembling methodological components
the IDM approach. Indeed, since IDM is fundamentally based currently remains predominant [20], however, we wonder
on the extensive use of models in all phases of software about the models used to formalize the approach, to specify
development, the question quickly arose of how to execute a the content of a methodological component, and to express
model and how to express the executable semantics of that - the assembly process [21]. Finally, a new research
this. In [8], a first reflection on the link between meta-model perspective in this sense is that of defining the
and the expression of the executability of models was made methodological components as being services [22].
on Petri nets. The authors supplement the static meta-model A methodological component is directly executable since it is
which describes the fixed structures in a model (arcs and a service that is written with SOA standards, while the
transitions in a Petri net), by a dynamic meta-model that composition of services is expressed informally at a high
describes the data structures necessary during the execution level of abstraction using intentional Map. Banana et al. [23]
of an instance of this model (markings and token proposed a metamodel of Hive [24] and Spark [25], as well
movements). The authors recognize however that this as another metamodel [26] of MongoDB query language. The
addition is not sufficient to express all the executability, IoT is responsible for the collection and / or creation of a large
because the formalism used (UML class diagram) does not volume of data. This enormous volume of data, known by the
itself have executable semantics, and they thus call for the term "Big Data" [27,28], allows, on the one hand, to have an
creation of an Executable UML. And it is probably as a result incredible wealth in terms of information allowing the offer
of these preliminary reflections on the expression of of advanced services. On the other hand, this volume of data
executability that the Kermeta language was developed [9]. creates new challenges to be considered such as securing,
Kermeta complements UML meta-diagrams in the form of a processing, and real-time accessibility of this data [29,30].
directly operational meta-specification. Combined with the To end this overview, we must mention the metalisms and
principle of dynamic meta-classes introduced previously [8], meta-modeling languages offered tools and meta-CASE type
this approach makes it possible to construct a complete meta- environment. MetaEdit is a well-known and popular tool
specification for a model. The Kermeta language has so far [31], it allows you to specify a static meta-model with the
been used to construct a complete and executable description GOPRR2 formalism, and immediately generate a graphic
of a simple model, that of the finite state machine [10], and editor for the model [32]. It is intended for the creation and
tested in the context of embedded software systems to specify tooling of new domain-specific languages [33]. As far as the
meta-models of UML 2.0 in Kermeta [11]. "processing" and "behavior" perspectives are concerned, they
Other work in the software engineering and IDM community are relegated to the code generation phase where a scripting
has focused on engineering process models (called software language (called Merle) makes it possible to browse and
process models), given the importance of describing, manipulate the instances of the model and to generate
controlling, and automating the procedures with which instructions in n ' any target language (HTML, XML, C ++,
software systems are built. An important work in this register Java, etc.). The executable semantics is thus expressed by the
is that around UML4SPM, which defines an engineering transformation of the static and non-productive structures of
process modeling language which is based on UML and close the model into a set of software instructions (or components)
to the OMG SPEM model [12], [13]. Several experiments are in another language or another model whose executability is
made to specify the semantics and express the executability known and supported by a platform. target shape. While the
of this language: first, using the BPEL business process definition of the meta-model is done in a declarative way with
execution language [14], and then using the Kermeta meta- a graphical interface, the executability is expressed
specification language [15]. For these two approaches, the operationally with a programming type interface. This is the
issue of interaction with the outside world (the user or other main drawback of MetaEdit.
