Efficient Search and Lookup in Unstructured P2P Overlay Networks Inspired by Swarm Intelligence

Abstract

Finding an efficient way to locate incomplete data in complex distributed systems is a challenging task and, due to dynamic nature of the Internet, requires to be updated constantly. As the problem refers to selection of an efficient search algorithm, different types of algorithms are proposed up to now. A huge complexity and dynamics presented in such systems imply a necessity of usage of an intelligent, self-organized solution. However, such intelligent algorithm should not possess an additional complexity. In this paper, we propose a new, simple and effective swarm-based metaheuristic for search and lookup in an unstructured P2P system inspired by behavior of bark beetles in nature. Also, a Physarum Polycephalum mechanism is adapted for this purpose. Both algorithms are compared with Dictyostelium discoideum (Dd)-slime mold, Gnutella, AntNet and k-Walker search mechanisms and tested by using two different models, Actor and Peer. The benchmarks measured by different metrics cover a parameter sensitivity analysis, comparative analysis and scalability analysis. Both algorithms show very promising results in terms of performance and scalability.