Usine Logicielle

The objective of the sub project OpenDevFactory is to supply a standard platform for integrating technological developments for modelling software tools. This sub project produces technological components on top of which domain tools (automobile, security, telecommunication, aeronautical…) can be derived at a lesser effort. That platform will be built as an interoperable federation of tools which limited parts could be deployed to make specialised IDEs meeting the particular needs of different kinds of users.
The technological bricks are organized as fol-lows:

 

 

The integration structure of OpenDevFactory is build on top of the Eclipse framework. On this structure, the generic and specific engineering components are case by case integrated depend-ing on the needs of the industrial project’s case studies.
There is a significant difference between providing an environment to support case studies and providing it in an industrial operational context. Numerous questions, issues and opportunities are raised by realising true model driven engineering. They need a deep analysis which can be realised with the help of industrial experimentations. Thanks to OpenDevFactory, the emerging themes that we can investigate are numerous. Among these, the project will focus on the following:

 

Engineering heterogeneity:
The objective is to integrate in the same engineering process various types of competing engineering disciplines. The technical domain chosen for this experimentation is real time embedded. In order to design such systems and to satisfy the constraints of volume or of resources consummation (time, memory and energy…), it is necessary to make use of several specific engineering disciplines in the context of MDE such as performance engineering, test engineering and embarcability engineering.

 

Discontinuity of engineering processes :
Some system classes need to integrate in a same design process different interdependent engineering disciplines: mechanic, physic, electronic, computers etc. These kinds of complex systems are not only software predominant because each solution element brought to the system architecture by exercising an engineering discipline impacts the nature of the problem to deal with by the other engineering disciplines.
The issue which is discussed in the project is to make the iterative process resulting of this situation seamless by bringing a formal support to the continuity of the engineering processes.

Models heterogeneity:
The support for obsolescence management does not only concern the electronic components and computers, it also includes engineering tools used by companies in their system and software IDEs. For software embedded in long life time products such as nuclear centrals or aircrafts, the durability of these tools producing these software can be less important than the products inside of which they are integrated. We are then confronted to a problem in MDE which is to transform in the most efficient way possible the model of the embedded software from the obsolescent tool into a model for the replacement tool without modifying the behaviour of the software defined by the model to be translated. Just like the issue of interoperability between engineering disciplines is sorted out, the model transformation tools of the OpenDevFactory platform eases the development of translators thus allowing the migration of the software models.

City planning of technical information sys-tems:
The technical information system “carries” the virtual product throughout the life-cycle, crossing its different temporal states (before projects are completed up to ready for retirement products) and its different views for each discipline and this, in a distributed heterogeneous and multi partners environment. The objective here is to put into place solutions allowing 1) to assure the interoperability of the different technical information systems partners and 2) to allow the transparent evolution of the technological implementation tools (E.g. The replacement of one application with another.) This objective requires the capacity to formalize the management and evolution model of the technical information and to master the conversion procedures (link with heterogeneous engineering and platform heterogeneity.)

Platforms heterogeneity:
The objective is to integrate in the same engineering process design aspects and multi platforms distributed deployment aspects. The problem is to be able to state at a lesser cost the design of a real time system on different execution platforms not necessarily offering the same quality of service nor the same programming model. In order to design systems according to this approach, it is necessary to use different engineering disciplines stated in the MDE world such as execution platform modelling, real time aspects modelling, model transformation and completion by non functional elements, failure management; this in order to satisfy the targets performances constraints.

Multi formalisms semantic correspondence:
The design of components based system on type “models of numeric algorithms” is characterised by an heterogeneity of the formalisms used for describing their behaviour. In order to study and finally validate the global consistency of a GALS system (Generally Asynchronous Locally Synchronous) built with the help of these components, it is necessary to lean on a common point of description of their properties, in particular for their real-time properties. The models built on this meta model (or semantic pivot) must be consistent and meet by construction the desired properties for the considered system. Thus for the same functional architecture, different ordering for the model execution can be defined but only a few are consistent from the real time and behavioural point of view.
The objective will be to evaluate a prototype of semantic pivot meta model, consistent with the formalisms used for this class of systems and al-lowing to separate the functional design from the model of execution.