Traffic dynamics in scale-free networks with limited buffers and decongestion strategy

Abstract

We studied the information traffic in Barabási–Albert scale-free networks wherein each node has a finite queue length to store the packets. It is found that in the case of the shortest path routing strategy, the networks undergo a first-order phase transition, i.e. from a free flow state to a full congestion state, with increasing packet generation rate. We also incorporate the random effect (namely random selection of a neighbor to deliver packets) as well as a control method (namely the packet-dropping strategy of the congested nodes after some delay time T) into the routing protocol to test the traffic capacity of the heterogeneous networks. It is shown that there exists an optimal value of T for the networks to achieve the best handling ability, and the presence of an appropriate random effect also contributes to the performance of the networks.