车载异构网络中基于PSRBP的多径TCP路径调度优化

Multipath TCP Scheduling Optimization Based on PSRBP in Vehicular Heterogeneous Network

车载异构网络使用多路径传输控制协议(multipath transmission control protocol,MPTCP)进行传输时,由于路径差异性会引起接收缓冲区阻塞问题,导致MPTCP传输性能不佳.针对MPTCP接收缓冲区阻塞问题,提出一种基于路径选择和接收缓冲区预测(path selection and receive buffer prediction, PSRBP)的MPTCP路径调度优化方法.该方法引入吞吐量预测模型,同时引入各条路径和最大吞吐量路径的相似性参数与既定阈值对比分析,从而选择出可以使用MPTCP传输的路径.所提算法通过判断无序包的数量来预测缓冲区阻塞问题,并停止不良路径上的数据传输以提高MPTCP的性能.仿真结果表明,该算法能够提高系统整体的吞吐量,并使网络利用率有所改善.

Nowadays, as a variety of network interfaces are deployed in vehicles for a heterogeneous vehicular network, the diversity of interfaces leads to a noticeable difference in characteristics between transmission paths. When the vehicular heterogeneous network uses the multipath transmission control protocol(MPTCP) for transmission, it may causes blocking problem in the receive buffer due to the differences between paths, resulting in poor MPTCP transmission performance. For MPTCP receives buffer blocking problem, this paper proposes an MPTCP path scheduling optimization method based on path selection and receive buffer prediction(PSRBP). This method introduces a throughput prediction model,then introduces the similarity parameter between each path and the maximum throughput path. The similarity parameter is compared with a predetermined threshold to select a path that can be transmitted using MPTCP. After that, the proposed algorithm predicts the buffer congestion problem by calculating the number of out-of-ordered packets, then stops data transmission through the bad paths to improve the performance of MPTCP.Simulation results show that the algorithm can improve the overall system throughput and improve the network utilization.