Key Technologies Study On Medical Clond For Big Medical Data Processing

Traditional hospital information system architecture has its disadvantages and is not able to deal with the people’s more and more vigorous health service demand under the new medical reform policy. With the advent of the era of big data, it is also unable to face the opportunities and challenges of big medical data.Introducing the core technologies and core concepts of cloud computing and big data into hospital informatization constructions, can improve the flexibility and efficiency of medical informatization. On one hand, it helps hospital information systems to realize overall deployment, intensive management, on-demand configuration, and satisfy its large-scale use with scalability, reliability and on-demand services. On the other hand, it offers the abilities of elastic storage, centralized management and effective utilization for big medical data.This thesis designed a three-layer application model of medical cloud and big data. Firstly, we designed and implemented a private hospital cloud based on virtualization technology to blend the existing hospital information systems into cloud architecture. Secondly, we designed and implemented a community hospital cloud based on virtual desktop infrastructure to a new medical resources delivery way as "cloud service", namely, hospital information system as a service. Thirdly, we designed and developed a big medical data processing system based on Hadoop to effectively use the big medical data. Combined with the application mode of hospital informatization, we designed a three-layer application model of medical cloud and big data based on core technologies and concepts of cloud computing and big data to realize the adoption of medical cloud and big data step by step.In order to change the status quo of "chimney type" hospital information system architecture, we designed a hospital information platform based on cloud computing to bulid a hospital private cloud with the characteristics of virtualization technology. On the basis of hospital private cloud, we used VDI to build a medical community cloud, which pushes hospital information systems as cloud services to more medical institutions. According to the principle of "centralized computing, distributed display", Various authorized medical institutions accessed medical community cloud for medical information system as SaaS, so it abandoned the original own hospital IT construction to simplify the hospital information platform architecture and save the cost.Medical community cloud highlights the heterogeneity and complexity and differences between hospital information systems of different medical institutions. The traditional integration way may make the integration platform more complicated and difficult to maintain, and also cannot fulfil large-scale medical data interaction in cloud environment. Combining with the service mode of medical community cloud, we used some medical information standards to design a flat hospital information system integration platform based on cloud computing. In order to solve the problem of big medical data in medical cloud, we designed and developed a big medical data processing system based on Hadoop to use cloud-based distributed computing to solve the problems of big medical data collection, storage and analysis.The research of this paper is a prototype for cloud computing and big data technologies being applied to hospital informatization. Virtualization technology and distributed computing technology are both key technologies of cloud computing but two distinct technical direction. This paper reasonably used these two technologies in different stages of medical cloud construction. The prototype was deployed in some hospital of Nanjing Military Command, and the feasibility of medical cloud was empirically proved. In the practical application and system evaluation research of medical cloud, cloud computing is proved to have advantages on technical architecture and economic benefit of hospital informatization, and makes up for the shortages of the traditional data processing methods in big medical data processing.