利用Betweenness Centrality计算网络流量矩阵的新算法

Novel Algorithm of Introducing Betweenness Centrality into Traffic Matrix Computing

引入Betweenness Centrality中间度核心性作为候选快照的选择指标,特别是以其中的GBC群组中间度核心性作为考量多链路权重改变时各链路的选取问题,实验结果表明,BC的引入加快了秩的提高,而GBC可以衡量群组大小不同时会对原系统产生影响的程度;同时指出将GBC作为唯一指标在实际操作层面存在问题,需要综合考虑其他因素。最后提出将来结合序列RBC与GBC进行计算的研究方向。

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.