Self-Organizing Networked Systems

A distributed approach
A self-organizing system typically consists of many networked entities that organize themselves and cooperate through the exchange of information without the need of a centralized control instance but using a distributed approach. Information is exchanged locally among individual entities in the frame of the fulfillment of a certain global objective. Some simple and high-level rules in the individual entities lead to sophisticated Favorable propertiesSelf-organizing systems have various favorable properties:

• They typically adapt very easily to changes from inside and outside the system.
  
  
• Additional entities can be added and will be assimilated into the global system.
  
  
• Entities may be removed without too much affect on the global system, and other entities may take over crucial tasks of them.
  
  
• Furthermore, self-organizing systems scale very well and there is no bottleneck of a central authority.

The cross-disciplinarity
Research into self-organizing networked systems not only has technical and user-oriented aims, it also enables a high degree of interdisciplinary.

We encounter self-organizing systems on an almost daily basis in:

• the formations of swarms of fish and migratory birds
  
  
• the interplay of termites when they build their hills
  
  
• the activity of body cells during the healing of wounds.

In many areas of nature, single individuals or organisms work together without central coordination, but in perfect harmony. Large areas of the economy have already been functioning for many years according to this paradigm.

Valuable characteristics
In self-organizing systems, the instances involved form decisions based on limited local knowledge. This leads to a desired emergent behavior of the entire system. Naturally self-organizing systems also possess many characteristics that are of value in technical systems. They are:

• flexible and reliable
  
  
• very adaptable and can be extended at any time.