Congestion Control Technology Based On Cross-layer ECN In Wireless Networks

With the rapid development of traditional Internet, wireless networks enter people’s daily lives with their unique flexibility, convenience and advantages of high efficiency more and more frequently. However, channel interference, low bandwidth, high error ratio and other features of wireless networks make TCP protocol meet a series of questions while used in such networks:First, TCP is in the absence of effective loss differentiation mechanism;Second, TCP can not fully use existing wireless bandwidth;In addition, energy efficiency is not high while TCP is used in wireless network. Therefore, the TCP applied in traditional wired networks can not directly be transferred to the wireless network.Firstly, this paper introduces and classifies the existing congestion detection methods, and then the main of them are analyzed and compared. In this basis, it discusses the technology that cross-layer ECN technology could transmit information to the sender timely and effectively, which is the development trend of the design for wireless congestion control algorithm.In the following, a cross-layer ECN algorithm using the number of RTS retransmission as congestion detection method is presented to solve the problem that traditional ECN using queue length as congestion measurement can not effectively reflect the state of wireless channel conflict. Experimental results show that the scheme could be integrated with the existing ECN mechanism, complement it, and promote TCP throughput and fairness between flows in wire/wireless hybrid networks.A congestion control algorithm WM-TCP (Wireless MAC TCP) with packet loss information, RTT and explicit congestion notification is proposed to distinct wireless packet loss and congestion packet loss.It judges congestion with three kinds of information, distinguishes packet loss caused by different problems effectively and improves network utilization and throughput. Meanwhile, the combination could avoid wrong detection of single mode in some environment and reduce the reliability influence of the algorithm on the single information error. Finally, this paper summarizes the research work and describes the future interesting directions of wireless network congestion control technology.