Designing peer-to-peer networks to mimic nature
The Technical University of Dresden in Germany received funding from the Information Society Technologies Programme to apply lessons learned from biological systems to computer applications. They set their sights on Peer-to-peer (P2P) networks, which offer a number of advantages over client-server architectures, but are limited in terms of speed and robustness. An innovative search algorithm was designed based on behaviour observed in immune systems in nature. The software engineers with the Technical University of Dresden replicated the proliferation and mutation response, which enabled faster transmission of message packets associated with the query. In addition, the peers comprising the P2P network were encouraged to form clusters around common types of data. This allowed the network to develop a memory over time, which was subsequently exploited to further reduce the amount of time necessary to execute queries. The new software, which has been protected by copyright, has been published to the website of the project. It is accompanied by a report composed by the Technical University of Dresden that describes the theoretical background behind the design in detail.
