Towards an Organic Network Control System Sven Tomforde, Marcel Steﬀen, J¨org H¨ ahner, and Christian M¨ uller-Schloer Leibniz Universit¨ at Hannover Institute of Systems Engineering Appelstr. 4, 30167 Hannover, Germany [tomforde | steffen | haehner | cms]@sra.uni-hannover.de Abstract. In recent years communication protocols have shown an in- creasing complexity – especially by considering the number of variable parameters. As the performance of the communication protocol strongly depends on the conﬁguration of the protocol system, an optimal param- eter set is needed to ensure the best possible system behaviour. This choice does not have a static character as the environment changes over time and the inﬂuencing factors of the optimisation are varying. Due to this dynamic environment an adaptation depending on the current situ- ation on the particular node within a communication network is needed. This paper introduces an Organic Network Control system which is able to cover this task and it also demonstrates the strengths of the pro- posed approach by applying the system to a Peer-to-Peer protocol and evaluating the achieved results. 1 Introduction Due to the continuously increasing interconnectedness and integration of large distributed computer systems and the dramatical growth of communication need new protocols are being developed continuously. Simultaneously, researchers and engineers try to guarantee the sustainability of such systems by optimising and enhancing existing algorithms. This leads to a growing complexity of the partic- ular methods and a rapidly increasing number of possibilities to conﬁgure the resulting systems. This complexity of the conﬁguration task leads to the need of new ways to guarantee a good approximation of the optimal behaviour of partic- ular nodes within a network and provide an automatic adaptation to the needs of the system’s user. Although the system should be able to cope with diﬀerent situations, which might not have been foreseen during the development, the user does not want to have additional eﬀort to conﬁgure or administrate his system. Organic Computing (OC - cf. [1]) is a recent research area which focuses on building self-organised systems to solve complex problems. Autonomous entities are acting without strict central control and achieve global goals although their decisions are based on local knowledge. The authors assume that due to the complexity of the particular tasks not all situations can be foreseen during the development process of the system. Therefore, the system must be adaptive and equipped with learning capabilities, which leads to the ability to learn new actions and strategies for previously unknown situations. The self-control of network entities is also part of the focus of Autonomic Computing (AC - cf. [2]).