作者: Mingzhe Li , Mark Claypool , Robert Kinicki
关键词:
摘要: Most streaming rate selection and buffer optimization algorithms are developed for wired networks can perform poorly over wireless networks. Wireless MAC layer behavior, such as adaptation, retransmissions, medium sharing, significantly degrade the effectiveness of current algorithms. This article presents Buffer Rate Optimization Streaming (BROS) algorithm to improve performance. BROS uses a bandwidth estimation tool designed specifically models relationship between size, data rate, available distribution. optimizes initial resulting in high but with few frame losses underflow events, while still keeping small delay. is implemented Emulated (EmuS) client-server system evaluated on an IEEE 802.11 testbed various conditions. The evaluation shows that effectively optimize size based network conditions, thus achieving better performance than static or jitter removal buffers.