Automated Refinement of High-Level Policies in Self-Organizing Networks

Diplomarbeit, November 2012

The increasing number of mobile devices as well as the resulting increased usage of the internet on these devices leads to a greater complexity of the mobile communications networks and the underlying systems managing these. To get this complexity under control self-organization helps to reduce the manual work. Thereby, policies play a decisive role. Policies can be called guidelines or constraints on a system. By specifying them on a high level of abstraction and refining them automatically to executable machine-oriented policies, the system’s behavior can be adapted at any time. In this thesis the model-driven approach for policy refinement is introduced and how it can be applied to five exemplary policies of a self-organizing long term evolution network. However, a multitude of such policies occur in complex systems. Thus, refinement code can not always be written manually for them. That’s why this thesis also deals with the automated code generation for further policies. A final chapter presents some methodologies extending the existing approach and achieving automation to a greater extent.