Self-Balanced and Self-Adaptive Routes in Unstructured P2P Networks

Abstract

A simple strategy for routing and load balancing in unstructured P2P network based on ant-like agents is presented. In [Unger, Wulff], the authors present an algorithm, which enables to build a sorted structure on top of a P2P Network in order to support searching in such systems and to reduce maintenance times. Based on this structure, a routing mechanism that allows the user to send messages between two peers in finite time will be provided The current work adopts ant systems as introduced by Dorigo, Maniezzo and Colorini and looks for alternative paths near to the optimal route in case of congestion conditions. Although ANTs take advantage of collaborative behavior and sign the closest paths, it is also possible to rapidly saturate the communication channels between source and target nodes. Therefore, a load balancing criterion is considered, as well.