Adaptive Link Rate Switching Based on Traffic Splitting Method for Power Saving
Adaptive Link Rate Switching Based on Traffic Splitting Method for Power Saving
复制标题
基于流量分割方法的自适应链路速率切换以实现节能
DOI:
10.1109/candarw51189.2020.00025
复制
发表时间:
2020
期刊:
影响因子:
--
通讯作者:
Kimura Shigetomo
中科院分区:
文献类型:
--
作者:
Uchida Shingo;Kimura Shigetomo
In recent years, power consumption from network devices has increased as Internet traffic has grown. In order to address this problem, many researchers have proposed schemes that switch the link rate for OpenFlow networks. For example, we previously proposed a method based on the adaptive link rate switching concept which switches between high and low rates based on the usage of the link. Furthermore, the method provided a middle rate, which is constructed from multiple low rate links but consumes less power than that of a high rate link. It showed that power consumption can be reduced from that of simple adaptive link rate switching, but provided only for communications over MPTCP (Multipath TCP). To provide the method for other protocols such as TCP and UDP, this paper proposes a new adaptive link rate switching method based on traffic splitting, which uses OpenFlow meter tables. The meter tables measure traffic speeds and can change the DSCP (Differentiated Services Code Point) field in IP headers if the speeds exceed a pre-defined value such as a link bandwidth. The traffic splitting method assigns a different route for each DSCP value depending on the traffic volume, so dividing a traffic flow into multiple routes. On dividing a traffic flow, the proposed method switches the adaptive link rate from the low rate to the middle rate. When the traffic speed exceeds the middle rate, the proposed method also switches the rate of the shortest link to the high rate and aggregates all divided links onto the high rate link. The paper verifies the feasibility of the proposed method on an emulated network and also calculates the power consumption reduction over a typical traffic pattern to show that power consumption may be reduced from that of a simple adaptive link rate.
登录
查看更多内容
DOI:
10.1109/nbis.2015.15
发表时间:
2015
期刊:
2015 18th International Conference on Network-Based Information Systems
影响因子:
--
作者:
S. Sasaki;Kanayo Ogura;B. B. Bista;T. Takata
通讯作者:
T. Takata
DOI:
10.7763/jocet.2013.v1.63
发表时间:
2013
期刊:
Journal of Clean Energy Technologies
影响因子:
--
作者:
B. B. Bista;Masahiko Takanohashi;T. Takata;D. Rawat
通讯作者:
D. Rawat
DOI:
10.1109/cloudnet.2014.6968985
发表时间:
2014
期刊:
2014 IEEE 3rd International Conference on Cloud Networking (CloudNet)
影响因子:
--
作者:
Adam Markiewicz;P. N. Tran;A. Timm
通讯作者:
A. Timm
DOI:
10.1109/candar.2016.0067
发表时间:
2016
期刊:
2016 Fourth International Symposium on Computing and Networking (CANDAR)
影响因子:
--
作者:
Masato Nishiguchi;S. Kimura
通讯作者:
S. Kimura