Breakdown of the Internet under Intentional Attack

摘要: We study the tolerance of random networks to intentional attack, whereby a fraction $p$ most connected sites is removed. focus on scale-free networks, having connectivity distribution $P(k)\ensuremath{\sim}{k}^{\ensuremath{-}\ensuremath{\alpha}}$, and use percolation theory analytically numerically critical ${p}_{c}$ needed for disintegration network, as well size largest cluster. find that even with $\ensuremath{\alpha}\ensuremath{\le}3$, known be resilient removal sites, are sensitive attack. also argue that, near criticality, average distance between in spanning (largest) cluster scales its mass, $M$, $\sqrt{M}$, rather than ${log}_{k}M$, expected away from criticality.