Research On Performance And Security Of Electronic Health Record System Based On Blockchain

The traditional electronic health medical record system adopts the concept of centralized construction,with the hospital as the center,and the patient’s medical data is independently maintained by each hospital.Such a centralized storage solution causes patients to form medical data islands,which brings inconvenience in daily behavior such as referral.Generally speaking,traditional systems have many problems in medical data sharing,data security,and privacy protection.With the vigorous development of blockchain technology in the past decade,its decentralization and non-tampering characteristics have brought new ideas to how to solve the problems of traditional electronic health medical record s ystem,such as difficulty sharing,poor security,and inadequate privacy protection.This paper proposes a new electronic health medical record system model based on blockchain technology,and initially realizes the design and implementation of the syste m prototype.Considering that the current performance bottleneck of blockchain technology is the key factor hindering the new system to the actual application,firstly,this paper analyzes the business process related to the medical blockchain network based on the system prototype,and then performs fine-grained performance analysis and bottleneck positioning based on then test tool.Finally,the key performance bottlenecks are located in the consensus phase and the block synchronization phase.Aiming at the performance problems in the consensus stage,this paper proposes a scheme based on VRFs for Byzantine consensus algorithm.From the overall idea,algorithm design,algorithm implementation,comparative test analysis,the performance of the new algorithm is gradually verified.The experimental analysis in the final stage shows that the average delay of the new consensus algorithm is reduced to 30 ms,the overall Invoke throughput is increased to 680 tps.Another advantage about performance is that the strategy of electing the consensus node avoids the problem of the rapid decline of the performance of the traditional PBFT consensus algorithm in large-scale network scenarios.In addition to performance optimization,the algorithm utilizes the random verifiable feature of VRFs to make the consensus network hard to be identified and then enhances consensus security.Aiming at the performance problems in the block synchronization phase,this paper proposes a storage optimization scheme based on improved LSM-Tree.Firstly,the key-value pair separation storage is used to reduce the LSM-Tree scale to solve the write amplification problem,and then the multi-thread parallel execution of the Compact operation is used to improve the data compression efficiency.From the overall idea,algorithm design,algorithm implementation and comparison test analysis,this paper verifies the performance optimization of the new algorithm in the block synchronization.The experimental analysis in the final stage shows that the optimized system has an Invoke throughput of 1400 tps and a Q uery throughput of 2200 tps,which is 2-3 times different performance improvement than the system prototype.This paper starts with the obvious performance problems in the consensus stage and the block synchronization stage,and combines the characteristics of the business data of the prototype of the system,and proposes the corresponding optimization plan,which not only improves the overall throughput rate of the new electronic health record s ystem.And taking into account the security issues in the consensus stage,making the medical data management system based on blockchain technology more practical and valuable.