摘要： 多路径传输控制协议通过聚合多路径带宽提高资源利用率及网络吞吐量。在无线异构网络环境中，由于网络性能不稳定、传输路径性能差异等因素容易导致数据包乱序及缓冲区阻塞，对网络多路传输性能造成负面影响。为在有限的接收缓冲区内对数据包实施合理调度，基于灰色预测模型GM （1，N）与马尔科夫优化的前向传输时间（FTT）预测模型，提出一种自适应多路传输数据调度算法GMM-S。通过对未来时刻子流FTT的准确预测实现子流传输性能的有效评估，并以此作为数据分发依据进行传输数据的多子流动态调整。仿真实验结果表明，与RR和LowRTT算法相比，该算法可有效解决接收端数据包乱序问题，同时提升网络吞吐量。

关键词: 多路径传输控制协议, 灰色预测模型, 马尔科夫算法, 数据调度, 丢包

Abstract: Multi-Path Control Transmission Protocol (MPTCP) improves resource utilization and network throughput by aggregating multi-path bandwidth.However,in the wireless heterogeneous network environment,out-of-order packets and buffer block are serious due to many factors,such as unstable network performance and differences between the transmission paths,which reduces the performance of multi-path transmission.In order to schedule packet reasonably in the limited receiving buffer,this paper propose an adaptive multi-path data scheduling algorithm GMM-S based on the Grey prediction Model GM(1,N) and Markov optimized Forward Transmission Time (FTT) prediction mode.The algorithm evaluates the transmission performance of sub-flows through the accurate prediction of FTT of sub-flows in a future moment,and takes it as a basis for data distribution to realize dynamic sub-flows adjustment of transmitted data.The simulation results show that compared with the RR algorithm and LowRTT algorithm,the proposed algorithm can effectively solve the packet disorder problem at the receiving end,and improve the network throughput.

Key words: Multi-Path Transmission Control Protocol(MPTCP), Grey prediction Model(GM), Markov algorithm, data scheduling, packet loss

中图分类号: 

• TP393