Large software projects are characterized by distributed environments consisting of teams at different organizations and geographical locations. These teams typically build multiple overlapping models, representing different perspectives, different versions across time, different variants in a product family, different development concerns, etc. Keeping track of the relationships between these models, constructing a global view, and managing consistency are major challenges. Model Management is concerned with describing the relationships between distributed models, i.e., models built in a distributed development environment, and providing systematic operators to manipulate these models and their relationships. Such operators include, among others, Match, for finding relationships between disparate models, Merge, for combining models with respect to known or hypothesized relationships between them, Slice, for producing projections of models and relationships based on given criteria, and Check-Consistency, for verifying models and relationships against the consistency properties of interest. In this thesis, we provide automated solutions for two key model management operators, Merge and Check-Consistency. The most novel aspects of our work on model merging are (1) the ability to combine arbitrarily large collections of interrelated models and (2) support for toleration of incompleteness and inconsistency. Our consistency checking technique employs model merging to reduce the problem of checking inter-model consistency to checking intra-model consistency of a merged model. This enables a flexible way of verifying global consistency properties that is not possible with other existing approaches. We develop a prototype tool, TReMer+, implementing our merge and consistency checking approaches. We use TReMer+ to demonstrate that our contributions facilitate understanding and refinement of the relationships between distributed models.

Large software projects are characterized by distributed environments consisting of teams at different organizations and geographical locations. These teams typically build multiple overlapping models, representing different perspectives, different versions across time, different variants in a product family, different development concerns, etc. Keeping track of the relationships between these models, constructing a global view, and managing consistency are major challenges. Model Management is concerned with describing the relationships between distributed models, i.e., models built in a distributed development environment, and providing systematic operators to manipulate these models and their relationships. Such operators include, among others, Match, for finding relationships between disparate models, Merge, for combining models with respect to known or hypothesized relationships between them, Slice, for producing projections of models and relationships based on given criteria, and Check-Consistency, for verifying models and relationships against the consistency properties of interest. In this thesis, we provide automated solutions for two key model management operators, Merge and Check-Consistency. The most novel aspects of our work on model merging are (1) the ability to combine arbitrarily large collections of interrelated models and (2) support for toleration of incompleteness and inconsistency. Our consistency checking technique employs model merging to reduce the problem of checking inter-model consistency to checking intra-model consistency of a merged model. This enables a flexible way of verifying global consistency properties that is not possible with other existing approaches. We develop a prototype tool, TReMer+, implementing our merge and consistency checking approaches. We use TReMer+ to demonstrate that our contributions facilitate understanding and refinement of the relationships between distributed models.

A runtime file system checker protects file-system metadata integrity. It checks the consistency of file system update operations before they are committed to disk, thus preventing corrupted updates from reaching the disk. In this thesis, we describe our experiences with building Brunch, a runtime checker for an emerging Linux file system called Btrfs. Btrfs supports many modern file-system features that pose challenges in designing a robust checker. We find that the runtime consistency checks need to be expressed clearly so that they can be reasoned about and implemented reliably, and thus we propose writing the checks declaratively. This approach reduces the complexity of the checks, ensures their independence, and helps identify the correct abstractions in the checker. It also shows how the checker can be designed to handle arbitrary file system corruption. Our results show that runtime consistency checking is still viable for complex, modern file systems.

A runtime file system checker protects file-system metadata integrity. It checks the consistency of file system update operations before they are committed to disk, thus preventing corrupted updates from reaching the disk. In this thesis, we describe our experiences with building Brunch, a runtime checker for an emerging Linux file system called Btrfs. Btrfs supports many modern file-system features that pose challenges in designing a robust checker. We find that the runtime consistency checks need to be expressed clearly so that they can be reasoned about and implemented reliably, and thus we propose writing the checks declaratively. This approach reduces the complexity of the checks, ensures their independence, and helps identify the correct abstractions in the checker. It also shows how the checker can be designed to handle arbitrary file system corruption. Our results show that runtime consistency checking is still viable for complex, modern file systems.

