Research On The Virus Percolation In Social Internet Of Things In The Presence Of Search Engine

Social Internet of Things(SIoT)integrates the social network with the Internet of Things(IoT)and has become a hot research issue for its potential to support novel IoT applications and networking services in more effective and efficient ways.The search engine,which is primarily designed for social network users to acquire interested information,also plays an important role on SIoT.There even exists search engines specially designed for SIoT,such as Shodan and Thingful.With this type of dedicated search engine,network users can easily locate most smart devices in SIoT.Specifically,a search engine assists the information dissemination,i.e.,enabling users(both humans and things)to access interested objects(both humans and things)with keywords-searching and transferring contents from the source directly to potential interested users.Accompanying such processes,the SIoT evolves as new links emerge between users and their interested objects.Seen from the information dissemination process,the search engine is also a tool for spreading viruses.Recent years have witnessed a rapid development of viruses,and the wide variety of security threats caused by viruses heighten the need for studying virus propagation.Studying the relationship between the virus propagation process and the search engine is vital.In this thesis,we aim to quantitatively characterize how a search engine influences the social relationship in the IoT and quantitatively discuss the characteristics of the virus propagation process in the presence of a search engine.First,it is pointed out that the search engine serves as the medium between the social network and the IoT,and then a Search Social Internet of Things(SSIoT)model based on classical bipartite graph theory is proposed to describe SIoT evolution.Second,six performance metrics are adopted,namely,degree distribution,network diameter,average distance,network density,the giant components,and user betweenness.Theoretically,it is proved that the degree distribution follows an intensified power-law,the network diameter and the average distance shrink,the network density,the giant components,and the user betweenness are greater in SSIoTs.Third,an iterative process of a cascade of virus propagation based on the classical percolation theory is modelled and it is revealed that increasing the searching ratio would lead to a change from a second to a first order percolation phase transition in the virus propagation process.Taking the Poisson distribution network as an example,the crucial searching ratio that leads to the transition from a second to a first order is discussed.This method is applicable to other networks with different distributions.Fourth,it is proved that the search engine accelerates virus propagation and expands the virus propagation area by increasing the infection density.Finally,the effects of the search engine on the resilience of SIoT are quantitatively characterized by two metrics: the crucial initial infection ratio(the jump point)and the functional node density.Based on real-world data sets(i.e.,DBLP,Facebook,Weibo,Slashdot,and P2P),our theoretical findings are verified.Experiment results well shown that the search engine promotes the evolution of SIoT and brings new structural characteristics.It is further proved that the search engine makes the SIoT more vulnerable to virus infection and the resilience of network decreases.