A PhD position on middleware and SOA (service oriented architectures) is open at Thales Communication France, Palaiseau (near Paris).
The position is co-advised by Thales Communication France (TCF) and the INRIA (French National Research Institute on Computer Science and Electrical Engineering) ADAM team. The thesis is funded by a CIFRE grant. The work will take place in the context of the ANR ITEmIS project in which both TCF and ADAM are participating.
Although some basic notions of French would be preferred, it is not mandatory to speak French for applying. A good level of written and spoken English is mandatory for applying.
Candidates may apply by sending a CV and a motivation letter to the following 3 persons:
- Hugues Vincent <Hugues.VINCENT@fr.thalesgroup.com>
- Lionel Seinturier <Lionel.Seinturier@univ-lille1.fr>
- Philippe Merle <Philippe.Merle@inria.fr>
Subject
Service-oriented architectures (SOA) are more and more the basis on which modern applications are built in many application domains (information systems, grid computing, embedded systems, etc.). Associated to SOA, the component-based paradigm seems to be the solution of choice for implementing these architectures. This PhD topic deals with component-based service-oriented software architectures.
In our view, two main requirements must be addressed by these architectures: heterogeneity and evolvability. The former point stems from the fact that even if a mainstream component model such as SCA seems to emerge as the leading framework for implementing SOA, it is certainly unlikely that it will be the only one on which SOA will be implemented. For example, OSGi appears as a challenger, and in the future, it is very likely that other models will be proposed. Hence, we believe that SOA will be based on several different models. This brings a strong need for supporting heterogeneity, not only at the middleware communication level (which is the solution currently considered by these frameworks) but also at the level of the architecture description level, which is a challenge we propose to tackle in this subject.
Evolvability of SOA is the second point that we propose to address in this subject. It is a matter of providing agile systems which can respond to changes in the requirements or in the execution environment. Many solutions, at different stages of the software lifecycle (from analysis to execution), have already been proposed for that. Yet, we believe that the fact that these solutions are only focused on one stage hinders their applicability. We believe that a holistic approach of adaptability must be considered, from modelling to runtime infrastructure. We then propose to design a model for adaptability, a runtime infrastructure and to provide some means by which these two levels can be causally connected and kept consistent.
As a concluding remark, one should note the two properties which are sought, heterogeneity and evolvability, cross fertilize themselves. Indeed, the fact that heterogeneous architectures are considered from the beginning will be a mean to avoid that the solutions designed for evolvability be too specific. Conversely, studying evolvability will be a strong incentive in finding the right artefacts which are needed to handle heterogeneity.
This Ph.D. subject will be conducted in the context of ANR ITEmIS project within an industrial/academic Ph.D. fellowship (CIFRE) at Thales Communications France.
Links
Thales Group: http://www.thalesgroup.com/?locale=EN-gb
INRIA ADAM: http://adam.lille.inria.fr
-- Lionel Seinturier
Univ. Lille (USTL) - Lab. LIFL - INRIA ADAM