Behavior Analysis of Internet Traffic Using Graph and Similarity Matrix

As system activity of utilizations, for example, video spilling and distributed applications proceeds to grow it is difficult task to understand behavior patterns of end host and other network application. Keeping in mind the end goal to comprehend conduct closeness of Internet end has in the same system prefixes bipartite diagrams are utilized to model system activity, and after that one-mode projection diagrams are built for catching social-conduct closeness of end has. At that point by applying a basic and effective algorithm and similarity matrices and cluster coefficient of one mode projection graph, inherent clustered groups of internet application are discovered. Through a vector diagram including coefficient bunching that catch social practices of end hosts, the clusters are found that not just display comparable social conduct of end hosts, additionally have comparable attributes in the accumulated activity.