Research On Scalable Streaming Media Distribution Technology

With the rapid development of Internet technology, streaming media represented by network video has become one of the most popular applications on the Internet. Compared with traditional applications, streaming media occupies more bandwidth and lasts longer during transmission. One of the biggest challenges facing the widely used CDN(Content Delivery Network) is poor system scalability and high deployment costs. The P2P(Peer to Peer) streaming media distribution technology, with which peers share their media resources with each other, can effectively improve the scalability of the distribution system, and reduce the deployment costs. However, pure P2 P streaming solutions are also more fragile due to churn. The recently emerged hybrid distribution system(P2P-CDN), incorporating the advantages of P2 P and CDN technology, has been a hot research topic.Scalable streaming media distribution technology is a key research issue of hybrid distribution system, including two key challenges of forecasting the demand of cooperative nodes and selecting credible cooperative nodes. To estimate the quantity of demanded cooperative nodes in hybrid streaming system, this paper makes a deep research on cooperative node demand forecasting methods. According to the sampling theory, this paper establishes a forecasting model of the video demanding reaching rate, and provides a method of calculating the cooperative node demand based on Fluid model.To address the problem of selecting credible cooperative nodes in hybrid streaming system, this paper proposes a credible cooperative node selection model based on trust. In this model, servers primarily select nodes with higher trust, better overall performance and longer past online time as cooperative nodes. Deposit and digital signature mechanism further constrain the behavior of the selected cooperative nodes. Simulations demonstrate that the proposed model can effectively identify malicious nodes and improve the system's user experience. Finally, this paper shows the applying of the credible cooperative node selection model in hybrid CDN-P2 P streaming system.