On the efficiency of multi-domain routing

In order to cope with scalability issues, large networks are often divided into several domains where routing is performed only on the basis of their aggregating structures. Although scalability can be assured this way, further inaccuracies are introduced in the routing information affecting eventually the efficiency of routing. In our paper, a novel statistical threshold-based aggregation model is suggested with which we compare existing routing methods on the basis of the efficiency they provide. The method itself is a statistical extension of the standard poly-line segment method where re-aggregation is triggered on the basis of measurement data. According to the simulation results, our statistical method outperforms existing methods.