The way computer systems are built dramatically changed over time. Starting from huge monolithic systems for many users up to ubiquitous computer environments with a lot of distributed and embedded computing power. Also the way these systems are configured has changed. The Autonomic and Organic Computing initiatives try to solve the upcoming management of complexity problems by utilizing biologically or sociologically inspired methods. One of the demands to the systems is self-configuration. This paper proposes a self-configuration process for the networked nodes of the AMUN middleware based on a social behavior. Aim of the self-configuration is to find a good distribution of services by calculating a quality of service based on the given resources and the required resources of the services. A configuration specification is provided and flooded into the network to start the cooperative job assignment algorithm. A terminal verification step guarantees the correctness of the found configuration. Evaluations of the completely distributed self-configuration process are provided.