The advances in Cloud computing and in modern Web applications have raised the need for highly available and scalable distributed databases to accommodate the big data being created and consumed. Along with the explosion in data growth comes the necessity to rapidly evolve databases and schemas to meet user demands for new functionality. A special attention is being paid to the vast amounts of semi-structured and un-structured data, and the data management tools should reflect the support for these needs. This has lead to the development of new Cloud serving systems such as "Not Only" SQL (NoSQL) databases. NoSQL databases were driven by the scalability needs of the big companies, such as Google, Facebook, Amazon, and Yahoo. While the demands of these key players are different from those of small and medium enterprises in terms of scalability, the core problem is the same - storage arrays are not scalable and force you into expensive, forklift upgrades. These facts combined with changes in how IT resources are delivered and consumed through the Cloud computing paradigm, projects adopting NoSQL solutions are not a hype anymore. NoSQL databases are being offered as a service by the big Cloud providers, such as Google, Amazon, Microsoft, but by smaller vendors as well. In this master thesis we investigate the possibilities and limitations of mapping relational database schemas to NoSQL schemas when migrating the database layer to the Cloud. Based on literature research we provide recommendations and guidelines with regard to schema transformation and discuss the implications at other application architecture layers, such as business logic and data access layer. We extend an existing data migration tool and methodology for incorporating the migration guidelines and hints. Moreover, we validate our work based on a chosen sub-set of relational and NoSQL databases by using example data from the established TPC-H benchmark.