Research On Security Schemes Based On Blockchain And Trust Management,and Offloading Strategies In Internet Of Medical Things

With the rapid development of the Internet of Things(Io T),the Internet of Medical Things(Io MT),which is a typical application of Io T,has received extensive attention from all sectors of the society.While achieving great progress,Io MT also faces great challenges,including requirements of low latency,secure,trusted and reliable data transmission,storage and processing.These problems,if not well solved,will greatly restrict the development of Io MT.In order to address the above challenges of Io MT,this dissertation mainly does two aspects of work: Firstly,advanced security measures including trust management,blockchain,and security protocols are proposed to resist network insider and outsider attacks and ensure security of the system.Secondly,considering the specific application scenarios and requirements of Io MT,effective computation offloading strategies are put forward,with low-latency as the main optimization objective and multiple critical performance factors such as security,reliability,credibility,and energy-consumption jointly taken into account.The paper mainly makes the following contributions:1)In order to resist network insider attacks from malicious network devices,this dissertation first suggests an Software Defined Networking(SDN)-and fog-based network framework.On this basis,a trust management policy which consists of a multi-dimensional trust evaluation mechanism along with an objective-and subjective-based trust evidence fusion method is proposed.The proposed trust evaluation method can obtain trust evidences from multiple views to more comprehensively and legitimately evaluate the trustworthiness of the trustee.Then,a trust fusion algorithm called JSDM-AHP which combines Jensen-Shannon Divergence Measure(JSDM)and Analytic Hierarchy Process(AHP)is recommended.By fully considering the objective credibility and the subjective importance of the trust evidences,the weight of each trust indicator is more reasonably designed using the proposed fusion method.The proposed trust management strategy has advantages of high reliability and low complexity.2)In order to defend against network insider attacks from malicious service providers(SPs),a Stacking Ensemble-learning Method for Trustworthiness Prediction(SEM-TP)of fog services is proposed.The proposed algorithm takes Quality of Service(Qo S)indexes of fog services as trust attributes for trustworthiness prediction,and by integrating heterogenous individual machine learning classifiers,the divergence and accuracy of the proposed ensemble learning method are significantly enhanced.Besides,in order to improve the robustness of the prediction model as well as reduce time cost,a hybrid and voting-based feature selection strategy is proposed,which combines with diverse feature selection algorithms and uses voting mechanism to select the most significant feature subset for model training.Compared with individual classifiers and other compared stacking-based prediction methods,the proposed SEM-TP achieves better prediction performance.3)In order to better improve the security of Io MT systems,this dissertation also proposes a system-level security approach based on hashchain and blockchain,which can not only guarantee the secure transmission of data in Wireless Body Area Networks(WBANs),but also ensure the secure storage of data outside WBANs.First,to achieve secure transmission of human health data,a lightweight and secure data transmission protocol based on hash chain is proposed for intra-WBAN,which can protect the confidentiality,integrity,and privacy of human physiological data of WBANs,and can effectively resist various network security attacks.Next,in order to protect the secure storage of human health data for beyond-WBAN,a two-layer blockchain security mechanism which consists of a control blockchain and a data blockchain is proposed.The two types of blockchain can guarantee the secure storage of data in different network layers,respectively.Besides,an effective blockchain sharding strategy is proposed,which can well alleviate the inherent time-inefficiency problem of blockchain.4)Considering the different networking architecture of Io MT based on SDN and blockchain,effective computation offloading policies are proposed respectively to address the challenges of resource constraints of Io MT devices and the time-sensitivity of Io MT applications.On the one hand,for the SDN-based Io MT scenario,a multi-objective computation offloading algorithm with emergency handling capacity based on the network architecture of SDN and fog computing is proposed,which makes decision on the target fog node for task processing by taking into full consideration of performance metrics such as time latency and nodes’ credibility.The proposed method has the characteristics of low delay,high security and reliability,and can meet the requirement of processing emergent tasks with lowlatency in Io MT.On the other hand,for the blockchain-based Io MT,a distributed computation offloading strategy based on the network framework of blockchain and fog computing is proposed,which takes full consideration of the security requirements of Io MT systems.By combining trust mechanism with blockchain,the proposed approach can well resist network insider and outsider attacks,and also the inherent time-inefficiency problem can well be alleviated.Moreover,a utility-based decision-making algorithm is proposed,and on the basis of maximizing the utility of the system,multiple performance metrices including time delay,energy consumption,trust factor,and reliability are jointly taken into account in the decision-making process of the optimal fog node for task processing.The biggest innovation of the work lies in that,all the offloading requests from WBAN devices are published as blockchain transactions,and the rest workflows are securely and automatically executed by the well-designed blockchain smart contracts,requiring no intervention of any third party,thereby ensuring security of the system.This dissertation provides a comprehensive and systematic solution to the challenges of Io MT in security,trust,low latency,and reliability under different networking architectures and application scenarios,which has certain theoretical value and practical significance.