Improvement of the Prediction-Based Energy Efficient Ethernet Strategy

Numukobwa Segolene,Wanchun Jiang,Kaiqin Liao

doi:10.1109/access.2019.2948840

Numukobwa Segolene, Wanchun Jiang + Show 1 more

Open Access

https://doi.org/10.1109/access.2019.2948840

Copy DOI

Journal: IEEE Access	Publication Date: Jan 1, 2019
Citations: 19	License type: CC BY 4.0

Affiliation: Central South University

Abstract

Ethernet consumes maximum energy even when there is no data transmission. To reduce the power consumption, IEEE 802.3az standardizes the Energy Efficient Ethernet that enhances Ethernet with the low power idle state without data transmission. However, this standard does not describe the specific strategy about when the Ethernet link will enter or exit the low power idle state. Recently, they proposed the EEEP strategy for the 1–10Gbps EEE to reduce power consumption. Specifically, EEEP predicts the future traffic in a time window by the Autoregressive Integrated Moving Average (ARIMA) model and determines when to enter or exit the low power idle state according to the prediction results. However, the EEEP strategy relies on the prediction accuracy of the ARIMA model for good energy saving. This paper proposes to use the Long Short Term Memory (LSTM) model for EEEP to improve the prediction accuracy. Owning to the historic traffic information, the LSTM model can achieve about 11% improvement on the accuracy compared to ARIMA, and thus helps EEEP to achieve better energy saving, according to our trace-driven simulation results.

Highlights

Today, network engineers are working towards expanding the bandwidth and increasing the speed of the network to satisfy user’s requirements [1]
SOLUTION we present the process of building a prediction model based on Long Short Term Memory (LSTM) for the EEEP strategy which includes the Data preparation, the structure of LSTM, the parameters configuration and how the model updates the network after each prediction
We set the maximum number of epochs to be 250, gradient threshold as 1 to avoid gradient vanishing, initial learn rate as 0.005, learn rate schedule as piecewise, learn rate drop period as 125, learn rate drop factor as 0.2, and after training using settings cited above, we initialized the LSTM model for prediction

Summary

INTRODUCTION

Network engineers are working towards expanding the bandwidth and increasing the speed of the network to satisfy user’s requirements [1]. Many studies such as previous studies [8], [9] investigated the implementation and performance analysis for typical real-time industrial communication systems In these investigations, specific traffic features and performance requirements of Ethernet networks have a deep impact on the exploitation of EEE strategy. In the recent study [17], they implemented a strategy called Energy Efficient Ethernet strategy based on traffic prediction and shaping This strategy uses ARIMA (1, 1) model to get the predicted number of packets, and uses this predicted number packets to estimate the time the link should stay in the active state to increase energy savings. LSTM use this historical information to capture well the characteristics and behavior of data traffic, which leads to an accurate prediction regardless of the unlikelihood of traffic.

AND RELATED WORK

WHY USING LSTM

CONCLUSION