Measurement And Analysis Of Device-to-device Content Sharing In Mobile Social Networks

The heavy multimedia traffic produced by mobile users poses great challenges for the mobile network operators.Recently,many studies demonstrate that exploiting the device-to-device(D2D)content sharing in offline Mobile Social Networks is a promising solution to cellular data offloading and can solve the problem of mobile traffic explosion to some extent,which is also a key technique of 5G cellular networks.In this paper,we perform large-scale data measurement and multi-feature forecast concentrating on D2 D content sharing.Many relevant approaches are based on either unrealistic assumptions,or limited data analytics caused by small data size(e.g.hundreds of MSN users)or single dimensional feature(e.g.human mobility only),which severely restrict their applications in practice.To address these issues,this paper performs large-scale data analysis and prediction on the potential key nodes of sharing content in the near future.Specifically,(1)we build a big data process platform on Spark and analyze the behaviors of 30 million users(with 40 billion D2 D transmissions and 16 million content files)of Xender,a leading global D2 D sharing platform.(2)Several important features are studied,including performance basics,content properties,location relations,meeting dynamics,and social characteristics.As a proof-of-concept study of our analytics,we also predict D2 D sharing activities using machine learning methods based on the measurement results and the accurate rate is 60%.(3)We also explore the users' social groups and design the CPT algorithm for key nodes' selection referring to complex network theory,and then we are able to realize effective content propagation by making full use of key nodes' social impact.To sum up,we try to predict users' behaviors and key nodes in D2 D sharing networks based on the D2 D content sharing data,in order to realize cellular data offloading and solve the issue of mobile traffic explosion.