| With the deepening of people’s attention to data and the improvement of sensors and Internet of things technology,a massive of data is produced by human beings.Recently,because of the explosive growth of data scale,data security accidents have occurred frequently.Specifically,the behaviors of tampering,destruction,malicious leaking on data were reported almost every week.The behavior of maliciously tampering with some sensitive data like economic data may cause serious economic losses.According to the above description,the security vulnerabilities of traditional data storage system are emerging,and a single point of failure problem of large-scale distributed storage systems also makes the stability of the system is difficult to guarantee.With the advancement of technology,blockchain was proposed and has attracted increasing attention from industry and academia.The inherent attributes like decentralization and tamper-proof of blockchain technology have brought tremendous changes to some traditional business models.As a decentralized technical solution,it has been widely applied in many financial fields.Based on this,the hope that introducing blockchain technology to storage systems and realize the progress of data storage system is growing.The currently popular consensus algorithm is proof-of-work(Po W),which can be applied to a very large scale blockchain node network.However,its waste of computation and low response speed makes it unable to be directly applied to basic application systems requiring fast response.While the traditional distributed consistency algorithm of practical byzantine fault toll(PBFT)can solve the problem of serious resource waste exist in the Po W consensus algorithm.But it still has the problems of high communication complexity,poor performance of large-scale network,and large number of voting signature messages.These disadvantages make the algorithm difficult to satisfy the production requirements of high performance for some applications.Based on the above problems,this paper carries out the following research work:1.Propose an improved Byzantine fault-tolerant(WBFT)algorithm based on light and heavy client nodes.This method improves the PBFT consensus process and reduces the communication complexity of algorithm fromO(n~2)to O(n).We implement the WBFT consensus algorithm with Go language and the experiment results demonstrate that the average transaction throughput of algorithm is improved by 10%.2.Design and implement a trusted data storage system based on WBFT algorithm.By testing and experimenting this system,we demonstrate that the system can guarantee the security and reliability of stored data in the case of some malicious nodes. |