systems) is underlined as that which distinguishes the two Another meta-modeling formalism supported by a meta-
approaches. Despite several advantages (notably the CASE is the ConceptBase environment [34] built around the
existence of reliable technical systems for the execution of Telos model and language [35]. Built with the declarative
BPEL models), the use of BPEL is not considered satisfactory logic language Datalog, ConceptBase is an extremely
because of the absence of concepts to take into account the powerful meta-modeling environment which allows you to
interaction with the user. Other research, more oriented specify any number of abstraction levels and to express
towards the comparative study of approaches, completes this constraints and requests on several of these levels (and not on
work around the executability of a process model [16, 17]. one level as in OCL). The specification is entered verbatim,
In the field of IS engineering and that of method engineering, but the content of the repository is displayed graphically at the
in particular, few works to our knowledge have addressed the request of the user. In terms of "processing" and "behavior"
question of the explicit expression of executability in a perspectives, ConceptBase has introduced in its most recent
process meta-model. The definition of a method being the versions, rules of the Event-Condition-Action (ECA) form to
express the dynamics of a meta-model. An illustration is A metamodel is a modeling language. It defines the concepts
proposed in [36] with the rules for the execution of a Petri net. as well as the relationships between concepts necessary for
the expression of models. The notion of metamodel leads to
III. BACKGROUND the identification of a second relationship, linking the model
In this section, we explain the fundamentals of the two axes: and the language used to build it, called "conformA". These
model-driven engineering (MDE) and the Internet of Things two relations thus make it possible to clearly distinguish the
(IoT). language which plays the role of the system, from (or)
metamodel (s) which play the role of model (s) of this
A. MDE
language.
Model Driven Engineering (MDE) [37] has allowed several 3) Meta-metamodel concept
significant improvements in the development of complex A meta-metamodel is a metamodeling language used to
systems by allowing to focus on a more abstract concern than express metamodels. This is the case, for example, for the
classical programming [38]. This development vision based MDA (Model Driven Architecture) proposal [45].
on “everything is model” is framed by a few concepts. In fact, Metamodeling is considered to be the activity of representing
the main modeling and meta-modeling work was carried out modeling languages through models. Indeed, this activity has
by the working groups of the various standardization bodies enabled the emergence of the notion of DSMLs (Domain
(OMG, UML [39], conceptual graphs [40]) and by groups Specific Modeling Languages, literally domain-specific
interested in exchanging models. [41], or the transformation modeling languages). We currently have many languages and
of models [42]. The needs of common language and the needs environments for metamodeling, we can cite for example
of exchange of information between the modeling tools led to MOF (Meta Object Facilities) and its variations (EMOF: for
a reflection on the modeling objects and also on their meta- Essential MOF, CMOF - for Complete MOF), EclipseEMF /
models. The different definitions presented in this section are Ecore.
established with reference mainly to the work of [40]; 4) Model transformation concept
[41,42,43] and obviously the MDA specification (Diaw et al. A transformation is the automatic generation of a target
[44] Favre and [45]) from the OMG. We also refer to the model from a source model, according to a transformation
summary work on IDM proposed by Diaw et al [44]. definition. A transformation definition is a set of
For several years, there has been consensus on an architecture transformation rules that describe how a model expressed in
with four levels of modeling: the source language can be transformed into a model of the
 M0 : Data. target language. A transformation rule is a description of how
 M1 : Model. one or more elements of the source language can be
 M2 : Metamodel. transformed into one or more elements of the target language.
 M3 : Meta-metamodel. Therefore, the model transformation consists in manipulating
the models through transformations, these transformations
 Meta-metamodel (M3)
ensure the passage from one or more source models
the meta-metamodel (M3) represents the most abstract level,
(represented in a given level of abstraction) to one or more
it represents the language for defining meta-models. It
target models (represented ( s) in a new scale of abstraction).
defines basic concepts that will allow the representation of all
The transformed model is called the source model and the
the other levels as well as themselves. Examples are Meta-
model resulting from the transformation is called the target
class, Meta-attribute, Meta-Operation in the case of UML;
model.
MetaEntity with (meta) attribute, and meta-Relation with
The model transformation is therefore carried out according
(meta) attribute for CDIF; and concept, conceptual
to three basic principles:
relationship, and graph for conceptual graphs.
• a source meta-model which defines the syntax of the
The metamodel (M2) defines the language of representation
instance models of the source formalism;
or formalism of the models. Examples are Class, Attribute,
• a target meta-model which defines the syntax of the
Operation for UML; Entity, Attribute, Relation for the
instance models of the target formalism;
formalism Entity-relation; and Concept type, and
• a transformation definition, which specifies the rules
Relationship type for conceptual graphs.
for syntactic matching of the model elements of the
The model (M1) defines the representation language of the
source meta-model to those of the target meta-
domain studied. Examples are employees, organization,
model.
