Cascading-failure-aware Disaster Recovery in Optical Cloud Networks

In optical cloud networks, data center nodes process cloud services requested by users. Disasters can damage critical infrastructure elements and possibly trigger overloading or malfunctioning of nearby apparatuses generating cascading failures with severe consequences for the operation of the entire infrastructure. Providing optical connectivity services in the presence of post-disaster cascading failures is crucial. This paper proposes a restoration strategy that combines the ability to recover service after a disaster event while leveraging information about potentially correlated cascading failures. Simulation results show how this strategy successfully reduces the chances of a service being disrupted multiple times by a disaster event and its cascading failures.