Diseño y evaluación de sistemas de estimación de ancho de banda disponible para servicios adaptativos de vídeo streaming

Abstract

The transmission of multimedia content over the network is one of the most demanded applications. Not surprisingly, more than half of Internet traffic corresponds to video streaming services in the "peak hours". Guarantee the quality of the contents in spite of the great demand of the service is the main challenge that video streaming services face. Moreover, access to multimedia content can be done through various kinds of networks and a wide variety of terminals with different constraints. Therefore, the adaptation of the contents to a heterogeneous environment such as the Internet is a key process in improving the perceived quality of the user. This thesis addresses this issue, providing solutions for three different ways to carry out the adaptation process. First, a non intrusive estimator, based on transcoding techniques, for streaming technology is presented. Analyzing the video packets received by the client application, the estimator is capable of selecting the most suitable encoding bitrate for the available bandwidth in the end-to-end path. Estimation of the available bandwidth is performed with non-intrusive methods based on conventional metrics such as jitter or packet loss but also in novel metric as the linearity of the instants of reception of RTP packets. The estimator has been integrated in a real client / server architecture and evaluated with different network traffic situations. The second solution to an adaptive streaming system is based on the use of Scalable Video Coding (SVC) technology, recently standardized as an extension of H.264 / AVC. Using feedback information from clients about the transmission status, the server is able to select the most suitable combination of SVC layers for the available bandwidth. Bandwidth estimation metrics are adapted to the SVC technology, improving the estimation algorithm. The system is implemented in real equipment and the results show the correct operation and accuracy of the system when adapting to different variations of the available bandwidth. Also, the system scalability is improved when clients access the service simultaneously. Finally, we can not ignore the increasing popularity of HTTP-based streaming solutions. Although protocols based on TCP were not considered suitable for the transmission of multimedia content in real time, the recent development of the standard Dynamic Adaptive Streaming over HTTP (DASH) provides a solution. Thus, based on widely used protocols on the Internet, the reuse of deployed network infrastructure is provided. Here, in this thesis a final adaptive streaming system is proposed using such technology. As in previous proposals, the system is implemented in real equipment and the improvements will be discussed.