https://doi.org/10.1140/epjb/e2016-60564-2
Regular Article
A dynamic routing strategy with limited buffer on scale-free network
1 School of Electronic and Information
Engineering, Beihang University, Beijing
100191, P.R.
China
2 Collaborative Innovation Center of
Geospatial Technology, Wuhan
430079, P.R.
China
3 Beijing Key Laboratory (NO:BZ0272),
Beijing
100191, P.R.
China
4 Beijing Laboratory for General
Aviation Technology, Beijing
100191, P.R.
China
a e-mail: yufeiwang@buaa.edu.cn
Received:
13
July
2015
Received in final form:
2
December
2015
Published online:
13
April
2016
In this paper, we propose an integrated routing strategy based on global static topology information and local dynamic data packet queue lengths to improve the transmission efficiency of scale-free networks. The proposed routing strategy is a combination of a global static routing strategy (based on the shortest path algorithm) and local dynamic queue length management, in which, instead of using an infinite buffer, the queue length of each node i in the proposed routing strategy is limited by a critical queue length Qic. When the network traffic is lower and the queue length of each node i is shorter than its critical queue length Qic, it forwards packets according to the global routing table. With increasing network traffic, when the buffers of the nodes with higher degree are full, they do not receive packets due to their limited buffers and the packets have to be delivered to the nodes with lower degree. The global static routing strategy can shorten the transmission time that it takes a packet to reach its destination, and the local limited queue length can balance the network traffic. The optimal critical queue lengths of nodes have been analysed. Simulation results show that the proposed routing strategy can get better performance than that of the global static strategy based on topology, and almost the same performance as that of the global dynamic routing strategy with less complexity.
Key words: Statistical and Nonlinear Physics
© EDP Sciences, Società Italiana di Fisica, Springer-Verlag, 2016