嵌入式网络通信中RED算法的优化

The Optimization of RED Algorithm in the Embedded Network Communication

在嵌入式网络通信中,主要采用RED算法来解决网络拥塞,由于RED算法中丢包率与平均队列长度成线性关系,导致网络在拥塞并不严重的时候丢包率较大,在拥塞比较严重的时候丢包率较小,拥塞控制能力较低.经研究,发现EXPRED(Exponent Random Early Detection)算法能解决这个问题,避免了网络的全局同步.通过NS-2.30仿真证实:EXPRED算法可以提高网络吞吐量,减少延时抖动,使网络比较稳定.

The RED algorithm used to solve network congestion in the embedded network communication.But drop packet ratio is proportional to the average queue length,as a result,drop packet ratio is high when the network congestion is not serious and drop packet ratio is low when the network congestion is serious,so the congestion ability is not effective.To study to find that the exponential function can solve that problem,in the EXPRED（Exponent Random Early Detection）,the drop packet ratio is smaller around the Minimum threshold,and it is more around the Maximum Threshold,to avoid the global synchronization of network.with the NS-2.30 simulation results show that： EXPRED algorithm can improve network throughput,reduce delay jitter,making the network more stable.