Sync-TCP: A new approach to high speed congestion control

Wu, Xiuchao; Chan, Mun Choon; Ananda, A.L.; Ganjihal, Chetan

doi:10.1109/icnp.2009.5339684

Cited by 15 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The compound TCP (CTCP) congestion control algorithm [28] combines loss-based and delay-based congestion control methods, allowing the algorithm to quickly increase the sending rate while obtaining high bandwidth scalability and improved TCP fairness. CTCP is designed and usually used in high bandwidth environment, and it is disabled by default [29]. eBCC adjusts the algorithm transition state and congestion control window using loss and queuing status as decision factors.…”

Section: Enhanced Bbr Congestion Control Algorithmmentioning

confidence: 99%

A QoS-Based Fairness-Aware BBR Congestion Control Algorithm Using QUIC

Han

Zuo

Yuan

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Congestion control is a fundamental technology to balance the traffic load and the network. The Internet Engineering Task Force (IETF) Quick UDP Internet Connection (QUIC) protocol has flexible congestion control and at the same time possesses the advantages of high efficiency, low latency, and easy deployment at the application layer. Bottleneck bandwidth and round-trip propagation time (BBR) is an optional congestion control algorithm adopted by QUIC. BBR can significantly increase throughput and reduce latency, in particular over long-haul paths. However, BBR results in high packet loss in low bandwidth and low fairness in multi-stream scenarios. In this article, we propose the enhanced BBR congestion control (eBCC) algorithm, which improves the BBR algorithm in two aspects: (1) 10.87% higher throughput and 74.58% lower packet loss rate in the low-bandwidth scenario and (2) 8.39% higher fairness in the multi-stream scenario. This improvement makes eBCC very suitable for IoT communications to provide better QoS services.

show abstract

Section: Enhanced Bbr Congestion Control Algorithmmentioning

confidence: 99%

A QoS-Based Fairness-Aware BBR Congestion Control Algorithm Using QUIC

Han

Zuo

Yuan

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

show abstract

“…More precisely, as for the size of the contents, we think that small packet transfers as the ones performed in actual IP networks should be avoided in CSN, because of their granularity and the waste of bandwidth they implies. On the contrary, due to its content abstraction, the Content-based Switching Network provides a way to avoid complicated transport layer functions, like TCP [6,7] if they are not essential. However, CSN could be used to transfer packet bursts [8] when small packets cannot be avoided.…”

Section: Common Technologiesmentioning

confidence: 99%

Deployment of the Content-Based Switching Network

Michard¹,

Tanigawa²,

Tode³

2014

View full text Add to dashboard Cite

In this paper, we propose a hybrid network architecture, called Content-based Switching Network (CSN), and its signaling scheme, which addresses the issues inherent to conventional hybrid networks which implement a horizontal separation over the entire network (from edge to edge). We will show how CSN nodes can flexibly choose their switching paradigm (store-and-forward, optical bypass, electrical bypass) during a path establishment. Contents being transferred in one piece from end-to-end, the concept of packet can be eluded in our network, and, in particular, the user is able to avoid complicated transport layer functions, like TCP, if they are not essential. In CSN, very large contents have a special status, since they cannot be store-and-forwarded. We will show how the resource management has been designed in order to deal with such contents. A section is dedicated to deployment and feasibility issues. Simulation results will show that CSN can successfully transfer contents at 1 Gbps and 10 Gbps, the maximum speed being limited by the state-of-the-art device technologies when buffering is required (memory speed), while no major limit is observed in the case of all-optical transfers other than the optical fiber speed. Other results concern the deployment of CSN from an unclean slate approach. They will show how beneficial can be the deployment of CSN from an Optical Circuit Switching network.

show abstract

“…Protocols based on the AIMD paradigm include the transmission control protocol TCP (Jacobson 1988). Recently, in the context of high-speed communication networks, it was realised that standard AIMD is not always suitable for deployment in some networks (Floyd 2003;Xu, Harfoush, and Rhee 2004;Kun, Jingmin, Qian, and Sridharan 2005;Li, Leith, and Shorten 2007;Konda and Kaur 2009;Molnar, Sonkoly, and Trinh 2009;Stewart, Armitage, and Huebner 2009;Wu, Chan, Ananda, and Ganjihal 2009;Eckart, He, Wu, and Xie 2010;Masaki, Nishantha, and Hayashida 2010;Trinh, Sonkoly, and Molnar 2010;Weigle, Cheng, Kaur, and Kulkarni 2010;Zhang, Piunovskiy, Ayesta, and Avrachenkov 2010). This observation has motivated a host of (similar) modifications to the basic AIMD algorithm.…”

Section: Introductory Remarksmentioning

confidence: 95%