基于HDFS的区域医学影像分布式存储架构设计-南方医科大学学报.PDFVIP

2011;31(3) 南方医科大学学报（J South Med Univ ） ·495 · ·基础研究· 基于HDFS 的区域医学影像分布式存储架构设计 1 2 1 李彭军 ，陈光杰 ，郭文明 1 2 南方医科大学 网络中心，生物医学工程学院，广东 广州 510515 摘要：构建区域医学影像协作平台是均衡医疗资源、提高基层医院诊疗水平、降低医疗费用的有效途径，但是构建区域化 影像平台在技术和成本上还存在巨大的挑战。本文详细分析了传统集中式存储和HDFS （Hadoop Distributed File System）分布式存储系统的优缺点，设计了一种适合HDFS 特点的S-DICOM文件格式，以及集中式存储（FC SAN）和分布 式存储（HDFS 集群）结合的统一存储架构，开发了一套SDFO（S-DICOM File Operator）中间件，为上层的PACS 应用组件 提供透明的存储访问接口。测试结果表明此架构可以满足海量医学影像资料的快速存取和处理需求。 关键词：图像归档与传输系统；云计算；软件即服务；Hadoop 分布式文件系统 中图分类号：TP391 文献标志码：A 文章编号：1673 －4254（2011）03-0495-04 A distributed storage architecture for regional medical image sharing and cooperation based on HDFS 1 2 1 LI Peng-jun , CHEN Guang-jie , GUO Wen-ming 1Network Center, 2 School of Biomedical Engineering, Southern Medical University, Guangzhou 510515, China Abstract: Given the importance of regional centers for medical image sharing and cooperation is important for resource balancing, healthcare service enhancement and medical expense reduction, building such regional medical image sharing and cooperation centers faces huge challenges. In this paper we analyze the advantages and weakness of two storage architectures, and designed a hybrid storage architecture combining FC SAN and Hadoop HDFS. A HDFS suitable medical image file format, called S-DICOM, and a set of S-DICOM operating middleware, SDFO (S-DICOM File Operator), was developed. The results of performance testing indicated that this hybrid storage a