Discriminating between Faults and Attacks in Secure Optical Networks

Security over the optical communications network presents many challenges to both attackers and network providers. The optical medium is not as simple to intrude because attackers must have specialized know-how, and it is not as easy to secure because of the humongous information each fiber carries and its very long length that is difficult to guard from bad actors who may have the required know-how. Conversely, optical system designers do not have the adequate tools in their disposal that are specially designed for securing the optical network and for monitoring and detecting intrusions. Even though intrusion detection in optical networks is a well understood operation in ITU standards, it is also recognized that optical technology cannot accomplish this yet and that the only tools available are simple optical power detectors and error-detection correction codes that monitor the incoming optical power level and the integrity of the signal. Thus, intrusion detection relies on signal processing in the electronic regime and by mapping in the transported frame test signals in order to accomplish this; such operations however are time consuming, inefficient, they may disrupt client service and inconclusive. In this paper, we present a method with which we discriminate between intrusions, and degradations and faults in-service and real-time. That is, we are able to detect intrusion as it occurs with no service disruption, and we believe that our method is fast, is conclusive, is efficient and it does not add to the network cost.