Experimental Analysis of Bittorrent Traffic based on Heavy-Tailed Probability Distributions
Aishwarya Gaikwad, Rupesh Chandrakant Jaiswal · International Journal of Computer Applications · 2016
Complexity involved in measuring and analyzing the BitTorrent traffic has led to various studies in this direction.Challenges involved are related to storage, data retrieval, location of content, topological features, privacy, copyright issues along with analysis of data and modeling of traffic.Internet Traffic, earlier thought to be of Poisson, is bursty in nature.In this paper, BitTorrent traffic for applications like video is observed by means of distributions, that best represent their nature.Inter-arrival times and lengths of packets are the parameters used to plot cdf so that the best distribution is determined.This analysis can further help in exploring various fractal characteristics [1], as the alpha value obtained is crucial in determining the heavy tailed-ness, responsible for impacting network performance and creating obstacles in maintaining the desired QoS.