International audience

In the context of complex system development, heterogeneous modeling responds to the need to integrate several domains. This need requires the use of the most appropriate formalism and tooling for each domain to be efficient. Model federation promotes the semantic interoperability of heterogeneous models by providing the means to reify correspondences between different model elements, add custom behaviors and bridge the gap between technological spaces. As such, it can be used as an infrastructure to address many different system engineering problems. This is what we have been doing for over a decade, as part of a close collaboration between a small software engineering startup and academia. This paper reports on this experience.

International audience

With the ever growing digitization of activities, software systems are getting more and more complex. They must comply with new usages, varied needs, and are permanently exposed to new security vulnerabilities.

Cybersecurity is becoming vital as industries are gradually moving from automating physical processes to a higher level automation using cyber physical systems (CPS) and internet of things (IoT). In this context, security is becoming a continuous process that runs in parallel to other processes during the complete life cycle of a system. Traditional threat analysis methods use design models alongside threat models as an input for security analysis, hence missing the life-cycle-based dynamicity required by the security concern. In this paper, we argue for an attacker-aware systems modeling language that exposes the systems attack surfaces. For this purpose, we have designed Pimca, a domain specific modeling language geared towards capturing the attacker point of view of the system. This study introduces the formalism along with the Pimca workbench, a framework designed to ease the development and manipulation of the Pimca models. Finally, we present two relevant use cases, serving as a preliminary validation of our approach. © Copyright 2020 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved.

International audience

Modern software and system collaborative process involves various teams in different development phases thus need efficient solutions for tools integration. In Model-Driven Development, transformation technique is used to allow exchanging models created by different tools. However, in a process, transformation are often defined manually for a tool-incompatible point and rarely reusable. To facilitate the automatic generation of transformation rules for tool integration, we propose to use process ontology together with process editor when modelling process. The idea is using ontology to stock process assets from various sources so that the relations between similar elements in different technical spaces can be established automatically. The process editor enriches the ontology by process elements captured from modelling activities. Then the integrated ontology helps the editor detect tool integration points and complete the process model as well as generate the mappings between concerned process elements.

The concept of System of Systems (SoS) has become a key to handle the complexity of large scale systems designed in domains such as aerospace or defense. Virtual prototyping, by providing modelling and execution, increases the success of the SoS. Large scale systems involve heterogeneous modelling and execution formalism. Typical approaches based on types fail in offering the needed dynamicity to answer the virtual prototyping requirements. In order to bridge the existing gap between system models and execution tools, we suggest an approach based on roles and parser combinatory. Roles provide a dynamic coupling between modelling and execution allowing independent evolution of modelling and execution. Parser combinatory ease the integration of new modelling formalisms. This approach has been validated on a seafloor observatory project and lead to the implementation of a framework in Smalltalk.

Model-Based Engineering (MBE) and Product Line Engineering (PLE) have been combined, to handle new system development constraints like: increasing complexity, higher product quality and cost reduction. Many authors have pointed out the need of modularization in the variability and in the core assets space. Existing approaches focus on separating and delimiting concerns or providing generic composition mechanisms. In Model-Based Product Lines, with an increasing number of models to manage, organizing the modeling space becomes central to support product line consistency, maintenance and product derivation process. To organize the modeling space, we propose to precisely describe the dependencies among modeling artifacts and clarify their use. Thus, we introduce the Product Line Modeling Space (PLiMoS) language that specializes relationships, based on an intentional framework, for the product line domain. The Domain Specific Language (DSL) provides a solution to model the modeling space and preserves independence with the product line tooling.

Existing approaches to tool integration based upon metamodels encounter the problem that common (integration) metamodels become too large and complicated, due to the large number of different tools that have to be integrated, each with their specific metamodel defining models handled by each tool. The approach taken here is radically different in that it distinguishes between the model for integration and the tool metamodels, and that it provides mechanisms by which the common integration model and the various tool metamodels can be related. In addition the integration model covers the need for expressing commonalities between models in different languages. Integration is done by means of Artifact resources that are representatives of the real models and model elements. Pure Artifact resources have properties that are required for simple tool integration like trace ability. In order to cover transformations, special Artifact resources for models and model elements in languages defined by metamodels will have properties identifying the metamodel of the language and the type of model element. In cases where integration is based upon concepts that are common to a set of languages, the common concepts are represented by roles played by Artifact resources, in general the same real model element may therefore play different roles for different integration purposes. The integration approach is validated by three experiments: trace ability, transformation, and common concepts for the purpose of software-hardware co-design. The approach is independent of specific implementation platforms. The paper demonstrates how the approach is realized based upon OSLC [24].

Model-Based Engineering (MBE) and Product Line Engineering (PLE) have been combined, to handle new system development constraints like: increasing complexity, higher product quality, faster time-to-market and cost reduction. As observed by some authors, the derivation of a product from product line shared core assets has been insufficiently addressed and can remain tedious in practice. We cope with this issue focusing on having a flexible and reactive model-based derivation, and propose an incremental evolution by extension of the product line coupled with this derivation activity. Process and tools bridge the gap between Application and Domain Engineering introducing a semi-automatic feedback to benefits from the developments made in the Application Engineering. The approach is applied to a model-based product line dedicated to Class diagrams, and is tooled within the Eclipse environment.

In this paper, we present different modeling and execution frameworks that allow us to efficiently analyze, design and verify complex systems, mainly to cope with the specific concerns of the Real-time and embedded systems (RTE) domain. First we depict a UML /MARTE based methodology for executable RTE systems modeling with a framework and its underlying model transformations required to execute UML models conforming to the MARTE standard. The advantages of adopting a more generic action language with formal features are highlighted, in order to raise the level of abstraction with formal features. Then, we investigate how MARTE, with its Time Model facilities, can be made to represent faithfully AADL periodic/aperiodic tasks communicating through event or data ports, in an approach to end-to-end flow latency analysis. An analytical framework allows us to optimize port-based communication by generating a run time executive that utilizes shared data areas where appropriate, while ensuring the timing semantic assumed by the control application. An analysis of the AADL mode change protocol is also provided, exposing a translation process that takes as an input an AADL model and produces as an output a time Petri net. We show how an AADL model transformation provides a formal model for model checking activities and we suggest that model transformation provides useful support to improve the integration of formal verification in an industrial engineering process. As a case study we use an implementation of a UDP /IP protocol stack.

The paper presents an AADL model transformation. This transformation provides a formal model for model checking activities. This transformation covers three aspects: structure, behaviour description and execution semantics. The necessity to complete AADL metamodel is shown in order to improve system behaviour. Transformation rules take into account these aspects. The Kermeta metamodeling platform is used to implement these rules which are applied to a case study. In conclusion, we suggest that model transformation provides useful support to improve the integration of formal verification in an industrial engineering process.