N° d'ordre

:

N° d’ordre : 2546







THESE


Préparée au Laboratoire d’Analyse et d’Architecture des Systèmes du CNRS

pour obtenir

LE TITRE DE DOCTEUR DE L’UNIVERSITÉ DE TOULOUSE,
DÉLIVRÉ PAR L’INSTITUT NATIONAL POLYTECHNIQUE DE TOULOUSE



École doctorale : Systèmes
Spécialité : Systèmes Informatiques


Par Ana–Elena Rugina


Modélisation et évaluation de la sûreté de fonctionnement –
De AADL vers les réseaux de Petri stochastiques

Dependability modeling and evaluation –
From AADL to stochastic Petri nets




Soutenue le 19 novembre 2007 devant le jury composé de :


M. P. Feiler Président
Mme. K. Kanoun Directeur de thèse
Mme. F. di Giandomenico Rapporteur
M. A. van Moorsel
M. M. Kaâniche Examinateur C. Lemercier Examinateur

















To all those who supported me.

















Every day you may make progress. Every step may
be fruitful. Yet there will stretch out before you an ever-
lengthening, ever-ascending, ever-improving path. You
know you will never get to the end of the journey. But
this, so far from discouraging, only adds to the joy and
glory of the climb.

Sir Winston Churchill
Acknowledgements
This work has been carried out at the Laboratory of Analysis and Architecture of Systems
of the French National Research Center (LAAS-CNRS). I wish to express my gratitude to the
successive directors of the LAAS-CNRS, Mr. Malik Ghallab and Mr. Raja Chatila, for the
facilities provided in the laboratory.
I also thank Mr. Jean Arlat, CNRS Research Director and head of the Dependable
Computing and Fault Tolerance research group (TSF), for having allowed me to carry out this
work in this reach research environment.
This thesis has been financed by a scholarship from the European Social Fund and has been
partially conducted in the context of (1) the ASSERT European Project (Automated Proof-
Based System and Software Engineering for Real-Time Applications) (2) the ReSIST
Network of Excellence (Resilience for Survivability in IST).
I would like to thank all committee members, for having attentively read my dissertation:
- Peter Feiler, senior researcher at the Carnegie Mellon Software Engineering
Institute (Pittsburgh, USA), who honored me by chairing this committee.
- Felicita di Giandomenico, senior researcher at “Istituto di Elaborazione del
l’Informazione” – CNR (Pisa, Italy).
- Aad van Moorsel, reader at the University of Newcastle upon Tyne, (UK).
- Karama Kanoun, senior CNRS researcher.
- Mohamed Kaâniche, CNRS researcher.
- Christophe Lemercier, head of departement “Data Processing & On-board
Software & Dependability” of ASTRIUM Satellites, (Toulouse).
My special thanks go to Mrs. Felicita di Giandomenico and Mr. Aad van Moorsel, who
accepted the charge of being “rapporteurs”. I wish to thank them for their relevant comments.
I am most grateful to my supervisors, Mrs. Karama Kanoun and Mr. Mohamed Kaâniche,
for their passion, technical and human advice, for having devoted me an important amount of
their time and for the evenings spent at LAAS before my oral defense.
I warmly thank all members of the Dependable Computing and Fault Tolerance research
group: permanent researchers, PhD students and interns. I am very grateful to them for their
precious advice, support and friendship. I have much appreciated the family-like atmosphere
and the open discussions on different topics. My experience in this group has been extremely
enriching. I am very pleased to mention here my officemates (Eric, Géraldine, Ossama,
Carlos and Magnos) with whom I shared unforgettable moments. I also think of previous PhD
students (Ali, Taha, Cristina, Eric M., Guillaume, Nicolas), who offered me unconditional
support during my moments of doubt at the beginning of my experience in LAAS-CNRS. I
thank Marilena Bruffa, who contributed directly to the achievement of this work, and Eric
Marsden, whose attentive reading has contributed to the improvement of my dissertation. My
thanks also go to Gina, for her availability, kindness and precious help in organizing the
defense.
I wish to extend my thanks to all members of the service departments of LAAS-CNRS
