Reducing the impact of targeted attacks in interdependent telecommunication networks

In targeted attacks the network elements are removed with the purpose of maximizing the impact of the attack in the network. When one network interacts with another, the network's critical parts may change due to failure spreading between them. Thus, it is interesting to identify changes arising in robustness when interdependent networks with similar topological properties interact. We focus on interdependent telecommunication networks modeled as two Erdös-Rényi (ER) random graphs. These graphs are still the most common models for physical or control plane topologies. Previous works have identified that single random networks are highly vulnerable to sequential targeted attacks based on nodal betweenness centrality. In order to interconnect the two ER networks, three link patterns for the interdependency matrix are proposed and the network robustness for each matrix is analyzed in targeted attacks. Results indicate that different interdependency matrices are able to identify new critical parts of the network. Moreover, by selecting a suitable interdependency matrix, the robustness level under targeted attacks can be also improved.