Characteristics of live media sharing communities

Social networking applications are one of the main sources of content in the current internet, where currently the most of the traffic is composed of transferring text and still images. In the near future it is expected that members (users) of social networks will move to richer media, generating and consuming a large variety of multimedia content within these networks, thus Future Internet should be prepared to handle such demands. The underlying user community drives the media consumption and determines the network conditions in such cases. Proper modeling of the user community within such a system would make possible the accurate evaluation of the impact of these applications. Content generated by live media sharing communities differs from stored video content available online for streaming or download (e.g., NetFlix, YouTube, etc.) and the underlying user pattern is different, as well. In this paper we identify the key aspects of user communities in live multimedia distribution systems and investigate statistical features of each aspect. We have collected and analysed data from different live video portals and show that the statistical features of different portals follow similar trends.