Congestion control in TCP/IP networks: a combined ECN and BECN approach

  • Authors:
  • Frank Akujobi;Nabil Seddigh;Biswajit Nandy;Rupinder Makkar;Ioannis Lambadaris

  • Affiliations:
  • Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada;Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada;Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada;Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada;Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada

  • Venue:
  • MILCOM'03 Proceedings of the 2003 IEEE conference on Military communications - Volume I
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper a novel algorithm is proposed which combines the merits of Explicit Congestion Notification (ECN) and Backward Explicit Congestion Notification (BECN) mechanisms for congestion control in TCP/IP networks. A comparative performance evaluation of the combined ECN and BECN mechanism is carried out using both long-lived FTP flows and short-lived web traffic. The simulation results show that the combined ECN+BECN mechanism benefits from BECN's early notification under heavy congestion and ECN's reliable delivery of congestion notification. It is observed that the ECN+BECN mechanism significantly reduces queue fluctuations due to early congestion indication compared to ECN. The loss of BECN Internet Control Message Protocol (ICMP) Source Quenches on the reverse path does not adversely impact performance due to this combined approach. It is also shown that the ECN+BECN scheme can significantly reduce the ICMP Source Quench reverse traffic in a network compared to a BECN only network. Experiments with webtraffic workloads show measurable improvement in both average object transfer delay and fairness over ECN.