Selbstorganisation durch dezentrale Koordination in Verteilten Systemen

Development and configuration of today’s distributed systems are challenged by their increasing complexity, e.g. the spatial distribution of cooperation partners, the heterogeneity of components, and dynamic changes of connection links. Complexity issues arise, in particular, with the use of intelligent, autonomous components that run independently from each other and that act in a constantly changing environment where also the availability of resources and components is changing continuously. Therefore, the effective coordination of system entities often has to be realized in fully decentralized system architectures with specific emphasis on scalability and robustness.

Thus, SodekoVS examines how self-organizing processes can serve as tools in the development of distributed software applications. Here, "self-organization" describes physical, biological and social phenomena, where new global structures arise from the local interactions of individuals (e.g. particles, cells, agents, etc.). Self-organizing processes allow systems to better adapt to varying environments and/or to maintain structures while being subject to perturbations. Enabling similar dynamics in software engineering contexts promises inherently adaptive and robust applications. In consequence, self-organizing systems promise new software quality attributes that are hard to obtain using standard software engineering approaches. In accordance with the visions of, e.g., "autonomic" and "organic" computing, self-organizing systems promote self-adaptability as one major property which leads to software systems with so called "self*"-properties. In particular the self-organizing systems under consideration exhibit self-adaptability which can be seen as a necessary foundation for realizing systems than can manage themselves at runtime. However, the systematic development of systems with such properties still challenges current development practices.

So, SodekoVS addresses the purposeful utilization of self-organizing dynamics to engineer adaptive, distributed software systems. A new development approach is proposed that considers the system architecture as well as the software development methodology as integral intertwined aspects for system construction. Following the proposed process, self-organized coordination dynamics inspired, e.g., by biological, physical and social systems, can be integrated into distributed applications by composing modules that distribute feedback control structures among system entities. These compositions support hierarchical as well as completely decentralized solutions without a single point of failure. This novel development conception is supported by a reference architecture, a tailored programming model as well as a library of ready to use self-organizing patterns. Finally, the methodical application of these development tools is evaluated in SodekoVS by respective case studies based on agent-based, decentralized management systems as, e.g., web service architectures, traffic control, etc.