IPTV streaming source classification

Abstract

In the last years video streaming over IP networks has changed the entertainment habits of society. Video on demand and multicast services have proliferated, bringing new challenges. Although the majority of the new service providers are compliant with copyright protection policies, some services are streamed without the proper rights, leading to a new kind of content piracy. Therefore, content right owners are interested in finding out the distribution channel followed by these unauthorized contents, in order to learn about illegal distributor sources. We present the first approach to this problem, in which we deal with the classification of IPTV streamed contents on satellital (DVB-S) or terrestrial (DVB-T) sources, both for live and delayed streaming. Our proposal is based on analyzing the time distribution of IP packet dispatches, extracting high-order statistics, and performing the source classification using a SVM. The reported results show the goodness of the proposed approach.