Using Bio-inspired Models to Design Peer-to-Peer Overlays

Abstract

Peer-to-peer systems are subject to frequent changes, fact which makes the resource management a very complex task. With the aim of reaching extreme scalability, centralized models no longer represent feasible solutions. Thus, we need to build flexible, adaptive, robust, and self-organizing systems, able to overcome all these challenges. We propose the design of a peer-to-peer overlay inspired from the layout of a honeycomb. In contrast with the common 2-dimensional grid, this structure reduces the number of neighbors that each node is connected to. Therefore, reorganizing the system in case of failures is done faster and with lower overhead. To generate and maintain the hexagonal overlay, the nodes rely on specific mechanisms explained and analyzed in this work.