It is only a matter of time before the Swedish Companies Registration Office makes digital registration of annual reports/auditors reports mandatory. At the time of writing, there are only ten public integrators currently able to handle this requirement. In collaboration with one of the big four accounting firms, this project aims to evaluate performance of the response time while using GraphQL versus REST in Web APIs. The application under test is a .NET application integrating to the Swedish Companies Registration Office API. Through a controlled experiment using two different GraphQL frameworks, HotChocolate and GraphQL for .NET, this thesis provides a knowledge base for the partnered accounting firm, developers and other stakeholders that are evaluating the use of GraphQL in their future applications. Results  from the experiment indicate that HotChocolate in its current version is not only faster than its competitor GraphQL for .NET, but also faster than REST. This is surprising, given that other related work seems to suggest that this is not always the case. Testing of GraphQL for .NET gives a more traditional result when compared to other related work. Given the results, a senior developer of HotChocolate was contacted to gain insight to why the framework outperforms not only GraphQL for .NET but also REST. The senior developer states that a large amount of effort has been put in to make the GraphQL execution engine more optimized, something that is corroborated by this thesis experiment. HotChocolate is also periodically measuring and comparing performance benchmarks to other libraries to conclude on its performance in different scenarios. The analysis of the experiment concluded that there exists another important variable previously not identified in other research, more precisely the chosen framework, that has a large impact on performance and can impact both memory allocation and response time.

The current push from the industry for Head-Mounted Display based wearable computers to the masses for everyday use suggests that academia's goals of a tight human and computer integration is achievable. Although promising, many interaction challenges remain to achieve a deep integration. Primarily that of active cognitive support in the form of focusing user attention. This thesis investigates to what degree a system can unobtrusively, perhaps even subliminally, guide attention of a user by directing their gaze from a current point to a selected point in a virtual environment. A subjective evaluation is made regarding the characteristics of visual sensory stimuli in order to possibly achieve an unobtrusive (or even subliminal) gaze guidance. The attention guidance system developed is deployed in a controlled experiment with 30 participants. Participants are asked to fire a longbow at targets in a virtual reality environment while their point of gaze is monitored with an eye-tracker. The experiment produced interesting results, but no conclusive evidence that the system was able actively guide a user's attention nor influence their choice could be observed. The conclusion of this thesis is that further development is needed towards a system that is capable of objectively calibrating the stimuli characteristics for each individual participant.

Web application frameworks have become a common part of software development, with each programming language having its own specific frameworks. However, choosing the right framework can be a daunting task when there are so many frameworks on the market and factors to consider. Previous research has attempted to solve this problem by evaluating frameworks based on several different metrics. While the results offer a way to rank the frameworks, they do not tell us much about the actual process of evaluating a framework, or why certain metrics matter and in what situation. This project aimed to answer these questions by conducting a literature review to gather information about the existing evaluation methods and metrics. Based on these results, an evaluation criteria was developed to provide an overview of when to evaluate what. Furthermore, a user survey was done to measure the usability of three popular JavaScript frameworks. Lastly, a controlled experiment was conducted to test the performance of the three frameworks. The results show that different frameworks have different strengths and weaknesses as they received different scores from the methods. Moreover, the results show that when choosing a framework, one has to consider which factors matter most as it may not always be possible to find a framework that has it all.

