Research On Massive Medical Devices Data Storage Based On Relation Database And NoSQL

In the era of big data,e-commerce,social networking sites and other websites generate huge amounts of data every day.Prior to using the relational database for data storage,query,the face of massive amounts of data,relational databases seem powerless.A medical device management system chose the relational database to storage data at the beginning of the design.As the system grows in use,the amount of accumulated data grows larger,making it difficult to handle massive amounts of data due to the inability of relational databases to scale with the simple addition of devices.And NoSQL database data model is simple,it can be linearly extended,etc.,which has great advantages in dealing with big data problems.Therefore,how to combine the advantages of the two and apply them to the system at a small cost is of great significance.Based on the business of medical device management system,combined with the advantages of relational database and NoSQL,this paper proposes a system architecture for storing data according to the type.We use relational databases to store dealer-related data and NoSQL to store medical device-related data.Through such a design,the data storage problem of mass medical devices can be solved,at the same time,the safety and transactional performance of the customer data can be guaranteed.In this paper,a hybrid SQL query system is designed and implemented.Through the analysis and optimization of SQL statements,it is distributed to the corresponding system according to different data sources for query,and the final result is obtained according to the query data.Without changing the structure of the system,making the original system can support the combination of relational database and NoSQL system architecture.This paper designed and implemented an SSD-based hybrid cache architecture.Through the use of SSDs,the hit ratio of RowCache is increased and the read performance of Cassandra is improved.The entire system architecture is divided into the memory layer and SSD layer,and each layer is divided into different regions.Through the hot data selection algorithm,the data in the hybrid cache is stored in different regions according to the heat level and transferred according to the corresponding data transfer mechanism to exert the respective advantages of the memory and the SSD.The entire system was designed and implemented on the basis of the cache interface provided by Cassandra and will not affect the normal use of Cassandra.Tested,the system is operating in good condition.And the performance has improved compared to the original system.