Bandwidth Estimation of Rate Adaption Algorithm in DASH

Nowadays, adaptive streaming is quite popular and is expected to develop at a high rate of speed. Dynamic Adaptive Streaming over HTTP (DASH) is a new 3GPP and MPEG standard for adaptive steaming over HTTP which is currently being extensively accepted in the industry. DASH requests best quality of media under available bandwidth with adaptive algorithm which is implemented in clients. To obtain the appropriate quality of media, an accurate estimate of the bandwidth is needed. In this paper, we present a novel algorithm to estimate the bandwidth even more accurately. This bandwidth estimation algorithm involves bandwidth detective method and bandwidth computing method. Outliers within a short term, slight fluctuations and persistent variations can be distinguished through our proposed bandwidth detective method. Hereafter, we compute the estimated bandwidth in different methods based on the result of bandwidth detection. We build a DASH platform and compare our proposed algorithm to the VLC adaptive algorithm. According to the results, it is demonstrated that our algorithm possesses more accurate bandwidth estimation and higher buffer level. Our proposed algorithm is sensitive to persistent variations and robust to slight fluctuations.