Application Specific Distributed Processing Networks

在过去的十年中，网络上视频信号产生的流量急剧增长。预计IPTV和视频流服务的年增长率将达到两位数，它们将在网络流量中占据越来越重要的地位。像视频信号这样的内容具有对失真的容忍度，这与文件传输应用程序本质上是不同的，在文件传输应用程序中，必须精确地再现内容。本提案的目的是利用这种固有的损失容忍度与利用网络多样性的新路由机制相结合，以增加视频网络的整体鲁棒性。

Over the last decade, the traffic generated by video signals over networks has grown tremendously. With year-to-year growth predictions in the double digits, IPTV and video streaming services will occupy a more and more prominent place in network traffic. Contents such as video signals are characterized by a tolerance to distortion that is inherently different from file transfer applications, in which bit-exact reproduction of contents is a necessity. The aim of this proposal is to exploit this inherent loss tolerance in conjunction with new routing mechanisms that exploit network diversity, in order to increase the overall robustness of video networks. The main question that this proposal will try to answer is how we can improve networked video transmission beyond what is currently feasible, through diversity techniques that take advantage of the signal's inherent loss tolerance. In search for an answer to this question, we will explore several novel architectures for video transmission, that we call Application Specific Distributed Processing Networks, including energy-efficient Multiple Description delivery, Network-Coded Multiple Descriptions and distributed coding with real-field codes. The results obtained in this program will find applications in video contents delivery over the internet, video surveillance systems, healthcare monitoring networks and information networks for the power grid. In parallel, the program will train highly qualified personnel in areas of interest to several potential industrial partners.