This dissertation proposes the Weaves runtime framework for the execution of large scale parallel programs over lightweight intra-process threads. The goal of the Weaves framework is to help process-based legacy parallel programs exploit the scalability of threads without any modifications. The framework separates global variables used by identical, but independent, threads of legacy parallel programs without resorting to thread-based re-programming. At the same time, it also facilitates low-overhead collaboration among threads of a legacy parallel program through multi-granular selective sharing of global variables. Applications that follow the tenets of the Weaves framework can load multiple identical, but independent, copies of arbitrary object files within a single process. They can compose the runtime images of these object files in graph-like ways and run intra-process threads through them to realize various degrees of multi-granular selective sharing or separation of global variables among the threads. Using direct runtime control over the resolution of individual references to functions and variables, they can also manipulate program composition at fine granularities. Most importantly, the Weaves framework does not entail any modifications to either the source codes or the native codes of the object files. The framework is completely transparent. Results from experiments with a real-world process-based parallel application show that the framework can correctly execute a thousand parallel threads containing non-threadsafe global variables on a single machine - nearly twice as many as the traditional process-based approach can - without any code modifications. On increasing the number of machines, the application experiences super-linear speedup, which illustrates scalability. Results from another similar application, chosen from a different software area to emphasize the breadth of this research, show that the framework's facilities for low-overhead collaboration among parallel threads allows for significantly greater scales of achievable parallelism than technologies for inter-process collaboration allow. Ultimately, larger scales of parallelism enable more accurate software modeling of real-world parallel systems, such as computer networks and multi-physics natural phenomena. / Ph. D.

Reuse is not a new concept in software engineering. Ideas, abstractions, and processes have been reused by programmers since the very early days of software development. In the beginning, since storage media was very expensive, software reuse was basically to serve computers and their mechanical resources, as it substantially conserved memory. When the limitations on physical resources started to diminish, software engineers began to invent reuse approaches to save human resources as well. In addition, as the size and complexity of software systems constantly grow, organized and systematic reuse becomes essential in order to develop those systems in timely and cost-effective fashion. That is one main reason why new technologies and approaches for building software systems, such as object-oriented and component-based development, emerged in the last two or three decades. The focus of this thesis is on software components as building blocks of today's software systems. We consider components as software black boxes whose specification and external behavior are known. We assume that this information can somehow be extracted for each deployed software component. The first and basic assumption then would be the availability of a searchable repository of software components and their external behavioral specifications. Web services are a good example of such components. The most important advantage of software components is that they can be reused repeatedly in building different software systems. Reuse presents challenging problems, one of which is studied in this thesis. This problem, the composition problem, simply is creating a composite component from a collection of available components that, by interacting with each other, provide a requested functionality. When there are a large number of components available to be reused, finding a solution to the composition problem manually would require a considerable time and human effort. This could make the search practically impossible or unwieldy. However, performing the search automatically would save a significant amount of development time, cost and human effort. Solving this problem would be a huge step forward in the component-based software development. In this thesis, we concentrate on a subproblem of the composition problem, composition planning or synthesis, which is defined as finding a collection of useful components from the repository and the necessary communications among them to satisfy a requested functionality. For scalability purposes, we study automatic solutions to composition planning and propose two approaches in this regard. In one, we take advantage of graphs to model the repository, which is the collection of available components along with their behavioral specification. Graph search algorithms and a few composition-specific algorithms are used to find solutions for given component requests. In the other approach, we extend a logical reasoning algorithm and come up with algorithms for solving the composition planning problem. In both approaches we provide algorithms for finding the possibility of a composition, as well as finding the composition itself. We propose different types of composition and show how applying each would impact the behavior of a composite component. We provide the necessary formalism for capturing these types of composition through two different models: interface automata and composition algebra. Interface automata is an automaton-based model for representing the behavior of software components. The other model in this regard is composition algebra, which is an algebraic model based on CSP (Communicating Sequential Processes), CCS (Calculus of Communicating Systems), and interface automata. These formal models are used to validate the results returned by the composition approaches. We also compare the two composition approaches and show why each of them is suitable for specific types of the problem according to the repository attributes. We then evaluate the performance of the reasoning-based approach and provide some experimental results. In these experiments, we study how different attributes of the repository components could impact the performance of the reasoning-based approach in solving the composition planning problem.

