An Energy Efficient Ethernet Strategy Based on Traffic Prediction and Shaping

TL;DR

This paper proposes an energy-efficient Ethernet strategy that predicts traffic patterns to optimize power consumption, leveraging the IEEE 802.3az standard, and demonstrates significant energy savings with minimal delay through simulations.

Contribution

It introduces a novel traffic prediction-based energy saving strategy for Ethernet that combines statistical analysis with IEEE 802.3az features, supported by theoretical bounds and simulation validation.

Findings

Significant energy savings achieved with the strategy.

Limited bounded delay in data delivery.

Effectiveness demonstrated on real and synthetic traffic traces.

Abstract

Recently, different communities in computer science, telecommunication and control systems have devoted a huge effort towards the design of energy efficient solutions for data transmission and network management. This paper collocates along this research line and presents a novel energy efficient strategy conceived for Ethernet networks. The proposed strategy combines the statistical properties of the network traffic with the opportunities offered by the IEEE 802.3az amendment to the Ethernet standard, called Energy Efficient Ethernet (EEE). This strategy exploits the possibility of predicting the incoming traffic from the analysis of the current data flow, which typically presents a self-similar behavior. Based on the prediction, Ethernet links can then be put in a low power consumption state for the intervals of time in which traffic is expected to be of low intensity. Theoretical bounds are derived that detail how the performance figures depend on the parameters of the designed strategy and scale with respect to the traffic load. Furthermore, simulations results, based on both real and synthetic traffic traces, are presented to prove the effectiveness of the strategy, which leads to considerable energy savings at the cost of only a limited bounded delay in data delivery.