(“Informatique et Instrumentation, Documentation, Magasin, Entretien, Direction–Gestion, Réception–Standard, Communication”), who always allowed me to work in the best
conditions.
I would also like to acknowledge here the Zonta International and the members of the
Zonta club of Muret (France), for having supported my application for the Amelia Earhart
fellowship. I am also very grateful to Peter Feiler (Carnegie Mellon Software Engineering
Institute), Eric Conquet (European Space Agency), Bruce Lewis (US Army) and Jean Arlat
(LAAS-CNRS), for their support in this context.
During my PhD, I have spent six weeks at the Carnegie Mellon Software Engineering
Institute in the “Performace Critical Systems” team. I thank Peter Feiler and all the team for
welcoming me. I learned many lessons, both technical and human during my stay in
Pittsburgh. I am grateful to Madelaine Dusseau for the support provided in the many
administrative tasks associated with this stay.
I am particularly thankful to Jean-Paul Blanquart and Dave Thomas (from ASTRIUM
Satellites) for the enriching exchanges that we had along my PhD. Their thoughts helped me
to mature my ideas.
In these important moments of my life, I think of my family and friends. I owe them my
sincerest gratitude for their support and confidence. Finally, my warmest thanks go to Florin,
who shared everything with me during the last years. I thank him for his understanding, his
patience and for having supported my moods, sometimes anxious lately. Remerciements
Les travaux présentés dans ce mémoire ont été effectués au Laboratoire d’Analyse et
d’Architecture des Systèmes du CNRS. Je remercie Messieurs Malik Ghallab et Raja Chatila,
qui ont assuré la direction du LAAS-CNRS depuis mon entrée, de m’avoir accueilli au sein de
ce laboratoire.
Je remercie également Monsieur Jean Arlat, Directeur de Recherche au CNRS, responsable
du groupe de recherche Tolérance aux fautes et Sûreté de Fonctionnement informatique
(TSF), de m’avoir permis de réaliser ces travaux dans ce groupe.
Les travaux développés dans cette thèse ont été financés par une bourse du Fond Social
Européen et effectués partiellement dans le cadre du projet européen ASSERT (Automated
Proof-Based System and Software Engineering for Real-Time Applications) et du réseau
d’excellence ReSIST (Resilience for Survivability in IST).
Je tiens à remercier tous les membres du jury, pour leur lecture attentive du mémoire :
- Peter Feiler, chercheur au « Carnegie Mellon Software Engineering Institute »
(Pittsburgh, Etats-Unis), qui m’a fait l’honneur de présider ce jury.
- Felicita di Giandomenico, chercheur à « Istituto di Elaborazione del
l’Informazione » – CNR (Pise, Italie).
- Aad van Moorsel, professeur à « University of Newcastle upon Tyne »
(Royaume-Uni).
- Karama Kanoun, Directrice de Recherche au CNRS.
- Mohamed Kaâniche, Chargé de Recherche au CNRS.
- Christophe Lemercier, directeur du département « Data Processing & On-
board Software & Dependability » de ASTRIUM Satellites, (Toulouse).
Je remercie tout particulièrement Madame Felicita di Giandomenico et Monsieur Aad van
Moorsel, qui ont accepté la lourde tâche de rapporteur. Je les remercie vivement pour leurs
commentaires avisés.
J’exprime ma profonde reconnaissance à Madame Karama Kanoun (Directrice de
Recherche au CNRS) et Monsieur Mohamed Kaâniche (Chargé de Recherche au CNRS),
pour avoir dirigé mes travaux de thèse, pour leur passion, leurs conseils tant sur le plan
technique que humain, pour m’avoir consacré une partie importante de leur temps et
notamment pour les soirées passées au LAAS avant ma soutenance.
Je remercie sincèrement tous les membres du groupe TSF, permanents, doctorants et
stagiaires. Je leur suis très reconnaissante pour leurs conseils, support et amitié. J’ai beaucoup
apprécié l’ambiance « familiale » et les discussions très ouvertes sur divers thèmes. Mon
expérience au sein du groupe a été extrêmement enrichissante. Il m’est particulièrement
agréable de remercier mes collègues de bureau (Eric, Géraldine, Ossama, Carlos et Magnos)
avec qui