Http-Based Adaptive Streaming Platform Building And Algorithm Design In Multi-User Condition

Nowadays, adaptive streaming develops at a high rate of speed. Dynamic adaptive streaming over HTTP (DASH) is widely adopted in industry as a new MPEG standard. DASH requests best quality of media based on current available bandwidth with adaptive algorithm. To obtain the appropriate quality of media, an accurate estimation of bandwidth is needed. This paper presents a novel bandwidth estimation method, which consists of bandwidth detective method and bandwidth computing method. By distinguishing different bandwidth variation patterns based on bandwidth detective method, we calculate the bandwidth in a more stable and accurate way. A DASH platform is built which offers adaptive streaming service with optional adaptive algorithms. With popularity of adaptive streaming, it is more likely to share a network link with multiple clients. Thus adaptive algorithm design for multi-use has significant meanings. Different algorithms are analyzed in multi-user condition with efficiency, fairness and stability indicator, and existing problems are located.