Entropy analysis for inter-domain Internet application flows

Internet applications provide various communication patterns to meet a wide range of user needs. Understanding the characteristics of Internet applications and traffic is essential for network engineering and supervision. In this paper, based on detailed flow data collected from public network of Internet service provider, the statistical characteristics of applications are investigated from complex network perspective. Flow graphs are constructed for five types of applications: web, P2P download, P2P stream, video stream and instant messaging. By introducing entropy measures of complex network, heterogeneity of these flow graphs are analyzed and compared. The dynamics of entropy within a 24-hour period is also studied for traffic prediction. Analysis results show different characteristics of Internet applications and traffic, which implies the heterogeneous behavior of users, and will be useful for traffic modeling and identification.