Research Article
A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks
@INPROCEEDINGS{10.1145/1234161.1234165, author={B. Sadeghi and A. Yamada and A. Fujiwara and L. Yang}, title={A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks}, proceedings={2nd International ICST Conference on Wireless Internet}, publisher={ACM}, proceedings_a={WICON}, year={2006}, month={8}, keywords={}, doi={10.1145/1234161.1234165} }- B. Sadeghi
A. Yamada
A. Fujiwara
L. Yang
Year: 2006
A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks A simple and efficient hop-by-hop congestion control protocol for wireless mesh networks
WICON
ACM
DOI: 10.1145/1234161.1234165
Abstract
We provide a design framework for hop-by-hop congestion control algorithms in wireless mesh networks. Mesh networks are becoming increasingly popular and standardization process for development of IEEE 802.11 based mesh networks is an ongoing effort (in 802.11 Task Group S). However, by use of an existing CSMA/CA MAC in such networks, application sources inject as many packets as possible into the network with no regard for whether the packets reach their final destination; hence, in presence of a bottleneck link, queue build-up and packet-drop happen which result in waste of the system resources utilized to deliver packets halfway through the network. We design a simple and efficient hop-by-hop congestion control mechanism which relies on local information available at nodes to detect congestion and modifies 802.11e channel access parameters to adjust the MAC transmission rate of congested flows to a value that removes the congestion. We examine the self-stabilizing characteristics of such an algorithm through an analytical framework and use simulations to show that it results in significant performance gain while maintaining the fairness properties of the underlying MAC.