Schedulable Mixed-critical Multi-core Systems Design

par Rany Kahil

Projet de thèse en Informatique

Sous la direction de Saddek Bensalem.

Thèses en préparation à Grenoble Alpes , dans le cadre de École doctorale mathématiques, sciences et technologies de l'information, informatique (Grenoble) , en partenariat avec VERIMAG (laboratoire) et de Rigorous System Design (equipe de recherche) depuis le 31-10-2015 .


  • Résumé

    This is a PhD project Rigorous System Design team at Verimag. The sutdent will be involved in solving practical research problems of programming many-core systems while satisfying real-time constraints. We consider industrial case studies from avionics and spacecraft domains and target state-of-the-art multi-core platforms. Though the real-time systems research offers efficient schedulability algorithms, their practical use encounters serious challenges: 1) Lack of consolidation in programming 2) Mixed-criticality 3) Complexity of timing analysis The first challenge concerns with the lack of formal development techniques for software-intensive systems with real-time constraints. Our work is based on an expressive formal programming and modeling framework BIP, providing refinement, code generation and verification of functional and timing properties for embedded systems. The PhD candidate is expected to develop BIP runtimes and code generation tools for multi-cores, extending our previous work in [1]. The next primary challenge is mixed criticality. The traditional static safety-critical scheduling has to be extended by mixed-critical scheduling, which re-shapes the schedulability analyses in a non-trivial way. The candidate is expected to extend our scheduling algorithms and tools [2] to take into account a set of multi-core platform constraints and evaluate them in different application case studies. The last but not the least, the conservative timing analysis is a challenging problem, involving two major aspects: the execution time uncertainty in multi-path programs and on cache/bus/network interference between different tasks. The candidate is expected to adapt our statistical model-checking techniques [3] to the analysis of execution times and interferences in real-time multi-core systems.

  • Titre traduit

    Schedulable Mixed-critical Multi-core Systems Design


  • Résumé

    This is a PhD project Rigorous System Design team at Verimag. The sutdent will be involved in solving practical research problems of programming many-core systems while satisfying real-time constraints. We consider industrial case studies from avionics and spacecraft domains and target state-of-the-art multi-core platforms. Though the real-time systems research offers efficient schedulability algorithms, their practical use encounters serious challenges: 1) Lack of consolidation in programming 2) Mixed-criticality 3) Complexity of timing analysis The first challenge concerns with the lack of formal development techniques for software-intensive systems with real-time constraints. Our work is based on an expressive formal programming and modeling framework BIP, providing refinement, code generation and verification of functional and timing properties for embedded systems. The PhD candidate is expected to develop BIP runtimes and code generation tools for multi-cores, extending our previous work in [1]. The next primary challenge is mixed criticality. The traditional static safety-critical scheduling has to be extended by mixed-critical scheduling, which re-shapes the schedulability analyses in a non-trivial way. The candidate is expected to extend our scheduling algorithms and tools [2] to take into account a set of multi-core platform constraints and evaluate them in different application case studies. The last but not the least, the conservative timing analysis is a challenging problem, involving two major aspects: the execution time uncertainty in multi-path programs and on cache/bus/network interference between different tasks. The candidate is expected to adapt our statistical model-checking techniques [3] to the analysis of execution times and interferences in real-time multi-core systems.