Research On Blockchain-Based Medical Data Management Methods

The current centralized management model of medical institutions requires high trust costs.As a trust mechanism,blockchain can establish decentralized trust among entities and provide a good solution for medical data management.On the one hand,blockchain can use its tamperproof properties to record proof of data integrity and provide reliable storage of data evidence.On the other hand,the decentralized nature of blockchain,combined with cryptographic techniques,prevents single point risk and enables secure data exchange.However,both of these have shortcomings:the generic hash function is used by blockchain deposition,which is inefficient when dealing with large files in medical data;secondly,regarding node selection in multi-party collaboration tasks,the available solutions fail to sufficiently consider the difference in capabilities among nodes,resulting in a poor user experience.In response to the above issues,this thesis investigates blockchain-based medical data management methods and optimizes data evidence deposition and data sharing,respectively.The specific contents of the research are as follows:(1)To address the problem of inefficiency when processing large files in medical data evidence deposition scenarios,this thesis proposes a parallel hash deposition method based on an adaptive strategy.This method splits individual computational tasks and merges each subtree task through a tree structure.Moreover,each subtask is executed in parallel by multiple cores,which improves the efficiency of file processing.In addition,to generalize the acceleration effect,an adaptive strategy is proposed that can dynamically allocate computational parallelism according to file size.It can be adapted to generalized deposition scenarios and provides better acceleration for files of all sizes.Experimental results show that this method has excellent generalizability for accelerating data of varying sizes on a sufficient number of processor cores and achieves a significant improvement in processing efficiency for large files.(2)To address the problem of poor user experience caused by unreasonable node selection in medical data sharing scenarios,this thesis proposes a sortition selection method based on quality of service assessment.The method integrates factors affecting user experience,i.e.,service availability,timeliness and reliability,and combines individual preferences with collective feedback to dynamically assess node service quality.To ensure fair and verifiable selection,this thesis designs a sortition scheme based on the verifiable random function and assigns selection probabilities based on evaluation that match the capabilities of nodes.The experimental results show that the method improves the overall user experience compared with existing solutions in a diversified node environment,and also achieves a reasonable task allocation among nodes.