TCP Congestion Window Analysis of Twitter with Exponential Model

摘要

In this work, an analysis of Transmission Control Protocol (TCP) Congestion Window with traffic model of twitter using exponential distribution is presented. TCP is a popular reliable protocol used by many internet service. As social medias emerge and introduce new characteristic of traffic, it is interesting to evaluate the performance of TCP in the latest social media environment scenario. Twitter as top social media shows its influence by taking part of many political event in the globe. An exponential distribution traffic model of Twitter traffic is proposed and an experiment on evaluation of TCP congestion window is presented in this paper. A congestion window represents efficiency of TCP to adapt network condition since the metric shows the number of data transmitted in a single transmission. In stable condition, shared bandwidth carry out higher congestion window which makes transmission more efficient. Combination of high traffic flow and high bandwidth comes with lower congestion window variation with 0.58 where low traffic flow obtains higher variance with 1.15.