The Transmission Control Protocol (TCP) is the dominating end-to-end protocol on the internet today but still it faces congestion problems in some cases. To overcome congestion problems, several congestion control and avoiding mechanisms namely: Tahoe, Reno, Vegas, and Sack etc. all with different features and advantages but with maximal throughput as main objective, which are termed as the clones of TCP, have been incorporated into TCP/IP protocol for handling congestion efficiently in different network scenarios. However, one clone cannot be suitable for each case. So this paper has investigated the characteristics of the mentioned clones and calculated throughputs of them in simulated environment varying various performances metrics such as delay, buffer size, error rate, number of traffic and bandwidth for finding which one is the best for what scenario. The performance of these clones for varying network conditions and settings can effectively be evaluated using NS-2. In this work, by doing simulation in NS-2 environment the throughputs of some exiting TCP implementations are calculated considering various metrics and then the calculated throughputs are compared among one another. These comparisons show that which one is suitable in which cases.