ABSTRACT
DITEN-University of Genoa / CNIT - University of Genoa Research Unit, Genoa, Italy
Roberto Bruschi
CNIT - University of Genoa Research Unit, Genoa, Italy
Franco Davoli
DITEN-University of Genoa / CNIT - University of Genoa Research Unit, Genoa, Italy
Paolo Lago
DITEN-University of Genoa / CNIT - University of Genoa Research Unit, Genoa, Italy
24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694 24.2 The Devices’ Internal Sources of Energy Consumption . . . . . . . . . 696 24.3 Power Scaling Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 698 24.4 A Theoretical Model of the Trade-Off Between Energy
Consumption and Quality of Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . 703 24.4.1 Introducing Energy-Aware Parameters . . . . . . . . . . . . . . . . . . 703 24.4.2 The Traffic Arrival Process Model . . . . . . . . . . . . . . . . . . . . . . 705 24.4.3 The Queuing System Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705 24.4.4 Model Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 708
24.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715 Author Contact Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
Fundamentals,
In the last few years, the research field of “green” and energy-efficient networking has gained great interest on the part of network providers and equipment manufacturers. Such interest springs from heavy and critical economical needs, since both energy cost and network electrical requirements exhibit an almost steadily growing trend. For example, as shown in [1], energy consumption of the Telecom Italia network had reached more than 2 TWh (about 1% of the total Italian energy demand) in 2006, increasing by 7.95% with respect to the previous year. Similar trends can be generalized to a large part of the other telecoms and service providers; recent studies done by the Global e-Sustainability Initiative (GeSI) [2] foresee a jump in the overall network energy requirement of European Telcos from about 21.4 TWh in 2010 to 35.8 TWh in 2020 if no Green Network Technologies (GNTs) would be adopted. This alarming growth in energy network requirements is essentially a consequence of the increase in data traffic volume (which follows Moore’s law, by doubling every 18 months [3]), causing an even larger increase in the number and capacity of deployed network devices. For instance, high-end IP routers are even more based on complex multi-rack architectures, which provide more and more network functionalities and continue to increase their capacities with a factor of 2.5 every 18 months [4]. At the same time, as shown in Figure 24.1, based on the data from [4], and as suggested by Dennard’s scaling law [5], silicon technologies (e.g., CMOS) improve their energy efficiency at a lower rate with respect to routers’ capacities and traffic volumes, by increasing of a factor of 1.65 every 18 months. Though core routers represent a minority of deployed
in Next-...
