ABSTRACT

Transmission Control Protocol (TCP) is widely used over Internet because it has security features and consistently while data transfer from source to destination. So most of the Internet communications are based on TCP. TCP implements a reliable source of information transmission over the network with accuracy and also provide the congestion control mechanism. When congestion occurs, the nodes or links carries so much data that they decline the allocation of bandwidth of the network. It is due to many reasons such as packet overloading, queuing interruption and blocking of the new connections due to the receiver buffer overflow. Further, the other minor errors in the data transmission causes the increased latency also due to the network
congestion.

This work explains the famous TCP congestion control variants such as TCP Tahoe, New Reno, Sack and Vegas. The various experiments are carried on multiple scenarios of network such as with 4, 11 and I 0 nodes. In these three scenarios the nodes are arranged in different topologies. In the experiments the TCP Vegas showed a good performance than other TCP variants in terms of throughput and end to end delay. When the bandwidth is large the most the variants reach to approximately the similar throughput, however with the small bandwidth the Vegas outperforms the other variants.