Green Traffic Engineering over Polyhedral Demand Uncertainty

Green Traffic Engineering (TE) solutions play a central role in minimizing the energy expenditure of a network where they seek the minimal links/routers/switches required to support a given traffic demand. A key limitation, however, is that they are not designed to be robust against random traffic demands. To this end, this paper reports the first robust, green TE solution, called Green-PolyH, that considers demands defined by a given polyhedral set. Advantageously, Green-PolyH ensures all such demands do not cause the utilization of links to exceed a given threshold. Our experiments over well known topologies show that savings above 80% are achievable whilst remaining robust against traffic changes.