|
RECENT EVENTS
THE PROGRAMME
THE CARROLL AGREEMENT
CURRENT PROJECTS
CONTACTS:
THALES,
CEA,
INRIA |
|
The architects and those in charge of the development of THALES
systems are faced with the fact that:
 |
Not only is proprietary middleware expensive to
maintain and keep up to date (mixture of business code and technical
code which
does not evolve the same pace, difficulty in finding
developers trained in ageing technologies and the increasing complexity
of the technological requirements which requires ever greater expertise)
|
 |
But also the rapid rate of development of commercial
middleware technology (massive investments by software publishers and
large communities of users and service providers) which is
generally supported by standards organisations.
|
It is important for THALES to be in harmony with these major trends
and it is thus that four middleware research areas have been identified:
2.1 Container engineering
Recent middleware enables component-based applications to be used. In middleware supporting
the deployment and execution of components, the container is the mediator between the functional components
("business centred") and the non-functional services (fault tolerance, safety,
persistence etc.). This guarantees proper separation of the business code
(which is where the competence of THALES lies) and the technical code.
2.2 From platform specific models to platforms
This area relates to transition from the domain of model-driven software engineering to that of
middleware and code. It constitutes the link between the areas of model-driven engineering and middleware
technology.
2.3 Framework for component-based real-time and embedded applications,
and real-time interaction models
This field covers study of the middleware architectures dedicated to real-time and
embedded
applications which use the best properties of the platforms as components (this corresponds
to the concept of the container and that of interaction between components). Here it is
sought to isolate the business code from concerns such as management of the
sequencing of tasks, communication protocol selection or implementation of trace
functions.
2.4 Network-Centric Middleware
The new application field of "systems of systems" is important for THALES, and will require
the investment of a substantial amount of work in coming years. This issue, essentially
arising from the C4ISR domain (Command, Control, Communications, Computers, Intelligence,
Surveillance, and Reconnaissance), is characterised by the building of applications resulting
from the reciprocal discovery of mobile systems. This leads to the concept of the system of systems,
for which special middleware is required. This is referred to as network-centric, and is required
to guarantee safety and quality of service, while offering a level of heterogeneous technology
integration and dynamics without equivalent in conventional middleware.
Last update:
07/01/2004
-
Legal provisions |


The software development cycle
seen by MIRROR and ALICE,
two pilot programmes conducted by THALES relating to model-driven software engineering and middleware.
The search carried out by Carroll corresponds to the requirements of these pilot programmes. |