Paper Info
Title
A dynamic local method for bandwidth adaptation in bundle links to conserve energy in core networks
Abstract
Energy savings in bundle links of the core network has been investigated recently. The bundle link technique is widely used in current core networks to provide higher bandwidth and more resilience. Basically a bundle link is composed of several high-speed physical sublinks which could be SONET connections, Ethernet circuits, etc. in order to make them work together as a virtual connection. In current network operations, all of the sublinks are activated if the bundle link is powered on although the sublinks could be shut down or brought up independently. Smartly and dynamically shutting down and bringing up sublinks and their attached ports according to the traffic demand or estimation could greatly increase the link utilization and save a large amount of energy. In previous work, we proposed a local heuristic threshold-based method to explore the potential energy savings in core networks by adjusting the number of active sublinks in the bundle links. In this paper, we further explore the possibilities to reach a better tradeoff between energy-saving performance and congestion risk by setting different value combinations of two parameters: the utilization threshold and sublink-adding strategy. From that, we propose a dynamic and hybrid local heuristic threshold-based algorithm, which we call HDLHT algorithm, to achieve a better tradeoff between energy-savings and congestion risk. In HDLHT algorithm, the bundle links are given different combinations of these two parameters according to the burstiness of their traffic. A simulation experiment deploying HDLHT algorithm was conducted on a synthetic network and the results show that HDLHT can greatly reduce the number of congestion occurrences with limited decrease in energy savings.
Year
DOI
Venue
2013
10.1109/ANTS.2011.6163644
Optical Switching and Networking
Keywords
DocType
Volume
Green networking,Bundle link,Bin packing,Load balance,Distributed algorithm
Journal
10
Issue
ISSN
ISBN
4
2153-1676
978-1-4673-0093-3
Citations 
PageRank 
References 
7
0.53
11
Authors
2
Name
Order
Citations
PageRank
Liu Lin15710.59
Byrav Ramamurthy21283110.58