Measurement and analysis of content diffusion characteristics in opportunity environments with Spark

• Published in: 信息与电子工程前沿（英文版） Vol. 20; no. 10; pp. 1404 - 1414
• Main Authors: Xiao-hong ZHANG, Kai QIAN, Jian-ji REN, Zong-pu JIA, Tian-peng JIANG, Quan ZHANG
• Format: Journal Article
• Language: English
• Published: College of Computer Science and Technology, Henan Polytechnic University, Jiaozuo 454000, China%Beijing Anqi Zhilian Technology Co., Ltd., Beijing 100000, China%Department of Computer Science, Wayne State University, Detroit 48202, USA 2019
• Subjects: Linear threshold model; Opportunity network; Device-to-device communication; Content dissemination
• ISSN: 2095-9184; 2095-9230

TP393.0; Opportunity networks provide a chance to offload the tremendous cellular traffic generated by sharing popular content on mobile networks. Analyzing the content spread characteristics in real opportunity environments can discover important clues for traffic offloading decision making. However, relevant published work is very limited since it is not easy to collect data from real environments. In this study, we elaborate the analysis on the dataset collected from a real opportunity environment formed by the users of Xender, which is one of the leading mobile applications for content sharing. To discover content transmission characteristics, scale, speed, and type analyses are implemented on the dataset. The analysis results show that file transmission has obvious periodicity, that only a very small fraction of files spread widely, and that application files have much higher probability to be popular than other files. We also propose a solution to maximize file spread scales, which is very helpful for forecasting popular files. The experimental results verify the effectiveness and usefulness of our solution.