'This thesis examines the possibilities introduced by the integration of CLR in Microsoft SQL Server 2005. The focus of the examination is to determine how the CLR integration affects the design of a distributed system concerning the performance and scalability criteria. The goal of the thesis is to define a metric empirically, which is able to determine whether a given functionality achieve the best performance and scalability on the application server or database. In order to define the metric, the parameters affecting performance and scalability are determined. A subset of these parameters is varied in a full factorial experimental design to determine the impact of each parameter. The measurements obtained from the performed experiments are compared statistically. Thereby, determining the differences between code deployed on the application server and the two technologies - T-SQL and SQLCLR. Subsequently, the obtained result is used to define the metric. In order to make the metric easy to use for the developer, it has been implemented into a tool with a simple user interface, where the different parameters can be specified.'