Test Driven Development (TDD) é uma prática ágil que ganhou popularidade ao ser definida como parte fundamental na eXtreme Programming (XP). Essa prática determina que os testes devem ser escritos antes da implementação do código. TDD e seus efeitos têm sido amplamente estudados e comparados com a prática Test Last Development (TLD) em diversos trabalhos. Entretanto, poucos estudos abordam TDD no desenvolvimento de Web Services (WS), devido à complexidade em testar as dependências entre os componentes distribuídos e as particularidades da Service Oriented Architecture (SOA). Este trabalho tem por objetivo definir e validar uma abordagem para o desenvolvimento de WS baseada na prática de TDD, denominada WS-TDD. Essa abordagem guia os desenvolvedores no uso de TDD durante o desenvolvimento de WS, sugerindo ferramentas e técnicas para lidar com as dependências e as particularidades de SOA, com foco na criação dos testes unitários e integrados automatizados na linguagem Java. No intuito de definir e validar a abordagem proposta, quatro métodos de pesquisa foram executados: (i) questionário presencial; (ii) experimento; (iii) entrevista presencial com cada participante do experimento e (iv) triangulação dos resultados com as pessoas que participaram nos três métodos anteriores. De acordo com os resultados obtidos, a WS-TDD mostrou-se mais eficiente quando comparada a TLD, aumentando a qualidade interna do software e a produtividade dos desenvolvedores. No entanto, a qualidade externa do software diminuiu, apresentando um maior número de defeitos quando comparada a TLD. Por fim, é importante destacar que a abordagem proposta surge como uma alternativa simples e prática para a adoção de TDD no desenvolvimento de WS, trazendo benefícios a qualidade interna e contribuindo para aumentar a produtividade dos desenvolvedores. Porém, a qualidade externa do software diminuiu ao utilizar a WS-TDD. / Test Driven Development (TDD) is an agile practice that gained popularity when defined as a fundamental part in eXtreme Programming (XP). This practice determines that the tests should be written before implementing the code. TDD and its effects have been widely studied and compared with the Test Last Development (TLD) in several studies. However, few studies address TDD practice in the development of Web Services (WS), due to the complexity of testing the dependencies among distributed components and the specific characteristics of Service Oriented Architecture (SOA). This study aims to define and validate an approach to develop WS based on the practice of TDD, called WS-TDD. This approach guides developers to use TDD to develop WS, suggesting tools and techniques to deal with SOA particularities and dependencies, focusing on the creation of the unitary and integrated automated tests in Java. In order to define and validate the proposed approach, four research methods have been carried out: (i) questionnaire; (ii) practical experiment; (iii) personal interview with each participant in the experiment and (iv) triangulation of the results with the people who participated in the three previous methods. According to the obtained results, WS-TDD was more efficient compared to TLD, increasing internal software quality and developer productivity. However, the external software quality has decreased due to a greater number of defects compared to the TLD approach. Finally, it is important to highlight that the proposed approach is a simple and practical alternative for the adoption of TDD in the development of WS, bringing benefits to internal quality and contributing to increase the developers’ productivity. However, the external software quality has decreased when using WS-TDD.

