Congestion Control based analysis of High Delay Tolerant network Transport Protocol

Abstract

The current congestion control mechanism used in TCP has difficulty reaching full utilization on high speed links, particularly on wide-area connections. One of reason is that regular TCP lays constraint on the congestion windows in order to keep packet drop rate in control, and the packet drop rate needed to fill a Gigabit pipe using the present TCP protocol is below the currently achievable fiber optic error rates. High Speed TCP and Cubic proposed as modifications of TCP’s congestion control mechanism to allow it to achieve reasonable performance in high-speed wide are links. Compound TCP which was proposed as modifications of TCP’s Congestion control mechanism, and was implemented on windows operating systems. In this project, simulations results showing the performance of High Speed TCP & Cubic and their impact of its use on the current implementations of TCP are presented. Network conditions including different degrees of congestion, different levels of loss rate and two distinct router queue management policies are simulated. We are also demonstrating that link utilization and performance of protocols clearly affected by the variability of RTT of each flows and fairness of other flows on congestion control are different. The performance and Fairness of High Speed TCP and Cubic are compared to the existing TCP. The project also involves a Linux kernel level implementation of Compound TCP along with already provided CUBIC and High Speed TCP, to provide a concrete base to the observations that were made by simulation results.