Customer, Mission.
1) Model concept
A model is an abstract and simplified representation (i.e.
which excludes certain details) of an entity that can represent
a process or a system of the real world in order to describe,
explain or predict it. A model is also understood as an
abstraction of a system in the sense that it reflects one or more
aspects 2 of a system or a part of it. In other words, a model
reflects in a few spells what the designer sees as important for
understanding and predicting the future system. Also, an
abstract model of the system makes it possible to reflect on Fig. 2. Model transformation concept.
it.
2) Metamodel concept
 5) Model-to-model transformations (M2M) areas. They are based on different techniques such as content-
Model-to-model transformations (i.e. from an abstract syntax based filtering, collaborative filtering, and hybrid filtering.
to an abstract syntax) consist of transforming a source model Content filtering offers a service similar to the services the
into a target model. These models can be instances of user has used in the past. This is hardly suitable for proposing
different metamodels. They offer more modular and easy-to- new services.
maintain transformations. In MDA cases for example where Collaborative filtering methods consist of recommending
there is a large abstraction space between PIMs (Platform services to a user based on those of other users with similar
Independent Model) and PSMs (Platform Specific Model), it behavior. These are slow when the size of the data becomes
is easier to generate intermediate models than to go directly large. Hybrid filtering combines the two previous
to the target PSM. Intermediate models can be useful for approaches. A new approach to recommending IoT services,
optimization or for bug fixing purposes. In addition, model- based on modeling with a tripartite graph linking objects,
to-model transformations are useful for calculating the users, and services, is proposed in [46]. This method ranks
different system views and their synchronizations. services to recommend them to the user. However, it does not
6) Model to code transformations (M2C) solve the problem of ambiguity for different services using
In this category we seek to scan the source model to the same types of connected objects.
generate an output code which is generally described in a
C. Devops
programming or structuring language. Here, we are talking
about transformation from an abstract syntax to a concrete DevOps is a software development and distribution process.
syntax. In this type, there are two approaches: approaches It combines an entire culture of collaboration and fusion
based on the visitor principle (Visitor-based approach) and between management approach, technologies of coding and
those based on the boss principle. technics of integration. DevOps is a new paradigm and novel
concept that is introduced in different industries that concerns
B. IoT
software and embedded development like robotics and smart
The Internet of Things (Internet of Things, IoT) is a paradigm agent.[47]
based on the interconnection of connected objects deployed
in different physical spaces. These objects communicate with IV. PROPOSED IOT METAMODEL
each other while interacting with the outside world in order The engineering approaches developed within the framework
to offer IoT services to users such as medical control, house of Information Systems Engineering (ISI) have often
control and monitoring, temperature control, etc. In this consisted of proposing models, procedures, and tools to
article, we consider an IoT service as a set of connected facilitate their use. In this area, the models offered are
objects deployed in one or more spaces that interact with the traditionally contemplative, they are only used to facilitate
physical world through their physical interfaces and between communication between the different actors of the project.
them through their network interfaces. These connected Productive models naturally emerged with process
objects are more and more numerous and diversified and engineering and became widespread with the adoption of
many new IoT services are introduced. As a result, the user is OMG work on MDA architecture). Indeed, the software
not aware of all the objects connected around him or of all the engineering community has massively refocused on the
possible IoT services from which he could benefit. It is concept of model in order to exploit their productive nature
therefore essential to define a recommendation system by advocating a model-centered vision of engineering tools,
offering the user personalized IoT services and helping them which has enabled the emergence of a new field: model-
choose the appropriate connected objects for the composition driven engineering (IDM). Figure 2 illustrates the proposed
of these services, according to those available to them. IOT meta-model:
Service referral systems have been widely used in many

Fig. 3. IoT metamodel proposed.

V. PROPOSED DEVOPS METAMODEL Continous Delivery incorporates release step in addition to

Continous Integration. Continous Deployment contains steps
As depicted in the Figure 1 Agile Development contains
of deployment in addition to Continous Delivery.
