Un objectif principal du génie logiciel est de pouvoir produire des logiciels complexes, de grande taille et fiables en un temps raisonnable. La technologie orientée objet (OO) a fourni de bons concepts et des techniques de modélisation et de programmation qui ont permis de développer des applications complexes tant dans le monde académique que dans le monde industriel. Cette expérience a cependant permis de découvrir les faiblesses du paradigme objet (par exemples, la dispersion de code et le problème de traçabilité). La programmation orientée aspect (OA) apporte une solution simple aux limitations de la programmation OO, telle que le problème des préoccupations transversales. Ces préoccupations transversales se traduisent par la dispersion du même code dans plusieurs modules du système ou l’emmêlement de plusieurs morceaux de code dans un même module. Cette nouvelle méthode de programmer permet d’implémenter chaque problématique indépendamment des autres, puis de les assembler selon des règles bien définies. La programmation OA promet donc une meilleure productivité, une meilleure réutilisation du code et une meilleure adaptation du code aux changements. Très vite, cette nouvelle façon de faire s’est vue s’étendre sur tout le processus de développement de logiciel en ayant pour but de préserver la modularité et la traçabilité, qui sont deux propriétés importantes des logiciels de bonne qualité. Cependant, la technologie OA présente de nombreux défis. Le raisonnement, la spécification, et la vérification des programmes OA présentent des difficultés d’autant plus que ces programmes évoluent dans le temps. Par conséquent, le raisonnement modulaire de ces programmes est requis sinon ils nécessiteraient d’être réexaminés au complet chaque fois qu’un composant est changé ou ajouté. Il est cependant bien connu dans la littérature que le raisonnement modulaire sur les programmes OA est difficile vu que les aspects appliqués changent souvent le comportement de leurs composantes de base [47]. Ces mêmes difficultés sont présentes au niveau des phases de spécification et de vérification du processus de développement des logiciels. Au meilleur de nos connaissances, la spécification modulaire et la vérification modulaire sont faiblement couvertes et constituent un champ de recherche très intéressant. De même, les interactions entre aspects est un sérieux problème dans la communauté des aspects. Pour faire face à ces problèmes, nous avons choisi d’utiliser la théorie des catégories et les techniques des spécifications algébriques. Pour apporter une solution aux problèmes ci-dessus cités, nous avons utilisé les travaux de Wiels [110] et d’autres contributions telles que celles décrites dans le livre [25]. Nous supposons que le système en développement est déjà décomposé en aspects et classes. La première contribution de notre thèse est l’extension des techniques des spécifications algébriques à la notion d’aspect. Deuxièmement, nous avons défini une logique, LA , qui est utilisée dans le corps des spécifications pour décrire le comportement de ces composantes. La troisième contribution consiste en la définition de l’opérateur de tissage qui correspond à la relation d’interconnexion entre les modules d’aspect et les modules de classe. La quatrième contribution concerne le développement d’un mécanisme de prévention qui permet de prévenir les interactions indésirables dans les systèmes orientés aspect. / One of the main goals of software engineering is to enable the construction of large, complex and reliable software in timely fashion. Object-oriented (OO) technology has provided modeling and programming principles and techniques that allow developing complex software systems both in academic and industrial areas. In return, experience gained in OO system development has allowed discovering some limitations of object technology (e.g., code scattering and poor traceability problems). Aspect Oriented (AO) Technology is a post-object-oriented technology emerged to overcome limitations of Object Oriented (OO) Technology, such as the crosscutting concern problem. Crosscutting concerns are scattered and tangled concerns. Major goals of Aspect Oriented Programming (AOP) include improving modularity, cohesion, and overall software quality. Aspect Oriented Programming results in the evolution of programming activities to fullblown software engineering processes, to preserve modularity and traceability, which are two important properties of high-quality software. Yet, there are also many challenges in AO Technology. Reasoning, specification, and verification of AO programs present unique challenges especially as such programs evolve over time. Consequently, modular reasoning of such programs is highly attractive as it enables tractable evolution, otherwise necessitating that the entire program be reexamined each time a component is changed or is added. It is well known in the literature, however, that modular reasoning about AO programs is difficult due to the fact that the aspects applied often alter the behavior of the base components [47]. The same modular reasoning difficulties are also present in the specification and verification phases of software development process. To the best of our knowledge, AO modular specification and verification is a weakly covered subject and constitutes an interesting open research field. Also, aspect interaction is a major concern in the aspect-oriented community. To deal with these problems, we choose to use category theory and algebraic specification techniques. To achieve the above thesis goals, we use the work of Wiels [110] and other contributions such as the one described in [25]. We assume at the beginning that the system under development is already decomposed into aspect and class components. The first contribution of our thesis is the extension of the algebraic specification technique to the notion of aspect. Secondly, we define a logic, LA that is used in specification bodies to describe the behavior of these components. The third contribution concerns the defini tion of the weaving operator corresponding to the weaving interconnection relationship between aspect modules and class modules. The fourth contribution consists of the design of a prevention policy that is used to prevent or avoid undesirable aspect interactions in aspect-oriented systems.

