Intelligent download and cache management for interactive non-linear video

Abstract

Since the invention of digital video significant progress has been made in reducing the amount of data needed to be transferred in the World Wide Web while improving viewing experience. However the paradigm of linear behavior has not changed at all. While the feature set of traditional digital video may be sufficient for some applications, there are several use cases where a significantly improved way of interacting with the content is highly desirable. It is possible to organize a video in an interactive and non-linear way. Additional information (for example high resolution images) can be added to any scene of the video. The non-linearity of the video flow and the implementation of additional content not found in traditional videos may lead to an increased download volume and/or a playback with many breaks for downloading missing elements. This paper describes a player framework for interactive non-linear videos. We developed the framework and its associated algorithms to simulate the playback of non-linear video content. It minimizes interruption when the sequence of scenes is directly influenced by interaction, while the traditional viewing experience is not altered. The evaluation showed that it is possible to reach shorter startup times using our strategies than using the strategies of other players. Furthermore, we demonstrated that a prefetching of elements does not lead to an increased download volume in every case. In contrast, it can even decrease the download volume if the right delete strategy is selected. It can be noted that the knowledge of the structure of interactive non-linear videos can be used to minimize startup times at the beginning of scenes while the download volume is not increased.