Spark-based Content Propagation Maximization In Large-scale Mobile Social Networks

With the increasing number of smart mobile equipments and the developing of multimedia services for mobile devices,traffic explosion has become a big challenge for the Mobile Network Operators(MNOs),while it also brings worse experiences for users.As one of the key techniques of 5G,the Device-to-Device(D2D)is promising in offloading traffic and alleviating the traffic explosion phenomenon.The analysis of mobile social networks formed by users' information exchanging through D2 D is essential to further encourage users to use D2 D for content sharing.Mobile social networks based on D2 D is more complex than the online social networks for its multiple dimensions and spatio-temporal features.Most classic content propagation models for online social networks are not suitable in the D2 D scenario.Besides,the existing researches on D2 D are mostly simulation on small scale data sets based on unconsolidated hypotheses and measurements of data sets,which lack models on realistic large-scale mobile social networks,confining their APPlications in life.In this paper,we firstly make comprehensive and large-scale measurements on 3.56 TB of real data sets related to Device-to-Device(D2D)content sharing activities,which are traces from a popular D2 D sharing APPlication(APP).The mobile social networks generated by the offline content deliveries between users are presented and analyzed from time series,contents,location,social network structures and propagation patterns.Focusing on the seeding users' selection problem in social networks,we propose algorithms of weighted SeedRanks(SRs)to select the seeding users with accuracy.The algorithms are adapted to the parallel computing platform of Apache Spark.The results of the recurrent experiments on the large-scale D2 D data sets prove the efficiency of our algorithms outperforming the PageRank and Greedy algorithms on the covered users' number,as well as the effectiveness of enlarging the propagation range of content in D2 D mobile social networks.