Resolving epidemic network failures through differentiated repair times

In this study, the authors investigate epidemic failure spreading in large-scale transport networks under generalised multi-protocol label switching control plane. By evaluating the effect of the epidemic failure spreading on the network, they design several strategies for cost-effective network performance improvement via differentiated repair times. First, they identify the most vulnerable and the most strategic nodes in the network. Then, via extensive event-driven simulations they show that strategic placement of resources for improved failure recovery has better performance than randomly assigning lower repair times among the network nodes. They believe that the event-driven simulation model can be highly beneficial for network providers, since it could be used during the network planning process for facilitating cost-effective network survivability design.