A Swarm Inspired Method for Efficient Data Transfer

Abstract

Yutaka KAWAI†a) , Student Member, Adil HASAN † †b) , Go IWAI †c) , Takashi SASAKI †d) , and Yoshiyuki WATASE †e) , Nonmembers SUMMARYIn this paper we report on an approach inspired by Ant Colony Optimization (ACO) to provide a fault tolerant and efficient means of transferring data in dynamic environments. We investigate the problem of distributing data between a client and server by using pheromone equations. Ants choose the best source of food by selecting the strongest pheromone trail leaving the nest. The pheromone decays over-time and needs to be continually reinforced to define the optimum route in a dynamic environment. This resembles the dynamic environment for the distribution of data between clients and servers. Our approach uses readily available network and server information to construct a pheromone that determines the best server from which to download data. We demonstrate that the approach is self-optimizing and capable of adapting to dynamic changes in the environment.