Network constrained smoothing: enhanced multiplexing of MPEG-4 video

C. Bewick1, R. Pereira1, M. Merabti1
1Distributed Multimedia Systems Research Group, School of Computing & Mathematical Sciences, Liverpool John Moores University, Liverpool, UK

Tóm tắt

Interactive multimedia applications over multi-hop cell or packet-switched networks have strict bounds on delay, loss, and jitter. Much research has focused on how best to handle the networking issues related to supporting quality of service (QoS) contracts without compromising network efficiency, i.e. improving network utilisation, and reducing connection packet loss and late delivery. This paper describes our novel technique, network constrained smoothing, for smoothing video traffic into a server-side transmission schedule. Our approach combines QoS metrics about video connections together with network loading information in an algorithm that divides the schedule into equal-sized intervals of constant bit-rate. We evaluate performance by simulation and compare with an existing smoothing technique, optimal smoothing. We show the benefits of using network constrained smoothing to synchronise independent multimedia streams. The results show reductions in network loading at peak times and provide manageable information for routers about currently supported multimedia traffic.

Từ khóa

#Smoothing methods #MPEG 4 Standard #Quality of service #Streaming media #Telecommunication traffic #Traffic control #Spread spectrum communication #Jitter #Contracts #Scheduling algorithm

Tài liệu tham khảo

10.1117/12.264304 10.1109/NOSDAV.1997.629390 10.1109/90.720873 10.1109/ISCC.2001.935398 bewick, 2000, A framework for load balancing and network constrained smoothing in multimedia networks, Gemisis grossglauser, 1996, RCBR: A simple and efficient service for multiple time-scale traffic, IEEE\ACM Transactions on Networking hadar, 2000, Statistical Multiplexing and Admission Control Policy for Smoothed Video Streams Using e-PCRTT Algorithm, ITCC'00 fitzek, 2000, MPEG-4 and H 263 video traces for network performance evaluation