We study traffic dynamics in growing scale-free networks. Both the scale-free structure of the network and the adaptive nature of the dynamics which controls traffic in the network are considered in the model. The model is investigated with computer simulations and analytically for the case of a scale-free tree. For the scale-free tree, an exact formula and its power law approximation of the complementary cumulative distribution function of link load (edge betweenness) is presented. We examine whether the scaling properties of the network affect the performance of the transport mechanism and estimate the average number of competing transport mechanisms at bottlenecks. We find that bottlenecks tend to appear on the periphery of the network as the performance increases. Various bandwidth allocation strategies are compared. We show that the best performance is achieved when capacity is distributed proportionally to the expected load of links. We demonstrate that it is necessary to study both the topology and the dynamics of the transport mechanism to understand the whole system.
- Scale-free network
- Transmission Control Protocol
ASJC Scopus subject areas
- Agricultural and Biological Sciences(all)