A Flooding Method for Exchanging Routing Information in Power -Law Networks

Recent measurement studies of the Internet topology show that the connectivities of nodes exhibit power–law attributes. This topology has two main characteristics: many nodes have a small number of links while a few nodes have a large number of links; and there are fewer hop–counts between nodes. Previous studies of routing mechanisms have evaluated their validity or effectiveness in relatively small networks. However, an evaluation of power–law networks is also needed to clarify their actual validity. In this paper, we evaluate some of the flooding mechanisms used in routing protocols. Our simulation results show that in a power–law network, the flooding mechanism does not scale well due to the concentration of message, whereas random networks that do not have power–law attributes actually scale well. To reduce the concentration of messages, we propose an efficient flooding method for power–law networks. Our method uses probabilistic flooding in which each node relays routing information with a certain probability. Routing information is also exchanged periodically to prevent information mismatches between nodes. The simulation results showed that, compared to conventional flooding approaches, our method reduced the amount of traffic by 50%.