A model-driven approach for developing a model repository: Methodology and tool support

Abstract

h i g h l i g h t s• Repository design: we describe a design framework and infrastructure to support the model-based repository lifecycle.• Modeling: we propose a modeling language to specify a model-based repository independent from end-development applications and execution platforms.• Tooling: we describe an operational architecture for development tools to support the proposed approach.• Validation: we apply the approach to a resource-constrained embedded system (RCES) in the context of the TERESA project : Railway Systems. Keywords:Modeling artifact Repository Meta-model Model-driven engineering Security & dependability patterns a b s t r a c t Several development approaches have been proposed to cope with the increasing complexity of embedded system design. The most widely used approaches are those using models as the main artifacts to be constructed and maintained. The desired role of models is to ease, systematize and standardize the approach to the construction of software-based systems. To enforce reuse and interconnect the process of model specification and system development with models, we promote a model-based approach coupled with a model repository. In this paper, we propose a model-driven engineering methodological approach for the development of a model repository and an operational architecture for development tools. In addition, we provide evidence of the benefits and feasibility of our approach by reporting on a preliminary prototype that provides a model-based repository of security and dependability (S&D) pattern models. Finally, we apply the proposed approach in practice to a use case from the railway domain with strong S&D requirements. * Fax: +33 (0)5 6150 4173. E-mail address:Hamid@irit.fr. the industrial efficiency, fostering technology reuse across domains (reuse of models at different levels), and thus reducing the amount of effort and time needed to design a complex system.Repositories of modeling artifacts have gained increasingly attention recently to enforce reuse in software engineering. In fact, repository-centric development processes are often adopted in software/system development, such as architecture-or patterncentric development processes.According to Bernstein and Dayal [2], a repository is a shared database of information on engineered artifacts. These researchers state that a repository has (1) a Manager for modeling, retrieving, and managing the objects in a repository, (2) a Database to store the data and (3) Functionalities to interact with the repository. In our work, we go one step further by proposing a model-based repository to support the specifications, definitions and packaging of a set of modeling artifacts to assist the developers of trusted applications for embedded systems. Closely related to our vision is the approach for specifying, designing and implementing a reuse repository presented in [3].

Reusable Security Requirements Repository Implementation Based on Application/System Components

“…Model-driven engineering approaches also commonly depend on the repository usage. Hamid [21] used models to ease, systematize and standardize the software development process, not only the requirements phase. The repository stored domain-specific modeling artifacts.…”

Engineering secure systems: Models, patterns and empirical validation

“…The repository is implemented using the Eclipse CDO 8 framework. We use the same process flows for the design and implementation of a reuse model repository such as the one described in Hamid (2017) . We apply metamodeling techniques that enable the specification of the repository structure and interfaces to content in the form of modeling artifacts and model transformation techniques for the purpose of generation.…”

“…All metamodels are specified using the EMF. The repository is implemented using the approach described in ( Hamid, 2017 ), which is based on the Eclipse CDO 2 framework. In our work, the target domain development environment is IBM Rational Rhapsody 3 , and the descriptions of the model transformations are based on the QVT operational language ( OMG, 2011 ).…”