Remarks on self-organization and trust in organic computing systems

Abstract

The vision of Organic Computing postulates the advent of multitudes of services provided by collections of intelligent devices by means of selforganized cooperation. Due to their large numbers and their versatile interactions in potentially unlimited networks, it will be unfeasible to explicitly control the behavior of these (partially mobile) devices and their services. Therefore, they will have to respond autonomously in an intelligent way to changing parameters in their environment in order to guarantee appropriate degrees of behavioral robustness and flexibility. Because of these life-like properties, they are called Organic Computing systems. Apparently, the behavior of adaptive, selforganized systems might be hard to predict. At the same time, these systems will have to be trustworthy to be accepted by human users, otherwise, their potential benefits would not be exploited. A necessary prerequisite for establishing and maintaining trust will be the possibility to influence the behavior of Organic Computing systems in a controlled way whenever the system is moving into behavioral regions that are viewed to be unacceptable by human users or by their current execution environment. Hence, an important facet of Organic Computing is the presence of controlled self-organization, enabled by appropriately designed observer/controller mechanisms and methods of data analysis. The talk will elaborate on the state of the art in the area of Organic Computing and, in particular, will focus on possibilities and problems for the engineering of trustworthy organic systems. © Springer-Verlag Berlin Heidelberg 2007.