Context. Software systems, becoming larger and more complex day-by-day, have resulted in software development processes to become more complex to understand and manage. Many companies have started to adapt distributed software engineering practices that would allow them to work in distributed teams at different organizations and/or geographical locations. For example, model-driven engineering methods are being used in such global software engineering projects. Among the activities in model-based software development, consistency checking is one of the widely known ones. Consistency checking is concerned with consistent models; in particular, having a consistent group of multiple models for a whole system, e.g., multiple models produced by distributed teams. Objectives. This thesis aims to find out how ‘Global Consistency Checking (GCC)’ can be utilized for exploring inconsistency problems between distributed models; particularly among UML class diagram relationships (in terms of consistency), as well as how GCC can be scaled with large number of models and relationships. Thereby, these inconsistencies are also aimed to incrementally resolve in our approach. Methods. We made a review in distributed software development domain and model management, in particular, methods of consistency checking between ‘Distributed Models (DM)’. Next, we conducted two case studies in two problem domains in order to apply our ‘consistency checking methodology’. We concurrently constructed and implemented new consistency rules, most of which are gathered from literatures and brainstorming with our coordinators. Generally, the method contains implementing different models of the case studies with a tool support and trying to figure out overlaps, merging models and checking the merged model against the consistency rules, and evaluating the results of GCC. We mainly addressed issues focused on consistency checking of individual models and the mapping between them e.g., pair-wise consistency checking (PCC), which are incapable of fully addressing problems against any consistency rules encountered in distributed environments. Results. We have identified seven types of inconsistency, which are divided in two groups named ‘Global inconsistency’ and ‘Pair-wise inconsistency’. In the first case study, we have 94 global inconsistencies and 73 pair-wise. In the second one, 14 global and 25 pair-wise inconsistencies are resulted. During ‘Resolution approach’, we followed six steps as a ‘systematic procedure’ for resolving these inconsistencies and constructed new merged model in each iteration. The initial merged model (inconsistent model) as an input for the first step has 1267 elements, and the consistent merged model (the output) from the sixth step has 686 elements. ‘time duration’ and ‘required effort’ for checking consistency against each ‘consistency rule’ were recorded, analyzed and illustrated in Sections 4.1.5 and 4.2.4. Conclusions. We concluded that GCC enables us to explore the inconsistencies, inclusive of resolving them and therefore, refining the relationships between different models, which are difficult to detect by e.g., a pair-wise method. The most important issues are: The number of model comparisons conducted by PCC, The inability of PCC for identifying some inconsistencies, Model relationships refinement and classification based on PCC approach will not lead to a final consistent DM, whereas, GCC guarantees it. Consistency rules application, inconsistency identification and resolving them could be generalized to any UML class diagram model representing a problem domain within the fields of consistency checking in software engineering. / 0046760850792, 0046737749752

Durant le développement de grandes lignes de produits logiciels, un ensemble de techniques d’implémentation traditionnelles, comme l’héritage ou les patrons de conception, est utilisé pour implémenter la variabilité. La notion de feature, en tant qu’unité réutilisable, n’a alors pas de représentation de première classe dans le code, et un choix inapproprié de techniques entraîne des incohérences entre variabilités du domaine et de l’implémentation. Dans cette thèse, nous étudions la diversité de la majorité des techniques d’implémentation de la variabilité, que nous organisons dans un catalogue étendu. Nous proposons un framework pour capturer et modéliser, de façon fragmentée, dans des modèles techniques de variabilité, la variabilité implémentée par plusieurs techniques combinées. Ces modèles utilisent les points de variation et les variantes, avec leur relation logique et leur moment de résolution, pour abstraire les techniques d’implémentation. Nous montrons comment étendre le framework pour obtenir la traçabilité de feature avec leurs implémentations respectives. De plus, nous fournissons une approche outillée pour vérifier la cohérence de la variabilité implémentée. Notre méthode utilise du slicing pour vérifier partiellement les formules de logique propositionnelles correspondantes aux deux niveaux dans le cas de correspondence 1–m entre ces niveaux. Ceci permet d’obtenir une détection automatique et anticipée des incohérences. Concernant la validation, le framework et la méthode de vérification ont été implémentés en Scala. Ces implémentations ont été appliquées à un vrai système hautement variable et à trois études de cas de lignes de produits. / When large software product lines are engineered, a combined set of traditional techniques, such as inheritance, or design patterns, is likely to be used for implementing variability. In these techniques, the concept of feature, as a reusable unit, does not have a first-class representation at the implementation level. Further, an inappropriate choice of techniques becomes the source of variability inconsistencies between the domain and the implemented variabilities. In this thesis, we study the diversity of the majority of variability implementation techniques and provide a catalog that covers an enriched set of them. Then, we propose a framework to explicitly capture and model, in a fragmented way, the variability implemented by several combined techniques into technical variability models. These models use variation points and variants, with their logical relation and binding time, to abstract the implementation techniques. We show how to extend the framework to trace features with their respective implementation. In addition, we use this framework and provide a tooled approach to check the consistency of the implemented variability. Our method uses slicing to partially check the corresponding propositional formulas at the domain and implementation levels in case of 1–to–m mapping. It offers an early and automatic detection of inconsistencies. As validation, we report on the implementation in Scala of the framework as an internal domain specific language, and of the consistency checking method. These implementations have been applied on a real feature-rich system and on three product line case studies, showing the feasibility of the proposed contributions.