Defensive resource allocations with security chokepoints in IPv6 networks

Abstract

Securely configured Internet Protocol version 6 networks can be made resistant to network scanning, forcing attackers to propagate following existing benign communication paths. We exploit this attacker limitation in a defensive approach in which heightened security measures are deployed onto a select group of chokepoint hosts to enhance detection or deter penetration. Chokepoints are chosen such that, together, they connect small isolated clusters of the communication graph. Hence, attackers attempting to propagate are limited to a small set of targets or have to penetrate one or more chokepoints. Optimal placement of chokepoints requires solving an NP-hard problem and, hence, we approximate optimal solutions via a suite of heuristics. We test our algorithms on data from a large operational network and discover that heightened security measures are only needed on 0.65% of the nodes to restrict unimpeded attacker propagation to no more than 15% of the network.