O processo de automatização de teste de software possui alto custo envolvido em sistemas de larga escala, pois exigem cenários de teste complexos e tempos de execução extremamente longos. Além disso, cada etapa do processo de teste requer recursos computacionais e um tempo considerável para a execução de muitos casos de teste, tornando-se um gargalo para as empresas de Tecnologia da Informação (TI). Neste contexto, os benefícios e oportunidades oferecidos pela combinação da computação em nuvem com o Teste como Serviço (Testing as a Service, TaaS), que é considerado um novo modelo de negócio e de serviço atraente e promissor, podem proporcionar um impacto positivo na redução do tempo de execução dos testes de maneira custo-efetiva e aumentar o retorno sobre o investimento ou Return on investment (ROI). Todavia, existe o problema de vendor lock-in, que é o aprisionamento do usuário à plataforma de um fornecedor específico ou serviço de teste, ocasionado pela dificuldade de migrar de um fornecedor TaaS para outro, limitando a utilização dessas novas tecnologias de maneira efetiva e eficiente, impedindo assim, a ampla adoção do TaaS. Como os estudos existentes não são rigorosos ou conclusivos e, principalmente, devido à falta de evidência empírica na área de serviço de teste, muitas questões devem ser investigadas na perspectiva da migração entre os provedores de TaaS. O objetivo deste trabalho é reduzir o impacto ocasionado pelo problema de vendor lock-in no processo de automatização de testes na computação em nuvem, na escrita, configuração, execução e gerenciamento dos resultados de testes automatizados. Neste contexto, foi desenvolvido o protótipo da abordagem intitulada Multi-TaaS por meio de uma biblioteca Java como prova de conceito. A abordagem Multi-TaaS é uma camada de abstração e a sua arquitetura permite abstrair e flexibilizar a troca de fornecedores de TaaS de forma portável, pois permite encapsular toda a complexidade da implementação do engenheiro de software ao desacoplar o teste automatizado de qual plataforma TaaS ele será executado, bem como abstrair os aspectos da comunicação e integração entre as APIs REST proprietárias dos diferentes fornecedores de TaaS. Além disso, a abordagem Multi-TaaS possibilita também sumarizar os resultados dos testes automatizados de forma independente das tecnologias da plataforma TaaS subjacente. Foram realizadas avaliações comparativas da eficiência, efetividade, dificuldade e do esforço de migração entre as abordagens Multi-TaaS e abordagem convencional, por meio de experimentos controlados. Os resultados deste trabalho indicam que a nova abordagem permite facilitar a troca do serviço de teste, melhorar a eficiência e, principalmente, reduzir o esforço e os custos de manutenção na migração entre fornecedores de TaaS. Os estudos realizados no experimento controlado são promissores e podem auxiliar os engenheiros de software na tomada de decisão quanto aos riscos associados ao vendor lock-in no TaaS. Por fim, a abordagem Multi-TaaS contribui, principalmente, para a portabilidade dos testes automatizados na nuvem e da sumarização dos resultados dos testes e, consequentemente, possibilita que o modelo de serviço TaaS na computação em nuvem seja amplamente adotado, de forma consciente, no futuro. / The automation of software testing involves high costs in large-scale systems, since it requires complex test scenarios and extremely long execution times. Moreover, each of its steps demands computational resources and considerable time for running many test cases, which makes it a bottleneck for Information Technology (IT) companies. The benefits and opportunities offered by the combination of cloud computing and Testing as a Service (TaaS), considered a new business and service model, can reduce the execution time of tests in a cost-effective way and improve Return on Investment (ROI). However, the lock-in problem, i.e., the imprisonment of the user in the platform of a specific vendor or test service caused by the difficult migration from one TaaS provider to another limits the effective use of such new technologies and prevents the widespread adoption of TaaS. As studies conducted are neither rigorous, nor conclusive, and mainly due to the lack of empirical evidence, many issues must be investigated from the perspective of migration among TaaS providers. This research aims at reductions in the impact of the vendor lock-in problem on the automation process of testing in cloud computing, writing, configuration, execution and management of automated test results. The prototype of the Multi- TaaS approach was developed through a Java library as a proof of concept. The Multi-TaaS approach is an abstraction layer and its architecture enables the abstraction and flexibilization of the exchange of TaaS providers in a portable way, once the complexity of the software engineers implementation can be encapsulated. The two main advantages of Multi-TaaS are the decoupling of the automated test from the TaaS platform on which it will be executed and the abstraction of the communication and integration aspects among the proprietary REST APIs of the different TaaS providers. The approach also enables the summarization of automated test results independently of the underlying TaaS platform technologies. A comparative evaluation between Multi-TaaS and conventional migration approaches regarding the difficulty, efficiency, effectiveness and effort of migration among TaaS providers was conducted through controlled experiments.The results show the approach facilitates the exchange of test service, improves efficiency and reduces the effort and maintenance costs of migration among TaaS providers. The studies conducted in the controlled experiment are promising and can assist software engineers in decision-making regarding the risks associated with vendor lock-in in TaaS. The Multi-TaaS approach contributes mainly to the portability of automated tests in the cloud and summarization of their results. Finally, this research enables also the widespread adoption of the TaaS service model in cloud computing, consciously, in the future.

