Research Article
A link adaptive transport protocol for multimedia streaming applications in multi hop wireless networks
@INPROCEEDINGS{10.4108/ICST.MOBIMEDIA2007.1580, author={Pirabakaran Navaratnam and Haitham Cruickshank and Rahim Tafazolli}, title={A link adaptive transport protocol for multimedia streaming applications in multi hop wireless networks}, proceedings={3rd International ICST Conference on Mobile Multimedia Communications}, proceedings_a={MOBIMEDIA}, year={2010}, month={5}, keywords={multi hop wireless networks medium contention transport protocols multimedia streaming rate control TFRC.}, doi={10.4108/ICST.MOBIMEDIA2007.1580} }
- Pirabakaran Navaratnam
Haitham Cruickshank
Rahim Tafazolli
Year: 2010
A link adaptive transport protocol for multimedia streaming applications in multi hop wireless networks
MOBIMEDIA
ICST
DOI: 10.4108/ICST.MOBIMEDIA2007.1580
Abstract
Transport layer performance in multi hop wireless networks has been greatly challenged by the intrinsic characteristics of these networks. In Particular, the nature of congestion, which is mainly due to medium contention in multi hop wireless networks, challenges the performance of traditional transport protocols in such networks. In this paper, we first study the impact of medium contention on transport layer performance and then propose a new transport protocol for supporting quality of service requirements in multi hop wireless networks. Our proposed protocol, Link Adaptive Transport Protocol provides a systemic way of controlling end-to-end rate for multimedia streaming applications, based on the degree of medium contention information received from the network. Simulation results show that the proposed protocol provides an efficient scheme to support quality of service requirements, such as end-to-end delay, jitter, packet loss rate, throughput smoothness and fairness for media streaming applications. In addition, our scheme requires few processing cycles and minimum overhead and does not maintain any perflow state table at intermediate nodes. This makes it less complex and more cost effective.