Abstract
This study presents an evaluation of the Stream Transmission Control Protocol (SCTP) for the transport of the scalable video codec (SVC), proposed by MPEG as an extension to H.264/AVC. Both technologies fit together properly. On the one hand, SVC permits to split easily the bitstream into substreams carrying different video layers, each with different importance for the reconstruction of the complete video sequence at the receiver end. On the other hand, SCTP includes features, such as the multi-streaming and multi-homing capabilities, that permit to transport robustly and efficiently the SVC layers. Several transmission strategies supported on baseline SCTP and its concurrent multipath transfer (CMT) extension are compared with the classical solutions based on the Transmission Control Protocol (TCP) and the Realtime Transmission Protocol (RTP). Using ns-2 simulations, it is shown that CMT-SCTP outperforms TCP and RTP in error-prone networking environments. The comparison is established according to several performance measurements, including delay, throughput, packet loss, and peak signal-to-noise ratio of the received video.
Highlights
Quick developments in network infrastructure, processing power, and storage capability in recent years are making possible the growth of multimedia services through the Internet, giving rise to a bunch of applications that include video streaming, video conference, and high-definition broadcasting
Evaluation scenarios we aim to describe the simulation environment employed to demonstrate the advantages of using Stream Transmission Control Protocol (SCTP) as the scalable video codec (SVC) video transport protocol in comparison with other protocols (RTP and Transmission Control Protocol (TCP)) currently used
The generated traces contain the size of each Realtime Transmission Protocol (RTP) packet, the packet number within the frame, the frame number, the sequence number, and a stream identifier that is used within SCTP
Summary
Quick developments in network infrastructure, processing power, and storage capability in recent years are making possible the growth of multimedia services through the Internet, giving rise to a bunch of applications that include video streaming, video conference, and high-definition broadcasting. Most of these multimedia services are currently implemented using technologies that are not properly designed to cope with the strong variability in the quality of transmission, which is experienced preeminently in wireless and mobile networks. The usual alternative to the lack of a multimedia-friendly transport protocol is the adoption of an end-to-end model in which the Realtime Transmission Protocol (RTP)[1] is encapsulated in UDP, and any auxiliary apparatus for the video transmission is performed in a non-standard way at the application level, including flow and congestion control, retransmission, and redundancy, if any
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call