Title :
An assessment of power-load proportionality in network systems
Author :
Ricca, Marco ; Francini, Andrea ; Fortune, Steven ; Klein, Thomas
Author_Institution :
Politec. di Torino, Turin, Italy
Abstract :
Power-load proportionality is a necessary feature in networks that aim at maximizing energy efficiency. Even in networks that handle near-capacity loads very efficiently, energy savings increase substantially if the power consumption closely follows the offered load. Rate adaptation is a common denomination for a set of technologies that operate at different timescales to establish power-load proportionality. To propel their deployment, it is useful to assess how compatible they are with existing network systems and identify the design upgrades that can maximize their energy savings in future networks. The formulation of accurate energy profiles for current equipment is a first step in this direction. We run extensive power measurement experiments to compile the energy profiles of five network systems from multiple vendors. Our results show only negligible signs of power-load proportionality in all five cases: to really make a dent in the carbon footprint and operational cost of packet networks, future system designs must pervasively deploy rate adaptation technologies, especially those that control power state transitions at the packet timescale.
Keywords :
distribution networks; energy consumption; load management; power measurement; carbon footprint; energy efficiency; energy profiles; energy savings; near-capacity loads; operational cost; packet networks; power consumption; power measurement experiments; power-load proportionality; rate adaptation; Hardware; Load modeling; Ports (Computers); Power demand; Power measurement; Switches;
Conference_Titel :
Sustainable Internet and ICT for Sustainability (SustainIT), 2013
Conference_Location :
Palermo
DOI :
10.1109/SustainIT.2013.6685190