Legal regulations require companies to preserve all business related data (documents, mails etc.) for a certain time. The amount of information is growing continuously. In order to handle this huge variety of unstructured data Enterprise Content Management systems are built. The resource infrastructure of such a system must be adjusted to peak workloads. As consequence, high acquisition and maintenance costs have to be expended by a company without actually achieving a high utilization of the resources most of the time. Content Management as a Service provides ECM functionality applying the Cloud Computing paradigm. This enables customer to use such a service in a cost-effective manner. A maximal business-effective use of the resource infrastructure on provider side can only be achieved by adapting it dynamically depending on actual workloads and Service Level Agreements. Furthermore, resources have to be shared by multiple tenants. Enterprise Archive Management systems cover a subset of ECM functionality. IBM is developing such a system with the intention to offer it through the cloud. In the context of this project, a multi-tenancy capable dynamic provisioning solution is developed within this thesis. Heuristics are used to resolve performance information about resources and derive the amount of capacity needed in order to satisfy actual user demands. A combination of a straight forward algorithm and the definition of a Constraint Optimization Problem is used to calculate an optimized resource topology and tenant assignments. Furthermore, it will be described how steps which are necessary to transform the old resource topology into then new one can be derived.