Neste trabalho foi realizado um levantamento GPR 4D em laboratório no qual foi simulado o vazamento de óleo de uma ETD - Estação de Transmissão e Distribuição de Energia Elétrica. Os estudos consistiram em derramar 15 litros de óleo em três experimentos: tanque contendo areia seca, tanque contendo areia úmida mais gradiente hidráulico e tanque contendo areia úmida. Em todos os experimentos o objetivo foi detectar a pluma de contaminação e avaliar a migração desse óleo com o tempo. Os dados GPR 4D foram adquiridos utilizando-se uma antena blindada de 400 MHz modelo SIR-3000 equipamento da GSSI. O monitoramento no tempo de aquisição dos dados variou de 2 minutos até 12 dias. No experimento com o tanque preenchido com areia seca não foi possível determinar a pluma de contaminação devido ao baixo contraste entre as propriedades físicas do meio e a pluma contaminante. No experimento em que o tanque foi preenchido com areia úmida e havia a presença de um gradiente hidráulico, foi possível determinar a migração da pluma de contaminação a partir do 5º dia. Para finalizar, o 3º experimento onde o tanque estava cheio de areia úmida também não foi possível identificar a pluma, devido ao baixo contraste entre as propriedades físicas. Nesta pesquisa também foram realizadas modelagens numéricas 2D utilizando o método FDTD, onde foram simulados os três experimentos, considerando posições intermediárias da pluma de contaminação em função do tempo. Os resultados das modelagens foram concordantes com os resultados obtidos com dados reais e ajudaram a definir o padrão de reflexão da pluma contaminante. Os promissores resultados indicam que a determinação efetiva da pluma contaminante de óleo numa ETD é possível desde o meio que esteja saturado com água. Portanto, para garantir o sucesso das pesquisas GPR para fins de mapeamento e delineamento de plumas de contaminação em ETDs recomenda-se que a aquisição dos dados seja feita após um período de chuvas. / In this work a 4D GPR survey was accomplished in a laboratory in which a oil leak from a ETD Estação de Transmissão e Distribuição of electric energy was simulated. The studies consisted in an oil spill (15 liters) in three different experiments: i) tank containing dry sand, ii) tank containing wet sand with a hydraulic gradient and iii) tank containing wet sand. In all experiments the objective was detect the contamination plume and evaluate the migration of the oil with the time. The 4D GPR data was acquired using a shielded antenna of 400 MHz model SIR-3000 of the GSSI equipment. The time monitoring of the data acquisition ranged from 2 minutes up to 12 days. In the experiment-1 with the tank filled with dry sand wasnt possible to determine the plume contamination due to the low contrast of the physical properties between the environment and the plume. In the experiment-2 where the tank filled with the dry sand in the presence of a hydraulic gradient, was possible to determine the migration of the plume from the fifth day. Finally, the experiment-3 where the tank was filled with wet sand wasnt possible to identify the plume, due to the low contrast between the physical properties. This research also conducted 2D numerical modeling using the FDTD method, where the three experiments were simulated, considering intermediary positions of the contamination plume with respect with time. The results of the modeling were consistent with the results obtained with real data and helped to define a pattern of reflection of the contamination plume. The promising results indicate that the effective determination of the contaminating plume of oil in an ETD is possible provided that the environment be saturated with water. Therefore, to guarantee the success of researches with GPR for mapping and delineation of contamination plumes in ETDs its recommended that the data acquisition be realized in the rainy season.

