Keywords:

Keywords:

Keywords:

Keywords: model-driven engineering, executable model, executable modeling language, metamodeling, software language engineering, behavioral semantics, fUML
Astract: Model-driven engineering (MDE) is a software development paradigm aiming to cope with the growing complexity of software systems by raising the level of abstraction. In this paradigm, a system is defined by means of models using modeling languages that enable developers to abstract away from implementation and platform details. From the models, complete implementations may be (semi-)automatically generated by utilizing model transformation techniques. As MDE puts models into the center of software development, adequate methods for creating, analyzing, and utilizing models are crucial. Due to the large body of used modeling languages, means for efficiently developing adequate tool support for modeling languages are needed. To address this need, the automation techniques provided by MDE may also be applied to automate the development of such tool support. This is current practice for developing syntax-based tools. However, the automated development of semantics-based tools has not reached the same level of maturity yet. The goal of this thesis is to fill this gap and provide a solution for automating the development of semantics-based tools for executable modeling languages. Therefore, a language and methodology for developing behavioral semantics specifications based on the standardized language fUML are proposed. To provide the basis for developing semantics-based tools, the execution environment of fUML was extended with means for execution control, runtime observation, and runtime analysis. Based on these extensions, a generic model execution environment for modeling languages whose behavioral semantics is defined with fUML was developed. This environment provides the foundation for developing semantics-based tools for executable modeling languages, which has been shown by the implementation of a semantic model differencing tool and other semantics-based tools.

Keywords: Cloud computing, Model-Driven-Engineering, UML
Astract: Cloud computing had and still has a major impact on how applications are made accessible for the users. Due to the advantages cloud computing has, there is a demand to migrate applications to the cloud. Unfortunately there does not exist general guidelines how to define the required application execution environments and deployment requirements so that they can be interpreted by any arbitrary cloud provider. In the last years, cloud providers came up with approaches to be able to describe cloud resources in form of an interpretable template. Just recently, in November 2013, OASIS published the open standard TOSCA [44], which aims to unite existing proprietary approaches and standardise them. Approaches following a declarative way of describing orchestrated cloud resources are quite recent and are extended frequently, as it is a promising possibility of illustrating complex dependencies and limitations of computing resources in a way that can be read by human beings as well. This thesis firstly discusses model driven engineering and cloud computing separately and afterwards, how they can be combined. The main aim is to create a model that contains enough information about dependencies, limitations and application specific requirements, which can support the migration of the application to the cloud. Furthermore, the master-s thesis proposes a process, which is subdivided into two parts: Deployment and Provisioning. The first step is about creating UML models and refining them with UML extensions (classifiers, profiles and stereotypes), which consists out of cloud computing specific attributes. The second step converts the model into a template, by means of applying model to text transformations, in order to be interpretable and executable by cloud providers. Existing solutions only address partial aspects of the whole problem, focusing on other objectives. One of the main goal of this thesis is the creation of a unified and model-based solution, whose processes and tools support the application modeler and make a (semi-)automatic execution of the deployment and provisioning of an application in the cloud possible.