Implementation method of prediction-based 40-100Gbps Ethernet energy saving strategy

Abstract

The invention discloses an implementation method of a prediction-based 40-100Gbps Ethernet energy saving strategy. An IEEE 802.3bj standard defines two energy-saving states of Fast Wake and Deep Sleepfor 40-100Gbps Energy Efficient Ethernet (EEE), and energy saving quantity and state transition delay corresponding to the different energy saving states are different. The prediction-based 40-100Gbps Ethernet energy saving strategy periodically predicts the arrival number of data frames in a next cycle according to historical information on the premise of ensuring that data frame tail delay doesnot exceed predetermined expected delay, selects an energy saving state used in the next cycle, controls, on the basis of a current prediction result, timing of leaving an energy saving state by EEE,and thus obtains a better compromise on two aspects of increasing energy saving quantity and reducing tail delay time. Experimental results show that compared with an existing energy-saving strategy,the prediction-based 40-100Gbps Ethernet energy saving strategy disclosed by the invention can achieve a better energy saving effect while the tail delay of the data frames is controlled to be in expected values.

Description

technical field [0001] The invention relates to the field of 40-100Gbps energy-saving Ethernet, in particular to a prediction-based 40-100Gbps Ethernet energy-saving policy implementation method. Background technique [0002] In recent years, as 40-100 Gbps Ethernet appears in commercial applications, the IEEE802.3bj working group promulgated a corresponding Energy Efficient Ethernet (EEE) standard in September 2014. In this standard, the link power consumption in the deep sleep state is only 10% of the peak value, and the transition time from the deep sleep state to the active state is 5.5μs, while the link power consumption in the fast wake-up state is usually 70% of the peak power consumption, but the transition time from the fast wake state to the normal operation state is shorter at 0.34μs. EEE uses these two states to save power consumption. When there is no data frame to be transmitted in the link, the link will choose to enter one of the low power consumption state...

AI Technical Summary

Problems solved by technology

However, none of the existing strategies achieve a good compromise in terms of energy saving and latency control.

Images