In current practice, business processes modeling is done by trained method experts. Domain experts are interviewed to elicit their process information but not involved in modeling. We created a haptic toolkit for process modeling that can be used in process elicitation sessions with domain experts. We hypothesize that this leads to more effective process elicitation. This paper brakes down "effective elicitation" to 14 operationalized hypotheses. They are assessed in a controlled experiment using questionnaires, process model feedback tests and video analysis. The experiment compares our approach to structured interviews in a repeated measurement design. We executed the experiment with 17 student clerks from a trade school. They represent potential users of the tool. Six out of fourteen hypotheses showed significant difference due to the method applied. Subjects reported more fun and more insights into process modeling with tangible media. Video analysis showed significantly more reviews and corrections applied during process elicitation. Moreover, people take more time to talk and think about their processes. We conclude that tangible media creates a different working mode for people in process elicitation with fun, new insights and instant feedback on preliminary results. / Heute, werden Geschäftsprozesse durch speziell ausgebildete Modellierungsexperten erstellt. Wissensträger der Domäne werden befragt, aber sie werden nicht in die Modellierung einbezogen. Wir haben ein haptisches Werkzeug zur Prozessmodellierung entwickelt welches bei der Prozesserhebung zusammen mit den Domänenexperten eingesetzt wird. Wir vertreten die Hypothese, dass dies zur effizienteren Prozesserhebung beiträgt. In diesem Artikel schlüsseln wir "effiziente Prozesserhebung" in 14 operationalisierte Hypothesen auf. Diese werden in einem kontrollierten Experiment mit Hilfe von Fragebögen, Tests und Videoanalyse untersucht. In dem Experiment vergleichen wir unseren Ansatz mit strukturierten Interviews in einem Messwiederholungsdesign. Das Experiment wurde mit 17 Schülern kaufmännischer Berufe in einem Oberstufenzentrum durchgeführt. Sie repräsentieren potentielle Anwender unseres Ansatzes zu Prozesserhebung. Sechs der vierzehn Hypothesen zeigten einen signifikanten Unterschied zwischen den beiden Ansätzen. Teilnehmer berichteten mehr Spass an und mehr Erkenntnisse durch greifbare Prozessmodellierung zu haben. In der Videoanalyse beobachteten wir, dass Teilnehmer häufiger Ihre Aussagen prüfen und korrigieren. Außerdem wurde mehr Zeit für das Reden und Nachdenken über die Prozesse verwendet. Wir schlussfolgern, dass greifbare Medien ein anderes Arbeitsumfeld begünstigen mit mehr Spaß, neuen Erkenntnissen und direktem Feedback auf vorläufige Ergebnisse.

Neste trabalho foi realizado um levantamento GPR 4D em laboratório no qual foi simulado o vazamento de óleo de uma ETD - Estação de Transmissão e Distribuição de Energia Elétrica. Os estudos consistiram em derramar 15 litros de óleo em três experimentos: tanque contendo areia seca, tanque contendo areia úmida mais gradiente hidráulico e tanque contendo areia úmida. Em todos os experimentos o objetivo foi detectar a pluma de contaminação e avaliar a migração desse óleo com o tempo. Os dados GPR 4D foram adquiridos utilizando-se uma antena blindada de 400 MHz modelo SIR-3000 equipamento da GSSI. O monitoramento no tempo de aquisição dos dados variou de 2 minutos até 12 dias. No experimento com o tanque preenchido com areia seca não foi possível determinar a pluma de contaminação devido ao baixo contraste entre as propriedades físicas do meio e a pluma contaminante. No experimento em que o tanque foi preenchido com areia úmida e havia a presença de um gradiente hidráulico, foi possível determinar a migração da pluma de contaminação a partir do 5º dia. Para finalizar, o 3º experimento onde o tanque estava cheio de areia úmida também não foi possível identificar a pluma, devido ao baixo contraste entre as propriedades físicas. Nesta pesquisa também foram realizadas modelagens numéricas 2D utilizando o método FDTD, onde foram simulados os três experimentos, considerando posições intermediárias da pluma de contaminação em função do tempo. Os resultados das modelagens foram concordantes com os resultados obtidos com dados reais e ajudaram a definir o padrão de reflexão da pluma contaminante. Os promissores resultados indicam que a determinação efetiva da pluma contaminante de óleo numa ETD é possível desde o meio que esteja