Coding and building technologies. Continuous Integration
This description is depicted in this metamodel:
Agile Processes in addition to test and integration steps.
 VII. CONCLUSION
In recent years connected objects have experienced veritable
democratization with more and more connected accessories
arriving in our daily lives, our homes, our work environment
but also cities. Without knowing it, we come across everyday
connected objects integrated into our environment. In this
paper, we have proposed a new metamodel of the Internet of
Things, this proposal aims to standardize existing IoT
architectures and those that may be proposed in the future.
Fig. 4. Devops metamodel proposed.
REFERENCES
VI. DISCUSSION
[1]. Mojtaba Shahin, Muhammad Ali Babar, and Liming Zhu.
The proposal for an Internet of Things metamodel has many Continuous integration, delivery and deployment: A
of the advantages that are listed as follows: standardizing IoT systematic review on approaches, tools, challenges and practices.
architectures with micro-services, improving interoperability IEEE Access, 5:3909–3943, 2017.
[2]. Eero Laukkanen, Juha Itkonen, and Casper Lassenius. Problems,
between different architecture packages. causes and solutions when adopting continuous delivery—a
We offer a meta-model for an ecosystem of connected systematic literature review. Information and Software Technology,
objects. This meta-model is an improvement in the meta- 82:55–79, feb 2017.
model which is cited in the state of the art. We have added [3]. Daniel Ståhl, Kristofer Hallén, and Jan Bosch. Achieving traceability
some classes which will be useful in the approach based on in large scale continuous integration and
the meta-models proposed in our work. It is a meta-model of delivery deployment, usage and validation of the eiffel framework.
Empirical Software Engineering, 22(3):967–
connected objects that will be manipulated using Docker 995, oct 2016.
containers that will be deployed at the level of [4]. Rolland, Colette. "L'ingénierie des méthodes : une visite guidée."
microcontrollers. We have kept some existing components in (2005).
the meta-model cited in the state of the art. they are as [5]. Rolland C., « Capturing System Intentionality with Maps », dans
Krogstie J.; Opdahl A., Brinkkemper S., Conceptual Modelling in
follows: Information Systems Engineering, SpringerVerlag, Berlin
Component is the basis of connected objects, it is either a Heidelberg, Germany, 2007, pp. 141 - 158.
sensor or an actuator. Sensor according to the Instrument [6]. Jeusfeld, Manfred A., Matthias Jarke, and John Mylopoulos.
Society of America, a sensor can be defined as "a device that Metamodeling for method engineering. the MIT Press, 2009.
provides usable output in response to a specified [7]. Irwin, Christine E. "Information Systems Methodologies: A
measurement. Actuator is a component of a machine that is Framework for Understanding." Journal of the Operational Research
responsible for moving or controlling a mechanism or a Society 43.3 (1992): 286-286.
system. IoTNode is a robot or a smart object which is made [8]. Breton, Erwan, and Jean Bézivin. "Towards an understanding of
up of sensors and actuators. Controller This is the device on model executability." Proceedings of the international conference on
Formal Ontology in Information Systems-Volume 2001. 2001.
which the IoTNode is connected. Communication is the
[9]. Muller, Pierre-Alain, Franck Fleurey, and Jean-Marc Jézéquel.
protocol adopted by the microcontroller such as: WIFI, "Weaving executability into object-oriented meta-languages."
Bluetooth, ZigBee etc. Domain is the domain to which the International Conference on Model Driven Engineering Languages
IoT ecosystem belongs, such as Smart Industry, Smart Farm, and Systems. Springer, Berlin, Heidelberg, 2005.
Smart Home, Smart Vehicle. IoTEcosystem is a set of [10]. Online http://www.kermeta.org.
interconnected smart objects forming a large We have added [11]. Koudri, Ali, Joël Champeau, and Denis Aulagnier. "Une sémantique
the following components: opérationnelle pour une meilleure méta-modélisation." 3e journées sur
l’Ingénierie Dirigée par les Modèles, France (2007): 223-228.
