In recent years, artificial intelligence ( AI) has played an increasingly important role in the medical field. AI based on big data analysis and deep learning algorithms has become the core driving force for the future development of the medical industry, but the biggest obstacle to its development is incompleteness and inaccuracy of medical data. The main reason is that medical big data is difficult to achieve sharing. This paper studies and analyzes the problems of medical big data sharing, proposes a medical data sharing platform based on permissioned blockchains (BCs), which uses dual-BCs architecture (Account BC and transaction BC), Concurrent Byzantine Fault Tolerance (CBFT) consensus mechanism, encryption technology and smart contract technology. Account BC (ABC) is used to store user data hash after encryption and transaction BC (TBC) is used to process computing tasks. The use of encryption technology ensures that users have full autonomy in their own medical data. Smart contract technology is used to enable users to set up different access permissions of medical data. A TBC can be used by multiple ABCs and TBC does not save users' medical data, but only obtains users' medical data from ABCs when needs computation. When the computing task is completed, the TBC deletes all users' medical data and packages the operation record to save into ABCs, thereby it realizes data sharing while protecting users' privacy. In addition, this paper compares the functions with other medical data systems, analyzes and compares their advantages and disadvantages. Finally, the conclusions and future works of the medical data sharing platform based on BC technology are prospected.