We study the problem of secure key establishment. We critically examine the security models of Bellare and Rogaway (1993) and Canetti and Krawczyk (2001) in the computational complexity approach, as these models are central in the understanding of the provable security paradigm. We show that the partnership definition used in the three-party key distribution (3PKD) protocol of Bellare and Rogaway (1995) is flawed, which invalidates the proof for the 3PKD protocol. We present an improved protocol with a new proof of security. We identify several variants of the key sharing requirement (i.e., two entities who have completed matching sessions, partners, are required to accept the same session key). We then present a brief discussion about the key sharing requirement. We identify several variants of the Bellare and Rogaway (1993) model. We present a comparative study of the relative strengths of security notions between the several variants of the Bellare-Rogaway model and the Canetti-Krawczyk model. In our comparative study, we reveal a drawback in the Bellare, Pointcheval, and Rogaway (2000) model with the protocol of Abdalla and Pointcheval (2005) as a case study. We prove a revised protocol of Boyd (1996) secure in the Bellare-Rogaway model. We then extend the model in order to allow more realistic adversary capabilities by incorporating the notion of resetting the long-term compromised key of some entity. This allows us to detect a known weakness of the protocol that cannot be captured in the original model. We also present an alternative protocol that is efficient in both messages and rounds. We prove the protocol secure in the extended model. We point out previously unknown flaws in several published protocols and a message authenticator of Bellare, Canetti, and Krawczyk (1998) by refuting claimed proofs of security. We also point out corresponding flaws in their existing proofs. We propose fixes to these protocols and their proofs. In some cases, we present new protocols with full proofs of security. We examine the role of session key construction in key establishment protocols, and demonstrate that a small change to the way that session keys are constructed can have significant benefits. Protocols that were proven secure in a restricted Bellare-Rogaway model can then be proven secure in the full model. We present a brief discussion on ways to construct session keys in key establishment protocols and also prove the protocol of Chen and Kudla (2003) secure in a less restrictive Bellare-Rogaway model. To complement the computational complexity approach, we provide a formal specification and machine analysis of the Bellare-Pointcheval-Rogaway model using an automated model checker, Simple Homomorphism Verification Tool (SHVT). We demonstrate that structural flaws in protocols can be revealed using our framework. We reveal previously unknown flaws in the unpublished preproceedings version of the protocol due to Jakobsson and Pointcheval (2001) and several published protocols with only heuristic security arguments. We conclude this thesis with a listing of some open problems that were encountered in the study.

The growing need for computer software in different sectors of activity, (health, agriculture, industries, education, aeronautic, science and telecommunication) together with the increasing reliance of the society as a whole on information technology, is placing a heavy and fast growing demand on complex and high quality software systems. In this regard, the anticipation has been on non-functional requirements (NFRs) engineering and formal methods. Despite their common objective, these techniques have in most cases evolved separately. NFRs engineering proceeds firstly, by deriving measures to evaluate the quality of the constructed software (product-oriented approach), and secondarily by improving the engineering process (process-oriented approach). With the ability to combine the analysis of both functional and non-functional requirements, Goal-Oriented Requirements Engineering (GORE) approaches have become de facto leading requirements engineering methods. They propose through refinement/operationalisation, means to satisfy NFRs encoded in softgoals at an early phase of software development. On the other side, formal methods have kept, so far, their promise to eliminate errors in software artefacts to produce high quality software products and are therefore particularly solicited for safety and mission critical systems for which a single error may cause great loss including human life. This thesis introduces the concept of Complementary Non-functional action (CNF-action) to extend the analysis and development of NFRs beyond the traditional goals/softgoals analysis, based on refinement/operationalisation, and to propagate the influence of NFRs to other software construction phases. Mechanisms are also developed to integrate the formal technique Z/Object-Z into the standardised User Requirements Notation (URN) to formalise GRL models describing functional and non-functional requirements, to propagate CNF-actions of the formalised NFRs to UCMs maps, to facilitate URN construction process and the quality of URN models. / School of Computing / D. Phil (Computer Science)