Font Size: a A A

Research On Key Issues For Trust Modeling In Mobile Distributed Environments

Posted on:2018-09-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z Q LiuFull Text:PDF
GTID:1368330542492879Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
With the continuous development of computer hardware and software as well as network communication technology,Internet of Things(Io T)has become the 3rd ware of information industry after computer and the Internet.In essence,Io T is a kind of distributed system which connects a large number of sensing devices to the Internet so as to achieve intelligent identification,monitoring and management.In Io T,owing to the huge size of network and the heterogeneous,autonomic and mobile characteristics of entities,the interactions among entities are of high uncertainty and uncontrollability.How to evaluate the trustworthiness of other entities and how to select the best interaction partner(s)when faced with lots of candidates are both difficult problems for traditional security technologies.Trust management provides a kind of novel solution to the security and reliability issues in mobile distributed environments.It not only enables each domain and each entity to build and maintain the trust relationships with others but also motivates them to work together.In recent years,a vast number of trust theories,models and systems have been proposed by researchers,and in which cross-domain trust evaluation,Location Based Service(LBS)and Vehicular Ad-hoc Network(VANET)have become hot research areas,and the lightweight and fast trust evaluation in highly mobile distributed environments(e.g.VANET)is the difficult point.This dissertation first analyzes the consistency of trust issue and being trusted issue,and then refines the self-certified trust theory and applies it to cross-domain trust evaluation,LBS and VANET.Moreover,this dissertation analyzes the consistency of trust issue and influence spreading issue,and then proposes the trust cascading theory and applies it to VANET.In detail,this dissertation mainly focuses on four key issues for trust modeling in mobile distributed environments,and the major achievements can be summarized as follows:1)We first pay attention to cross-domain trust evaluation.In mobile distributed environments,an entity may move across domains with great frequency.How to utilize the trust information in previous domains and quickly establish trust relationships with others in the current domain remains a challenging issue.The classic trust models do not support crossdomain and the existing cross-domain trust models require super nodes or trusted third par- ties.This work proposes a Lightweight Cross-domain Trust(LCT)model for mobile distributed environments in a fully-distributed manner.The trust certifications contain a variety of trust aspects with different interest preference weights,and they are collected and provided by trustees.Furthermore,three factor weights are comprehensively considered to ease the collusion attack.Finally,a distributed cross-domain scenario is deployed,and a series of simulations and analysis are conducted.The results demonstrate that LCT model obviously outperforms Bayesian Network(BN)model and Certificated Reputation(CR)model in the distributed cross-domain scenario.2)In addition,we focus on the trust modeling for LBS Recommendation(LBSR).In recent years,a vast number of trust models for LBSR have been proposed.These models are centralized in essence,and the trusted third party may collude with malicious service providers and cause the single-point failure problem.This work proposes a novel Fully-distributed Context-aware Trust(FCT)model for LBSR.In FCT model,the recommendation operations are conducted by service providers directly and the trusted third party is no longer required.In addition,FCT model supports the movements of service providers due to its self-certified characteristic.Moreover,four kinds of factor weights are comprehensively considered for easing the collusion attack and value imbalance attack.Finally,a distributed service recommendation scenario is deployed,and comprehensive simulations and analysis are conducted.The results show that FCT model significantly outperforms CR model in terms of the robustness and service recommendation performance.3)Moreover,we are concerned with entity-oriented trust evaluation in VANET.With the advances in automobile industry and wireless communication technology,VANET has attracted the attention of a large number of researchers.Trust management plays an important role in VANET,whereas it is still at the preliminary stage.This work proposes a novel Lightweight Self-Organized Trust(LSOT)model which includes both trust certificationbased and trust recommendation-based evaluations.Both the super nodes and trusted third parties are not needed in LSOT model.In trust certification-based evaluation,three factor weights are comprehensively considered to ease the collusion attack,while in trust recommendation-based evaluation,the testing interaction method is used to build and maintain the trust network and the Maximum Local Trust(MLT)algorithm is utilized to identify trustworthy recommenders.Finally,a distributed VANET scenario is deployed,and a series of simulations and analysis are conducted.The results demonstrate that LSOT model significantly outperforms Experience-Based Trust(EBT)model and LCT model in terms of evaluation performance and robustness.4)Furthermore,we work on data-oriented trust evaluation in VANET.Trust management is essential for VANET,whereas the existing trust models cannot totally conform to the characteristics of VANET.This work reveals the trust cascading phenomenon in VANET and proposes a novel Lightweight Trust Cascading(LTC)model which is built in a self-organized way and integrates entity-oriented and data-oriented evaluations.In entity-oriented evaluation,the trust relationships among nodes can be built and updated by leveraging the testing interaction method before the occurrence of event.When an event occurs,the entity-oriented trust values can be efficiently incorporated into data-oriented trust evaluation.Finally,a distributed VANET scenario is deployed based on a public dataset,and comprehensive simulations and analysis are conducted.The results illustrate that LTC model significantly outperforms Majority Voting(MV)model and Countering Information Oversampling(CIO)model in terms of efficiency of message spreading and robustness against malicious nodes.
Keywords/Search Tags:trust model, self-certified trust, trust cascading, cross-domain trust, location based service, service recommendation, vehicular ad-hoc network
PDF Full Text Request
Related items