saturado com água. Portanto, para garantir o sucesso das pesquisas GPR para fins de mapeamento e delineamento de plumas de contaminação em ETDs recomenda-se que a aquisição dos dados seja feita após um período de chuvas. / In this work a 4D GPR survey was accomplished in a laboratory in which a oil leak from a ETD Estação de Transmissão e Distribuição of electric energy was simulated. The studies consisted in an oil spill (15 liters) in three different experiments: i) tank containing dry sand, ii) tank containing wet sand with a hydraulic gradient and iii) tank containing wet sand. In all experiments the objective was detect the contamination plume and evaluate the migration of the oil with the time. The 4D GPR data was acquired using a shielded antenna of 400 MHz model SIR-3000 of the GSSI equipment. The time monitoring of the data acquisition ranged from 2 minutes up to 12 days. In the experiment-1 with the tank filled with dry sand wasnt possible to determine the plume contamination due to the low contrast of the physical properties between the environment and the plume. In the experiment-2 where the tank filled with the dry sand in the presence of a hydraulic gradient, was possible to determine the migration of the plume from the fifth day. Finally, the experiment-3 where the tank was filled with wet sand wasnt possible to identify the plume, due to the low contrast between the physical properties. This research also conducted 2D numerical modeling using the FDTD method, where the three experiments were simulated, considering intermediary positions of the contamination plume with respect with time. The results of the modeling were consistent with the results obtained with real data and helped to define a pattern of reflection of the contamination plume. The promising results indicate that the effective determination of the contaminating plume of oil in an ETD is possible provided that the environment be saturated with water. Therefore, to guarantee the success of researches with GPR for mapping and delineation of contamination plumes in ETDs its recommended that the data acquisition be realized in the rainy season.

Context: Software engineering researchers are guided by research principles to conduct any type of research. Though, there are many guidelines to detail how a particular research method can be applied, there is always a necessity to continue and to improve the existing empirical research strategies. The context of this thesis is to address guidelines for conducting controlled experiments in Global Software Engineering (GSE). With this thesis, the state-of-the-art of conducting experiments in GSE research has been explored. Objectives: The goal of this thesis is to analyze the existing experiments in GSE research. Research problems addressed with GSE experiments and the state-of-the-art of overall GSE experiment design need to be analyzed. Appropriate guidelines should be drawn in order to provide strategies to future GSE researchers in mitigating or solving GSE specific experimentation challenges. Methods: A systematic literature review (SLR) is conducted to review all the GSE experiments that are found in the literature. The search process was done on 6 databases. A specific search and quality assessment criterion is used to select these GSE experiments. Furthermore, scientific interviews are conducted with GSE research experts to evaluate a set of guidelines (thesis author’s recommendations) that address the challenges when conducting GSE experiments. Thematic analysis has been performed to analyze the evaluation results and to further improve or implement any suggestions given by the interviewees. Conclusions: The results obtained from the SLR have provided a chance to understand the state-of-the-art and to analyze the challenges or problems when conducting controlled experiments in GSE. The challenges that are identified in GSE controlled experiments are found to be with experiment study-setting, involving subjects and addressing GSE relevant threats to validity in a GSE experiments. 9 guidelines are framed, with each guideline addressing a specific challenge. The final guidelines (that resulted after interviews) provide effective recommendations to GSE researchers when conducting controlled experiments. / +919441308167

