摘要: 为解决传统传输控制协议(TCP)在异构网络中效率较低的问题,给出网络状态认知的定义,建立以瓶颈链路队列长度为目标量的网络认知模型,采用TCP-L传输控制协议进行仿真。实验结果证明,该模型能准确估计瓶颈链路队列长度的动态变化趋势。TCP-L协议能在不同负荷的背景流条件下稳定发送窗口、往返传输时延和队列长度,减少拥塞发生,提高吞吐量和传输效率。
关键词:
瓶颈链路队列长度,
拥塞控制,
网络认知,
异构网络,
吞吐量,
传输控制协议
Abstract: In view of the traditional Transmission Control Protocol(TCP) control protocol efficiency in heterogeneous network, this paper gives the definition of the cognition of the network status, proposes and constructs a network cognitive model which makes the length of bottleneck link queue as the target parameter. This model uses TCP-L transmission control protocol to finish simulation experiments. Experimental result shows that this model can more accurately estimate the dynamic change trend of bottleneck link queue’s length, and its estimate can be more accurately located in the gathering area of true values. Under different background flow conditions, the TCP-L protocol can stabilize sending window, round trip transmission delay and queue length, reduce congestion and improve throughput and transmission efficiency.
Key words:
bottleneck link queue length,
congestion control,
network cognition,
heterogeneous network,
throughput,
Transmission Control Protocol(TCP)
中图分类号:
赵淼, 王辉. 瓶颈链路队列长度的网络认知模型[J]. 计算机工程, 2012, 38(19): 60-63.
DIAO Miao, WANG Hui. Network Cognitive Model of Bottleneck Link Queue Length[J]. Computer Engineering, 2012, 38(19): 60-63.