A Delivery System for Scalable Video Streaming Using the Scalability Extension of H.264/AVC over DiffServ Networks

With the rapid development of multimedia applications over Internet, the requirement of high bandwidth and stringent Quality of Services is more and more urgent. In this paper, we present a delivery system for streaming video using the scalability extension of H.264/AVC over diffserv networks. It first maps the Network Abstraction Layer Unit of the SVC from different spatial-temporal layers and different quality layers using fine grain SNR scalability (FGS) to 16 types of source marks. Then it uses the Improved Two Rate Three Color Marker (ITRTCM) at theingress of a diffserv domain to map these source marks to three drop precedence according to current network conditions. The core routers of diffserv networks delivery those packets according the predefined Per-Hop Behaviors (PHBs) associated with a Differentiated Services Code Points (DSCP). Experimental results show that the proposed delivery system can provide better guarantee of QoS to the streaming applications using SVC.