Premium
The heterogeneity of inter‐domain Internet application flows: entropic analysis and flow graph modelling
Author(s) -
Yu Ke,
Li Frank Y.,
Wu Xiaofei,
Di Jiaxi
Publication year - 2015
Publication title -
transactions on emerging telecommunications technologies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.366
H-Index - 47
ISSN - 2161-3915
DOI - 10.1002/ett.2706
Subject(s) - the internet , computer science , popularity , entropy (arrow of time) , internet traffic , flow network , theoretical computer science , data mining , world wide web , mathematics , mathematical optimization , physics , psychology , social psychology , quantum mechanics
The growing popularity of the Internet has triggered the proliferation of various applications, which possess diverse communication patterns and user behaviour. In this paper, the heterogeneous characteristics of Internet applications and traffic are investigated from a complex network and entropic perspective. On the basis of real‐life flow data collected from a public network provided by an Internet service provider, flow graphs are constructed for five types of applications as follows: Web, P2P Download, P2P Stream, Video Stream and Instant Messaging. Three types of entropy measures are introduced to the flow graphs, and the heterogeneity of applications within a 24‐h period is analysed and compared. By using the strict timing information of the flow records, the growth process and the entropy fluctuation of flow graphs are observed. On the basis of the analysis of the flow records, a complex network model is developed for the flow graph growth. Simulation results show that the novel model matches well the flow graphs in terms of both degree distribution and entropy dynamics. Our analysis results and the proposed model shed light on traffic identification and modelling for inter‐domain traffic in today's Internet. Copyright © 2013 John Wiley & Sons, Ltd.