A scalable formal framework for analyzing the behavior of nature-inspired routing protocols

Abstract

Nature-inspired routing algorithms for fixed networks is an active area of research. In these algorithms, ant- or bee-agents are deployed for collecting the state of a network and providing them to autonomous and fully distributed controllers at each network node. In these routing systems the agents, through local interactions, self-organize to produce system-level behaviors which show adaptivity to changes and perturbations in the network environment. The formal modeling of such fully self-organizing, distributed and adaptive routing systems is a difficult task. In this paper, we propose a scalable formal framework that has following desirable features: (1) it models important performance metrics: throughput, delay and goodness of links, (2) it is scalable to any size of topology, (3) it is robust to changing network traffic conditions. The proposed framework is utilized to model a well-known BeeHive protocol which is further validated on NTTNeT (a 57 node topology). To the best of our knowledge, this is the first formal framework that has been validated on such a large topology. © 2008 Springer-Verlag Berlin Heidelberg.