Novel Algorithm of Introducing Betweenness Centrality into Traffic Matrix Computing

Traffic matrix estimation problem remains one of the research focus for network designer and administrator for many years, especially the traffic estimation of back-bone networks for ISPs. In this article, we introduce betweenness centrality as the measure index of candidate snapshots and group betweenness centrality (GBC) particularly for choosing multiple link weight changes. Our experiments show that the introduction of BC actually accelerates the increase of ranks, and GBC reflects the influences of different group sizes. Some considerations are suggested for further research.