Design of Distributed Medical Architecture for Regional Medical Image Based on HDFS

Abstract: Building a regional medical imaging collaboration platform is an effective way to balance medical resources, improve the level of diagnosis and treatment of primary hospitals, and reduce medical expenses. However, there are still huge challenges in the construction of regionalized imaging platforms in terms of technology and cost. This paper analyzes the advantages and disadvantages of traditional centralized storage and HDFS (Hadoop Distributed File System) distributed storage system, and designs a S-DICOM file format suitable for HDFS, as well as centralized storage (FC SAN) and distributed storage. (HDFS cluster) combined with a unified storage architecture, developed a set of SDFO (S-DICOM File Operator) middleware to provide a transparent storage access interface for the upper PACS application components. The test results show that this architecture can meet the rapid access and processing requirements of massive medical image data.

With the extensive clinical application of X-ray machines, CT, magnetic resonance and other large-scale imaging equipment, image examination has become one of the most important basis for clinical diagnosis. But expensive imaging equipment and repetitive image inspections have also become an important part of hospital and patient care spending. At the same time, the imaging diagnosis is difficult, the operation complexity is high, and the professionalism is strong. The grassroots hospitals are extremely lacking in excellent imaging diagnostic talents. The imbalance of medical equipment and talents is also an important reason for the current "difficult to see a doctor, expensive to see a doctor." The construction of a regionally integrated medical collaboration platform is an important way to balance medical resources, improve the level of diagnosis and treatment in primary hospitals, and achieve “ordered medical care”. The construction of regional medical imaging collaboration platform is an important part of regional medical collaboration, but the construction of regional medical imaging collaboration platform still has great challenges in technology and cost.

1 Challenges in building a regional medical imaging collaboration platform

Digital medical imaging technology currently has a mature international standard, namely DICOM 3.0. The PACS system constructed in accordance with its standards has gradually evolved from a single machine to a department to a whole hospital or region. At present, many large-scale top three hospitals in China have carried out PACS applications in the whole hospital, which has achieved no filming in hospitals. Regionalization of PACS systems will be the main research goal of the next phase of government health departments and medical institutions, but the construction of large regional medical imaging centers and collaboration platforms is currently facing enormous challenges.

1.1 High construction costs

The data volume of PACS is much larger than other medical systems such as HIS and LIS. The regional medical image data reaches hundreds of terabytes or even PB level, and the traditional storage architecture (such as FC SAN/iSCSI) is extremely expensive.

1.2 There is a bottleneck in transmission bandwidth

Even with high-performance FC SANs, its network bandwidth and processing power are difficult to meet the fast processing and transmission requirements of PB-level data.

1.3 Limited availability

Large hospital PACS systems often use the "online-nearline-offline" storage mode. Most of the offline data is stored in the tape library, and its usability is poor, and the data cannot be obtained in real time.

1.4 Lack of integrated application platform

The current medical image collaboration, such as remote imaging consultation, basically adopts a "peer-to-peer" mode, and lacks an integrated, cross-platform, highly available regional medical imaging collaborative application software. With the rapid development of cloud computing technology, it provides an effective way to build a low-cost, high-availability and high-performance regional medical imaging collaboration platform. Cloud computing is a new technology and operating model that Google has come up with. From the application scope, cloud computing can be divided into public cloud, private cloud and hybrid cloud. From the perspective of service models, cloud computing can be divided into IaaS, SaaS and PaaS. The regional medical imaging cloud computing platform belongs to the category of hybrid cloud. The subject we undertake is to study the transmission media provided by various medical institutions through high-speed metropolitan area networks, medical insurance networks, e-government extranets, and the Internet between medical hospitals. SaaS mode medical imaging collaboration application system, including Web DICOM terminal, image consultation, image referral, distance education, digital film storage and other services. High-performance, high-reliability mass image storage system will be the foundation and key of medical image cloud computing platform. This paper mainly introduces the design of a storage architecture based on Hadoop platform-based distributed storage and traditional centralized storage (FCSAN). And implementation.

2 Introduction to the Hadoop Platform

Hadoop is one of the most widely used open source distributed storage and computing platforms. It is an open source platform developed based on Google's GFS distributed file system and Map/Reduce distributed computing technology. Its design goal is to build high-capacity, high-performance, high-reliability distributed storage and distributed on common hardware platforms. Computing architecture. Hadoop has been widely used in Yahoo, Facebook, Amazon, Baidu, China Mobile and other companies. Yahoo, FaceBook and other companies have built large Hadoop application clusters of thousands to tens of thousands of ordinary servers. The amount of image data stored on FaceBook has now exceeded 1 PB or 1024 TB.

2.1 Characteristics and Applicability of Hadoop Clusters

The Hadoop HDFS distributed file system has the following characteristics: (1) It is very suitable for the storage and processing of massive data; (2) It has high scalability, and it can achieve linear growth of storage capacity and computing power by simply adding the number of servers; 3) High data redundancy. By default, each data is kept on backup on 3 servers; (4) It is suitable for “streaming” access, that is, one write, multiple read, and rarely modified after data is written. It is very suitable for the characteristics of medical image files; (5) In addition to data storage capabilities, Hadoop MapReduce distributed computing framework can also make full use of the computing resources of each server CPU, which is convenient for image fusion and image content retrieval based on massive medical image data. Data-intensive calculations such as 3D reconstruction.

Magnetic Bead Acid Nucleic Extraction Kit

In order to implement the requirements for the prevention and control of the new coronavirus pneumonia epidemic, all parts of the country are actively cooperating in the normalization of epidemic prevention and control. Each city has arranged multiple nucleic acid testing sampling points, and requires a negative nucleic acid test result within a few days. Public places are accessible, and the importance of nucleic acid testing for epidemic prevention and control is self-evident.
As of now, there are 13,000 medical and health institutions in the country that can carry out nucleic acid testing, and 153,000 professional and technical personnel are engaged in nucleic acid testing technology. The nucleic acid testing capacity has reached 57 million tubes per day. In order to help the prevention and control of the new crown epidemic, Yisheng Biology quickly launched the nucleic acid extraction reagent + 96-channel automatic nucleic acid extraction instrument. It only takes about 10 minutes to extract 96 nucleic acid samples at one time, and extract nucleic acid in a faster and more accurate way. Work becomes more efficient.
Magnetic Bead Swab Sample Viral DNA/RNA Extraction Kit can be used to isolate and purify high-quality viral DNA/RNA from various swabs and virus culture supernatants. By using magnetic beads with unique separation function and carefully optimized buffer system to efficiently capture the released nucleic acid, the extracted nucleic acid has high yield and good quality, which is suitable for various downstream application experiments, such as RT-qPCR, qPCR and other experiments.

Magnetic Bead Acid Nucleic Extraction Kit,Viral Dna Extraction Kit,Viral Rna Extraction Kit,Rapid Viral Rna Extraction Kit

Jilin Sinoscience Technology Co. LTD , https://www.jlgkscience.com