Techniques for improving multimedia applications in heterogeneous networks

Universal access to multimedia information is now the principal motivation behind the design of next-generation communication networks. However, there are many challenges that need to be solved before transporting high quality multimedia in heterogeneous packet-based networks becomes a reality.; Low channel bandwidth is common in many wireless networks as well as on many Internet paths today. We present the design and evaluation of a transport layer solution for aggregating the bandwidth of multiple channels to form one logical wide channel for the application. The mechanism is scalable with the number of transmission channels. Also it is robust to the asymmetric channel characteristics as well as channel failures.; The high error rate and the rapidly changing characteristics of wireless channels pose a challenge for the transport of compressed video in mobile wireless networks. In motion compensated coding, errors due to packet losses are propagated from reference frames to dependant frames causing lasting visual effects. In addition, the bounded playout delay for interactive video limits the effectiveness of retransmission-based error control. We present a mechanism that combines retransmission-based error control with path diversity in wireless networks, to provide different levels of protection to packets according to their importance to the reconstructed video quality. The new multi-path retransmission mechanism maintains the video quality under different loss rates and mobility speeds, with less overhead compared to error control techniques that depend on reference frame updates.; Although Additive-Increase Multiplicative-Decrease (AIMD) based transport protocols, such as Transmission Control Protocol (TCP), are the dominant transport layer protocols in current Internet, streaming video over such protocols is a challenging problem. As the available bandwidth to the connection changes significantly over medium and long time scales, it is desirable that the application adapts the video quality as a function of the available bandwidth. We present and evaluate adaptation strategies for real-time video, where we switch among several versions of the coded video to match the available network bandwidth accurately and meet client delay constraints.