Pragmatic Analysis of Diffusion in Computer Networks
Social networks is a content transportation medium, that has become more popular than the conventional fax and email system in the current world even shadowing the online messengers. With social websites like Facebook, Twitter etc., one can spread any type of information whether it be text, picture or video. Information of even an enormous size can spread like an epidemic on social networks. The researchers are intrigued by the interesting behaviors such networks are exhibiting and are delving in these types of networks. Content diffusion in computer networks is among trusted networks, just like information sharing on social websites where you are only allowed to share / view / comment only among trusted members. Among trusted networks one can spread information to achieve a diversity of tasks like marketing products or diffusion of job vacancies through pop-ups. On the other hand there can also be a malware that can utilize the same network and infect the trusted network. With the help of study of diffusion one can predict how to benefit from spreading of information in a network. Our interest in this paper is to study influence which is amongst the origins of diffusion for information spreading in a network. We have analyzed three real computer networks and compared them with artificially generated complex networks of random and scale free equivalent to each of them. Our experiments are based simultaneously on the concepts of Linear Threshold and Independent Cascade Model. We have used five different methods / metrics of selecting initial seed nodes and then calculated influence for each of them. Our experiments also include comparing of the metrics on each of the real networks and its corresponding random and scale free networks. Our experiments not only show that small amount of initial seed can infect maximum network but also that some metrics have same effect on equivalent networks while some donot.