Markov Model Based Congestion Control for TCP

  • Authors:
  • Shan Suthaharan

  • Affiliations:
  • -

  • Venue:
  • ANSS '04 Proceedings of the 37th annual symposium on Simulation
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Random Early Detection (RED) scheme forcongestion control in TCP is well known over a decade.Due to a number of control parameters in RED, it cannotmake acceptable packet-dropping decision, especially,under heavy network load and high delay to provide highthroughput and low packet loss rate. We propose asolution to this problem using Markov chain baseddecision rule. We modeled the oscillation of the averagequeue size as a homogeneous Markov chain with threestates and simulated the system using the networksimulator software NS-2. The simulations show that theproposed scheme successfully estimates the maximumpacket dropping probability for Random Early Detection.It detects the congestion very early and adjusts thepacket-dropping probability so that RED can make wisepacket-dropping decisions. Simulation results show thatthe proposed scheme provides improved connectionthroughput and reduced packet loss rate.