Energy-Aware IP Traffic Engineering with Shortest Path Routing E. Amaldi a , A. Capone a , L.G. Gianoli a,b,∗ a Politecnico di Milano, Dipartimento di Elettronica e Informazione Italy b ´ Ecole Polytechnique de Montr´ eal, D´ epartement de g´ enie ´ electrique, Canada Abstract Internet energy consumption is rapidly becoming an issue due to the exponential traffic growth and the rapid expansion of commu- nication infrastructures worldwide. We address the problem of energy-aware intra-domain traffic engineering in networks operated with a shortest path routing protocol. We consider the problem of switching off (putting in sleep mode) network elements (links and routers) and of adjusting the link weights so as to minimize the energy consumption as well as a network congestion measure. To tackle this multi-objective optimization problem with priority (first minimize the energy consumption and then the congestion), we propose a Mixed Integer Linear Programming based algorithm for Energy-aware Weights Optimization (MILP-EWO). Our heuris- tic exploits the Interior Gateway Protocol Weight Optimization (IGP-WO) algorithm for optimizing the OSPF link weights so as to minimize the network congestion. The computational results obtained for seven real network topologies and different types of traffic matrices show that it is possible to switch off a substantial number of nodes and links during low and moderate traffic periods, while guaranteeing that network congestion is low enough to ensure service quality. For comparison purposes, we also report the results obtained with (randomized) greedy procedures. The proposed approach is also validated on two networks of emulated Linux routers. Keywords: Energy-Aware, Traffic Engineering, Shortest Path, OSPF, Weights Optimization 1. Introduction In the last years, the exponential growth of Internet has been accompanied by a considerable increase in its energy consump- tion. Data reported in [1] show that in 2007 Internet has been re- sponsible for 5.5% of the total energy consumption in the world and that the annual increment rate can be estimated around 20- 25%. Other estimates attribute to the ICT sector a compre- hensive energy consumption equal to 156 GW in 2007, with 22 GW consumed exclusively by network equipments (exclud- ing servers in data centers) [2]. Since the annual growth rate is evaluated around 12%, consumptions may reach 95 GW in 2020 [2]. Detailed data reported in [3] about the annual con- sumptions of very important Internet Service Providers (ISP), show that in 2009 the annual power consumptions of the biggest ISPs overcame the 10 GWh per year. Always in [3] it is also shown that in large-scale and medium size ISPs like Telecom Italia and GRNET the total energy consumptions of access and core networks should exceed 400 GWh per year in 2015. For these reasons, the issues of energy saving in IP networks and of power awareness in network design have recently become of great interest in the scientific community and have attracted the attention of device manufacturers and ISPs [4]. Since the seminal work by Gupta and Singh [5], the research community has started developing technologies for manufactur- ing energy efficient network devices, methodologies for power ∗ Corresponding author Email addresses: amaldi@elet.polimi.it (E. Amaldi), capone@elet.polimi.it (A. Capone), gianoli@elet.polimi.it (L.G. Gianoli ) aware network design, as well as energy management strategies for reducing energy wastes of networks in operation [6, 7]. In this paper we focus on energy management of the network and its elements. Energy management in the Internet exploits the fact that net- works are designed and dimensioned to serve the estimated peak traffic demand. During network operation, traffic load varies remarkably over time and even during peak hours it is usually well below network capacity [8]. Unfortunately, cur- rent network device architectures and transmission technolo- gies make their power consumption almost independent of the traffic load [9, 10], with a static power consumption (measured at zero load) of the same order of the peak power (measured at full load). A natural approach to improve the Internet energy perfor- mance is to switch off or put to sleep mode the unnecessary network devices (routers and links) so as to adapt the network consumption to the traffic load, while guaranteeing that traffic demands can still be routed in the remaining sub-network. The advantages of this approach with respect to those based on load balancing are discussed in [11]. Coordinated energy manage- ment strategies can clearly achieve a better performance than those dealing with individual network devices [12]. The routing protocol plays a key role in the energy manage- ment and it affects possible strategies. We consider the most widely used Internal Gateway Protocol (IGP) in IP networks, namely the Open Shortest Path First (OSPF) protocol. The link state approach adopted by OSPF to distribute topology informa- tion and to perform local calculation of shortest paths is based Preprint submitted to Elsevier January 10, 2013