基于流媒体业务的低解码时延网络编码构造方法

Constructing Network Code with Low Decoding Delay

  • 摘要: 针对网络编码在实际应用中存在较大的解码时延问题,提出了一种低时延的网络编码构造方法,即从源节点到目的节点的多条互不相交的路径中选择一条作为便捷路径,在便捷路径中传输不作编码的原始数据包,从而减小目的节点的解码时延。为尽量多的目的节点分配便捷路径的问题可归结为图的最大完全子图问题。理论分析和实例计算的结果表明,该方法在不损失网络编码的吞吐量增益的前提下,可有效地减小目的节点的解码时延,从而保证网络编码能更好地应用于流媒体等时延敏感业务的传输。

     

    Abstract: This paper proposes a network code construction method which has low decoding delay. The key idea of this method is that, among all the disjoint paths from the source to a destination, one path is selected as “convenient path”, on which packets are delivered directly without coding. Allocating maximum number of convenient paths for all the destinations in a multicast network can be modeled as the problem of maximal complete sub-graph. Both theoretical analysis and numerical results from example networks show that the proposed network code construction method can reduce the decoding delay effectively without any reduction in throughput benefit of network coding. The method introduced by this paper can serve as an efficient way for transmitting data of delay sensitive applications such as media streaming.

     

/

返回文章
返回