Com a necessidade cada vez maior de agilizar a entrega de novos desenvolvimentos ao cliente e de diminuir o tempo de desenvolvimento das aplicações, a priorização de casos de teste possibilita a detecção das falhas presentes na aplicação mais rapidamente por meio da ordenação dos casos de teste a serem executados. E, com isso, possibilita também que a correção destas falhas inicie o mais brevemente possível. Entretanto, quando os casos de teste a serem priorizados são testes automatizados de sistema, critérios tradicionais utilizados na literatura como cobertura de código ou modelos do sistema deixam de ser interessantes, dada a característica inerente deste tipo de teste na qual a organização e a modelagem adotadas são ignoradas por se tratarem de testes de caixa preta. Considerando a hipótese de que casos de teste automatizados grandes testam mais partes da aplicação e que casos de teste similares podem estar testando a mesma área da aplicação, parece válido crer que a execução dos casos de teste de sistema priorizando os testes mais complexos pode alcançar resultados positivos quando comparada à execução não ordenada dos casos de teste. É neste cenário que este trabalho propõe o uso dos grafos de chamadas dos próprios casos de teste como critério para priorização destes, priorizando assim a execução dos casos de teste com a maior quantidade de nós no seu grafo. A abordagem proposta neste trabalho mostrou, por meio de dois estudos de caso, ser capaz de melhorar a taxa de detecção de falhas em relação à execução não ordenada dos casos de teste. Além disso, a abordagem proposta obteve resultados semelhantes as abordagens tradicionais de priorização utilizando cobertura de código da aplicação. / With the increasing need to streamline the delivery of new developments to the customer and reduce application development time, test case prioritization allows a quicker detection of faults present in the application through the ordering of test cases to be executed. Besides that, a quicker detection enables also the correction of these faults to start as soon as possible. However, when the test cases to be prioritized are automated system tests, traditional criteria used in the literature like code coverage or system models become uninteresting, given that this type of test case, classified as black box test, ignores how the application was coded or modeled. Considering the hypothesis that bigger automated test cases verify more parts of the application and that similar test cases may be testing the same application areas, it seems valid to believe that giving a higher priority to more complex test cases to be executed first can accomplish positive results when compared to the unordered execution of test cases. It is on this scenario that this project studies the usage of call graphs from test cases as the criterion to prioritize them, increasing the priority of the execution of test cases with the higher number of nodes on the graph. The approach proposed in this document showed through two case studies that it is capable of improving fault detection rate compared to unordered test cases. Furthermore, the proposed approach achieved similar results when compared to a traditional prioritization approach using code coverage of the application.

Indiana University-Purdue University Indianapolis (IUPUI) / Evolutionary testing (ET) techniques (e.g., mutation, crossover, and natural selection) have been applied successfully to many areas of software engineering, such as error/fault identification, data mining, and software cost estimation. Previous research has also applied ET techniques to performance testing. Its application to performance testing, however, only goes as far as finding the best and worst case, execution times. Although such performance testing is beneficial, it provides little insight into performance characteristics of complex functions with multiple branches. This thesis therefore provides two contributions towards performance testing of software systems. First, this thesis demonstrates how ET and genetic algorithms (GAs), which are search heuristic mechanisms for solving optimization problems using mutation, crossover, and natural selection, can be combined with a constraint solver to target specific paths in the software. Secondly, this thesis demonstrates how such an approach can identify local minima and maxima execution times, which can provide a more detailed characterization of software performance. The results from applying our approach to example software applications show that it is able to characterize different execution paths in relatively short amounts of time. This thesis also examines a modified exhaustive approach which can be plugged in when the constraint solver cannot properly provide the information needed to target specific paths.

Indiana University-Purdue University Indianapolis (IUPUI) / Current mainstream computing systems have become increasingly complex. Most of which have Central Processing Units (CPUs) that invoke multiple threads for their computing tasks. The growing issue with these systems is resource contention and with resource contention comes the risk of encountering a deadlock status in the system. Various software and hardware approaches exist that implement deadlock detection/avoidance techniques; however, they lack either the speed or problem size capability needed for real-time systems. The research conducted for this thesis aims to resolve issues present in past approaches by converging the two platforms (software and hardware) by means of the Graphics Processing Unit (GPU). Presented in this thesis are two GPU-based deadlock detection algorithms and one GPU-based deadlock avoidance algorithm. These GPU-based algorithms are: (i) GPU-OSDDA: A GPU-based Single Unit Resource Deadlock Detection Algorithm, (ii) GPU-LMDDA: A GPU-based Multi-Unit Resource Deadlock Detection Algorithm, and (iii) GPU-PBA: A GPU-based Deadlock Avoidance Algorithm. Both GPU-OSDDA and GPU-LMDDA utilize the Resource Allocation Graph (RAG) to represent resource allocation status in the system. However, the RAG is represented using integer-length bit-vectors. The advantages brought forth by this approach are plenty: (i) less memory required for algorithm matrices, (ii) 32 computations performed per instruction (in most cases), and (iii) allows our algorithms to handle large numbers of processes and resources. The deadlock detection algorithms also require minimal interaction with the CPU by implementing matrix storage and algorithm computations on the GPU, thus providing an interactive service type of behavior. As a result of this approach, both algorithms were able to achieve speedups over two orders of magnitude higher than their serial CPU implementations (3.17-317.42x for GPU-OSDDA and 37.17-812.50x for GPU-LMDDA). Lastly, GPU-PBA is the first parallel deadlock avoidance algorithm implemented on the GPU. While it does not achieve two orders of magnitude speedup over its CPU implementation, it does provide a platform for future deadlock avoidance research for the GPU.