Anycast (re)routing of multi-period traffic in dimensioning resilient backbone networks for multi-site data centers

摘要: We consider the problem of dimensioning resilient backbone networks for cloud-like scenarios where demand is to be served at one among several candidate data centers (DCs), and that varies over time, which we assume slotted. thus multi-period traffic, need find routes both a primary DC backup (in case primary, or network connection towards it, fails) account also synchronization traffic (following its own routes) between two chosen DCs. propose path formulation adopt column generation approach: (restricted) master (RMP) selects “configurations” use each in time periods, while pricing problems (PPs) construct new, potentially cost-reducing configurations given demand. Our model allows PPs solved parallel, demonstrate savings achieved by doing so. compare anycast (re)routing strategies, allow spans multiple periods either (i) not rerouted different (ii) only change routes, (iii) freely choices them.