Network Energy Efficiency Through Matching and Reconfiguration

Abstract

Energy efficiency has become a major concern in the design and operation of fixed networks. Energy-Efficient Ethernet (EEE) came as a response to reduce the energy consumption of Ethernet technology. The performance of EEE depends on the link utilization and may degrade when the utilization increases. This is especially true when an energy-aware routing (EAR) protocol is deployed on top of EEE as it tends to aggregate traffic over a subset of links, thereby increasing their utilization. This behavior introduces a tradeoff between EEE and EAR. As network traffic patterns change frequently, it requires a significant effort to configure the EAR protocol on an EEE network in order to reduce the downside effects of the tradeoff, thus benefiting from both the EEE and EAR energy efficiency. In this work, we propose a network matching and reconfiguration algorithm (NetRec), which dynamically adapts the topology control (ESTOP) settings of an EEE network to increase energy savings. Our reasoning is that larger energy savings can be obtained if the network topology is dynamically reconfigured according to traffic demand. NetRec is designed in the context of Software-Defined Networking (SDN). We evaluate NetRec's energy performance in networks of different sizes under varying traffic conditions.