Performance Measurement and Bottleneck Analysis for Streaming Media Servers
He Yan, Huang Haocheng, Jinyao Yan · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2013
Traditionally, the quality of service for streaming media systems depends on the network bandwidth, network delay, efficiency of media coding compression and Abstr act.With the rapid development of Internet and the popularity of media contents, the performance of stream media server is becoming an important factor for stream-media systems.By measuring and analyzing the performance constraints of the media server, we find the way to improve the performance of stream media server, and then improve media service.In this paper, we discuss the performance metrics and constraints of stream media server based on the analysis of media streaming system and service.To evaluate the performance of stream media servers, we change the following conditions as the performance metric: the number of connections, the number of concurrent active media files, the format of the media and the bit-rate of the stream.We measure the CPU idle time, IO wait, memory usage, network bandwidth as metrics to find the bottleneck and constraints for streaming server.By increase the capacity of identified bottleneck, we may further improve the performance of steam media servers.