Keywords:
Astract: Cloud computing became the dominant computing model in the IT industry today. With this rapid adoption new questions arise in the area of software testing. Many of the existing functional testing methods can be transferred to cloud-based environments. Though, non-functional testing became a many-faceted problem due to the distributed and dynamic nature of cloud systems. With Google App Engine as an example for an established and widespread implementation of the Platform as a Service model, this thesis examines two in cloud-based software important non-functional requirements: scalability and elasticity. For this goal three questions have been elaborated: Is JMeter a load generator tool suitable for cloud-based applications? How does Google App Engine handle high traffic scenarios? What are the consequences of the limitations enforced by the Google Cloud Datastore, e.g. the entity group write limit? Thereby a special focus is on the migration from existing Java-based applications into the Cloud.

Keywords:
Astract: The capability of UML profiles to serve as annotation mechanism has been recognized in both industry and research. With JUMP, we have presented a fully automatic approach to generate profiles from annotation-based Java libraries. We have demonstrated the practical value of JUMP by contributing profiles that facilitate reverse-engineering and forward-engineering scenarios for the Java platform. Its evaluation shows that automatically generated profiles are equal or even improved in quality compared to profiles currently used in practice.

Keywords: Transformation, SiTra, Java 8, Legacy Code
Astract: Although many different model transformation languages (MTLs) and model transformation frameworks (MTFs) exist, each of them is requiring a developer to learn an additional language. Moreover since many of them follow a declarative approach, they are often less intuitive for programmers familiar with imperative language. Therefore a Java based imperative library called jTL has been developed which enables the user to write the whole transformation process in pure Java. The resulting jTL is applicable to a wide range of transformation tasks that include classical graph based source models or especially source code translation, because transformation of legacy code is still a not yet solved challenge in software migration. Since source code transformation is a quite complex and performance intensive task, the jTL represents an easy to use approach, which focuses on goals like maintainability, readability, reusability and performance improvements. To achieve these goals two similar existing approaches were examined to uncover weaknesses and deduce possible improvements for the resulting jTL. Moreover jTL further benefits from new Java 8 features like bulk operation, lambda expression, functional interface and parallel stream processing. Although jTL has been kept as simple as possible, it is still a complete transformation approach, since it also supports further important features like tracing or reuse mechanisms like rule inheritance which should be part of every state of the art transformation language. As an evaluation, two concrete transformation examples are given that make use of jTL. This should give an idea how a transformation with jTL can be established and what differences and improvements come up with jTL in contrast to other Java based transformation solutions.

Keywords:

Keywords:
Astract: The capability of UML profiles to serve as annotation mechanism has been recognized in both industry and research. Today's modeling tools offer profiles specific to platforms, such as Java, as they facilitate model-based engineering approaches. However, at the modeling level, the set of available profiles is considerably smaller compared to the number of existing Java libraries using annotations at the programming level. This is because an effective mapping between Java and UML to generate profiles from annotation-based libraries is missing. In this paper, we present JUMP to overcome this limitation, thereby continuing existing mapping efforts by emphasizing on annotations and profiles. We demonstrate the practical value of JUMP by contributing profiles that facilitate reverse-engineering and forward-engineering scenarios for the Java platform. The evaluation of JUMP shows that profiles can be automatically generated from Java libraries exhibiting equal or even improved quality compared to profiles currently used in practice.