Due to the large set of service offerings in cloud computing, developers frequently face the challenge of selecting appropriate cloud services for their cloud applications. Moreover, with the advent of hybrid cloud computing, it becomes even more complex to design and implement cloud applications, since hybrid cloud applications combine services across multiple target clouds, and possible across multiple cloud providers.

Several cloud modeling languages such as Blueprint, CloudDSL and CAML aim to aid the developer in designing applications for the cloud. However, these models still require, that the developer has knowledge on specific service offerings of his target clouds, or they do not support hybrid cloud scenarios.

To this end, this thesis uses cloud modeling language Clams, since it supports cloud agnostic modeling techniques by using cloud computing patterns, which define best practice solutions to recurring development problems. The Clams model this pattern representation in combination with UML sequence diagrams as placeholders, which get replaced by suitable service offerings according to a given selection criterion, e.g. services with highest availability. Thus, Clams gives the developers the possibility to design and evaluate cloud applications without the knowledge of the particular service offerings.

However, Clams does not support the design of hybrid cloud applications. Therefore, the contribution of this thesis is to extend the modeling capability of Clams to model hybrid cloud applications. The contributions of are twofold. First, extending the Clams modeling language, in order to support multiple cloud providers, which offer their services at multiple cloud sites. Second, finding suitable cost models to assess the costs between the service offering of different cloud providers.

In this master thesis an approach is presented in which the data model of OpenClams is extended. The extension includes the modeling of hybrid cloud applications in OpenClams and the estimation of the monthly costs of a modeled cloud application based on the regions of the services. Furthermore, the approach is implemented and a prototypical evaluation is carried out. This evaluation of the approach shows that hybrid cloud applications can be modeled in OpenClams and that the estimation of the monthly costs of a cloud application delivers correct results with minimal deviation. An analysis of the deviation shows that the deviation in costs is related to free quotas of services and further discounts of the cloud provider.