Network virtualization for disaster resilience of cloud services

Today's businesses and consumer applications are becoming increasingly dependent on cloud solutions, making them vulnerable to service outages that can result in a loss of communication or access to business-critical services and data. Are we really prepared for such failure scenarios? Given that failures can occur on both the network and data center sides, is it possible to have efficient end-to-end recovery? The answer is mostly negative due to the separate operation of these domains. This article offers a solution to this problem based on network virtualization, and discusses the necessary architecture and algorithm details. It also answers the question of whether it is better to provide resilience in the virtual or physical layer from a cost effectiveness and failure coverage perspective.