Data Storage,Access Control And Computation Resource Optimization For Blockchain Network

In recent years,there has been an explosive growth of data generated by mobile devices.The applications of machine learning,big data analytics and other techniques have provided a rich experience for users,while the resources-limited local terminals cannot meet the demands for storing all data.Cloud storage services can provide efficient support for mobile terminal data storage,analysis and access management.However,in the centralized cloud storage platform,user data are still under threat from the illegal access and privacy leakage.The application of blockchain provides a new perspective for solving the problem of "trust"in the cloud storage.However,there are still some problems in existing blockchain-based cloud data management,such as imbalanced storage,unauthorized data access control and computing resource optimization for mobile terminals.These problems are produced mainly by uneven user distribution,heterogeneous terminals and blockchain computing-intensive tasks.This thesis focuses on the problems of storage imbalance and unauthorized access of user data in the cloud storage network.Meanwhile,efficient offloading approaches for computing-intensive tasks are also investigated in the blockchain network.The main contents and contributions of this thesis are as follows.(1)The Gini coefficient is introduced to measure the balance of storage for the problem of imbalanced storage of outsourced data in the blockchain network.The problem is modeled and formalized,and its NP-hardness is proved.An efficient heuristic algorithm,a customized genetic algorithm and a Tabu search algorithm are designed to achieve storage balance and improve network availability and stability.Experimental results show that,the proposed heuristic algorithm performs better in terms of accuracy,computation overhead and storage overhead,compared with the existing works.When the initial solution of the Tabu search is generated by the proposed heuristic algorithm,the Tabu search algorithm can obtain the best performance in terms of accuracy.Meanwhile,the customized genetic algorithm can obtain the best performance,in comparison to existing work in terms of accuracy.(2)An efficient blockchain-based cloud computing access control scheme is proposed for the problem of unauthorized access to cloud data.The data owner maintains an access matrix to describe the access strategy,and then stores the access matrix and the public keys of all nodes in the blockchain,to ensure its consistency and integrity.The symmetric key and asymmetric key are adopted to encrypt data in our scheme which can reduce the key generation overhead of data owner.Data owner efficiently uses the symmetric key to encrypt the shared files,and uses the public key of the authorized user to encrypt the symmetric key in parallel.Hence,the authorized users obtain a symmetric key by their local private keys,and data owner does not need to maintain multiple encryption keys.Security analysis shows that our scheme can effectively prevent unauthorized access and collusion attacks on outsourced files.Experimental results show that the computation overhead of our scheme is 25.37%,45.46%and 36.44%lower than that of the three existing schemes,respectively.The communication overhead of our scheme is 17.16%lower than that of an existing scheme The proposed scheme is of higher security than the other two existing schemes,although the communication overhead increase by 5.88%and 39.05%,respectively.Moreover,the storage overhead of our scheme is 59.36%,20.25%and 61.88%lower than that of existing three schemes,respectively(3)A smart-contract-based long-term double auction is proposed to optimize the mobile device computing resources in the blockchain network.To optimize allocation resources,we transform this problem to the multi-choice knapsack problem which is of NP-hard.A long-term double auction algorithm is proposed to offload subtasks from mobile devices to heterogeneous edge servers,to improve the efficiency and long-term performance.The long-term performance of the algorithm enables users to participate in the offloading until they leave,when users have an adequate budget and computational demand.The algorithm can increase the total utility of buyers and sellers and it also satisfies computational efficiency,individual rationality,budget balance and truthfulness.Moreover,a smart contract scheme is proposed to ensure the automatic execution of algorithm in a safe and controllable environment.Experimental results show that,the utility and utilization ratio of our algorithm are increased by 130.55%and 138.64%,respectively,.compared with the existing auction algorithm.