Quantitative and qualitative network robustness analysis under different multiple failure scenarios

Society depends now more strongly than ever on large-scale networks. Failures of great significance have occurred in the last years. Thus, it has become of vital importance to define robustness metrics. Classical robustness analysis has been focused on the evaluation of topological characteristics. In this paper we extend this analysis introducing two new robustness metrics that include traffic service requirements: QuaNtitative Robustness Metric (QNRM) and QuaLitative Robustness Metric (QLRM). A review of some well known graph robustness metrics is provided. In order to compare these metrics with the proposals, a set of six topologies (random, small-world and scale-free) is evaluated. Finally, it is shown that our two new metrics are able to evaluate the performance of a network under a given kind of impairment.