Topology-based Proactive Network Protection in Large-scale Failure Scenarios

Abstract

Natural disasters are among the main causes of large-scale failure scenarios. The dynamic behaviour of natural disasters and their probabilistic failure pattern indicates a need for a dynamic probabilistic protection approach to address the issue and reduce the number of disrupted connections in the network. Appropriate data protection against a time-varying destructive phenomenon serves to prevent damage before it occurs. In this case, the endangered route should be detected and the data passing through should be rerouted through more reliable paths prior to failure. The high-risk paths can be detected based on the decision parameters in the dynamic preventive approach. In this article, we develop a new method to improve the assignment procedure of decision factors in preventive protection models using network topological properties and links failure probability. The proposed decision parameters are adaptive with the network conditions and updateable while the disaster spreads through the region. The adaptive characteristic of the rerouting decision parameter will enhance network efficiency and facilitate network management.