ABSTRACT


This thesis analyzed the performance of various algorithms of TCP and queue management techniques for congestion control. Queue management is a process which controls the queue from sender to the receiver manage the size of buffer. For example, when too much data arrived from sender side, the queue is full. It requires some strategy to either delete the data from queue or rearrange it. If length of buffer is full due to the network congestion. So, the actual goal of the queue management is to improve the performance and fairness of the network.
Queuing in a network directly relates to the bottleneck. So, for that various queuing techniques used to manage the Buffers length on the network and also queue management algorithms used to manage queues size, released more packets, resolve network congestion and improve network performance as well. When channels with limited buffer space shared multiple data sources then some of arrival packets remain lost, and the buffer becomes full.
Queue management analyzed some techniques using TCP such as Drop tail, RED and REM for better performance, after the experimental analysis of these three algorithms of TCP, it is observed that RED Queue management technique one of the best technique to maintain the efficiency of buffer size and packets loss ratio is very low as compared other techniques such as Drop tail and REM. The results indicate that RED technique of TCP has superiority with higher throughput, less packets delay and less packet loss compared to techniques of TCP such as DropTail and REM with different numbers of nodes probability in wired network.