Design And Implementation Of ECG Data Acquisition And Storage System Based On Hadoop

With the continuous advancement of information technology and the rapid development of computer technology,the era of "big data" in medical institutions has arrived.However,due to the single data structure,small storage capacity,slow calculation speed and poor fault-tolerant recovery capability of the traditional storage mode,the storage and calculation of big data become a bottleneck,which seriously restricts the application of big data.The emergence of Hadoop provides ideas for solving this problem.How to use Hadoop to build a big data platform for medical institutions has become a hot research direction for in-depth reform of medical institutions.This thesis takes the practical application of Hadoop big data in medical institutions as the research direction,designs and implements an ECG data acquisition and storage system,and focuses on solving the problem of collection and storage of ECG big data.The main research work of this thesis is as follows:(1)Determine the basic requirements of ECG data acquisition and storage system,and design the overall architecture of system.Aiming at the requirements of real-time interaction and mass storage of ECG data acquisition and storage system,a hybrid architecture scheme combining MYSQL database.(2)The detailed design of two core functional modules of data acquisition and data storage in ECG data acquisition and storage system is given,including the data model and interface scheme of data interaction.Combined with the application characteristics of ECG data,a kind of application is proposed.Based on the HBase-based storage scheme and the low performance of HBase in non-primary key query,a second-level index implementation scheme based on Coprocessor is proposed.After verifying the establishment of the secondary index,the query performance of non-primary key columns is improved by more than 80%.(3)Completed the defect analysis and design of ECG data acquisition and storage system.Aiming at the single point of failure of ECG data acquisition and storage system,a "hot backup" solution based on Backup NameNode is proposed.For the high latency problem of ECG data acquisition and storage system,the default storage policy of Hadoop is analyzed.The limitation is to propose an improved copy placement strategy that reduces the write latency by approximately 41.2%.(4)Based on the above research,the prototype of ECG data acquisition and storage system was realized,and the system function and performance test were completed.The experimental results show that the ECG data acquisition and storage system designed and implemented in this paper runs reliably.Performance in terms of storage capacity,data access speed,and concurrent ability have met design goals to meet the needs of real-world applications.