GeoPlace is the geographic location where a [12]. Bendraou, Reda, Marie-Pierre Gervais, and Xavier Blanc.
microcontroller is positioned. ActuatorCategory is the "UML4SPM: A UML2. 0-based metamodel for software process
category to which an actuator belongs such as Motor, Bulb, modelling." International Conference on Model Driven Engineering
Screen, Intelligent Hand, Wheel. SensorCategory is the Languages and Systems. Springer, Berlin, Heidelberg, 2005.
category to which a sensor belongs such as humidity, [13]. Bendraou, Reda, et al. "Software process modeling and execution: the
temperature, acidity. UML4SPM to WS-BPEL approach." 33rd EUROMICRO Conference
on Software Engineering and Advanced Applications (EUROMICRO
Devops tools are in service of Internet of Things 2007). IEEE, 2007.
ecosystem to guarantee the continuous integration, delivery [14]. Bendraou, Reda, et al. "Definition of an Executable SPEM 2.0." 14th
and deployment of programs that process smart agent Asia-Pacific Software Engineering Conference (APSEC'07). IEEE,
2007.
blonging to Internet of Things Ecosystem. There is a
[15]. Bendraou, Reda, et al. "Vers l'Exécutabilité des Modèles de Procédés
dependency between the two metamodels as depicted in the Logiciels." 2008.
figure 5: [16]. Combemale, Benoît, et al. "Expériences pour décrire la sémantique en
ingénierie des modèles." Hermes SCIENCES/LAVOISIER, éditeur:
2ième journées sur l’Ingénierie Dirigée par les Modèles (IDM)
(2006): 17-34.
[17]. Combemale, Benoit, et al. "Towards Rigorous Metamodeling."
MDEIS 6 (2006): 23-27.
[18]. Harmsen, Frank, and Motoshi Saeki. "Comparison of four method
engineering languages." Working Conference on Method
Fig. 5. Devops metamodel proposed. Engineering. Springer, Boston, MA, 1996.
[19]. Rolland, Colette. "Capturing system intentionality with maps." [40]. Bézivin, J. and Gerbée, O. (2001). Towards a precise definition of the
Conceptual modelling in Information Systems engineering. Springer, omg/mda frameword. the 16th International Conference on
Berlin, Heidelberg, 2007. 141-158. Automated Software Engineering, ASE, page 273–280.
[20]. Henderson-Sellers, Brian, and Jolita Ralyté. "Situational method [41]. Bézivin, J. Who is afraid of ontologies ? In OOPSLA’98 Workshops,
engineering: state-of-the-art review." Journal of Universal Computer Vancouver, Canada. (1998)
Science (2010). [42]. Bézivin, J., Lennon, Y. and Nguyen, C. From cobol to omt : A
[21]. Seidita V., Ralyté J., Henderson-Sellers B., Cossentino M., Arni- reengineering workbench based on semantic networks. In Tools
Bloch N., « A comparison of deontic matrices, maps and activity USA’95, Santa Barbara, CA, USA, pages 137–152.
diagrams for the construction of situational methods », CAiSE Forum, [43]. Kleppe, A.G., Warmer, J. and Bast, W. (2003). Mda explained : The
19th Int. Conf. on Advanced Information Systems Engineering, model driven architecture : Practice and promise.
Trondheim, Norway, 11-15 June, 2007
[44]. Diaw, S., Lbath, R. and Coulette, B. (2010). Etat de l’art sur le
[22]. Rolland C., « Method engineering: towards methods as services », développement logiciel basé sur les transformations de modèles.
Software Process: Improvement and Practice, vol. 14, n°3, pp. 143- Technique et Science Informatiques, Ingénierie Dirigée par les
164, 2009 Modèles, 29(4-5):505–536.
