Computer and Network Sciences Department
Complementing the work on task scheduling, I am additionally exploring the use of deterministic and time-predictable computing platforms. The goal of this work is to improve the analysability of critical software in terms of its worst-case behaviour, while ensuring competitive performance in the average case. Predictability can be ensured in software using hypervisors on off-the-shelf hardware, as demonstrated in a collaboration with Thales Avionics (thesis Jean). The hypervisor intercepts accesses to shared resources (e.g., memory) and thus can manage these accesses to avoid contention and improve isolation. The quality of this work has been acknowledged through a best Paper Award and resulted in a joint patent with Thales.
Another means to improve the predictability of a platform is through specialised hardware. Current work (thesis Hebbache) aims at improving the average-case performance of the memory hierarchy of the Patmos multi-core platform by using dynamic arbitration schemes that converge to predictable Time-Division Multiplexing (TDM) in the worst-case.
|Critical Systems Design Process|
The research cited above aims at providing a suitable computing platform (OS, runtime, hardware) for critical real-time systems. A third pillar of my research activities are techniques and tools that help to design such systems via model-based design — exploiting, among others, the aforementioned computing platforms. Etienne Borde and myself strongly contribute to the state-of-the-art in model-based design at various steps of the design process (model refinement, test, code generation, ...) and at various technical and formal levels (thesis Richa), but also the definition of modelling languages. These contributions on model-driven engineering have been acknowledged by the scientific community, e.g., through a Best Paper Award at ICMT 2015
RAMSES allows to automatically derive implementation models from high-level design models (thesis Cadoret) in order to validate low-level software implementations, to analyse the availability as well as the schedulability of mixed-criticality applications (thesis Medina), or, even, to explore the design space of implementations through model transformation (thesis Rahmoun). Parts of this work have been carried out with industrial partners such as Alstom in the context of the IRT SystemX as well as the chair « Ingénierie de Systèmes Complexes » with academic partners from École Polytechnique and ENSTA ParisTech. A follow-up project ISC recently started (thesis Hassine) aiming at managing uncertainty in model-based design exploration. The RAMSES tool has also been selected as a showcase project by Institute.
|Safety and Security Systems|
Following up the project ELA, the project CTI (thesis Oudot) is currently starting, again in the context of the IRT SystemX, and allowed to establish a new collaboration with A. Easwaran (NTU, Singapore). This work aims at addressing security issues that may impact safety and realiability issues by improving the critical system design process.
|Projects and PhD thesis|
|Real-time Embedded Systems|
|Java Programming (first year)|
System Programming (first year)
|Paris Saclay University|