Synaptic Connection Autonomic Networks

Konferenz: EURESCOM Summit 2005 - Ubiquitous Services and Applications - Exploiting the Potential
27.04.2005 - 29.04.2005 in Heidelberg, Germany

Tagungsband: EURESCOM Summit 2005 - Ubiquitous Services and Applications

Seiten: 9Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Wang, Fang; Ghanea-Hercock, Robert (Pervasive ICT Research Center, British Telecom, Orion 1/12, Ipswich IP5 3RE, UK)
Sun, Yaoru (Behaviour and Brain Science Centre, School of Psychology, University of Birmingham, Birmingham B15 2TT, UK)

This paper proposes a novel approach to form weighted peer-to-peer networks in a self-organising and decentralised way, termed Synaptic Connection Autonomic Networks (SCAN). Distributed peers in SCAN establish and update their connections or associations based on resource sharing results by following Hebbian learning in a related manner to that in biological neural systems. The strengths of peer associations reflect the utility of one peer to another and continuously adapt over time. In operation SCAN constructs resilient peer-to-peer networks in real time. The result is a more efficient and effective resource sharing mechanism between distributed peers. Simulated experiments verified that SCAN successfully formed P2P networks with correct peer associations and the resource search based on it was continuously improving as the networks were correctly formed.