Dans la plupart des sociétés développées, en France aussi, on assiste aujourd’hui à une augmentation des disparités sociales et économiques entre les territoires, à l’apparition et à la persistance de quartiers « en difficulté », concentrant les populations les plus fragiles. La « ségrégation spatiale » apparait comme une norme d’organisation urbaine et le territoire comme un enjeu de cohésion sociale. De nombreuses études issues de l’économie et de la sociologie ont cherché à expliquer les écarts de situations observés, notamment sur le marché du travail, entre ces populations. Trois mécanismes principaux expliquent l’impact du lieu d’habitation sur la réussite sur le marché du travail des individus : le spatial mismatch, l’environnement social et la discrimination en raison du lieu de résidence. Dans cette thèse de doctorat nous nous intéressons à ce dernier mécanisme à travers lequel le lieu de résidence affecte les performances sur le marché du travail des individus. Au travers de trois études couvrant trois chapitres, nous mettrons en évidence trois éléments importants dans l’analyse de la discrimination liée au lieu de résidence : le sentiment de discrimination, la discrimination salariale et la discrimination à l’embauche. Ceux-ci sont illustrés par des méthodologies différentes : une analyse de données d’enquêtes subjectives, une analyse de données d’enquêtes objectives et enfin une analyse de données expérimentales. Les résultats principaux sont les suivants. Nous avons montré qu’il y a des effets spécifiques du lieu de résidence sur le ressenti de discriminations des jeunes. Nous avons montré que le lieu de résidence a une influence directe sur la probabilité de trouver un emploi à la sortie des études ainsi que sur le salaire associé à cet emploi. A l’aide d’une décomposition des écarts salariaux nous avons montré qu’il existe une probable discrimination relative au lieu de résidence qui affecte notamment les jeunes femmes. Nous avons trouvé un effet déterminant du département de résidence sur les chances d’accès à l’embauche des individus. Le même effet existe aussi, de manière plus faible, au niveau de la réputation du quartier de résidence des candidats. Une analyse de l’ordre des réponses faites aux candidats a confirmé ces résultats et a même révélé une discrimination encore plus forte. A partir de ces résultats, nous proposons des recommandations de politiques publiques liées aux différents aspects de la discrimination liée au lieu de résidence. / In most of the developed societies, in France too, we are now witnessing an increase of the social and economic disparities between regions, an occurrence and persistence of neighborhoods "in difficulty", concentrating the most vulnerable populations. "Spatial segregation" appears as a norm of urban organization and territory as an issue of social cohesion. Many economic and sociological studies have sought to explain the differentials observed between these populations, particularly in the labor market. Three main mechanisms explain the impact of the place of residence on the individuals' labor market achievement: the spatial mismatch, the social environment and the discrimination based on place of residence. In this doctoral dissertation we focus on this last mechanism through which the place of residence affects the individuals' labor market achievement. Across three studies covering three chapters, we highlight three important elements in the analysis of discrimination related to place of residence: the feeling of discrimination, wage discrimination and employment discrimination. These are illustrated by different methods: an analysis of subjective survey data, an analysis of objective survey data and, finally, an analysis of experimental data. The main results are as follows. We have shown that there are specific effects of place of residence on the feeling of discrimination among young people. We have shown that the place of residence has a direct influence on the probability of finding a job after the school as well as on the salary associated with the job. Using a decomposition of wage differentials we have shown that there is a potential discrimination related to the place of residence that affects especially young women. We found a major impact of the residential county on the individuals’ chances of access to hiring. The same effect also exists, more weakly, at the level of the neighborhood reputation. An analysis of the order of answers to job candidates confirmed these results and revealed an even stronger discrimination. On the basis of these results, we propose policy recommendations related to different aspects of discrimination based on place of residence.

