TCP NRT: a new TCP algorithm for differentiating non-congestion retransmission timeouts over multihop wireless networks

Abstract

In multihop wireless networks, reliable data transfer is one of the most difficult tasks. When transmission control protocol (TCP) operates in multihop wireless networks, the performance of TCP reduces drastically. TCP retransmission timeouts (RTOs) related to non-congestion events such as spurious and random packet losses have been reported as one of the main problems in the performance degradation of TCP in multihop wireless networks. The RTOs triggered by random packet losses due to transmission errors lead to unnecessary reduction of TCP congestion window size, and the spurious RTOs due to sudden delay of packets on the network paths often cause unnecessary retransmissions as well as reduction of congestion window size. Existing solutions for detecting noncongestion RTOs have no mechanism to differentiate the spurious RTOs from RTOs caused by random packet loss. In this paper, we introduce an efficient algorithm called non-congestion retransmission timeouts (TCP NRT) which is capable of recovering packets after RTOs by reducing unnecessary retransmissions and needless reduction of congestion window size in order to improve the performance of TCP in multihop wireless networks. TCP NRT consists of three key components: NRT-detection, NRT-differentiation, and NRT-reaction. We implemented the algorithm in Qualnet network simulator and compared its performance to existing TCP versions. Results from the experiments show that our algorithm achieves significant performance improvement in terms of throughput and accuracy. Also, the results showed that our algorithm, TCP NRT, maintains a fair and friendly behavior compared to the most widely deployed TCP, NewReno.

A survey on congestion control mechanisms in packet switch networks

“…An efficient algorithm called non-congestion retransmission timeouts (TCP NRT) [20] was proposed for the solution of this problem. NRT is capable of detecting spurious RTOs and needless reducing of window.…”

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.…”

“…In such scenario, the transitory delay variations are mostly due to link retransmissions as the majority of users are pedestrians (i.e. Many studies [3,13,26] treat for sacrifice in TCP performance on account of undue reduction in cwnd, but they do not address the issues related to the reduced protocol responsiveness on account of link retransmissions; although very common in recent wireless architectures [20]. Hence, it becomes more important to address the issue of delay variability due to link retransmissions in WLAN.…”

“…The aforementioned results into additional performance degradation because of the reduced protocol responsiveness to wireless packet losses [25]. Many studies [3,13,26] treat for sacrifice in TCP performance on account of undue reduction in cwnd, but they do not address the issues related to the reduced protocol responsiveness on account of link retransmissions; although very common in recent wireless architectures [20]. In fact, all major TCP proposals implemented on the Internet fail in achieving the goal of high bandwidth utilization with fast convergence (required especially with frequent wireless losses) when the end-to-end network path composes a wireless link [27].…”

Adaptive TCP: A Sender Side Mechanism with Dynamic Adjustment of Congestion Control Parameters for Performance Improvement in WLAN

“…With FRR, conventional TCP sets (a) Loss Window and (b) ssthresh (that describes estimation of the bandwidth availability) to an arbitrary value (which is half of the current size of cwnd) [1]. When the packet loss is due to the non-congestion reason, this reduction using a stagnant rule is unfair [5]. TCP schemes with Loss Differentiation Algorithms (LDA) [6] [7] advocate for retaining of the aforesaid parameters after loss recovery.…”