TCP’s dynamic adjustment of transmission rate to packet losses in wirelessnetworks

Abstract

Based on the assumption of transmission control protocol (TCP) that packets are lost due to congestion, TCP's congestion control algorithms such as fast retransmit/recovery (FRR) and retransmission timeouts (RTO) unconditionally reduce the transmission rate for every packet loss. When TCP operates in wireless networks, however, FRRs/RTOs are often triggered regardless of congestion due to sudden delay and wireless transmission errors. The congestion irrelative FRRs/RTOs incur TCP's misbehavior such as blindly halving the transmission rate, unnecessarily retransmitting the outstanding packets which may be in the bottleneck queue. Although many previous studies have been proposed to detect the congestion irrelative FRRs/RTOs, they paid little attention on effectively adjusting the transmission rate for the detected congestion irrelative FRRs/RTOs. In this article, we propose an enhanced TCP to dynamically adjust its transmission rate according to network conditions. Our scheme adjusts the transmission rate in proportion to the available bandwidth in order to quickly utilize the available bandwidth, and also re-adjusts it in inverse proportion to the loss rate in order to avoid burst losses and long go-back-N retransmissions. By doing so, our scheme has significant effects to avoid the performance degradation caused by the congestion irrelative FRRs/RTOs. Throughout the extensive experiments, we evaluate our scheme and compare it with previous works in terms of goodput, fairness, and friendless under various network topologies. The results show that our scheme significantly outperforms previous studies while it maintains the fair and friendly behavior to other TCP connections.

Refining TCP's RTT dependent mechanism by utilizing link retransmission delay measurement in Wireless LAN

“…The majority of recent wireless TCP proposals (e.g., TCP-WELCOME [31], TCP-NRT [26], Dynamic TCP [10], and those proposed in [3,5]) target for improvement in TCP performance by providing treatment for inappropriate reduction in the sending rate with wireless loss recovery. Unfortunately, like the major TCP schemes deployed in the current Internet (e.g., TCP Westwood [36], TCP Newreno [6], TCP CUBIC [27]), they do not address the impact of link ARQs on their performance.…”

“…TCP achieves reliable data delivery by retransmitting the lost packet. In this way, both of the aforementioned incidences,that is, packet loss detection using timeout and the growth of cwnd on arrival of each TCP-ack, are greatly dependent on RTT, estimated by the TCP sender on end-to-end basis [6,10]. This procedure is also called the additive increase and multiplicative decrease (AIMD) algorithm [6].…”

“…This procedure is also called the additive increase and multiplicative decrease (AIMD) algorithm [6]. In this way, both of the aforementioned incidences,that is, packet loss detection using timeout and the growth of cwnd on arrival of each TCP-ack, are greatly dependent on RTT, estimated by the TCP sender on end-to-end basis [6,10].…”

“…For that, TCP regulates its sending rate using one of its state variables congestion window (cwnd), which revises either on detection of a packet loss or on arrival of a positive acknowledgment (TCP-ack) from the receiver [6]. Both of the aforementioned incidences are greatly dependent on round trip time (RTT), estimated by the TCP sender on end-to-end basis [10,11]. Because the TCP throughput is limited by the augmentation in cwnd per RTT [12], the efficiency claimed by TCP is essentially depending upon its accuracy for RTT estimation, in addition to the size of cwnd [13].…”

On the performance of asynchronous TCP connections to Wireless Sensor Network over WiFi

“…However, when r=4, caching worsens the average response time of the gateway by a factor of 2. We can argue that, in bursty traffic conditions 802.11 cannot handle the multiclient request traffic because of high collision rates in the network and TCP resends the segments for false timeouts which is a known problem for TCP over unreliable wireless links [12].…”