How would you design a resilient network that can survive hardware failures without impacting production?
