Improved transmission control protocol congestion control technique for high bandwidth long distance networks

Abstract

Transmission Control Protocol (TCP) is responsible for reliable communication of data in high bandwidth long distance networks. TCP is reliable because of its congestion control technique. Many TCP congestion control techniques for different operating systems have been developed previously. TCP Compound and TCP CUBIC are current congestion control techniques being used in Microsoft Windows and Linux operating systems respectively. TCP Reno is Standard TCP congestion control technique. TCP CUBIC does not perform well in high bandwidth long distance networks due to its exponential growth and less reduction in congestion window size. This leads to burst packet losses, unfair allocation of unused link bandwidth, long convergence time, and poor TCP friendliness among competing flows. The aim of this research work is to develop an improved congestion control technique based on TCP CUBIC for high bandwidth long distance networks. This improved technique is based on three components which are Congestion Control Technique for Slow Start (CCT-SS), Congestion Control Technique for Loss Occurrence (CCT-LO), and Enhanced Response Function of TCP CUBIC (ERFC). CCT-SS is proposed which increases the lower boundary limit of congestion window, which in turn, decreases the packet loss rate. CCT-LO is proposed which introduces a new congestion window reduction parameter in order to achieve fairer and quicker allocation of link bandwidth among the competing flows. ERFC is proposed which reduces the average congestion window size of TCP CUBIC in order to improve the TCP friendliness. As a conjunctive result of this research work, an improved congestion control technique is developed by combining the CCT-SS, CCT-LO and ERFC components. Network Simulator 2 is used to evaluate the performance of the proposed congestion control technique and to compare it with the current and other congestion control techniques. Results show that the performance of the proposed congestion control technique outperforms by 8.4% as compared to current congestion control technique.