An algorithm for calculating web video download rate based on data acquisition system

摘要

In the era of information explosion, a monotonically increasing amount of video data is produced every day. Calculating the web video download rate to simulate user perception quickly and accurately has become a valuable issue. The web video traffic was captured from broadband network users and analyzed by our own data acquiring system. In this paper, we described the data acquiring system and proposed four algorithms to calculate the video download rate based on the text log data, and then analyzed the accuracy of the algorithms to select the most reasonable one through the comparisons with conventional method. In addition, this paper also proposed a scheme that can quickly distinguish video data from other types of web traffic with a few log information instead of parsing the HTTP content. Through our experiments, the selected algorithm was proved to be reasonable, convenient and efficient.