Faced with the increasing volumes of discrete and multi-source heterogeneous healthcare big data,traditional integration platform architectures are challenged by their limited data processing capacity,low efficiency,poor flexibility,and difficulties in storing and analyzing unstructured data.To address these issues,a method for constructing a healthcare big data integration platform is pro-posed,which utilizes a hybrid architecture based on Hadoop and Missively Parallel Processing(MPP)databases.This approach combines the technical advantages of both architectures.The MPP relational architecture is utilized for performing logical processing scenarios involving complex queries,multi-table associations,and self-service analysis of structured data.On the other hand,the Hadoop distrib-uted architecture is employed for parallel computation of large-scale unstructured data.For the con-struction of this integrated platform,a strategy is followed that involves logical stratification and physi-cal partitioning to achieve centralized collection,classified storage,and effective integration of health and medical big data.This ensures high-quality governance over the data while improving processing efficiency,providing an efficient support platform for clinical practice as well as scientific research.
health medical big dataHadoopMPP databasemixed architectureintegrated plat-form
国家科技期刊平台
NSTL
万方数据
