Energy-Aware Networks: Reducing Power Consumption By Switching Off Network Elements Luca Chiaraviglio * , Marco Mellia * , Fabio Neri * * Dip. di Elettronica, Politecnico di Torino, Italy, Email: {last name}@tlc.polito.it Abstract—According to several studies, the power consumption of the Internet accounts for up to 10% of the worldwide energy consumption, and several initiatives are being put into place to reduce the power consumption of the ICT sector in general. To this goal, we propose a novel approach to switch off network nodes and links while still guaranteeing full connectivity and maximum link utilization. After showing that the problem falls in the class of capacitated multi-commodity flow problems, and therefore it is NP-complete, we propose some heuristic algorithms to solve it. Results shows that it is possible to reduce the number of links and nodes currently used by up to 25% and 10% respectively while offering the same service quality. Keywords: green networks, power-aware, optimization I. I NTRODUCTION Power consumption in general, and of ICT technologies in particular, has become a key issue during the last few years. The ratio of power demand versus power resources is constantly growing, and energy cost is increasing at a constant rate. The electricity jumped up of about 35% in Italy during the period 2004-2007 [1]. Moreover, Green House Gases (GHG) emissions have a negative impact on the world climate [2], and people are becoming more conscious about the problems that will arise in the near future due to this. According to a number of studies, ICT alone is responsible for a percentage which varies from 2% to 10% of the world power consumption [3], due to the ever increasing diffusion of electronic devices. In this scenario, the power consumption of telecommunication networks, and of the Internet in particular, is not negligible. For example, considering a data center, the network infrastructure alone is responsible of 23% of the overall power consumption, even without taking into account the energy necessary for equipment cooling [4]. The study of power-saving network devices has been intro- duced over these years, starting from the pioneering work of [6]. In [7] the ideas of Adaptive Link Rate (ALR) and protocol proxying are proposed. The ALR technique is the capability to dynamically adapt the channel transmission rate according to the actual traffic demand. Proxying instead leverages on the idea of turning off high performance devices during low activity periods while moving offered services to low power (and low performance) components. Both these techniques require the change of protocols, and both consider a single pair of devices, i.e. routers sharing the same link, or a couple of high/low performance servers. More recently, some effort was devoted to investigate how to reduce the power consumption of the entire network in- frastructure, and not of single components only. In [5] some simple measurements about power consumption of networking devices are first presented; then authors consider a network topology and evaluate the total network consumption given the power footprint of each element. They consider two scenarios: in the first one all devices are turned on, while in the second one only the minimum number of elements are actually powered on to guarantee the service. The reduction of the corresponding power consumption is finally evaluated. In this paper, we consider a wide area network scenario. Given the network topology and a traffic demand, we evaluate the possibility of turning off some elements (nodes and links) under connectivity and Quality of Service (QoS) constraints. The goal is to minimize the total power consumption of a large network, in which usually resource overprovisioning is large. We investigate some simple optimization algorithms. In particular, we selectively power off nodes and links of the topology following different strategies. Results show that it is possible to reduce the percentage of powered nodes and links up to 10% and 25% respectively, while guaranteeing that the resource utilization is still below a given threshold. e.g., 80%. II. PROBLEM FORMULATION An informal description of the design problem studied in this paper is the following: Given i) a physical network topology comprising routers and links, in which links have a known capacity, ii) the knowledge of the average amount of traffic exchanged by any source/destination node pair, iii) the maximum link utilization that can be supported, iv) the power consumption of each link and node, Find the set of routers and links that must be powered on so that the total power consumption is minimized, Subject to flow conservation and maximum link utilization constraints. We provide an Integer Linear Programming (ILP) formula- tion of the problem to precisely define it. Let us represent the network infrastructure as a graph G =(V,E), where V is the set of vertices and E is the set of edges. Vertices represent network nodes, while edges represent network links, being N =| V | and L =| E | the total number of nodes and links respectively. Let c ij be the capacity of link from node i