期刊文献+

一种动态门限变速增加的TCP慢启动算法 被引量:5

A dynamic threshold and varying rate TCP slow-start algorithm
下载PDF
导出
摘要 提出了一种动态门限变速增加的慢启动算法,将Vegas和TCP Westwood的关键技术应用到了TCP首次慢启动过程,并提出了动态设定慢启动门限和分段的窗口增加粒度的算法.动态门限变速增加通过对链路可用带宽的估计来设置慢启动门限,使其不会出现多包丢失,从而过早地进入拥塞避免阶段,所以动态门限变速增加能够迅速到达稳态.利用Vegas瓶颈链路队列长度的值,对窗口增加粒度进行分段控制,能够根据瓶颈缓冲区的队列情况来调整窗口增加粒度.该算法能够在多种网络环境下,避免慢启动阶段的多包丢失,迅速到达稳态,提高了慢启动阶段的吞吐量.同时该算法对链路瓶颈缓冲有较好的鲁棒性,对WEB业务适应性强,对现有网络协议保持友好性,只需在发送端修改. DSVR, a dynamic threshold and verying rate TCP Slow-Start algorithm, is proposed. The key technologies of Vegas and TCP Westwood are applied to the first slow start process in DSVR, which dynamically configures the TCP Slow-Start threshold and adaptively adjusts the increasing rate of TCP transmitting windows. DSVR can reach the steady state rapidly because its configuration of the slow-start threshold is based on the bandwidth estimation, and thus the lost packages will be limited and the entrance of congestion avoidance stage will not be too early. Simulation result proves that this alogrithm can avoid the loss of data packets, reach the steady state quickly, and improve TCP throughput in the complex network. This algorithm is robust to the bottleneck buffer, adaptable to WEB service, and compatible with the present TCP protocol It is simple and practical in that it only modifies the sender of TCP.
出处 《西安电子科技大学学报》 EI CAS CSCD 北大核心 2006年第4期640-645,共6页 Journal of Xidian University
基金 国家自然科学基金资助项目(60573161) 陕西省"十五"科技攻关项目(2000K08-12)
关键词 拥塞控制 慢启动 动态门限变速增加 congestion control slow-start DSVR
  • 相关文献

参考文献9

  • 1Barakat C, Altman E. Analysis of the Phenomenon of Several Slow Start Phases in TCP[A]. International Conference on Measurement and Modeling of Computer[C]. California: ACM Press, 2000. 102-103.
  • 2Allman M. Improving TCP Performance Over Satellite Channels[D]. Ohio: Ohio University, 1997.
  • 3李云,陈前斌,隆克平,吴诗其.一种基于链路带宽估计的TCP慢启动算法[J].计算机学报,2003,26(6):693-700. 被引量:7
  • 4Hoe J C. Improving the Start-up Behavior of a Congestion Control Scheme for TCP[A]. The ACM SIGCOMM[C].Stanford: ACM Press, 1996. 270-280.
  • 5Lai Ko Bake M. Nettimer: a Tool for Measuring Bottleneck Link Bandwidth[A]. The USENIX Symposium on Internet Technologies and Systems[C]. San Francisco: USITS Press, 2001. 123-134.
  • 6Barakat C,Ahman E. Performance of Short TCP Transfers[A]. Proceedings of NETWORKING 2000[C]. Paris:Springer, 2000. 567-579.
  • 7Casetti M S, Gerla C, Sanadidi M, et al. TCP Westwood: End-to-End Bandwidth Estimation for Efficient Transport Over Wired and Wireless Networks[A]. The MOBICOM[C]. Rome: ACM Press, 2001. 287-297.
  • 8Lakshman T V, Madhow U. The Performance of TCP/IP for Networks with High Bandwidth-delay Products and Random Loss[J]. IEEE/ACM Trails on Networking. 1997. 5(3):336-350.
  • 9NS2(Newtork Smulator version 2), Lawrence Berkeley Laboratory[CP/OL]. http://www.isi. edu. nsnarn/ns/, 1997-07-23.

二级参考文献8

  • 1Allman et al. TCP congestion control. Internet Eningeering Task Force,RFC 2581, 1999.
  • 2Mahdavi J, Mathis M, Podolsky M. An extension to the selective acknowledgement (SACK) option for TCP. Internet Eningeering Task Force, RFC 2883, 2000.
  • 3Mathis M, Mahadavi J. Forward acknowledgment: Refining TCP congestion control. In: Proceedings of the ACM SIG-COMM'96, Stanford, 1996. 281-291.
  • 4Floyd S, Henderson T. The newreno modification to TCP's fast recovery algorithm. Internet Eningeering Task Force,RFC.2582, 1999.
  • 5Hoe J C. Improving the start-up behavior of a congest- ion control scheme for TCP. In: Proceedings of the ACM SIGCOMM'96, Stanford, 1996. 270-280.
  • 6Ewerlid A. Reliable communication over wireless links. In:Proceedings of the Nordic Radio Symposium, Saltsjobaden,Sweden, 2001.40-45.
  • 7Zhang H. WF^2Q: Worst-case fair weighted fair queue. In:Proceedings of IEEE INFOCOM' 96, San Francisco, California, 1996. 120- 128.
  • 8van Jacobson, Karels M J. Congestion avoidance and control.In: Proceedings of the SIGCOMM'88 Symposium, 1988, 18:314-329.

共引文献6

同被引文献37

引证文献5

二级引证文献11

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部