Abstract
Today modern multimedia delivery services usually distribute their contents by means of the streaming technology. In conventional platform, the point-to-point delivery model is adopted but also known as less efficient. To extent scalability, some services apply periodic broadcast to provide an efficient platform that is independent of the number of clients. Fast Broadcasting, Staircase Broadcasting, Harmonic Broadcasting and Skyscraper Broadcasting are some novel schemes in this subject. For a popular video, these periodic broadcast services can significantly improve viewer initial delay by distributing cyclic video segments through multiple channels. However, the requirement of a large amount of client buffers is a major drawback of them. Additionally, video segments are broadcasted periodically by a per-determined schedule. If the prediction of the popularity of a video is incorrect, some bandwidth will be wasted. Moreover, period broadcasting also should not be adequate to carry out on heterogeneous networks. In this thesis, we propose a novel dynamic periodic broadcasting scheme that gets the benefit of less buffer usage and flexibility for bandwidth adaptation. That is, to prevent unnecessary bandwidth wasted, only requisite video segments will be transmitted. We also integrate a receiver-driven bandwidth adaptation strategy on our dynamic periodic broadcasting scheme to fit the channel bandwidth of each client. We then design our channel adjustment scheme. Instead of using a fixed number of channels to serve all clients, we can adjust the transmission rate as well as the required channel number owing to the available bandwidth on the path from the server to the client. Our experimental results show that this dynamic scheme can perform well in heterogeneous networks.