Research On A Cloud-Based Trust Evaluation Scheme Using For Vehicular Social Network Environments

Vehicular Social Networks(VSNs)use vehicular terminals and mobile phones as their main carriers for data interaction.Meanwhile,VSNs take advantages of advanced network access technologies and service technical supports to realize intelligent collaboration as well as interaction among humans,vehicles,Road Side Units(RSUs)and cloud services.Along with the innovation of information technologies as well as the intellectualization,networking,and informatization of automobile,there also exist some security threats which hinder the wide development of VSNs.On one hand,although some new-emerging technologies such as cloud services and abundant on-board applications have promoted the large-scale data sharing and provided diverse information services,the large amount of data from multiple sources put a heavy burden to the cloud,which greatly affects the VSNs network performance.On the other hand,VSNs also exist security issue,such as privacy security and data reliability.While the traditional Public Key Infrastructure(PKI)is effective to guarantee the integrity and confidentiality of data transmitted with digital signature and public-key cryptography;however,it cannot distinguish the data reliability,and extra cost may be incurred.As an alternative,trust evaluation scheme can be used to solve these problemsTargeting the network performance and information security issues,in this thesis,we comprehensively analyze and summarize the differences as well as difficulties existing in the present schemes;while aiming to investigate an effective and reliable trust evaluation scheme for VSNs.Besides improving the VSNs network performance,low-quality messages can be effectively identified which further ensures that more reliable data are shared and transmitted Specifically,our main contributions are list as follows(1)Based on the concept of trust,we propose a reliable trust management framework for VSNs,which aims to solve the data collection,storage and management issues during the vehicular data sharing as well as interaction.The proposed framework is a three-layer framework,including the Global Trust Management(GTM),Domain Trust Management(DTM)and Vehicular Trust Monitor(VTM).As to each layer,it is responsible for the management and storage of corresponding vehicular data.While it clearly defines the interlinked and mutually support among three levels in the framework,and the reliability of proposed framework is also analyzed in the thesis(2)A trust evaluation algorithm based on trust management framework is proposed to solve the problem of information security in VSNs.This algorithm comprehensively considers the influence of various nodes in VSNs,depending on the direct/indirect connections and intra-domain/out-of-domain connections between nodes,the trust values of nodes will be considered according to different weights in the evaluation process.Specifically,the algorithm is divided into five sub-modules:neighbor trust assessment,friend trust assessment,RSU trust assessment,historical trust assessment and global trust assessment;depending on the influence of different modules,the global trust evaluation will comprehensively combine neighbor trust,friend trust,RSU trust and history trust according to different weights.Finally,we give the security analysis of proposed scheme,and an experimental scenario is constructed by Network Simulator 2(NS2),so as to verify the effectiveness and feasibility of proposed scheme(3)This thesis proposes a VSNs architecture based on cloud services,which can effectively deploy trust evaluation scheme and improve the network performance of VSNs.The proposed architecture is a three-layer cloud network architecture,including central cloud,RSU cloud and vehicular cloud.Each of them has its own functions,and they are interrelated.In addition,the Vehicular Virtual Machine(VVM)is a transient cloud,which is built on the physical resources of related vehicles.Since VVM has both computational and memory power,it will be used to assist the RSU cloud in completing trust evaluations.Finally,the simulation environment of vehicular network was built with NS2,and the effectiveness and feasibility of proposed scheme were verified based on the evaluation of network performance parameters.