Currently, in many corporate environments and scientific institutions, a number of different IT-applications runs on a number of different computer systems. Computer systems on client-/server-architecture, which meanwhile have established themselves traditionally on an equal basis, carry most of the burden and, due to the heterogenity of system landscapes and computer systems, present operators of data centers with major challenges that, combined with a persistently high level of growth, lead to enormous costs in the fields of administration and operation.

This project aims to analyze patterns for the improvement of the operation of data centers through virtual-, adaptive-, and grid computing. These fields of computing, which are still new in mass operation, are being defined and analyzed according to possible areas of application within data centers. Therefore, a design- and architectural model is developed, which is meant to show how the capacity requirement for the IT infrastructre in data centers can be made more efficient through the application of these techniques. The main thesis of this project is that a reduced application of infrastructure ressources is possible. The precondition is the determination of the actually required capacity.

For the analysis of the capacity requirement, in the case study of SAP, empirical data from server landscapes of a data center will be gathered in the context of traditional operation. These data serve the development of a model for the capacity determination for the consolidated infrastructure operation and the exmination of this thesis.