Green Networks: Energy Efficient Design for Optical Networks Balagangadhar G. Bathula, Jaafar M. H. Elmirghani School of Electronic and Electrical Engineering University of Leeds, LS2 9JT, United Kingdom Email: {b.bathula, j.m.h.elmirghani}@leeds.ac.uk Abstract—In this paper we discuss new energy efficient paradigms for optical networks. With the increasing number of high bandwidth applications, devices used in backbone networks such as optical networks increase. Energy consumption of optical networks is an important issue that has to be addressed. In this work we propose novel routing algorithms for decreasing the energy consumption of optical networks. We propose sleep cycle protocols for use in the network nodes. Energy-Aware optical network protocols can impact the Quality of Services (QoS) such as bit-error-rate (BER) and delay. Our proposed algorithm maintains a trade-off between energy consumption and the QoS. Keywords: Energy management, QoS, Grid networks. I. I NTRODUCTION Information and communication technology (ICT) has a profound impact on the economy and the environment. A study estimated that the Internet equipment consumed roughly 8 % of the total energy (i.e. electricity) in the United States with the prediction of growth to 50 % within a decade [1]. The development of faster communication links is likely to con- tribute to the demand for faster computers, which is likely to increase energy consumption. In addition computer networks at present, require additional power demanding equipment, such as servers, amplifiers, routers, filters, storage devices and communication links. These communication components con- sume significant amounts of energy. With the ever-increasing demand for bandwidth, these communication components tend to increase and hence energy efficiency is an important issue. WDM networks can provide a huge amount of bandwidth for present and future Internet applications. These networks are deployed in large scale to form the backbone network. Dynamic optical capabilities in WDM can be achieved with the advent of optical-cross connect (OXC) nodes, which can switch the wavelengths completely in the optical domain. Optical networks are evolving into a complex interconnection of circuit-switched networks due to the continued growth in high-bandwidth applications. The E-Science community is a fine example of such applications, which has already started using optical networks as the backbone network to support multi-giga bit connectivity. These developments led to research in the area of intelligent optical control plane [2]. Using intelligent optical control planes lightpaths, (or wave- length channels) can have dynamic route selection polices. All-optical networks (AON) have to maintain a wavelength continuity constraint. Lightpaths established in wavelength routed networks can be maximized with the help of dynamic wavelength discovery paths [3]. Constraint-based path selec- tion policies help to meet the QoS demands of the service effectively. Contention resolution schemes such as deflection routing are used in wavelength routed optical burst switched networks (WR-OBS) [4]. In this paper we propose a multi path selection approach to minimize the energy consumption of the optical core network. These wavelength routed paths may have to forgo minimum distance paths and choose a path which is at a larger distance. This tends to degrade the QoS like BER and delay. Given the service requirement conditions, we propose to select the paths such that the overall energy consumed by the optical network decreases and at the same time maintain the service threshold conditions. We propose a clustered node architecture similar to the one proposed in [5]. It has to be noted however that in [5] the clustering approach was proposed to reduce packet loss, here we utilize a similar approach with a different objective: energy minimization. The selection of these clusters can be static or dynamic. Dynamic cluster partitioning of the core network can be based on network load or load balancing. However we restrict our study only to the static case. By using an efficient optical control management mechanism, these clusters can be set to ON or OFF states. During the OFF cycle the nodes that belong to the cluster, adopt a sleep mode, cutting down the traffic routed through them. Thus a cluster isolates itself from the network. The energy reduction achieved due to a sleep cycle is at the cost of decrease in QoS. Hence the traffic that is by-passed from the cluster during its OFF cycle should be aware of the service threshold conditions. Thus there is a need to develop an intelligent and efficient control plane and associated algorithms for the implementation of Energy- Aware Optical networks (EAON). The remainder of the paper is organized as follows: in Section II we explain the optical cross-connect (OXC) node and compute the total energy consumed by each OXC switch. In Section III we describe the network architecture for Energy Efficient Optical networks (EEON). We discuss the proposed energy-efficient routing algorithm with the help of a network example in Section IV. Finally in Section V we conclude this paper with possible future extensions. II. ENERGY CONSUMED IN OPTICAL NETWORKS In this section we calculate the energy required to transmit an optical bit and the energy spent by each OXC shown in Fig. 1. The OXC consists of mux/de-mux and a wavelength cross-connect switch. An OXC can also have the functionality to add/drop channels, using the transmitter and receiver array shown in Fig. 1. Energy is defined as the product of power and 978-1-4244-3474-9/09/$25.00 ©2009 IEEE Authorized licensed use limited to: University of Leeds. Downloaded on June 19, 2009 at 08:13 from IEEE Xplore. Restrictions apply.