实时流媒体在目前的因特网中传输经常会因为网络带宽的不足或数据包丢失严重,使得接收方播放质量受到严重影响。基于跨层设计的思想,在应用层使用自适应前向纠错算法,即使流媒体数据有一定的丢包率,接收方仍然能完整地恢复出原视频序列。在数据链路层采用有利于流媒体传输的区分服务模型,用以增大高优先级数据流(如实时流媒体数据流)传输过程的吞吐量。数学分析和仿真结果均表明,该策略能使接收方获得最大的可播放帧率,从而有效提高流媒体传输的可靠性和实时性。
Media streaming applications over networks are often disturbed by data packet loss and errors as well as the limited network bandwidth, which degrades the video quality on the receiving side. Based on cross-layer design, an adaptive forward error correction algorithm at the application layer, in the hope that the media could be played on receivers with high quality. On the other hand, a diffserv model used for the data link layer is proposed, so that high-priority data streams (such as media streaming data streams) can be assigned more bandwidth. The simulation results and mathematical analysis demonstrate that the proposed mechanism achieves better quality of media streaming, in terms of playable frame rate, reliability and real-time performance on the receiving side.