O processo de automatização de teste de software possui alto custo envolvido em sistemas de larga escala, pois exigem cenários de teste complexos e tempos de execução extremamente longos. Além disso, cada etapa do processo de teste requer recursos computacionais e um tempo considerável para a execução de muitos casos de teste, tornando-se um gargalo para as empresas de Tecnologia da Informação (TI). Neste contexto, os benefícios e oportunidades oferecidos pela combinação da computação em nuvem com o Teste como Serviço (Testing as a Service, TaaS), que é considerado um novo modelo de negócio e de serviço atraente e promissor, podem proporcionar um impacto positivo na redução do tempo de execução dos testes de maneira custo-efetiva e aumentar o retorno sobre o investimento ou Return on investment (ROI). Todavia, existe o problema de vendor lock-in, que é o aprisionamento do usuário à plataforma de um fornecedor específico ou serviço de teste, ocasionado pela dificuldade de migrar de um fornecedor TaaS para outro, limitando a utilização dessas novas tecnologias de maneira efetiva e eficiente, impedindo assim, a ampla adoção do TaaS. Como os estudos existentes não são rigorosos ou conclusivos e, principalmente, devido à falta de evidência empírica na área de serviço de teste, muitas questões devem ser investigadas na perspectiva da migração entre os provedores de TaaS. O objetivo deste trabalho é reduzir o impacto ocasionado pelo problema de vendor lock-in no processo de automatização de testes na computação em nuvem, na escrita, configuração, execução e gerenciamento dos resultados de testes automatizados. Neste contexto, foi desenvolvido o protótipo da abordagem intitulada Multi-TaaS por meio de uma biblioteca Java como prova de conceito. A abordagem Multi-TaaS é uma camada de abstração e a sua arquitetura permite abstrair e flexibilizar a troca de fornecedores de TaaS de forma portável, pois permite encapsular toda a complexidade da implementação do engenheiro de software ao desacoplar o teste automatizado de qual plataforma TaaS ele será executado, bem como abstrair os aspectos da comunicação e integração entre as APIs REST proprietárias dos diferentes fornecedores de TaaS. Além disso, a abordagem Multi-TaaS possibilita também sumarizar os resultados dos testes automatizados de forma independente das tecnologias da plataforma TaaS subjacente. Foram realizadas avaliações comparativas da eficiência, efetividade, dificuldade e do esforço de migração entre as abordagens Multi-TaaS e abordagem convencional, por meio de experimentos controlados. Os resultados deste trabalho indicam que a nova abordagem permite facilitar a troca do serviço de teste, melhorar a eficiência e, principalmente, reduzir o esforço e os custos de manutenção na migração entre fornecedores de TaaS. Os estudos realizados no experimento controlado são promissores e podem auxiliar os engenheiros de software na tomada de decisão quanto aos riscos associados ao vendor lock-in no TaaS. Por fim, a abordagem Multi-TaaS contribui, principalmente, para a portabilidade dos testes automatizados na nuvem e da sumarização dos resultados dos testes e, consequentemente, possibilita que o modelo de serviço TaaS na computação em nuvem seja amplamente adotado, de forma consciente, no futuro. / The automation of software testing involves high costs in large-scale systems, since it requires complex test scenarios and extremely long execution times. Moreover, each of its steps demands computational resources and considerable time for running many test cases, which makes it a bottleneck for Information Technology (IT) companies. The benefits and opportunities offered by the combination of cloud computing and Testing as a Service (TaaS), considered a new business and service model, can reduce the execution time of tests in a cost-effective way and improve Return on Investment (ROI). However, the lock-in problem, i.e., the imprisonment of the user in the platform of a specific vendor or test service caused by the difficult migration from one TaaS provider to another limits the effective use of such new technologies and prevents the widespread adoption of TaaS. As studies conducted are neither rigorous, nor conclusive, and mainly due to the lack of empirical evidence, many issues must be investigated from the perspective of migration among TaaS providers. This research aims at reductions in the impact of the vendor lock-in problem on the automation process of testing in cloud computing, writing, configuration, execution and management of automated test results. The prototype of the Multi- TaaS approach was developed through a Java library as a proof of concept. The Multi-TaaS approach is an abstraction layer and its architecture enables the abstraction and flexibilization of the exchange of TaaS providers in a portable way, once the complexity of the software engineers implementation can be encapsulated. The two main advantages of Multi-TaaS are the decoupling of the automated test from the TaaS platform on which it will be executed and the abstraction of the communication and integration aspects among the proprietary REST APIs of the different TaaS providers. The approach also enables the summarization of automated test results independently of the underlying TaaS platform technologies. A comparative evaluation between Multi-TaaS and conventional migration approaches regarding the difficulty, efficiency, effectiveness and effort of migration among TaaS providers was conducted through controlled experiments.The results show the approach facilitates the exchange of test service, improves efficiency and reduces the effort and maintenance costs of migration among TaaS providers. The studies conducted in the controlled experiment are promising and can assist software engineers in decision-making regarding the risks associated with vendor lock-in in TaaS. The Multi-TaaS approach contributes mainly to the portability of automated tests in the cloud and summarization of their results. Finally, this research enables also the widespread adoption of the TaaS service model in cloud computing, consciously, in the future.