Reuse is not a new concept in software engineering. Ideas, abstractions, and processes have been reused by programmers since the very early days of software development. In the beginning, since storage media was very expensive, software reuse was basically to serve computers and their mechanical resources, as it substantially conserved memory. When the limitations on physical resources started to diminish, software engineers began to invent reuse approaches to save human resources as well. In addition, as the size and complexity of software systems constantly grow, organized and systematic reuse becomes essential in order to develop those systems in timely and cost-effective fashion. That is one main reason why new technologies and approaches for building software systems, such as object-oriented and component-based development, emerged in the last two or three decades. The focus of this thesis is on software components as building blocks of today's software systems. We consider components as software black boxes whose specification and external behavior are known. We assume that this information can somehow be extracted for each deployed software component. The first and basic assumption then would be the availability of a searchable repository of software components and their external behavioral specifications. Web services are a good example of such components. The most important advantage of software components is that they can be reused repeatedly in building different software systems. Reuse presents challenging problems, one of which is studied in this thesis. This problem, the composition problem, simply is creating a composite component from a collection of available components that, by interacting with each other, provide a requested functionality. When there are a large number of components available to be reused, finding a solution to the composition problem manually would require a considerable time and human effort. This could make the search practically impossible or unwieldy. However, performing the search automatically would save a significant amount of development time, cost and human effort. Solving this problem would be a huge step forward in the component-based software development. In this thesis, we concentrate on a subproblem of the composition problem, composition planning or synthesis, which is defined as finding a collection of useful components from the repository and the necessary communications among them to satisfy a requested functionality. For scalability purposes, we study automatic solutions to composition planning and propose two approaches in this regard. In one, we take advantage of graphs to model the repository, which is the collection of available components along with their behavioral specification. Graph search algorithms and a few composition-specific algorithms are used to find solutions for given component requests. In the other approach, we extend a logical reasoning algorithm and come up with algorithms for solving the composition planning problem. In both approaches we provide algorithms for finding the possibility of a composition, as well as finding the composition itself. We propose different types of composition and show how applying each would impact the behavior of a composite component. We provide the necessary formalism for capturing these types of composition through two different models: interface automata and composition algebra. Interface automata is an automaton-based model for representing the behavior of software components. The other model in this regard is composition algebra, which is an algebraic model based on CSP (Communicating Sequential Processes), CCS (Calculus of Communicating Systems), and interface automata. These formal models are used to validate the results returned by the composition approaches. We also compare the two composition approaches and show why each of them is suitable for specific types of the problem according to the repository attributes. We then evaluate the performance of the reasoning-based approach and provide some experimental results. In these experiments, we study how different attributes of the repository components could impact the performance of the reasoning-based approach in solving the composition planning problem.

[pt] A engenharia de software baseada em componentes é uma abordagem que prevê a reutilização de artefatos de software na geração de um conjunto de aplicações. Ao desenvolver aplicações com esta abordagem, é preciso reunir ou compor componentes de software já existentes. Após compor estas unidades, é necessário coordenar as interdependências estabelecidas entre elas para adequar a aplicação em desenvolvimento à resolução do problema. Esta dissertação propõe uma arquitetura de software que separa e estrutura os conceitos de coordenação, composição e componentes de software em camadas arquiteturais. A partir desta estrutura, espera-se que modificações específicas em construções de uma camada tenham o mínimo de influência sobre as demais. ACCA (Arquitetura para a Coordenação e a Composição de Artefatos de Software) deve ser entendida como uma estrutura conceitual utilizada para organizar o desenvolvimento de software baseado em componentes. Além disto, são apresentados um framework para ilustrar a realização da camada de composição de ACCA, o processo de reificação de ACCA e um processo de desenvolvimento de software utilizando a abordagem proposta. / [en] Component Based Software Engineering is an approach for reusing software artifacts when developing applications. In order to develop solutions using this approach, it is necessary to compose software using components that have already been developed. After putting those pieces together, we need to coordinate the interdependencies established among those compositions to fulfill the requirements, needed to resolve a problem. This dissertation proposes a software architecture that separates and structures the concepts of coordination, composition and software components in different architectural layers. Using this approach, we expect that specific modifications in layer constructions have the minimum impact on the others layers. ACCA (Architecture for Coordination and Composition of software Artifacts) must be understood as a conceptual structure that is used to organize component based software development. It also presents a composition framework, the reification process for ACCA and a software development process organized using this approach.