[23]. BANANE, Mouad, ERRAISSI, Allae, et BELANGOUR, [45]. Favre, J.M., Estublier, J. and Blay-Fornarino, M. (2006). L’ingénierie
Abdessamad. SPARQL2Hive: An approach to processing SPARQL dirigée par les modèles : au-delà du mda. page 273–280.
queries on Hive based on meta-models. In : 2019 8th International
Conference on Modeling Simulation and Applied Optimization [46]. Ibrahim Masha, Tein-Yaw Chung, and Osama Alsaryrah. Toward
(ICMSAO). IEEE, 2019. p. 1-5. service recommendation in Internet of Things. In 2015 Seventh
International Conference on Ubiquitous and Future Networks, pages
[24]. Banane, Mouad, and Abdessamad Belangour. "A new system for
massive RDF data management using Big Data query languages Pig, 328–331, July 2015
Hive, and Spark." International Journal of Computing and Digital [47]. Len Bass. The software architect and DevOps. IEEE Software,
Systems 9.2 (2020): 259-270. 35(1):8–10, jan 2018.
[25]. BANANE, Mouad et BELANGOUR: A Big Data Solution To Process
Semantic Web Data Using The Model Driven Engineering Approach,
International Journal of Scientific & Technology Research. vol. 9, no
02, 2020.
[26]. Banane, Mouad, and Abdessamad Belangour. "Towards a New
Scalable Big Data System Semantic Web Applied on Mobile
Learning." International Journal of Interactive Mobile Technologies
(iJIM) 14.01 (2020): 126-140.
[27]. A. Erraissi, B. Mouad and A. Belangour, "A Big Data visualization
layer meta-model proposition," 2019 8th International Conference on
Modeling Simulation and Applied Optimization (ICMSAO),
Manama, Bahrain, 2019, pp. 1-5. doi:
10.1109/ICMSAO.2019.8880276
[28]. Erraissi Allae, et Abdessamad Belangour. « A Big Data Security
Layer Meta-Model Proposition ». Advances in Science, Technology
and Engineering Systems Journal 4, nᵒ 5 (2019).
https://doi.org/10.25046/aj040553.
[29]. Erraissi, Allae, and Abdessamad Belangour. "An Approach Based On
Model Driven Engineering For Big Data Visualization In Different
Visual Modes." International Journal of Scientific & Technology
Research.
[30]. Fatima Kalna, Allae Erraissi, Mouad Banane, Belangour “A Scalable
Business Intelligence Decision-Making System in The Era of Big
Data” International Journal of Innovative Technology and Exploring
Engineering 2019. https://doi.org/10.35940/ijitee.L3251.1081219
[31]. Kelly, S., Lyytinen, K., Rossi, M., « MetaEdit+: A Fully Configurable
Multi-User and MultiTool CASE Environment », Proceedings of 8th
Int. Conf. on Advanced Information Systems Engineering
(CAiSE'96), LNCS 1080, Springer-Verlag, pp. 1–21.
[32]. Online http://www.metacase.com/.
[33]. Kelly S., Tolvanen J., Domain-Specific Modeling: Enabling Full
Code Generation, Wiley & Sons, New Jersey, 2008
[34]. Jarke M., Jeusfled M.A., Nissen H.W., Quix C., Staudt M., «
Metamodelling with Datalog and Classes: ConceptBase at the Age of
21 » 2nd Int. Conf. Object Oriented Data Bases (ICOODB’09), LNCS
5936, pp.95-112, Springer, 2010
[35]. Mylopoulos J., Borgida A., Matthias J., Koubarakis M., « Telos:
representing knowledge about information systems », ACM
Transactions on Information Systems, vol. 8, Issue 4
[36]. Jeusfled M.A., Jarke M., Mylopoulos J., (Eds), Metamodeling for
Method Engineering, The MIT Press, 2009.
[37]. Lämmel, R., Saraiva, J. and Visser., J. Generative and
transformational techniques in software engineering. 4143:36–64.
(2006)
[38]. Combemale, B. Ingénierie dirigée par les modèles (idm) État de l’art.
Rapport technique, Institut de Recherche en Informatique de Toulouse
.2008.
[39]. Booch, G., Rumbaugh, J. and Jacobson, I. (1994). Unified modeling
language. Rapport technique, Rational Software Corporation.